Single-channel properties of synaptic and extrasynaptic GABAA receptors suggest differential targeting of receptor subtypes

Abstract

Many neurons express a multiplicity of GABAA receptor subunit isoforms. Despite having only a single source of inhibitory input, the cerebellar granule cell displays, at various stages of development, more than 10 different GABAA subunit types. This subunit diversity would be expected to result in significant receptor heterogeneity, yet the functional consequences of such heterogeneity remain poorly understood. Here we have used single-channel properties to characterize GABAA receptor types in the synaptic and extrasynaptic membrane of granule cells. In the presence of high concentrations of GABA, which induced receptor desensitization, extrasynaptic receptors in outside-out patches from the soma entered long-lived closed states interrupted by infrequent clusters of openings. Each cluster of openings, which is assumed to result from the repeated activation of a single channel, was to one of three main conductance states (28, 17, or 12 pS), the relative frequency of which differed between patches. Such behavior indicates the presence of at least three different receptor types. This heterogeneity was not replicated by individual recombinant receptors (alpha1beta2gamma2S or alpha1beta3gamma2S), which gave rise to clusters of a single type only. By contrast, the conductance of synaptic receptors, determined by fluctuation analysis of the synaptic current or direct resolution of channel events, was remarkably uniform and similar to the highest conductance value seen in extrasynaptic patches. These results suggest that granule cells express multiple GABAA receptor types, but only those with a high conductance, most likely containing a gamma subunit, are activated at the synapse.

Fig. 1.

Activation of extrasynaptic GABA_Areceptors by a high concentration of GABA evokes clusters of channel openings. A, Application of 50 μm GABA (solid bar) to an outside-out patch from the somatic membrane of a P7 internal granule cell caused a large inward current that rapidly desensitized in the continued presence of agonist (holding potential −60 mV). B, Discrete single-channel clusters were separated by prolonged closed periods. C, Expansion of trace in B (open bar) showing two types of single-channel clusters that differ in their main conductances. D, Further expansion oftrace in C (open bars) showing more clearly the conductance difference between the two clusters and highlighting an apparent difference in kinetic behavior. For display, records were filtered at 1 kHz.

Fig. 2.

Patches from somatic membrane display different proportions of channel openings to three main conductance states.A, Current–voltage relationships from channels in three different patches. The left-hand panel illustrates data from a patch that exhibited clusters of one type, giving a slope conductance of 28 pS. The middle panel shows data from a patch in which clusters opened to one of two current levels, giving slope conductances of 28 and 17 pS. The right-hand panelshows data from a third patch in which clusters exhibited openings to one of three current levels, giving slope conductances of 26, 18, and 12 pS. B, The distribution of different cluster types between patches. High-conductance clusters (●) were recorded in all 13 patches and had a mean single-channel conductance of 27.6 pS, shown by the solid line (dotted lines indicate ±1 SEM). Ten of 13 patches also contained clusters of openings that had a mid conductance (■; 16.9 pS). Four patches displayed clusters of openings that had a low conductance (▴; 11.6 pS). Filled bars indicate the four different combinations of cluster types recorded in different patches and their relative frequency (as a percentage).

Fig. 3.

Extrasynaptic receptors exhibit channel clusters that fall into distinct classes with different main conductance states and open probabilities. A, Examples of high-, mid-, and low-conductance channel clusters recorded from a single somatic patch from a P7 internal granule cell (−60 mV; for display, records were filtered at 1 kHz). The integral is plotted below each current record. Shown to the right of each cluster is its corresponding all-point amplitude histogram with the calculated chord conductance and P_o (see Materials and Methods). B, The distribution of all chord-conductance estimates from 310 individual clusters recorded from a total of 21 patches. The histogram was fitted with the sum of three Gaussian distributions, with peaks at 12, 17, and 28 pS (compare with the slope conductance estimates in Fig. 2). The symbolssuperimposed on the distribution indicate the conductance for each of the three clusters shown in A. C, The relationship between cluster main conductance and clusterP_o. The distinction between low-conductance (▴), mid-conductance (■), and high-conductance clusters (●) is based on the distribution in B and calculated by the method of minimum misclassification (Colquhoun and Sigworth, 1995).Superimposed bars indicate the mean and SD of main conductance and P_o for each cluster type.

Fig. 4.

