Voltage-dependent inactivating single-channel potassium currents were recorded in cell-attached and inside-out patches from embryonic Xenopus myocytes differentiating in culture. 2. Channels with rapid inactivation (time constants < 25 ms) and with slow inactivation (time constants > 80 ms) recorded after one day in vitro appear to belong to two functionally different classes. Rapidly and slowly inactivating channels show steady-state inactivation with potentials of half-inactivation of -74 +/- 7 and -44 +/- 9 mV. They exhibit voltage-dependent activation, with times to half-maximal activation of 0.79 +/- 0.09 and 1.17 +/- 0.22 ms when stepped from -120 to +40 mV. Rapidly inactivating channels also have a lower open probability than slowly inactivating ones. The channels have similar conductances of 23 +/- 6 and 17 +/- 4 pS and extrapolated reversal potentials close to the potassium equilibrium potential. 3. In cell-attached patches, inactivation behaviours of channels with rapid or slow inactivation do not change during recording. After patch excision, rapidly inactivating channels usually switch to a slow inactivation mode. Slowly inactivating channels derived from rapidly inactivating channels after patch excision retain their conductance and extrapolated reversal potential, but are not distinguishable from native slowly inactivating channels with respect to steady-state inactivation, activation and inactivation times, as well as open probabilities. 4. The change in inactivation behaviour of rapidly inactivating channels after patch excision is reversed by application of reduced dithiothreitol (DTT). In contrast, channels with slow inactivation in the cell-attached mode do not change in to rapidly inactivating channels after application of DTT in the excised configuration, suggesting that these channels belong to a structurally different class. 5. Frequent observation of superposing channel openings indicates clustering of inactivating potassium channels in the myocyte membrane, since many patches lack channel activity. Clustering does not depend on the presence of differentiating neurones. 6. Channels with rapid inactivation increase 6-fold in density during the first day in culture in the presence of neurones; channel density decreases in their absence. Channels with slow inactivation increase 2-fold in density in the presence or absence of differentiating neurones during this period. 7. Channels with rapid or slow inactivation in cell-attached membrane belong to functionally distinct classes that are developmentally regulated differently. Reversible changes from rapid to slow inactivation mode after patch excision suggest that the channels may be structurally related.
