Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The benzodiazepines exert their pharmacological actions by interacting with the benzodiazepine binding sites associated to the -aminobutyric acid type A (GABA_A) receptor complex. The GABA_A receptor complex is a ligand-gating chloride channel that mediates the inhibitory effects of the neurotransmitter GABA .

The GABA_A receptor complex is a heteromeric protein, probably pentameric, composed of different subunits from five different families (, , , , and ; Sieghart, 1995; Hedblom and Kirkness, 1997). Most of these families are also heterogeneous, and different members have been identified (1-6, 1-3, 1-3). However, the exact molecular composition of the native GABA_A receptor complex is still controversial. In recombinant receptors, the presence of a minimum of three different subunits, , , and , are needed to confer all pharmacological properties of native GABA_A receptor complexes (Pritchett et al., 1989b). The stoichiometry of the native GABA_A receptors is also unknown. Evidence in native and recombinant receptors indicates the existence of two , two , and one  subunit in a single receptor complex (Pollard et al., 1995; Araujo et al., 1996; Tretter et al., 1997; Gorrie et al., 1997; Jechlinger et al., 1998). However, other stoichiometries have also been proposed (Backus et al., 1993; Benke et al., 1996). On the other hand, it has been suggested that the benzodiazepine binding sites of the GABA_A receptor are formed at the interface of the  and  subunits (Smith and Olsen, 1995; Sigel and Buhr, 1997) and it is known that the  subunits determine the benzodiazepine binding properties. In this sense, three different benzodiazepine binding sites have been identified in recombinant receptors (Pritchett et al., 1989a; Pritchett and Seeburg, 1990) based on the differential affinity for the imidazopyridine zolpidem. The coexpression of 1, 1/2/3, and 2 determines high-affinity binding sites for zolpidem, whereas the presence of 2, 3, or 5 determines medium- or low-affinity binding sites, respectively (Pritchett et al., 1989a; Pritchett and Seeburg, 1990). Similarly, in isolated rat cerebral membranes or in rat brain slices, three benzodiazepine binding sites (type I, type II_M, and type II_L) were observed based on the affinity for zolpidem (Ruano et al., 1992; Benavides et al., 1993). For zolpidem, the type I benzodiazepine binding sites display high affinity (K_d/i, 10-20 nM; 1 binding properties), the type II_M sites display medium affinity (K_i, 200-300 nM; 2 and/or 3 binding properties), and the type II_L benzodiazepine binding sites are characterized by low-affinity binding sites (K_i, 1-3 µM; 5 binding properties; Ruano et al., 1992; Benavides et al., 1993). Furthermore, the regional distribution of the different benzodiazepine binding sites and the expression of the different  subunits are also similar (Ruano et al., 1992; Benavides et al., 1993; Fritschy and Mohler, 1995).

Two different  subunits can coexist in a single GABA_A receptor complex from whole rat brain (Duggan et al., 1991; Pollard et al., 1993; Mertens et al., 1993; Khan et al., 1994b). In this sense, we have recently identified and characterized the pharmacological properties of 1-3-containing GABA_A receptors from rat cortex (Araujo et al., 1996). The 1 subunit coassembled with the 3 subunit displayed predominantly type I benzodiazepine binding sites (Araujo et al., 1996). However, it has also been reported that the association between the 1 and 6 subunits from rat cerebellum displayed predominantly 6 binding properties (Pollard et al., 1995). Thus, the benzodiazepine binding properties of the native GABA_A receptor complexes are not exclusively determined by the presence of a particular  subunit.

