Four Amino Acids in the α Subunits Determine the γ-Aminobutyric Acid Sensitivities of GABAA Receptor Subtypes*

GABAA receptors, mediators of fast inhibitory neurotransmission, are heteropentameric assemblies from a large array of subunits. Differences in the sensitivity of receptor subtypes to endogenous GABA may permit subunit-dependent finely tuned responsiveness to the same GABAergic inputs. Using both radioligand binding and electrophysiology combined with mutagenesis, we identified a domain of four amino acids within the α subunits that mediates the distinct sensitivities to GABA allowing their selective switch between αβ3γ2 combinations. Replacing this domain in α3 by the corresponding segments of α1–α5 resulted in mutant receptors displaying the GABA EC50 values of the respective wild-type receptors. Vice versa, the α3 motif forced the low sensitivity to GABA of α3 upon α1β3γ2, α4β3γ2, and α5β3γ2. Binding of the GABA agonist [3H]muscimol was not affected by the exchange of the motif between α1 and α3 subunits. Thus, the equilibrium binding pocket is maintained upon replacement of the four amino acids. Taken together our data suggest that the identified motifs contribute to a structure involved in the transduction of the binding signal rather than to the binding itself.

GABA A receptors, mediators of fast inhibitory neurotransmission, are heteropentameric assemblies from a large array of subunits. Differences in the sensitivity of receptor subtypes to endogenous GABA may permit subunit-dependent finely tuned responsiveness to the same GABAergic inputs. Using both radioligand binding and electrophysiology combined with mutagenesis, we identified a domain of four amino acids within the ␣ subunits that mediates the distinct sensitivities to GABA allowing their selective switch between ␣␤3␥2 combinations. Replacing this domain in ␣3 by the corresponding segments of ␣1-␣5 resulted in mutant receptors displaying the GABA EC 50 values of the respective wild-type receptors. Vice versa, the ␣3 motif forced the low sensitivity to GABA of ␣3 upon ␣1␤3␥2, ␣4␤3␥2, and ␣5␤3␥2. Binding of the GABA agonist [ 3 H]muscimol was not affected by the exchange of the motif between ␣1 and ␣3 subunits. Thus, the equilibrium binding pocket is maintained upon replacement of the four amino acids. Taken together our data suggest that the identified motifs contribute to a structure involved in the transduction of the binding signal rather than to the binding itself.
Fast inhibitory neurotransmission in the mammalian central nervous system is mediated mainly by the GABA A 1 receptor, a ligand-gated chloride channel. The receptor complex presumably is composed of five protein subunits, each consisting of an extracellular N-terminal domain with a putative cysteine loop, four largely conserved transmembrane segments (TM), and a variable intracellular region between TM3 and TM4. This topology is characteristic for members of the superfamily of ligand-gated ion channel receptors (1,2). Several GABA A receptor subunits (␣1-6, ␤1-3, ␥1-3, ␦, ⑀, , , 1-3) have been cloned from mammalian brain (3,4). Thus, the genetic diversity of multiple GABA A receptor subunits permits the assembly of a vast number of receptor heteromeric isoforms. Apparently, the subunit composition determines the pharmacological profile of the resulting receptor subtype (5), but the physiological significance of the GABA A receptor heterogeneity in the central nervous system remains largely unknown. Most GABA A receptor subunits exhibit distinct regional and cellular distribution throughout the brain, with their expression patterns changing during pre-and postnatal development (6). In addition, in neurons expressing different receptor isoforms, a subunit-selective targeting to cellular domains has been observed (7)(8)(9). These findings further strengthen the notion that the receptor subtypes subserve individual functions in GABAergic neurotransmission. Hence, they may contribute to synaptic variety, as depending on their subunit composition recombinant GABA A receptors differ in their sensitivity to the endogenous agonist GABA (10 -12). Especially in view of the existence of extrasynaptic GABA A receptors (7,13), a 20-fold difference in GABA sensitivity may play a decisive role in neuronal transmission irrespective of the assumed saturating GABA concentration in the synaptic cleft. Therefore, various GABA A receptor subtypes may allow finely tuned responsiveness to the same GABAergic inputs (14).
