Impact of intracellular domain flexibility upon properties of activated human 5-HT3 receptors

Background and Purpose

It has been proposed that arginine residues lining the intracellular portals of the homomeric 5-HT₃A receptor cause electrostatic repulsion of cation flow, accounting for a single-channel conductance substantially lower than that of the 5-HT₃AB heteromer. However, comparison of receptor homology models for wild-type pentamers suggests that salt bridges in the intracellular domain of the homomer may impart structural rigidity, and we hypothesized that this rigidity could account for the low conductance.

Experimental Approach

Mutations were introduced into the portal region of the human 5-HT₃A homopentamer, such that putative salt bridges were broken by neutralizing anionic partners. Single-channel and whole cell currents were measured in transfected tsA201 cells and in Xenopus oocytes respectively. Computational simulations of protein flexibility facilitated comparison of wild-type and mutant receptors.

Key Results

Single-channel conductance was increased substantially, often to wild-type heteromeric receptor values, in most 5-HT₃A mutants. Conversely, introduction of arginine residues to the portal region of the heteromer, conjecturally creating salt bridges, decreased conductance. Gating kinetics varied significantly between different mutant receptors. EC₅₀ values for whole-cell responses to 5-HT remained largely unchanged, but Hill coefficients for responses to 5-HT were usually significantly smaller in mutants. Computational simulations suggested increased flexibility throughout the protein structure as a consequence of mutations in the intracellular domain.

Conclusions and Implications

These data support a role for intracellular salt bridges in maintaining the quaternary structure of the 5-HT₃ receptor and suggest a role for the intracellular domain in allosteric modulation of cooperativity and agonist efficacy.

Linked Article

This article is commented on by Vardy and Kenakin, pp. 1614–1616 of volume 171 issue 7. To view this commentary visit http://dx.doi.org/10.1111/bph.12550.

Importance of recognition loops B and D in the activation of human 5-HT₃ receptors by 5-HT and meta-chlorophenylbiguanide

The 5-HT₃ receptor is a cation selective member of the pentameric Cys-loop ligand-gated ion channels. While five subunits are known to exist, only two receptor subtypes have been significantly characterized: the homomeric receptor consisting of five A subunits and the heteromeric receptor containing both A and B subunits. The agonist recognition and activation of these receptors is orchestrated by six recognition loops three, A-C, on the principal subunit, and three, D-F, on the complementary subunit. In this study we have focused on the B loop of the principal subunit and loop D of the complementary subunit where aligned amino acids differ between the two subunits. A mutational analysis has been carried out using both 5-HT and m-chlorophenylbiguanide (mCPBG) to characterize receptor activation in the mutant receptors using two-electrode voltage clamp in Xenopus oocytes. The results show that the B loop W178I mutation of the 5-HT3A subunit markedly reduces the efficacy of mCPBG in both the homomeric and heteromeric receptors, while activation by 5-HT remains intact. Replacement of the D loop amino acid triplet RQY of the 5-HT3A subunit, with the aligned residues from the 5-HT3B subunit, QEV, converts 5-HT to a weak partial agonist in both the homomer and heteromer, but does not compromise activation by mCPBG. Exchange of the RQY triplet for the 5-HT3B subunit homologue, QEV, increases the Hill coefficient and decreases the EC₅₀ of this mutant when expressed with the wild type 5-HT3A subunit.

Changes in GABA(A) receptor subunit expression in the midbrain during the oestrous cycle in Wistar rats

In women, the late luteal phase or "premenstrual" period is commonly associated with psychological disturbances, which include mood changes and increased aggression. The underlying cause is unknown but one possibility is that fluctuations in levels of neuroactive steroids precipitate changes in expression of GABA(A) receptor subunits that result in functional changes in inhibitory control systems. The present study investigated the levels of expression of alpha4, beta1 and delta GABA(A) receptor subunits in the periaqueductal gray matter (PAG) in rats and whether plasticity occurs during the oestrous cycle in females. In male rats alpha4, beta1 and delta subunit immunoreactive neurones were present throughout the PAG in similar numbers. In female rats in proestrus, oestrus and early dioestrus, the density of alpha4, beta1 and delta subunit immunoreactive cells was similar to males. However, in late dioestrus, the numbers increased significantly, especially in the dorsolateral PAG, a region which is particularly rich in GABAergic interneurones. These parallel changes may reflect an increase in expression of the alpha4beta1delta GABA(A) receptor subtype. Recombinant alpha4beta1delta receptors, expressed in Xenopus oocytes, exhibited and EC(50) for GABA an order of magnitude lower (2.02+/-0.33 microM; mean+/-S.E.M.) than that found for the most ubiquitous alpha1beta2gamma2 GABA(A) receptor (32.8+/-2.5 microM). Increased expression of alpha4beta1delta GABA(A) receptors in the interneurones of the PAG could render the panic circuitry abnormally excitable by disinhibiting the ongoing GABAergic inhibition. Similar changes in neuronal excitability within the PAG in women consequent to falling steroid levels in the late luteal phase of the menstrual cycle could contribute to the development of pre-menstrual dysphoria.

Effects of the antidepressant/antipanic drug phenelzine on alanine and alanine transaminase in rat brain

1. Phenelzine (PLZ) is an antidepressant with anxiolytic properties. Acute and chronic PLZ administration increase brain GABA levels, an effect due, at least in part, to an inhibition of the activity of the GABA metabolizing enzyme, GABA transaminase (GABA-T). 2. Previous preliminary reports have indicated that acute PLZ treatment also elevates brain alanine levels. As with GABA, the metabolism of alanine involves a pyridoxal phosphate-dependent transaminase. 3. In the study reported here, the effects of acute PLZ treatment on the levels of various amino acids, some of which are also metabolized by pyridoxal phosphate-dependent transaminases were compared in rat whole brain. Of the 6 amino acids investigated, only GABA and alanine levels were elevated (in a time- and dose-dependent manner). 4. The elevation in brain alanine levels could be explained, at least in part, by a time- and dose-dependent inhibitory effect of PLZ on alanine transaminase (ALA-T), although as with GABA the increases are higher than expected from the degree of enzyme inhibition produced. In addition, we also showed that the elevation in alanine levels and the inhibition of alanine transaminase in the brain are retained after 14 days of PLZ treatment, and that PLZ produces a marked increase in extracellular levels of alanine. 5. These results are discussed in terms of their relevance to synaptic function and to the pharmacological profile of PLZ.

