The expression of GABA_Aβ subunit isoforms in synaptic and extrasynaptic receptor populations of mouse dentate gyrus granule cells

Abstract

The subunit composition of GABA_A receptors influences their biophysical and pharmacological properties, dictates neuronal location and the interaction with associated proteins, and markedly influences the impact of intracellular biochemistry. The focus has been on α and γ subunits, with little attention given to β subunits. Dentate gyrus granule cells (DGGCs) express all three β subunit isoforms and exhibit both synaptic and extrasynaptic receptors that mediate ‘phasic’ and ‘tonic’ transmission, respectively. To investigate the subcellular distribution of the β subunits we have utilized the patch-clamp technique to compare the properties of ‘tonic’ and miniature inhibitory postsynaptic currents (mIPSCs) recorded from DGGCs of hippocampal slices of P20–26 wild-type (WT), β₂^−/−, β_2N265S (etomidate-insensitive), α₁^−/− and δ^−/− mice. Deletion of either the β₂ or the δ subunit produced a significant reduction of the tonic current and attenuated the increase of this current induced by the δ subunit-preferring agonist 4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol (THIP). By contrast, mIPSCs were not influenced by deletion of these genes. Enhancement of the tonic current by the β_2/3 subunit-selective agent etomidate was significantly reduced for DGGCs derived from β_2N265S mice, whereas this manipulation had no effect on the prolongation of mIPSCs produced by this anaesthetic. Collectively, these observations, together with previous studies on α₄^−/− mice, identify a population of extrasynaptic α₄β₂δ receptors, whereas synaptic GABA_A receptors appear to primarily incorporate the β₃ subunit. A component of the tonic current is diazepam sensitive and is mediated by extrasynaptic receptors incorporating α₅ and γ₂ subunits. Deletion of the β₂ subunit had no effect on the diazepam-induced current and therefore these extrasynaptic receptors do not contain this subunit. The unambiguous identification of these distinct pools of synaptic and extrasynaptic GABA_A receptors should aid our understanding of how they act in harmony, to regulate hippocampal signalling in health and disease.

In mammalian brain, ‘fast’ neuronal inhibition is predominantly mediated by γ-aminobutyric acid (GABA), acting to transiently activate synaptic ionotropic GABA_A receptors and cause ‘phasic’ inhibition. However, in some neurones, GABA additionally activates a topographically discrete pool of extrasynaptic receptors to mediate a persistent ‘tonic’ form of inhibition, which functions in partnership with phasic inhibition, to influence neuronal excitability (Farrant & Nusser, 2005; Walker & Semyanov, 2007). In certain neurones phasic and tonic inhibition are mediated by distinct GABA_A receptor isoforms, with each subtype assembled from a possible palette of 19 subunits (α_1–6, β_1-3, γ_1-3, δ, θ, π, ɛ, ρ_1-3) to form a pentamer. Such receptor isoforms have distinct physiological and pharmacological properties and their expression may be differentially influenced by psychiatric and neurological disorders (Fritschy & Brunig, 2003; Whiting, 2003; Mohler, 2007).

Although receptors incorporating the γ₂ subunit may occur both synaptically and extrasynaptically (Devor et al. 2001; Caraiscos et al. 2004; Luscher & Keller, 2004; Zhang et al. 2007), δ-subunit-containing receptors are primarily located extra- or peri-synaptically (Nusser et al. 1998; Wei et al. 2003). For dentate gyrus granule cells (DGGCs) and thalamocortical relay neurones, the α₄ subunit is the principal δ-subunit partner (Wisden et al. 1992; Sur et al. 1999; Chandra et al. 2006), whereas in cerebellar granule cells (CGCs) and particular hippocampal interneurones, it coassembles with α₆ and α₁ subunits, respectively (Nusser et al. 1998; Glykys et al. 2007). By contrast, relatively little is known regarding the identity of the β subunit isoforms that coassemble with the δ subunit. However, this issue is of considerable physiological importance, given that several intracellular proteins, known to regulate receptor trafficking, membrane insertion and surface stability, can differentially interact with different β subunit isoforms (Chen & Olsen, 2007; Michels & Moss, 2007). Furthermore, the large intracellular loop of the β subunit can be phosphorylated by a variety of kinases, in an isoform selective manner, to consequently influence the trafficking and function of GABA_A receptors (Brandon et al. 2003; Kittler & Moss, 2003; Houston et al. 2007). Pharmacologically, the enhancement of receptor function by certain general anaesthetics, anticonvulsants, neurosteroids and potentially ethanol is also influenced by the nature of the β subunit isoform (Belelli et al. 1997; Hill-Venning et al. 1997; Wallner et al. 2003; Shen et al. 2007).

We previously demonstrated that both synaptic and extrasynaptic receptors of thalamic ventrobasal (VB) neurones contain the β₂ subunit (Belelli et al. 2005; Peden et al. 2008). DGGCs, in common with VB neurones, express extrasynaptic GABA_A receptors composed of α₄ and δ subunits, although the identity of the partner β subunit is not known. To characterize the synaptic and extrasynaptic GABA_A receptors of DGGCs we have utilized an electrophysiological approach, in combination with mice genetically engineered to either lack specific GABA_A receptor subunits (β₂, α₁, δ), or to express an etomidate-insensitive mutant β₂ subunit (β_2N265S).

We report that in contrast to thalamic VB neurones, the β₂ subunit is not expressed at DGGC synapses and alternatively we present evidence for synaptic β₃-GABA_A receptors and the presence of a minor population of synaptic β₁-GABA_A receptors. However, in common with thalamic relay neurones, extrasynaptic diazepam-insensitive δ-GABA_A receptors of DGGCs incorporate the β₂ subunit. Therefore, the α₄β₂δ isoform is an important receptor subtype in two distinct neuronal populations. A component of the tonic current is mediated by diazepam-sensitive extrasynaptic receptors containing α₅ and γ₂ subunits (Zhang et al. 2007). In contrast to the extrasynaptic δ-GABA_A receptors, we find that these extrasynaptic α₅-GABA_A receptors do not contain the β₂ subunit.

In conclusion, within the neuronal cell soma of DGGCs, evidence is presented for the differential expression of β subunit isoforms, not only between the synaptic and extrasynaptic pools, but additionally within distinct populations of extrasynaptic receptors. The unambiguous identification of the subunit composition of these heterogeneous extrasynaptic receptors will be an important step in advancing our understanding of how their properties are tailored to act in concert with their synaptic counterparts to influence neuronal signalling in both physiological and pathophysiological scenarios.

