Biochem. J. (1995) 310,1-9 (Printed in Great Britain) 1 REVIEW ARTICLE The GABAA receptors F. Anne STEPHENSON School of Pharmacy, University of London, 29/39 Brunswick Square, London WC1 N 1AX, U.K.
INTRODUCTION although the classical benzodiazepines, valium and librium, bind y-Aminobutyric acid (GABA) is the major inhibitory neuro- to a single high-affinity site in the brain, other related compounds, transmitter in the mammalian brain. It mediates its effects via which are again thought to act at the benzodiazepine or an the specific interaction with the integral membrane proteins, the adjacent overlapping site, show shallow displacement curves GABA receptors. GABA receptors can be classified according to in radioligand-binding studies. It is now known that these their respective transduction mechanisms following activation, compounds, for example the fi-carbolines, zolpidem and as the GABA and the GABAB receptor respectively. The focus oxoquazepam, discriminate between GABAA receptor subtypes of this review is the GABAA receptor. GABAA receptors are fast- (see below). acting ligand-gated chloride ion channels. Thus receptor ac- Other allosteric modulators of GABAA receptors include the tivation in the brain is followed within milliseconds by the gating barbiturate drugs, non-competitive chloride channel blocking or opening of an integral chloride ion channel which results, in agents [including picrotoxin and the cage convulsant compound, general, in the hyperpolarization of the recipient neuronal cell. t-butylbicyclophosphorothionate (TBPS)], certain neurosteroids, The GABAA receptors are of importance because of both the the anthelminthic agents, the avermectins, Zn2+, ethanol and the pivotal role that they play in the regulation of brain excitability anti-convulsant drug, loreclezole. Each of these compounds and the fact that their function is allosterically regulated by modulates GABAA receptor function by binding to a distinct site several distinct classes of therapeutic compounds. These include within the receptor complex. Figure 1 is a diagrammatic rep- anxiolytic benzodiazepines such as valium, barbiturates, neuro- resentation of the complex pharmacology of the mammalian steroids and some volatile anaesthetics. The first GABAA receptor GABAA receptors. gene sequences were reported in 1987 by Schofield et al. [1]. Since that time, a plethora of homologous genes encoding GABAA MOLECULAR BIOLOGY OF GABAA RECEPTORS receptor polypeptides have been identified. These polypeptides At the current time, 15 mammalian GABAA receptor genes have are thought to co-assemble, predominantly in the brain, to form been identified (for review see [6]). These are classified with heteromeric receptor proteins with distinct biophysical. and respect to the conservation in amino acid sequence homology of pharmacological properties. This review will focus on the their gene products. Thus there are five GABAA receptor subunit developments in this field since the GABAA receptor genes were types a, fi, y, a and p. Within four of these, isoforms exist with first cloned, addressing in particular, the structural and functional the now accepted nomenclature of al-a6, fll-,83, yl-y3 and significances of mammalian GABAA receptor heterogeneity. pl-p2. Isoforms of a single subunit type, for example comparing all the a subunits, have at least 70 % amino acid sequence GABAA RECEPTOR PHARMACOLOGY identity, whereas if the conservation in primary structure is The rich pharmacology ofthe GABAA receptors has had a major compared across subunit classes, for example a versus , subunits, impact in the elucidation of the biochemical properties of the the percentage identity falls to within the range 30-40 %. Some GABAA receptors. Before describing the current developments in of the GABAA receptor genes undergo alternative splicing. The the understanding of their structures, it is necessary to have an most prevalent of these is the y2 subunit which exists in two appreciation of their pharmacological properties. Perhaps the forms, y2Short (y2s) and y2lOng (y2L) [7,8]. But also, alternative most important of these was the realization that the anxiolytic splicing of the human ,B3- [9] and the rat a6-subunit gene have benzodiazepine drugs (e.g. valium and librium) exerted their been described [10]. In the chick, the GABAA receptor subunit action by a facilitation ofGABA neurotransmission (for detailed genes fl2 and f4 have been found to undergo splicing to yield reviews, see [2,3]). Thus it was found that benzodiazepines bound ,f2' and f84' [11,12]. (Note that for the chick, a fl4 and a y4 with high affinity to an allosteric modulatory site of the GABAA GABAA receptor subunit gene have been reported but it is receptor. In the presence of GABA, they potentiated the in- probable that these are the avian homologues of the mammalian hibitory response by an increase in the frequency of chloride ion ,f1- and y3-subunit genes respectively, [13].) Two invertebrate channel opening. The benzodiazepine mode of action permitted GABAA receptor genes have been cloned. They are the Drosophila the purification of the GABAA receptors by benzodiazepine Rdl gene [14], and a GABAA receptor gene from the pond snail, affinity chromatography (e.g. [4,5]). Benzodiazepines such as Lymnaea stagnalis [15]. Interestingly, in amino acid sequence valium, are anxiolytic. Although they are unable to activate comparisons with vertebrate GABAA receptor genes, these are GABAA receptors by themselves, by convention they are termed both most closely related to the , subunits, with identities of the agonists. Other benzodiazepines and related compounds thought order of 28 %. This low percentage ofconservation is in contrast to act at similar sites within the GABAA receptors have been to comparisons between the subunits of various vertebrate identified which are anxiogenic and, in the presence of GABA, species, where values of greater than 90 % are found. Figure 2 decrease the frequency of chloride channel opening. These shows the evolutionary tree of the GABAA receptor and the compounds are termed, by convention again, 'inverse agonists'. related glycine receptor subunit genes reported to date. The The benzodiazepine Ro 15-1788 has no or very low intrinsic lineage is consistent with the evolution from a single ancestral efficacy and is recognized as an antagonist. Furthermore, gene [16].
Abbreviations: GABA, y-aminobutyric acid; y2L, y2Long; y2s, y2Short; TBPS, t-butylbicyclophosphorothionate. 2 F. A. Stephenson
(a) 1 226 429 NH2 -A 0 Cys-Cys Ml M2 M3 M4 I 139 153 Clomethiazol ? 313 319 Splice variant y2L insert in a6 Avermectin ? Splice variant in ,3 (b)
'hols Channel-openers o (barbiturates, steroids, [Ca2+1; alcohols) Frusemide (anion-channel blockers)
Figure 1 Schematic presentation of the various binding sites associated with the GABAA receptors of mammalian brain Figure 3 Schematic representation of the pertinent features of GABAA receptors Abbreviations: PKC, protein kinase C; PTZ, pentylenetetrazol; [Ca2+]i, intracellular free Ca2+ concentration; misc., miscellaneous. Reprinted from [3] with the kind permission of (a) This shows the numbered amino acid sequence of the bovine al subunit highlighting F. Hoffmann-La Roche Ltd, Basel, Switzerland. domains conserved between all GABAA receptor polypeptides. V, consensus sequences for N-glycosylation; Ml-M4, transmembrane domains; Cys-Cys, conserved extracellular motif common to all members of the ligand-gated ion channel superfamily. (b) shows a schematic view of the receptor oligomer as viewed perpendicular to the plane of the membrane. It represents each of the five subunits of the receptor as a segment within the annular structure where the hole in the middle represents the chloride ion channel. The four membrane-spanning regions within each polypeptide are shown as filled circles with the predicted a helix of M2 lining the wall of the channel. The subunit complements and their ordering around the rosette are not definitively established, but see the text for discussion.
