Immunological Identification of Multiple α-Like Subunits of the γ-Aminobutyric AcidA Receptor Complex Purified from Neonatal Rat Cortex

Antibodies were prepared against a synthetic peptide corresponding to amino acid sequences 174-203 of the bovine gamma-aminobutyric acidA (GABAA) receptor alpha 1-subunit. The antibodies recognized this synthetic alpha 1-peptide, but failed to react with the homologous peptide sequence, 170-199, of the bovine beta 1-subunit. On Western blots, anti-alpha 1-subunit antibody recognized a 50-kilodalton (kDa) protein in affinity-purified receptor preparations from adult rat cortex and cerebellum. In receptor purified from neonatal cortex, the anti-alpha 1-antibody reacted with 50-kDa, 53-54-kDa, and 59-kDa proteins. After digestion with endoglycosidase F, these three protein bands retained differing electrophoretic mobilities. The 50-kDa and 59-kDa subunits of affinity-purified neonatal receptor, which were photoaffinity-labeled with [3H]flunitrazepam, were immunoprecipitated to different extents by alpha-subunit antibody. These data suggest the existence in GABAA receptor from neonatal cortex of three proteins (50 kDa, 53 kDa, and 59 kDa) which have immunological homology to alpha 1-subunit of bovine GABAA receptor. The presence of an alpha- and a beta-like subunit with similar mobility on sodium dodecyl sulfate-polyacrylamide gel electrophoresis may account for the relatively high concentration of protein in the 53-54-kDa band which has been observed in receptor purified from neonatal cortex. The presence of multiple alpha-like subunits may be related to the presence of a relatively high concentration of type II GABA receptor in this tissue.     

Quantitative Immunoprecipitation Studies with Anti-γ-Aminobutyric AcidA Receptor γ2 1–15 Cys Antibodies

Antibodies raised against the synthetic peptide NH2-QKSDDDYEDYASNKTC-COOH (gamma 2 1-15 Cys), which corresponds to the N-terminal amino acid sequence with a C-terminal cysteine of the human gamma 2 subunit of the gamma-aminobutyric acidA (GABAA) receptor, were used to study the quantitative immunoprecipitation of agonist benzodiazepine binding sites from bovine brain. Anti-gamma 2 1-15 Cys antibodies were found to immunoprecipitate specifically in parallel [3H]flunitrazepam- and [3H]muscimol-reversible binding sites in a dose-dependent manner. The maximum percentages of [3H]flunitrazepam binding sites immunoprecipitated from detergent extracts of bovine cerebral cortex, cerebellum, and hippocampus were 68, 77, and 83%, respectively. Immunoprecipitation studies with anti-alpha 1 324-341 antibodies carried out in parallel with anti-gamma 2 1-15 Cys antibodies provided evidence for the promiscuity of the gamma 2 subunit within native GABAA receptors. These results substantiate the association of the gamma 2 polypeptide with native GABAA receptors.     

Antibodies Specific for α-Subunit Subtypes of GABAA Receptors Reveal Brain Regional Heterogeneity

Abstract: Antisera were produced in rabbits against synthetic peptides based on subtype-specific regions of the cDNA sequences of the α1, α2, α3, and α4 (also termed α5) subunits of mammalian GABA_A receptors. The antigen peptides were chosen from the putative cytoplasmic loop between the proposed third and fourth membrane spanning helices; they were not only subtype-specific sequences, but also their hydrophilicity and predicted secondary structures suggested high potential antigenicity. In all cases, antipeptide antisera recognized on western blots the corresponding α-subunit polypeptide of the GABA_A receptors purified from bovine brain by benzodiazepine-affinity chromatography, and were able to immunoprecipitate binding activity from detergent-solubilized purified receptors. The four antisera each recognized a unique polypeptide, and only one, in the purified receptor, with α1, α2, α3, and α4 identified at 51, 52, 56, and 57 kDa, respectively. This represents the first identification of the α4 gene product on a gel. Both the relative amount of staining in immunoblots and the fraction of receptor binding that could be immunoprecipitated by saturating concentrations of each of the four subtypespecific antibodies varied in a consistent manner between receptors purified from different brain regions. Thus, cerebral cortex receptor contained all four α polypeptides on western blots, and significant activity could be precipitated by all four. Hippocampal receptor lacked α3 immunoreactivity on blotting and by immunoprecipitation; α1 was less, whereas both α2 and α4 were more abundant in hippocampus than in cortex by both techniques. Cerebellum receptor contained only α1 of the four α subunits tested, and the anti-α1 antibodies immunoprecipitated >90% of the binding activity. The variable amounts of staining and immunoprecipitation from the three brain areas by the four antisera demonstrate the presence of heterooligomeric receptor complexes with different α-subunit constituents in cortex, hippocampus, and cerebellum. The sum of cortical receptor activity precipitated individually by the four anti-α antisera was > 150%, indicating that some heterooligomers are likely to contain more than one class of α subtype, although most receptor complexes probably contain only one α subtype. These α-subunit subtype-specific antibodies should be useful in analyzing structure, function, and localization of GABA_A/benzodiazepine receptors in mammalian brain.     