In recombinant ternary receptor complexes the sensitivity to GABA is most apparently influenced by the ␣ subunit, with ␣3-containing isoforms consistently reported to display the lowest sensitivities of all GABA A receptor subtypes (10, 14 -17). In addition, the GABA-dependent [ 35 S]TBPS binding of ␣3␤3␥2/3 assemblies deviates remarkably from all other subunit combination, that is, high concentrations of GABA are needed to significantly reduce binding of the cage convulsant [ 35 S]TBPS as compared with its respective binding maximum (18). As the deflection point of the concentration-response curves of [ 35 S]TBPS binding for a given subunit combination correlates well with its EC 50 value for GABA (18), these data confirm the low sensitivity of ␣3-containing receptors and recommend ␣3 as the starting point to identify molecular determinants of GABA potency in the ␣ subunits.
Here we describe the identification and functional characterization of four-amino acid motifs specific for individual ␣ subunits of the GABA A receptor, which control the sensitivity to GABA of the resulting receptor. Exchange between subunits of the whole motif but not of single amino acids resulted in transfer of the respective GABA sensitivities for most receptors of the general composition ␣i␤3␥2, where i ranges from 1 to 6. Thus, our findings may provide insight into the mechanistic features responsible for the molecular diversity of GABAergic neurotransmission and deliver tools to further dissect the biological significance of GABA A receptor heterogeneity.

EXPERIMENTAL PROCEDURES
Construction of Chimeras-The coding strands of the rat GABA A receptor ␣1 and ␣3 subunits were excised from expression vectors pRK5-r␣1 and pRK5-r␣3, respectively, and subcloned into pBluescript SK(Ϫ) (Stratagene, Amsterdam, Netherlands). Two conserved NcoI re-striction sites (Fig. 1) were employed to exchange a 414-bp fragment between ␣1 and ␣3, spanning the region from the putative cysteine loop to the TM3. We named the chimeras ␣1(TM12␣3) and ␣3(TM12␣1), respectively, with the characters in parentheses indicating the foreign origin of TM1 and TM2. To generate the chimera ␣1(TM14)␣3 we made use of the XbaI site in the multiple cloning site of pBluescript and the first NcoI site of the coding strand to excise a 590-bp fragment from ␣1. This fragment was cloned into an ␣3 construct cut with XbaI and partially digested with NcoI. The final chimera contains the N-terminal part of ␣1 followed by an ␣3 stretch extending from the cysteine loop to the C terminus. A reverse chimera, designated ␣3(TM14)␣1, was constructed likewise.
In a first series of PCR we used pRK5-r␣1 as template with primer pairs cis55/a13up and a31down/a1Nco to generate fragments of 416 and 216 bp in length. Likewise, fragments of 635 and 216 bp were amplified from pRK5-r␣3 with primer pairs cis55/a31up and a13down/a3Nco. In a second PCR setting, we mixed the 416-bp ␣1 fragment with the 216-bp ␣3 product and combined the 635-bp ␣3 fragment with the 216-bp amplificat of ␣1.
Because of the partial complementarity of primers a31down/a31up and a13down/a31up, the fragments hybridize at their respective 3Јends and thus allow extension of the double strand by Vent polymerase.
The resulting hybrid PCR products were double digested with XbaI and NcoI and cloned into pBluescript-r␣3, which had been cut with XbaI and partially digested with NcoI. The chimeras were named r␣13-␣3 and r␣31-␣3, respectively. The sequences of all chimeras ( Fig.  1) were verified by DNA sequencing and subcloned into expression vector pRK5.
