Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

GABA_B receptors mediate metabotropic functions of the inhibitory neurotransmitter GABA via activation of G-proteins of the G_i/o class. They inhibit adenylyl cyclase and modulate the activity of calcium and potassium channels, thereby controlling presynaptic neurotransmitter release and postsynaptic inhibitory potentials (Bowery, 1993). Impaired GABA_B receptor function is associated with pathophysiological effects, including pain, epilepsy, spasticity, and cognitive deficits (Couve et al., 2000), which was recently highlighted in mice deficient for a GABA_B receptor subunit (Prosser et al., 2001; Schuler et al., 2001).

Thus far, two GABA_B receptor subunits, termed GABA_B(1) and GABA_B(2), have been cloned (Kaupmann et al., 1997, 1998; Jones et al., 1998; White et al., 1998; Kuner et al., 1999; Ng et al., 1999). Both subtypes exist in several splice forms, the most prominent being two N-terminal splice forms of GABA_B(1), GABA_B(1a), and GABA_B(1b), whereby the first 147 residues in GABA_B(1a) are replaced with 18 different residues in GABA_B(1b) (Kaupmann et al., 1997). GABA_B receptors belong to the C family of G-protein coupled receptors (GPCRs), which also comprises metabotropic glutamate, Ca²⁺-sensing, and vomeronasal receptors. They all have an unusually large N-terminal, extracellular ligand binding domain that shows similarity to bacterial periplasmic amino acid binding proteins (Couve et al., 2000). Although GPCRs were earlier generally believed to function as monomers, several GPCRs have been recently shown to exist as homo- as well as heterodimers that are functional and play important roles in signaling (Salahpour et al., 2000). C-family GPCRs are likely to function as dimers (Romano et al., 1996; Bai et al., 1999); also, A-family GPCRs (e.g., opioid receptor subtypes or dopamine and somatostatin receptors) have been shown to form homodimers and heterodimers, thereby creating new receptors with altered ligand binding and functional properties (Jordan and Devi, 1999; Rocheville et al., 2000a,b). GABA_B receptors are unique in that complete functional activity depends on the formation of heteromers between GABA_B(1) and GABA_B(2). The monomeric receptor molecules alone are not able to reproduce the pharmacological properties of native GABA_B receptors. Both splice forms GABA_B(1a) and GABA_B(1b) show lower affinity for agonists than native receptors but bind GABA_B receptor antagonists with native affinity. GABA_B(2) has not yet been shown to bind GABA-ergic ligands with appreciable affinity (Jones et al., 1998; Kaupmann et al., 1998). However, coexpression of GABA_B(2) with GABA_B(1) restores high-affinity agonist binding as well as regulation of intracellular effector systems (Jones et al., 1998; Kaupmann et al., 1998; White et al., 1998; Kuner et al., 1999; Ng et al., 1999). G-protein coupling seems to mainly involve intracellular loops of GABA_B(2), but efficient G-protein coupling requires allosteric interactions of GABA_B(1) with GABA_B(2) (Galvez et al., 2001; Robbins et al., 2001).

GABA_B(1) and GABA_B(2) associate, at least partly, through a coiled-coil motif forming a parallel coiled-coil helix found in their respective C termini (Kammerer et al., 1999; Kuner et al., 1999). When expressed alone in human embryonic kidney (HEK) 293 cells or sympathetic neurons, GABA_B(1) is retained in the endoplasmic reticulum (ER) (Couve et al., 1998) due to a C-terminal retention motif [RXR(R)] (Margeta-Mitrovic et al., 2000), which is located C-terminal of its coiled-coil region. Through this coiled-coil interaction, GABA_B(2) masks the retention signal present in GABA_B(1), thus enabling surface expression of functional GABA_B receptors (Margeta-Mitrovic et al., 2000; Calver et al., 2001; Pagano et al., 2001).

Despite the discovery of structural elements for retention and interaction of GABA_B receptor subunits, a detailed understanding of the functional role of GABA_B receptor C termini is still lacking. The apparent interdependence of heterodimerization, surface trafficking and efficient G-protein coupling in vivo made it difficult to assess the exact molecular function of single domains or interactions within the GABA_B receptor dimer. Thus, the aim of our study was to dissect these functions and to examine the role of defined C-terminal domains in heteromerization and signaling. We constructed several C-terminal mutant proteins of GABA_B(1) and GABA_B(2) and analyzed heterodimerization, cell surface expression, G-protein coupling, and constitutive activity in cotransfected HEK293 cells and in sympathetic neurons of the superior cervical ganglion (SCG).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of Mutant Receptor Subunits. Cloning of full-length rat GABA_B(1a) and GABA_B(2) sequences was described in Kuner et al. (1999). The truncated and chimeric versions of GABA_B(1a) were generated by PCR and subcloning of the PCR products into the PstI (bp 2281) and SalI sites of GABA_B(1) in pBlueScript (Stratagene, La Jolla, CA). In GABA_B(1a)CT and GABA_B(1a)924-960, respectively, either the whole C terminus (amino acids R862 to K960) or amino acids R924 to K960 after the coiled-coil region were deleted. In the chimeric constructs GABA_B(1a)_2CT817-940 and GABA_B(1a)_2CT, either the C-terminal sequence of GABA_B(1a) after the coiled-coil sequence (i.e., amino acids R924-K960) or the whole GABA_B(1a) C terminus (amino acids R857-K960) was exchanged by overlap extension PCR for the equivalent sequences in GABA_B(2) (i.e., amino acids D817-L940 or amino acids I744-L940, respectively). GABA_B(1a)cc was constructed by ligating annealed complementary oligonucleotides with appropriate overhangs into the BclI (bp 2576) and NarI (bp 2772) sites of GABA_B(1a), yielding deletion of the coiled-coil encoding sequence (bp 2659-2763 encoding amino acids S887-L921). To allow surface detection of GABA_B(1a) mutants, the sequence for the c-myc epitope was inserted behind the sequence of the predicted signal peptide by overlap extension PCR resulting in the encoded protein sequence ¹⁴LGAEQKLISEEDLNGGA¹⁹ (c-myc epitope, including an additional N printed in bold).