In the present work, we investigated the association of the 1 and 5 subunits and the benzodiazepine binding properties of immunopurified GABA_A receptors. The 5 subunits and, by consequence, the type II_L binding sites, are a relatively minor component of the total population of native GABA_A receptors that are expressed in restricted brain areas, such as hippocampus and olfactory bulb. Therefore, we used subunit specific antibodies and specific ligands for immunopurification and pharmacological experiments of the 5-containing receptors from rat hippocampus.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. [³H]Zolpidem (58.0 Ci/mmol), [³H]flumazenil (FMZ; 75.2 Ci/mmol), [³H]Ro15-4513 (24.1 Ci/mmol), and [³H]muscimol (19.1 Ci/mmol) were purchased from New England Nuclear (Boston, MA). [³H]L-655,708 (83.0 Ci/mmol) was purchased from Amersham (Arlington Heights, IL). Zolpidem was synthesized in the preclinical research department of Synthélabo Recherche (Paris, France). Other benzodiazepines were purchased from Hoffmann-La Roche (Nutley, NJ).

Antibodies. The anti-5 antibody was prepared using a synthetic peptide from the COO terminus of the rat 5 subunit (amino acids 424-433; PVIKGATSPK), coupled to keyhole limpet hemocyanin, via an extra cysteine located at the N-terminal. The immunizations were done as described previously (Araujo et al., 1996).

Membrane Preparation and Receptor Solubilization. Membranes from 3-month-old Wistar rat hippocampi were prepared as described elsewhere (Ruano et al., 1994a,b and references therein) in presence of protease inhibitors; 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 1 mM benzamidine, 50 µg/ml trypsin inhibitor type II-S, and 50 µg/ml bacitracin.

Antibody Purification and Immunoprecipitations. The anti-1 and -5 antibodies were purified through peptide affinity columns. The peptides were coupled to CNBr-activated Sepharose 4B (Amersham Pharmacia). One milliliter of the different antisera (diluted 1/5 in PBS) were recirculated overnight at 4°C in the corresponding column (1 ml). After washing with 150 ml of PBS, the antibodies were eluted with 50 mM glycine-HCl, pH 2.3, and the fractions (0.5 ml) were neutralized by 1 M Tris, pooled, and dialyzed in 1 liter of PBS (overnight at 4°C).

Immunoaffinity Chromatography. The immunoaffinity columns were synthesized as previously described (Araujo et al., 1996). Briefly, 1 to 2 mg of each purified antibody was absorbed to 0.5 ml of protein A-Sepharose. The IgG protein A-Sepharose complex was washed with 40 ml of PBS followed by 5 ml of 0.2 M triethanolamine, pH 8.3. The column was then treated with 1.5 ml of 20 mM dimethyl pimelimidate in 0.2 M triethanolamine, pH 8.3, for 30 min at room temperature. After incubation, the medium was replaced by 1 ml of 0.2 M ethanolamine, pH 8.3, and incubated for 5 min. After coupling, the column was packed and washed, at 10 ml/h, with: 1) 40 ml of PBS; 2) 2 ml of 140 mM NaCl, 1 mM EDTA, 1 mM EGTA, and 50 mM sodium phosphate, pH 11.5; and 3) 20 ml of PBS. The columns were pre-equilibrated with 10 ml of solubilization buffer.

Pharmacological Properties of Immunoprecipitated or -Purified Receptors. Displacement experiments were performed using 1.3 nM [³H]FMZ and 13 or 11 different concentrations of zolpidem (ranging from 0.5 nM to 100 µM) or CL 218,872 (ranging from 5 nM to 100 µM) as described previously (Araujo et al., 1996). Briefly, aliquots of the immunoprecipitated or immunopurified receptors were incubated in 50 mM Tris-HCl, pH 7.4, and 0.05% of Na deoxycholate, 0.05% CHAPS (0.75 ml final volume) with different concentrations of Cl218,872 or zolpidem. The saturation analysis of the anti-5 immunoprecipitated receptors were done using five different concentrations of [³H]muscimol (ranging from 5-60 nM), [³H]FMZ, or [³H]L-655,708 (0.5 to 20 nM). Nonspecific binding was determined in presence of 5 µM clonazepam or 1 mM GABA. The displacement and saturation curves were fitted using LIGAND (Munson and Rodbard, 1980).