The GABAA/benzodiazepine receptor is a heterotetramer of homologous alpha and beta subunits.

The GABAA receptor has been purified to homogeneity from bovine cerebral cortex. Under stringent conditions of isolation, the GABAA receptor was shown to consist only of alpha (Mr 53 000) and beta (Mr 57 000) subunits. A densitometric scan of SDS-PAGE gels under reducing conditions showed that these subunits were present in a 1:1 ratio. A model of the receptor as a heterologous tetramer alpha 2 beta 2 is proposed. Monoclonal antibodies have been raised to the purified bovine GABAA receptor. One of these antibodies, 1A6, was shown to react with both the alpha and beta subunits of the purified receptor. The subunits were still positive in immunoblots following the removal of the carbohydrate moieties of the respective polypeptides by endoglycosidase F treatment. This antibody has been employed to demonstrate antigenic cross-reactivity between the GABAA receptors of three vertebrate species. It is further proposed that there is partial amino acid sequence homology between the alpha and beta polypeptides and hence that they are derived from a single ancestral gene.     

Beta 1 subunits of sodium channels. Studies with subunit-specific antibodies

The purified Na+ channel from rat brain consists of alpha (260 kDa), beta 1 (36 kDa), and beta 2 (33 kDa) subunits. Pure beta 1 subunits were prepared from purified rat brain Na+ channels which had been adsorbed to hydroxylapatite resin and used to prepare specific anti-beta 1 subunit antiserum. Antibodies purified from this antiserum by antigen affinity chromatography immunoprecipitate 125I-labeled, purified beta 1 subunits and purified Na+ channels (measured as high affinity [3H] saxitoxin binding sites) and recognize beta 1 subunits on immunoblots of solubilized rat brain membranes. The affinity-purified anti-beta 1 antibodies recognize beta 1 subunits in rat spinal cord, heart, skeletal muscle, and sciatic nerve, but do not detect immunoreactive beta 1 subunits in eel electroplax or eel brain. The developmental time course of expression of immunoreactive beta 1 subunits in rat forebrain was measured by immunoprecipitation followed by immunoblotting with affinity-purified anti-beta 1 antibodies. The amount of immunoreactive beta 1 subunits increased steadily to adult levels during the first 21 days of postnatal development.     

Subunit selectivity and epitope characterization of mAbs directed against the GABAA/benzodiazepine receptor

mAbs bd 17, bd 24, and bd 28 raised against bovine cerebral gamma-aminobutyric acid (GABAA)/benzodiazepine receptors were analyzed for their ability to detect each of 12 GABAA receptor subunits expressed in cultured mammalian cells. Results showed that mAb bd 17 recognizes epitopes on both beta 2 and beta 3 subunits while mAb bd 24 is selective for the alpha 1 subunit of human and bovine, but not of rat origin. The latter antibody reacts with the rat alpha 1 subunit carrying an engineered Leu at position four, documenting the first epitope mapping of a GABAA receptor subunit-specific mAb. In contrast to mAbs bd 17 and bd 24, mAb bd 28 reacts with all GABAA receptor subunits tested but not with a glycine receptor subunit, suggesting the presence of shared epitopes on subunits of GABA-gated chloride channels.     