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Schematic outline of GABAB receptor subunit constructs. N, large extracellular N-terminal domain; 7TM, domain including the seven transmembrane helices; cc, predicted coiled-coil region in both the C terminus of GABAB(1) (R1) and GABAB(2) (R2); RSRR, endoplasmic reticulum retention signal; F, Flag-tag (red); C, C terminus, which is depicted as enlarged relative to other domains.

Antibodies. The GABA_B(1)-specific polyclonal antiserum was bought from BD PharMingen (San Diego, CA) and is directed against the last 20 amino acids (941-960) in GABA_B(1a). GABA_B(2) -specific antisera were raised against two synthetic peptides, one derived from amino acids 203 to 222 in the N-terminal extracellular domain (DVQRFSEVRNDLTGVLYGED-amid) and the second corresponding to amino acids 831 to 850 in the C terminus of GABA_B(2) (QELNDILSLGNFTESTDGGK-amid). Peptide synthesis, antibody generation and purification were done by ARK Scientific GmbH Biosystems (Darmstadt, Germany). The titers of the crude sera (as defined in an enzyme-linked immunosorbent assay test as the reciprocal of the serum dilution resulting in an OD₄₉₂ of 0.2 when using 1 µg of free antigen on the solid phase per well) were 66,000 for the serum recognizing the N-terminally located peptide sequence and 300,000 for the serum recognizing the C-terminally located peptide sequence, respectively. For detection of the myc epitope tag, a rabbit polyclonal anti-Myc tag IgG antibody (Upstate Biotechnology, Lake Placid, NY) was used. Peroxidase- and tetramethylrhodamine B isothiocyanate-conjugated secondary antibodies were bought from Jackson ImmunoResearch Laboratories (West Grove, PA).

Cell Culture and Transfection. HEK293 cells were maintained in modified Eagle's medium (MEM; Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum and PenStrep at 37°C with 5% CO₂. Cells were transfected at 70% confluence with 25 µg of total DNA for a 15-cm plate or 2.5 µg of DNA for a 24-well coverslip by calcium phosphate precipitation. Cells were incubated for 2 days before testing expression of recombinant proteins or performing functional assays.

Receptor Binding Assays. Transfected cells were harvested, washed once in phosphate-buffered saline (PBS) or Krebs-Tris buffer (20 mM Tris-HCl, pH 7.4, 118 mM NaCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 4.7 mM KCl, 1.8 mM CaCl₂, and 5.6 mM glucose) and incubated in PBS or Krebs-Tris-buffer with 10 nM [³H]CGP54626A (Tocris Cookson Ltd., Bristol, UK) for 1 h at room temperature (RT). The incubation was terminated by addition of 3 ml of cold washing buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.4) and rapid vacuum filtration over Whatman GF/C filters prewashed with 2 ml of washing buffer. The filters were washed twice with 3 ml of cold washing buffer and counted for tritium using a Tri-Carb scintillation counter (Canberra EuriSys GmbH, Rüsselsheim, Germany). Nonspecific binding was determined either with 10 µM CGP54626 or with 2 mM SCH50911 (Tocris Cookson).