The mature GABAA receptor subunits have similar molecular masses as deduced from their respective cDNA sequences. These range from 48000 (y subunits) to a maximum of 64000 (a4 subunit). They all share a similar predicted domain structure not only between themselves but also with other members of the ligand-gated ion channel superfamily, most notably the proto- typic peripheral nicotinic acetylcholine receptor (cf. [17]). A schematic diagram of the pertinent features of a typical GABAA receptor subunit is shown in Figure 3. Thus it is predicted that each subunit has an extended, extracellular, hydrophilic N- terminal domain of the order of 220 amino acids. Within this region are consensus sequences for N-glycosylation. There are L GABc four putative hydrophobic domains, M1-M4, which are pre- GABrG ' GABm dicted to span the membrane and form the chloride ion channel. M2, again by analogy with nicotinic acetylcholine receptors and | ~~GABsB3 GABhRl also by some mutagenesis studies (see below), is thought to form GABhR2 GABcB4 the inner lining of the channel. The M I-M3 regions are adjacent ,GABrB2 to the N-terminal domain whereas M4 is at the C-terminal end GABbBl GABrBl of the proteins. This transmembrane topology model predicts 'GABhBl GABhB3 that both the N- and C-terminal regions are extracellular but, GABrB3 surprisingly, this remains unproven. Separating the M3 and M4 A GABcB3 transmembrane-spanning regions is a second hydrophilic region often referred to as the cytoplasmic or, alternatively, the intracellular loop domain. This region contains consensus sequences for phosphorylation by various protein kinases. It is Figure 2 Evolutionary tree of the GABAA and glycine receptor gene family also where the insertion of the eight amino acids in y2L is found. The Figure shows an evolutionary tree of published GABAA and glycine receptor genes. It is Splicing of the /33- and x6-subunit genes yields changes in the constructed from the DNA sequences of 270 shared codons derived from an initial protein extracellular N-terminal domain. For the /33 subunit, variant sequence alignment. The alignment includes the majority of the extracellular domain and the signal peptides were found [9]; for the a6 subunit, a deletion of transmembrane-spanning regions, M1-M4. Gene sequences are labelled, GAB for GABAA, GLY ten amino acids, rat numbering, a6 57-66, was discovered [10]. for glycine, organism, subunit type and number, where b = bovine; c = chicken; d = Another feature of note in the GABAA receptor polypeptides Drosophila; h = human; m = mouse; n = nematode; s = snail; A = a; B = /8; D = d; which is also conserved across all members of the G = y; R = p. Thus, GABcA1 = chicken GABAA receptor al subunit, etc. Further details of highly ligand- the analysis can be found in [16]. gated ion channel superfamily is a subdomain of the N-terminal GABAA receptors 3
GAA1 $BOVIN GAA2 $BOVIN region. This is a sequence of 15 amino acids which is predicted GAA3 $BOVIN QVESRRQEPGDFVKQDIGGLSP 22 to form a Cys-Cys , loop. Modelling studies suggested that this GAA5$RAT Q - 1 GAA4$BOVIN Q - 1 domain may be involved with agonist binding ([18]; but see also GAA6$RAT Q- 1 section on localization ofligand-binding domains within GABAA GAA1$BOVIN --QPSLQDELKDNTVFTRILDRLLDGYDNRLRPGLGERVTVKTDI VT 75 receptors). GAA2$BOVIN ---NIQEDEAKNNITIFTRILDRLLDGYDNRLRPGLG0SITWVFTNIYVT 75 Between the different subunit types, the regions of amino acid GAA3$BOVIN KRAPDIPPDDSTDNITIFTRIWRLL.DGYDNRLRPGLGDAVTEVKTDIYVT GAA5$RAT MPTSSVQDETNDNITIFTRILDGLLDGYDNRLRPGGERITQVRTDIYVT 82 conservation are found within the transmembrane-spanning GAA4$BOVIN ----ENQK(E LCPENFTRILDSLLDGYDNRLRPGFGGPVTEVKTDZYVT GAA6$RAT ----- LEDEGNFYSENVSRILDNJLEGYDNRJRPGFGGAVTEVKTDIYVT 6583. domains and also within defined regions of the N-terminal region. For the isoforms of one particular subunit type, in the GAA1 $BOVIN 125 alignment of the sequences, the divergent regions are found at GAA2$BOVIN GAA3 $BOVIN 150 the extreme N- and C-terminal domains and notably within the GAA5$RAT SFGPVSDTYIDFFFQSWID3RQI,8KGPMQRILPNNLLASKIWTPD 132 GAA4$BOVIN SFGPVSDVB4EYD FFRQITKRt.KYDGPITEILRNOOVTKVWTPD respective cytoplasmic loops. Figure 4 shows an alignment of GAA6$RAT SFGPVSDVENEYIDVFFROTWTDERLKFKGPAEItSINNL)WSKIWTPD 105 mammalian al-ax6 subunit primary structures to illustrate the of between subunit isoforms. GAAI$BOVIN HNGKKSVAHNMTMPNKLLRZTEDGMLYTYRLTRAECP8IHLEDFPM 1 I r, regions divergence GAA2$BOVIN TsFFHW HNTPNKLIW QDDGT YS TQBNLEDFPM 175 GAA3$BOVIN T FHNKKVHNTPNLESLRLDGLYMLTIHA MEDF F4 200 GAA5$RAT TFFH SKKSIAHNKPNKlLRLEDDGTLLYT14RLTISABCPMQLEDFPM 182 THE OLIGOMERIC STRUCTURES OF THE GABAA RECEPTORS GAA4$BOVIN T?FFRKKSVSHTAPNFRIXTLYTLTISSA3CYK LVDFP 181 GAA6$RAT TFFRNGKKSIA TPKLNRTNENGTLY'RTADCP RLVNFPM 155 It is well established by classical protein chemical methods and electron microscopy studies that the peripheral nicotinic acetyl- GAA1$BOVIN DAHACPLKFGSYAY PEVVYEWTREPARSVVVAEDGSRLNQYtLLGQTV GAA2 $BOVIN DAHSCPLKFGSYAYTTSEVTrYIWTYNASDSVQVAPFGSRLNQYDLPGQSI choline receptor has a pentameric quaternary structure, a2fly'. GAA3 $BOVIN DVHACPLKFGSYAYTTAVY ASWTLGKNKSVEVAtNDSRLNQYDLLGHVV GAA5 $RAT DAHACPLKFG8YAY'PNSEVVYVWTNGSTKSVVAEDGSRtLNQYHLMGQTV More recently, other members of the ligand-gated ion channel GAA4$BOVIN DGHACPLKFGSYAYPKSEMIY'rKGPEKSVEVPKESSSLVQYnLTGQTv superfamily including the strychnine-sensitive glycine receptor GAA6$RAT DGHACPLKFGSYAYPKS3ErIYTKKGPLYSVEVPEESSSLLQYDLIGQIV and a cloned neuronal nicotinic acetylcholine receptor were also
GAA1$BOVIN Vi;Wl Vyi~ JzMD VVUv Is-nrn.. J..* vVs&
Table 1 A summary of the characteristics of GABAA receptor polypeptides co-expression of a, /3, y and a subunit has been reported to from studies of cloned receptor proteins eliminate the benzodiazepine facilitation of the GABA-gated Abbreviation: BZ, benzodiazepine. Note: x can be 1, 2 or 3. chloride ion conductance [30], although a more recent report did observe diazepam potentiation of alflly2L, receptors [37]. Subunit type Comments Further, single-channel properties and comparative pharma- cological characterization of cells co-expressing these four sub- al Type-I BZ pharmacology in al/3xy2 combinations units show that they all do co-assemble rather than forming two a2 Type-It BZ pharmacology in a2flxy2 combinations distinct GABAA receptor subpopulations [37]. a3 Type-Il BZ pharmacology in a3,8xy2 combinations To summarize briefly, the results of the expression studies a4 a4flxy2 receptors have high affinity for Ro 15-4513 only suggest that brain GABAA receptors comprise probably penta- a5 Type-Il BZ pharmacology and very low affinity for Zolpidem meric receptors of either (a,3), (a/Jy), possibly (a,88), (afly8) and a6 a6,/xy2 receptors have high affinity for Ro 15-4513 only (p) subunit combinations. The ratios of the different subunits in N-terminal splice variant in rat is not known but will be discussed below. fit axflly2 receptors are loreclezole-insensitive each pentamer ,B2 ax,x2y2 receptors are loreclezole-sensitive Similarly, account must be taken of the variants within a single fl3 axxf3y2 receptors are loreclezole-sensitive subunit class i.e. (alfly) compared with (a2fly) etc. Again, N-terminal