Other Methods. Photolabeling, immunoblots, protein determination, and SDS-polyacrylamide gel electrophoresis were done as described elsewhere (Araujo et al., 1996)

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Association between 1 and 5 Subunits in Native GABA_A Receptors from Rat Hippocampus. We first tested the association between 1 and 5 subunits by immunoaffinity chromatography using subunit specific anti-1 and -5 antibodies. The specificity of these antibodies is shown in Fig. 1A (see also Araujo et al., 1996, 1998). For the immunopurification experiments, the solubilized GABA_A receptors (20-30 pmol of [³H]FMZ binding activity) were immunoadsorbed to either anti-1 or -5 immunoaffinity columns, extensively washed and eluted with 2% of SDS. The anti-1 and -5 immunoaffinity columns retained 55.6 ± 6.7 and 22.5 ± 2.5% (n = 3) of the total binding activity applied to the columns, respectively. As shown in Fig. 1B, in the anti-1 immunopurified receptors, anti-1 and -5 antibodies immunoreacted with peptides of 51 kDa (1) and 55 kDa (5), respectively, demonstrating the copurification of these two  subunits. Reciprocally, anti-1 and -5 antibodies immunoreacted with 51 and 55 kDa peptides in the anti-5 immunopurified receptors. Thus, these results demonstrated the association of these two  subunits in the native hippocampal receptor complexes (see also Duggan et al., 1991; Mertens et al., 1993). The association was also quantified by immunodepletion experiments. The results indicated that the 15 combination represented the 20.4 ± 2.6 and 8.6 ± 1.2% (n = 4) of the 5 and 1 populations, respectively.

View larger version (40K): [in this window] [in a new window]   Fig. 1.   Specificity of antibodies (A) and copurification of the 1 and 5 subunits in the same GABAA receptor (B). A, cerebral membranes (75 µg of protein/lane) were subjected to SDS-polyacrylamide gel electrophoresis, blotted, and immunostained with 1 to 5 µg/ml of purified anti-1, -5, or -2 or the mAb 62-3G1 (1/5 dilution; specific for 2/3 subunits). The anti-1 and -5 antibodies immunoreacted with 51- and 55-kDa bands, respectively. In some experiments anti-1 also immunoreacted with an unidentified 100-kDa band. B, GABAA receptors were immunopurified by anti-1 or -5 immunoaffinity columns. After washing, the immunoaffinity columns were treated with SDS and the eluted receptor was analyzed by immunoblot using 5 µg of purified anti-1 or -5 antibodies. For all immunoblots, the affinity-purified antibodies were labeled with digoxigenin. Numbers on the left indicate the Mr of the bands.

Pharmacological Properties of Immunoprecipitated Receptors. The association between 1 and 5 subunits should be reflected in the pharmacological properties of the immunoprecipitated receptors. Thus, we first quantified the immunoprecipitation by anti-1 and -5 antibodies of the total (labeled by 5 nM [³H]FMZ), high-affinity (labeled by 7 nM [³H]zolpidem), and low-affinity (labeled by 5 nM [³H]FMZ plus 10 µM zolpidem) benzodiazepine binding sites from rat hippocampus. As shown in Fig. 2A, the anti-1 and -5 antibodies immunoprecipitated 50.4 ± 4.1 and 20.0 ± 1.2% of the total benzodiazepine binding sites (these values are similar to those obtained by immunoaffinity columns; see above). As expected (see Ruano et al., 1994a,b; Mertens et al., 1993; Sur et al., 1998), anti-1 and -5 immunoprecipitated most (85-95%) of the high- and low-affinity binding sites, respectively. Furthermore, anti-1 also immunoprecipitated a consistent 22.5 ± 7.0% (n = 4) of the low-affinity binding sites. This value is in close agreement to that calculated by immunodepletion experiments (see above). However, no significant immunoprecipitation of the high-affinity binding sites was detected by anti-5 antibody (3.2 ± 5.2%, n = 5; Fig. 2A). Thus, the absence of [³H]zolpidem binding activity immunoprecipitated by anti-5 antibody could be due to the absence of high-affinity binding sites for this ligand, in spite of the association of the 5 subunit with the 1 subunit in the same receptor complex. Alternatively, the absence of high- affinity binding sites immunoprecipitated by anti-5 antibody could be also due to the low proportion of the 15 receptors, over the total 1 population (8.2 ± 1.5%).