Coimmunoprecipitation and Immunoblot Analysis. Transfected HEK293 cells were collected 48 h after transfection and solubilized over night by gentle rotation at 4°C with 2% Triton X-100 in PBS containing 400 mM NaCl and 10% glycerol as well as the protease inhibitors phenylmethylsulfonyl fluoride (1 mM), leupeptin (10 µg/ml), aprotinin (10 µg/ml), pepstatin A (5 µg/ml), and chymostatin (5 µg/ml). After ultracentrifugation at 100,000g for 1 h at 4°C, the supernatant was diluted 2-fold in PBS and rotated with 1/100 volume (i.e., 8 µl) of preimmuneserum for 1 h at 4°C. For clearance of unspecific immunocomplexes 20 µl of protein A-Sepharose (Sigma, St. Louis, MO) were added and incubation continued for 1 h. Then, the supernatant of the protein A-Sepharose beads was rotated with 1/100 volume (i.e., 8 µl) of the GABA_B(2) -C terminus-specific anti-peptide antiserum for 1 h at 4°C. Immunocomplexes were precipitated for 1 to 2 h at 4°C again with 20 µl of protein A-Sepharose. The captured immunocomplexes of both the incubation with preimmuneserum and antiserum were washed 4 times with PBS containing 1% Triton X-100 and once with PBS containing 0.5% Triton X-100 before eluting into sample buffer (50 mM Tris-HCl, pH 8.5, 200 mM dithiothreitol, 6% SDS, 10% glycerol, 7 M urea, 0.01% bromphenol blue) at 42°C for 10 min. Samples were resolved through 6% SDS-polyacrylamide gel electrophoresis and electrophoretically transferred to nitrocellulose membranes. After blocking with 5% nonfat dry milk powder in PBS containing 0.2% Tween 20, the nitrocellulose membranes were probed with the GABA_B(1)-specific antiserum, diluted 1:2000 in blocking solution, for 1 h at room temperature, followed by three 5-min washes in PBS containing 0.2% Tween 20 and incubation with horseradish peroxidase-labeled anti-guinea pig antibodies. Bound antibody was visualized by enhanced chemiluminescence (enhanced chemiluminescence; Amersham Biosciences, Freiburg, Germany).

Immunocytochemistry. Cells were seeded and grown on sterile glass coverslips coated with either fibronectin (10 µg/ml) or poly-D-lysine (5 µg/ml) and transfected with various receptor mutant constructs 2 days before immunocytochemical analysis. For analyzing receptor expression, cells were washed twice with PBS, fixed on ice for 10 min in 2% paraformaldehyde/0.2% glutaraldehyde, incubated twice for 5 min each in 50 mM glycine in PBS, and washed with PBS. For permeabilization, cells were incubated with 0.2% Triton X-100 in PBS for 10 min on ice and blocked for 30 min at RT with 4% normal goat serum (NGS) and 0.2% Triton X-100 in cold PBS. Primary antibody incubations (anti-Myc IgG, 1:200; anti-GABA_B(2) antibodies, 1:100 or 1:200) were in 2% NGS and 0.1% Triton X-100 in cold Tris-buffered saline (TBS; 10 mM Tris-HCl, 100 mM NaCl, pH 7.6) for 2 to 3 h at RT. After three 10 min washes with 1% NGS in TBS at RT, the cells were incubated with donkey-anti-rabbit tetramethylrhodamine B isothiocyanate-conjugated antibody (1:400) in 1.5% NGS in TBS in the dark for 30 min at RT. After several washing steps, cover slips were mounted and analyzed using a confocal microscope (Leica Mikrosysteme Vertrieb GmbH, Bensheim, Germany).

Cellular cAMP Accumulation Assay. Two days before the assay, HEK293 cells plated onto 15-cm plates were cotransfected with various combinations of GABA_B receptor subtype mutants. On the day of the assay, the medium was replaced by MEM medium supplemented with 1 mM 3-isobutyl-1-methylxanthine (Sigma) and the cells were incubated for 15 min at 37°C and 5% CO₂. The cells were harvested and resuspended to 2.5 × 10⁶ cells/ml in Krebs-Tris buffer supplemented with 1 mM 3-isobutyl-1-methylxanthine. These cells (200 µl; 0.5 × 10⁶ cells) were added to tubes containing either no drugs (basal activity), 2 µM forskolin (FSK), or 2 µM FSK plus GABA-ergic agonists [e.g., 100 µM (R)-baclofen (Tocris Cookson)] or antagonists [e.g., 10 µM CGP54626 (Tocris Cookson)] as indicated in the figure legend. After 20 min at 37°C, the tubes were placed on ice and lysed in 167 mM HCl. The lysate was incubated on ice for at least 1 h and, after neutralization with NaOH, was centrifuged at 15,000g to remove cell debris. The cellular production of cAMP was measured in the supernatant using the commercially available cAMP ³H assay system from Amersham Biosciences. Experiments were performed at least twice in duplicate. The significance of differences was evaluated using unpaired, two-tailed Student's t-tests.

Coupling to GIRKs. Coexpression of GIRK1 and GIRK2 with GABA_B(1a) and GABA_B(2) wild-type or mutant constructs was as described by Kuner et al. (1999). Transfected cells on coverslips were transferred to the stage of an inverted microscope and were continuously perfused with a solution containing 115 mM NaCl, 25 mM KCl, 1.8 mM CaCl₂, 1 mM MgCl₂, 5 mM HEPES, and 10 mM glucose, pH 7.25, adjusted with NaOH. The whole-cell configuration of the patch-clamp technique was applied at RT with pipette solutions containing 115 mM K-gluconate, 20 mM KCl, 10 mM HEPES, 5 mM EGTA, 4 mM MgATP, and 0.3 mM NaGTP, pH 7.25, adjusted with KOH. GABA (100 µM) or (R)-baclofen (50 µM) was dissolved in the extracellular solution and was applied steadily to single cells by a double-barreled application pipette. The inward rectifier currents were activated by voltage ramp generation. For analysis, GIRK-currents in the absence and presence of agonist were compared at 135 mV. Values are expressed as mean ± S.E.M. and p-values represent the result of independent two-tailed t-tests.