Mutagenesis-Eukaryotic expression vectors (pRK5 or pRK7) for the rat GABA A receptor ␣1-6 subunits (19) were used in this study. The desired point mutations and multiple amino acid changes were introduced into the coding strands using the QuikChange site-directed mutagenesis kit (Stratagene). As applicable we interchanged nucleotides to match human codon preferences for enhanced expression rates in HEK 293 cells. Successful mutagenesis was verified by restriction analysis and DNA sequencing.
Electrophysiology-Two days after transfection single coverslips containing HEK 293 cells were placed in a recording chamber mounted on the movable stage of a fluorescence microscope (Olympus IX70) and perfused at room temperature with a defined saline solution containing (in mM): 130 NaCl, 5.4 KCl, 2 CaCl 2 , 2 MgSO 4 , 10 glucose, 5 sucrose, and 10 HEPES (free acid), pH adjusted to 7.35, with about 35 mM NaOH. Transfected cells were identified by their fluorescence (enhanced green fluorescent protein), and ligand-mediated membrane currents of these cells were studied in the whole-cell configuration of the patch clamp technique (21). Patch clamp pipettes were pulled from hard borosilicate capillary glass (0.5 mm ID, 1.5 mm OD, Vitrex, Science Products GmbH, Hofheim, Germany) using a horizontal puller (model P-97, Sutter Instruments, CA) in a multi-stage process. The pipettes had an initial resistance of 2-4 megohms when filled with a solution containing (in mM): 90 KCl, 50 KOH, 2 CaCl 2 , 2 MgCl 2 , 10 EGTA, 3.1 ATP (di-potassium salt), 0.4 GTP (tri-sodium salt), and 10 HEPES (free acid), pH 7.35.
The junction potential between the pipette and the external solution was less than 2.3 mV and therefore was disregarded. Seal resistances of Ͼ1 gigohm were routinely obtained by applying gentle suction to the pipettes. Membrane rupture was monitored electrically as an increase in capacity. Pipette capacitance, membrane capacitance, and series resistance were compensated for electronically to achieve minimal capacitive transients. A series resistance compensation of Ͼ60% was regularly used. Using a fast Y-tube application system, the recombinant receptors were tested at GABA concentrations of 0.01, 0.03, 0.1, 0.3, 1, 3, 5, 10, 30, 60, 100, 300, and 1000 M. The responses of the cells were recorded by a patch clamp amplifier (EPC-8, HEKA-Electronic, Lambrecht, Germany) in conjunction with a standard personal computer and the pClamp 8.1 software package (Axon Instruments, Foster City, CA). The standard holding potential for the cells was Ϫ40 mV. Whole cell currents were low pass-filtered by an eight-pole Bessel filter at 5 or 3 kHz before being digitized by a Digidata 1322A interface (Axon Instruments) and recorded by the computer at a sampling rate of at least 1 kHz.
Binding Assays-Resuspended cell membranes (50 -200 g of protein/tube) were incubated in a final volume of 0. 5  S]TBPS, respectively), the assay mixtures were rapidly diluted to 5 ml with ice-cold 10 mM Tris/HCl, pH 7.4, filtered through glass fiber filters, and washed once with 5 ml of 10 mM Tris/HCl, pH 7.4. Filters were immersed in 4 ml of Zinsser Aqua-Solv scintillation fluid, and radioactivity was determined in a Beckman liquid scintillation counter using external standardization. Statistical calculations were performed using the GraphPad Prism program (GraphPad Software, San Diego, CA).

GABA Potency of Wild-type ␣i␤3␥2
Receptors-We transiently expressed ternary rat GABA A receptors in HEK 293 cells configured from any of the six ␣ subunits with the ␤3 and ␥2S variants. Transfected cells were analyzed by whole-cell patch clamping. When expressed in HEK 293 cells, the EC 50 of GABA-induced chloride current properties varied with the employed ␣ subunit in a range of Ͻ1 to Ͼ50 M in the rank order (Table I). Thus we have confirmed the exceptionally low sensitivity to GABA conferred by the ␣3 subunit.

