Intracellular Retention of Recombinant GABABReceptors*

γ-Aminobutyric acid type B (GABAB) receptors mediate the transmission of slow and prolonged inhibitory signals in the central nervous system. Two splice variants of GABAB receptors, GABABR1a and GABABR1b, were recently cloned from a mouse cortical and cerebellar cDNA library. As predicted, these receptors belong to the G protein-coupled receptor superfamily. We have used epitope-tagged versions of GABABR1a receptors to study the cellular distribution of these proteins in a variety of non-neuronal and neuronal cell types. Here we report that recombinant GABAB receptors fail to reach the cell surface when expressed in heterologous systems and are retained in the endoplasmic reticulum when introduced into COS cells. In addition, we prove that recombinant GABAB receptors are excluded from the cell surface when overexpressed in ganglion neurons and we further demonstrate that they fail to activate in superior cervical ganglion neurons. Together our observations suggest that recombinant GABAB receptors require additional information for functional targeting to the plasma membrane.

␥-Aminobutyric acid type B (GABA B ) receptors mediate the transmission of slow and prolonged inhibitory signals in the central nervous system. Two splice variants of GABA B receptors, GABA B R1a and GABA B R1b, were recently cloned from a mouse cortical and cerebellar cDNA library. As predicted, these receptors belong to the G protein-coupled receptor superfamily. We have used epitope-tagged versions of GABA B R1a receptors to study the cellular distribution of these proteins in a variety of non-neuronal and neuronal cell types. Here we report that recombinant GABA B receptors fail to reach the cell surface when expressed in heterologous systems and are retained in the endoplasmic reticulum when introduced into COS cells. In addition, we prove that recombinant GABA B receptors are excluded from the cell surface when overexpressed in ganglion neurons and we further demonstrate that they fail to activate in superior cervical ganglion neurons. Together our observations suggest that recombinant GABA B receptors require additional information for functional targeting to the plasma membrane.
␥-Aminobutyric acid (GABA) 1 is the main inhibitory neurotransmitter in the central nervous system. The fast inhibitory activity of GABA is mediated by the activation of ionotropic GABA A (and GABA C ) receptors which are members of the ligand gated ion channel superfamily (1). The slower inhibitory actions of GABA are mediated by the activation of GABA B receptors which are G protein-coupled receptors (2,3). GABA B receptors couple to various effector systems mainly through the activation of the specific G proteins G ␣ o, G ␣ q, G ␣ i, and G ␤␥ . Via these signal transduction molecules they inhibit adenylyl cyclase, inhibit histamine-and serotonin-induced accumulation of inositol triphosphate, inhibit voltage-gated calcium channels and activate potassium channels (3). In general, activation of GABA B receptors results in modulation of neurotransmitter release from pre-synaptic neurons or hyperpolarization of post-synaptic membranes. These two mechanisms contribute significantly to long-term inhibition of synaptic transmission (3) and their inappropriate function may be implicated in a number of diseases of the central nervous system (3)(4)(5).
Two amino-terminal splice variants of GABA B receptors, GABA B R1a and GABA B R1b, were recently cloned from a mouse cortical and cerebellar cDNA library (6). As expected, the proteins belong to the seven transmembrane, G proteincoupled receptor superfamily. In particular, the two clones showed significant homology to metabotropic glutamate receptors (7,6), a Ca 2ϩ sensing receptor (8), and vomeronasal receptors (9,10). Receptors belonging to this family contain an intracellular carboxyl-terminal tail, characteristic seven-membrane spanning domains, and a large amino-terminal extracellular domain, which in metabotropic glutamate receptors and GABA B receptors includes a region structurally related to bacterial amino acid-binding proteins (6,11). Binding properties to antagonists revealed strong similarities between expressed recombinant GABA B R1a and endogenous rat cortex GABA B receptors. However, recombinant receptors had a 100-fold reduced affinity for full agonists as compared with native receptors (6). The significance of this discrepancy is still unclear. In terms of coupling to effector systems, GABA B R1a mediated the inhibition of adenylyl cyclase in transfected HEK 293 cells. Nevertheless, the receptors failed to couple to K ϩ channels in Xenopus oocytes (6).
