Diversity of Structure and Function of α1α6β3δ GABAA Receptors

δ subunit-containing γ-aminobutyric acid, type A (GABAA)receptors are expressed extrasynaptically and mediate tonic inhibition. In cerebellar granule cells, they often form receptors together with α1 and/or α6 subunits. We were interested in determining the architecture of receptors containing both subunits. We predefined the subunit arrangement of several different GABAA receptor pentamers by concatenation. These receptors composed of α1, α6, β3, and δ subunits were expressed in Xenopus oocytes. Currents elicited in response to GABA were determined in the presence and absence of 3α,21-dihydroxy-5α-pregnan-20-one (THDOC) or ethanol, or currents were elicited by 4,5,6,7-tetrahydroisoxazolo[5,4-c]-pyridin-3-ol (THIP). Several subunit configurations formed active channels. We therefore conclude that δ can assume multiple positions in a receptor pentamer made up of α1, α6, β3, and δ subunits. The different receptors differ in their functional properties. Functional expression of one receptor type was only evident in the combined presence of the neurosteroid THDOC with the channel agonist GABA. Most, but not all, receptors active with GABA/THDOC responded to THIP. None of the receptors was modulated by ethanol concentrations up to 30 mm. Several observations point to a preferred position of δ subunits between two α subunits in α1α6β3δ receptors. This property is shared by α1β3δ and α6β3δ receptors, but there are differences in the additionally expressed isoforms.

␦ subunit-containing ␥-aminobutyric acid, type A (GABA A )receptors are expressed extrasynaptically and mediate tonic inhibition. In cerebellar granule cells, they often form receptors together with ␣ 1 and/or ␣ 6 subunits. We were interested in determining the architecture of receptors containing both subunits. We predefined the subunit arrangement of several different GABA A receptor pentamers by concatenation. These receptors composed of ␣ 1 , ␣ 6 , ␤ 3 , and ␦ subunits were expressed in Xenopus oocytes. Currents elicited in response to GABA were determined in the presence and absence of 3␣,21-dihydroxy-5␣-pregnan-20-one (THDOC) or ethanol, or currents were elicited by 4,5,6,7-tetrahydroisoxazolo[5,4c]-pyridin-3-ol (THIP). Several subunit configurations formed active channels. We therefore conclude that ␦ can assume multiple positions in a receptor pentamer made up of ␣ 1 , ␣ 6 , ␤ 3 , and ␦ subunits. The different receptors differ in their functional properties. Functional expression of one receptor type was only evident in the combined presence of the neurosteroid THDOC with the channel agonist GABA. Most, but not all, receptors active with GABA/THDOC responded to THIP. None of the receptors was modulated by ethanol concentrations up to 30 mM. Several observations point to a preferred position of ␦ subunits between two ␣ subunits in ␣ 1 ␣ 6 ␤ 3 ␦ receptors. This property is shared by ␣ 1 ␤ 3 ␦ and ␣ 6 ␤ 3 ␦ receptors, but there are differences in the additionally expressed isoforms.
Synaptic receptors mediate phasic inhibition, whereas extrasynaptic receptors mediate tonic inhibition (8 -10). The ␦ subunit is part of the extrasynaptically located GABA A receptors, which have been shown to be operational in many regions of the brain (for review, see Ref. 10), among them cerebellar granule cells (11). The ␦ subunit is co-assembled either with ␣ 4 or ␣ 6 subunits (12,13) or with ␣ 1 subunits, at least in hippocampal interneurons (14).
Several studies suggest that ␦ and ␥ 2 subunits do not coexist in the same receptor (15)(16)(17). Therefore, ␦ has generally been considered as a substitute of the ␥ 2 subunit. ␦ subunit-containing receptors and their sensitivity to neurosteroids have been functionally characterized (18 -20). Interestingly, ␦ subunitcontaining GABA A receptors have been implicated in altered seizure susceptibility and altered states of anxiety during ovarian cycle (21) and in postpartum depression (22). ␣ 6 subunits are exclusively expressed in cerebellar granule cells (6). By immunogold staining, this subunit has been shown to be concentrated at Golgi synapses and at mossy fiber synapses and at a lower density in the extrasynaptic membrane (23). The ␦ subunit has been found exclusively in extrasynaptic locations, in the soma and on dendritic membranes (24). Thus, ␣ 6 and ␦ subunits may co-localize in extrasynaptic membranes. ␣ 6 knock-out mice display a strongly reduced expression of ␦ (12), suggesting an association of the two subunits. Immunoprecipitation experiments provided direct evidence for this (25). In whole cerebellum, GABA A receptor subtypes have been quantified using sequential immunoaffinity adsorption (26). ␣ 6 and ␦ subunits may be associated with the ␣ 1 subunit. The receptors composed of ␣ 1 , ␣ 6 , ␤ x (x ϭ 2,3), and ␦ subunits have been estimated to constitute 6% of all receptors in rat cerebellum and 10% in mouse cerebellum, respectively. As the ␣ 6 subunit is exclusively expressed in granule cells, the percentage of ␣ 1 ␣ 6 ␤ x ␦ receptors is substantially higher in these cells. Evidence for the simultaneous presence of ␣ 1 and ␣ 6 in the same receptor pentamer in recombinant systems has also been provided by functional experiments at least for ␥ 2 -containing receptors (27). Their relative position in the pentamer has been shown to strongly affect the pharmacological properties in the case of the ␣ 1 ␣ 6 ␤ 2 ␥ 2 receptor (7).
