Fjt XJ




Antagonists bo uj OS

Figure 11.6 Schematic representation of the GABAa receptor complex. Examples of the many structurally diverse compounds that act at different sites Z

on the receptor (see text for details). Picrotoxinin, the active component of picrotoxin, and TBPS act as non-competitive antagonists. The barbiturates, Q

steroids and anaesthetics are positive allosteric modulators, as are the benzodiazepine site ligands shown, with the exception of DMCM (negative q allosteric modulator) and flumazenil (benzodiazepine site antagonist) 2!

effects ranging from anticonvulsant and anxiolytic (benzodiazepine-like) to proconvulsant/ convulsant and anxiogenic. Between these two extremes are compounds, such as the imidazodiazepine flumazenil, that display only a limited degree of intrinsic activity but which are capable of antagonising the effects of the clinically useful benzodiazepines as well as those of the convulsant ligands or so-called 'inverse agonists' (negative allosteric modulators). All the compounds appear to act at the same or overlapping sites on the receptor complex. In studies of GABAA receptor single-channel currents, anxiolytic benzodiazepines, such as diazepam, increase the response to GABA but do not generally change the conductance of individual Cl" channels. Instead they increase the affinity of the receptor for GABA and, in steady-state experiments, increase the frequency of channel opening, in a manner equivalent to increasing the concentration of GABA. At GABAergic synapses such compounds prolong the decay of the postsynaptic current and may also increase its peak amplitude. Inverse agonists such as DMCM reduce the response to GABA by decreasing the frequency of channel opening.


Like benzodiazepine agonists, barbiturates possess sedative, anxiolytic and anti-convulsant properties. Although certainly not their only site of action, sedative barbiturates, such as phenobarbitone or the clinically used intravenous anaesthetic thiopentone, cause a marked potentiation of GAB A responses. Unlike benzodiaze-pines, barbiturates increase the time for which GABA-activated channels are open and increase the length of bursts of openings. At higher concentrations barbiturates can activate Cl" channels even in the absence of GABA. Neither effect is due to an action at the benzodiazepine binding site, as they are not blocked by the benzodiazepine antagonist flumazenil.


It has been known since the 1940s that steroids of the pregnane series have anaesthetic properties. These early studies led to the development of the intravenous anaesthetic althesin, the active component of which is alphaxolone (3a-hydroxy-5a-pregnan-11,20 dione). This compound has barbiturate-like actions and is also able to directly activate Cl" channels in the absence of GABA. These properties are shared by several endogenous steroids (synthesised in the brain or adrenal glands), the most potent being the reduced metabolite of progesterone, 3a-hydroxy-5a-pregnan-20-one (allopregnanolone or 3a,5a-THP) and the reduced metabolite of dexoycorticosterone, 3a,21-dihydroxy-5a-pregnan-20-one (allotetrahydrodeoxycorticosterone or a-THDOC). Particular interest in these compounds stems from the fact that they may act as endogenous modulators of GABAA receptors and their levels are altered by stress as well as during the menstrual cycle and pregnancy. For example, during menstruation decreasing levels of progesterone result in a decline in the production of allopregnanolone. Recently it has been demonstrated that such an abrupt decline (akin to drug withdrawal) can cause changes in the properties of GABAA receptors that may underlie the symptoms associated with premenstrual syndrome, including increased susceptibility to seizures and insensitivity to benzodiazepine agonists. Steroids appear to act at a distinct site on the GABA-receptor complex, as flumazenil does not block their action, and the Cl" currents they evoke directly can be potentiated by barbiturates (and vice versa). The 3^-methyl-substituted synthetic analogue of allopregnanolone, ganaxolone (3a-hydroxy-3^-methyl-5a-pregnan-20-one) is less easily metabolised than its endogenous parent compound, allowing activity following oral administration, and is currently under investigation as an anticonvulsant.


Steroids, such as alphaxolone, and barbiturates, such as thiopentone, represent only two classes of the many structurally diverse molecules found to induce general anaesthesia. Although a number of these clearly have actions on a range of targets, including glycine, 5-HT3, nicotinic and glutamate receptors, all, with the exception of the dissociative anaesthetic ketamine, have an effect on GABAA receptors at relevant concentrations. For example, the intravenous anaesthetic agents propofol, propanidid and etomidate markedly enhance responses to GABA (apparently by prolonging bursts of Cl~ channel openings) and are capable of directly evoking Cl~ currents. The currents produced by these agents at high doses, as well as those caused by steroids and barbiturates, are blocked by bicuculline, indicating that they are due to activation of the Cl~ channel associated with the GABAA receptor. It is also now clear that volatile anaesthetics such as halothane and isoflurane as well as alcohols (including ethanol), rather than having non-specific membrane-disrupting actions, owe at least some of their properties to a potentiation of GABA responses, through a direct interaction with sites on GABAa receptors.


