SUMMARY1. Pinched-off cholinergic nerve terminals (synaptosomes) prepared from the electric organ of Torpedo ocelata were fused into large structures (> 20 ,tm) using dimethyl sulphoxide and polyethylene glycol 1500, as previously described for synaptic vesicles from the same organ.2. The giant fused synaptosomes were easily amenable to the patch clamp technique and 293 seals with a resistance greater than 4 GQ were obtained in the ' cell-attached' configuration. In a large fraction of the experiments, an 'inside-out' patch configuration was achieved.3. Several types of unitary ionic currents were observed. This study describes the most frequently observed single-channel activity which was found in 247 out of the 293 membrane patches (84-3 %).4. The single-channel current-voltage relation was linear between -60 and 20 mV and showed a slope conductance of 23-8 + 1-3 pS when the pipette contained 350-390 mM-Na+ and the bath facing the inside of the synaptosomal membrane contained 390 mM-K+.5. From extrapolated reversal potential measurements, it was concluded that this channel has a large selectivity for K+ over Na+ (70 4 + 11-5, mean + S.E.M.). Chloride ions are not transported significantly through this potassium channel.6. This potassium channel has a low probability of opening. The probability of being in the open state increases upon depolarization and reaches about 1 % when the inside of the patch is 20 mV positive compared to the pipette side.7.The mean channel open time increases with depolarization; thus the product current x time (= charge) also increases upon depolarization, showing properties of an outward rectifier.8. The potassium channel in the giant synaptosome membrane has a bursting behaviour. Open-time distribution, closed-time distribution and a Poisson analysis indicate that the minimal kinetic scheme requires one open state and three closed states.

1. Periodic oscillations were detected in the activity of single macromolecules: potassium channels. 2. When potassium channels are repeatedly activated in isolated patches from fused synaptosomes of Torpedo electric organ, their behavior exhibits a departure from random activation. 3. The departure from random behavior is demonstrated by the runs test and by the lack of fit to Poisson distribution. 4. Under appropriate experimental conditions, the channels display periodic oscillations with a periodicity of approximately 20 s when activated at a rate of 1.25 Hz. 5. The oscillations do not arise from sampling, recording, or computational artifacts. 6. It is conceivable that single-channel oscillations play a role in the generation of membrane oscillations and thus may contribute to the oscillatory behavior of the nervous system.

1. The statistical properties of the bursting potassium channel from fused Torpedo synaptosomes were studied by using patch-clamp recording and time series analysis. 2. Voltage steps produce channel openings; the number of channels opening fluctuates from trial to trial. The maximal current observed in each trial is strongly dependent on the previous history of the membrane patch. Trials with no activity are frequently clumped together and so are trials with intense activity. 3. Autocorrelation analysis reveals a strong interdependence of successive responses, which is voltage dependent. 4. We propose that the strong statistical interdependence of responses to successive depolarizing pulses (the statistical "memory") is a manifestation of a long lived closed state. We speculate that this statistical memory may be of significance in frequency modulation of transmitter release.

SUMMARY1. The voltage dependence of the bursting potassium channel in fused synaptosomes from Torpedo electric organ was studied in vitro, using the inside-out and the cell-attached configurations of the patch clamp technique.2. The patch of membrane was held at various holding potentials (-140 to -50 mV) and then stepped to test potentials (-50 to +40 mV) for periods ranging from 5 to 300 ms. Each potential step was repeated 200-600 times. After subtraction of the capacitative transients and the leakage currents, an ensemble-averaged current was obtained. This ensemble current showed a marked activation upon depolarization, followed by an inactivation.3. The activation of the bursting potassium channel is markedly dependent on the voltage step. Activation was detected at voltages positive to -50 mV. The peak of the ensemble current increases with the degree of depolarization, while the time to the peak decreases. With progressively larger depolarization, there is a shortening in the delay between the onset of the voltage step and the opening of the bursting potassium channels.4. The inactivation phase of the ensemble current could be described adequately in most of the experiments, as a single exponential decay to a steady-state inactivation level. The time constant of inactivation was not markedly voltage dependent. by 5 ms pulses produces a progressive decline in the peak ensemble current amplitude. The decline is larger at higher stimulation frequencies.9. The voltage-and time-dependent activation and inactivation properties of the bursting potassium channel make it a possible candidate for participating in frequency modulation of transmitter release and thus of synaptic transmission. We propose the potassium inactivation hypothesis for frequency modulation, which states that the inactivation of the potassium channel by previous stimulation could cause a broadening of the subsequent action potential and hence augmentation of calcium entry and transmitter release.

No abstract