Recombinant α₁β₂γ_2S receptors exhibit a single type of channel cluster. A, Clusters of channel openings recorded in an outside-out patch from an HEK-293 cell transiently transfected with α₁, β₂, and γ_2S subunits in the continued presence of 50 μm GABA (−60 mV). All clusters contain high-conductance openings. B, An individual cluster recorded from the same patch. Inset shows an expanded view of part of the cluster (solid bar) illustrating the presence of clear sub-level transitions (for display, records were filtered at 1 kHz). The dotted linesindicate the closed, sub, and main conductance states, as determined from fits to an all-point histogram constructed for this patch (data not shown). C, The distribution of all chord-conductance estimates (main conductance) from 190 individual clusters recorded from a total of 13 patches. Most of the clusters are of the high-conductance type such that a single Gaussian adequately describes this distribution. Five cluster measurements lay outside the single Gaussian distribution. In three of these, high-conductance openings were present within the clusters, but most of the openings were to the subconductance state. D, The relationship between the cluster main conductance and the cluster P_osuggests the existence of a single main population of cluster types for this recombinant GABA_A receptor.

Fig. 5.

Spontaneous IPSCs display a rapid rise and slow, multi-exponential decay. A, Continuous record showing spontaneous IPSCs recorded from a P7 internal granule cell in the presence of 5 μm CNQX, 10 μm AP5, and 200 nm strychnine (−70 mV; 2 kHz filtering). B, The initial phase of some IPSCs shows obvious inflections caused by superimposition of events (asterisks; 5 kHz filtering).C, Most IPSCs exhibit monotonic rises with no obvious inflections. Only IPSCs that exhibited such a monotonic rise were included in any further analysis. D, The distribution of the resulting 10–90% rise times is well described by a single Gaussian. The inset illustrates the lack of relationship between rise time and peak amplitude for IPSCs recorded in one cell.E, An example of an averaged IPSC with its decay fitted by the sum (solid line) of three exponentials (dotted lines). The inset shows the average waveform (dots) and single (1), double (2), or triple (3) exponential fits (solid lines) displayed with current on a log scale; only the triple exponential fit adequately describes the current decay (2 kHz filtering).

Fig. 6.

Peak-scaled nonstationary fluctuation analysis of IPSCs reveals uniform weighted-mean single-channel conductance of synaptic receptors. A, The average waveform of 74 IPSCs (solid line) is shown scaled to the peak of an individual IPSC (gray line; 2 kHz filtering). Thehorizontal dotted lines represent the 30 amplitude bins used to determine values of mean current and variance.B, A plot of mean current against variance for the cell shown in A. The plot is fitted with a parabolic relationship to give the weighted-mean single-channel current (i; also given as conductance, γ) and the number of channels open at the peak (N_p). Thedotted line indicates the baseline variance.C, Histogram of weighted-mean single-channel current obtained from a total of 10 cells illustrating the uniform distribution of conductance. D, A plot of normalized current against normalized variance with data pooled from all 10 cells (error bars represent SEM).

Fig. 7.

Unitary current steps can be resolved in the decay of spontaneous IPSCs. A, Successive enlargements (solid bars) of the decay phase of a spontaneous IPSC recorded from a P7 internal granule cell (−70 mV). The second enlargement clearly shows the stepped nature of the IPSC decay (theequally spaced dashed lines illustrate the approximate amplitude levels of the steps). B, Four representative IPSCs showing the consistent presence of a channel-like step at the end of each IPSC. The peaks of the IPSCs have been truncated for display, and the dotted line indicates the pre-event baseline current. C, Two selected IPSCs, representative of the smallest events recorded. In each case the peak amplitude is an integer multiple (5 and 7) of the last step amplitude (dotted lines). For display, records were filtered at 1 kHz.

Fig. 8.

Analysis of resolvable current steps in IPSCs reveals a uniform single-channel conductance and a rare lower conductance state. A, Quantitative analysis of a stepped IPSC waveform recorded from a P7 internal granule cell (1 kHz filtering). The IPSC has five steps in its decay, seen as clearly resolved peaks in its all-point amplitude histogram (inset). The positions of the baseline (C) and the numbered dashed lines(1–5) are taken from the fit of six Gaussian components to the amplitude histogram. B, Plots of the inter-peak step size for the last two steps (S1 andS2) calculated for 50 consecutive IPSCs recorded in the same cell as A. The bottom panelillustrates the consistency of the step size for the last five channel closures (S1–S5) recorded from a total of eight cells. The mean step sizes (open symbols) and the SD are plotted for each step. The dotted line shows the mean value for S1. C, A histogram of IPSC step size (calculated as a single value for each of 234 IPSCs recorded from 8 cells) fitted with a single Gaussian. D, A single IPSC in which different sized steps were observed in the decay (1 kHz filtering). The inset was obtained after smoothing of the trace (see Materials and Methods) to better illustrate the steps in the waveform. The dotted lines superimposed on the trace show current levels obtained from peaks of the fit to the all-point histogram. C indicates the final closed state, and the dashed lines 1–3 represent the values of the three open levels observed in the decay. The final step size was calculated from the inter-peak interval between C and2, and the preceding smaller step size was calculated between C and 1.