Receptor Subtype-Induced Targeting and Subtype-Specific Internalization of Human α₂-Adrenoceptors in PC12 Cells

Abstract

The three α₂-adrenergic receptor subtypes have distinct tissue distributions, desensitization properties, and, in some cell types, subtype-specific subcellular localization and trafficking properties. The subtypes also differ in their neuronal physiology. Therefore, we have investigated the localization and targeting of human α₂-adrenoceptors (α₂-AR) in PC12 cells, which were transfected to express the α₂-AR subtypes A, B, and C. Inspection of the receptors by indirect immunofluorescence and confocal microscopy showed that α_2A-AR were mainly targeted to the tips of the neurites, α_2B-AR were evenly distributed in the plasma membrane, and α_2C-AR were mostly located in an intracellular perinuclear compartment. After agonist treatment, α_2A- and α_2B-AR were internalized into partly overlapping populations of intracellular vesicles. Receptor subtype-specific changes in PC12 cell morphology were also discovered: expression of α_2A-AR, but not of α_2B- or α_2C-AR, induced differentiation-like changes in cells not treated with NGF. Also α_2B-AR were targeted to the tips of neurites when they were coexpressed in the same cells with α_2A-AR, indicating that the targeting of receptors to the tips of neurites is a consequence of a change in PC12 cell membrane protein trafficking that the α_2A-subtype induces. The marked agonist-induced internalization of α_2A-AR observed in both nondifferentiated and differentiated PC12 cells contrasts with earlier results from non-neuronal cells and points out the importance of the cellular environment for receptor endocytosis and trafficking. The targeting of α_2A-AR to nerve terminals in PC12 cells is in line with the putative physiological role of this receptor subtype as a presynaptic autoreceptor.

Three genes encoding adrenergic α₂-receptors (α₂-AR) have been cloned (α_2A/D, α_2B, and α_2C) in human and rodents (Kobilka et al., 1987; Regan et al., 1988; Lomasney et al., 1990; Zeng et al., 1990; Lanier et al., 1991; Chruscinski et al., 1992;Link et al., 1992). The α₂-AR are G-protein-coupled receptors and mediate effects of endogenous catecholamines and many therapeutic drugs via G-proteins to a variety of effectors, including adenylyl cyclases and ion channels (MacDonald et al., 1997). The three α₂-AR subtypes have quite similar ligand binding properties but different desensitization properties, as well as distinct subcellular and tissue distributions.

In the rodent brain, α_2A-AR expression is widely distributed in noradrenergic projection areas but is also abundant in the locus ceruleus and other noradrenergic nuclei, in line with the presynaptic autoreceptor function of this subtype (Nicholas et al., 1993; Scheinin et al., 1994; Talley et al., 1996;Wang et al., 1996; MacDonald et al., 1997). α_2A-AR are also present in the medulla and spinal cord in areas involved in control of autonomic functions and nociception (Rosin et al., 1993; Nicholas et al., 1996). In studies on genetically engineered mice, the α_2A-AR has been shown to be crucial in the central hypotensive (MacMillan et al., 1996), sedative (Lakhlani et al., 1997; Sallinen et al., 1997; Stone et al., 1997), and analgesic effects (Lakhlani et al., 1997; Stone et al., 1997) of α₂-AR agonists. In the CNS, α_2B-AR are only expressed in the thalamus (McCune et al., 1993; Nicholas et al., 1993; Scheinin et al., 1994;Winzer-Serhan and Leslie, 1997), but they are present in many tissues outside the CNS, and their physiological significance has been clearly established in the regulation of peripheral blood vessel tone (Link et al., 1996). α_2C-AR are mainly localized in the basal ganglia, olfactory tubercle, hippocampus, and cerebral cortex (Nicholas et al., 1996; Rosin et al., 1996), and behavioral studies with transgenic mice suggest that α_2C-ARs may serve important functions in sensorimotor integration (Sallinen et al., 1998a,b).

