The dynamics of synchronized neurotransmitter release determined from compound spontaneous IPSCs in rat dentate granule neurones in vitro.

1. The properties of GABAA receptor-mediated spontaneous IPSCs generated in hippocampal dentate granule neurones were analysed using whole-cell voltage-clamp techniques in order to explore the functional consequences of the low number (6-12) and close proximity of synaptic contacts made by single GABAergic interneurones. 2. Spontaneous IPSCs (sIPSCs) occurred with a frequency of 14.0 +/- 9.1 Hz (n = 31) and revealed a multi-modal positively skewed amplitude distribution (39.0 +/- 19.8 pA, median values). 3. The variance of 10-90% rise times and decay kinetics between IPSCs decreased with increasing peak amplitude. Larger amplitude events had significantly faster rise times, consistent with their site of generation being proximal to the soma. The decay kinetics of sIPSCs did not significantly change with amplitude. 4. Large amplitude sIPSCs occurred singularly or in discrete bursts, repeated regularly at low frequency. The rising phase of such sIPSCs were multi-phasic, composed of clear step-like inflections that were not a product of noise. The variability between the rising phase of individual sIPSCs was quantified by calculating their standard deviation, which produced fast rising (0.22 +/- 0.05 ms time to peak, n = 16) functions with half-widths of 0.38 +/- 0.10 ms, which declined to plateaux. 5. Computer simulations demonstrated that IPSCs with properties similar to those recorded experimentally could be generated by the linear summation of groups of temporally dispersed component events. Standard deviation functions of the rising phase of simulated IPSCs accurately described distributions of the temporal dispersion of unitary components. 6. The GABA uptake inhibitor (R)-N[4,4-bis(3-methyl-2-thienyl)but-3-enl-yl] nipecotic acid (tiagabine) (10 microM, n = 12) significantly prolonged the decay of mIPSCs (6.5 +/- 0.8 to 8.7 +/- 1.0 ms, median values) and sIPSCs (6.2 +/- 0.4 to 7.3 +/- 1.2 ms, median values), but failed to alter the frequency of occurrence, 10-90% rise times or peak amplitude of events. The application of flurazepam (30 microM, n = 7; 50 microM, n = 4) prolonged the decay of sIPSCs regardless of their amplitude. 7. These data indicate that sIPSCs are formed by the summation of unitary components that occur asynchronously and that GABA released from multiple sites has independent post-synaptic actions.