Kv4 channels represent the main class of brain A-type K+ channels that operate in the subthreshold range of membrane potentials (Serodio, P., E. Vega-Saenz de Miera, and B. Rudy. 1996. J. Neurophysiol. 75:2174- 2179), and their function depends critically on inactivation gating. A previous study suggested that the cytoplasmic NH2- and COOH-terminal domains of Kv4.1 channels act in concert to determine the fast phase of the complex time course of macroscopic inactivation (Jerng, H.H., and M. Covarrubias. 1997. Biophys. J. 72:163-174). To investigate the structural basis of slow inactivation gating of these channels, we examined internal residues that may affect the mutually exclusive relationship between inactivation and closed-state blockade by 4-aminopyridine (4-AP) (Campbell, D.L., Y. Qu, R.L. Rasmussen, and H.C. Strauss. 1993. J. Gen. Physiol. 101:603-626; Shieh, C.-C., and G.E. Kirsch. 1994. Biophys. J. 67:2316-2325). A double mutation V[404,406]I in the distal section of the S6 region of the protein drastically slowed channel inactivation and deactivation, and significantly reduced the blockade by 4-AP. In addition, recovery from inactivation was slightly faster, but the pore properties were not significantly affected. Consistent with a more stable open state and disrupted closed state inactivation, V[404,406]I also caused hyperpolarizing and depolarizing shifts of the peak conductance-voltage curve ( approximately 5 mV) and the prepulse inactivation curve (>10 mV), respectively. By contrast, the analogous mutations (V[556,558]I) in a K+ channel that undergoes N- and C-type inactivation (Kv1.4) did not affect macroscopic inactivation but dramatically slowed deactivation and recovery from inactivation, and eliminated open-channel blockade by 4-AP. Mutation of a Kv4-specific residue in the S4-S5 loop (C322S) of Kv4.1 also altered gating and 4-AP sensitivity in a manner that closely resembles the effects of V[404, 406]I. However, this mutant did not exhibit disrupted closed state inactivation. A kinetic model that assumes coupling between channel closing and inactivation at depolarized membrane potentials accounts for the results. We propose that components of the pore's internal vestibule control both closing and inactivation in Kv4 K+ channels.