Differences in receptor regulation have been described for the three α₂-AR subtypes. Long-term desensitization has been observed in fibroblast cell lines for all α₂-AR subtypes, but the extent of desensitization is greater for the α_2A- and α_2B-AR subtypes compared with α_2C-AR (Eason and Liggett, 1992, 1993). The subtypes also show distinct subcellular localization and targeting patterns in some non-neuronal cell types (von Zastrow et al., 1993;Daunt et al., 1997). Because the α₂-AR have fundamental functions in neuronal physiology and pharmacology, we decided to examine their subcellular distributions and the effects of agonist activation on their trafficking within a neuronal cell type. As a model, we used the rat pheochromocytoma cell line PC12, which has many similarities with postganglionic sympathetic neurons (Lee et al., 1977, 1980), especially after differentiation (Greene and Tischler, 1976). Our hypothesis was that the α_2A-AR subtype would be specifically targeted to nerve terminals, as expected for an autoreceptor, and that the localization of α_2A-AR would thus differ from the other two α₂-AR subtypes.

MATERIALS AND METHODS

Antibodies and chemicals. Rabbit polyclonal antisera against the C termini of α_2A- and α_2C-AR (Daunt et al., 1997) were kindly provided by Dr. B. K. Kobilka (Stanford University, Stanford, CA). A monoclonal antibody against α_2B-AR (Liitti et al., 1997) was kindly provided by S. Liitti and H. Frang (Center for Biotechnology, Turku, Finland). Hygromycin B was from Boehringer Mannheim (Indianapolis, IN), FITC-conjugated sheep anti-mouse IgG, (−)norepinephrine, and the neomycin analog G418 were from Sigma (St. Louis, MO), and the FITC-conjugated sheep anti-rabbit IgG was from Silenius Laboratories (Victoria, Australia). The anti-neurofilament (NF) 160 kDa antibody and the radioligand [³H]RX821002 [2-(2-methoxy-1,4-benzodioxan-2-yl)-2-imidazoline] were from Amersham Pharmacia Biotech (Buckinghamshire, UK). RX821002 was from Research Biochemicals (Natick, MA), atipamezole and dexmedetomidine were gifts from Orion Pharma (Turku, Finland), and rauwolscine was from Carl Roth KG (Karlsruhe, Germany). Cell culture reagents were from Life Technologies (Gaithersburg, MD) unless mentioned otherwise. Nerve growth factor (NGF) was from Promega (Madison, WI). Other chemicals were of analytical or reagent grade and were purchased from commercial suppliers.

Cell culture. PC12 cells (American Type Culture Collection, CRL 1721) were grown in collagen-coated [1% Vitrogen 100 (Collagen Corporation, Fremont, CA) and 0.1% BSA] 25 and 75 cm² cell culture flasks (Falcon; Becton Dickinson, Meylan, France) in DMEM supplemented with 2.5% heat-inactivated fetal bovine serum, 12.5% heat-inactivated horse serum (HS) (Kraeber GmbH & Co., Hamburg, Germany), 292 mg/ll-glutamine, 100 U/ml penicillin, 50 μg/ml streptomycin, 1 mm sodium pyruvate, and 20 mm NaHCO₃, in water-saturated 5% CO₂ at 37°C. Two-thirds of the growth medium was changed every third day, and cells were replated every 5–6 d. Passages 4–20 of stable transfected clones were used. Before plating the cells onto glass coverslips, the cells were trypsinized, because harvesting with a rubber policeman altered the microscopic morphology and slowed the differentiation of PC12 cells. A rubber policeman was used for harvesting cells for radioligand binding studies. Differentiation was induced by incubating cells with 50–100 nm 2.5S NGF in DMEM plus 1–3% HS on collagen or 0.1 mg/ml poly-l-lysine-coated (Sigma) glass coverslips or cell culture plastics. Non-transfected PC12 cells did not express endogenous α₂-AR as evidenced by immunocytochemistry and radioligand binding.

Stable expression of human α₂-ARs.The cDNAs encoding human α₂-AR subtypes were a gift from Dr. R. J. Lefkowitz (Duke University, Durham, NC) (Kobilka et al., 1987; Regan et al., 1988; Lomasney et al., 1990). The construction of the pMAMneo- and pREP-based (Clontech, Palo Alto, CA) expression vectors for stable or semistable expression, respectively, of human α_2A-, α_2B-, and α_2C-AR (Marjamäki et al., 1992, 1993) was performed with standard methods. The PC12 cells were plated onto collagen-coated tissue culture dishes and transfected 24 hr later by the calcium phosphate precipitation method (Chen and Okayama, 1987,1988). The selection of stable or semistable transfectants was performed with 500 μg/ml the neomycin analog G418 (Greene et al., 1991) or Hygromycin B, respectively. Resultant clones were screened for α_2A-, α_2B-, and α_2C-AR expression by immunofluorescent staining and radioligand binding. Only weak immunostaining was observed in clones expressing <500–800 fmol α₂-AR/mg protein. Stable clones expressing receptor densities of 1.0–2.5 pmol/mg total cellular protein were chosen for further studies. The double transfections for colocalization studies were done by pREP-mediated transfection of stable PC12α_2Acells with α_2B-cDNA and of stable PC12α_2B cells with α_2A-cDNA. The coexpression of α_2A- and α_2B-AR was detected by double-staining with subtype-specific antibodies.