To enable an approach to establish the number and molecular location of pharmacologically important drug binding sites of ␣ 1 ␣ 6 ␤ 3 ␦ GABA A receptors, it is important to understand their architecture. It is unfortunately not possible to determine membrane protein architecture directly in neurons. Thus, model systems have to be used. We combined subunit concat-enation with functional expression in Xenopus oocytes. Thus, we covalently linked ␣ 1 , ␣ 6 , ␤ 3 , and ␦ subunits to force a defined arrangement of different subunits in a pentamer (28) and characterized the concatenated receptor in detail. From our work with ␣ 1 ␤ 3 ␦ (29) and ␣ 6 ␤ 3 ␦ GABA A receptors (30), we anticipated that positional roles of subunits were not as well defined as in ␣ 1 ␤ 2 ␥ 2 receptors (31,32). We indeed provide evidence for the facts that the ␦ subunit can assume different positions in the receptor pentamer and that neurosteroids strongly enhance currents elicited by GABA in most receptor forms and are co-agonists at one of the receptors. Ethanol, at concentrations below 30 mM, fails to modulate all functional receptors.
Expression in Xenopus Oocytes-Capped cRNAs were synthesized (Ambion, Austin, TX) from the linearized vectors containing different non-concatenated and concatenated subunits. A poly(A) tail of about 400 residues was added to each transcript using yeast poly(A) polymerase (USB Corp., Cleveland, OH) to stabilize them. The concentration of the cRNA was quantified on a formaldehyde agarose gel using Radiant Red stain (Bio-Rad) for visualization of the RNA with known concentrations of RNA ladder (Invitrogen) as standard on the same gel. The cRNAs were dissolved in water and stored at Ϫ80°C. Isolation of oocytes from the frogs, culturing of the oocytes, injection of cRNA, and defolliculation were done as described earlier (35). cRNA coding for each dual and triple subunit concatemer was injected either alone or in different combinations in oocytes, resulting in a total of nine different concatenated receptors. Oocytes were injected with 50 nl of RNA solution containing each construct at 50 nM. Keeping the amount of injected RNA constant is crucial for the concatenation approach (36). The combination of ␣ 1 , ␣ 6 , ␤ 3 , and ␦ subunits was expressed at a ratio of 10:10:10:50 nM. The injected oocytes were incubated in modified Barth's solution (35) at 18°C for about 72 h for the determination of I max and for at least 24 h before the measurements for the detailed characterization of the functional receptors.
Two-electrode Voltage Clamp Measurements-All measurements were done in medium containing 90 mM NaCl, 1 mM MgCl 2 , 1 mM KCl, 1 mM CaCl 2 , and 5 mM HEPES, pH 7.4, at a holding potential of Ϫ80 mV. For the determination of maximal current amplitudes, 1 mM GABA (Fluka) was applied in the absence and presence of 1 M THDOC (Sigma) for 20 s. THDOC was prepared as a 10 mM stock solution in dimethyl sulfoxide (DMSO) and was dissolved in external solution resulting in a maximal final DMSO concentration of 0.5%. The perfusion solution (6 ml/min) was applied through a glass capillary with an inner diameter of 1.35 mm, the mouth of which was placed about 0.4 mm from the surface of the oocyte (5). Non-concatenated and concatenated receptors containing the ␦ subunit showed a pronounced decrease in response to GABA with time. This decrease amounted to about 50% and did not recover. The experiments were performed after the measured currents became constant. Concentration-response curves for GABA were fitted with the equation where c is the concentration of GABA, EC 50 is the concentration of GABA eliciting half-maximal current amplitude, I max is the maximal current amplitude, I is the current amplitude, and n is the Hill coefficient. Two component curves were fitted with the equation where I a and I b are the maximal current amplitudes of component a and b, and EC 50a and EC 50b are the concentrations of GABA eliciting half-maximal current amplitudes in component a and b, respectively. The individual were curves were standardized to the sum of I a and I b and subsequently averaged.