In order to investigate the possible causes of the inability of recombinant receptors to couple properly to effector systems, we have used epitope-tagged versions of recombinant GABA B receptors to examine for the first time their subcellular distribution in heterologous systems and neuronal cell types. We have found that recombinant GABA B receptors fail to reach the cell surface in several cellular contexts and remain inactive in superior cervical ganglion (SCG) neurons. Our results suggest that intracellular retention constitutes a plausible reason for the inactivity of recombinant receptors in the systems studied so far.

EXPERIMENTAL PROCEDURES
Plasmids and DNA Constructions-pBlueScript based plasmids containing the sequence for GABA B R1a receptors were tagged with the 10-amino acid 9E10 epitope (EQKLISEEDL) from c-Myc (12) by site-directed mutagenesis (13). Oligonucleotide GABA B -9E10 -1 (GGTGGC-GTTGGGGGTTAGGTCTTCTTCTGATATTAGCTTTTGTTCCTGCGC-CCCGCCAGC) was used to tag the receptor between the fourth and fifth amino acid of the mature protein and GABA B 9E10-3 (CTCCATG-CCCCCTCATAGGTCTTCTTCTGATATTAGCTTTTGTTCCTTGTAAA-GCAAATG) was used to tag the receptor at the carboxyl terminus. All mutations and the fidelity of the final expression constructs were verified by DNA sequencing. SpeI-NdeI fragments of these constructs or of untagged sequences were subcloned into a CMV based expression vec-tor to constitute the plasmids AC5-Gbm (amino-terminal 9E10, N-GABA B ), AC10-Gbm (carboxyl-terminal 9E10, C-GABA B ), and AC25-Gb (untagged receptor, GABA B ). GABA A receptor subunit expression constructs have been previously described (14). A reporter plasmid encoding for the S65T mutant of the jellyfish Aequorea victoria green fluorescent protein (GFP) was used as described previously (15).
Cell Culture and Transfection-COS, human embryonic kidney (293 HEK), and baby hamster kidney cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum. Exponentially growing cells were transfected with 10 g of DNA by electroporation using a Bio-Rad Gene-Pulser H according to instructions from the manufacturer. Cells were incubated 12-18 h to allow expression of recombinant proteins.
Antibodies-9E10 antibody was obtained from 9E10 hybridoma cells and used as supernatant for immunofluorescence or as affinity purified antibody for immunoprecipitations and 125 I-surface binding experiments. Polyclonal anti-Myc antibody was purchased from Santa Cruz Biotechnology. Anti-FLAG mouse monoclonal antibody was purchased from IBI Ltd. The secondary antibodies, anti-mouse Texas Red (TR), anti-mouse fluorescein isothiocyanate, and anti-rabbit fluorescein isothiocyanate were purchased from Pierce.
Metabolic [ 35 S]Methionine Labeling and Immunoprecipitations-COS cells were cultured as described, transfected by electroporation, and plated onto 6-cm dishes. Cells were washed twice with methioninefree DMEM and L-methionine starved for 30 min at 37°C. Cells were washed twice and incubated in 1 ml of methionine-free DMEM containing 0.5 mCi [ 35 S]methionine (Easytag, NEN Life Science Products Inc.) for 4 h at 37°C. For pulse-chase experiments cells were labeled for 2 h and further incubated in complete DMEM supplemented with 3 mg/ml L-methionine for the indicated period of time. For glycosylation analysis, cells were exposed to tunicamycin (5 g/ml) during the 4-h [ 35 S]methionine labeling period. After incubation periods, cells were washed twice with phosphate-buffered saline (PBS) and removed in 0.6 ml of lysis buffer (50 mM sodium fluoride, 10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 5 mM EGTA, 5 mM EDTA, 16.9 mM Na 2 HPO 4 , 3.1 mM NaH 2 PO 4 , 0.5% deoxycholic acid, 1% Nonidet P-40, 0.1% phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml pepstatin, 10 g/ml antipain). Nuclei were removed from cell lysates by centrifugation at 14,000 rpm ϫ 5 min at 4°C in a microcentrifuge. Cell lysates were preabsorbed with previously equilibrated 25 l of protein A-Sepharose for 1 h at 4°C. Protein A-Sepharose beads were then removed by centrifugation at 14,000 rpm ϫ 30 s at 4°C, and lysates were incubated with 5 g of 9E10 monoclonal antibody for 1 h at 4°C. Immune complexes were precipitated with 25 l of protein A-Sepharose for 1 h at 4°C. Sepharose beads were washed twice in lysis buffer, twice in high salt lysis buffer (lysis buffer containing 0.5 M NaCl), and again in lysis buffer. Bead pellets were resuspended in 40 l of SDS-polyacrylamide gel electrophoresis (PAGE) loading buffer, boiled for 3 min, and analyzed by 8% SDS-PAGE and autoradiography.