Ca²⁺-Channel Current Recordings. Superior cervical ganglion (SCG) neurons were isolated and injected with cDNA as described previously (Ikeda, 1996; Ikeda, 1997). Briefly, SCG neurons were enzymatically dissociated from adult rats and then plated onto polystyrene culture dishes (35 mm), coated with poly-L-lysine, and maintained in a humidified atmosphere of 95% air/5% CO₂ at 37°C. GABA_B receptor subunits were expressed by intranuclear microinjection (0.1 µg/µl per cDNA construct). Electrophysiological recordings were made within 24 h after injection of vectors. Injected neurons were identified by fluorescence from coexpressed jellyfish green fluorescent protein (enhanced green fluorescent protein; BD Biosciences Clontech, Palo Alto, CA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Interactions between Coiled-Coil Deletion Mutants of GABA_B(1) and GABA_B(2)

To assess the role of C-terminal domains of GABAB_(1a) and GABA_B(2) for interaction, trafficking, and function of the GABA_B receptor, we generated a series of deletion mutants (Fig. 1). These include specific deletions of the 35 and 32 amino acids in GABA_B(1a) and GABA_B(2), respectively, which have been shown to mediate a coiled-coil interaction (Kuner et al., 1999; GABA_B(1a)cc and GABA_B(2)cc). To test the importance of the C-terminal coiled-coil interaction for the integrity of the GABA_B receptor, we performed coimmunoprecipitations using full-length proteins. Combinations of wild-type and coiled-coil deletion constructs were cotransfected into HEK293 cells and GABA_B(2) or GABA_B(2)cc was immunoprecipitated from Triton X-100 extracts with preimmuneserum (PIS) followed by an antiserum (AS) directed against the C terminus of GABA_B(2). The immunoprecipitated protein was blotted and analyzed for coprecipitation of GABA_B(1a) or GABA_B(1a)cc (Fig. 2A). GABA_B(2) not only precipitated GABA_B(1a) but also GABA_B(1a)cc. Furthermore, GABA_B(2)cc was also able to precipitate GABA_B(1a) and GABA_B(1a)cc. In all cases, the appropriate preimmuneserum did not lead to precipitation of GABA_B(1a) or GABA_B(1a)cc (Fig. 2A, PIS). Coimmunoprecipitation of GABA_B receptor wild-type proteins was most efficient (Fig. 2A, AS). We also analyzed the interaction of the soluble C-terminal fragments of GABA_B(1) and GABA_B(2). In both yeast two-hybrid assays and glutathione S-transferase pull-down assays, deletion of the coiled-coil region in one of the fragments abolished the interaction (not shown). These data suggest that interaction domains outside the cytoplasmic C termini of GABA_B(1) and GABA_B(2), in addition to the coiled-coil domains within the cytoplasmic C termini, mediate assembly of GABA_B receptor heteromers.

View larger version (86K): [in this window] [in a new window]   Fig. 2.   A, coimmunoprecipitation of GABAB(1a) (R1a) and GABAB(2) (R2) coiled-coil deletion mutants. Extracts from HEK293 cells cotransfected with the indicated constructs were immunoprecipitated with the GABAB(2) -specific antiserum recognizing the C terminus (AS) or the appropriate preimmuneserum (PIS). The immunoprecipitates were immunoblotted for GABAB(1) (C, SDS-cell extract; S, solubilized protein fraction used for immunoprecipitation). B to D, surface trafficking of mutated GABAB receptor subunits. HEK293 cells grown on glass coverslips were transfected with constructs encoding either wild-type or mutant forms of GABAB(1a) (B), GABAB(2) (C), or both (D). All GABAB(1a) constructs were tagged N-terminally with a c-myc epitope recognized by a polyclonal anti-myc antibody (mycR1a). Two days after transfection, the cells were examined by immunofluorescence using the anti-myc antibody or a GABAB(2)-specific polyclonal antiserum directed against the N terminus without [surface staining (ST)] or with permeabilization (P). The bottom panels in B and C are controls (R1a, nontagged wild-type GABAB(1a) ; PIS, preimmuneserum). D, insets, immunofluorescence of the soluble Flag-tagged GABAB(2) constructs using the GABAB(2)C terminus-specific antiserum. Immunofluorescence analysis was done on at least three independent transfections for all combinations shown with consistent results. Scale bar, 10 µm.

C-Terminal Determinants of GABA_B Receptor Surface Expression

In the absence of GABA_B(2), GABA_B(1a) is retained in the ER (Couve et al., 1998). However, GABA_B(1a) efficiently reaches the cell surface after coexpression of GABA_B(2) (White et al., 1998; Margeta-Mitrovic et al., 2000).

We tested the relevance of C-terminal stretches of the two subunits for GABA_B receptor trafficking. The plasma membrane targeting of GABA_B(1a) and GABA_B(2) mutants expressed in HEK293 cells was examined by immunocytochemistry using antisera directed against extracellular N-terminal epitopes. This enabled staining of receptors targeted to the cell membrane in nonpermeabilized cells [surface staining (ST)] compared with those present at the membrane and in the cell interior in detergent permeabilized cells (P) (Fig. 2, B-D).