Decreased protein levels of nicotinic receptor subunits in the hippocampus and temporal cortex of patients with Alzheimer's disease

Deficits of cortical nicotinic acetylcholine receptors (nAChRs) have been observed in Alzheimer's disease (AD) by receptor binding assays. Little is known about the receptor subunit specificity influenced by AD, and it might be of importance for therapeutic strategies. In the present study, the protein levels of nAChR alpha3, alpha4, alpha7, and beta2 subunits were investigated using western blot analysis on postmortem brains of patients with AD and age-matched controls. The results showed that in human postmortem brain samples, bands with molecular masses of 52, 42, and 50 kDa were detected by anti-alpha4, anti-alpha7, and anti-beta2 antibodies, respectively. When anti-alpha3 antibody was used, one major band of 49 kDa and two minor bands of 70 and 38 kDa were detected. In AD patients, as compared with age-matched controls, the alpha4 subunit was reduced significantly by approximately 35 and 47% in the hippocampus and temporal cortex, respectively. A significant reduction of 25% in the alpha3 subunit was also observed in the hippocampus and a 29% reduction in the temporal cortex. For the alpha7 subunit, the protein level was reduced significantly by 36% in the hippocampus of AD patients, but no significant change was detected in the temporal cortex. In neither the hippocampus nor the temporal cortex was a significant difference observed in the beta2 subunit between AD patients and controls. These results reveal brain region-specific changes in the protein levels of the nAChR alpha3, alpha4, and alpha7 subunits in AD.     

Bovine γ-Aminobutyric AcidA Receptor Sequence-Specific Antibodies: Identification of Two Epitopes Which Are Recognised in Both Native and Denatured γ-Aminobutyric AcidA Receptors

Polyclonal antibodies have been raised against synthetic peptides whose sequences correspond to the N-terminal 15 amino acids and the C-terminal 17 amino acids of the bovine gamma-aminobutyric acidA (GABAA) receptor alpha 1 subunit. These antibodies were shown to react with the denatured GABAA receptor alpha subunit, Mr 53,000, in Western blots with both purified receptor and brain membranes as antigens. Also, both antibodies recognised both the purified and detergent-solubilised GABAA receptor as demonstrated by dose-dependent specific immunoprecipitation of the GABA and benzodiazepine binding sites from solution. Evidence is also presented to show brain-regional distribution of the expression of the alpha 1 subunit.     

Identification of the alpha 3-subunit in the GABAA receptor purified from bovine brain

Polyclonal antibodies have been raised to synthetic amino acid sequences of the bovine GABAA receptor alpha 1 and alpha 3 subunits. Anti-alpha 1 subunit antibodies recognise a polypeptide of 53 kDa whereas anti-alpha 3 subunit antibodies recognise a polypeptide of 59-60 kDa, in Western blots of GABAA receptor purified from adult bovine cerebral cortex, cerebellum and 12-day calf cerebral cortex.     

Subtype-Selective Immunoprecipitation of the β2-Adrenergic Receptor

Most antibodies known to interact with beta-adrenergic receptors do not exhibit subtype selectivity, nor do they provide quantitative immunoprecipitation. A monoclonal antibody, G27.1 raised against a synthetic peptide corresponding to the C-terminus of the beta 2-adrenergic receptor of hamster, is selective for the beta 2 subtype. G27.1 provides nearly quantitative immunoprecipitation of the beta 2-adrenergic receptor from hamster lung that has been photoaffinity-labeled and solubilized with sodium dodecyl sulfate. Immunoprecipitation is completely blocked by nanomolar concentrations of the immunizing peptide. This antibody interacts with beta 2-adrenergic receptors from three rodent species, but not with those from humans. When C6 glioma cells, which contain both beta 1- and beta 2-adrenergic receptors, are photoaffinity-labeled in the absence or presence of subtype-selective antagonists, subtype-selective photoaffinity-labeling results. G27.1 can immunoprecipitate beta 2-, but not beta 1-, adrenergic receptors from these cells. Similar results were obtained following subtype-selective photoaffinity-labeling of membranes from rat cerebellum and cerebral cortex. The beta-adrenergic receptors from C6 glioma cells and rat cerebral cortex exist as a mixture of two molecular weight species. These species differ in glycosylation, as shown by endoglycosidase F digestion of crude and immunoprecipitated receptors.     