Endoglycosidase H Treatment-Transfected COS cells were metabolically labeled for 4 h and immunoprecipitated with 9E10 antibodies as described above. Bead pellets were washed three times in lysis buffer and three times in high salt lysis buffer. Sepharose beads were then resuspended in 50 l of endoglycosidase H reaction buffer (40 mM phosphate buffer, 50 mM EDTA, 0.5% Nonidet P-40, 0.01% SDS, 1% ␤-mercaptoethanol, pH 5.5) and boiled for 3 min. 1 milliunit of endoglycosidase H (Boehringer Mannheim) and protease inhibitors (0.1% phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml pepstatin, 10 g/ml antipain) were added and digestions were carried at 37°C for 14 h. After the incubation period 25 l of SDS-PAGE loading buffer were added, reactions were boiled for 4 min, and analyzed by 8% SDS-PAGE and autoradiography.
Immunofluorescence-Transfected cells were plated onto 6-cm dishes containing 5 g/ml fibronectin and 10 g/ml poly-L-lysine-coated coverslips. Cells were washed twice with PBS, fixed for 30 min in 4% paraformaldehyde, washed twice with 0.5 M NH 4 Cl, and blocked for 30 min in immunofluorescence solution (0.25% bovine serum albumin, 10% fetal bovine serum in PBS). For detection of surface proteins in intact cells, cells were incubated with 9E10 supernatant diluted 1:1 in immunofluorescence solution for 45 min at room temperature and with 1 g/ml TR or fluorescein isothiocyanate-conjugated secondary antibodies for 45 min. For treatment of permeabilized cells, the procedure described above was followed by incubating the cells in immunofluorescence solution containing 0.5% Nonidet P-40 for 20 min at room temperature. Cells were washed twice with 0.5 M NH 4 Cl and blocked as described above. Incubation with 9E10 was performed with 1:1 diluted antibody in immunofluorescence solution containing 0.1% Nonidet P-40. Different secondary antibodies were used for detection of intra-cellular epitopes in order to compare surface and intracellular protein localization. All samples were washed with PBS before mounting and were examined using a confocal microscope (MRC1000, Bio-Rad).
125 I-Surface Binding-Transfected COS and HEK 293 cells were plated onto uncoated or 5 g/ml fibronectin and 10 g/ml poly-L-lysine coated 6 ϫ 4-cm well plates, respectively. Prior to incubation with 9E10 antibody, cells were chilled on an ice bed, washed three times with ice-cold PBS, and twice with iodination medium (0.25% BSA NaHCO 3free DMEM, pH 7.4). Cells were then incubated in iodination medium containing 9E10 iodinated to a specific activity of 500 Ci/mmol (Amersham International) and used at 10 -15 g/ml for 90 min on an ice bed. After incubation, medium was removed and cells were washed five times with ice-cold PBS. Cells were removed with 1 ml of trypsin, transferred to Röhren tubes, and counted in a ␥-counter (Wallac, 1261 Multigamma).
Neuron Primary Cultures, Microinjection, and Recording of Ca 2ϩ Channel Currents-For electrophysiology, single SCG neurons were dissociated from 15-19-day-old rats and plated on poly-D-lysine-coated glass coverslips bordered by 2-cm plastic rings as described previously (16). 4 h after plating, neurons were microinjected with an equal mixture of plasmids carrying cDNA for the N-GABA B or GABA B receptor (final pipette concentration of 0.5 g/l dissolved in water) and cDNA for S65T mutant of GFP or with GFP cDNA alone. DNA (1.2 l) was loaded into pre-pulled high resistance (ϳ30 megaohm) "Pyrex" glass pipettes and injected into the nucleus of single neurons as described previously for injection of antisera or cRNA (17,18). After injections, cells were incubated for 14 -24 h in a humidified incubator (5% CO 2 ) at 37°C. Injected neurons which successfully expressed cDNA were identified as bright fluorescent cells using an inverted microscope (Diaphot 200, Nikon, Japan) equipped with an epifluorescent block N B2E (Nikon). Electrophysiological recordings were made at room temperature 14 -24 h after injection. SCG and DRG primary neurons for immunofluorescence, kindly provided by Mariza Dayrell and Kenji Okuse, respectively, were obtained and cultured as described previously (19,20). Microinjection of neurons was performed as described previously (21) using an Eppendorf Transjector (5246), an Eppendorf Micromanipulator (5171), and a Nikon Diaphot 300 microscope.