GABA(A) receptor gene expression in rat cortex: differential effects of two chronic diazepam treatment regimes

Diazepam is widely prescribed as an anxiolytic but its therapeutic application is limited because with daily use tolerance develops to certain aspects of its pharmacological profile. We compared the effects of two dosing paradigms on GABA(A) receptor gene expression and benzodiazepine binding characteristics. Equivalent daily doses of 15 mg/kg/day diazepam were delivered either via constant infusion or daily subcutaneous injection for 14 days. The two distinct treatment regimes produced significantly different changes in GABA(A) receptor alpha4-, beta2-, beta3- and gamma1-subunit mRNA steady-state levels. Similar changes in the GABA enhancement of flunitrazepam binding and the BZ3/BZ2 subtype ratio determined ex vivo were produced, however, significant differences were found in [(3)H]-Ro 15-4513 binding between cortical tissue from diazepam injected animals compared with diazepam infused animals. Our data suggest that it is the diurnal fluctuations in receptor occupancy that are responsible for the different effects produced by these two dosing regimes.

Chiari type I malformation and postoperative respiratory failure

PURPOSE

To present a case of respiratory failure following suboccipital craniectomy for Chiari type I malformation.

CLINICAL FEATURES

A 22-yr-old man presented with a two year history of symptoms and signs suggestive of brainstem compression at the level of the foramen magnum. This was confirmed with magnetic resonance imaging. The procedure of suboccipital craniectomy, upper cervical laminectomy and fourth ventricle exploration was performed. Three hours postoperatively the patient experienced episodes of apnea and subsequently became drowsy. Blood gas analysis revealed hypercapnic respiratory failure. Chest X-ray revealed evidence of pulmonary aspiration. The trachea was re-intubated and the lungs ventilated in intensive care for 72 hr. He was discharged home two weeks postoperatively.

CONCLUSION

Chiari type I malformation is associated with a number of associated anomalies. These patients are at considerable risk of respiratory depression and bulbar dysfunction in the perioperative period. The anesthetic issues are reviewed and discussed.

Characterization of the interaction of zopiclone with gamma-aminobutyric acid type A receptors

Zopiclone is a cyclopyrrolone that is used clinically as a hypnotic. Although this drug is known to interact with neuronal gamma-aminobutyric acid type A receptors, its binding site(s) within the receptor oligomer has been reported to be distinct from that of the classical benzodiazepines. After photoaffinity labeling with flunitrazepam, receptors in rat cerebellar membranes showed differentially reduced affinity for flunitrazepam and zopiclone by 50- and 3-fold, respectively. Because histidine 101 of the alpha-subunit is a major site of photolabeling, we have made specific substitutions of this residue and studied the consequences on the binding properties of zopiclone and diazepam using recombinant alpha1beta2gamma2-receptors transiently expressed in tsA201 cells. Both compounds showed similar binding profiles with receptors containing mutated alpha-subunits, suggesting a similar interaction with the residue at position 101. At alpha1beta2gamma3-receptors, flunitrazepam affinity was dramatically decreased by approximately 36-fold, whereas the affinity for zopiclone was decreased only 3-fold, suggesting a differential contribution of the gamma-subunit to the binding pocket. Additionally, we used electrophysiological techniques to examine the contribution of the gamma-subunit isoform in the receptor oligomer to ligand recognition using recombinant receptors expressed in Xenopus oocytes. Both compounds are agonists at alpha1beta2gamma2- and alpha1beta2gamma3-receptors, with flunitrazepam being more potent but less efficacious. In summary, these data suggest that histidine 101 of the alpha1-subunit plays a similar role in ligand recognition for zopiclone, diazepam, and flunitrazepam.

Intracerebral hemorrhage and Moyamoya disease in pregnancy

PURPOSE

To present a case of Moyamoya disease with intracranial hemorrhage complicating pregnancy.

CLINICAL FEATURES

A 36-yr-old parturient at 34 wk gestation presented with left hemiparesis, headache, nausea and vomiting. Subsequent deterioration in level of consciousness and the development of a dilated right pupil necessitated immediate intubation. Urgent non-contrast CT scan revealed a large right intracerebral hematoma with transtentorial herniation. The patient underwent simultaneous emergency Cesarean section and craniotomy. A postoperative angiogram revealed findings consistent with Moyamoya disease. The neonate survived but the patient developed severe cerebral edema and died eleven days postoperatively.

CONCLUSION

Adult patients with Moyamoya disease often present with intracranial hemorrhage which poses unique anesthetic challenges. We report a case of intracerebral hemorrhage during pregnancy, which is known to be associated with high morbidity and mortality. The anesthetic techniques are reviewed and discussed.

Importance of phenylalanine 107 in agonist recognition by the 5-hydroxytryptamine(3A) receptor

The 5-hydroxytryptamine (5-HT)(3) receptor is a member of the ligand-gated ion channel receptor family with significant homology to the nicotinic acetylcholine, gamma-aminobutyric acid(A), and glycine receptors. In this receptor class, the agonist binding site is formed by parts of the extracellular amino-terminal region. This study examines the effects of altering phenylalanine 107 (F107) of the 5-HT(3AL) subunit, obtained from NG108-15 cells, using site-directed mutagenesis. The wild-type (WT) and mutant receptors were expressed in HEK 293 cells and characterized using both whole-cell patch-clamp and radioligand binding. The tyrosine mutant F107Y exhibits a significantly lower affinity for the agonist 5-HT (K(i) = 203 versus 15.6 nM) and an increase of similar magnitude in the EC(50) value (10.6 versus 1.2 microM) compared with WT. The activation kinetics of the maximal currents generated by 5-HT with this mutant were markedly slower than those of the WT receptor, but application of supramaximal concentrations of the agonist markedly decreased the time to half-peak. The asparagine mutant F107N displayed a significantly higher affinity for 5-HT than the WT receptor (1.62 versus 15.6 nM), which was mirrored in direction and magnitude by changes in the EC(50) value for this agonist (0.2 versus 1.2 microM). In contrast to the WT receptor, the mutant F107N was activated by acetylcholine (EC(50) = 260 microM). The response to acetylcholine was blocked by the 5-HT(3) receptor antagonist renzapride with a similar IC(50) value as that determined against currents generated by 5-HT in the WT receptor. These data suggest that F107 is an important determinant of agonist recognition at the 5-HT(3) receptor.