Type I and Type II γ-Aminobutyric Acid/Benzodiazepine Receptors: Purification and Analysis of Novel Receptor Complex from Neonatal Cortex

The gamma-aminobutyric acid (GABA) type A receptor was purified several thousandfold by affinity chromatography from rat cerebellum, adult cortex, and neonatal cortex. Competition for the benzodiazepine binding site by CL 218872 indicated that cerebellar receptors were predominantly type I, adult cortical receptors were a mixture of subtypes, and neonatal cortex was enriched in type II receptor. The receptor purified from neonatal cortex contained predominantly a 54-kilodalton (kDa), beta-subunit-like protein, whereas receptors from cerebellum and adult cortex contained nearly equal amounts of a 50-kDa, alpha-subunit-like protein and a 54-kDa polypeptide. Peptide maps of trypsin-digested 54-kDa subunits from cerebellum, adult cortex, and neonatal cortex exhibited very similar profiles, a result indicating considerable homology between these proteins in the receptor subtypes. A 59-kDa subunit protein was detected in the receptor complex purified from neonatal cortex. Like the 50-kDa, alpha-subunit of the type I receptor, this protein was photolabeled with [3H]flunitrazepam. The photolabeled peptide fragments, produced by trypsin digestion of these alpha 50- and alpha 59-subunits, exhibited the same retention times on reverse-phase HPLC. A less highly purified GABAA receptor preparation from adult rat spinal cord possessed characteristics that were very similar to those of the receptors purified from neonatal cortex.     

Sequence and Regional Distribution of the mRNA Encoding the α2 Polypeptide of Rat γ-Aminobutyric AcidA Receptors

A cDNA from a rat hippocampal cDNA library encodes an isoform of the alpha polypeptide of the gamma-aminobutyric acid (GABA)/benzodiazepine (BZ) receptor. Its deduced amino acid sequence is 96% identical to that of the alpha 2 polypeptide of the bovine GABAA receptor. The polypeptide has features shared by all previously reported GABAA receptor alpha polypeptides and shares 71-76% identity with previously described rat alpha polypeptides. Most of the differences lie in the presumed extracellular and intracellular domains. On Northern blots, the alpha 2 cDNA detects two mRNAs, which are found in cortex, hippocampus, and striatum, brain regions enriched in pharmacologically defined "BZ type II" receptors. Other workers have previously shown that the alpha polypeptides of the GABAA receptor largely determine the BZ binding properties of reconstituted receptors. The distribution of alpha 2 mRNAs in rat brain suggests that the alpha 2 subunit may indeed be involved in the BZ type II receptors.     

Evidence for the Existence of Differential O-Glycosylated α5-Subunits of the γ-Aminobutyric AcidA Receptor in the Rat Brain

Abstract: Polyclonal antibodies were raised to synthetic peptides having amino acid sequences corresponding with the N- or C-terminal part of the γ-aminobutyric acid_A (GABA_A) receptor α₅-subunit. These anti-peptide α₅(2–10) or anti-peptide α₅(427–433) antibodies reacted specifically with GABA_A receptors purified from the brains of 5–10-day-old rats in an enzyme-linked immunosorbent assay and were able to dose-dependently immunoprecipitate up to 6.3 or 13.1% of the GABA_A receptors present in the incubation, respectively. In immunoblots, each of these antibodies reacted with the same two protein bands with apparent molecular mass of 53 or 57 kDa. After exhaustive treatment of purified GABA_A receptors with N -Glycanase, each of these antibodies identified two proteins with apparent molecular masses of 46 and 48 kDa. Additional treatment of GABA_A receptors with neuraminidase and O -Glycanase resulted in an apparently single protein with molecular mass of 47 kDa, which again was identified by both the anti-peptide α₅(2–10) and the anti-peptide α₅(427–433) antibody. These results indicate the existence of at least two different α₅-sub-units of the GABA_A receptor that differ in their carbohydrate content. In contrast to other α- or β-subunits of GABA_A receptors so far investigated, at least one of these two α₅-subunits contains O-linked carbohydrates.     

Identification of the subunits of GTP-binding proteins coupled to somatostatin receptors.