Ca 2ϩ Channel Current Recordings-Currents through voltage-gated Ca 2ϩ channels were recorded using the conventional whole cell patchclamp method as described previously (17). Cells were superfused (20 -25 ml/min) with a solution consisting of 120 mM tetraethylammonium chloride, 3 mM KCl, 1.5 mM MgCl 2 , 5 mM BaCl 2 , 10 mM HEPES, 11.1 mM glucose, and 0.5 mM tetrodotoxin. The pH was adjusted to 7.35 with NaOH. Patch electrodes (2-3 megaohm) were filled with a solution containing 110 mM CsCl, 3 mM MgCl 2 , 40 mM HEPES, 3 mM EGTA, 2 mM Na 2 ATP, 0.5 mM Na 2 GTP (pH adjusted to 7.4 with CsOH). Neurons were voltage-clamped using a discontinuous ("switching") amplifier (Axoclamp 2B) sampling voltage at 6 -8 kHz (50% duty cycle). Commands were generated through a Digidata 1200 interface using "pClamp 6" computer software (Axon Instruments, Foster City, CA). Ca 2ϩ channel currents were routinely evoked every 20 s with 100-ms depolarizing rectangular test pulse to 0 mV from a holding potential of Ϫ90 mV. Currents were digitized and stored on a computer for later analysis using pClamp 6 software. Ca 2ϩ channel current amplitudes were measured isochronally 10 ms from the onset of the rectangular test pulse, i.e. near to the peak of the control current. To eliminate leak currents, Ca 2ϩ was substituted for Ba 2ϩ in the external solution at the end of each experiment to block all Ca 2ϩ channel currents and the residual current was digitally subtracted from the corresponding currents in Ba 2ϩ solution. Data are presented as mean Ϯ S.E. The statistical significance of the results was determined by Student's t test (unpaired).

Expression of Recombinant GABA B Receptors-To initiate
our studies of subcellular distribution, targeting and cellular trafficking of recombinant GABA B receptors, we introduced the Myc epitope tag at the amino-(N-GABA B ) and carboxyl-terminal (C-GABA B ) domains of the murine GABA B R1a protein.
These constructs were used to analyze the expression of GABA B receptors in heterologous systems. To achieve this, COS cells were transfected with N-GABA B and metabolically labeled with [ 35 S]methionine. GABA B receptors were immunoprecipitated using 9E10 monoclonal antibodies directed against the Myc epitope. A predominant protein of approximately 120 -130 kDa was specifically immunoprecipitated from lysates derived from N-GABA B transfected cells and was absent from mock transfected controls (Fig. 1A, lanes 1 and 3). The size of the immunoprecipitated protein is in agreement with the molecular mass of recombinant GABA B R1a (6). Notably, a fraction of the immunoprecipitated product migrated above 200 kDa in N-GABA B -transfected cells and was also absent from mock transfected controls (Fig. 1A, lanes 1 and 3). The presence of these bands suggest the existence of GABA B receptor aggregates resistant to reducing conditions. The existence of metabotropic glutamate receptors dimers sets a precedent for such observation (22). Analysis of the GABA B R1a DNA sequence indicates that the gene product contains at least seven putative N-glycosylation sites. To determine whether N-GABA B was glycosylated in this particular expression system, cells were exposed to tunicamycin during the 35 S labeling period and immunoprecipitated as described above. N-GABA B was sensitive to tunicamycin as indicated by a significant decrease in the molecular mass of the 120 -130 kDa immunoprecipitated protein (Fig. 1A, lanes 3 and 4) indicating that recombinant GABA B receptors are glycoproteins.