splice variant in human expression studies have yielded insight into the characteristics y1 axflxyl receptors bind BZ agonists only endowed by the isoforms of a subunit type. These studies are y2 This subunit confers a complete spectrum of BZ activities with described in the next section. al-3,5flx combinations Reduces Zn2+ sensitivity Splice variants y2s and y2L Characteristic features of GABAA receptor subunit isoforms as axflxy2L receptors are ethanol-sensitive y3 This subunit confers a complete spectrum of BZ activities with determined by the study of recombinant receptors al-3,5flx combinations As above, it is now accepted from expression studies that the pre- Can confer Zolpidem low affinity in the absence of a5 sence ofboth an a and a y subunit is required for benzodiazepine a6 No known characteristics p Forms robust homo-oligomers that are bicuculline-insensitive potentiation of the various cloned GABAA receptors. However, the type of a or y subunit has been shown to influence the subtype of benzodiazepine pharmacology. The aIflxy2 receptors have a high affinity for compounds such as CL 218 872 and the /3-carbolines, both of which were historically classified as benzo- biochemical methods. The properties of the various cloned diazepine receptor type-I ligands [25]. The a2flxy2, a3,8xy2 and receptors seem to be independent of the expression system a5,8xy2 have a low affinity for these ligands and therefore they employed. However, it should be noted that with the exception correlate with the benzodiazepine type-II site. A further sub- of single channel recordings, which have provided evidence for division has been made since the a5flxy2 receptors have an even functional heterogeneity, these studies have characterized the lower affinity for the drug Zolpidem [38]. The a4/3xy2 and properties of a 'population' of cloned receptors. Detailed mol- a6,8xy2 receptors bind only the benzodiazepine partial inverse ecular comparisons of recombinant receptors expressed in the agonist Ro 15-4513 with high affinity, whereas classic benzo- different systems has so far not been carried out. diazepines have very low affinity for these putative receptors Each GABAA receptor polypeptide, when expressed alone in [39,40]. This pharmacology correlates with the previously de- Xenopus oocytes, can form a GABA-gated homo-oligomeric scribed 'benzodiazepine-' or 'diazepam-insensitive' [3H]Ro 15-
chloride ion channel [28-30]. In general, however, the a, /3, y and 4513-specific binding sites [41]. There have been reports recently a subunits expressed alone assemble at low efficiency and it is of the co-expression of different a subunits together with fly probable that they do not form homo-oligomers in vivo. An which suggest that the four different types of subunit are co- exception is the p subunit. This does assemble efficiently to form assembling: ala3 by Verdoorn [42] and ala6 by Mathews et al. a functional GABA-gated ion channel [31]. Moreover, the p [43]. However, the benzodiazepine pharmacological profile of homomeric channels are reminiscent of invertebrate-type GABA these GABAA receptors has not been described. receptors in that they are insensitive to the classical GABAA For the y subunits, the y2 and y3 polypeptides behave receptor antagonist, bicuculline [32]. They are also barbiturate- identically in that their presence in (a/3y) constructs is required and benzodiazepine-insensitive [32]. The p-subunit mRNAs are for a full benzodiazepine agonist and inverse agonist profile found predominantly in the retina [31]. It may be that they form [29,35]. But in these trimeric receptors containing yl subunits, a subset of GABAA receptors distinct from those found in the the benzodiazepine inverse agonists behave as positive facilitators brain; they are sometimes termed GABAC receptors. [44]. Also relevant is that a recent report suggested that the y3 The co-expression of a single type of a and , subunits forms subunit could influence the benzodiazepine pharmacology since robust GABA-gated ion channels which are potentiated by al,8xy3 receptors behaved as a5,8xy2 recombinants with a low barbiturates and neurosteroids (e.g. [1,33]). But, although there affinity for zolpidem [45]. are some reports that these GABA currents are modulated by The presence of y subunits additionally reduces the sensitivity benzodiazepines [1,34], the consensus view currently is that for a of the cloned receptors to antagonism by Zn2+ [46,47]. Fur- full spectrum of reproducible benzodiazepine responsiveness the thermore, the presence of the y2L splice variant renders the co-expression of an a, and either y2 or y3 subunit is requisite receptors susceptible to potentiation by ethanol [48]. It is not [29,35,36]. Pairwise combinations such as ay, fly do not form clear whether this is a direct effect on the receptor itself or channels at high efficiency. The d-subunit has not been identified whether it occurs via a second messenger mechanism resulting in so far with a particular pharmacological characteristic so it is the activation of protein kinase C and subsequent more difficult to assess co-expression studies including this phosphorylation of the y2L subunit [48]. subunit. Notwithstanding this caveat, reports are consistent with Until recently, it was thought that the presence of the variant the idea that co-expression ofa subunit-containing paired subunit of the ,B subunit did not influence the physiological or pharma- combinations do not assemble with high efficiency [37]. The cological properties of the cloned GABAA receptors. GABAA receptors 5
Interestingly, Wafford et al. [49] have shown that the anti- Table 2 A summary of the subunit complements of native GABAA receptors convulsant drug loreclezole binds to a novel site of the GABAA * is the antibody specificity that was used for either immunoaffinity purification or receptor. They have localized this site of action to the TM2 immunoprecipitation studies. The co-associated subunits were then identified by immunoblotting region of the ,82 and 83 subunits. with the specific antibodies as indicated. Positive results only are shown, i.e. since each group A summary of the pertinent features of the GABAA receptor does not have the full repertoire of GABAA receptor subunit antibodies, not all polypeptides were subunits is given in Table 1. screened for in each case. ND is not detectable and parentheses indicate a low concentration. LOCALIZATION OF GABAA RECEPTOR mRNAS BY IN SITU Primary Experimental HYBRIDIZATION antibody* paradigm Co-associated subunits Reference In situ hybridization with GABAA receptor subunit-specific al Purification a1,ox2,cx3,fl3,y2 [57,100] oligonucleotide probes has shown that each respective mRNA Purification al,x3 [60] has a distinct pattern of expression. With the exceptions of the Precipitation al ,a3 [101] a6-subunit mRNA which is localized almost exclusively in a Precipitation a1,a3,sx5,,f2/3,y2,8 [62] single cell type, the mature cerebellar granule cell, and the p Precipitation a1,,82/3,y2L,y2s,a6 (implied) [64] subunits which are found in the different a2 Purification al,c2,a 3,/13,y2 [57,100] predominantly retina, a3 Purification o1,a2,a3,fl3,y2 [57,100] GABAA receptor subunit mRNAs have a heterogeneous dis- Purification a1,a3,(a5),,f2/3,y2,8 [62] tribution across many different cell types in adult brain. Fur- Precipitation al,a3 [58] thermore, each also undergoes developmental changes in patterns a4 Purification ca ,a4,,f2/3,y2 [61] of expression. There appears to be a more widespread het- a6 Purification al,a6,fl2/3,y2 ND a2,a3 [63] erogeneity in mRNAs expressed early in development compared Precipitation A6,y2,A ND al,a2,a3 [66] with the adult stage. In situ hybridization studies are not ,B2 Purification al,(a2),,#2,y2 [69] readily ND x3,c5,fl1 quantifiable and there is the further caveat that for heteromeric yl Purification a2,yl [71] receptors the levels of mRNA are not necessarily representative ND c1,a3,x6,y2,y3,d1 of functional receptors. Also, Primus et al. [50] reported that a Purification 1,ot2,a3,oc5,fl2/3,y1 [73] significant proportion of rat fl-subunit mRNA was non-poly- ND y2,y3 adenylated, thus introducing an extra variable. Nevertheless the Sy2 Purification a1,a3,(a5),,82/3,y2,dl [62] and can be made. y3 Purification a1,a2,a3,a4,a6,,f2/3,y2 following generalizations points ND oe5,y2 [45] In adult brain, of the a-subunit mRNAs, the al is the most Purification y2,y3 ND yl [71] ubiquitously and abundantly expressed, followed by the a2 and Purification al,x3,,f2/3,y2,8 [62] a3 mRNAs. As stated above, the a6 mRNA is confined to the cerebellar granule cell, the a5 mRNA has a low overall dis- tribution but is enriched particularly in the CAl and CA3 of the while regions hippocampus a4 transcripts are the rarest of The antigens used for the production of such antibodies have the a subunits but are enriched in thalamic regions. Specific neuronal cell have been identified been either synthetic peptides or bacterially derived fusion types which show the co- proteins, both of which correspond to the divergent regions of expression of at least two different a subunits; for example, the mitral cells of the olfactory bulb (al and a3) and the cerebellar the various GABAA receptor subunits. In general, the peptide granule cell (acl and a6). The a2-subunit mRNA is notably found sequences that have been employed successfully for GABAA in both neuronal cells and the Bergmann glial cells of the receptor subunit-specific antibody production are from the cerebellum. The most divergent N-terminal and C-terminal regions whereas the bac- abundant of the ,f-subunit mRNAs are the terial p32 and fl3, which both show a widespread distribution; the /31 fusion protein antigens are larger and encompass the mRNA is lower. Again there are identified cell types where the respective cytoplasmic loops (cf. Figure 4). The al-a6, fl1-,f3, different ,-subunit mRNAs co-localize, an example being once yl-y3 and a subunits have all been identified in brain membrane more the mitral cells of the olfactory bulb. The y2 subunit is the preparations by immunoblotting, by several different groups. most abundant of the y-subunit mRNAs, followed by yl, and The molecular masses of the immunoreactive polypeptides agree the y3 transcript is rare. The 8-subunit mRNA is enriched in the fairly well with the values predicted from the cDNA sequences cerebellar granule cell and the dorsolateral and ventrolateral once account is taken of potential mass contributions by N- nuclei of the thalamus. most linked carbohydrates (summarized in [55]). Similarly, the overall geniculate Thus, overall, the brain regional distributions and abundances of the different ubiquitously expressed subunits are the acl, a2, a3, p32, 8l3 and y2 subunits. Other isoforms show more restricted patterns of polypeptides determined qualitatively by immunoblotting and distribution. Furthermore, it is clear that some cell types express quantitatively by immunoprecipitation agree well with the more than five different GABAA receptor subunit transcripts, respective GABAA subunit mRNA localizations. However, the thus implying more than one receptor subtype per cell. (For more interesting question is which of the subunits co-assemble? more detailed analyses and correlates of the in situ hybridization Two approaches have been employed. GABAA receptor sub- studies see and for a populations have been purified by immunoaffinity chromato- [51-53] developmental profile [54].) graphy and the co-associating subunits identified by immuno- blotting. Alternatively, receptor subpopulations have been THE POLYPEPTIDE COMPLEMENTS OF NATIVE GABAA immunoprecipitated from detergent extracts of different brain RECEPTORS DETERMINED USING SUBUNIT-SPECIFIC regions, then the immune pellets were analysed for co-immuno- ANTIBODIES precipitated subunits. So far, most information has been The distributions of respective subunit mRNAs are not necess- forthcoming for the different a subunits of the GABAA receptors arily representative ofthe localizations offunctional, heteromeric because high-affinity, specific antibodies were readily obtained. receptor proteins. Thus a natural progression in the deter- The salient features of all these studies are summarized below. mination of the subunit complements ofnative GABAA receptors Table 2 displays the results from different laboratories in more has been the production and use of subunit-specific antibodies. detail. 6 F. A. Stephenson
With respect to the a subunits, there is universal agreement and by analogy with the experiments carried out for the a that the majority of native GABAA receptors contain a single subunits proposed the existence of two y subunits per receptor type of a subunit (e.g. [56-58]). The most abundant of the a molecule. This is in agreement with two other studies. A subunit subunits in the adult rat brain is the al subunit [59]. The ratio of (al)2(/1)1(y2)2 was deduced in an elegant study by benzodiazepine pharmacology of the various GABAA receptor Backus et al. [72] by comparing the conductance properties of a-subunit-enriched receptors was similar in profile to that found (acl/32y2) wild-type and mutant cloned receptors. Similarly, Khan for the corresponding (acx/y) cloned receptors [58] and further et al. [64] deduced by immunoprecipitation studies that one correlation was thus with the Type-I, Type-IT pharmacological subpopulation of receptors in the cerebellum comprised subclasses (see above and [25]). However, in some studies it was ala6/32/3y2Ly2S. Conflictingly, however, Togel et al. [45] and reported that two different a subunits could be detected in Mossier et al. [73] purified y3 and yl GABAA subunit-containing immunoaffinity-purified preparations or immune precipitates. receptors respectively, but in these studies no evidence was found For example, the co-existence of the al subunit with either the for the co-existence of the different y polypeptides. In the y3- a2, a3, a4, a5 or a6 subunit and the a2 with the a3 subunit has subunit-enriched receptors no c5-subunit immunoreactivity was been described ([57,60-63]; Table 2). GABAA receptors in detected either, yet a low affinity for zolpidem was found ([45] detergent solution have a tendency to aggregate [21]. However, and see above). control experiments using different-specificity antibody columns For the a subunit, again