The effect of treatment regimen on the development of tolerance to the sedative and anxiolytic effects of diazepam

RATIONALE

Chronic treatment with benzodiazepines results in tolerance to their sedative and anxiolytic effects and there is considerable evidence that different mechanisms underlie the development of tolerance to different behavioural effects.

OBJECTIVE

The purpose of the present experiment was to compare the behavioural effects of chronic treatment with diazepam (15 mg/kg per day) given as daily subcutaneous injections or by osmotic minipump. Both regimens resulted in continual receptor occupancy, but the daily injections also provided a period of higher brain concentrations.

METHODS

Rats were tested in the holeboard, which provides measures of exploration and locomotor activity, and in the elevated plus-maze and social interaction tests of anxiety. For those in the subcutaneous injection group the tests were 2 h after injection, when brain concentrations were highest.

RESULTS

Despite a higher brain concentration in the injected group, both groups showed tolerance to diazepam's sedative effects, after 7 days of treatment. In contrast, in the elevated plus-maze, there was tolerance to the anxiolytic effects in the pump group after 14 days, but a persisting anxiolytic effect in the injected group at 14 and 28 days. Whilst higher brain concentrations could explain this result in the plus-maze, they cannot account for the pattern observed in the social interaction test, where the injection group showed a significant anxiogenic effect at 28 days.

CONCLUSIONS

Whereas the mechanism underlying tolerance to the sedative effects of diazepam was insensitive to the different treatment regimens, the results suggest that different adaptive mechanisms were triggered in the two tests of anxiety with a differential sensitivity to the treatment regimen. The adaptive mechanism predominating in the social interaction test was favoured by the injection regimen which produced intermittent peak concentrations. This mechanism seems to be an oppositional one, leading to an anxiogenic response, which was manifest despite high brain concentrations of diazepam at the time of testing.

Decreased GABA enhancement of benzodiazepine binding after a single dose of diazepam

The GABA and benzodiazepine binding sites on GABA(A) receptors are allosterically coupled. The in vitro binding of 2 nM [3H]flunitrazepam to cortical and cerebellar membranes prepared from drug-naive rats was potentiated approximately 1.6-fold by 100 microM GABA. Potentiation in both regions was significantly reduced 4 or 12 but not 24 h after a single dose of 15 mg/kg diazepam. At 24 h after the last of 14 daily doses of diazepam, no differences in GABA potentiation were observed. Diazepam-induced changes in GABA(A) receptor gamma2-subunit gene transcription and alpha1-, beta2-, and gamma2-subunit steady-state mRNA levels did not appear to be temporally related to allosteric uncoupling.

Rapid, sensitive procedure to determine buspirone levels in rat brains using gas chromatography with nitrogen-phosphorus detection

Reported here is a rapid, sensitive and relatively inexpensive procedure using gas chromatography with nitrogen-phosphorus detection (GC-NPD) to quantify buspirone levels in brains of rats. The analyte was directly extracted from brain homogenate with toluene after basification and then subjected to GC-NPD analysis using a capillary column. The calibration curves were linear over the range of 10 to 320 ng per 2 ml of brain homogenate, with typical r2 values >0.99. The assay was highly reproducible and gave peaks with excellent chromatographic properties.

Chronic diazepam exposure decreases transcription of the rat GABA(A) receptor gamma2-subunit gene

The rate of transcription of the GABA(A) receptor gamma2-subunit gene in rat cortex has been measured using the nuclear run-off transcriptional assay. Exposure of rats to diazepam (15 mg/kg/day for 14 days) caused a significant reduction in the level of nascent GABA(A) receptor gamma2-subunit transcripts. Therefore, a component of the cellular response to chronic benzodiazepine exposure includes events which take place at the level of transcription of a GABA(A) receptor gene.

Chronic zolpidem treatment alters GABA(A) receptor mRNA levels in the rat cortex

The effect of chronic zolpidem treatment on the steady-state levels of gamma-aminobutyric acidA alpha1-6, beta1-3 and gamma1-3 subunit mRNAs in rat cortex has been investigated. Male Sprague-Dawley rats were injected once daily, for 7 or 14 days, with 15 mg/kg of zolpidem in sesame oil vehicle. The levels of the alpha4 and beta1 subunit mRNAs were significantly increased after 7 days of treatment and the level of alpha1 subunit mRNA was significantly decreased after 14 days of treatment, as determined by solution hybridization. These results are compared to the previously determined effects of an equivalent schedule of treatment with diazepam.

Analysis of the ligand binding site of the 5-HT3 receptor using site directed mutagenesis: importance of glutamate 106

The 5-HT3 receptor is a ligand-gated ion channel with significant structural similarity to the nicotinic acetylcholine receptor. Several regions that form the ligand binding site in the nicotinic acetylcholine receptor are partially conserved in the 5-HT3 receptor, presumably reflecting the conserved signal transduction mechanism. Specific amino acid differences in these regions may account for their distinct ligand recognition properties. Using site-directed mutagenesis, we have replaced one of these residues, glutamate 106 (E106), with aspartate (D), asparagine (N), alanine (A) or glutamine (Q) and characterized the ligand-binding and electrophysiological properties of the mutant receptors after transient expression in HEK-293 cells. The affinity for the selective 5-HT3 receptor antagonist [3H]GR65630 was decreased 14-fold in the mutant E106D (Kd = 3.69 +/- 0.32 nM) when compared to wildtype (WT, E106) 5-HT3 receptor (0.27 +/- 0.03 nM), while the affinity for E106N was unchanged (0.42 +/- 0.07 nM, means +/- SEM, n = 3-10). Decreased affinities for both E106D and E106N were observed for the antagonists granisetron, ondansetron and renzapride and for the agonists 5-HT (130- and 30-fold) and 2-methyl-5-HT (250- and 20-fold), respectively. Both mutants still formed 5-HT-activatable ion channels, but the high Hill coefficient of the concentration effect curves in wildtype (2.0) was decreased to unity in both cases. The EC50 of 5-HT was increased seven-fold in E106N (8.7 microM) when compared to wildtype (1.2 microM), but unchanged in E106D, and the potency of the antagonist ondansetron for both mutants was decreased. E106A and E106Q expressed poorly preventing a detailed characterization. These data suggest that E106 contributes to the ligand-binding site of the 5-HT3 receptor and may form an ionic or hydrogen bond interaction with the primary ammonium group of 5-HT.