In order to confirm that the immunoprecipitated protein corresponded to the full-length GABA B receptor, cells were transfected with C-GABA B . Since this receptor contains the epitope tag at the carboxyl terminus, only translation of the complete gene would ensure detection of a protein of a mass of 120 -130 kDa. Mock and C-GABA B -transfected cells were metabolically labeled and the receptor was immunoprecipitated as described above. A prevailing protein was immunoprecipitated from C-GABA B lysates and was absent from lysates of mock transfected cells (Fig. 1B, lanes 1 and 2). Importantly, the migration of C-GABA B and N-GABA B were indistinguishable by SDS-PAGE (Fig. 1, A, lane 3, and B, lane 2). Together these results demonstrate that full-length recombinant GABA B receptors are expressed in COS cells and that they are modified by glycosylation, presumably as they enter and mature through the secretory pathway.
GABA B Receptors Fail to Reach the Cell Surface in Heterologous Systems-The predicted membrane topology of GABA B receptors and other G protein-coupled receptors suggests that the amino-terminal domain of the protein is exposed to the extracellular environment (6,23). This implies that the 9E10 epitope in cell surface N-GABA B receptors will be accessible to antibodies in intact cells. Taking advantage of this positionspecific marker, we sought to determine the cellular distribution of GABA B receptors in COS cells using immunofluorescence. As a control for the assay, cells were transfected with an amino-terminal 9E10-tagged version of the ␤3 subunit of the GABA A receptor (␤3-9E10). The ␤3 subunit is targeted to the cell surface and forms functional Cl Ϫ channels when expressed homomerically in a variety of cell types (24,25). In addition, its amino-terminal domain is exposed extracellularly following targeting to the plasma membrane. In agreement with these observations, ␤3-9E10 was expressed in COS cells and detected at the cell surface of intact cells ( Fig. 2A, upper left  panel). Upon detergent permeabilization, significant amounts of the protein were detected in intracellular compartments confirming previous reports (26) (Fig. 2A, lower left panel). Surprisingly, when recombinant GABA B receptors were expressed in the same system, no surface staining was detected in intact cells ( Fig. 2A, upper right panel). However, upon permeabilization, large amounts of the protein were detected intracellularly ( Fig. 2A, lower right panel). To confirm these results, N-GABA B was expressed in HEK 293 and baby hamster kidney cells. The receptors failed to reach the plasma membrane in either cell type (Fig. 2, B and C, upper panels) despite high intracellular accumulation of the protein (Fig. 2, B and C, lower  panel). These results suggest that recombinant receptors are excluded from the plasma membrane when expressed in heterologous systems.
To further examine these results, the presence of surface proteins was evaluated by measuring the amount of surfacebound 125 I-conjugated 9E10 antibodies in intact COS and HEK 293 cells. As a control, the ␣1 and ␤2 subunits of GABA A receptors were used. The subcellular localization of these proteins has been extensively studied in our laboratory. Briefly ␣1 and ␤2 fail to reach the cell surface when expressed independently, but coexpression overcomes this deficiency and the ␣1␤2 dimer is found predominantly at the cell surface (14). To follow the distribution of ␣1 and ␤2, an amino-terminal 9E10-tagged ␣1 subunit was utilized (␣1-9E10). Importantly, the presence of an amino-terminal epitope tag does not modify the function of the ␣1 subunit and allows detection of the protein at the cell surface of intact cells (14). As expected, surface levels of ␣1- 9E10 were very low and were significantly increased by coexpression of ␤2 (Fig. 3A). In agreement with our immunofluorescence results, surface levels of N-GABA B were similar to those of ␣1-9E10 (Fig. 3A). More importantly, N-GABA B and C-GABA B levels did not differ significantly. Since the carboxylterminal epitope in C-GABA B is predicted to reside in the cytoplasm, these results clearly indicate that both tagged receptors are equally inaccessible to 9E10 antibodies in intact cells. The same experiment was performed in HEK 293 cells and similar results were obtained (Fig. 3B). These observations further suggest that recombinant GABA B receptors fail to reach the plasma membrane in heterologous systems.
The possibility that the 9E10 epitope might interfere with receptor localization or function seems unlikely because the 9E10-tagged full-length receptor is expressed abundantly, modified by glycosylation, and presents antagonist binding properties which are very similar to the ones observed for untagged receptors, suggesting that N-GABA B receptors maintain the appropriate conformation (not shown).
