NOISE-ANALYSIS OF MINIATURE IPSCS IN ADULT-RAT BRAIN-SLICES - PROPERTIES AND MODULATION OF SYNAPTIC GABA(A), RECEPTOR CHANNELS

The properties of synaptic gamma-aminobutyric acid (GABA), receptor channels were resolved by using tight-seal, whole-cell recordings from granule cells of the dentate gyrus in adult rat hippocampal slices and by applying the technique of nonstationary noise analysis to study miniature inhibitory postsynaptic currents (mIPSCs) recorded in the presence of tetrodotoxin (TTX), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), and D-2-amino-5-phosphonovaleric acid (D-APS). This technique allowed us to extract information about the conductance, the number, and the kinetics of ligand gated channels underlying elementary synaptic currents. To ascertain the validity of the nonstationary noise analysis method we have first tested it on computer simulated mIPSCs with different channel activation, lifetime kinetics, and opening probabilities. Using intraburst mean open times, shorter than the time to the first opening following activation, caused a large variance at the peak due to the stochastic channel properties. This resulted in a skewed mean current-variance relationship, which precluded proper estimation of unit conductance and especially the number of channels open at the peak of mIPSCs. Regardless of the probability of channel opening, accurate estimates of the unit conductance and the number of channels underlying each simulated mIPSC were obtained when channels had mean open times longer than the time to first opening. Once the validity of the nonstationary analysis had been ascertained, it was used on mIPSCs recorded at 35 degrees C. The unit conductance of the synaptic GABA(A) channels was 28 +/- 1 (SE)pS and the average number of channels underlying mIPSCs was 46 +/-. The mean current-variance relationship was not skewed at higher amplitudes, suggesting that the intrinsic variance at the peak of the GABA, mIPSCs is low and that the open time of the channels is longer than the time to first opening. The estimated unit conductance of the channels was constant over a wide range of holding potentials. The amplitude distribution of mIPSCs with rapid 10-90% rise times (290 +/- 20 mu s) was clearly skewed towards low values. This skew was not due to filtering of electrotonically distant currents. Current-variance analysis revealed that the skewness resulted from differences in the number of GABA, receptor channels and not from the heterogeneity of unitary conductances at various synapses. Selection of mIPSCs with slower rise times yielded smaller unit conductance estimates. Thus electrotonic filtering may lead to underestimation of unit conductance, or slower rising currents may arise from the activation of a different class of GABA(A) receptor channels. Spectral analysis of mIPSCs recorded at 35 degrees C revealed two main corner frequencies corresponding to time constants of 5.8 and 0.3 ms. Considering simple three state models, these time constants are consistent with intraburst mean open times >1.0 ms and a mean closed time of similar to 0.4 ms. These estimates of channel kinetics also explain why the intrinsic variance due to stochastic properties of the channels is low at the peak of mIPSCs. The unit conductance of GABA(A) receptor channels activated by exogenously applied GABA (500 mu M) was compared with that of receptor channels activated during mIPSCs. The estimated unit conductance from whole-cell stationary noise was 24 +/- 1 pS, an intermediate value between the conductance estimated from fast and slowly rising mIPSCs. Spectral analysis of whole-cell noise induced by exogenous GABA application revealed three major corner frequencies reflecting time constants of 6.6, 1.7, and 0.3 ms, different from those obtained from the analysis of mIPSCs. Thus exogenous GABA may activate different GABA, receptor channels from those at the synapse, or the method of GABA delivery may cause the channels to open differently. At 22 degrees C, the estimated unit conductance of the GABA(A) receptor channels underlying mIPSCs was 20 +/- 1 pS corresponding to a Q(10) of 1.2. Spectral analysis revealed three time constants of 22.0, 1.6, and 0.3 ms. Pentobarbital and the benzodiazepine zolpidem both prolonged the decay time constants of mIPSCs in a dose-dependent manner, with no change in the peak conductance of synaptic currents, the estimated unit conductance, or the number of underlying synaptic GABA, receptor channels. These findings are consistent with the hypothesis that most channels are open at the peak of mIPSCs, and that synaptic receptor channels are pharmacologically saturated by the GABA released into the cleft. Spectral analysis of mIPSCs modulated by zolpidem and pentobarbital at 35 degrees C showed a threefold increase in the two time constants of the spectrum which was also reflected in the threefold prolongation of mIPSC decay time constant. In conclusion, only a small number of synaptic GABA, receptor channels appears to be activated by saturating amounts of transmitter. Under these conditions, the kinetics and saturation of synaptic receptor channels permit only a prolongation but not an increase in amplitude of the mIPSCs. This may not be the case for excitatory synapses, but at GABA(A) synapses, increasing the amplitude of synaptic currents can be best achieved by increasing the number of available synaptic receptor channels rather than by altering channel kinetics or receptor affinity.