- Top of page
Delta (δ) subunit containing GABAA receptors are expressed extra-synaptically and mediate tonic inhibition. In cerebellar granule cells, they often form a receptor together with α6 subunits. We were interested to determine the architecture of these receptors. We predefined the subunit arrangement of 24 different GABAA receptor pentamers by subunit concatenation. These receptors (composed of α6, β3 and δ subunits) were expressed in Xenopus oocytes and their electrophysiological properties analyzed. Currents elicited in response to GABA were determined in presence and absence of 3α, 21-dihydroxy-5α-pregnan-20-one and to 4,5,6,7-tetrahydroisoxazolo[5,4-c]-pyridin-3-ol. α6-β3-α6/δ receptors showed a substantial response to GABA alone. Three receptors, β3-α6-δ/α6-β3, α6-β3-α6/β3-δ and β3-δ-β3/α6-β3, were only uncovered in the combined presence of the neurosteroid 3α, 21-dihydroxy-5α-pregnan-20-one with GABA. All four receptors were activated by 4,5,6,7-tetrahydroisoxazolo[5,4-c]-pyridin-3-ol. None of the functional receptors was modulated by physiological concentrations (up to 30 mM) of ethanol. GABA concentration response curves indicated that the δ subunit can contribute to the formation of an agonist site. We conclude from the investigated receptors that the δ subunit can assume multiple positions in a receptor pentamer composed of α6, β3 and δ subunits.
α6 subunits are exclusively expressed in cerebellar granule cells (Sieghart and Sperk 2002). 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 extra-synaptic membrane (Nusser et al. 1996). In contrast, the δ subunit has been found exclusively in extrasynaptic locations, in the soma and on dendritic membranes (Nusser et al. 1998). Thus, both α6 and δ have been found in extrasynaptic membranes. In whole cerebellum, GABAA receptor subtypes have been quantified using sequential immunaffinity adsorption (Pöltl et al. 2003). The receptor composed of α6, β2/3 and δ subunits has been estimated to constitute 11% and 18% of all receptors, in rat and mouse cerebellum, respectively. As the α6 subunit is exclusively expressed in granule cells, the percentage of α6βxδ receptors is substantially higher in these cells. α6 less mice display a substantially reduced expression of δ (Jones et al. 1997), suggesting a direct association of these two subunits. This was actually demonstrated with co-immunoprecipitation (Khan et al. 1996).
It is unfortunately not possible to determine membrane protein architecture in situ in the neurons. Thus, model systems have to be used. In the present study, we have focused on the architecture of α6β3δ GABAA receptors expressed in Xenopus oocytes at the functional level. To investigate active channels we used covalently linked α6, β3 and δ subunits to have a defined arrangement of different subunits in a pentamer (Minier and Sigel 2004a). The concatenated receptors were characterized in detail using the agonist GABA, the neurosteroid 3α, 21-Dihydroxy-5α-pregnan-20-one (THDOC) and THIP and their properties were compared with those of non-concatenated receptors. In the present study, we provide evidence for the facts that (a) the δ subunit can assume different positions in the α6β3δ receptor pentamer, similarly to the previously described α1β3δ GABAA receptors (Kaur et al. 2009), (b) that the presence of neurosteroids strongly enhances or even uncovers currents mediated by functional receptors, and (c) that ethanol fails to modulate these receptors.
- Top of page
Our aim was to determine the architecture of α6β3δ GABAA receptors. Functional expression of non-concatenated α6β3δ might result in the formation of receptors with multiple subunit arrangements. In order to predefine the subunit arrangement, we used subunit concatenation (Baumann et al. 2001, 2002, 2003; Minier and Sigel 2004a; Baur et al. 2006). To construct all possible subunit arrangements as pentamers clearly exceeded our work capacity. Therefore, we chose to investigate all variants of the major GABAA receptor isoform γβαβα, where one of the α, or one of the β subunits or the γ subunit was replaced by the δ subunit. If a α or a β subunit was substituted, the γ subunit position was occupied by a β subunit, as is the case in αβ receptors. In addition, we used the constructs to form 19 alternative receptors.
Four of the δ subunit containing receptors studied show functional expression
First we analyzed the receptor in which the δ subunit was placed in all five possible positions in a pentamer composed of the consensus arrangement of α and β subunits. All receptors responded with current amplitudes <50 nA to the exposure of 1 mM GABA. If 1 μM of the neurosteroid THDOC was added to GABA, β3-α6-δ/α6-β3 (R1), α6-β3-α6/β3-δ (R2) and β3-δ-β3/α6-β3 (R3) receptors produced currents >300 nA (Fig. 1b). 1 μM THDOC alone elicited currents <10 nA in these receptors. 100 μM THIP resulted in much larger currents than observed on application of GABA, but this current amounted to only 25–40% of that induced by GABA + THDOC.
Secondly, we combined the concatenated constructs either with other available concatenated constructs or loose subunits to form 19 additional receptors. Only one of them, α6-β3-α6/δ, resulted in the expression of currents amounting to >200 nA. Interestingly this receptor was also activated by GABA alone, to an extent amounting to about 12%.