Rapid recycling of receptors between an intracellular compartment and the plasma membrane, with short permanence at the cell surface, would result in intracellular accumulation of most of the protein. To address this possibility, putative cell surface receptors were labeled with antibodies on live N-GABA B -transfected COS cells. Cells were exposed to 9E10 monoclonal antibodies and incubated at 37°C to allow endocytosis of the potential antibody-receptor complexes. Subsequently, immunofluorescence was performed to detect these complexes in vesicular structures underlying the plasma membrane. Although this technique has been applied successfully in detecting the fast recycling of the GABA A ␥2 subunit, 2 no signal was obtained upon examination of N-GABA B transfected cells that had been fed with 9E10 antibodies (not shown).
The inability of recombinant GABA B receptors to reach the cell surface under our experimental conditions may result from a very slow transport machinery. To address this issue the half-life of N-GABA B was determined in COS cells by a pulsechase experiment. Briefly, cells were transfected with N-GABA B and labeled with a pulse of [ 35 S]methionine. After 2 h, cells were incubated in regular media and lysates were prepared at different time points after the [ 35 S]methionine pulse. N-GABA B was immunorecipitated from cell lysates and analyzed by SDS-PAGE. Confirming the results described above, N-GABA B was specifically immunoprecipitated from lysates of transfected cells and was absent from lysates of mock transfected controls (Fig. 4, lanes 1 and 2). The protein was stable for approximately 4 h and declined thereafter (Fig. 4,  lanes 2-5). The amounts of [ 35 S]methionine-labeled protein at each time point were used to estimate the half-life of the receptor which was approximately 10 h. The time required for the receptor to reach the cell surface cannot exceed the life span of the protein. Thus, absence of surface receptors does not seem to result from slow transport, since detection of receptors was always performed 12-18 h after transfection. The outcomes of these last two experiments indicate that recombinant GABA B receptors do not reach the plasma membrane at detectable levels, and strongly suggest that they are retained in some intermediate compartment of the secretory pathway.
GABA B Receptors Are Retained in the Endoplasmic Reticulum-Our results concerning sensitivity to tunicamycin indicated that recombinant receptors entered the endoplasmic reticulum (ER). To determine whether they were specifically retained in the ER or whether they trafficked to successive compartments, N-GABA B and an ER resident protein were detected simultaneously in COS cells. Previous studies in our laboratory have conclusively demonstrated that the GABA A ␤2 subunit is retained in the ER when expressed alone (14). To determine if N-GABA B and the ␤2 subunit were retained in the same subcellular compartment, COS cells were co-transfected with N-GABA B and a ␤2 subunit tagged in the extracellular amino-terminal domain with the FLAG epitope (FLAG-␤2). As 2 C. Connolly, unpublished results. Values for mock transfected control cells were considered background and therefore subtracted before conversion to percentages. Each bar represents the average of three independent binding assays; standard deviations are indicated. B, the experiment described above was performed under identical conditions in HEK 293 cells. expected, neither N-GABA B nor FLAG-␤2 reached the cell surface in intact cells (not shown). Interestingly, upon detergent permeabilization both proteins colocalized intracellularly in a reticular structure (Fig. 5A).
Taking advantage of the presence of N-linked oligosaccharides in GABA B receptors, the subcellular localization of recombinant receptors was confirmed using an endoglycosidase H sensitivity assay. Glycosylated proteins processed in the ER contain a characteristic high mannose oligosaccharide that is sensitive to digestion by the highly specific enzyme endoglycosidase H (Endo H). In the Golgi stack the oligosaccharide is modified and becomes resistant to Endo H digestion. Sensitivity to Endo H is, therefore, a good indicator of ER localization for N-linked glycosylated proteins (27). To address the sensitivity of recombinant GABA B receptors to Endo H digestion, COS cells were transfected with N-GABA B and metabolically labeled with [ 35 S]methionine. Immunoprecipitated receptors were digested with Endo H and analyzed by SDS-PAGE. A predominant form of GABA B was immunoprecipitated from lysates derived from N-GABA B transfected cells and was absent from lysates derived from mock transfected controls (Fig.  5B, lanes 1 and 2). Upon Endo H treatment the electrophoretic mobility of the majority of the immunoprecipitated receptor was significantly altered (Fig. 5B, lane 3). This decrease in molecular weight indicates that the N-GABA B receptor is sensitive to Endo H digestion. Taken together these observations strongly suggest that N-GABA B enters the secretory pathway and is specifically retained in the ER when expressed in COS cells.