Preparation of cell homogenates and saturation binding assays. Cells were harvested into chilled PBS, pelleted, and frozen at −70°C. Saturation binding assays with cell homogenates and [³H]RX821002 were performed in K⁺-phosphate buffer as described previously (Halme et al., 1995).

Receptor activation. Cells were plated on collagen- or poly-l-lysine-coated glass coverslips at 1–2 × 10⁴cells/cm². After 4–6 d of culture, the differentiated and nondifferentiated cells were treated with serum-free DMEM containing 10 μm norepinephrine or 10 nm dexmedetomidine for 30 min at 37°C, in 5% CO₂. The medium was then aspirated, and the cells were rinsed once with 4°C PBS and fixed.

Immunocytochemistry. Cultures of PC12 cells on glass coverslips were fixed for 20 min at room temperature or for 5 min at 4°C and 15 min at room temperature (for agonist-treated cells) with 4% paraformaldehyde in PBS. After fixation, the cells were either stained immediately or stored in PBS at 4°C for later staining. The staining was performed at room temperature. Nonspecific binding was blocked by incubating the cells for 45 min with blocking buffer containing 0.2% Nonidet P-40 (Calbiochem, La Jolla, CA) as permeabilizing agent (not added in all experiments) and 5% non-fat dry milk in 50 mm Tris-HCl, pH 7.6. The cells were incubated for 45 min with primary antibodies or antisera diluted in the same buffer (anti-α_2A and -α_2C, 1:500; anti-α_2B, 10 μg/ml, anti-NF, 1:100). After incubation, the cells were rinsed three times with PBS, followed by 5 min blocking and incubation with secondary antibody diluted 1:500 [FITC-conjugated anti-mouse IgG and FITC- or tetramethylrhodamine isothiocyanate (TRITC)-conjugated anti-rabbit IgG] in the blocking buffer (for 30 min) in darkness. After rinsing three times with PBS, the coverslips were mounted for fluorescent microscopy onto a drop of anti-fade mounting medium containing 50% glycerol, 100 mg/ml 1,4-diazabicyclo-[2.2.2.]octane (Sigma) and 0.05% sodium azide in PBS on microscope slides. As negative controls, transfected PC12 cells were stained in a similar manner but without the primary antibody, or nontransfected PC12 cells were stained as above. Stained cells were not observed under these conditions. The specificity of the receptor antibodies was verified by cross-staining the α₂-AR subtypes in PC12 cells. Single- and double-labeled immunofluorescent microscopy was performed using a conventional fluorescence microscope (Olympus BHS, Olympus 100×/D Plan Apo, 100 UV, 1.30 objective; Olympus Opticals, Tokyo, Japan) and a laser scanning confocal microscope (Leica DM RXA, 100×/1.4 oil ICT:D objective; Leica, Nussloch, Germany).

Immunocytochemistry of cells expressing both α_2A- and α_2B-AR was performed using simultaneously the above described rabbit polyclonal antibody against α_2A-AR and the mouse monoclonal antibody against α_2B-AR. Anti-rabbit TRITC and anti-mouse FITC were used for visualization.

Electron microscopy. PC12 cells were cultured and fixed as described above and post-fixed with potassium ferrocyanide-osmium fixative (Karnovsky, 1971). The cells were embedded in epoxy-resin (Glycidether 100; Merck, Darmstadt, Germany) and then sectioned for electron microscopy. Ultrathin sections were stained with 12.5% uranyl acetate (Stempak and Ward, 1964) and 0.25% lead citrate (Venable and Coggeshall, 1965) and examined in a Jeol JEM-100SX electron microscope (Jeol, Akishima, Japan).