Methods

The α₁^−/−, β₂^−/− and β_2N265S mice were generated on a mixed C57BL6–129SvEv background at the Merck Sharp & Dohme Research Laboratories at the Neuroscience Research Centre in Harlow as previously described (Sur et al. 2001; Reynolds et al. 2003). The δ^−/− mice were also generated on a mixed C57BL/6J–129Sv/SvJ background at the University of Pittsburgh as previously described (Mihalek et al. 1999). Experiments were conducted on slices prepared from the first two generations of WT, α₁^−/−, β₂^−/−, β_2N265S and δ^−/− breeding pairs derived from the corresponding heterozygous ^+/− mice bred at the University of Dundee.

Slice preparation

Hippocampal, or cerebellar slices were prepared from mice of either sex (P20–26) according to standard protocols as previously described (Belelli & Herd, 2003; Reynolds et al. 2003). Animals were killed by cervical dislocation in accordance with Schedule 1 of the UK Animals (Scientific Procedures) Act 1986. The brain was rapidly dissected and placed in oxygenated ice cold artificial cerebrospinal fluid (aCSF) solution containing (mm): 225 sucrose, 2.95 KCl, 1.25 NaH₂PO₄, 26 NaHCO₃, 0.5 CaCl₂, 10 MgSO₄, 10 d-glucose (pH 7.4; 330–340 mosmol l⁻¹). The tissue was maintained in ‘ice-cold’ aCSF. Hippocampal horizontal (300–350 μm thick), or cerebellar parasagittal (250 μm thick) slices were cut using a Vibratome (St Louis, MO, USA). The slices were incubated at room temperature in an oxygenated, extracellular solution (ECS) containing (mm): 126 NaCl, 2.95 KCl, 26 NaHCO₃, 1.25 NaH₂PO₄, 2 CaCl₂, 10 d-glucose and 2 MgCl₂ (pH 7.4; 300–310 mosmol l⁻¹) for a minimum of 1 h prior to experimentation.

Electrophysiology

Whole-cell patch clamp recordings were performed at 35°C from DGGCs or Purkinje neurones visually identified with an Olympus BX51 (Olympus, Southall, UK) microscope equipped with DIC/IR optics. Patch electrodes were prepared from thick walled borosilicate glass (Garner Glass Co., Claremont, CA, USA) and had open tip resistances of 4–5 MΩ when filled with an intracellular solution that contained (mm): 135 CsCl, 10 Hepes, 10 EGTA, 2 Mg-ATP, 1 CaCl₂, 1 MgCl₂, 5 QX-314 (pH 7.3 with CsOH, 300–305 mosmol l⁻¹). Miniature inhibitory postsynaptic currents (mIPSCs) were recorded using an Axopatch 1D or Axopatch 200B amplifier (Molecular Devices, Union City, CA, USA) at a holding potential of −60 mV in ECS additionally containing 2 mm kynurenic acid (Sigma-Aldrich, UK) and 0.5 μm tetrodotoxin (TTX; Tocris Bioscience, Bristol, UK) to block ionotropic glutamate receptors and sodium-dependent action potentials, respectively.

Drug application

R-(+)-Etomidate (3 × 10⁻³m), 4,5,6,7-tetrahydro-isoxazolo[5,4-c]pyridin-3-ol (THIP, gaboxadol; 10⁻³m) and bicuculline methobromide (3 × 10⁻²m) were prepared as concentrated aqueous stock solutions, whereas diazepam (10⁻³m) was dissolved in DMSO. Stock solutions (1000×) were diluted in ECS to the desired final concentration. The final maximum DMSO concentration (0.1%) had no effect on the properties of the mIPSCs, or tonic current (data not shown). All modulatory agents were applied via the perfusion system (2–4 ml min⁻¹) and allowed to infiltrate the slice for a minimum of 10 min before recordings were acquired in the presence of the drug. With the exception of THIP, which was a generous gift of Dr Bjarke Ebert (Lundbeck), and etomidate (Organon), all drugs tested were obtained from either Sigma-Aldrich (Poole, Dorset, UK), or Tocris Bioscience (Bristol, UK).

Data analysis

Data were recorded onto a digital audio tape using a Biologic DTR 1200 recorder and analysed offline using the Strathclyde Electrophysiology Software, WinEDR/WinWCP (Belelli et al. 2005). Individual mIPSCs were detected using a low amplitude (−4 pA, 4 ms duration) threshold detection algorithm followed by visual scrutiny to avoid spurious detections. Accepted events were analysed for peak amplitude, 10–90% rise time, charge transfer and time for events to decay by 50% (T50) and 90% (T90). To minimize the contribution of dendritically generated currents, which are subject to cable filtering, the analysis was restricted to events with a rise time ≤ 1 ms. A minimum of 50 accepted events were digitally averaged by alignment at the mid-point of the rising phase, and the mIPSC decay fitted (98–5% of the peak amplitude) by either monoexponential (y(t) =Ae^(−t/τ)), or biexponential (y(t) =A₁e^(−t/τ₁)+A₂e^(−t/τ₂)) functions using the least squares method, where A is amplitude, t is time and τ is the decay time constant. Analysis of the s.d. of residuals and use of the F test to compare goodness of fit revealed that the decay of the average mIPSC waveform was always best fitted with the sum of two exponential components. Thus, a weighted decay time constant (τ_w) was also calculated according to the equation: τ_w=τ₁P₁+τ₂P₂, where τ₁ and τ₂ are the decay time constants of the first and second exponents and P₁ and P₂ are the proportions of the synaptic current decay described by each function. The mIPSC frequency was determined over a 2 min recording period (20 s bins) with the EDR program using a detection method based on the rate of rise of the slowest events (35–40 pA ms⁻¹) and subsequent visual scrutiny. The efficacy of the phasic current was calculated by multiplying the charge transfer associated with the mean individual mIPSC (Q_mIPSC) by the number of mIPSCs per unit of time (effectively the frequency).

The tonic current was calculated as the difference between the holding current before and after application of 30 μm bicuculline methobromide. The holding current and RMS (i.e. s.d.) were sampled every 51.2 ms over a 2 min period for each recording condition. At the sampling rate of 10 kHz, 512 baseline points for each 51.2 ms provided one data point. Epochs containing synaptic events or an unstable baseline were excluded from the analysis. A minimum of 100 data points were measured for each recording condition (i.e. in the absence and presence of drug). The significance of a drug effect upon the tonic conductance was calculated by a paired comparison of the holding current before and after drug application.

