We report here that in cerebellar Purkinje cells from which the axon has been removed, positive voltage steps applied to the voltage-clamped soma produce spikes of active current. The spikes are inward, are all-or-none, have a duration of -1 ms, and are reversibly eliminated by tetrodotoxin, a Na channel poison. From cell to cell, the amplitude of the spikes ranges from 4 to 20 nA. Spike latency decreases as the depolarizing step is made larger. These spikes clearly arise at a site where the voltage is not controlled, remote from the soma. From these facts we conclude that Purk' je cell dendrites contain a sufficient density of Na channels to generate action potentials. Activation by either parallel fiber or climbing fiber synapses produces similar spikes, suggesting that normal input elicits Na action potentials in the dendrites. These findings greatly alter current views of how dendrites in these cells respond to synaptic input. (2)(3)(4)(5). In experiments on isolated neocortical pyramidal cells from newborn rats, differential microperfusion was used to show the presence of Na channels in dendritic stumps that stretched as far as 100 ,um from the soma (6).For neurons in situ, the distribution of spike-generating ion channels has been inferred primarily from intradendritic recordings (3-5). Such measurements suggested that Purkinje cell dendrites could generate Ca action potentials but were devoid of Na channels (2, 3). Dendrites of hippocampal pyramidal cells have action potentials produced by both Na channels and Ca channels (4, 5). More recently, microfluorometric imaging techniques have also provided some evidence regarding channel distribution (7-9).The recent application of whole-cell patch-clamp techniques to brain slices holds promise for helping with these questions. Here we have used the voltage clamp in an unconventional way to show the presence of Na spikes in the dendrites of cerebellar Purkinje cells.
MATERIALS AND METHODSExperiments were done on 200-,m sagittal slices of cerebellar vermis from 22-to 55-day rats, by using methods similar to those described (10, 11). In brief, the cerebellum was quickly removed from the animal and placed in ice-cold saline for 1-2 min. A Vibratome was used to cut slices, which were then allowed to incubate for 1-5 hr at 350C. A slice was placed in a simple chamber on a Zeiss Axioskop and viewed with a 40x water immersion objective. We elected to view the preparation with bright-field rather than Nomarski optics; it was possible to resolve individual Purkinje cells, their dendrites, and, in many cases, the axon. Cells near the surface of the slice were selected for recording and were gently cleaned as described by Edwards et al. (10).In cases where axotomy was done, a well-defined procedure was adopted to ensure removal of the axon. Two investigators carefully inspected the cells at various stages throughout the procedure to minimize the possibility oferror. We selected a Purkinje cell with a clearly visible axon. Pressure was applied to the cleaning pipet...