RESULTS

α_2A-AR are distinctly targeted in differentiating PC12 cells

During NGF-induced differentiation, neurites were seen as soon as after 2 d treatment in α_2A-, α_2B-, and α_2C-AR-transfected PC12 cells. Growth cones were seen on plasma membranes of transfected and untransfected PC12 cells as soon as after the first day of NGF treatment. During differentiation, the cells flattened out, and there was an increase in the number of growth cones and in the number and length of neurites.

PC12 cells transfected with the α_2A-AR plasmid acquired an altered phenotype also without NGF treatment (see below). In PC12α_2A cells not treated with NGF, α_2A-AR staining was observed on plasma membranes all over the cell (Fig.1a). Abundant α_2A-AR staining was present in the filopodia and the growth cones of PC12α_2A cells, which remained brightly stained during the entire course of differentiation. After 5–6 d of NGF-induced differentiation and neurite outgrowth, the distal segments of the neurites stained more brightly compared with the weaker staining of the plasma membrane over the cell body, suggesting that the receptors are clustered to the tips of neurites (Fig.1b).

Fig. 1.

Scanning confocal images of nondifferentiated (a, c, e) (100×) and differentiated (b, d,f) (50×) PC12 cells expressing human α₂-AR subtypes, immunostained with α₂-AR subtype-specific antibodies. The images are summations of 12 midsections of permeabilized cells. In nondifferentiated PC12α_2A cells, α_2A-AR staining is seen on plasma membranes all over the cells, and clusters of α_2A-AR are seen on filopodia and growth cones (a, arrow). After NGF-treatment, clustering of the α_2A-AR staining to the tips (arrows) of the neural extensions is seen (b). α_2B-AR staining is localized evenly on plasma membranes (arrows) in nondifferentiated (c) and differentiated (d) cells, and α_2C-AR staining is localized mainly intracellularly (arrows) before (e) and after (f) differentiation.

α_2B-AR were localized evenly on plasma membranes in undifferentiated (Fig. 1c) and differentiated (Fig. 1d) PC12 cells, and there was no enrichment of α_2B-AR staining in growth cones or neurites during NGF-induced differentiation (Fig. 1d). α_2C-AR staining was found mainly intracellularly and perinuclearly in undifferentiated (Fig.1e) and differentiated (Fig. 1f) PC12 cells, and the plasma membrane staining was weaker compared with the α_2A- and α_2B-AR subtypes.

α_2A- and α_2B-AR internalize after agonist activation

Receptor activation was induced with the α₂-AR agonists norepinephrine (10 μm, 30 min) and dexmedetomidine (10 nm, 30 min), in both undifferentiated and differentiated cells. Clear intracellular punctate staining (Fig.2a) could be seen in agonist-treated PC12α_2A cells; this was not observed in cells not treated with α₂-AR agonists. These clustered immunoreactive puncta were considered to represent internalized receptors in some intracellular compartment of the cells, because no punctate staining was observed in nonpermeabilized cells. Also, confocal microscopy showed clusters of immunoreactive material in the midsections of the cells, confirming their intracellular nature. The growth cones and the filopodia of PC12α_2A cells remained brightly stained after agonist treatment (Fig. 2a,b).

Fig. 2.

Scanning confocal images of nondifferentiated (a, c, e) (100×) and differentiated (b, d,f) (50×) PC12 cells expressing human α₂-AR subtypes after norepinephrine (10 μm) treatment. The images are summations of 12 midsections of permeabilized cells. The cells were immunostained with α₂-AR subtype-specific antibodies. Similar receptor trafficking is observed in nondifferentiated and differentiated cells: α_2A- and α_2B-AR are internalized into intracellular vesicles (arrowheads) after norepinephrine-induced receptor activation (a–d), although the growth cones and neural tips of PC12α_2A cells remain brightly stained (a, b, arrows). The localization of α_2C-AR is not changed by agonist treatment (e, f). More numerous and brighter intracellular vesicles are seen in PC12α_2Bcells (c, d) than in PC12α_2A cells (a, b), suggesting that internalization of α_2B-AR is stronger than that of α_2A-AR.