All GABA_B(1a) deletion and chimeric mutants were expressed to a similar extent and, except for the full-length GABA_B(1a) protein, were all found to be located at the plasma membrane (Fig. 2B). All GABA_B(2) mutants were expressed at the cell surface, but in GABA_B(2)CT transfections, comparably fewer cells were found to be surface stained and staining was less intense (Fig. 2C). The total expression of GABA_B(2)748-780 and GABA_B(2)CT+cc seemed to be lower than that of the remaining GABA_B(2) mutants (Western blot data not shown). In cotransfected HEK293 cells, all GABA_B(2) mutants which still possessed the coiled-coil region (i.e., GABA_B(2)820-940, GABA_B(2)_1CT922-960, GABA_B(2)748-780, and GABA_B(2)CT+cc) were able to traffick GABA_B(1a) to the cell surface; GABA_B(2)820-940 was the most efficient (Fig. 2D and Table 1). Interestingly, GABA_B(2) subunits lacking the coiled-coil domain were able to traffick GABA_B(1a) to the cell surface as well, yet to a smaller extent. Even in GABA_B(1a) + GABA_B(2)CT cotransfections, a few cells per dish showed surface staining (Fig. 2D). The observed plasma membrane targeting of GABA_B(1a) by GABA_B(2)cc and GABA_B(2)CT may result from both additional interaction sites between GABA_B(1) and GABA_B(2) located outside their C termini and the strong overexpression of the two membrane proteins in transiently transfected cells. Furthermore, the C terminus of GABA_B(2) expressed as a cytoplasmic protein (Flag-GABA_B(2)CT) prevented retention of GABA_B(1a), whereas the C terminus of GABA_B(2) lacking the coiled-coil sequence (Flag-GABA_B(2)CTcc) did not (Fig. 2D). These data substantiate recent reports on a role for the coiled-coil interaction in masking the retention signal in GABA_B(1) (Calver et al., 2001; Margeta-Mitrovic et al., 2000; Pagano et al., 2001).

A new aspect of the molecular determinants of GABA_B(1) retention emerged when studying surface trafficking of two of our mutants. GABA_B(1a)cc and GABA_B(2)_1CT922-960 have an intact ER retention signal RSRR but are still very efficiently transported to the cell surface (Fig. 2, B and C). In the construct GABA_B(1a)cc, the 35 codons encoding the putative coiled-coil sequence, directly upstream of the sequence for the retention signal, have been deleted. In GABA_B(2)_1CT922-960, the C terminus of GABA_B(1a) distal of the coiled-coil region but including the retention signal was fused to the end of the coiled-coil region in GABA_B(2) (Fig. 1). These data suggest that sequences upstream of the RSRR signal are required for intracellular retention of the receptor.

The GABA_B Receptor C Termini Are Not Necessary for Agonist-Mediated Effector Coupling

Coupling to Adenylyl Cyclase. We next examined whether C-terminal domains in GABA_B(1) and GABA_B(2) play a role in effector coupling. HEK293 cells were transiently cotransfected with combinations of wild-type or deletion mutants of the GABA_B receptor subunits, and the effect of receptor activation on FSK-induced intracellular cAMP accumulation was measured. In all experiments, protein levels of GABA_B(1a) and GABA_B(2) mutants were controlled by Western blotting (data not shown) and [³H]CGP54626 binding. Binding-active protein levels of GABA_B(1a) and derived mutants in cotransfected HEK293 cells were comparable with 0.5 to 1.5 × 10⁶ receptors/cell.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Coupling of GABAB(1a) and GABAB(2) coiled-coil deletion mutants to effector proteins. A, inhibition of adenylyl cyclase. HEK293 cells were cotransfected with the indicated constructs. Two days later the influence of 500 µM (R)-baclofen (R-Bac) or 500 µM GABA in the absence or presence of 2 mM SCH50911 (SCH) on cAMP levels induced by 2 µM FSK was measured as outlined under Materials and Methods. "Basal" indicates cAMP levels in the absence of any drug. FSK-induced cAMP levels were set to 100%. The bar graph shows the percentage of FSK-stimulated cAMP production under the indicated condition. Error bars indicate the S.D. resulting from at least three independent transfections. The asterisks (*p < 0.01) refer to the effect of R-Bac and GABA on FSK-induced cAMP levels. B, GIRK-currents. Representative whole-cell recordings are shown from transfected HEK293 cells expressing functional GABAB(1a) + GABAB(2) receptors and GIRK1 and GIRK2. GIRK-currents (Control), which are barium sensitive (Ba2+), were activated by 50 µM R-Bac. The voltage ramp used for current activation is shown above the current-voltage relationship. The currents represent averages of five current traces. Arrowhead, membrane potential (135 mV) at which the currents were analyzed. The bar graph compares the GIRK-current increases in the presence of different GABAB deletion mutants (see A) caused by 50 µM R-Bac or 100 µM GABA. Error bars indicate mean ± S.E.M. (n, number of cells recorded). *, p < 0.01 compared with wild-type GABAB receptors. C, calcium channel currents. Representative whole-cell recordings are shown from rat SCG neurons heterologously expressing GABAB(1a) + GABAB(2) in the absence (Control) or presence of 300 µMbaclofen (Bac). Currents were evoked with the double pulse voltage protocol illustrated. Current amplitude was determined 10 ms after the onset of the first test pulse (arrowhead). Horizontal and vertical calibration bars represent 20 ms and 0.5 nA, respectively. The bar graph indicates inhibition of calcium current in the presence of 300 µM Bac after expression of the indicated GABAB combination (see inset in A) in SCG neurons. Error bars indicate mean ± S.E.M. (n, number of neurons recorded). *, p < 0.05 for double-injected neurons compared with GABAB(2)-injected neurons; , p < 0.05 compared with neurons injected with wild-type receptors; #, p < 0.05 for GABAB(2)-injected neurons compared with control cells.