The expression of the concatenated dual subunit construct α6-β3 alone resulted in about 180 nA current upon application of GABA + THDOC, but only in about 30 nA on application of THIP. It is not clear whether this construct is able to form tetramers or hexamers, or whether one of the subunits is hanging out, not being incorporated in the pentamer (Minier and Sigel 2004a). β3-α6-δ/α6-β3 (R1) and β3-δ-β3/α6-β3 (R3) receptors contain this dual subunit construct, but respond with 19- and fold-fold lager currents on application of THIP. In addition, the KD to GABA was significantly smaller (p < 0.025) in β3-α6-δ/α6-β3 (R1), and β3-δ-β3/α6-β3 (R3) receptors as compared with the dual subunit construct (Fig. 1a and b). Therefore, we conclude that these two receptors do not represent an artifact but are really formed. However, we cannot exclude that a fraction of current expressed from β3-δ-β3/α6-β3 (R3) receptors is due to α6-β3.
Thus, the δ subunit can occupy both positions occupied by β subunits, the position of the α subunit between the two β subunits or both positions corresponding to the γ and the adjacent β subunit in the major adult isoform receptors, corresponding to γβαβα. The fact that the δ subunit can assume different positions is in contrast to our observations with the γ subunit that exclusively can only adopt a single specified position (Baumann et al. 2001, 2002).
The δ subunit contributes to the formation of an agonist site
The GABA concentration response curve for α6-β3-α6/β3-δ (R2) receptors is clearly bi-phasic (Fig. 3c). Replacement of the β3 subunit adjacent to the δ subunit by a δ subunit to form α6-β3-α6/δ receptors converts this bi-phasic curve into a mono-phasic curve with a Hill coefficient <1. The concentration response curve for α6-β3-α6/β3-δ (R2) receptors was carried out in the presence of THDOC. It could be argued that THDOC is responsible for the low affinity component. Therefore, we carried out the experiment with α6-β3-α6/δ receptors in the presence and absence of THDOC. In both cases, the Hill coefficient was <1. The presence of THDOC shifted the curve about two-fold to the left without resulting in a low affinity phase. Thus, at least in this receptor, THDOC does not induce a second phase.
Previously, we have presented evidence for the presence at the β3/δ subunit interface in α1-β3-α1/β3-δ receptors of an GABA agonist site (Kaur et al. 2009). The present observation strongly supports the proposition that the δ subunit is able to assume the role of the α subunit in an agonist site that is normally formed at the β/α subunit interface. While this agonist site has an EC50 of about 8 μM in α1-β3-α1/β3-δ receptors, its EC50 in α6-β3-α6/β3-δ (R2) is about 400 μM. Please note that the EC50 does not reflect the affinity of GABA to the site.
Comparison to currents mediated by non-concatenated α6β3δ receptors
It has previously been observed that the properties of α6β3δ receptors depend on the expression conditions, i.e., on the amount and the ratio of mRNA coding for the different subunits that is injected into Xenopus oocytes (Hadley and Amin 2007). Under certain expression conditions, bi-phasic GABA concentration response curves were obtained, supporting further a heterogeneity of expressed receptors. We also noticed here that injection of 2.5 fmol: 2.5 fmol: 2.5 fmol mRNA coding for α6, β3 and δ, respectively, resulted in spontaneous current in the absence of GABA and in current amounting to about 1100 nA upon application of 1 μM THDOC. In contrast, injection of 0.5 fmol : 0.5 fmol : 2.5 fmol mRNA did not result in spontaneous current and the induced current amounted to only about 50 nA on application of 1 μM THDOC. All these observations indicate that different expression conditions results in the formation of different receptors.
Thus, any comparison of non-concatenated with concatenated receptors should be done carefully. In spite of this difficulty, the properties of the non-concatenated receptors observed here were compared with those of concatenated β3-α6-δ/α6-β3 (R1), α6-β3-α6/β3-δ (R2), β3-δ-β3/α6-β3 (R3) and α6-β3-α6/δ receptors. Concatenated receptors resulted in lower current amplitudes elicited by 1 mM GABA than non-concatenated receptors. This phenomenon may be due to contamination of α6β3δ receptors by α6β3, and/or due to the additional formation of a receptor with a subunit arrangement not studied here. The dose–response curves for all concatenated receptors in respect to the major components were in a similar range as the one obtained for non-concatenated α6β3δ receptors. A low affinity component as observed in α6-β3-α6/β3-δ (R2) and β3-δ-β3/α6-β3 (R3) receptors was not evident in non-concatenated α6β3δ receptors, as expressed under our conditions. Two factors may contribute to this absence. First, properties of non-concatenated receptors depend strongly on the expression conditions and we may be chosen conditions that do not result in expression of low affinity receptors. Second, assuming equal contribution to the current of all four successfully expressed receptor configurations, the low affinity component would amount to only 14% of the total current. 100 μM THIP elicited similar current amplitudes as 1 mM GABA/1 μM THDOC in non-concatenated receptors. In concatenated receptors β3-α6-δ/α6-β3 (R1), α6-β3-α6/β3-δ (R2), β3-δ-β3/α6-β3 (R3) and α6-β3-α6/δ receptors, 100 μM THIP elicited 20–40% of the current elicited by 1 mM GABA/1 μM THDOC. This might be due to a small shift to the right in the THIP dose–response curves in concatenated receptors.