All results are reported as the arithmetic mean ± standard error of mean (s.e.m.). The properties of phasic (i.e. mIPSC properties) and tonic (i.e. RMS) GABA_A receptor-mediated inhibition were not different between the WT strains originally generated at the Neuroscience Research Centre in Harlow and at the University of Pittsburgh and hence the data from both WT strains were pooled. When data are presented as normalized, the mean value is calculated by averaging the normalized change for each cell following drug application. Statistical significance of the data was assessed with Student/s t test (paired or unpaired), and repeated measures ANOVA (one- or two-way, RM ANOVA) followed post hoc by the Newman–Keuls test as appropriate, using the SigmaStat software package (Systat Software Inc., San Jose, CA, USA). The large sample approximation of the Kolmogorov–Smirnov (KS) test (SPSS, SPSS Inc., Chicago, IL, USA) was used to compare the distribution of individual mIPSC parameters. For a stringent comparison, the level of significance was set at P < 0.01 for KS tests.

Results

Extrasynaptic GABA_A receptors of DGGCs

Deletion of the β₂ subunit reduces the THIP-sensitive, but not the diazepam-sensitive, tonic conductance

Using bicuculline to define the resting tonic current (I_tonic) this antagonist revealed an I_Tonic of −16.6 ± 2.7 pA (n = 14) in DGGCs derived from WT mice (P < 0.001, paired Student/s t test, Fig. 1A, left panel). Concomitantly, bicuculline (30 μm) reduced membrane (root mean square, RMS) noise from 4.2 ± 0.2 pA to 2.8 ± 0.1 pA (P < 0.001, paired t test). By comparison, for β₂^−/− granule cells, the outward current induced by 30 μm bicuculline was significantly reduced compared with WT (β₂^−/−I_Tonic=−7.5 ± 1.3 pA, n = 13, P < 0.01 versus WT; unpaired t test; Table 1, Fig. 1A, right panel and Fig. 1C). Similarly, the membrane noise of the β₂^−/− neurones was significantly lower than that of WT (β₂^−/−, 3.4 ± 0.2 pA, P < 0.01), which was now only modestly, albeit significantly reduced by bicuculline (2.6 ± 0.1 pA, P < 0.001, paired t test). These results suggest that a significant proportion of extrasynaptic receptors in DGGCs contain the β₂ subunit (see also Discussion).

Figure 1. For DGGCs deletion of the β₂ subunit decreases the tonic current and the THIP-induced current, but does not influence the enhancement of the current by diazepam.

A, whole-cell recordings obtained from exemplar DGGCs of WT (left panel) and β₂^−/− (right panel) mice before and after the bath application of bicuculline (30 μm). B, whole-cell recordings and the corresponding all-points histograms (normalized to the holding current recorded in the presence of bicuculline, right panel) obtained from DGGCs of WT (left panel) and β₂^−/− (right panel) mice under control conditions (black), in the presence of THIP (1 μm, white) and subsequent to the application of bicuculline (30 μm, grey). Application of 30 μm bicuculline (grey) reveals a GABA_A receptor-mediated tonic current, which is significantly reduced by deletion of the β₂ subunit. Note also the significant decrease in the magnitude of the THIP-evoked current upon deletion of the β₂ subunit. C, bar graph summarizing the effect of the β₂ subunit deletion upon the tonic current and the THIP and diazepam-evoked current. Note that the amplitude of the current evoked by diazepam (1 μm) is not influenced by deletion of the β₂ subunit. Data were obtained from 5–14 cells. Error bars indicate the s.e.m.*P < 0.05, **P < 0.01, unpaired t test.

Table 1.

Summary of the properties of the phasic and tonic GABA_AR-mediated transmission in mouse WT and β₂^−/− DGGCs

	WT	β2−/−
Rise time (ms)	0.37 ± 0.01	0.39 ± 0.02
Peak amplitude (pA)	−68.8 ± 2.4	−65.2 ± 3.9
τw (ms)	8.9 ± 0.4	10.3 ± 0.6
Charge transfer (fC)	−582.4 ± 33.3	−623.1 ± 43.4
Frequency (Hz)	3.2 ± 0.3	2.7 ± 0.6
Tonic current (pA)	−16.6 ± 2.7	−7.5 ± 1.3*
Phasic charge (fC)	−1864	−1682
Tonic charge (fC)	−16600	−7500
Tonic contribution to total charge transfer	90%	82%

n = 13–29. *P < 0.05.

Tonic inhibition in dentate granule cells is mediated, at least in part, by peri- or extra-synaptic GABA_A receptors composed of α₄, β and δ subunits (Nusser & Mody, 2002; Stell & Mody, 2002; Farrant & Nusser, 2005; Chandra et al. 2006), although evidence for an additional population of α₅βγ₂ receptors has recently been presented (Zhang et al. 2007). Given that I_Tonic is reduced in β₂^−/− DGGCs, we considered whether this manipulation reduced the number of δ- and/or γ₂-GABA_A receptors. Therefore, we investigated the actions of the GABA_A receptor agonist THIP on the I_Tonic of β₂^−/− DGGCs. THIP is selective for recombinant receptors lacking a γ subunit, e.g. extrasynaptic δ-GABA_A receptors (Brown et al. 2002; Storustovu & Ebert, 2006), and, in agreement, this agonist selectively increases the I_Tonic of thalamic, hippocampal and cortical neurones known to express the δ subunit (Belelli et al. 2005; Cope et al. 2005; Maguire et al. 2005; Drasbek & Jensen, 2006). Consistent with previous studies, for WT DGGCs, THIP (1 μm) induced a significant inward current (WT, −38 ± 5 pA, n = 14, P < 0.001, paired t test; Fig. 1B and C) that was sensitive to bicuculline. However, in β₂^−/− mice, the inward current induced by 1 μm THIP was significantly reduced relative to WT (β₂^−/−, −19 ± 6 pA, n = 8 versus WT, P < 0.05, unpaired t test; Fig. 1B and C).