Dimethyl sulfoxide/propylene glycol is a suitable solvent for the delivery of diazepam from osmotic minipumps

The functional integrity of Alzet osmotic minipumps was assessed using two organic solvents (50% (v/v) dimethyl sulfoxide (DMSO)/50% (v/v) propylene glycol (PG) and 100% tetraglycol) which dissolve diazepam, an aqueous insoluble benzodiazepine. Both solvents showed a significant decrease in output rate over time: the decline with tetraglycol was, however, more marked and variable. Further, the DMSO/PG vehicle demonstrated a comparable decline in rate (1.45%) to that of the control vehicles saline and water (1.12%). DMSO/PG is therefore a suitable solvent for the chronic delivery of diazepam from osmotic minipumps.

The major site of photoaffinity labeling of the gamma-aminobutyric acid type A receptor by [3H]flunitrazepam is histidine 102 of the alpha subunit

The alpha subunit of the gamma-aminobutyric acid type A (GABA(A)) receptor is known to be photoaffinity labeled by the classical benzodiazepine agonist, [3H]flunitrazepam. To identify the specific site for [3H]flunitrazepam photoincorporation in the receptor subunit, we have subjected photoaffinity labeled GABA(A) receptors from bovine cerebral cortex to specific cleavage with cyanogen bromide and purified the resulting photolabeled peptides by immunoprecipitation with an anti-flunitrazepam polyclonal serum. A major photolabeled peptide component from reversed-phase high performance liquid chromatography of the immunopurified peptides was resolved by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. The radioactivity profile indicated that the [3H]flunitrazepam photoaffinity label is covalently associated with a 5.4-kDa peptide. This peptide is glycosylated because treatment with the enzyme, peptide-N4-(N-acetyl-beta-glucosaminyl)asparagine amidase, reduced the molecular mass of the peptide to 3.2 kDa. Direct sequencing of the photolabeled peptide by automated Edman degradation showed that the radioactivity is released in the twelfth cycle. Based on the molecular mass of the peptides that can be generated by cyanogen bromide cleavage of the GABA(A) receptor alpha subunit and the potential sites for asparagine-linked glycosylation, the pattern of release of radioactivity during Edman degradation of the photolabeled peptide was mapped to the known amino acid sequence of the receptor subunit. The major site of photoincorporation by [3H]flunitrazepam on the GABA(A) receptor is shown to be alpha subunit residue His102 (numbering based on bovine alpha 1 sequence).

Identification of domains in human recombinant GABAA receptors that are photoaffinity labelled by [3H]flunitrazepam and [3H]Ro15-4513

We have used [3H]flunitrazepam and [3H]Ro15-4513 as photoaffinity labelling agents in combination with a chemical cleavage technique to localize the benzodiazepine recognition sites of specific human recombinant alpha 1 beta 1 gamma 2, alpha 1 beta 3 gamma 2 and alpha 6 beta 3 gamma 2 GABAA receptor subtypes. The chemical agent utilized was hydroxylamine, whose substrate is a relatively rare asparagine-glycine amide bond that occurs only in the alpha subunits of the receptors examined in this study. Cleavage products were resolved using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). The results of these experiments show that, in the alpha 1 subunit-containing receptors, incorporation of [3H]flunitrazepam occurs within residues 1-103 of the alpha 1 subunit, while incorporation of [3H]Ro15-4513 occurs within the region of the alpha 1 subunit that lies between residue 104 and the C-terminus. Photolabelling of membranes prepared from the alpha 6 beta 3 gamma 2-expressing cell line with [3H]Ro15-4513 resulted in the incorporation of radiolabel into two major protein species of M(r) 56,000 and M(r) 48,000, indicating incorporation into the alpha 6 subunit and possibly also the gamma 2 subunit. Hydroxylamine cleavage of alpha 6-containing receptors labelled with [3H]Ro15-4513 produced a gel profile consistent with the incorporation of the label occurring between residue 125 and the C-terminal. Thus, we have shown that the recognition sites for the agonist [3H]flunitrazepam and the inverse agonist [3H]Ro15-4513 occur within distinct domains of the human GABAA receptor.

Chronic treatment with diazepam or abecarnil differently affects the expression of GABAA receptor subunit mRNAs in the rat cortex

Diazepam and abecarnil produce their overt effects by interaction with the GABAA receptor. Chronic treatment with abecarnil, however, does not induce diazepam-like tolerance. This study investigates the effects of chronic diazepam and abecarnil treatment on expression of GABAA receptor alpha 1-6 beta 1-3 and gamma 1-3 subunit isoform mRNAs in rat cortex. Male Sprague-Dawley rats were injected subcutaneously once daily for 7 or 14 days with 15 mg/kg diazepam or 6 mg/kg abecarnil in sesame-oil vehicle, and steady-state levels of GABAA receptor subunit mRNAs were quantified by solution hybridization. The levels of alpha 4- and alpha-, beta 1- and gamma 3-subunit mRNAs were significantly increased after 7 days of diazepam treatment, and this effect was maintained at 14 days. A significant increase in alpha 3-subunit mRNA was apparent only after 14 days of diazepam treatment and a significant decrease in beta 2-subunit mRNA was seen only after 14 days of abecarnil treatment. Gamma 2-Subunit mRNA was significantly decreased after 14 days of either diazepam or abecarnil exposure. A degree of association between a particular drug treatment and changes in the levels of mRNAs arising from a given gene cluster was noted. Our results are consistent with a model of diazepam dependence based on GABAA receptor subunit isoform switching.

Ultrastructure of the 5-hydroxytryptamine3 receptor

We have determined the ultrastructure of 5-hydroxytryptamine3 (5-HT3) serotonin receptors purified from NG108-15 mouse neuroblastoma x rat glioma cells by electron microscopic examination of receptor particles embedded in uranyl acetate stain and metal replicas of rapidly frozen receptors. The 5-HT3 receptor can be modelled as a cylinder 11 nm in length and 8 nm in diameter with a closed end and a central cavity 3 nm in diameter. Analysis of the rotational symmetry of single receptor particles indicates that the 5-HT3 receptor is composed of five subunits arranged symmetrically around a central cavity. Together with evidence obtained for related proteins in other studies using ultrastructural, biochemical, or electrophysiological methods, our observations suggest that all members of the ligand-gated ion channel superfamily may possess a pentameric quaternary structure.