The agonist treatment also induced internalization of α_2B-AR, which was stronger than that of α_2A-AR; in confocal microscopy, most of the receptor staining of agonist-treated PC12α_2Bcells was observed intracellularly, and plasma membrane staining was almost invisible (Fig. 2c). No change was observed in the localization of agonist-activated α_2C-AR compared with cells not treated with α₂-AR agonists (Fig. 2e). There was no difference in the agonist-induced internalization in nondifferentiated versus differentiated PC12α_2B and PC12α_2C cells (Fig.2d,f). All differentiation and receptor activation experiments have been repeated with 2–4 independent clones of the three different human α₂-AR subtypes, 3–30 times per clone.

α_2A-and α_2B-AR internalize into partly separate intracellular vesicles

Immunostaining and scanning confocal microscopy of agonist-treated double-transfected PC12α_2A-B cells showed that α_2A- and α_2B-AR are colocalized in plasma membrane and neural tips before agonist treatment (Fig. 3a,c). After agonist treatment, clear internalized vesicles of both α_2A- and α_2B-AR were seen in 0.16 μm slice thin images (Fig. 3b,d). Intracellular staining of vesicles containing these two internalized receptor subtypes was partly overlapping (yellow in confocal images) and partly distinct (red andgreen, respectively), suggesting that α_2A- and α_2B-AR would be internalized into partly distinct populations of intracellular vesicles. However, both receptor subtypes remained in neural tips after agonist treatment (Fig. 3b,d).

Fig. 3.

Scanning confocal images of PC12α_2A-B cells coexpressing α_2A- and α_2B-AR immunostained with subtype-specific rabbit anti-α_2A antiserum–anti-rabbit TRITC (red) and mouse anti-α_2Bantibody–anti-mouse FITC (green). Images (a, b, 100×; c,d, 50×) represent single 0.16 μm thin midsections from scanned cells. Overlapping α_2A- and α_2B-AR staining (yellow) is seen on the plasma membrane and in the tips (arrows) of neural extensions (a–d), indicating colocalization of α_2A- and α_2B-AR in these locations. After agonist treatment (b, d), α_2A- and α_2B-AR are internalized into partly separate (separate red and greenpuncta) (arrowheads) and partly shared (yellow puncta) intracellular vesicles. The tips of the neurites remain yellow (b,d) (arrows), indicating that the two subtypes are still colocalized and not redistributed by agonist treatment in these parts of the cells.

Expression of α_2A-AR induces differentiation of PC12 cells

The culture of PC12 cells was performed according to Greene and Tischler (1976), using collagen-coated Falcon cell culture flasks after selection of the transfected clones. The morphology of the cells was not changed during 3–20 passages when the cells were replated every 5–6 d and divided 1:4–6.

Receptor subtype-specific changes in PC12 cell shape and growth type were discovered after the selection of the transfected clones. The shape of PC12α_2A cells was clearly different from clones expressing α_2B- and α_2C-AR. The PC12α_2Acells flattened out, and their cytoplasm was thinner compared with the other clones. The cells tended to differentiate without NGF, and they developed growth cones and short neurites (Fig.4b). The extent of this tendency to differentiate was proportional to the receptor expression level of individual α_2A-AR-expressing clones (n = 5) (data not shown). To examine the possibility that activation of α_2A-AR by endogenous norepinephrine produced by the PC12α_2A cells would induce this morphological change, we cultured the cells in medium supplemented with selected α₂-AR antagonists to block α_2A-AR. Treatment of PC12α_2A cells with the α₂-AR antagonists atipamezole, rauwolscine, or RX821002 (5 μm) did not influence the tendency of PC12α_2A cells to undergo morphological differentiation. The shape of PC12α_2B and PC12α_2C cells (Fig. 4c,d) remained similar to that of untransfected PC12 cells. Both untransfected and transfected PC12 cells were successfully stained with an anti-neurofilament antibody to confirm the preserved neuronal phenotype of the cells (data not shown).

Fig. 4.

Phase-contrast microscopic images (10×) of living nontransfected (a) and α₂-AR expressing (b–f) PC12 cells in collagen-coated Falcon cell culture flasks. Compared with nontransfected PC12 cells (a), the α_2A-AR expressing cells (b) showed morphological changes under normal culturing conditions during the first few (1–3) passages after transfection; the PC12α_2A cells flattened out and their cytoplasm was thinner compared with clones expressing other α₂-AR subtypes, and they developed growth cones and short neurites without NGF-treatment (b). The morphology of PC12α_2B and PC12α_2C cells remained similar to the nontransfected cells (c,d). The different morphology of PC12α_2Acells (α_2A-B) did not change by the coexpression of α_2B-AR (e), but expression of α_2A-AR in PC12α_2B(α_2B-A) cells induced similar morphological changes (f) as after transfection of PC12 cells with α_2A-AR alone (b). Culture from passage 3 to passage 20 did not change the acquired morphology of the clones.