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Benzodiazepine binding properties of the anti-1- and -5-immunoprecipitated GABAA receptors from rat hippocampus. A, solubilized GABAA receptor (0.7-0.8 pmol of [3H]FMZ binding activity) was incubated with saturating amounts of anti-1 or -5 antibodies and the immunoprecipitation of the different benzodiazepine binding sites was tested in immunopellets and in supernatant using 5 nM [3H]FMZ (total binding sites; open columns), 7 nM [3H]zolpidem (type I binding sites; right-hatched columns), or 5 nM [3H]FMZ + 10 µM zolpidem (type IIL binding sites; left-hatched columns). Results are expressed as a percentage of each corresponding binding type and are means ± S.D. from five independent experiments. B and C, solubilized GABAA receptor (0.7-1.0 pmol of [3H]FMZ binding) was immunoprecipitated by anti-1 () or -5 () antibodies. The benzodiazepine binding properties were determined in immunopellets by displacement experiments of the [3H]FMZ (1.3 nM) binding with CL218,872 (B) or zolpidem (C). Results are expressed as percentage of control in absence of drug (10,000 or 7,000 cpm for anti-1 and -5, respectively) and are means ± S.D. of three independent experiments. D, fluorographs of [3H]Ro15-4513 photolabeled and immunoprecipitated GABAA receptors from rat hippocampus. Hippocampal membranes were photolabeled with 20 nM [3H]Ro15-4513, solubilized, and immunoprecipitated by anti-1 or -5 antibodies. Results are representative of at least three independent experiments. Fluorographs shown were exposed for 2 (lane 1) or 6 weeks (lanes 2 and 3) .

View this table: [in this window] [in a new window]   TABLE 1 Pharmacological characterization of the anti-1 and -5 immunoprecipitated GABAA receptors from rat hippocampus Solubilized receptor was immunoadsorbed to the different IgG-protein A-Sepharose columns, washed extensively, and aliquots of anti-1 or -5 immunobeads (0.4-0.6 pmol of [3H]FMZ binding activity/tube) were used. Displacement experiments were performed by determining binding activity of 1.2 nM [3H]FMZ and 13 or 10 different concentrations of zolpidem (ranging from 5 1010 to 104M) or CL 218,872 (ranging from 5 109 to 104M), respectively. Displacement curves were fit (LIGAND) to a one-, two-, or three-binding site model. Proportion of different biding sites is given within parenthesis. Results are mean ± S.D. of three experiments.

Pharmacological Properties of 1-5 Double- Immunopurified Receptors. The pharmacological analysis of the immunoprecipitated receptors indicated the existence of pharmacologically active and inactive  subunits coassembled in a single receptor complex. Thus, to confirm this possibility, we further investigated the benzodiazepine binding properties of the 15 containing GABA_A receptors isolated by two sequential immunopurifications in series (see Duggan et al., 1991; Pollard et al., 1995; Araujo et al., 1996). For these experiments, the solubilized receptors were first immunopurified by anti-1 immunoaffinity columns, eluted at high (11.5) pH and the eluted receptor (5-6 pmol of [³H]FMZ binding activity, 21.3 ± 3.1% of the bound receptor, see also Araujo et al., 1996) was immunoadsorbed to anti-5 immunobeads. The GABA_A receptors immunoadsorbed to anti-5 columns were either eluted by SDS and analyzed in Western blots, used in displacement experiments using Cl218,872 or used for radioligand assays.