Quaternary structure of the native GABAA receptor determined by electron microscopic image analysis

In the transmitter-gated ion channel class of receptors, the members of which are all believed to be heterooligomers, the number and arrangement of the subunits are only known with any certainty for the nicotinic acetylcholine receptor from Torpedo electric fish. That receptor has been shown to possess a pentameric rosette structure, with five homologous subunits (alpha 2, beta gamma delta) arranged to enclose the central ion channel. The data were obtained by electron image analysis of two-dimensional receptor arrays, which form as a consequence of that receptor's exceptionally high abundance in the Torpedo membranes and are therefore not attainable for other receptors. We have applied another direct approach to determine the quaternary structure of native ionotropic GABA receptors. We have purified those receptors from porcine brain cortex and analysed the rotational symmetry of isolated receptors visualized by electron microscopy. The results show the receptor to have a pentameric structure with a central water-filled pore, which can now be said to be characteristic of the entire superfamily.

Molecular properties of 5-hydroxytryptamine3 receptor-type binding sites purified from NG108-15 cells

5-Hydroxytryptamine3 (5-HT3) receptor-type binding sites were solubilised from NG108-15 mouse neuroblastoma x rat glioma hybrid cells using five different detergents [n-octyl-beta-D-glucoside, Triton X-100, 3-[3-(cholamidopropyl)dimethylammonio]-1-propanesulphonate (CHAPS), sodium cholate, and deoxycholate] and the solubilisation efficiencies compared. The equilibrium binding, kinetic properties, and pharmacological profile of solubilised binding sites were similar to those of 5-HT3 receptor-type binding sites (5-HT3R) in membrane preparations determined using [3H]GR65630. The solubilised binding sites were purified using an affinity column constructed by coupling the high-affinity antagonist GR119566X to an Affi-Gel 15 resin. The affinity of purified 5-HT3R for [3H]-GR65630 was reduced threefold compared to the crude soluble preparation, but the pharmacological profile was similar. The sedimentation coefficient of the purified protein (11S, detergent: CHAPS) was determined by sucrose density gradient centrifugation. The apparent molecular mass of the detergent/binding site complex (370 kDa) was determined by size exclusion chromatography in the presence of n-dodecyl-beta-D-maltoside. Gel electrophoresis of the purified protein revealed bands at apparent molecular masses of 36, 40, 50, and 76 kDa. Electron microscopy of the negatively stained purified protein showed the presence of round particles of 8-9 nm diameter with a 2-nm stained pit in the centre, closely resembling the doughnut shapes described for nicotinic acetylcholine receptors.

Effect of free fatty acids on GABAA receptor ligand binding

Phospholipase A2 (PLA2) treatment of synaptosomal membranes, which causes the release of fatty acids, particularly unsaturated fatty acids, inhibits the flux of chloride ions through the gamma-aminobutyric acid (GABA) benzodiazepine receptor ion channel in response to activation by agonists. PLA2 treatment has also been shown to affect ligand binding to the receptor. In the present study, we have investigated the effect of unsaturated free fatty acids, arachidonic acid and oleic acid and saturated free fatty acids, arachidic acid and stearic acid on various characteristics of GABAA receptor ligand binding. Only the unsaturated fatty acids showed any effect: arachidonic acid and oleic acid enhanced flunitrazepam binding and muscimol binding but inhibited tert-butylbicyclophosphorothionate (TBPS) binding in a dose-dependent manner. The effects on muscimol and TBPS binding were shown to be due to changes in receptor density by saturation analysis. Oleic acid and arachidonic acid also decreased the enhancement of flunitrazepam and muscimol binding by cartazolate and pentobarbital but did not affect GABA enhancement of flunitrazepam binding. These data indicate that unsaturated free fatty acids can mimic the effects of PLA2 treatment and underline the importance of the lipid microenvironment on ligand binding to the GABAA receptor.

Effects of the antidepressant/antipanic drug phenelzine on GABA concentrations and GABA-transaminase activity in rat brain

The effects of long-term (28-day) administration of several antidepressant/antipanic drugs [imipramine, desipramine, tranylcypromine and phenelzine (PLZ)] on gamma-aminobutyric acid-tranaminase (GABA-T) activity and GABA levels were investigated in rat frontal cortex. Of the drugs investigated, only PLZ inhibited GABA-T and elevated GABA levels. Additional short-term experiments were conducted with PLZ, and they demonstrated a dose-dependent inhibition of GABA-T in rat whole brain. Time-response studies on inhibition of GABA-T in whole brain demonstrated that at a dose of PLZ of 15 mg/kg i.p. inhibition of GABA-T remained relatively constant from 1 to 8 hr and that the enzyme was still inhibited by 23% at 24 hr after PLZ administration.

5-HT3 receptors in NG108-15 neuroblastoma x glioma cells: effect of the novel agonist 1-(m-chlorophenyl)-biguanide

The effect of the novel agonist, 1-(m-chlorophenyl)-biguanide (mCPBG) was examined on 5-HT3 receptors in NG108-15 mouse neuroblastoma x rat glioma hybrid cells, using whole-cell voltage-clamp and radioligand binding on intact cells. Electrophysiological studies showed that mCPBG is a partial agonist, with an EC50 of 3.1 microM. Displacement of the selective 5-HT3 receptor antagonist [3H]GR65630 by mCPBG revealed a Ki of 14.2 nM. The study suggests that mCPBG may have a high affinity for desensitized 5-HT3 receptors and also revealed some differences between 5-HT3 receptors in NG108-15 and N1E-115 cells.

Multiple benzodiazepine receptors: no reason for anxiety

Since the introduction of the benzodiazepines into clinical practice in 1960, these drugs have been widely employed as anxiolytics, sedative/hypnotics and anticonvulsants. In recent years, concern has been expressed about their side-effects, and their use has declined. During this latter period many advances have been made in understanding the molecular mechanisms by which these drugs produce their effects. Adam Doble and Ian Martin review this progress and highlight the possibilities that these advances may hold for the development of more efficacious and specific medicines.