Properties of ␣1-␣3 Chimeric Subunits in [ 35 S]TBPS Bind-
ing-To initialize the search for the molecular determinants of GABA potency in the ␣ subunits, we chose the ubiquitous ␣1 as a reference to the ␣3 subunit and constructed chimeras containing variable segments of each subunit (Fig. 1). We analyzed the GABA-dependent [ 35 S]TBPS binding of these chimeric subunits and compared them with recombinant receptors containing either wild-type ␣1 or ␣3 in addition to ␤3␥2S. These findings warranted further dissection of this segment to pin down the decisive amino acid residues. The sequence identity of ␣1 and ␣3 in this stretch of 62 amino acids is 87%, and thus a maximum of 8 amino acid residues could cause the functional characteristics. Based on the assumption that the particular low sensitivity conferred by the ␣3 subunit is due to a unique structural feature, we aligned the 62-amino acid stretch of all ␣ subunits (Fig. 3). The ␣3 subunit differs in only two positions from all other ␣ subunits, i.e. Val 174 and His 188 (relating to the unequivocal numbering of the unprocessed protein), recommending both as prime candidates for subsequent mutagenesis studies.
Single Amino Acids Are Involved but Are Insufficient to Determine the Sensitivity to GABA-Using site-directed mutagenesis we substituted His 188 in ␣3 for the corresponding arginine of ␣1. Expression of the mutant ␣3H188R subunit together with ␤3␥2S did not alter the sensitivity of the resulting receptors to GABA (data not shown), i.e. the EC 50 value and the Hill coefficient were identical to those of wild-type ␣3␤3␥2S. Likewise, the dose response to GABA of receptors containing the reverse mutant ␣1R162H was indistinguishable from that of the ␣1␤3␥2S assemblies.
To examine the impact of ␣3Val174 on the sensitivity to GABA, we mutated this residue into its ␣1 corresponding amino acid, threonine. Fig. 4 illustrates the effect on the GABA sensitivity of the resulting ␣3V174T␤3␥2S receptors with the dose-response curve of ␣3V174T␤3␥2S being intermediate between the wild-type ␣1and ␣3-containing receptors but with an increase in ligand-induced current at the highest concentration of GABA employed (1 mM). Therefore, residue Val 174 in ␣3 can be considered a necessary determinant of the GABA sensitivity although insufficient to convert the ␣3 subunit into a subunit with the characteristics of ␣1, thus predicting additional factors.
We therefore mutated the neighboring amino acids Leu 173 , Asp 175 , and Asn 176 into their ␣1 counterparts isoleucine, glutamic acid, and aspartic acid, respectively. When expressed together with ␤3 and ␥2S, ␣3D175E was indistinguishable from wild-type ␣3, whereas the EC 50 values for GABA of ␣3L173I and ␣3N176D were slightly decreased to 15 Ϯ 1 and 18 Ϯ 2 M, respectively (Fig. 4). However, the EC 50 values of ␣1-containing receptors were not achieved by any of these mutant subunits.
Individual Four-amino Acid Motifs Determine the GABA Sensitivity of the ␣ Subunits-From the sequence alignment of the ␣ subunits and the above cited results it became apparent that ␣3 Leu 173 , Val 174 , and Asn 176 are part of a four-amino acid (4aa) motif highly variable between the six subunits. The variability appeared even more significant because of the conserved residues flanking the motif N-and C-terminally. We therefore proceeded to exchange the complete motif between ␣1 and ␣3.