View larger version (61K): [in this window] [in a new window]   Fig. 4.   Functional characterization of GABAB receptors consisting of GABAB(1a) (A) or GABAB(1a)CT (B) and GABAB(2) mutants. cAMP accumulation was measured in HEK293 cells cotransfected with the indicated constructs as outlined under Materials and Methods. cAMP production was stimulated using 2 µM FSK and inhibited with 100 µM R-baclofen (R-Bac). R-Bac was antagonized using either 2 mM SCH50911 or 10 µM CGP54626 (Antagonist). Results are presented as described in the legend of Fig. 3A and represent at least two independent experiments. Inhibition was significant (p < 0.01) for all cotransfected combinations as opposed to single GABAB(1a) or GABAB(1a)CT transfections. GIRK-activation and inhibition of Ca2+ channels was done as outlined in the legend to Fig. 3. Asterisks [*, p < 0.01 (GIRKs) and *, p < 0.05 (cAMP level, Ca2+-channels)] refer to reduced effector coupling compared with wild-type.

Coupling to K⁺ Currents. To investigate GABA_B receptor regulation of potassium currents, we reconstituted GABA_B receptor heterodimers of wild-type and/or mutated GABA_B receptor subunits with GIRK1 and GIRK2 in HEK293 cells. Baclofen (or GABA) application to HEK293 cells led to robust increases in potassium conductance through GIRK-activation in the presence of all tested combinations by 57 to 150% (Figs. 3B and 4), except for GABA_B(1a) + GABA_B(2)cc (Fig. 3B), GABA_B(1a) + GABA_B(2)CT, and GABA_B(1a)CT + GABA_B(2)CT (Fig. 4 and Table 1). The smaller (p < 0.01) conductance increases compared with wild-type may have resulted from reduced surface expression, as shown for GABA_B(2)CT when expressed alone (Fig. 2C) or for GABA_B(1a) in the combinations mentioned above (Fig. 2D). However, 6 of 25 cells transfected with GABA_B(1a) + GABA_B(2)CT and two of nine cells transfected with GABA_B(1a)CT + GABA_B(2)CT showed GIRK-current activation comparable with other deletion mutants [68.1 ± 20% (n = 6) and 49.4% (n = 2)]. This is consistent with the observation that only a small fraction of cells transfected with GABA_B(1a) + GABA_B(2)CT expresses GABA_B(1a) on the cell surface. Thus, GIRK-activation is also possible in the absence of GABA_B receptor C termini.

Coupling to Ca²⁺ Currents. To investigate GABA-ergic inhibition of neuronal calcium channel currents, we microinjected plasmids encoding wild-type and/or mutated GABA_B receptor subunits into the nucleus of SCG neurons. In SCG neurons that endogenously express GABA_B(1a) and calcium channels, a significant (p < 0.05) baclofen-mediated inhibition of calcium conductance was obtained only when GABA_B(2) was injected (Fig. 3C; see also Filippov et al., 2000). An increased (p < 0.05) calcium current inhibition was observed when coinjecting GABA_B(1a) + GABA_B(2), GABA_B(1a)cc + GABA_B(2), and GABA_B(1a)cc + GABA_B(2)cc (Fig. 3C). Similar to the results in GIRK channel coupling in HEK293 cells, injection of GABA_B(1a) + GABA_B(2)cc (Fig. 3C) or GABA_B(1a) + GABA_B(2)CT (Fig. 4A) resulted in a smaller (p < 0.05) calcium current inhibition in SCG neurons (see Discussion). Thus, the cytoplasmic C-terminal domains of GABA_B(1) and GABA_B(2) do not seem to be involved in the agonist-activated G-protein coupling of GABA_B receptors in either HEK293 cells or neurons.