Over the past decade or so significant advances have been made in our understanding of the structure of the GABAA receptor, which is now known to be formed by the assembly of multiple subunit proteins. In 1987 two subunits of the receptor, designated a and p, were cloned (Schofield et al. 1987). Following on from this work, 16 mammalian subunits encoded by distinct genes have now been identified. These genes encode proteins of approximately 450-550 amino acids (depending on the species) which, according to their sequence similarities, have been grouped into seven families — a, p, y, 8, e, % and 0 (Barnard et al. 1998; Bonnert et al. 1999). The a, p and y families contain multiple isoforms (a1-a6, P1-P3 and y1-y3) and in a number of cases additional complexity is generated by alternative mRNA splicing.

The subunits share varying degrees of sequence identity but have a similar predicted tertiary structure. This consists of four membrane-spanning a-helices (M1-M4), a large extracellular N-terminal region, a large intracellular domain between M3 and M4 and a short extracellular C-terminal portion (Fig. 11.7). The highest degree of conservation is in the transmembrane regions and the greatest variation in the intracellular loop between M3 and M4. The extracellular domain contains potential N-linked glycosylation sites and a p-loop formed by a disulphide bridge between two cysteine residues. The intracellular loops of p and y subunits contain sites for phosphorylation by a variety of protein kinases, including cAMP-dependent protein kinase, cGMP-dependent protein kinase, protein kinase C, Ca2+/calmodulin-dependent protein kinase and tyrosine kinase, which may be important in the regulation of receptor function. These general features are very similar to those of two other ligand-gated ion channels, the nicotinic acetylcholine receptor and the glycine receptor (see below) and there is a considerable degree of sequence homology among these proteins. By analogy with the nicotinic acetylcholine receptor, it is thought that the GABAA receptor is formed by the assembly of five subunits around a central ion channel, with the M2 region of each subunit forming the lining of the channel (Fig. 11.7).

Figure 11.7 Presumed arrangement of GABAa receptor subunits to form a receptor-channel complex, (a) Diagrammatic representation of an individual subunit with four transmembrane regions, extracellular sites for glycosylation and a site for phosphorylation on the intracellular loop between M3 and M4, (b) Association of five subunits to form a central ionophore bounded by the M2 region of each subunit, The suggested stoichiometry of the most widely expressed form of receptor is 2a, 26 and ly. Shown below are the possible subunit combinations of one such benzodiazepine-sensitive receptor together with a benzodiazepine-insensitive receptor in which the y subunit is replaced by a 5, and a re-containing receptor with four different subunit types

Figure 11.7 Presumed arrangement of GABAa receptor subunits to form a receptor-channel complex, (a) Diagrammatic representation of an individual subunit with four transmembrane regions, extracellular sites for glycosylation and a site for phosphorylation on the intracellular loop between M3 and M4, (b) Association of five subunits to form a central ionophore bounded by the M2 region of each subunit, The suggested stoichiometry of the most widely expressed form of receptor is 2a, 26 and ly. Shown below are the possible subunit combinations of one such benzodiazepine-sensitive receptor together with a benzodiazepine-insensitive receptor in which the y subunit is replaced by a 5, and a re-containing receptor with four different subunit types

Subunit combinations and receptor function

Expression studies in Xenopus oocytes or transfected cell lines originally suggested that functional GABA-activated chloride channels could be formed by receptor subunits of each class in isolation. However, much better expression occurs with two or more subunit types in combination and it is likely that most native receptors contain at least three different subunits. Co-expression of a and ¿6 subunits results in the assembly of functional receptors that can be activated by GABA and are sensitive to the antagonists bicuculline and picrotoxin and show modulation by barbiturates. But only when a y subunit is expressed in conjunction with an a and a 6 subunit is benzodiazepine binding and potentiation of GABA seen. As benzodiazepines do not bind to y subunits alone, it is likely that the conformation of the receptor is appropriate for benzodiazepine binding only when all three subunit types are present.