Replacement of the 4aa motif between the ␣1 for the ␣3 subunits had a tremendous effect on the sensitivity to GABA. When substituting residues LVDN in ␣3 by ␣1ITED, the doseresponse curve of the mutant receptor subtype ␣3ITED␤3␥2S shifted to the left (Fig. 5A), resulting in the same EC 50 to GABA of wild-type ␣1and mutant ␣3-containing receptors (2.9 Ϯ 0.1 and 3.2 Ϯ 0.1 M for ␣1 and ␣3ITED, respectively). The opposite effect was obtained with the converse mutant ␣1LVDN, as both the dose-response curve (Fig. 5A) and the EC 50 (48 Ϯ 2 and 55 Ϯ 2 M for wild-type and mutant receptors, respectively) to GABA were identical to those of the wild-type ␣3␤3␥2S receptors. Thus, exchange of the 4aa motif resulted in a complete transfer of the GABA sensitivities from ␣1 upon ␣3 and vice versa (Table I, Fig. 5A).
When expressing ␤3␥2S together with ␣3IQDD, the doseresponse curve of the resulting receptors did not differ from that of ␣2␤3␥2S (Fig. 6A, Table I). Introducing the LVDN motif of ␣3 into the ␣2 sequence had a remarkable effect on the GABA potency, as the sensitivity for GABA of ␣2LVDN␤3␥2S assemblies was reduced more then 1000-fold. As the doseresponse curve of the mutant receptor complex did not even plateau at the highest GABA concentration tested, the EC 50 could only be extrapolated to 5 mM (Fig. 6A).
The dose-response curve for GABA of the mutant subunit ␣6LVDN␤3␥2S shifted in the direction of the dose-response curve of wild-type ␣3␤3␥2S (Fig. 6B, Table I), resulting in an EC 50 value of ␣6LVDN containing receptors 2.5-fold higher than of wild-type ␣6 receptors. The reverse mutant ␣3LMQN␤3␥2S was the only receptor of all wild-type and mutant GABA A subunit assemblies tested that exhibited a biphasic response to GABA with EC 50 values of 1.7 Ϯ 0.4 M for the high affinity and 101 Ϯ 56 M for the low affinity site (Fig. 6B). Thus, both values were double the magnitude of the ␣6 and ␣3 wild type-containing receptors, respectively. The fraction of the high affinity site was determined to be 63 Ϯ 5%. Based on the F-score (F 1,12 ϭ 10.97), the two-site fit was statistically superior to that of the Hill equation (p ϭ 0.006). In our hands, ␤3 and ␤3␥2S alone did not form functional channels (data not shown). Thus, we can exclude any contribution of homomeric ␤3 receptors or ␤3␥2S receptors to the biphasic nature of the ␣3LMQN␤3␥2S dose-response curve.  The four-amino acid motifs that determine the sensitivity to GABA are boxed. Residues previously implicated in agonist binding are highlighted in gray boxes. The corresponding sequences of rat nicotinic acetylcholine ␣1 and glycine receptor ␣1 subunits are shown with residues matching the consensus sequence of the GABA A receptor ␣ subunits as points and the gaps as dashes. Note that GABA A receptor residue r␣6Gln 140 was originally sequenced as His.
FIG. 4. Single amino acid residues are insufficient in determining GABA sensitivities. Amino acid residues Leu 173 , Val 174 , Asp 175 , and Asn 176 in ␣3 were substituted by their corresponding ␣1 counterparts. These single mutations were insufficient to convert the ␣3 subunit into a subunit with the characteristics of ␣1 when coexpressed with ␤3␥2S. Error bars represent S.E. ment of 6 nM [ 3 H]muscimol by GABA from wild-type and selected mutant ␣1 and ␣3 subunits coexpressed with the ␤3 and ␥2S subunits (Fig. 7). The GABA A receptor agonist [ 3 H]muscimol selectively labels the high affinity binding site. The IC 50 values conferred by the mutant subunits ␣3V174T and ␣1T148V were indistinguishable from those of the underlying wild-type ␣ subunits. The same held true for mutant subunits ␣3ITED and ␣1LVDN, respectively. Thus, displacement of neither the single amino acid residue nor the 4aa motif influenced the equilibrium binding parameters of the resulting receptors, indicating that the depicted segment does not contribute to the high affinity agonist binding site. DISCUSSION The structural heterogeneity of GABA A receptors is widely accepted, but the physiological significance of the receptor variety in vivo remains enigmatic, especially as there are no known endogenous analogues for most psychoactive drugs modulating GABAergic inputs. The fact that many receptor subtypes can be distinguished by their respective sensitivity to GABA in vitro argues strongly that differential expression of at least the ␣ subunits influences the neuronal responsiveness to this neurotransmitter in vivo. Thus, an array of different subunit isoforms would allow an adjustment to varying GABAergic inputs as well as differing responses to uniform signals (14,22). The diversity of GABA A receptor subtypes may hence form an important basis for synaptic plasticity in the inhibitory circuitry.
