Ion selectivity of voltage-activated sodium channels is determined by amino-acid residues in the pore regions of all four homologous repeats. The major determinants are the residues DEKA (for repeats I-IV) which form a putative ring structure in the pore; the homologous structure in Ca-channels consists of EEEE. By combining site-directed mutagenesis of a non-inactivating form of the rat brain sodium channel II with electrophysiological methods, we attempted to quantify the importance of charge, size, and side-chain position of the amino-acid residues within this ring structure on channel properties such as monovalent cation selectivity, single-channel conductance, permeation and selectivity of divalent cations, and channel block by extracellular Ca2+ and tetrodotoxin (TTX). In all mutant channels tested, even those with the same net charge in the ring structure as the wild type, the selectivity for Na+ and Li+ over K+, Rb+, Cs+, and NH4+ was significantly reduced. The changes in charge did not correlate in a simple fashion with the single-channel conductances. Permeation of divalent ions (Ca2+, Ba2+, Sr2+, Mg2+, Mn2+) was introduced by some of the mutations. The IC50 values for the Ca2- block of Na+ currents decreased exponentially with increasing net negative charge of the selectivity ring. The sensitivity towards channel block by TTX was reduced in all investigated mutants. Mutations in repeat IV are an exception as they caused smaller effects on all investigated channel properties compared with the other repeats.

Shaker potassium channels undergo a slow C-type inactivation which can be hastened dramatically by single-point mutations in or near the pore region. We found that the oxidizing agent chloramine-T (Chl-T) causes an irreversible loss of current for those mutants which show C-type inactivation. For several mutants at position T449, which show a wide spectrum of inactivation time constants, the time constant of current rundown induced by Chl-T correlated with the speed of inactivation. Rundown was accelerated when the channels were in the inactivated state but rundown also occurred when channels were not opened or inactivated. Apparently, only those channels which can undergo C-type inactivation are accessible to Chl-T. In order to gain information about the target amino-acid residue for the action of Chl-T and the structural rearrangements occurring during C-type inactivation, several mutant channel proteins were compared with respect to their response to Chl-T. Since Chl-T can oxidize cysteine and methionine residues, we mutated the possible targets in and close to the pore region, namely C462 to A, and M440 and M448 to I. While the residues M440 and C462 were not important for channel rundown, mutation of M448 to I made the channels more resistant to Chl-T by about one order of magnitude. While inactivation was accelerated upon application of Chl-T in most mutants, mutation of M448 to I abolished this effect on the time course of inactivation, indicating that M448 is one of the target residues for Chl-T.

This study tests the hypothesis that ion channels will be affected differently by external (extracellular) versus internal (cytoplasmic) exposure to hyperosmolar media. We looked first for effects on inactivation kinetics in wild-type Shaker B potassium channels. Although external hyperosmolar exposure did not alter the inactivation rate, internal exposure slowed both onset and recovery from fast inactivation. Differential effects on activation kinetics were then characterized by using a noninactivating Shaker B mutant. External hyperosmolar exposure slowed the late rising phase of macroscopic current without affecting the initial delay or early rising phase kinetics. By contrast, internal exposure slowed the initial steps in channel activation with only minimal changes in the later part of the rising phase. Neither external nor internal hyperosmolar exposure affected tail current rates in these noninactivating channels. Additionally, suppression of peak macroscopic current was approximately twofold smaller during external, as compared with internal, hyperosmolar exposure. Single-channel currents, observed under identical experimental conditions, showed a differential suppression equivalent to that seen in macroscopic currents. Apparently, during unilateral hyperosmolar exposure, changes in macroscopic peak current arise primarily from changes in single-channel conductance rather than from changes in equilibrium channel gating. We conclude that unilateral hyperosmolar exposure can provide information concerning the potential structural localization of functional components within ion-channel molecules.

Shaker potassium channels undergo a slow C-type inactivation which can be hastened dramatically by single-point mutations in or near the pore region. We found that the oxidizing agent chloramine-T (Chl-T) causes an irreversible loss of current for those mutants which show C-type inactivation. For several mutants at position T449, which show a wide spectrum of inactivation time constants, the time constant of current rundown induced by Chl-T correlated with the speed of inactivation. Rundown was accelerated when the channels were in the inactivated state but rundown also occurred when channels were not opened or inactivated. Apparently, only those channels which can undergo C-type inactivation are accessible to Chl-T. In order to gain information about the target amino-acid residue for the action of Chl-T and the structural rearrangements occurring during C-type inactivation, several mutant channel proteins were compared with respect to their response to Chl-T. Since Chl-T can oxidize cysteine and methionine residues, we mutated the possible targets in and close to the pore region, namely C462 to A, and M440 and M448 to I. While the residues M440 and C462 were not important for channel rundown, mutation of M448 to I made the channels more resistant to Chl-T by about one order of magnitude. While inactivation was accelerated upon application of Chl-T in most mutants, mutation of M448 to I abolished this effect on the time course of inactivation, indicating that M448 is one of the target residues for Chl-T.

1. Defolliculated Xenopus oocytes were voltage clamped in bathing solutions containing 115 mM KCl and 1.8 mM CaCl2. External application of H2O2 transiently elicited voltage‐dependent outward rectifying currents within several seconds. Upon depolarization to +50 mV these currents had an activation time constant of 370 ms and reached amplitudes of up to 70 microA. This current was also observed in oocytes without the vitelline membrane. 2. The current was abolished by 500 microM niflumic acid, by the replacement of external Cl‐ by methanesulphonate, or when extracellular Ca2+ was removed indicating the involvement of Ca2+‐activated Cl‐ channels, which are very abundant in Xenopus oocytes. 3. While the current could be recorded in bathing solutions containing Li+, K+, Rb+, Cs+ and NH4+, extracellular Na+ abolished the current completely (IC50 = 6 mM Na+). 4. The H2O2‐induced Cl‐ current was half‐maximally blocked by approximately 25 microM 2'4'‐dichlorobenzamil, 250 microM MgCl2, 100 microM CdCl2 and 100 microM NiCl2. These substances have been shown to block Na+‐Ca2+ exchangers in various tissues. 5. The data are consistent with the existence of an endogenous Na+‐Ca2+ exchanger in the plasma membrane of Xenopus oocytes, which runs in reverse mode in the absence of high external Na+ and the presence of external Ca2+. This endogenous component has to be considered when Xenopus oocytes are used for heterologous expression studies.