there are inconsistencies between in series demonstrated that the co-existence of two a-subunit groups. Quirk et al. [66] found an exclusive co-localization of the variants was not artefactual. For example, purification of a subunit with the a6 polypeptide in the cerebellum. This (c61) GABAA receptors on an a2- and an a3-subunit antibody column subunit population did not bind benzodiazepine agonists, in series yielded purified material containing a2 and a3 but antagonists or even inverse agonists with high affinity. No y2 not the most abundant, al, subunit immunoreactivity [63]. immunoreactivity was detected in these cerebellar receptors. This Furthermore, because only two different a-subunit immuno- is in contrast to the postulated co-existence of ac/y8 subunits by reactivities have been reported as co-existing within a single single-channel studies of cloned receptors (see above). Fur- receptor oligomer, this has formed the basis for the argument thermore, Mertens et al. [62] found that the a subunit was more that there are probably two a subunits per receptor molecule widely expressed and co-existed with al, a3, /32/3 and y2 [57,63]. It should be stressed, however, that there is no in- subunits; the a6 subunit was not investigated. formation as yet regarding the a-subunit stoichiometry in native It was cited earlier that Burt and Kamatchi [23] calculated a single a-variant GABAA receptors. The GABAA receptors con- possible existence of 151 887 different GABAA receptors. A taining two types of a subunit have been identified as minority subunit stoichiometry, (a)2(/)(y)2 is the most consistent with all subpopulations with the exception of the co-existence of the al published work (see above and also below for further discussion). and a6 subunits. Khan et al. [64] deduced that this population Thus, with the assumptions that within receptor oligomers, accounts for 500% of all the a6-subunit-containing GABAA different a subunits co-exist, different /3 and y subunits do not, receptors, which is in agreement with direct findings ofPollard et the a subunit substitutes for a y, and p subunits assemble as al. [65]. Although another group did not find evidence for the co- homo-oligomers, this value reduces to a more manageable 137. existence of al and a6 subunits this may be explained by the use (It is realised that the assumptions made are inconsistent with oflower-avidity antibodies [66]. There have recently been reports some reports and may be an underestimation; cf. Table 2. whereby two different a subunits were co-expressed with (fly) Nevertheless, the characterization of the native receptors has subunits and indeed the functional properties of these were significantly reduced receptor numbers.) distinct from single a-subunit receptors, thus suggesting the co- assembly of all four subunit types ([42,43] and see section on SIGNIFICANCES OF RECEPTOR characteristic features of GABAA receptor subunit isoforms as THE FUNCTIONAL GABAA determined by the study of recombinant receptors, above). The HETEROGENEITY? benzodiazepine pharmacologies of these mixed a-subunit re- The questions which immediately arise following the discovery of ceptors is not known. For the ala6/3y receptors, results are the multiple receptor subtypes is, why are there so many and consistent with a 'benzodiazepine-insensitive' pharmacology, i.e. what distinct functions, if any, do they subserve? It can be a dominance of the a6 subunit [64,65]. argued that all the different GABAA receptors serve the same Such detailed studies have not been carried out for the /3 function in that they are all fast-acting, ligand-gated chloride ion subunits, probably because it has been more difficult to raise channels and that GABAA receptor heterogeneity has evolved useful isoform-specific antibodies. (The intracellular loops of the with the conservation of the important part of the protein, the / subunits are more highly conserved than for the other subunit chloride ion channel, with all other changes being redundant. classes and, additionally, / subunits do not have a hydrophilic C- The diametrically opposed view, however, is that the GABAA terminus.) Many studies have therefore utilized the monoclonal receptor genes have indeed evolved to serve distinct functions. antibody bd-17, which recognizes both the /32 and /33 but not /31 Support for this argument stems from the facts that each GABAA subunits [67,68]. Benke et al. [69], however, recently reported the receptor gene shows a characteristic pattern of temporal and production of /31-, /32- and ,/3-subunit antibodies. In this study, spatial expression (see above) and additionally, that recombinant it was shown that the most prevalent / subunit was /2, which receptors do have subtly different properties. A complete under- was most often associated with the al subunit [69]. Moreover, standing of the role played by GABAA receptor subtypes will quantitative immunoprecipitation studies indicated the presence only come from a fully integrated approach involving molecular of a single type of / subunit per receptor [69]. For all the GABAA properties through to behavioural neuroscience. Some points are receptor subpopulations that have been identified, a / subunit made below. was also present (Table 2). This is consistent with the identi- Cloned receptors of defined subunit composition have fication of the / subunit as the high-affinity agonist-binding site differential sensitivities to activation by GABA and different ([70] and see below). desensitization rates (e.g. [24,74]). These properties may be Of the y subunits, the y2 subunit is the most prevalent. Quirk important during development where GABA acts as both a et al. [71] showed the co-association of the y2 and y3 subunits, trophic agent and as a neurotransmitter (see [54] for discussion). GABAA receptors 7
Recombinant GABAA receptors also have a differential sen- agonist benzodiazepine-binding sites per receptor molecule. sitivity to allosteric modulation by endogenous molecules found Specific activity determinations for GABAA receptors purified in the brain which include the neurosteroids and Zn2+ [33,46,47]. from mammalian brain by benzodiazepine affinity chromato- GABAA receptors which differ in their benzodiazepine phar- graphy range from 0.8-3 benzodiazepine-binding sites/mg of macological profile cannot be discussed in the context of their protein (summarized in [86]). These values are low for a pure physiological functions since the current consensus view is that protein with one binding site per 250000 Da receptor molecule there are no endogenous benzodiazepine-like compounds. These and are attributed to receptor inactivation during isolation. Two binding sites therefore remain serendipitous for both the mol- benzodiazepine-binding sites per receptor would require an even ecular dissection and classification of GABAA receptors and, higher theoretical specific activity for purification to homogen- importantly, for exploitation in the treatment of anxiety, etc. eity. Binding site stoichiometry remains an open question. Many of the GABAA receptor subunits contain consensus Specific amino acids within the a-subunit primary structures amino