Somatostatin (SRIF) induces its biological effects by interacting with membrane-bound receptors that are linked to cellular effector systems via G proteins. We have studied SRIF receptor-G protein associations by solubilizing the SRIF receptor from rat brain and AtT-20 cells and immunoprecipitating the receptor-G protein complex with peptide-directed antisera against the different subunits of the G protein heterotrimer. Antiserum 8730, which selectively interacts with all Gi alpha subtypes, maximally and specifically immunoprecipitated SRIF receptor-Gi alpha complexes. To identify the subtypes of Gi alpha that are coupled to SRIF receptors, the subtype-selective antisera 3646, 1521, and 1518, which specifically interact with Gi alpha 1, Gi alpha 2, and Gi alpha 3, respectively, were used to immunoprecipitate SRIF receptor-Gi alpha complexes. Antiserum 3646 immunoprecipitated SRIF receptor-Gi alpha 1 complexes from both brain and AtT-20 cells. Antiserum 1521 immunoprecipitated Gi alpha 2 from both brain and AtT-20 cells but did not immunoprecipitate SRIF receptors from these tissues. Antiserum 1518 immunoprecipitated AtT-20 cell SRIF receptors but uncoupled brain SRIF receptor-G protein complexes. This result was confirmed with another peptide-selective antiserum, SQ, directed against Gi alpha 3. The findings from these studies indicate that Gi alpha 1 and Gi alpha 3 are coupled to SRIF receptors, whereas Gi alpha 2 is not. Even though brain and AtT-20 cell SRIF receptors were both coupled to Gi alpha, the receptors from these tissues differed in their coupling to Go alpha. Antiserum 2353, which is directed against Go alpha, immunoprecipitated SRIF receptors from AtT-20 cells, but did not immunoprecipitate or uncouple SRIF receptor-G protein complexes from rat brain. To determine the beta subunits associated with the SRIF receptor, antisera directed against G beta 36 and G beta 35 were used to immunoprecipitate SRIF receptor-G protein complexes from brain. Peptide-directed antiserum against G beta 36 selectively immunoprecipitated solubilized brain SRIF receptors. However, antiserum directed against the G beta 35 subunit did not immunoprecipitate brain SRIF receptors, suggesting that brain SRIF receptors may preferentially associate with G beta 36. In addition to coimmunoprecipitating with Gi alpha and G beta, brain SRIF receptors coimmunoprecipitated the G protein gamma subunits, G gamma 2 and G gamma 3. These results provide the first evidence that SRIF receptors are coupled to different subunits of G proteins and suggest that selectivity exists in the association of different G protein subunits with the SRIF receptor.     

α4 but Not α3 and α7 Nicotinic Acetylcholine Receptor Subunits Are Lost from the Temporal Cortex in Alzheimer's Disease

Neuronal nicotinic acetylcholine receptors labelled with tritiated agonists are reduced in the cerebral cortex in Alzheimer's disease (AD), but to date it has not been demonstrated which nicotinic receptor subunits contribute to this deficit. In the present study, autopsy tissue from the temporal cortex of 14 AD cases and 15 age-matched control subjects was compared using immunoblotting with antibodies against recombinant peptides specific for alpha3, alpha4, and alpha7 subunits, in conjunction with [3H]epibatidine binding. Antibodies to alpha3, alpha4, and alpha7 produced one major band on western blots at 59, 51, and 57 kDa, respectively. [3H]Epibatidine binding and alpha4-like immunoreactivity (using antibodies against the extracellular domain and cytoplasmic loop of the alpha4 subunit) were reduced in AD cases compared with control subjects (p < 0.02) and with a subgroup of control subjects (n = 9) who did not smoke prior to death (p < 0.05) for the former two parameters. [3H]Epibatidine binding and cytoplasmic alpha4-like immunoreactivity were significantly elevated in a subgroup of control subjects (n = 4) known to have smoked prior to death (p < 0.05). There were no significant changes in alpha3- or alpha7-like immunoreactivity associated with AD or tobacco use. The selective involvement of alpha4 has implications for understanding the role of nicotinic receptors in AD and potential therapeutic targets.     

Identification and immunohistochemical mapping of GABAA receptor subtypes containing the delta-subunit in rat brain.

Synaptic inhibition in brain is mainly mediated via GABAA receptors which display a striking structural heterogeneity. A novel type of GABAA receptor subunit, the delta-subunit, has recently been described based on molecular cloning of its cDNA. To identify the prevalence and distribution of GABAA receptors which contain the delta-subunit protein in situ, polyclonal site-directed antisera were developed against three synthetic peptides derived form the rat delta-subunit cDNA-sequence. All antisera specifically recognized a 54 kDa protein in GABAA receptor preparations. Nearly 30% of the GABAA receptors contained the delta-subunit immunoreactivity and displayed high affinity GABA and high affinity benzodiazepine binding sites as shown by immunoprecipitation. Receptors which contain the delta-subunit were immunohistochemically shown to be restricted to a few brain areas such as the cerebellum, thalamus and dentate gyrus of the hippocampal formation. Thus, those neurons which express GABAA receptors with a delta-subunit have now been visualized and made accessible for a functional analysis of this GABAA receptor subtype in situ.     