The current responses to 1 M THDOC, 1 M THDOC ϩ 1 mM GABA, and 100 M THIP were determined on independent oocytes. The current amplitudes were averaged from oocytes from at least two batches. It should be noted that there is an up to 4-fold variation between two batches of oocytes in the mean current amplitude expressed from ␦ subunit-containing receptors.
Data are given as mean Ϯ S.E. for the I max values for GABA with and without THDOC and as mean Ϯ S.D. for analysis of properties of receptors using GABA. The perfusion system was cleaned between two experiments by washing with 100% DMSO after application of THDOC experiments to avoid contamination.
We characterized the response of concatenated receptors to THDOC, THIP, and ethanol, drugs thought to affect preferentially ␦ subunit-containing receptors. In addition, these receptors were compared with non-concatenated receptors.
Properties of Non-concatenated Receptors-Non-concatenated ␣ 1 /␤ 3 , ␣ 1 /␤ 3 /␦, ␣ 6 /␤ 3 , and ␣ 6 /␤ 3 /␦ receptors were expressed in Xenopus oocytes and characterized by electrophysiological techniques. Fig. 2 shows the concentration-response curves for GABA obtained from these receptors. They were characterized by an EC 50 and a Hill coefficient of 4.6 Ϯ 1. The non-concatenated receptor ␣ 1 /␣ 6 /␤ 3 /␦ was also studied (Fig. 3). This combination resulted in the expression of a large current amplitude, as did some concatenated receptors containing both ␣ 1 and ␣ 6 subunits. This subunit combination might additionally result in ␣ 1 /␤ 3 , Fig. 3) were expressed in Xenopus oocytes. We estimated receptor expression with 1 mM GABA in the absence and presence of the neurosteroid THDOC. Both concatenated receptors resulted in the absence of THDOC in currents Ͼ100 nA. THDOC enhanced the current in both receptors about 8-fold. None of the functional receptors was directly activated by 1 M THDOC alone (Fig. 3).
To characterize these receptors further, concentration-response experiments with GABA as the agonist were performed. Fig. 4A shows current traces from ␤ 3 -␣ 6 -␦/␣ 1 -␤ 3 (R13). Averaged GABA concentration-response curves for the concatenated ␤ 3 -␣ 1 -␦/␣ 6 -␤ 3 (R12) and ␤ 3 -␣ 6 -␦/␣ 1 -␤ 3 (R13) are illustrated in Fig. 4B. In the case of R13, the curve was well fitted with an EC 50 of 11.9 Ϯ 1.6 M and a Hill coefficient of 1.0 Ϯ 0.1, whereas in the case of R12, a better fit was obtained when two components with Hill coefficients of 1.0 each were assumed. The respective EC 50 values were 0.36 Ϯ 0.06 and 65 Ϯ 24 M, the first component accounting for about 72% and the second for about 28%. These values should be compared with that of non-concatenated ␣ 1 ␣ 6 ␤ 3 ␦ receptor of 2.0 Ϯ 0.9 M (Fig. 3) despite the fact that it is not clear in the latter case if all subunits are incorporated in all receptors formed (see above).
An averaged GABA concentration-response curve for the concatenated ␤ 3 -␣ 1 -␦/␤ 3 -␣ 6 (R52) is illustrated in Fig. 4B. As the current elicited by GABA alone was very small, the curves were obtained by simultaneous application of varying concentrations of GABA supplied with 1 M THDOC. The EC 50 for GABA amounted to 51 Ϯ 15 M (Fig. 3).