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Pharmacological properties of the anti-1 and -5 double-immunopurified receptors. GABAA receptors were immunopurified by anti-1 immunoaffinity columns. A and B, eluted receptors were immunoadsorbed to anti-5 immunobeads, washed, and either eluted by SDS, blotted, and immunostained with 2 µg of purified anti-1 or -5 antibodies (A) or immunopellets were used for displacement experiments of the [3H]FMZ (1.3 nM) binding by Cl218,872 (B). C and D, Percentage of 5 nM [3H]FMZ (total benzodiazepine binding sites, open columns), 5 nM [3H]zolpidem (type I, right-hatched columns), 5 nM [3H]FMZ + 10 µM zolpidem (type IIL, left-hatchet columns), and 10 nM [3H]L-655,708 (type IIL, double-hatched columns) was calculated (C), or specific binding activity (open columns) of 10 nM [3H]FMZ, 10 nM [3H]L-655,708, and 40 nM [3H]muscimol or the ratio over the [3H]FMZ binding (hatched columns) were tested (D). The results are means ± S.D. of two (B and D) or three (C) experiments and are expressed as percentage of control in absence of drug (3500 cpm) (B), percentage of immunoprecipitation [initial specific binding activity: 30,000 cpm, 6000 cpm, 4000 cpm, 12000 cpm for [3H]FMZ, [3H]zolpidem, [3H]FMZ + zolpidem, and [3H]L-655,708 binding sites, respectively (C)], or pmoles of [3H]ligand specifically immunoadsorbed to 20 µl of anti-5 immunobeads (D). A typical displacement of anti-1 immunoprecipitated receptors is shown in B for comparison.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

In the present studies we investigated the coexistence and the benzodiazepine binding properties of the 1 and 5 subunits in native GABA_A receptor from rat hippocampus. Evidence has indicated the existence of two homologous or different  subunits coassembled in a single GABA_A receptor complex from whole rat brain (Duggan et al., 1991; Pollard et al., 1993; Mertens et al., 1993), cortex (Araujo et al., 1996), and cerebellum (Khan et al., 1994b; Pollard et al., 1995). Similarly, our results demonstrated the association between 1 and 5 subunits in the hippocampal formation, confirming previous results using whole brain (Mertens et al., 1993).

However, the pharmacological properties of these receptor complexes, containing two different  subunits, were not previously examined in detail (see Mertens et al., 1993). Evidence indicates that the benzodiazepine binding site of the GABA_A receptor complex is formed by the interface between  and  subunits (Smith and Olsen, 1995; Sigel and Buhr, 1997). The presence of 2 subunit, the major expressed  subunit in the hippocampus (not shown, but see Ruano et al., 1994a; Fritschy and Mohler, 1995; Sur et al.,1998), in association with the different  subunits, determines the classical benzodiazepine pharmacology of the GABA_A receptor complex (i.e., 1, with high affinity for Cl218,872 and zolpidem; 2 or 3, with low or medium affinity for Cl218,872 and zolpidem and 5, with low affinity for both compounds; Pritchett et al., 1989a; Pritchett and Seeburg, 1990). Therefore, if both 1 and 5 subunits, coassembled in a single GABA_A receptor complex, display benzodiazepine binding activity, at least two different binding sites should be immunoprecipitated by each antibody. However, our results demonstrated that the 5 subunit is pharmacologically predominant over the 1 subunit, in spite of the coexistence of both 1 and 5 subunits in native GABA_A receptors (Figs. 1B and 3A). This conclusion is supported by 1) the absence of immunoprecipitation of [³H]zolpidem binding sites by anti-5 antibody (Fig. 2A); 2) the absence of high-affinity binding sites for neither Cl218,872 or zolpidem in displacement curves of the anti-5 immunoprecipitated receptors (Fig. 2, B and C and Table 1); 3) the presence of low-affinity binding sites for both Cl218,872 and zolpidem immunoprecipitated by the anti-1 antibody (Fig. 2, A-C and Table 1); and 4) the anti-5 antibody immunoprecipitated a single P55-kDa photolabeled peptide (Fig. 2D). Thus, for the benzodiazepine binding sites, these results, as a whole, demonstrated that 5 is pharmacologically predominant over the 1 and indicate the existence of pharmacologically active and inactive  subunits associated in a single GABA_A receptor complex.