The large number of cloned subunit proteins makes it clear that GABAA receptors themselves must be diverse. An illustration of this diversity is provided by the pharmacology of benzodiazepine ligands. Even before the existence of GABAA receptor subunits was recognised, variations in the binding of radiolabelled drugs to native benzodiazepine receptors had led to the suggestion that not all GABA receptors were the same. Two types of benzodiazepine receptor were postulated — BZI and BZII. These had similar affinity for agonists such as diazepam and antagonists such as flumazenil, but BZI receptors showed a higher affinity for triazolopyridazines (e.g. CL 218872) and 6-carbolines (e.g. ¿6-CCM). It is now clear that the molecular basis for these differences resides in the variety of a subunits. Thus, while y subunits are required for benzodiazepine binding, the precise nature of this interaction depends on the type of a subunit present. Heteromeric recombinant receptors (a6y) containing an a1 subunit exhibit BZI-type pharmacology, receptors containing a2, a3 or a5 subunits exhibit BZII pharmacology, while receptors containing a4 or a6 subunits have a low affinity for both benzodiazepines and 6-carbolines. Studies involving site-directed mutagenesis of the various subunits have narrowed down even further the precise amino acid residues responsible for these differences in benzodiazepine pharmacology, as well as those involved in the binding of GABA. Altogether, such data suggest that GABA molecules bind at the interface of a and 6 subunits while benzodiazepines bind at the interface of a and y subunits.

The complexity afforded by different a, 6 and y subunits is increased further by the existence of the 8, e, 0 and % subunits. The 8 subunit preferentially associates with the a4 and a6 subunits. Receptors containing this subunit are unusual in having a particularly high affinity for GABA and muscimol and a reduced sensitivity to benzodiazepines and neurosteroids. The most recently cloned subunits, e, 0 and %, are the least well understood. The sequence of the e subunit is most closely related to that of the y subunits but studies in recombinant expression systems show that it assembles with a and 6 subunits to form receptors that are insensitive to benzodiazepines and show altered sensitivity to anaesthetics (pregnanolone, pentobarbital and propofol). The 0 subunit is most closely related to the 6 subunits; it coassembles with a, 6 and y subunits to form receptors with a low affinity for GABA, although other subunit combinations (notably a^0e or a0e) have been suggested. The e and 0 subunits have a fairly restricted pattern of expression that includes the hypothalamus and brainstem nuclei such as the locus coeruleus. The sequence of the % subunit is most closely related to that of the 6 subunits. Unlike the other GABAA subunits it is principally found in peripheral tissues, including lung, thymus, prostate and particularly the uterus.

Heterogeneity of native GABAa receptors

Given that the pharmacological and biophysical properties of recombinant GABAA receptors have been shown to depend critically on their subunit composition, much effort has been directed towards understanding the assembly of native receptors. This could provide a rational basis for the design of compounds able to interact with specific subpopulations of GABAA receptors in different brain regions that may be involved in different aspects of brain function. Clearly, many hundreds of different receptor types could arise from the assembly of 16 different subunits into a pentameric structure. However, numerous studies, involving the use of subunit-specific antibodies to localise or to purify receptor populations, have suggested that the restricted distribution and preferential assembly of these subunits results in the generation of no more than a dozen favoured receptor types. Of these, the most common receptor type is composed of a1, 62 and y2 subunits. Several lines of evidence suggest that the most likely stoichiometry of these receptors is 2a, 26 and 1y (although assemblies containing 2a, 16 and 2y have also been described). As indicated above, in less widely expressed assemblies, the ^ or e subunits can substitute for the y subunit, while the % and 0 subunits may co-assemble with a, 6 and y subunits.

Advances in our understanding of the functional significance of GABAA receptor heterogeneity have also come from studies of mice lacking specific subunit genes or expressing altered receptor subunits. To date, mutant mice have been generated that lack the a1, a6,62,63, y2 or ^ subunits (Rudolph et al. 2001; Sur et al. 2001). In the case of the y2 subunit deletion, neurons cultured from newborn mice show a complete lack of sensitivity to benzodiazepines (Gunther et al. 1995). By introducing a histidine residue (instead of the normal arginine) at position 101 in the a1 subunit of mice—making receptors containing this subunit insensitive to benzodiazepines — it has also been possible to determine which of the various effects of benzodiazepines are mediated by a1-containing receptors and which by receptors containing a2, a3 or a5 subunits. This approach showed that a1-containing GABAA receptors are involved in the sedative and amnesic actions of benzodiazepines (McKernan et al. 2000; Rudolph et al. 1999). Complementary experiments have shown that the anxiolytic actions of benzodiazepines are mediated by a2-containing receptors and the muscle-relaxant actions by a2- and a3-containing receptors (Rudolph et al. 2001; Crestani et al. 2001).

Was this article helpful?

0 0

Post a comment