As mentioned above, recent observations have indicated a population of extrasynaptic α₅βγ₂ receptors in adult mouse DGGCs exposed to 5 μm GABA. To investigate the identity of the β subunit of these receptors the action of diazepam was compared between WT and β₂^−/− DGGCs. Diazepam (1 μm), in the presence of GABA (5 μm), produced a modest, albeit significant, inward current in both WT and β₂^−/− DGGCs (WT, −16.4 ± 6 pA, n = 6; β₂^−/−, −15.3 ± 2 pA, n = 5, P < 0.05 paired t test for both genotypes; Fig. 1C) but, in contrast to THIP, this effect was not significantly different between the two genotypes (P > 0.05, unpaired t test WT versusβ₂^−/−). In the absence of added GABA, diazepam produced an inward current of similar amplitude in WT DGGCs (WT, no added GABA, −13.1 ± 2 pA, n = 5, P > 0.05 versus WT with 5 μm GABA, unpaired t test) indicating that similar receptor isoforms are recruited under both experimental conditions.

Collectively, these results suggest the presence of extrasynaptic β₂-GABA_A receptors, which incorporate α₄ and δ subunits, although, in agreement with Zhang et al. (2007), a diazepam-sensitive extrasynaptic isoform incorporating β_1/3, α₅ and γ₂ subunits is also resident (see below).

The enhancement of the tonic current by etomidate is reduced by the β_2N265S mutation

The deletion of the β₂ subunit may induce compensatory gene transcription, which complicates interpretation of such experiments (Brickley et al. 2001; Sur et al. 2001; Ponomarev et al. 2006). Therefore, to investigate the veracity of the conclusions drawn from the β₂ null experiments, we recorded from DGGCs derived from β_2N265S‘knock-in’ mice. For these mice, the β₂ protein is replaced by a mutant subunit, engineered to be less sensitive to the positive allosteric effects of the general anaesthetic etomidate (Belelli et al. 1997; Reynolds et al. 2003; Belelli et al. 2005). Importantly, the mutation is functionally silent, until challenged with etomidate and consequently a comparison of the actions of this general anaesthetic on neurones derived from WT and mutant mice can act as a ‘finger print’ to infer the presence of the β₂ subunit.

In agreement with our studies on thalamic relay neurones (Belelli et al. 2005), the I_Tonic revealed by 30 μm bicuculline in β_2N265S DGGCs was not significantly different from that of WT neurones (β_2N265SI_Tonic=−26 ± 8.3 pA, n = 6, P > 0.05 versus WT, unpaired t test; Fig. 2B), i.e. the mutation appears functionally silent. However, this manipulation greatly influenced the interaction of etomidate with the extrasynaptic receptors. In WT DGGCs, etomidate (3 μm) induced a bicuculline-sensitive inward current of −42 ± 5 pA (n = 5, P < 0.001, paired t test; Fig. 2A and B). For β_2N265 neurones, the inward current induced by 3 μm etomidate was significantly reduced compared with WT (β_2N265S, −23 ± 4 pA, P < 0.05 versus WT, unpaired t test; Fig. 2A and B).

Figure 2. The etomidate-induced increase of the tonic conductance is reduced for DGGCs derived from β_2N265S mice.

A, whole-cell recordings and the corresponding all-points histograms (normalized to the holding current recorded in the presence of bicuculline) obtained from exemplar DGGCs of WT (left panel) and β_2N265S (right panel) mice under control conditions (black), in the presence of etomidate (3 μm, white) and subsequent to the application of bicuculline (30 μm, grey). Application of 30 μm bicuculline (grey) reveals a GABA_A receptor-mediated tonic current of similar magnitude in WT and β_2N265S DGGCs. By contrast, note that the amplitude of the etomidate-evoked current is reduced in DGGCs of β_2N265S compared with WT mice. B, bar graph summarizing the effect of the β₂ subunit mutation upon the tonic current and the etomidate-evoked current. Data were obtained from 5–14 cells. Error bars indicate the s.e.m.*P < 0.05, unpaired t test.

Deleting the δ subunit reduces the tonic current and the effect of etomidate

The interpretation of the β_2N265S experiments assumes the inward current induced by etomidate to be mediated, at least in part, by extrasynaptic δ-GABA_A receptors (see Discussion). However, DGGCs express both extrasynaptic δ- and α₅-GABA_A receptors (Zhang et al. 2007), and the α₅-GABA_A receptors native to hippocampal CA1 neurones are highly sensitive to etomidate (Caraiscos et al. 2004; Cheng et al. 2006). We therefore compared the actions of this anaesthetic on the tonic current of DGGCs derived from WT and δ^−/− mice. In agreement with others (Stell et al. 2003; Maguire et al. 2005), the outward current induced by 30 μm bicuculline was significantly reduced in δ^−/− DGGCs compared with WT neurones (I_Tonic: WT, −16.6 ± 2.7 pA, n = 14; δ^−/−, −6.2 ± 2.4 pA, n = 5, P < 0.05, unpaired t test; Fig. 3A and B). Additionally, the membrane RMS was significantly lower in recordings from δ^−/− neurones (WT, 4.2 ± 0.2 pA; δ^−/−, 3.0 ± 0.1 pA, P < 0.05, unpaired t test), reflecting a decreased number of extrasynaptic receptors. Indeed, the magnitude of the remaining tonic current of δ^−/− neurones was not significantly different from that of β₂^−/− neurones (I_Tonic: β₂^−/−, −7.5 ± 1.3 pA, n = 13, P > 0.05 versusδ^−/−, unpaired t test). Deletion of the δ subunit significantly reduced the inward current induced by 3 μm etomidate compared with WT neurones (WT, −42 ± 5 pA, n = 5; δ^−/−, −17 ± 5 pA, n = 6, P < 0.01, unpaired t test; Fig. 3A and B). Collectively, these data propose that DGGCs, in common with thalamic VB neurones, express a population of extrasynaptic receptors composed of α₄β₂δ subunits.

Figure 3. Deletion of the δ subunit reduces the tonic current and the etomidate-induced increase of the tonic conductance of DGGCs.

A, whole-cell recordings and the corresponding all-points histograms (normalized to the holding current recorded in the presence of bicuculline) obtained from exemplar DGGCs of WT (left panel) and δ^−/− (right panel) mice under control conditions (black), in the presence of etomidate (3 μm, white) and subsequent to the application of bicuculline (30 μm, grey). Application of 30 μm bicuculline (grey) reveals that the GABA_A receptor-mediated tonic current of δ^−/− DGGCs is significantly smaller versus WT DGGCs. Note that the amplitude of the etomidate-evoked current is also reduced in DGGCs of δ^−/−versus WT mice. B, bar graph summarizing the effect of the δ subunit deletion upon the tonic current and the amplitude of the etomidate-evoked current. Data were obtained from 5–14 cells. Error bars indicate the s.e.m.**P < 0.01, unpaired t test.