The α_2A -AR-induced differentiation allows targeting of α_2A- and α_2B-AR to the tips of neural extensions of PC12 cells

To explore whether the targeting of α_2A-AR to the tips of neural extensions of differentiated PC12 cells would be secondary to the changes induced by expression of the α_2A-subtype, cotransfection studies were performed. Expression of α_2A-AR in PC12α_2B cells (PC12α_2B-A cells) induced similar differentiation-like morphological changes, as seen after transfection with α_2A-AR alone. The morphological changes were obvious as soon as during the first passage (Fig.4f). Expression of α_2B-AR in PC12α_2A cells (PC12α_2A-B cells) did not change the α_2A-induced differentiated phenotype of the cells (Fig. 4e). Immunostaining of cotransfected PC12α_2A-B and PC12α_2B-Acells, visualized by conventional fluorescent microscopy and scanning confocal microscopy, showed overlapping α_2A- and α_2B-specific immunoreactivity in the tips of neural extensions (Figs. 3,5c–f). This indicates that expression of α_2A-AR permits targeting of α_2B-AR also to neurite tips in PC12 cells. The different distribution of α_2B-AR in PC12α_2B cells differentiated with NGF and in PC12α_2A-B/α_2B-A cells expressing both α_2A- and α_2B-AR is illustrated in Figure 5.

Fig. 5.

Scanning confocal images (50–100×) of non-NGF-differentiated (a, c,d, e, f) and NGF-differentiated (b) PC12 cells expressing human α_2B-AR alone (a, b) and coexpressing human α_2A- and α_2B-AR (c–f), immunostained with monoclonal anti-human α_2B-AR antibody (a, b,c, e) and polyclonal α_2A-AR antiserum (d, f). The images are summations of 12 midsections of permeabilized cells. When expressed alone, α_2B-AR are localized evenly on plasma membranes in nondifferentiated (a) and differentiated (b) PC12α_2B cells, and no concentration of staining into the tips of neural processes is seen (arrows). When α_2B-AR are expressed in PC12α_2A cells (PC12α_2A-B), the α_2B-AR are targeted into the growth cones and neural processes (c) (arrows), and overlapping staining can be seen with α_2A-AR (d). The expression of α_2A-AR in PC12α_2B cells (PC12α_2B-A) induced differentiation-like morphological changes and similar targeting of α_2B-AR (e) and α_2A-AR (f) into the tips (arrows) of developed neural processes.

To examine the possibility that extensive membrane ruffling in neural extensions of PC12 cells would create the image of receptor concentration, we prepared electron microscopic samples of PC12 cell neural extensions. Prominent membrane ruffling was seen in many growth cones and shorter extensions (average length in slice, 5.8 μm), but in longer extensions (average length in slice, 18.8 μm), similar to the neurite tips in which receptor concentration was observed, no apparent membrane ruffling was seen, indicating that the concentration of α₂-AR-specific staining in the tips of neural extensions was not an artifact induced by membrane ruffling.

DISCUSSION

The aim of this study was to examine the subcellular distribution of human α₂-AR subtypes in neuronal cells, because α₂-ARs are widely distributed in the brain and spinal cord and in peripheral sympathetic nerve cells and because the elucidation of their targeting and trafficking properties might provide important information about their function and regulation (Nicholas et al., 1996; Koenig and Edwardson, 1997). As a neuronal cell type, we used PC12 cells, which are derived from a rat pheochromocytoma tumor (Greene and Tischler, 1976) and are often used as a model of postganglionic sympathetic neurons (Lee et al., 1977; Youdim et al., 1986).

Receptor subtype-specific differences were demonstrated in the subcellular localization of transfected human α₂-AR subtypes in differentiating neuronal cells; α_2A-AR were distinctly targeted to the tips of the neural extensions, and both α_2A- and α_2B-AR were internalized after receptor activation. The results also show that expression of α_2A-AR induced differentiation-like changes in PC12 cell morphology and growth type, which also permitted targeting of α_2B-AR to the tips of the neural extensions.