Recombinant GABA B Receptors Fail to Function when Expressed in Ganglion Neurons-
The presence of functional GABA B receptors in SCG neurons (28,29,30) and dorsal root ganglion neurons (DRG) (31) has been reported extensively in recent years (3,32). To examine the activity of recombinant GABA B receptors in ganglion neurons, the inhibition of Ca 2ϩ channel currents in response to the GABA B receptor agonist L-baclofen was studied electrophysiologically. Primary SCG neurons were injected either with a reporter plasmid carrying the GFP or coinjected with GFP and N-GABA B . Upon exposure to L-baclofen, a 10.7 Ϯ 3.0% inhibition of Ca 2ϩ channel currents was observed in GFP injected SCG neurons indicating the presence of endogenous GABA B receptors coupled to Ca 2ϩ channels (Fig. 6, A and B, right). The weak inhibition obtained in these experiments suggests that the number of active endogenous receptors in SCG neurons might be small. In contrast, a 42.8 Ϯ 7% inhibition was obtained in the same neurons when Ca 2ϩ channel current inhibition was achieved through the activation of endogenous ␣ 2 adrenergic receptors upon exposure to noradrenaline. These results are in accordance with previous observations (17) and demonstrate that inhibition of Ca 2ϩ channel currents is not saturated at levels affected by L-baclofen. As expected, in cells expressing recombinant N-GABA B receptors, inhibition of Ca 2ϩ channel currents by NA did not differ significantly from the inhibition observed for GFP injected control neurons (Fig. 6B, left). Interestingly, when cells expressing recombinant untagged (not shown) or N-GABA B receptors were exposed to L-baclofen, Ca 2ϩ channel current inhibition was not significantly increased (2.9 Ϯ 1.4%) (Fig. 6,  A and B, left). These observations are considerably different from others reported for activation recombinant G proteincoupled receptors in sympathetic neurons, where 45-65% inhibitions of Ca 2ϩ channel currents have been observed above the endogenous response (33). These results clearly indicate that recombinant GABA B receptors failed to activate when expressed in SCG neurons. The significance of the lower inhibition of Ca 2ϩ channel current in N-GABA B injected cells is still unclear and probably resulted from high cell to cell variation of endogenous receptor activation in response to agonist.
To evaluate the possibility that inactivity of recombinant GABA B receptors resulted from intracellular retention as in heterologous systems, N-GABA B was microinjected into SCG and DRG neurons and the subcellular distribution of GABA B receptors was examined by immunofluorescence. The subcellular distribution in neurons was similar to that obtained for the three cell lines tested, namely, no cell surface receptors were detected in intact neurons (Fig. 7, A and B, upper panels), and large amounts of the protein were detected intracellularly when cells were permeabilized (Fig. 7, A and B, lower panels). This was observed when cell bodies or neuronal projections were examined (Fig. 7). The pattern of expression did not change when cells were examined after 5 h of injection, when small amounts of receptors were detectable by immunofluorescence, or 5 days after microinjection (not shown). These observations demonstrate the absence of cell surface staining in neurons and suggest that retention does not result from intracellular obstruction, due to overexpression of recombinant GABA B receptors, or to slow transport.

DISCUSSION
The study of the targeting mechanisms of GABA B receptors to the plasma membrane is of fundamental importance for the understanding of GABA-mediated signaling and for the development of novel therapeutic agents that modify the function of GABA B receptors. The results reported in this study suggest that recombinant murine GABA B R1a receptors fail to reach the cell surface when expressed in ganglion neurons, and are specifically retained in the ER when introduced in heterologous systems. Preliminary studies using GABA B R1b, an alternatively spliced receptor that differs from GABA B R1a in the extreme amino-terminal region, have produced similar results (not shown). Thus, exclusion from the plasma membrane does not appear to be restricted to a single class of GABA B receptor. We suggest that the difficulties in coupling recombinant GABA B receptors to known effector systems result mainly from inappropriate targeting. Interestingly, a 30% inhibition of forskolin-stimulated adenylyl cyclase activity was reported for one of the splice variants of recombinant receptors (6). However, this effect should be studied in more detail because the assay was performed in cell membranes, and many of them may have had ER origin.