Synaptic GABA_A receptors of DGGCs

The mIPSCs of DGGCs occurred with a frequency of 3.2 ± 0.3 Hz, exhibited a peak amplitude of −68.8 ± 2.4 pA and decayed with a time constant (τ_W) of 8.9 ± 0.4 ms. Hence, considering the frequency and total charge passed per mIPSC, under these experimental conditions (i.e. TTX treated) the current mediated by the extrasynaptic receptors is dominant, contributing ∼90% of the total charge transferred (Table 1). As described above, DGGCs express a population of diazepam-sensitive extrasynaptic GABA_A receptors. Additionally, in the presence of 5 μm GABA, diazepam (1 μm) prolonged the mIPSC decay of DGGCs, with no effect on their rise time, or amplitude (% increase, τ_w= 61 ± 7%, n = 6, P < 0.001, one-way RMA). As noted above, in common with the extrasynaptic conductance, the effect of diazepam on mIPSCs was not altered in the absence of added GABA (% increase, τ_w= 61 ± 9%, n = 5, P > 0.05 versus+ 5 μm GABA, two-way RMA). Hence, these synaptic receptors contain the γ₂ subunit, coexpressed with the α₁, α₂, α₃, or the α₅ subunit. By contrast, although THIP (1 μm) activated a population of δ-GABA_A receptors, it had no effect on the amplitude or kinetics of mIPSCs (% control, τ_w= 102 ± 2%, peak amplitude = 105 ± 2%, rise time = 103 ± 3%, n = 14, P > 0.05, one-way RMA).

The influence of deleting the β₂ or the α₁ subunit on DGGC mIPSCs

Deletion of the β₂ subunit had no significant effect on the frequency (2.7 ± 0.6 Hz), peak amplitude (−65.2 ± 3.9 pA), or decay kinetics (τ_W= 10.3 ± 0.6 ms) of DGGC mIPSCs and consequently no influence on the charge carried by synaptic GABA_A receptors (Fig. 4A; Table 1). However, given the considerable effect of deleting the β₂ subunit on the tonic current (Fig. 1), this manipulation greatly reduced the total (i.e. synaptic + extrasynaptic) charge transferred by GABA_A receptors by ∼50%, although given the dominant influence of the residual tonic conductance, extrasynaptic receptors still contributed ∼82% to the total charge transferred in β₂^−/− neurones (Table 1).

Figure 4. The effect of deletion of the β₂ and α₁ subunit upon the properties of DGGC mIPSCs.

A, representative, superimposed normalized ensemble averages of GABA_AR mIPSCs recorded from WT (black) and β₂^−/− (grey) and B, WT (black) and α₁^−/− (grey) DGGCs. For comparison a representative, superimposed normalized ensemble average of GABA_AR mIPSCs recorded from an exemplar WT Purkinje neurone (dotted line) is shown in B. Note that deletion of the β₂ subunit does not affect any of the DGGC mIPSC properties, whereas deletion of the α₁ subunit produces a modest prolongation of the decay kinetics.

Restriction of the β₂ subunit to an extrasynaptic receptor pool may be a consequence of the absence of an appropriate subunit partner. In contrast to DGGCs, we have previously reported the β₂ subunit deletion to greatly decrease the amplitude of mIPSCs of Purkinje and thalamocortical VB neurones (Haythornthwaite, 2004; Belelli et al. 2005). However, the synaptic GABA_A receptors of these neurones incorporate the α₁ subunit, i.e. α₁β₂γ₂ GABA_A receptors (Reynolds et al. 2003; Belelli et al. 2005; Kralic et al. 2006), the most prevalent subtype in the mammalian CNS (Whiting, 2003), whereas DGGCs express only relatively low levels of α₁ subunit mRNA and protein (Wisden et al. 1992; Sperk et al. 1997; Hutcheon et al. 2004). In agreement, deletion of the α₁ subunit had no effect on the frequency, peak amplitude, or rise time of DGGC mIPSCs, although these synaptic currents were modestly prolonged relative to WT mIPSCs (WT: τ_w= 9.9 ± 0.5 ms, n = 10; α₁^−/−: τ_w= 11.9 ± 0.6 ms, n = 15; P < 0.05, unpaired t test; see Fig. 4B for an exemplar neurone). Given that Purkinje neurone synapses contain α₁β₂γ₂ GABA_A receptors (Reynolds et al. 2003; Kralic et al. 2006), we determined the properties of their mIPSCs for comparison. Such mIPSCs decayed much more rapidly (τ_W= 3.1 ± 0.1 ms, n = 33 neurones) than those of DGGCs, consistent with synaptic expression of the α₁ subunit (Fig. 4B; Kralic et al. 2006). Furthermore, in contrast to DGGCs, α₁^−/− Purkinje neurones were devoid of mIPSCs (not shown – see also Kralic et al. 2006). Thus, in conclusion, although the β₂ subunit is clearly expressed extrasynaptically by DGGCs, it is conceivable that the lack of an appropriate α subunit partner precludes receptors incorporating this subunit to congregate at the synapse.

The influence of etomidate on the mIPSCs of WT and β_2N265S DGGCs

The β₂‘knock-out’ experiments suggest that DGGC synaptic GABA_A receptors contain the β₃ and/or the β₁ subunit, coupled with the γ₂ subunit (diazepam sensitive). Etomidate preferentially facilitates GABA responses mediated by β₃ or β₂ compared with β₁ subunit containing receptors (Belelli et al. 1997; Hill-Venning et al. 1997). Here, 3 μm etomidate had no effect on the rise time, or frequency, of WT DGGC mIPSCs and only modestly increased their amplitude (11 ± 6% increase, P > 0.05, one-way RMA). However, this anaesthetic clearly prolonged the mIPSC decay (% increase, τ_W= 75 ± 12%n = 6, P < 0.001, one-way RMA – see Fig. 5). These data, together with the lack of effect on mIPSCs of deleting the β₂ subunit, suggest the presence of synaptic β₃-GABA_A receptors.

Figure 5. The effect of etomidate to prolong mIPSCs is similar for DGGCs derived from WT and β_2N265S mice.

A, representative, normalized superimposed ensemble averages of GABA_AR mIPSCs from WT (left) and β_2N265S (right) DGGCs recorded before (black) and after (grey) application of etomidate (3 μm). B, bar graph summarizing the effect of etomidate (3 μm) upon the mIPSC properties (normalized to percentage of control) of WT and β_2N265S DGGCs. Data were obtained from 6–9 cells. Error bars indicate the s.e.m.**P < 0.01, ***P < 0.001, 1-way RM ANOVA.

