DENDRITIC EXCITABILITY AND A VOLTAGE-GATED CALCIUM CURRENT IN LOCUST NONSPIKING LOCAL INTERNEURONS

1. The active properties of axonless nonspiking interneurons in the thoracic nervous system of the locust Schistocerca americana were studied in vivo with the switched current-clamp technique from dendritic impalements, and in vitro with the whole-cell variation of the patch-clamp technique. 2. In 20% of in vivo recordings, depolarization of a dendrite to potentials more positive than about -40 mV evoked resonant behaviour and / or regenerative potentials. The latter were slow (half width: 20-30 ms), small (base-to-peak amplitude: 25-35 mV), and were often followed by a pronounced after hyperpolarization (AHP). 3. The slow regenerative potentials sometimes had multiple peaks separated by incomplete repolarizations. The voltage envelope of such potentials was always broader than that of spikes with single peaks. In other recordings, a same depolarizing pulse could evoke several regenerative potentials with different waveforms. These results suggested the presence of multiple dendritic initiation sites separated by regions of inexcitable membrane, allowing decremental conduction and the passive fusion of spike envelopes. 4. Graded active responses could also be evoked on rebound from short hyperpolarizations such as inhibitory postsynaptic potentials (IPSPs) provided that the membrane was already depolarized to about -40 mV. IPSPs evoked by several presynaptic interneurons differed in their ability to evoke rebound potentials suggesting that some synaptic sites were electrically closer than others to regions of active membrane. 5. Patch-clamp recordings from somata of nonspiking neurons isolated from 75% embryos and grown in culture medium for 1-2 days revealed the presence of an inactivating inward current resistant to 0.5-1 muM tetrodotoxin (TTX). The inward current was carried equally well by Ba2+, and sensitive to blockade by Cd2+ (0.5 mM), Ni2+ (0.75 mM), or Co2+ (2.5 mM). 6. The current activated around -40 mV, with voltage-dependent-activation (time-to-peak approximately 20 ms at -35 mV and 1-2 ms at 0 mV). Tail currents evoked upon repolarization were well fitted by a single exponential (tau = 1-2 ms). Deactivation time constants shorter than 300 mus, however, could not be measured. 7. The current inactivated rapidly in a voltage-dependent manner, following two-exponential kinetics. A very small persistent component could be explained by the overlap between activation and inactivation curves, greatest at approximately -20 mV. The voltage of half-inactivation was about -25 mV. At a resting potential of -58 mV, 90% of the current was available for activation. Recovery from steady-state inactivation followed the sum of two or more exponential processes. Analysis of the two inactivating components of the current in steps to different voltages or during the process of recovery did not allow the separation of currents with different activation thresholds. 8. We conclude that in addition to expressing voltage-gated K+ channels the dendrites of nonspiking interneurons express at least one Ca2+ current underlying active dendritic properties. The patch-clamp data closely match those obtained in vivo and explain the rising phase of the active potentials but not their repolarization. The current-clamp recordings suggest that the dendritic expression of the channels is heterogeneous and that active responses may be generated in multiple loci. The possible integrative functions of the current are discussed.