As data concerning specific expression, regulation, composition, and localization of receptor subtypes have to be interpreted in the light of their respective functional properties, it is essential to characterize these properties and to establish the underlying molecular determinants. In the present study we identified a segment of four amino acids in the extracellular N-terminal region of ␣ subunits decisive for the GABA sensitivity of GABA A receptors in the composition ␣i␤3␥2S. By exchanging the 4aa motif between the ␣ subunits, we were able to transfer the GABA sensitivities of the "donor" wild-type subunits upon most mutants. The exchange was incomplete with substitution of single amino acids, thus pointing to an orchestrated action of more than one amino acid. Although we cannot prove the absolute necessity of all four amino acids in the motif for the characteristics of GABA sensitivity, we provide evidence here for the contribution of three of four nonsequential amino acids. Thus, the use of the term 4aa motif is warranted. The subunit-specific 4aa motifs are conserved between the human, rat, mouse, bovine, and chicken ␣ subunits, suggesting their functional importance in the inhibitory circuitry of the brain throughout evolution.
The ␣1, ␣3, ␣4, ␣5, and most likely ␣2 subunits share the exact sequence position of the molecular determinant for GABA potency, whereas the GABA potency of ␣6 was only incompletely transferred by the 4aa motif. The mutated ␣3IQDD subunit fully adopted the sensitivity of wild-type ␣2, thus matching the effects produced with the corresponding ␣1, ␣4, and ␣5 mutants. In contrast, the converse mutant ␣2LVDN led to more than a 1000-fold rightward shift of the GABA doseresponse curve. This points to the 4aa motif as being crucial for GABA sensitivity in ␣2, but its exchange may disrupt the receptor conformation imperative for a directed shift of the sensitivity to GABA. However, this hypothesis would require the additional assumption that the GABA interaction site of the ␣2 subunit is more labile than that of all other ␣ subunits.
Among the ␣ subunits, ␣6 confers the highest sensitivity to GABA A receptors (11,12,23,24). Although the EC 50 to GABA was altered in ␣6LVDN and ␣3LMQN, the characteristics of the respective wild-type subunits were either incompletely transferred (␣6LVDN␤3␥2S) or resulted in a unique biphasic dose-response curve (␣3LMQN␤3␥2S) with the higher and lower affinity closely resembling those imposed by the original sequences ␣6 and ␣3, respectively (Fig. 6B). This is even more surprising as the exchanged domain comprises just two altered amino acids, with only one of these being unique to all six 4aa motifs (Q). As we can exclude any functional contribution of homomeric ␤3 receptors or ␤3␥2S receptors to the biphasic nature of the ␣3LMQN␤3␥2S dose-response curve (data not shown), our results suggest an additional mechanism in the ␣6 subunit, besides the 4aa motif, that underlies its extraordinarily high sensitivity. Using chimeric subunits of the ␣1 and ␣6 subunit, we now aim to identify its molecular basis.