Given that gene deletion may trigger compensatory changes, we additionally investigated the actions of etomidate on the mIPSCs of DGGCs derived from β_2N265S mice. We have previously reported the effect of this anaesthetic on the decay phase of mIPSCs of cerebellar Purkinje and thalamic VB neurones to be blunted by this mutation (Reynolds et al. 2003; Belelli et al. 2005). In common with WT, 3 μm etomidate had no effect on the rise time, or frequency of mIPSCs recorded from β_2N265S DGGCs and only caused a modest increase of amplitude (7 ± 3% increase, n = 9, P < 0.05, one-way RMA). However, the anaesthetic again produced a clear prolongation of their decay (% increase of τ_w= 78 ± 9%, n = 9, P < 0.001, one-way RMA – see Fig. 5), an effect not significantly different from that for WT neurones (WT versusβ_2N265SP > 0.05, two-way RMA; Fig. 5B). Collectively, these data suggest the presence of synaptic β₃-GABA_A receptors.

To investigate whether all DGGC synaptic GABA_A receptors are sensitive to etomidate, we determined the influence of this anaesthetic on individual mIPSC T₉₀ values (the time taken for the mIPSC to decay from peak amplitude to 10% of peak). Cumulative probability plots of T₉₀ values revealed a limited population of mIPSCs in all neurones tested that were insensitive to this concentration (3 μm) of etomidate (Fig. 6A, see shaded area). Similar results were obtained for β_2N265S neurones (Fig. 6B). The presence of a population of synaptic GABA_A receptors incorporating the β₁ subunit provides a parsimonious explanation for these data.

Figure 6. A small population of DGGC mIPSCs are insensitive to the action of etomidate both in WT and β_2N265S mice.

Cumulative probability plots of the T90 values of mIPSCs recorded before (black) and after application of 3 μm etomidate (grey) from DGGCs of WT (A) and β_2N265S (B) mice. C, cumulative probability plots of the T90 values of mIPSCs recorded before (black) and after application of 30 μm pentobarbitone (grey) from WT DGGCs. The plots are constructed from T90 values obtained from 826 to 1586 mIPSCs collected from 6–9 representative DGGCs. The grey shaded area illustrates the area of overlap between the T90 distributions following etomidate, but not pentobarbitone application (T90 ≤ 10 ms; control versus 3 μm etomidate). Note that the mIPSCs occurring within this region were insensitive to etomidate. By contrast, pentobarbitone prolongs the decay of all DGGC mIPSCs.

In contrast to etomidate, the GABA-modulatory effects of anaesthetic barbiturates are not influenced by the isoform of the β subunit (Hill-Venning et al. 1997). Pentobarbitone (30 μm) produced a significant prolongation of the mIPSC decay (% increase of τ_w= 123 ± 3%, n = 7, P < 0.001, one-way RMA), but in contrast to etomidate, all mIPSCs were sensitive to this anaesthetic (Fig. 6C).

The effect of etomidate on DGGC mIPSCs is enhanced by deletion of the δ subunit

As described above, deletion of the δ subunit decreased extrasynaptic receptor expression in DGGCs and consequently greatly reduced the effect of etomidate. By contrast, etomidate-induced anaesthesia (loss of the righting reflex) is little influenced by this genetic manipulation (Mihalek et al. 1999). We therefore determined the effect of this anaesthetic on DGGC synaptic GABA_A receptors. There were no significant differences in the properties of the mIPSCs of δ^−/− DGGCs compared with WT. However, surprisingly etomidate (3 μm) was significantly more effective in prolonging the mIPSCs of δ^−/− DGGCs than those of WT neurones (% increase τ_w, WT, 92 ± 12%, n = 7, P < 0.001, one-way RMA; δ^−/−, 159 ± 20%, n = 9, P < 0.001, one-way RMA; WT versusδ^−/−P < 0.001, two-way RMA, Fig. 7A and B). The mechanisms underlying this unexpected form of plasticity and the consequences for interpreting behavioural studies are discussed below.

Figure 7. The effect of etomidate to prolong DGGC mIPSCs is increased by deletion of the δ subunit.

A, representative, normalized superimposed ensemble averages of GABA_AR mIPSCs from WT (left) and δ^−/− (right) DGGCs recorded before (black) and after (grey) application of etomidate (3 μm). B, bar graph summarizing the normalized (expressed as percentage control) effect of etomidate (3 μm) upon the mIPSC properties of WT and δ^−/− DGGCs. Note that deletion of the δ subunit significantly increased the etomidate-induced prolongation of the mIPSC decay. Data were obtained from 7–9 cells. Error bars indicate the s.e.m.***P < 0.001, 1-way RM ANOVA; †††P < 0.001, 2-way, RM ANOVA.

Discussion

The subunit composition of GABA_A receptors influences their biophysical and pharmacological properties, dictates neuronal location and their interaction with associated proteins, and markedly influences the impact of intracellular biochemistry. In this regard, the nature of the β subunit isoform has received little consideration. DGGCs have both synaptic and extrasynaptic receptors and express all three β subunit isoforms (Wisden et al. 1992; Sperk et al. 1997; Miralles et al. 1999; Pirker et al. 2000). The present study reveals a highly specific designation of β subunits into synaptic, or extrasynaptic receptor pools, in an isoform-selective manner.

DGGC extrasynaptic GABA_A receptors

Previous studies have revealed that deletion of either the α₄ or the δ subunit greatly reduced the tonic current of DGGCs (Stell et al. 2003; Chandra et al. 2006), although the nature of the accompanying β subunit is not known. Immunocytochemistry reveals a diffuse labelling of the β₂ subunit in DGGCs (Sperk et al. 1997), and in agreement here we report that deletion of the β₂ subunit and, confirming earlier studies, the δ subunit greatly reduced both the magnitude and associated membrane noise of the tonic current of DGGCs. In DGGCs the inward current induced by the agonist THIP is attenuated by deletion of the δ subunit (Maguire et al. 2005) and here the effect of this agonist was also substantially reduced for neurones derived from β₂^−/− mice. Further proof that DGGCs express a population of extrasynaptic β₂ subunit-containing receptors is provided by the effect of etomidate. This anaesthetic is selective for GABA_A receptors containing either β₂ or β₃ subunits (Belelli et al. 1997; Hill-Venning et al. 1997). Here etomidate enhanced the tonic current and this effect was attenuated for DGGCs derived from mice expressing a mutant β₂ subunit (β_2N265S) engineered to be less sensitive to this anaesthetic (Reynolds et al. 2003; Belelli et al. 2005). Etomidate, by enhancing GABA sensitivity, may recruit additional populations of extrasynaptic GABA_A receptors, as suggested for the actions of diazepam on hippocampal CA1 neurones (Prenosil et al. 2006). Therefore, it cannot be assumed that etomidate acts primarily on the same population(s) of extrasynaptic receptors that mediate the tonic current in the absence of the anaesthetic. However, the effect of etomidate on the tonic current was also substantially decreased by δ subunit deletion. Collectively, these studies suggest a significant component of the tonic conductance in DGGCs is mediated by the α₄β₂δ receptor isoform. Thalamic VB neurones express α₄β₂δ receptors (Belelli et al. 2005) and therefore this extrasynaptic receptor isoform appears to reside in two distinct neuronal populations.