The existence of these pharmacologically active and inactive  subunits was directly proved by the double- immunopurification experiments. The 1-5 double- immunopurified receptor displays exclusively low-affinity binding sites for Cl218,872 and zolpidem (see Fig. 3, B and C and 4A for a summary). Also, [³H]L-655,708 (5- selective ligand) and a nonselective antagonist, [³H]FMZ, display the same binding activity in the double-immunopurifed receptors (Fig. 3D). Therefore, these experiments clearly demonstrated the presence of pharmacologically active 5 subunits associated with pharmacologically inactive 1 subunits, coassembled in the single native GABA_A receptor complex. It should be noted that 3 is expressed at very low levels in this brain region (an anti-3 antibody immunoprecipitated 3.0 ± 1.7% of the [³H]FMZ binding; data not shown) and the  subunit is not significantly associated with 2-containing receptors at the hippocampal formation (results not shown but see Quirk et al., 1994a, 1995; Araujo et al., 1998). The existence of these active and inactive  subunits was also suggested in cortical GABA_A receptor containing 1 and 3 subunits (Araujo et al., 1996) and the preponderance of the pharmacological properties of the 5 subunits was also postulated by Sur et al. (1998).

The stoichiometry of the native GABA_A receptors is unknown. Previous results in native and recombinant receptors indicated the existence of two  subunits (Pollard et al., 1995; Tretter et al., 1997; Araujo et al., 1996; Jechlinger et al., 1998; present work) and two  subunits (Tretter et al., 1997; Gorrie et al., 1997; Li and De Blas, 1997) in a single receptor complex. However conflicting results have been reported for the  subunit. The presence of one  subunit (Mossier et al., 1994; Togel et al., 1994; Tretter et al., 1997; Jechlinger et al., 1998) or two  subunits (Quirk et al., 1994b; Khan et al., 1994a; Benke et al., 1996) in the receptor complex has been proposed. As mentioned before, both the  and 2 subunits are implicated in the benzodiazepine binding sites, whereas GABA binding sites seem to be conformed by the  and  subunits (Smith and Olsen, 1995; Sigel and Buhr, 1997). Therefore, the association between 1 and 5 subunits, the existence of pharmacologically active and inactive  subunits together with the presence of two GABA binding sites in the 1-5 immunopurified receptors (see Fig. 4A) strongly support the presence of a single  subunit, the 2 subunit (our anti-2 antibody immunoprecipitated 90-95% of both total- and low-affinity binding sites for [³H]FMZ; data not shown but see also Ruano et al., 1994a) coassembled with two  subunits (1 and 5) and two  subunits (probably 2 and/or 3; see Fig. 4B for a model). Thus, the predominance of the pharmacological properties of the 5 subunit should implicate that only one of the two  subunits in the native GABA_A receptor is arranged in the appropriate location (Fig. 4B) forming the benzodiazepine binding site. In consequence, the benzodiazepine binding sites of the native GABA_A receptors should be determined by the presence and the arrangement of the different subunits in the pentameric GABA_A receptor complex.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Summary of pharmacological properties of the 1-5-immunopurified receptors (A) and an hypothetical model (B) of the molecular composition of the native 1-5 GABAA receptors from rat hippocampus. The model presented in B is based on the pharmacological properties of the 1-5 double- immunopurified receptors shown in A, and is also in concordance with the molecular data reported in this and other works (Mertens et al., 1993; Ruano et al., 1994a; Tretter et al., 1997; Sur et al., 1998)

Finally, the physiological consequences for the presence of the different  subunit pairs in the GABA_A receptors from the hippocampal formation are currently unknown. The absence of benzodiazepine binding sites in a particular  subunit does not preclude the influence of this subunit in the functional properties of the GABA_A receptor complexes. In this sense, recombinant GABA_A receptors containing 13, 2, and 2 subunits confers unique functional properties, distinctive from GABA_A receptors containing a single  subtype (Ebert et al., 1994; Verdoorn, 1994). Therefore, the presence of two different  subunit subtypes in native receptor complexes, irrespective to the presence of absence of benzodiazepine binding sites, could modify the functional properties of the GABA_A receptor.