In heteropentameric GABA A receptor complexes the putative GABA binding site is presumed to be located at the ␣-␤ interface. A variety of approaches revealed several residues in the ␣ and ␤ subunits that line the binding pocket. From these experiments it was concluded that in the ␣ subunits most amino acids relevant to the recognition of GABA cluster in two segments of the extracellular N-terminal domain, i.e. Phe 93 -Ser 97 (25)(26)(27)(28) and Val 205 -Asp 210 (29). Additionally, a conserved arginine N-terminally adjacent to the 4aa motif (Fig. 3) has been proposed to be a key factor for both binding and action of GABA, as mutant ␣1 (30) or ␣5 subunits (31) in which the respective residue was substituted by lysine showed a dramatic increase in the K ⌬ of the [ 3 H]muscimol binding as well as the EC 50 to GABA. All amino acid residues up to now associated with the formation of the GABA binding site are highly conserved among the ␣ subunits, most likely because of the search criteria employed. Because of the wide variance in EC 50 values of GABA A receptors containing different ␣ subunit isoforms, the sites involved in the action of GABA on the ␣ subunits need to contain highly variable residues in addition to the known set of conserved amino acids. We tentatively suggest that the 4aa motif is part of this variable segment.
Exchange of the 4aa motif was not accompanied by a corresponding shift in equilibrium binding parameters of the GABA agonist [ 3 H]muscimol. As ligand-binding affinity at steady state most likely recognizes receptors in a desensitized and high affinity state, we propose that the 4aa motif does not contribute to the high affinity binding site. However, it may contribute to the low affinity binding site, which has been interpreted as a conformational variant of the high affinity site (32).
A previous study (33) showed that exchange of the isoleucine belonging to the ␣1ITED motif by valine decreases both the sensitivity to GABA and the affinity of agonist GABA A receptor ligands in mutant ␣1␤2␥2 receptors. The discrepancies between these data and our results may be caused by the nature of the introduced amino acid. The first position of the 4aa motif in all known ␣ subunits is occupied either by leucine or isoleucine. Whereas replacement of isoleucine by leucine may maintain the structural features of the 4aa motif, introduction of valine could induce conformational changes in both the motif and its flanking segments. Furthermore, we cannot exclude a decisive role of the different ␤ subunits employed in the two studies.
Information about the three-dimensional arrangement of GABA A receptors is based primarily on the structure derived from the soluble molluscan acetylcholine-binding protein (AChBP), a homologue of the N-terminal ligand-binding domain of the nicotinic acetylcholine receptor ␣ subunits (34). By definition, acetylcholine receptors share structural and functional characteristics with other members of the superfamily of ligand-gated ion channels. Therefore, it is permissible to model the extracellular region of the GABA A receptor according to the crystal structure of AChBP (35,36). All available models include the 4aa motif in a secondary structure named "loop E," claimed to be part of the GABA binding pocket at an ␣-␤"plusminus" interface (35). Depending on the alignment chosen, these four residues are predicted either to link the ␤5Ј and ␤6 strands (34) or to participate in ␤5Ј (35,36). As all predictions rely on accurate sequence alignments anchored around con- served residues, they become more speculative in regions of low or absent homology. The 4aa motif is present only in the GABA A receptors and possibly the glycine receptors (see Fig. 3) but is missing in other members of the ligand-gated ion channel superfamily. Thus, the alignment of the 4aa motif remains uncertain, and extrapolations from the structure of the AChBP to the function of GABA A receptors should be handled with caution (34 -36).
The activation of GABA A receptors involves the processes of binding to the sensitized receptor, signal transduction, and finally channel opening. Our study was not geared toward identifying the precise molecular role of the 4aa motif. Rather, we expect that the data presented here will help to further elucidate the binding process and signal transmission mechanism of GABA A receptors. Moreover, the results of our study may encourage a knock-in approach in which exchange of the 4aa motif between ␣ subunit genes should result in the expression of mutant receptors with altered sensitivity to GABA. Resulting phenotypic abnormalities and possible compensations could be attributed directly to unbalanced neuronal responses to GABAergic inputs and hence might settle the significance of the ␣ subunit diversity in the physiology and pathophysiology of the brain.