Comparison of ␣ 1 ␤ 3 ␦, ␣ 6 ␤ 3 ␦, and ␣ 1 ␣ 6 ␤ 3 ␦ Receptors- Fig. 6 compares ␣ 1 ␤ 3 ␦, ␣ 6 ␤ 3 ␦, and ␣ 1 ␣ 6 ␤ 3 ␦ receptors in terms of functional expression in Xenopus oocytes and EC 50 for GABA. To denominate subunit positions, we allude to the major adult isoform that has been shown to be arranged ␥ 2 ␤ 2 ␣ 1 ␤ 2 ␣ 1 . Irrespective of the type of ␣ subunit in the two ␣ positions, ␦ can assume the position of the ␤ subunit between the two ␣ subunits (R11-R14). The corresponding EC 50 values for GABA range between 0.1 M for two ␣ 6 and near 10 M for two ␣ 1 reasonably close to non-concatenated ␣ 1 ␤ 3 ␦ and ␣ 6 ␤ 3 ␦ receptors, respectively. Irrespective of the type of ␣ subunit in the two ␣ positions, ␦ can assume the position of the ␤ subunit between ␥ and ␣ subunits (R21-R24), provided the two ␣ subunits are identical, producing receptors with the expected EC 50 . In the case of mixed ␣ subunits, if ␣ 1 neighbors the ␦ subunit, there is little channel activity observed; if ␣ 6 neighbors the ␦ subunit, channels with an unusually high EC 50 for GABA of about 150 M form. We next discuss channels in which the ␦ subunit replaces one of the ␣ subunits. Interestingly, if ␦ occupies the position between the two ␤ subunits (R31, R34) in the major GABA A receptor isoform, ␣ 6 is tolerated in the other ␣ position, but ␣ 1 is not. The functional receptor with an ␣ 6 subunit has an EC 50 that compares well with the corresponding non-concatenated receptor. Conversely, if ␦ occupies the ␣ position normally neighboring the ␥ subunit (R41, R44), ␣ 1 is tolerated in the other ␣ position, but ␣ 6 is not, but for reasons discussed in Kaur et al. (29), the receptor with ␣ 1 may not be formed. Finally, we consider the case when ␦ replaces the ␥ subunit (R51-R54). A functional receptor only results if ␦ is neighbored by an ␣ 1 subunit, irrespective of the second ␣ subunit. In both cases, the corresponding EC 50 is unexpectedly high. In addition we studied ␣ 1 -␤ 3 -␣ 1 /␦, ␣ 1 -␤ 3 -␣ 6 /␦, ␣ 6 -␤ 3 -␣ 1 /␦, and ␣ 6 -␤ 3 -␣ 6 /␦ receptors (Figs. 3 and 6). Only ␣ 6 -␤ 3 -␣ 6 /␦ was able to form a functional receptor (30).

DISCUSSION
We investigated the architectural role of the ␦ subunit in ␣ 1 ␣ 6 ␤ 3 ␦ GABA A receptor pentamers. Functional expression of non-concatenated ␣ 1 ␣ 6 ␤ 3 ␦ might theoretically result in a large variety of subunit arrangements. In such a case, subunit concatenation (28, 31-34) may be used to control arrangement. To construct all possible subunit arrangements clearly exceeded our work capacity. We investigated all variants of the major GABA A receptor isoform ␥␤␣␤␣, where the ␥ subunit, or one of the ␤ subunits, was replaced by the ␦ subunit. If a ␤ subunit was substituted, the ␥ subunit position was occupied by a ␤ subunit, as is the case in ␣␤ receptors.
Validity of the Approach-In the case of ␥ 2 subunit-containing receptors, there seems to be a single pentameric subunit arrangement (31,32). This pentameric arrangement may be produced with the help of a number of different sets of triple and dual subunit constructs. Concatenated receptors are characterized by a 2-3.5-fold higher EC 50 for GABA and a current amplitude that amounts to 25-100% of that expressed by nonconcatenated subunits (32). We assume, but cannot prove, that similar rules apply to ␦ subunit-containing receptors. As mentioned below, we cannot completely exclude that a specific linker promotes or inhibits the corresponding subunit assembly, thereby affecting expression levels.
Concatenated subunits, in some cases, produce, for poorly understood reasons, current by themselves. The presence of ␤ 3 in these constructs results in larger expression in comparison with ␤ 2 -containing constructs (36). The amplitude of these currents strongly depends on the amount of RNA coding for this subunit that is injected into an oocyte (36). In the case of the ␥ 2 subunit, we never prominently described these currents as we used relatively small amounts of RNA and therefore only observed small currents. For the expression of ␦ subunit-containing receptors, we used 5-fold higher concentrations of RNA and replaced the ␤ 2 subunit by ␤ 3 . In these cases, we observed for some constructs, but not for others, expression of currents (Refs. 30 and 39 and this study). When such a conspicuous subunit was used for the expression of a pentameric receptor, we required that the functional properties of this receptor pentamer had to differ from those of the individual construct. Alternatively, when such functional differences were not evident, we constructed the concatenated pentamer.
it cannot fully be excluded that at least part of this current is caused by ␤ 3 -␣ 1 -␦.
Summary-In conclusion, we have shown that GABA A receptors containing the ␣ 1 , ␣ 6 , ␤ 3 , and ␦ subunits exhibit the ability to promiscuously assemble into different arrangements, at least after expression in Xenopus oocytes. Further we show that one of the ␦ subunit-containing receptors remains relatively silent in the absence of neurosteroid. The architecture of ␣ 1 ␣ 6 ␤ 3 ␦ GABA A receptors in situ in cerebellar granule cells needs to be established. Furthermore, the mechanisms promoting the assembly of a defined receptor will have to be understood. It is intriguing to hypothesize that a neuron directs assembly of different receptor isoforms from the same set of subunits, depending on the actual functional needs. This would add to the plasticity of the neuronal system.