GABA_B Receptor Antagonists Increase cAMP Levels in Cotransfected HEK293 Cells

When performing cAMP assays, we observed that antagonists not only block GABA-ergic agonist-mediated cAMP decrease but also increase cAMP levels above those induced with FSK alone (Fig. 4). We therefore looked at the direct effect of the antagonists SCH50911 and CGP54626 on FSK-induced cAMP levels and found that both antagonists increased cAMP accumulation up to 100% (Fig. 5A). The EC₅₀ values of 1.67 ± 0.51 µM for SCH50911 (n = 4) and 6.7 ± 1.1 nM (n = 3) for CGP54626 were greater than their dissociation constants [IC₅₀ = 0.43 ± 0.04 µM for SCH50911 (Kaupmann et al., 1998), K_D = 1.3 ± 0.1 nM for CGP54626; data not shown]. Treating transfected HEK293 cells with 100 ng/ml pertussis toxin, which ADP-ribosylates and inactivates G_i and G_o, for 16 h before assay eliminated both the G_i-mediated inhibition of FSK-stimulated cAMP accumulation by GABA or (R)-baclofen (not shown) as well as the increase of FSK-stimulated cAMP accumulation by SCH50911 or CGP54626 (Fig. 5C). These antagonists also increased basal cAMP levels up to 60% (not shown). However, basal cAMP levels are near the detection limit of our assay, so dose-response curves were not measured.

View larger version (36K): [in this window] [in a new window]   Fig. 5.   Inverse agonism of GABAB receptor antagonists. A, HEK293 cells were cotransfected with the indicated constructs and FSK-stimulated cAMP production in the presence of increasing concentrations of the GABAB receptor antagonists SCH50911 (filled symbols) and CGP54626 (open symbols), respectively, was determined as outlined under Materials and Methods. cAMP production in the absence of ligand was set to 100% and is reported as a function of the antagonist concentration. Data points were fitted to the four-parameter logistic function using the program Origin (Microcal Software, Inc., Northampton, MA). The data shown are means of duplicates from one of three independent experiments. Expression levels and fitted parameters for this experiment are as follows: GABAB(1a) + GABAB(2): 900,000 [3H]CGP54626 sites/cell, pEC50(SCH50911) = 5.83 ± 0.09, pEC50(CGP54626) = 8.21 ± 0.03; GABAB(1a)cc + GABAB(2): 730,000 [3H]CGP54626 sites/cell, pEC50(SCH50911) = 6.67 ± 0.22, pEC50 (CGP54626) = 8.63 ± 0.11; GABAB(1a) + GABAB(2)cc: 920,000 [3H]CGP54626 sites/cell. B, Western blot analysis of the cotransfected cells measured in A. Cell lysate of 100,000 cells was loaded in each lane: 1, GABAB(1a)cc + GABAB(2) ; 2, GABAB(1a) + GABAB(2)cc; 3, GABAB(1a) + GABAB(2). The left three lanes were probed with a GABAB(1) -specific antiserum and the right three lanes with the GABAB(2) C terminus-specific antibody. C, the effect of 100 µM GABA, 1 mM SCH50911 (SCH) and 10 µM CGP54626 (CGP) on 2 µM FSK-stimulated cAMP levels was measured in HEK293 cells coexpressing GABAB(1a) and GABAB(2) with () and without () prior treatment with 100 ng/ml pertussis toxin for 16 h. D, HEK293 cells were cotransfected with the expression plasmids for GABAB(1a) (R1) or GABAB(1a)CT (R1CT) in combination with those for GABAB(2) (R2) andthe indicated deletion mutants. cAMP production was measured in the presence of 2 µM FSK and either 100 µM GABA or 2 mM SCH50911 or 10 µM CGP54626, respectively. FSK-induced cAMP levels were set to 100% (horizontal line). Data in C and D are shown as percentage of FSK-induced cAMP levels under the indicated conditions and represent mean ± S.D. of at least two independent experiments. *, p < 0.05, antagonist effect on FSK-induced cAMP levels; , p < 0.05, difference between GABAB(1a) and equivalent GABAB(1a)CT combinations.