Solubilization, purification, and functional reconstitution of 5-hydroxytryptamine3 receptors from N1E-115 neuroblastoma cells

5-Hydroxytryptamine3 (5-HT3) receptors from N1E-115 neuroblastoma cells were solubilized using 1.1% n-octylglucoside; five other detergents were less effective. Purification was achieved by affinity chromatography using immobilized GR119566X and biospecific elution with quipazine. Saturation analyses with [3H] GR67330 binding revealed high affinity binding to homogeneous populations of sites in both the solubilized (Kd = 0.05 +/- 0.02 nM) and purified (Kd = 0.10 +/- 0.04 nM) preparations. Competition experiments indicated that the solubilized and purified receptor preparations retained the characteristics observed in N1E-115 cells in vivo. Polyacrylamide gel electrophoresis of the purified receptor revealed a single protein band of 54.7 +/- 1.3 kDa. The purified receptor was incorporated into liposomes, and the functional integrity of the protein was demonstrated by measurement of m-chlorophenylbiguanide-stimulated 22Na uptake. Saturation analysis of the reconstituted preparation revealed a Kd of 0.24 +/- 0.07 nM and suggested that 0.2% of 5-HT3 receptors present in the original membrane preparation had been incorporated into liposomes.

The agonist properties of m-chlorophenylbiguanide and 2-methyl-5-hydroxytryptamine on 5-HT3 receptors in N1E-115 neuroblastoma cells

1. The effects of 5-HT3 selective agonists have been studied in whole-cell voltage-clamped N1E-115 neuroblastoma cells. 2. Application of 5-hydroxytryptamine (5-HT) results in the rapid development of a transient inward current at quasiphysiological membrane potentials. This current can be blocked by the 5-HT3 specific antagonists BRL 43694 and GR67330. 3. Application of 2-methyl-5-HT (2-Me5-HT), a 5-HT3 selective agonist, produced a qualitatively similar inward current, but with a maximum response only 20% of that produced by 5-HT. 4. In the presence of 100 microM 2-Me5-HT, the upper part of the 5-HT dose-response curve was shifted to the right but reached the same maximum value as in the absence of 2-Me5-HT. 2-Me5-HT appears therefore to be a partial agonist under these conditions. 5. The novel 5-HT3 agonist, meta-chlorophenylbiguanide (mCPBG) is a full agonist, but has a Hill coefficient (1.5) significantly less than that of 5-HT (2.3). 6. Comparison with radioligand binding data show that mCBPG is 100 times less potent than expected; it may therefore exhibit a high affinity for a desensitized state.

Characterization of 5-HT3 receptors in intact N1E-115 neuroblastoma cells

The highly selective 5-HT3 receptor antagonist [3H]GR65630 has been used to characterize 5-HT3 receptors in intact N1E-115 neuroblastoma cells. Equilibrium binding analysis demonstrated high-affinity binding to a single class of receptors with a Kd of 0.69 (+/- 0.12) nM and Bmax of 31.4 (+/- 11.4) fmol/10(5) cells, equivalent to approximately 200,000 sites per cell. Specific binding was displaced by low concentrations of 5-HT3-selective ligands, and by the nicotinic antagonist d-tubocurarine.

Target size of 5-HT3 receptors in N1E-115 neuroblastoma cells and rat brain

Radiation inactivation was used to determine the molecular target size of the binding site for [3H]GR65630, a specific 5-HT3 receptor ligand, in two different neuronal tissues. Using a calibration curve of known molecular weight enzymes, the target sizes of [3H]GR65630 binding sites in N1E-115 neuroblastoma cells and rat brain were 98,600 +/- 11,300 and 49,100 +/- 8,500 Da, respectively. The results suggest 5-HT3 receptors may be present as dimers in N1E-115 neuroblastoma cells.

Biochemical characterization of an isolated and functionally reconstituted gamma-aminobutyric acid/benzodiazepine receptor

We have solubilized, affinity-purified, and functionally reconstituted the gamma-aminobutyric acid/benzodiazepine (GABA/BDZ) receptor from rat brain into natural brain lipid liposomes. The detergent, 3-[(3-cholamidopropyl)-dimethylammonio] 1-propanesulphonate, was employed for the isolation of the receptor in the presence of a whole rat brain lipid extract supplemented with cholesteryl hemisuccinate. The soluble and reconstituted protein showed a homogeneous [3H]flunitrazepam binding population and the allosteric modulation of this binding site by GABA, by the pyrazolopyridine, cartazolate, and by the depressant barbiturate, pentobarbital. The purified GABA/BDZ receptor when incorporated into liposomes has been visualized by electron microscopy and reveals rosette structures, 8-9 nm in diameter, which appear to have a central pore. Sodium dodecyl sulphate-polyacrylamide gel electrophoresis of the reconstituted GABA/BDZ receptor reveals three major protein bands of 41, 52-56, and 59-62 kDa, the latter two of which appears as doublets. Functional receptor reconstitution is demonstrated by the measurement of GABA-stimulated 36Cl- flux into the purified GABA/BDZ receptor incorporated liposomes and its modulation by the BDZs, barbiturates, and pyrazolopyridines.

GABA preincubation of rat brain sections increases [3H]GABA binding to the GABAA receptor and compromises the modulatory interactions

Receptor autoradiography has been employed to investigate the effect of gamma-aminobutyric acid (GABA) preincubation on the interaction of the GABAA receptor with its ligands. [3H]GABA (50 nM) binding to the GABAA receptors is increased by 60% compared to control sections after GABA (100 microM) preincubation. Receptor autoradiography shows that the increase is more pronounced in certain brain areas. The allosteric interactions between the GABA and benzodiazepine recognition sites were also examined. An increase in [3H]GABA (50 nM) binding to rat brain sections by co-incubation with the benzodiazepine, flunitrazepam (FNZ) has been observed autoradiographically. This effect has been quantitated in several brain regions; the overall brain increase in [3H]GABA binding induced by 1 microM FNZ was 20%. The increase in [3H]FNZ (1 nM) binding by co-incubation with GABA has also been observed autoradiographically, and the effect quantitated in four brain regions. The overall brain increase in [3H]FNZ binding induced by 100 microM GABA was 34%. After GABA preincubation these allosteric responses are significantly reduced in size. The increase in the [3H]GABAA binding as a consequence of GABA preincubation appears to reflect an increase in receptor affinity for [3H]GABA with no significant change in the maximum number of binding sites. We suggest that GABA preincubation converts the GABAA receptor to a higher affinity desensitised receptor conformation.