The predominant α₂-AR subtype in the brain and spinal cord is the α_2A-AR (MacDonald et al., 1997; Lawhead et al., 1992; Nicholas et al., 1993), which mediates the hypnotic, sedative, sympatholytic, as well as analgesic, effects of α₂-AR activation (MacMillan et al., 1996;Lakhlani et al., 1997; Sallinen et al., 1997; Stone et al., 1997). In major noradrenergic nuclei of the rat brain, α_2A-mRNA is present in the cell bodies, which suggests that α_2A-AR are presynaptic autoreceptors, mediating inhibition of synaptic norepinephrine release (Scheinin et al., 1994). The α_2A-AR immunoreactivity in the brain is found on presynaptic axons and also in perikarya in which it is localized within intracellular structures involved in synthesis and trafficking of receptors (Milner et al., 1999). The CNS functions of α_2B-AR are still unknown (MacDonald et al., 1997) and α_2C-AR are thought to have modulatory effects on several different brain functions, e.g., sensorimotor integration (MacDonald et al., 1997;Sallinen et al., 1997, 1998a,b).

This study shows that, during PC12 cell differentiation, α_2A-AR are targeted to the developing neurites, concentrating into their tips, directed by functional changes induced by this receptor. This indicates that α_2A-AR have properties that enable the intracellular machinery of the cell to induce differentiation and to allow membrane-bound receptors to be transported along the neurites to the site of autoreceptor action. The targeting of α_2A-AR to the tips of the neurites during neuronal differentiation resembles that of the G-protein isoform G_o, which mediates α₂-adrenergic inhibition of neurotransmitter release. Also, G_o-proteins are targeted to the growth cones, filopodia, and the tips of neurites during NGF-induced differentiation of PC12 cells (Zubiaur and Neer, 1993). On the contrary, in Madin–Darby canine kidney (MDCK) cells, transfected α_2A-AR have been shown to be targeted to the basolateral surface, which is thought to correspond to the somatodendritic parts of neuronal cells (Keefer et al., 1994). It was, however, recently shown that trafficking of α_2A-AR in neurons could be predicted based on the trafficking of the endogenous apically expressed α_2A-AR in MDCK cells (Okusa et al., 1994). All α₂-AR subtypes were localized in the cell bodies and along neurites when transfected into cultured primary spinal cord neurons (Wozniak and Limbird, 1998). In our study, using a model of peripheral postganglionic neurons, the distinct targeting of α₂-AR subtypes in PC12 cells was found to be secondary to the functional changes induced by α_2A-AR in these cells. The distinct targeting of α_2A-AR gives evidence of the neuronal autoreceptor-like character of human α_2A-AR in peripheral sympathetic neurons, because α_2B-AR were evenly distributed in the plasma membrane and α_2C-AR were mainly intracellular when expressed in PC12 cells without α_2A-AR. This is in line with previous studies on non-neuronal cell types; in nonpolarized fibroblasts (von Zastrow et al., 1993; Daunt et al., 1997) and in polarized epithelial cells (Wozniak and Limbird, 1996), α_2A- and α_2B-AR resided primarily in the plasma membrane, whereas a large proportion of α_2C-AR was found in the endoplasmic reticulum and in cis/medial Golgi.

The mechanisms mediating targeting of receptors and other membrane-associated proteins in nerve cells are not fully understood. Mutagenesis studies of α_2A-AR have shown that the direct delivery of this receptor to the basolateral membrane in polarized MDCK cells most likely involves transmembranous structures and that the third cytoplasmic loop of the receptor probably contains structural elements important for the stabilization of the receptor in its subcellular locus (Keefer et al., 1994; Saunders et al., 1998). Synaptic proteins, e.g., postsynaptic density-95/synapse-associated protein 90 (PSD-95/SAP90), participate in targeting of many membrane-associated neuronal proteins (Kim et al., 1995; Kornau et al., 1995). The roles of these and other proteins in targeting the adrenergic α₂-receptors remain to be demonstrated.

