1. Excitatory postsynaptic currents (EPSCs) were recorded in CA3 pyramidal cells of hippocampal slices of 15- to 24-day-old rats (22-degrees-C) using the whole-cell configuration of the patch clamp technique. 2. Composite EPSCs were evoked by extracellular stimulation of the mossy fibre tract. Using the selective blockers 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and D-2-amino-5-phosphonopentanoic acid (APV), a major alpha-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA)/kainate receptor-mediated component and a minor NMDA receptor-mediated component with slower time course were distinguished. For the AMPA/kainate receptor-mediated component, the peak current-voltage (I-V) relation was linear, with a reversal potential close to 0 mV. The half-maximal blocking concentration of CNQX was 353 nm. 3. Unitary EPSCs of the mossy fibre terminal (MF)-CA3 pyramidal cell synapse were evoked at membrane potentials of -70 to -90 mV by low-intensity extracellular stimulation of granule cell somata using fine-tipped pipettes. The EPSC peak amplitude as a function of stimulus intensity showed all-or-none behaviour. The region of low threshold was restricted to a few micrometres. This suggests that extracellular stimulation was focal, and that the stimulus-evoked EPSCs were unitary. 4. Latency and rise time histograms of EPSCs evoked by granule cell stimulation showed narrow unimodal distributions within each experiment. The mean latency was 4.2 +/- 1.0 ms, and the mean 20-80 % rise time was 0.6 +/-0.1 ms (23 cells). When fitted within the range 0.7 ms to 20 ms after the peak, the decay of the EPSCs with the fastest rise (rise time 0.5 ms or less) could be described by a single exponential function; the mean time constant was in the range 3.0-6.6 ms with a mean of 4.8 ms (8 cells). 5. Peak amplitudes of the EPSCs evoked by suprathreshold granule cell stimulation fluctuated between trials. The apparent EPSC peak conductance in normal extracellular solution (2 mm Ca2+, 1 mM Mg2+), excluding failures, was 1 ns. Reducing the Ca2+ concentration and increasing the Mg2+ concentration reduced the mean peak amplitude in a concentration-dependent manner. 6. Peaks in EPSC peak amplitude distributions were apparent in low Ca2+ and high Mg2+. Using the criteria of equidistance and the presence of peaks and dips in the autocorrelation function, five of nine EPSC peak amplitude distributions were judged to be quantal. From the likelihood ratio when fitting non-quantal and quantal model functions to the peak current data, the probability of wrongly rejecting the non-quantal models was estimated to be in the range < 0.001-9.4%. The apparent quantal conductance change was in the range 105-177 pS with a mean of 133 pS in different experiments. The coefficient of variation of a quantal event was estimated to be 22 7. Spontaneously occurring miniature EPSCs were recorded at negative membrane potentials in the presence of 1 muM tetrodotoxin (TTX). Miniature EPSC 20-80% rise times varied between 0.2 and > 10 ms within each experiment. Peak amplitude distributions of the miniature EPSCs with rise times less than 0.8 ms (presumably arising from MF-CA3 synapses) were skewed. Mode and mean values of these distributions corresponded to apparent conductances of 106 +/- 19 and 251 +/- 22 pS, respectively (6 cells). 8. Fast application of 1-3 mm glutamate to outside-out patches isolated from the somata of CA3 pyramidal cells activated currents which were mediated by AMPA/kainate receptor channels. The elementary conductance of these channels estimated from non-stationary fluctuation analysis was 8.5 +/-2.1 pS (9 patches), and the maximal open probability with 3 mm glutamate was 0.71 +/-0.06. Extracellular divalent cation concentrations had only small effects on the recorded glutamate-activated currents. 9. In the whole-cell recording configuration, responses to short current pulses and a biocytin fill were obtained from a CA3 pyramidal neurone. A compartmental model was made, based on the cell morphology as reconstructed using a light microscope. The electrical parameters of the model were adjusted until its short pulse response gave the best fit to the measured response of the neurone. This gave a specific membrane capacitance (C(m)) of 0.683 muF cm-2, a specific membrane resistance (R(m)) of 164 000 OMEGA cm2, and a cytoplasmic resistivity (R(i)) of 294 OMEGA cm, with zero somatic shunt conductance. 10. The most proximal and the most distal mossy fibre synaptic conductances were simulated, with the soma voltage clamped via different series resistances. Both the dendritic cable and the series resistance attenuated and slowed the EPSCs. With the plausible range of series resistances (1.5-10 MOMEGA), the apparent peak conductance was reduced to 0.32-0.87 of the 'real' value, the 20-80 % rise time was increased by a factor of 1-2-4.5 and the effective decay time constant by a factor of 1.1-2.6. 11. The results indicate that unitary EPSCs of MF-CA3 synapses show a rapid rise and a fast decay, and that they are quantal in nature, at least in a subset of MF-CA3 synapses. We estimate that a typical unitary EPSC of the MF-CA3 synapse at a 'physiological' concentration of divalent cations has a quantal content between 2 and 16: Considering voltage clamp errors, a quantal event appears to be generated by the simultaneous opening of between fourteen and sixty-five glutamate receptor channels of the AMPA/kainate subtype.