Interestingly, both antagonists had no effect on FSK-induced cAMP levels in HEK293 cells cotransfected with GABA_B(1a) and GABA_B(2)cc (Fig. 5, A and D). In HEK293 cells expressing GABA_B(1a)cc together with GABA_B(2), the EC₅₀ values with 216 nM for SCH50911 and 2.4 nM for CGP54626 were near the dissociation constant values of these ligands for the wild-type receptor but the efficacy (maximal increase of FSK-stimulated cAMP accumulation) was about 4-fold lower (Fig. 5A). Protein levels of the various receptor subunits were similar in all three combinations as detected by Western blot (Fig. 5B) and [3H]CGP54626 binding. Because deletion of the coiled-coil region in GABA_B(2) abolished this inverse agonist effect, we analyzed various combinations of GABA_B(1a) and GABA_B(1a)CT with GABA_B(2) deletion mutants (Fig. 5D and Table 1). HEK293 cells cotransfected with GABA_B(1a) and GABA_B(2) deletions that lack the coiled-coil region (i.e., GABA_B(2)cc and GABA_B(2)CT) did not show an increase in FSK-stimulated cAMP levels, whereas cotransfection with GABA_B(2)820-940, GABA_B(2)748-780, or GABA_B(2)CT+cc showed an effect similar to that of the wild-type. If GABA_B(1a) was replaced by GABA_B(1a)CT in cotransfected HEK293 cells, the efficacy decreased to values similar to those with GABA_B(1)cc + GABA_B(2) (Fig. 5, A and D). In contrast to GABA_B(1a), however, GABA_B(1a)CT together with GABA_B(2)748-780 or GABA_B(2)CT+cc showed no significant inverse agonism (Fig. 5D and Table 1). In these two mutants, the coiled-coil region is located closer to the end of transmembrane domain 7 than in the wild-type GABA_B(2) protein (see Fig. 1) which might affect receptor conformation. Thus, the C termini of both GABA_B(1) and GABA_B(2), and especially their ability to form a coiled coil, are important for allowing the receptor to adopt an activated conformation without an agonist bound.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Physical Interaction And Trafficking of GABA_B Receptor Subunits. Despite recent advances in our understanding of GABA_B receptor heteromerization, the exact molecular determinants mediating the interaction between the two subunits are still unknown. Our data indicate that heteromerization and trafficking are strongly based on the coiled-coil interaction between the C termini of GABA_B(1) and GABA_B(2). At the same time, we found that deletion of the coiled-coil domains in GABA_B(1) and GABA_B(2) subunits did not completely abolish their physical interaction. Furthermore, GABA_B(2) mutants lacking the coiled-coil domain led to significant surface expression of GABA_B(1) and to a decreased, but significant amount of functional GABA_B receptor heteromers (Fig. 3 and 4, GABA_B(1a) + GABA_B(2)cc and GABA_B(1a) + GABA_B(2)CT). These data are in accordance with recent reports, which were published as this study was in progress (Calver et al., 2001; Pagano et al., 2001) and suggest that besides the coiled-coil regions, additional interaction sites outside the C termini are responsible for the coassembly of the two GABA_B receptor subunits. At least three observations speak in favor of an N-terminal interaction without excluding intramembranous interactions: GABA_B(1e), a splice variant encoding an N-terminal fragment of GABA_B(1), can compete with full-length GABA_B(1) for binding to GABA_B(2) (Schwarz et al., 2000); GABA_B(2) contributes to high-affinity agonist binding (Jones et al., 1998; Kaupmann et al., 1998); and the extracellular domains of both GABA_B(1) and GABA_B(2) are involved in the mechanisms of receptor activation (Galvez et al., 2001). Indeed, other C family GPCRs form intermolecular disulfide bridges between their N-terminal extracellular domains and also show noncovalent interactions to support dimerization (Romano et al., 2001; Zhang et al., 2001).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   The functionality of the retention signal RSRR depends on surrounding sequences. The sequences of the coiled-coil region of wild-type GABAB(1) and mutants are shown. In the chimeric constructs, sequences derived from GABAB(2) are italic. The retention signal reported in Margeta-Mitrovic et al. (2000) is bold. The functional retention sequence deduced from a combination of results described in this study and in constructs characterized by Pagano et al. (2001) is underlined. +, surface expression.

Molecular Determinants for Coupling of GABA_B Receptors with Effector Systems. Type C metabotropic receptors, such as the metabotropic glutamate receptors, couple to G-proteins via residues in their second and third intracellular loops (Gomeza et al., 1996). These residues are not conserved at parallel positions in GABA_B(1), thereby raising the possibility that other loci within the intracellular loops or in the intracellular C termini of GABA_B receptors mediate G-protein coupling. We therefore probed the ability of C-terminal mutants of GABA_B(1) and of GABA_B(2) to couple with each of their three main effector systems (i.e., adenylyl cyclase, GIRKs, and calcium channels).

Constitutive Activity of GABA_B Receptors. Our data indicate that CGP54626 and SCH50911, which are well known as GABA_B receptor antagonists, are also able to function as inverse agonists, because in GABA_B receptor-expressing HEK293 cells, both ligands increase FSK-stimulated cAMP levels up to 2-fold. Being sensitive to treatment with pertussis toxin, this effect depends on a functional G-protein of the G_i/o class. Our observations imply that CGP54626 and SCH50911 shift the GABA_B receptor to a conformation that is less accessible to G_i, thus allowing higher activation of adenylyl cyclase. These results indicate that the GABA_B receptor has the potential for constitutive activity. Galvez et al. (2001) observed a slight increase in the basal level of inositol phosphate formation in HEK293 cells expressing wild-type GABA_B receptors. A large increase in the basal inositol phosphate formation was detectable in cells expressing GABA_B receptor dimers consisting of GABA_B(1a) and a chimeric GABA_B(2) subunit, in which the extracellular domain has been replaced by that of GABA_B(1a). This large increase in inositol phosphate could be partly inhibited by GABA and this inhibition could be antagonized by CGP64213.

Conclusions. Interaction of the two subunits of the GABA_B receptor occurs not only through the coiled-coil domains but also through low affinity interaction sites, probably located within N-terminal extracellular domains and/or membrane spanning regions. Whereas the C termini of the two subunits are not necessary for functional heterodimerization, their coiled-coil domains are involved in stabilizing the active conformation of the heterodimer. Furthermore, efficient trafficking of GABA_B(1) to the cell surface requires interaction via the coiled-coil domain. Notably, the retention signal of GABA_B(1) requires a part of its coiled-coil sequence.