acid sequences for phosphorylation by protein kinase A, which determine the type of benzodiazepine pharmacology have protein kinase C and tyrosine kinase. Some of the subunits have been identified by site-directed mutagenesis. First, histidine 101 been shown to be phosphorylated in vitro with concomitant (rat amino acid numbering), which notably falls within the modulation of channel activity. Furthermore, native GABAA sequence identified from the protein chemistry studies, is requisite receptors show differential responses following activation of for high-affinity agonist benzodiazepine binding [87]. This histi- different protein kinases (see e.g. [75] for more detailed dis- dine is not conserved in the a4- and ac6-subunit sequences, which cussion), thus suggesting that regulation of receptor subtype have high affinity for the inverse agonist, Ro 15-4513, only. The activity by intracellular mechanisms may be important. mutation, R1OOH, increases the affinity of (a6RI00H)/3xy2 In whole-animal studies, the a6 GABAA receptor subunit has receptors for flumazenil and diazepam but the K; values de- been implicated in cerebellar motor control [76]. Interestingly termined are still 6-30-fold lower than found for the al/3xy2 also, the /33-subunit gene is deleted in the neurological disorder, wild-type receptor [86]. Secondly, Pritchett and Seeburg [88] Angelman's syndrome [77]. The mouse pink-eyed cleft-palate converted a cloned type-II receptor to a receptor with type-I (pCP) mutation has a deletion of a region which contains the at5, benzodiazepine pharmacology by a single point mutation. Thus /33, and y3 GABAA subunit genes [78]. This deletion is hom- glycine 201 (al numbering) was identified as a determinant of ologous to the one found for Angelman's syndrome in humans high-affinity binding for type-I-selective compounds [88]. More and these mice show neurological dysfunction. Interestingly, recently and interestingly, Mihic et al. [89] identified a residue in however, it was recently reported that one strain had deletions the y2 subunit, Thr-142, which was important for the de- which resulted in no expression of the a5- and y3-subunit genes, termination of benzodiazepine efficacy. Some previous publi- yet these mice were neurologically normal [79]. cations identified this amino acid as Ser-142 but these now appear to be incorrect [89]. Comparison of the mutant acl/I1y2LT142S with the wild-type acl11y2LT142 yielded receptors LOCALIZATION OF LIGAND-BINDING DOMAINS WITHIN GABAA where the antagonist, Ro 15-1788, and the partial inverse agonist, RECEPTORS Ro 15-4513, both acted as partial agonists; potentiation by conventional benzodiazepine agonists was increased 2-fold but The initial insight into the localization of drug-binding sites no effect was found on GABA affinity, efficacy, or barbiturate or within the GABAA receptors was gained from photoaffinity- alphaxolone modulation of GABA responses [89]. labelling studies. Thus it was shown that the benzodiazepine As described above, [3H]muscimol photoaffinity-labelled pre- [3H]flunitrazepam, and the high-affinity GABAA receptor dominantly a 58000 Da subunit which it was assumed was the agonist [3H]muscimol, photoaffinity-labelled two different subunit although this has not actually been proven. [3H]Muscimol GABAA receptor subunits. These were distinguished on the basis reportedly also specifically photoaffinity-labelled the 53000 Da of molecular mass [80,81] and they were at the time widely acl subunit, albeit at a much reduced efficiency compared with the recognized as the a (51 kDa; benzodiazepine-binding) and / 58000 Da subunit. Thus again the experimental evidence (58 kDa; GABA-binding) subunits respectively, suggesting that available is most consistent with a high-affinity binding site at the these allosterically coupled binding sites were located on different interfaces between subunits, this time the ac and /3. Amin and subunits of GABAA receptors. Heterogeneity of polypeptides Weiss [70] showed that two homologous domains of the photoaffinity-labelled by [3H]flunitrazepam was reported as early subunit in (alp12y2) cloned receptors are required for activation as 1980 [82]. This heterogeneity was particularly evident in of the receptors by GABA. These were identified as Binding neonatal rat brain and furthermore showed a pharmacological Domain I, 32 (157-160) sequence YGYT, and Binding Domain specificity of photolabelling reminiscent of type-I and type-II II, /32 (202-205) sequence TGSY. The tyrosine and threonine receptors (summarized in [83]). It is now known that this was due residues were identified by mutagenesis as points of interaction to the presence of the minor a-subunit isoforms. The discovery with GABA. Binding Domain I is completely conserved across of the requirement of the y2 subunit for benzodiazepine modu- the subunits of rat and cow. The motif, TGXY, of Binding lation of GABA responses led to a re-evaluation of the Domain II is also conserved. Weiss and Amin [90] have further [3H]flunitrazepam photoaffinity-labelling reaction. Using identified homologous amino acids in p1 homomeric receptors as subunit-specific antibodies, it was confirmed that the al, not the being important for GABA activation, i.e. Tyr-198, Thr-244 and y2, subunit was the site of the photoaffinity-labelling reaction Tyr-247. For the Binding Domain I homologous sequence in the [84]. Indeed, the site of labelling has been reduced to within the glycine receptor al subunit, Gly-160 has been identified as an acl (59-148) region, by cyanogen bromide cleavage and antibody important determinant of glycine and antagonist binding. Fur- mapping experiments [85]. Thus in order to explain the role of thermore and interestingly, mutations adjacent to Gly-160, i.e. both the a and y subunits in agonist benzodiazepine binding, it R159Y and Y161R, result in a glycine al homomeric receptor is proposed that this site is at the interface between these two which shows gating by GABA [91]. Sigel et al. [92], however, subunits. Note that with the working model of (a)2(,8)1(y)2 for showed that the mutation F64L in the al subunit of al/32y2 the most prevalent GABAA receptor of adult mammalian brain receptors resulted in a 200-fold reduction in agonist affinity. This (see above discussion), simplistically this would indicate two same residue, Phe-641, was shown to have incorporated radio- 8 F. A. Stephenson
end of TM2, as the determinant important for high-affinity loreclezole binding. The mutagenesis studies are summarized in diagrammatic form in Figure 5.