Controversial observations have been reported on the action of physiological concentrations of ethanol on α6β3δ receptors. While one group reported positive allosteric modulation (Wallner et al. 2003), a consortium of several groups was unable to reproduce the findings (Borghese et al. 2006). It is well documented that injection of the subunits in different ratios affects the expressed current properties (Hadley and Amin 2007, this paper). In principle, the observed discrepancy in ethanol effects could be due to formation of different receptor pentamers in different laboratories upon injection of genetic information coding for α6, β3 and δ subunits. With subunit concatenation, these different types of receptors can be expressed individually. However, we observed that none of the four receptors β3-α6-δ/α6-β3 (R1), α6-β3-α6/β3-δ (R2), β3-δ-β3/α6-β3 (R3) and α6-β3-α6/δ was modulated by ethanol in the concentration range of 7.5–30 mM. The latter concentration corresponds to 1.38‰ (w/v). We cannot fully exclude formation of an ethanol sensitive receptor configuration not studied here, but we consider this unlikely.
Relative abundance of the different δ subunit containing receptors
Our results on a functional study on α6β3δ GABAA receptors should be compared with a structural study on α4β3δ GABAA receptors. Using atomic force microscopy Barrera et al. (2008) determined stoichiometry and subunit arrangement of receptors expressed in tsA 201 cells. They showed that αβαδβ is the predominant subunit arrangement around the pore when viewed from the extra-cellular space, with 21% of the population exhibiting a distinct subunit arrangement of αβαβδ. Only a very small number of receptor entities were analyzed and these numbers should therefore be taken with care. In addition, unlike in other atomic force microscopical studies, the receptors were depicted here as spots and not as expected as rings with a central hole (Müller et al. 2002) and antibodies as spots and not as heart-shaped entities (Fritz et al. 1997), respectively. The above study was done at a structural level including receptors retained in the endoplasmic reticulum, whereas we focused on the channel function of δ subunit containing receptors located in the surface membrane. If it is assumed that α6 is similar to α4, αβαδβ receptors correspond to β3-α6-δ/β3-α6 (R5) and αβαβδ to α6-β3-α6/β3-δ (R2) in our study. We did not observe any functional expression of β3-α6-δ/β3-α6 (R5). From the present experiments it is difficult to conclude on relative abundance of the three expressing receptors. Subunit concatenation may affect expression levels. In addition, single channel conductance and open probability could differ between different receptors. Nevertheless, active, non-concatenated α6β3δ receptors probably constitute a mixture of β3-α6-δ/α6-β3 (R1), α6-β3-α6/β3-δ (R2), β3-δ-β3/α6-β3 (R3) and α6-β3-α6/δ where R1–R3 receptors are only active in the presence of neurosteroids. As discussed above, evidence for the expression of multiple receptors resulting from injection into Xenopus oocytes of genetic information coding for α6, β3 and δ has been presented (Hadley and Amin 2007; this paper).
Our findings indicate that the assembly properties of the δ subunit resemble that of the ε subunit (Bollan et al. 2008) in respect to the fact that both subunits can assume multiple positions in a receptor. However, the ε subunit apparently prefers alternative positions as compared with the δ subunit, except that both subunits seem to be able to occupy the position of the β subunit adjacent to the γ subunit in the major adult isoform.
The assembly properties of α6β3δ GABAA receptors should be compared with those of α1β3δ receptors (Kaur et al. 2009). While both β3-αx-δ/αx-β3 (R1) and αx-β3-αx/β3-δ (R2) are formed with x = 1 and 6, β3-αx-δ/β3-αx (R5) receptors are only formed with x = 1, and β3-δ-β3/αx-β3 (R3) are only formed with x = 6. Whether or not αx-β3-αx/δ receptors, which are formed with x = 6, are also formed with x = 1 remains to be shown. As we assume, on the basis of functional properties that R5 receptors may only be formed with low efficiency upon delivery of genetic information coding for the α1, β3 and δ subunits (Kaur et al. 2009), there may only be subtle assembly differences between α1β3δ and α6β3δ receptors.
In summary, we have shown that in GABAA receptors containing the α6, β3 and δ subunits, the δ subunit exhibits the ability to promiscuously assemble into different subunit arrangements at least after expression in Xenopus oocytes. Further we show that most of the δ subunit containing receptors remain relatively silent in the absence of neurosteroid and that all are insensitive to ethanol. The architecture of α6β3δ GABAA receptors and their positioning in cerebellar granule cells needs to be established.