Polyclonal antibodies to the glycine receptor antagonist strychnine

Polyclonal antibodies have been raised in rabbits against the glycine receptor antagonist strychnine, coupled through a 2-amino substituent to the antigenic protein key-hole limpet haemocyanin. Strychnine binding of the predominantly immunoglobulin G (IgG) class of antibodies was measured by incubation with [3H]strychnine, followed by adsorption of IgG onto Staphylococcus aureus cells and filtration through glass-fibre filters under vacuum. Only strychnine and structurally related alkaloids or derivatives were able to inhibit [3H]strychnine binding to the IgG. A significant rank correlation was found between the potencies of these compounds to inhibit [3H]strychnine binding to the antibodies and to the glycine receptor in mouse spinal cord membranes. In contrast, preincubation of strychnine antibodies with a variety of ligands at other neurotransmitter, drug, or hormone receptors in the CNS (at 10(-4) M) failed to inhibit binding significantly. The failure of glycine to inhibit strychnine antibody binding is consistent with previous suggestions that the recognition sites for this amino acid on the CNS receptor may be conformationally distinct from those for the antagonist alkaloid. Strychnine antibodies may now help in the identification and purification of possible endogenous ligands at this alkaloid binding site in the CNS.

Light microscopic autoradiographic localisation in rat brain of the binding sites for the GABAA receptor antagonist [3H]SR 95531: comparison with the [3H]GABAA receptor distribution

Quantitative receptor autoradiography has been employed to determine the binding distribution of [3H]SR 95531 (2'-(3'-carboxy-2',3'-propyl)-3-amino-6-p-methoxyphenylpyrazinium bromide), a GABAA receptor antagonist, in the rat brain. The location showed good correspondence, in most areas, with the binding of [3H]GABA to GABAA receptors. The major exceptions were binding to cerebellum and hippocampus. We also describe here a non-GABA-associated [3H]SR 95531 binding site which is found at its highest level in the hippocampal CA1 region but also in CA2, CA3, dentate gyrus and cerebral cortex. The results support the localisation of [3H]SR 95531 sites to the GABAA receptor complex.

Solubilisation of the gamma-aminobutyric acid/benzodiazepine receptor from rat cerebellum: optimal preservation of the modulatory responses by natural brain lipids

We have solubilised the gamma-aminobutyric acid/benzodiazepine (GABA/BDZ) receptor from rat cerebellum using 3-[(3-cholamidopropyl)dimethylammonio] 1-propane sulphonate (CHAPS) in the presence of a natural brain lipid extract and cholesteryl hemisuccinate. The soluble material shows a homogeneous [3H]flunitrazepam ([3H]FNZ) binding population with an equilibrium dissociation constant (KD) of 4.4 +/- 0.2 nM compared to a KD of 2.3 +/- 0.2 nM in cerebellar synaptosomal membranes. The receptor complex in solution retains the characteristic facilitation of [3H]flunitrazepam binding induced by GABA, the pyrazolopyridine cartazolate, and the depressant barbiturate pentobarbital to the same extent as that observed in synaptosomal membranes. Furthermore, these responses are retained both quantitatively and qualitatively when this preparation is stored for 48 h at 4 degrees C. This is contrary to the results obtained with a CHAPS-soluble preparation including asolectin in which these responses are anomalous and extremely labile on storage.

Polyclonal antibodies to agonist benzodiazepines

Benzodiazepine-binding, immunoglobulin G class antibodies have been raised in three rabbits immunised with a conjugate of kenazepine coupled to keyhole limpet haemocyanin. The antibodies were assayed by [3H]flunitrazepam binding, followed by adsorption onto Staphylococcus aureus cells. Measurement of the rates of association and dissociation of [3H]flunitrazepam binding, together with saturation analysis of equilibrium binding, revealed varying degrees of heterogeneity in the affinity constants of the three rabbit antisera (equilibrium KD values 0.18 to 4.13 nM at 20-22 degrees). Specificity of the antibodies was investigated by testing a wide variety of compounds (at concentrations of up to 10-100 microM) for their ability to inhibit [3H]flunitrazepam binding. Only benzodiazepines known to act as agonists at their receptor sites in the central nervous system (CNS) caused an inhibition of binding. The rank orders of the IC50 values of these drugs for inhibition of [3H]flunitrazepam binding to IgG from two out of the three rabbits correlated significantly with that previously published for displacement of CNS receptor binding. The agonist beta-carboline derivative ZK 93423, the anxiolytic cyclopyrrolones suriclone and zopiclone and the purines inosine and hypoxanthine all failed to inhibit antibody binding, supporting previous suggestions that these drugs may bind at non-benzodiazepine recognition sites on the CNS receptor. The antibodies described are expected to provide useful reagents for raising anti-idiotypic antibodies directed against the CNS receptor and for the identification and purification of possible endogenous benzodiazepine receptor agonists in the CNS.

The benzodiazepines and their receptors: 25 years of progress

The benzodiazepines were introduced into clinical practice in 1960 and since that time have become one of the most frequently prescribed of all psychotropic agents. Over the past 25 years the mechanism by which these compounds produce their effects has been revealed by a combination of electrophysiological and biochemical studies. It is now apparent that these drugs interact with a specific recognition site on the GABAA receptor complex and thereby modulate the gating of the chloride channel associated with this major inhibitory amino acid transmitter receptor. This modulation is novel in that while the clinically prescribed benzodiazepines increase the chloride current in response to a given GABA stimulus, agents are now known which produce the opposite effect. However, the details of the molecular mechanisms by which these events occur remain to be elucidated. Our progress toward this goal will rely on the acquisition of structural information about this receptor protein using the techniques of molecular biology and its marriage with new initiatives concerning the detail of ligand recognition. The interpretation of such information in the context of the complex pharmacology of this important receptor system is indeed a stimulating prospect.

Modulation of [3H]flunitrazepam binding to rat cerebellar benzodiazepine receptors by phosphatidylserine

The influence of phosphatidylserine on ligand binding to the benzodiazepine/GABA receptor complex was assessed in rat cerebellar synaptic membranes and in a detergent-solubilized membrane preparation. Intact synaptic membranes or membranes solubilized with the zwitterionic detergent CHAPS (3-[(3-cholamidopropyl)-dimethylammonio]propanesulfonate) were incubated with a range of concentrations of phosphatidylserine for 2 h at 4 degrees C, prior to use in radioligand binding assays. Phosphatidylserine, an endogenous membrane phospholipid, facilitated the site-specific binding of [3H]flunitrazepam to synaptic membranes and CHAPS-solubilized preparations. In addition, phosphatidylserine inhibited the facilitation of [3H]flunitrazepam binding induced by either cartazolate or gamma-aminobutyric acid (GABA). Although the maximum effect (38% facilitation of [3H]flunitrazepam binding; greater than 90% inhibition of the cartazolate action) was produced using 130 microM phosphatidylserine, a significant enhancement of [3H]flunitrazepam binding could be observed upon preincubation of synaptic membranes with concentrations of phosphatidylserine as low as 5 microM. These results suggest that endogenous phosphatidylserine may play a role in the regulation of benzodiazepine/GABA receptor function, possibly through modulation of the mechanisms which functionally link the various components of this complex receptor system.