A recent report revealed a component of the tonic current recorded from DGGCs (in the presence of GABA) to be enhanced by diazepam, an effect not evident in neurones derived from α₅^−/− mice, identifying an additional population of extrasynaptic GABA_A receptors incorporating α₅ and γ₂ subunits (Zhang et al. 2007). Here we confirm this effect of diazepam, and additionally demonstrate the benzodiazepine enhancement of the tonic current to be evident either in the presence or absence of added GABA. As noted above, for both β₂^−/− and δ^−/− DGGCs, there is a residual, albeit small, tonic current. Here, we report the tonic current of WT and β₂^−/− DGGCs to be similarly sensitive to diazepam, indicating that the benzodiazepine-induced enhancement of the tonic conductance is mediated by a population of extrasynaptic receptors that do not incorporate the β₂ subunit, presumably α₅β_1/3γ₂ receptors. Etomidate potently enhances the function of extrasynaptic α₅-GABA_A receptors of hippocampal CA1 neurones (Caraiscos et al. 2004; Cheng et al. 2006). Therefore, the residual etomidate-induced current evident in β_2N265S neurones may result from an interaction with extrasynaptic α₅β₃γ₂ receptors. However, the effect of etomidate on β_2N265S neurones must be interpreted with caution, as the mutation does not completely abolish the enhancement of GABA_A receptor function by this anaesthetic, nor are β₁-GABA_A receptors absolutely etomidate insensitive (Belelli et al. 1997, 2005; Hill-Venning et al. 1997; Reynolds et al. 2003).

Therefore, in common with hippocampal CA1 neurones (Caraiscos et al. 2004; Scimemi et al. 2005), DGGCs express distinct populations of benzodiazepine-insensitive (α₄β₂δ) and -sensitive (α₅β_1/3γ₂) extrasynaptic receptors. However, additional extrasynaptic receptor isoforms may contribute. In particular, for β₂^−/− DGGCs a low concentration of THIP still induced an inward current, albeit reduced in amplitude in comparison to WT neurones. THIP is a potent agonist of δ subunit-containing GABA_A receptors, and of receptors composed of only α and β subunits, but is a relatively ineffective agonist of γ₂ subunit-containing receptors (Storustovu & Ebert, 2006). Consequently, it is unlikely that the residual effect of THIP in β₂^−/− DGGCs is mediated by the proposed diazepam-sensitive α₅β_1/3γ₂ receptors and suggests the presence of a population of αβ_1/3 or αβ_1/3δ receptors.

In conclusion the extrasynaptic receptors of DGGCs are heterogeneous and a future challenge will be to understand how the properties of these distinct pools of receptors and their relative expression are tailored to act in concert to influence neuronal excitability.

DGGC synaptic GABA_A receptors

The prolongation of WT mIPSCs by etomidate is indicative of synaptic GABA_A receptors incorporating the β₂ or β₃ subunit. However, for each neurone, a small proportion of relatively ‘fast decaying’ etomidate-insensitive synaptic events were evident. Pentobarbitone does not discriminate between β subunit isoforms (Hill-Venning et al. 1997) and in contrast to etomidate, all DGGC mIPSCs were sensitive to this anaesthetic. These data advocate the presence of a limited population of ‘fast’ decaying mIPSCs mediated by synaptic β₁-GABA_A receptors. We previously demonstrated the effect of etomidate on mIPSCs recorded from Purkinje, or thalamic VB neurones derived from β_2N265S mice to be greatly reduced (Reynolds et al. 2003; Belelli et al. 2005). However, the effects of this anaesthetic on β_2N265S DGGC mIPSCs were indistinguishable from WT neurones, indicating a population of β₃-GABA_A synaptic receptors. Indeed, strong β₃ subunit immunoreactivity is evident in these neurones (Sperk et al. 1997; Miralles et al. 1999; Pirker et al. 2000). Importantly, in contrast to Purkinje and VB neurones, the β₂ subunit deletion had no effect on DGGC mIPSCs (Haythornthwaite, 2004; Belelli et al. 2005).

Adding complexity, immunoprecipitation studies suggest a substantial proportion of hippocampal GABA_A receptors may contain two different β subunit isoforms within the pentamer (Li & De Blas, 1997). Recombinant receptors incorporating β₁ with β₃ or β₂ subunits appear pharmacologically similar to β₁-GABA_A receptors (relatively insensitive to loreclezole, or etomidate –Fisher & Macdonald, 1997; Boulineau et al. 2005).

Therefore, in conclusion, DGGC inhibitory synapses predominantly contain β₃-GABA_A receptors that mediate relatively slowly decaying mIPSCs. Evidence is presented for a minor population of synaptic receptors composed of β₁ or potentially β₁ and β₃ subunits that mediate relatively fast decaying mIPSCs. Interestingly, thalamic reticular neurones exhibit two kinetically distinct populations of mIPSCs, which based on their differential sensitivity to loreclezole, also appear to be due to activation of synaptic β₁- and β₃-GABA_A receptors (Huntsman & Huguenard, 2006). Although DGGCs express the β₂ subunit, it appears to be restricted to a particular pool (α₄β₂δ) of extrasynaptic receptors. These data contrast with thalamic VB neurones, where β₂-GABA_A receptors are expressed both extrasynaptically (α₄β₂δ) and synaptically (α₁β₂γ₂–Belelli et al. 2005; Peden et al. 2008). The differential distribution of the β₂ subunit in DGGCs is intriguing given that synaptic GABA_A receptors are highly mobile, being recruited from, and dynamically interacting with, the extrasynaptic receptor pool (Thomas et al. 2005; Bogdanov et al. 2006; Charrier et al. 2006).