Conserved Organization of γ-Aminobutyric AcidAReceptor Genes: Cloning and Analysis of the Chicken β4-Subunit Gene

A series of genomic clones containing DNA that encodes the chicken gamma-aminobutyric acidA (GABAA) receptor beta 4 subunit have been isolated. These have been restriction mapped and partially sequenced to determine the structural organization and the size of the beta 4-subunit gene. This gene, which comprises nine exons, spans more than 65 kb. The organization of the chicken GABAA receptor beta 4-subunit gene has been compared to that of the murine GABAA receptor delta-subunit gene and to those of the genes that encode other members of the ligand-gated ion-channel superfamily, namely muscle and neuronal nicotinic acetylcholine receptors (AChRs). Although the positions of the intron/exon boundaries of GABAA receptor subunit genes are seen to be highly conserved, there are significant differences between the genes that encode GABAA receptor and AChR subunits. These results are discussed in relation to the proposal that this superfamily of ligand-gated ion-channel receptor genes arose by duplication of an ancestral receptor gene.     

Antibodies Recognising the GABAA/Benzodiazepine Receptor Including Its Regulatory Sites

Polyclonal antibodies have been raised against the GABA/benzodiazepine receptor purified to homogeneity from bovine cerebral cortex in deoxycholate and Triton X-100 media. Radioimmunoassay was applied to measure specific antibody production using the 125I-labelled gamma-aminobutyric acid (GABA)/benzodiazepine receptor as antigen. The antibodies specifically immunoprecipitated the binding sites for [3H]muscimol and for [3H]flunitrazepam from purified preparations. In addition, when a 3-[(3-cholamidopropyl)dimethylammonio] 1-propanesulphonate (CHAPS) extract of bovine brain membranes was treated with the antibodies, those sites as well as the [3H]propyl-beta-carboline-3-carboxylate binding, the [35S]t-butylbicyclophosphorothionate binding (TBPS), the barbiturate-enhanced [3H]flunitrazepam binding, and the GABA-enhanced [3H]flunitrazepam binding were all removed together into the immunoprecipitate. Western blot experiments showed that these antibodies recognise the alpha-subunit of the purified GABA/benzodiazepine receptor. These results further support the existence in the brain of a single protein, the GABAA receptor, containing a set of regulatory binding sites for benzodiazepines and chloride channel modulators.     

Immunochemical Characterization of the β2 Subunit of the GABAA Receptor

To date three β subunits of the GABA_A receptor have been identified in rat brain as a result of cDNA library screening. The β2 subunit has been reported to have a wide distribution in rat brain based on in situ hybridization studies quantifying β2 mRNA. To study the β2 subunit more directly, we have raised a polyclonal antibody to a synthetic peptide representing residues 315–334 of the intracellular loop of the β2 subunit. The antibody, which had been affinity-purified, recognized the β2 peptide but did not immunolabel homologous β1 and β3 subunit peptides, indicating that this antibody is specific for the β2 subunit of the receptor. In western blots of the purified receptor, the antibody recognized a major diffuse band of 54–58 kDa arid exhibited minor labeling of lower-molecular-mass polypeptides. In western blots of cortex homogenate, the antibody exhibited nervous system-specific labeling of a 55-kDa band that comigrated with the 55-kDa band of the purified receptor. Quantitative immunolabeling of this 55-kDa polypeptide permitted direct determination of the relative amounts of the β2 subunit in different brain regions. The brainstem contained the highest relative specific activity of the β2 subunit, followed by the inferior colliculus, olfactory lobe, and cerebellum. Lower levels of immunolabeling were seen in hypothalamus, hippocampus, thalamus, and cortex.     

GABAA receptor beta subunit heterogeneity: functional expression of cloned cDNAs.

Cloned cDNAs encoding two new beta subunits of the rat and bovine GABAA receptor have been isolated using a degenerate oligonucleotide probe based on a highly conserved peptide sequence in the second transmembrane domain of GABAA receptor subunits. The beta 2 and beta 3 subunits share approximately 72% sequence identity with the previously characterized beta 1 polypeptide. Northern analysis showed that both beta 2 and beta 3 mRNAs are more abundant in the brain than beta 1 mRNA. All three beta subunit encoding cDNAs were also identified in a library constructed from adrenal medulla RNA. Each beta subunit, when co-expressed in Xenopus oocytes with an alpha subunit, forms functional GABAA receptors. These results, together with the known alpha subunit heterogeneity, suggest that a variety of related but functionally distinct GABAA receptor subtypes are generated by different subunit combinations.