Solubilisation of the benzodiazepine receptor from rat cerebellum by the detergent n-octylpentaoxyethylene

The detergent n-octylpentaoxyethylene is one in the series of tenside detergents developed for membrane solubilisation. We have used this detergent to solubilise benzodiazepine receptors from rat cerebellum. The soluble receptor has an affinity (KD) for [3H]flunitrazepam of 1.8 nM +/- 0.2, which is not significantly different from that observed in synaptic membranes. Under optimal conditions (0.6% wt/vol), the number of soluble flunitrazepam binding sites (Bmax) of 0.35 pmol/mg protein suggests an apparent solubilisation of 40% of sites from the membrane. However, the absence of the characteristic facilitation of [3H]flunitrazepam binding by gamma-aminobutyric acid (GABA), cartazolate, and pentobarbital in this soluble receptor preparation suggests that such a preparation is unlikely to be a useful preparation for further studies on the molecular mechanisms underlying GABAA receptor function.

Evidence of strain differences in GABA-benzodiazepine coupling

The effects of a single dose of oxazepam on seizure threshold, receptor occupancy and brain oxazepam concentration were investigated at several time points after drug treatment in two inbred strains of mice (NIH and C3H/HE). The C3H/HE strain showed a greater sensitivity to the effects of both pentylenetetrazol and oxazepam. Furthermore, the C3H/HE strain showed decreases in both receptor occupancy and seizure threshold across time, whereas the NIH strain showed no change in either measure. Although within-strain correlations were observed between seizure threshold and receptor occupancy, the C3H/HE strain had similar seizure thresholds to the NIH strain throughout but lower percentage receptor occupancies, thus a between-strain correlation was not observed. The C3H/HE strain had a higher number of specific benzodiazepine binding sits and these results may reflect a strain difference in GABA-benzodiazepine receptor coupling. In a further experiment, the development of acute tolerance to the effects of oxazepam was investigated. Brain concentrations of oxazepam and receptor occupancies were determined for each strain of mouse, at two different time points (1.5 and 7.5 h after drug treatment), at which equivalent seizure thresholds were obtained by manipulating the starting dose of oxazepam. For each strain, when equivalent seizure thresholds were observed at different time points, equivalent receptor occupancies were also observed. However, a trend towards higher brain concentrations of oxazepam at the later time point was detected for the two strains, suggesting that there may be some decrease across time in the affinity of the receptor for oxazepam.

Solubilization of the benzodiazepine/gamma-aminobutyric acid receptor complex: comparison of the detergents octylglucopyranoside and 3-[(3-cholamidopropyl)-dimethylammonio]1-propanesulfonate (CHAPS)

Octylglucopyranoside (OCTG) was three times more efficient than 3-[(3-cholamidopropyl)-dimethylammonio]1-propanesulfonate (CHAPS) in solubilizing the benzodiazepine (BDZ)/gamma-aminobutyric acid (GABA) receptor complex from rat cerebellar synaptic membranes. OCTG-solubilized receptor preparations had ligand binding characteristics that were significantly different from those of the CHAPS-solubilized receptors. The inclusion of phospholipids in the solubilization media improved the binding characteristics of both soluble receptor preparations and appeared absolutely necessary for the maintenance of chloride facilitation of flunitrazepam (FNZ) binding to OCTG-solubilized receptors. FNZ and ethyl-beta-carboline-3-carboxylate bound to OCTG-solubilized preparations with equilibrium dissociation constants of 2.2 nM and 1.6 nM, respectively, and chloride (150 mM) and GABA (100 microM) + chloride facilitated the binding of FNZ by 15% and 55%, respectively; these ligand binding characteristics are similar to those of membrane-located BDZ receptors. Cartazolate, a pyrazolopyridine that facilitated the binding of FNZ to membrane-located and CHAPS-solubilized receptors, did not facilitate FNZ binding to OCTG-solubilized receptors. These results are discussed in terms of an interaction between the membrane lipid phosphatidylserine (PS) and cartazolate; PS appears to have the capacity to inhibit the effects of cartazolate on FNZ binding. Storage of the soluble receptor preparations for 24 h at 4 degrees C resulted in the loss of several characteristic BDZ receptor binding properties. Incorporation of the OCTG-solubilized receptor complex into liposomes prevented these losses but this procedure did not protect the CHAPS-solubilized receptors. We conclude that OCTG may have some advantages over CHAPS as the detergent of choice for the solubilization and reconstitution in liposomes of a functional BDZ/GABA receptor-chloride ionophore complex.

Long-lasting anticonvulsant effects of diazepam in different mouse strains: correlations with brain concentrations and receptor occupancy

There were marked strain differences in the duration of the protective effects of diazepam against the convulsant actions of penylenetetrazole and picrotoxin in mice. In no case was significant protection found at 12 h or longer, regardless of whether the incidence of or the latencies to myoclonus or tonic-clonic convulsions were considered. These behavioural differences could not be explained simply in terms of strain differences in benzodiazepine metabolism or in percent of receptor occupancy, as determined by the fractional displacement of 3H-flunitrazepam binding in vivo. It is suggested that there might be strain differences in the percent of receptor occupancy needed in order to produce an anticonvulsant effect.

Modification of pyrazoloquinolinone affinity by GABA predicts efficacy at the benzodiazepine receptor

A series of pyrazoloquinolinones CGS 9896, CGS 9895 and CGS 8216 have been reported to exhibit agonist, partial agonist and antagonist properties, respectively, at the benzodiazepine receptor. We have examined the effects of these compounds and of diazepam on pentylenetetrazole seizure thresholds in mice and found CGS 9896 to be a partial agonist and CGS 8216 a weak inverse agonist in this respect; CGS 9895 was essentially devoid of efficacy. In the presence of 100 microM GABA and 150 mM NaCl, the affinity of CGS 9896 for the benzodiazepine receptor at 37 degrees C was significantly increased whereas the affinity of CGS 8216 was similarly decreased and that of CGS 9895 was unchanged. The effect of GABA on the affinity of these ligands for the benzodiazepine receptor at 37 degrees C was thus predictive of the pharmacological efficacy which we observed.