CONCLUSIONS Seven years have elapsed since the first GABAA and glycine receptor gene sequences were reported. That led to the realization that there existed a ligand-gated ion channel superfamily [1]. Since then, a wealth ofinformation has been forthcoming on the molecular and pharmacological properties of GABAA receptor genes. The major challenge now, as indeed for many of the neurotransmitter receptor proteins, is to integrate all this detail into a cohesive understanding of the functional significance of GABAA receptor heterogeneity in normal and pathological conditions. An additional future direction will include an under- standing of the regulation of GABAA receptor gene expression. A developmental switch in the expression of GABAA receptor of receptor expressed subtypes in cerebellar granule cells has recently been described Figure 5 Schematic model of the major type GABAA BSF 1 has been identified as a in adult mammalian brain, hghlighting amino acids Implicated In ligand [43,98]. Further, the protein binding neuronal-cell-type-enriched factor involved in the differential transcription ofthe GABAA receptor 8-subunit gene [99]. Further The subunit stoichiometry most consistent with all published experimental results is investigation of these early reports together with the use of both (al)2(fl2/3)1(y2)2. The Figure shows an al (red), fl2 (light grey) and y2 (dark grey) subunit and knock-out mice should provide of the a and ,t subunits and residues GABAA receptor transgenic with the binding pocket for GABA at the interface into function and regulation of gene expression. important for benzodiazepine subpharmacology at the interface between the a and y subunits. insights Note that in accord with the studies of Unwin [17] on the peripheral nicotinic acetylcholine receptor, the GABA binding site is depicted as approximately half-way down the extracellular I would like to thank the members of my group, Simon Pollard, Dr. Chris L. region. The leucine residues (L) in the transmembrane region M2 are those conserved with the Thompson, Dr. Paul Chazot and Dr. Miroslav Cik, who have contributed to this review nicotinic receptor, which for that protein have been identified as the gate within the channel not only practically (S.P. and C.L.T.) but with many useful discussions. I thank lumen [17]. This GABAA receptor will have type-I benzodiazepine pharmacology and, Professor G. G. Lunt and Dr. M. Ortells for kindly providing Figure 3 and Dr. Cik for additionally, it is loreclezole-sensitive. help with Figures 4 and 5. Referenced work from the author's laboratory was funded by the Medical Research Council (U.K.). activity following [3H]muscimol affinity labelling in the sequence REFERENCES to al, a2, a3 and a5 subunits [93]. TIDVF, which is unique the 1 Schofield, P. R., Darlison, M. G., Fujita, N., Burt, D. R., Stephenson, F. A., Rodriguez, But note that it is unlikely that this is the major amino acid H., Ramachandran, J., Reale, V., Glencorse, T. A., Seeburg, P. H. and Barnard, E. A. photoaffinity-labelled by [3H]muscimol since predominant label- (1987) Nature (London) 328, 221-227 ling is associated with a 58000 Da band and although the a3 2 Olsen, R. W. and Venter, J. C. (eds.) (1986) in Benzodiazepine/GABA-Receptors and subunit has a molecular mass in this vicinity, it is not a major Chloride Channels: Structural and Functional Properties, Alan R. Liss, New York component ofbenzodiazepine affinity-purified GABAA receptors. 3 Haefely, W. E. (1994) in The Challenge of Neuropharmacology (Mohler, H. and Switzerland Picrotoxinin is a classical non-competitive antagonist of Da Prada, M., eds.), pp. 15-39, Editiones Roche, Basel, F. C. and Barnard, E. A. (1983) J. Biol. Chem. GABAA receptors and it has been proposed that it acts at the 4 Sigel, E., Stephenson, A., Mamalaki, studies have 258, 7219-7223 level of the chloride ion channel. Mutagenesis 5 Stauber, G. B., Ransom, R. W., Dilber, A. and Olsen, R. W. (1987) Eur. J. Biochem. identified the residues within the predicted M2 domain which are 167,125-133 important for picrotoxin-sensitivity. Thus, a GABAA receptor 6 Olsen, R. W. and Tobin, A. J. (1990) FASEB J. 4, 1469-1480 subunit of a Drosophila mutant Rdl which was resistant to the 7 Whiting, P. J., McKernan, R. M. and Iversen, L. L. (1990) Proc. Nati. Acad. Sci. insecticide dieldrin (and picrotoxin) had a single point mutation, U.S.A. 87, 9966-9970 A302S, in the putative TM2 region [94]. Furthermore, the co- 8 Kofuji, P., Wang, J. B., Moss, S. J., Huganir, R. L. and Burt, D. B. (1991) J. Neurochem. 56, 713-715 expression of the glycine receptor ft subunit with the al subunit channel picrotoxinin-insensitive [95]. 9 Kirkness, E. F. and Fraser, C. M. (1993) J. Biol. Chem. 268, 4420-4428 renders the glycine-gated 10 Korpi, E. R., Kuner, T., Kristo, P., Kohler, M., Herb, A., Luddens, H. and Seeburg, Residues within the putative TM2 domain of the subunit were P. H. (1994) J. Neurochem. 63, 1167-1170 subsequently identified as important for the endowment of this 11 Harvey, A. J., Chinchetru, M. A. and Darlison, M. G. (1994) J. Neurochem. 62, property [95]. More recently, Gurley et al. [96] showed directly 10-16 that the mutant receptor al(fi2T246F)y2 was picrotoxin-in- 12 Bateson, A. N., Lasham, A. and Darlison, M. G. (1991) J. Neurochem. 56, sensitive. These reports thus suggest that the TM2 domain is 1437-1 440 important for picrotoxinin binding and they can be extrapolated 13 Darlison, M. G. and Albrecht, B. E. (1995) Semin. Neurosci., in the press 14 R. H., D. P., Shaffer, C. D., Maclntyre, R. J. and Roush, as evidence for the M2 region forming the inner lining ffrench-Constant, Mortlock, providing R. T. (1991) Proc. Nati. Acad. Sci. U.S.A. 86, 7209-7213 of the chloride ion channel. 15 Harvey, R. J., Vreugdenhil, E., Zaman, S. H., Bhandal, N. S., Usherwood, P. N. R., The TM2 domain is also implicated in the binding ofloreclezole Barnard, E. A. and Darlison, M. G. (1991) EMBO J. 10, 3239-3245 to the GABAA receptor. As described above, loreclezole sen- 16 Ortells, M. 0. and Lunt, G. G. (1995) Trends Neurosci. 18, 121-127 sitivity was found with ,/2- and ,f3-subunit-containing receptors 17 Unwin, N. (1993) J. Mol. Biol. 229,1101-1124 [49]. Wingrove et al. [97] have further refined these studies and 18 Cockcroft, V. B., Osguthorpe, D. J., Barnard, E. A. and Lunt, G. G. (1990) Proteins 8, identified Asn-289 (p2) and Asn-290 (f3), located at the carboxyl 386-397 GABAA receptors 9
19 Langosch, D., Thomas, L. and Betz, H. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 59 Benke, D., Mertens, S., Trceziak, A., Gillessen, D. and Mohler, H. (1991) J. Biol. 7394-7398 Chem. 266, 4478-4483 20 Cooper, E., Couturier, S. and Ballivet, M. (1991) Nature (London) 350, 235-238 60 Zezula, J. and Sieghart, W. (1991) FEBS Lett. 284,15-18 21 Mamalaki, C., Barnard, E. A. and Stephenson, F. A. (1989) J. Neurochem. 52, 61 Kern, W. and Sieghart, W. (1994) J. Neurochem. 62, 764-769 124-134 62 Mertens, S., Benke, D. and Mohler, H. (1993) J. Biol. Chem. 268, 5965-5973 22 Nayeem, N., Green, T. P., Martin, I. L. and Barnard, E. A. (1994) J. Neurochem. 62, 63 Pollard, S., Duggan, M. J. and Stephenson, F. A. (1993) J. Biol. Chem. 268, 815-818 3753-3757 23 Burt, D. R. and Kamatchi, G. L. (1991) FASEB J. 5, 2916-2923 64 Khan, Z. U., Gutierrez, A. and De Blas, A. L. (1994) J. Neurochem. 