1. The patch clamp technique was used to investigate the K+ channels in the membranes of nerve terminals in thin slices prepared from the rat posterior pituitary. 2. Depolarization of the membrane produced a high density of K+ current. With a holding potential of -80 mV, test pulses to +50 mV activated a K+ current which was inactivated by 65% within 200 ms. Hyperpolarizing prepulses enhanced the transient K+ current, with half-maximal enhancement at -87 mV. Depolarizing prepulses reduced or eliminated the transient K+ current. 3. In cell-attached patches formed with pipettes containing 130 mm KCl, three types of K+ channel could be distinguished on the basis of single-channel properties. One channel had a conductance of 33 pS and was inactivated with a time constant of 18 ms. A second channel had a conductance of 134 pS and was inactivated with a time constant of 71 ms. A third channel had a conductance of 27 pS, was activated relatively slowly with a time constant of 65 ms, and was not inactivated during test pulses of up to one second in duration. 4. Inactivation of the whole-cell K+ current was a biphasic process with two exponential components. The fast component had a time constant of 22 ms (at +50 mV), corresponding well with the time constant of decay of average current in cell-attached patches containing only the rapidly inactivating K+ channel. The slow component of inactivation had a time constant of 104 ms (at +50 mV), which was similar to but slightly slower than the time constant of decay of the average current in cell-attached patches containing only the slowly inactivating K+ channel. Inactivation of the slow transient K+ current became more rapid with increasing depolarization. 5. The low-conductance rapidly inactivating K+ channel had a lower voltage threshold for activation than the other two K+ channels. 6. Both inactivating K+ channels were enhanced in a similar manner by prior hyperpolarization. There was no difference with regard to voltage mid-point or steepness. 7. The large-conductance slowly inactivating K+ channel was activated by Ca2+ at the inner membrane surface. The resting intracellular Ca2+ was sufficiently high to produce significant activation of this channel without depolarization-induced Ca2+ entry. 8. Removal of Ca2+ from the bathing solution produced a -10 mV shift in the voltage dependence of enhancement of both transient K+ currents by prior hyperpolarization. This could be explained as a surface charge effect. 9. Tetraethylammonium blocked all three K+ channels, but was most effective in blocking the rapidly inactivating low-conductance K+ channel. Charybdotoxin did not reduce K+ current in the preparation. Dendrotoxin selectively blocked the low-conductance, non-inactivating K+ channel. 10. The low-conductance, rapidly inactivating K+ channel has many properties that are similar to transient K+ channels described in nerve cell bodies. The low-conductance non-inactivating K+ channel does not resemble channels previously described in nerve cell bodies, including the hypothalamic cell bodies from which these nerve endings originate. This K+ channel may be a specific axonal or nerve terminal membrane component. 11. The complex voltage, time and Ca2+ dependence of these K+ channels is consistent with a function in the regulation of secretion from the posterior pituitary by frequency and pattern of electrical activity.