Subtype-specific desensitization properties of α₂-AR have been reported. In Chinese hamster ovary and COS cells, both α_2A- and α_2B-AR are effectively desensitized by phosphorylation, whereas α_2C-AR are not (Eason and Liggett 1992, 1993; Kurose and Lefkowitz, 1994). In a study on transfected human embryonic kidney 293 (HEK-293) cells, extensive agonist-induced internalization of mouse α_2B-AR was observed, whereas very limited internalization of mouse α_2A-AR was detected by ELISA but not by immunofluorescent methods (Daunt et al., 1997). This suggested that α_2A-AR internalization may occur by a nonendosomal mechanism, different from α_2B-AR internalization. In this study, we could visualize extensive internalization of human α_2A-AR after agonist-induced activation of the receptors in both nondifferentiated and differentiated PC12 cells. This emphasizes the importance of the cellular environment for receptor function. The neuronal, adrenergic PC12 cells contain an as yet unknown machinery that makes the rapid agonist-induced internalization of α_2A-AR possible. Internalization of this α₂-AR subtype may subserve some important regulatory function in neuronal cells not observed in some other types of cells. The extensive internalization of α_2B-AR observed in this study has also been reported in other cell types (Daunt et al., 1997). No α_2C-AR internalization was visualized in this study, because this receptor was mainly localized intracellularly already at baseline, and a further reduction in the small plasma membrane α_2C-AR population would not have been detectable.

A recent study of D1 and D2 dopamine receptors coexpressed in HEK-293 cells showed receptor subtype-specific endocytosis by distinct mechanisms (Vickery and von Zastrow, 1999). When coexpressing α_2A- and α_2B-AR in the same PC12 cells, the receptor subtypes were seen in partly separate populations of intracellular vesicles (Fig. 3), indicating partly distinct endocytotic mechanisms. The intracellular vesicles containing α₂-AR subtypes could also be in different stages of the endocytotic cycle, suggesting that the kinetics of internalization would differ between the α₂-subtypes.

In the present study, PC12 cells acquired morphological features dependent on the transfected α₂-AR subtype; especially expression of the α_2A-AR subtype clearly altered their morphology. The human α_2A-AR is known to control actin polymerization and focal adhesion assembly in preadipocytes but only after agonist stimulation of overexpressed α_2A-AR or by constitutively active α_2A-AR (Betuing et al., 1996). Overexpression of Ca²⁺/calmodulin-dependent protein kinase II has also resulted in altered PC12 cell growth and morphology (Masse and Kelly, 1997). Morphological changes of PC12 cells will not, however, happen along with expression of any protein at densities above physiological levels because, in this study, only PC12α_2A cells presented marked morphological differentiation-like changes, despite similar receptor densities. The morphological changes induced by α_2A-AR expression were not inhibited by α₂-AR antagonist treatment, which indicates that the change is not induced by α_2A-AR activation by norepinephrine produced and released by the cells. The morphological changes were seen in all PC12α_2A clones expressing α_2A-AR above the level of 500 fmol/mg protein, and the changes were more pronounced in cell clones with higher receptor densities. Therefore, it seems unlikely that our results were attributable to selecting G418-resistant PC12 clones that simply were more efficient in differentiation. The differentiating effect of α_2A-AR could be repeated with the expression of α_2A-AR in PC12α_2Bcells; the morphology of these PC12α_2B-A cells became similar to that of PC12α_2A cells. The mechanism of subtype-specific morphological change of PC12α_2A cells remains unclear, but constitutive activity of the transfected α_2A-AR is one possible mechanism mediating this effect.

The human α_2A-AR was described to induce membrane ruffling of preadipocytes (Betuing et al., 1996). To avoid misinterpretations of receptor localization induced by immunocytochemical staining of the ruffled plasma membrane, electron microscopy of PC12α_2A-_Band PC12α_2B-_A cells was performed. Membrane ruffling was detected in growth cones and short neural extensions but not in longer extensions. The concentration of α_2A-specific immunocytochemical staining seen in the long neural extensions is therefore not induced by membrane ruffling. Also, the concentrated staining of the neural tips in thin slices of confocal images in Figure 3 confirms this observation.

The specific targeting of α_2A-AR to developing nerve terminals indicates that its localization is appropriate for an autoreceptor. Extensive internalization of α_2A-AR was observed in nondifferentiated and differentiated sympathetic neuron-like cells, which indicated different sequestration of this receptor in PC12 cells compared with some non-neuronal cell types. The significance of this difference for functional α_2A-AR desensitization is still unknown. These present results point out the importance of the cellular environment for receptor endocytosis and trafficking.