63, 371-374 24 Levitan, E. S., Schofield, P. R., Burt, D. R., Rhee, L. M., Wisden, W., Kohler, M., 65 Pollard, S., Thompson, C. L. and Stephenson, F. A. (1995) J. Biol. Chem., in the Fujita, N., Rodriguez, H., Stephenson, F. A., Darlison, M. G., Barnard, E. A. and press Seeburg, P. H. (1988) Nature (London) 335, 76-79 66 Quirk, K., Gillard, N. P., Ragan, C. I., Whiting, P. J. and McKernan, R. M. (1994) 25 Pritchett, D. B., Luddens, H. and Seeburg, P. H. (1989) Science 245, 1389-1392 J. Biol. Chem. 269, 16020-16028 26 Hadingham, K. L., Harkness, P. C., McKernan, R. M., Quirk, K., Le Bourdelles, B., 67 Schoch, P., Richards, J. G., Haring, P., Takacs, B., Stahli, C., Staehelin, T., Haefely, Horne, A. L., Kemp, J. A., Barnard, E. A., Ragan, C. I. A. and Whiting, P. J. (1992) W. and Mohler, H. (1985) Nature (London) 314,168-171 Proc. Natl. Acad. Sci. U.S.A. 89, 6378-6382 68 Ewert, M., Shivers, B. D., Luddens, H., Mohler, H. and Seeburg, P. H. (1990) 27 Pregenzer, J. F., Im, W. B., Carter, D. B. and Thomsen, D. R. (1993) Mol. Pharmacol. J. Cell Biol. 110, 2043-2048 43, 801-806 69 Benke, D., Fritschy, J.-M., Trzeciak, A., Bannwarth, W. and Mohler, H. (1994) 28 Blair, L. A. C., Levitan, E. S., Marshall, J., Dionne, V. E. and Barnard, E. A. (1988) J. Biol. Chem. 269, 27100-27107 Science 242, 577-579 70 Amin, J. and Weiss, D. S. (1993) Nature (London) 366, 565-569 29 Pritchett, D. B., Sontheimer, H., Shivers, B. D., Ymer, S., Kettenmann, H., Schofield, 71 Quirk, K., Gillard, N. P., Ragan, C. I., Whiting, P. J. and McKernan, R. M. (1994) Nature (London) 338, 582-585 Mol. Pharmacol. 45,1061-1070 P. R. and Seeburg, P. H. (1989) H. and 30 Shivers, B. D., Killisch, I., Sprengel, R., Sontheimer, H., Kohler, M., Schofield, P. R. 72 Backus, K. H., Arigoni, M., Drescher, U., Scheurer, L., Malherbe, P., Mohler, 3, 327-337 Benson, J. A. (1993) NeuroReport 5, 285-288 and Seeburg, P. H. (1989) Neuron (1994) J. Biol. Chem. 269, 31 Cutting, G. R., Lu, L., O'Hara, F., Kasch, L. M., Montrose-Ratizadeh, C., Donovan, 73 Mossier, B., Togel, M., Fuchs, K. and Sieghart, W. W. B., Uhl, G. R. and Kazazian, 25777-25782 D. M., Shimada, S., Antonarakis, S. E., Guggino, and Barnard, E. A. (1988) Neuron 1, H. H. (1991) Proc. Nati. Acad. Sci. U.S.A. 88, 2673-2677 74 Levitan, E. S., Blair, L. A. C., Dionne, V. E. 773-781 32 Shimada, S., Cutting, G. and Uhl, G. R. (1992) Mol. Pharmacol. 41, 683-687 Moss, S. J. and Smart, T. G. P. H. and Costa, E. (1991) Mol. Pharmacol. 39, 75 Krishek, B. J., Xie, X., Blackstone, C., Huganir, R. L., 33 Puia, G., Vicini, S., Seeburg, (1994) Neuron 12, 1081-1095 691-696 P. H. (1993) Nature Neuron 5, 76 Korpi, E. R., Kleingoor, C., Kettenmann, H. and Seeburg, 34 Sigel, E., Baur, R., Trube, G., Mohler, H. and Malherbe, P. (1990) (London) 361, 356-359 703-711 F., Kirkness, E. F. and Lalande, M. (1993) and Seeburg, P. H. (1992) 77 Sinnett, D., Wagstaff, J., Glatt, K., Woolf, 35 Herb, A., Wisden, W., Luddens, H., Puia, G., Vicini, S. Am. J. Human Gen. 52, 1216-1229 Proc. NatI. Acad. Sci. U.S.A. 89, 1433-1437 78 Nakatsu, Y., Tyndale, R. F., DeLorey, T. M., Durham-Pierre, D., Gardner, J. M., 36 Knoflach, F., Rhyner, Th., Villa, M., Kellenberger, S., Drescher, U., Malherbe, P., McDanel, H. J., Nguyen. Q., Wagstaff, J., Lalande, M., Sikela, J., Olsen, R. W., Sigel, E. and Mohler, H. (1991) FEBS Lett. 293,191-194 Tobin, A. J. and Brilliant, M. H. (1993) Nature (London) 364, 448-450 37 Saxena, N. C. and MacDonald, R. L. (1994) J. Neurosci. 14, 7077-7086 79 Culiat, C. T., Stubbs, I. J., Montgomery, C. S., Russell, I. B. and Rinchik, F. M. 38 Pritchett, D. B. and Seeburg, P. H. (1990) J. Neurochem. 54,1802-1804 (1994) Proc. Natl. Acad. Sci. U.S.A. 91, 2815-2818 39 Luddens, H., Pritchett, D. B., Kohler, M., Killisch, I., Keinanen, K., Monyer, H., 80 Casalotti, S. 0., Stephenson, F. A. and Barnard, E. A. (1986) J. Biol. Chem. 261, Sprengel. R. and Seeburg, P. H. (1990) Nature (London) 346, 648-651 15013-15016 40 Wisden, W., Herb, A., Weiland, H., Keinanen, K., Luddens, H. and Seeburg, P. H. 81 Deng, L. R., Ransom, R. W. and Olsen, R. W. (1986) Biochem. Biophys. Res. (1991) FEBS Lett. 289, 227-230 Commun. 138, 1308-1314 41 Sieghart, W., Eichinger, A., Richards, J. G. and Mohler, H. (1987) J. Neurochem. 48, 82 Sieghart, W. and Karobath, M. (1980) Nature (London) 286, 285-287 46-52 83 Sieghart, W. (1989) Trends Pharmacol. Sci. 10, 407-411 42 Verdoorn, T. (1994) Mol. Pharmacol. 45, 475-480 84 Stephenson, F. A., Duggan, M. J. and Pollard, S. (1990) J. Biol. Chem. 265, 43 Mathews, G. C., Bolos-Sy, A. M., Holland, K. D., Isenberg, K. E., Covey, D. F., 21160-21165 Ferrendelli, J. A. and Rothman, S. M. (1994) Neuron 13, 149-158 85 Stephenson, F. A. and Duggan, M. J. (1989) Biochem. J. 264, 199-206 44 Ymer, S., Draguhn, A., Wisden, W., Werner, P., Keinenen, K., Schofield, P. R., 86 Stephenson, F. A. (1988) Biochem. J. 249, 21-32 Sprengel, R., Pritchett, D. B. and Seeburg, P. H. (1990) EMBO J. 9, 3261-3267 87 Wieland, H. A., Luddens, H. and Seeburg, P. H. (1992) J. Biol. Chem. 267, 45 Togel, M., Mossier, B., Fuchs, K. and Sieghart, W. (1994) J. Biol. Chem. 269, 1426-1 429 12993-1 2998 88 Pritchett, D. B. and Seeburg, P. H. (1991) Proc. Nati. Acad. Sci. U.S.A. 88, 46 Draguhn, A., Verdorn, T. A., Ewert, M., Seeburg, P. H. and Sakmann, B. (1990) 1421-1 425 Neuron 5, 781-788 89 Mihic, S. J., Whiting, P. J., Klein, R. L., Wafford, K. A. and Harris, R. A. (1994) 47 Smart, T. G., Moss, S. J., Xie, X. and Huganir, R. L. (1991) Br. J. Pharmacol. 103, J. Biol. Chem. 52, 32768-32773 1837-1839 90 Weiss, D. S. and Amin, J. (1994) Soc. Neurosci. Abstr. 20, 13.4 48 Wafford, K. A., Burnett, D. M., Leidenheimer, N. J., Burt, D. R., Wang, J. B., Kofuji, 91 Schmieden, V., Kuhse, J. and Betz, H. (1993) Science 262, 256-258 P., Dunwiddie, T. V., Harris, R. A. and Sikela, J. M. (1991) Neuron 7, 27-33 92 Sigel, E., Baur, R., Kellenberger, S. and Malherbe, P. (1992) EMBO J. 11, 49 Wafford, K. A., Bain, C. J., Quirk, K., McKernan, R. M., Wingrove, P. B., Whiting, P. J. 2017-2023 and Kemp, J. A. (1994) Neuron 12, 775-782 93 Smith, G. B. and Olsen, R. W. (1994) J. Biol. Chem. 269, 20380-20387 50 Primus R. J., Jacobs, A. A. and Gallager, D. W. (1992) Brain Res. 14, 179-185 94 ffrench-Constant, R., Rocheleau, T. A., Steichen, J. C. and Chalmers, A. E. (1993) 51 Wisden, W., Laurie, D. J., Monyer, H. and Seeburg, P. H. (1992) J. Neurosci. 12, Nature (London) 363, 449-451 1040-1062 95 Pribilla, I., Takagi, T., Langosch, D., Bormann, J. and Betz, H. (1992) EMBO J. 11, 52 Laurie, D. J., Seeburg, P. H. and Wisden, W. (1992) J. Neurosci. 12, 1063-1076 4305-4311 53 Persohn, E., Malherbe, P. and Richards, J. G. (1992) J. Comp. Neurol. 326, 96 Gurley, D. A., Amin, J., Ross, P., Weiss, D. and White, G. (1994) Soc. Neurosci. 193-216 Abstr. 20, 509 54 Laurie, D. J., Wisden, W. and Seeburg, P. H. (1992) J. Neurosci. 12, 4151-4172 97 Wingrove, P. B., Wafford, K. A., Bain, C. and Whiting, P. J. (1994) Proc. Natl. Acad. 55 McDonald, R. L. and Olsen, R. W. (1994) Annu. Rev. Neurosci. 17, 569-602 Sci. U.S.A. 91, 4569-4573 56 Duggan, M. J. and Stephenson, F. A. (1990) J. Biol. Chem. 265, 3831-3835 98 Thompson, C. L. and Stephenson, F. A. (1994) J. Neurochem. 62, 2037-2044 57 Duggan, M. J., Pollard, S. and Stephenson, F. A. (1991) J. Biol. Chem. 266, 99 Motejlek, K., Hauselmann, R., Leitgeb, S. and Luscher, B. (1994) J. Biol. Chem. 24778-24784 269, 15265-15273 58 McKernan, R. M., Quirk, K., Prince, R., Cox, P. A., Gillard, N. P., Ragan, I. C. and 100 Pollard, S., Duggan, M. J. and Stephenson, F. A. (1991) FEBS Lett. 295, 81-83 Whiting, P. J. (1991) Neuron 7, 667-676 101 Luddens, H., Killisch, I. and Seeburg, P. H. (1991) J. Receptor Res. 11, 535-551