Several explanations for intracellular retention such as interference of the 9E10 epitope, slow transport to the plasma membrane, or rapid endocytosis are unlikely in view of the results obtained in the present study. Nevertheless, other possibilities arise that are in agreement with our results. First, the receptor may be expressed at the plasma membrane at levels below the detection limits of the techniques used in this study. Zhang et al. (34) have used flow cytometry to demonstrate that only 0.3-1.0% of transiently transfected HEK 293 cells express the Caenorhabditis elegans ODR-10 odorant receptor at the cell surface. This small fraction of properly targeted receptors appears to be sufficient to mediate a significant response to the appropriate odorant. However, the physiological significance of such small numbers remains to be determined. In the case of recombinant GABA B receptors, the failure of them to couple to Ca 2ϩ or K ϩ channels argues against the possibility of a small fraction of receptors mediating the responses to GABA or GABA agonists.
Second, the receptor may require accessory molecules for maturation and targeting to the plasma membrane. This pos- Expression of cell surface molecules was examined in intact cell bodies (left-hand panels) and neuronal projections (right-hand panels) using 9E10 antibody and anti-mouse specific secondary antibodies (upper panels). Intracellular protein was detected in the same cells after permeabilization using 9E10 antibody and a different anti-mouse specific secondary antibody (lower panels). Ϫ and ϩ signs indicate nonpermeabilized and permeabilized cells, respectively. B, N-GABA B injected DRG neurons were analyzed for expression of recombinant GABA B receptors 18 h after injection. Expression of cell surface molecules was examined in intact cell bodies (left-hand panels) and neuronal projections (right-hand panels) using 9E10 antibody and anti-mouse specific secondary antibodies (upper panels). Intracellular protein was detected in the same cells after permeabilization using 9E10 antibody and a different anti-mouse specific secondary antibody (lower panels). The weak signal present in the upper panels results from bleed-through due to high levels of intracellular receptors. sibility needs to be studied in detail. The fact that the receptor is excluded from the cell surface in neurons that normally express GABA B -mediated responses argues against it. However, the putative accessory protein could be present in limiting amounts, could have a very slow turnover, and/or strong association to the receptor, or could regulate the targeting of GABA B receptors in response to activity or developmental cues. For example, several reports indicate that expression of presynaptic and postsynaptic GABA B receptors is extremely well regulated during postnatal ontogenesis and early stages of innervation (35)(36)(37)(38)(39). Our studies concerning expression of GABA B in different cell lines provide ideal heterologous systems to attempt to identify potential accessory molecules. Finally, the recombinant receptor may carry a cis element, such as a retention signal, that requires highly specific and defined conditions to allow the transport of the receptor and its release from the ER to its functional target localization. Alternatively, recombinant receptors may lack essential information for proper targeting and other splice variants, or other receptor subtypes, may represent the physiological receptors in the neuronal types studied here. No information is yet available for the subcellular distribution of native GABA B receptors in neurons. The production of antibodies against the amino-terminal domain of the receptor may help clarify some of these issues by comparing the subcellular distribution, trafficking, and membrane stability of endogenous and recombinant GABA B receptors.
Interestingly, in addition to GABA B receptors, other recombinant G protein-coupled receptors fail to reach the cell surface or couple effectively to effector systems. These include olfactory and calcitonin gene-related peptide receptors (40,41). Olfactory receptors are excluded from the plasma membrane when expressed in heterologous systems but proper targeting is restored if chimeras are made between olfactory receptors and ␤-adrenergic receptors (40). Calcitonin gene-related peptide receptors on the other hand become functional exclusively in the presence of accessory proteins (41). These observations strongly suggest that other molecules are involved in the regulation of G protein-coupled receptor targeting and localization. Last, although metabotropic glutamate receptors couple properly to effector systems, the majority of the protein is retained intracellularly in heterologous systems (42,43). Subcellular distribution of these receptors using amino-terminal antibodies should be examined in greater detail to determine their precise localization and evaluate the relevance of intracellular accumulation.
Future studies concerning endogenous and recombinant GABA B receptors should help clarify the importance of G protein-coupled receptors in synaptic transmission and the relevance of the regulation of receptor subcellular localization and targeting in neurons.