The exclusion of the β₂ subunit from the synapse may be a consequence of the lack of an appropriate α subunit partner. Although for recombinant receptors the β₂ subunit associates indiscriminately with various α subunits, in certain neurones it preferentially coassembles with α₁ and γ₂ subunits (Sur et al. 2001; Fritschy & Brunig, 2003; Whiting, 2003; Mohler, 2007). However, in the dentate gyrus, the expression of the α₁ subunit is limited, whereas the α₂ subunit, which preferentially coassembles with β₃ and γ₂ subunits, exhibits strong immunoreactivity and a clear synaptic localization (Fritschy & Mohler, 1995; Sperk et al. 1997; Fritschy & Brunig, 2003; Ramadan et al. 2003; Wei et al. 2003; Mohler, 2007). Furthermore, as reported here, deletion of the α₁ subunit only modestly influenced the mIPSC decay. Thus, a preferential association of the α₂ subunit with the β₃ and possibly the β₁ subunit may result in the exclusion of the β₂ subunit from DGGC synapses.

The β subunit isoform influences receptor expression and function

An array of GABA_A receptor associated proteins may regulate receptor trafficking, membrane stabilization and internalization and consequently influence neuronal excitability (Michels & Moss, 2007; Chen & Olsen, 2007). Many such interactions occur via an association with the large intracellular loop of the β subunit, a portion of the protein that differs substantially across the β subunit isoforms. As some interactions are β-subtype-specific, given the discrete subcellular distribution of β subunits in DGGCs, these associated proteins may differentially impact on tonic and phasic transmission.

The activity of resident kinases and phosphatases influence GABA_A receptor function and trafficking and may modify the response to allosteric modulators such as alcohol, benzodiazepines and neurosteroids (Hodge et al. 1999; Kittler & Moss, 2003; Herd et al. 2007; Qi et al. 2007). The large intracellular loop of the β subunit contains consensus sites for phosphorylation by several kinases and incorporates binding domains for intracellular proteins that influence kinase activity (Kittler & Moss, 2003; Houston et al. 2007). Importantly, both the association of such proteins, the kinase/phosphorylation profile and the functional consequences of phosphorylation are β isoform specific. For example, for β₃ or β₁ subunits a homologous serine residue can be phosphorylated by α-CAMK-II, whereas the β₂ subunit does not contain this site and consequently β₂-GABA_A receptors are insensitive to α-CAMK-II-dependent modulation (Houston et al. 2007). Clearly, future studies comparing the impact of manipulating neuronal kinase/phosphatase activity on extrasynaptic and synaptic receptor function and pharmacology in DGGCs will be of interest.

The β subunit isoform influences GABA_A receptor pharmacology

The use of β subunit ‘knock-in’ mice has revealed that β₂- and β₃-GABA_A receptors mediate distinct components of the behavioural repertoire produced by the general anaesthetic etomidate and the anticonvulsant loreclezole (Jurd et al. 2003; Reynolds et al. 2003; Cirone et al. 2004; Zeller et al. 2005; Groves et al. 2006). Extrasynaptic δ-GABA_A receptors are proposed targets for general anaesthetics, neurosteroids, ethanol and THIP (Stell et al. 2003; Wallner et al. 2003; Wei et al. 2003; Hanchar et al. 2005; Belelli et al. 2005). Indeed, although both synaptic and extrasynaptic GABA_A receptors of WT DGGCs are sensitive to etomidate, given the dominant influence (∼90%) of the tonic conductance, ∼99% of the anaesthetic-induced increase in the total charge transfer is mediated by extrasynaptic receptors. A similar differential effect of etomidate occurs in thalamic VB neurones (Belelli et al. 2005). The sedative effect of the anaesthetic is blunted in β_2N265S mice (Reynolds et al. 2003). However, this mutation did not influence the effect of the anaesthetic on DGGC synaptic GABA_A receptors, but considerably reduced (∼50%) the etomidate enhancement of the tonic conductance. Deletion of the δ subunit suppressed extrasynaptic receptor expression and therefore reduced (by ∼59%) the effect of the anaesthetic on the tonic conductance. In agreement with others (Wei et al. 2003), we found δ-GABA_A receptors to have little influence on DGGC mIPSCs (at physiological temperature). However, unexpectedly, etomidate was significantly more effective in prolonging the mIPSCs of δ^−/− DGGCs. By contrast, neurosteroid modulation of cerebellar granule cell (CGC) sIPSCs is reduced by δ subunit deletion, a scenario reversed by stimulation of PKC (Vicini et al. 2002). Whether the enhanced effect of etomidate on synaptic receptors is similarly a consequence of perturbations in phosphorylation, or alternatively, is due primarily to changes in the proportion of β₃ to β₁ synaptic GABA_A receptors is not known. Clearly, such plasticity warrants caution for interpretation of behavioural studies in δ^−/− mice.

The influence of the β isoform on receptor pharmacology is not restricted to general anaesthetics. Although the results of studies on the interaction of ethanol with δ-GABA_A receptors are controversial (Lovinger & Homanics, 2007), the effect of this alcohol on CGC extrasynaptic receptors is reportedly favoured by δ-GABA_A receptors incorporating β₃ compared with β₂ subunits (Hanchar et al. 2005). Furthermore, under conditions designed to mimic physiological intracellular chloride, neurosteroids reportedly inhibit GABA-evoked outward current responses mediated by α₄β₂δ receptors, but enhance equivalent responses mediated by α₄β₃δ receptors (Shen et al. 2007).

In conclusion, within the neuronal cell soma of DGGCs the β₂ subunit is selectively incorporated into a pool of benzodiazepine-insensitive extrasynaptic (α₄β₂δ) receptors, but is not resident in populations of benzodiazepine-sensitive extrasynaptic (α₅β_1/3γ₂) and synaptic receptors. To date our appreciation of the distinguishing properties of β subunit isoforms is somewhat rudimentary. However, it is already evident that the nature of the β subunit may influence receptor expression, biochemical control of receptor function and the pharmacology of the receptor. The categorical identification of GABA_A receptor isoforms within a neurone should aid future studies aimed at determining how these pools of distinct GABA_A receptors act in unison to influence DGGC excitability. Furthermore, such studies may be pertinent in aiding our understanding of certain neurological disorders, given that perturbations of DGGC GABA-ergic transmission have been reported in rodent models of Alzheimer/s disease, Down/s syndrome, anxiety, stress and epilepsy (Brooks-Kayal et al. 1998; Kleschevnikov et al. 2004; Maguire et al. 2005; Palop et al. 2007).