The dual role of calcium: Pore blocker and modulator of gating

Faced with the bewildering characteristics of the ionic currents that cause the action potential in squid axon, Hodgkin and Huxley in 1952 developed an elegant model (1) that remains one of the most insightful descriptions of the functional properties of voltage-gated ion channels. One ingredient in this conceptual wizardry was the strict separation between gating, the process responsible for activation and inactivation of sodium and potassium channels, and permeation, the nearly ohmic behavior of fully activated channels. This clean separation between gating and permeation has been a bedrock principle in the biophysical characterization of voltage-dependent ion channels and is supported by many single-channel studies in which the opening and closing of channels is distinct from the properties of open channels. The maxim is that the gates open and close channels and pay scant attention to the flow of ions through the open channel.

This dogma has some major exceptions, however, notably because of effects of permeant and pore-blocking ions on gating. Although such effects are quite variable among various classes of ion channels, the customary observation is that raising the concentration of either permeant or pore-blocking ions inhibits the gates from closing (2–15). The experimental data strongly suggest that, if an ion can bind deeply within the permeation pathway, it will tend to obstruct gate closure. This is the “foot-in-the-door” phenomenon originally described by Clay Armstrong to account for the effects of intracellular pore blockers on potassium channel gating (2, 3). The two papers from Armstrong’s laboratory in this issue of the Proceedings (16, 17) report precisely the opposite result. The binding of extracellular calcium within the pore of sodium channels has two consequences. Besides blocking current carried by sodium ions, it enhances the rate of closing of the activation gates. This raises two intriguing possibilities. First, the binding of extracellular calcium within the pore may be a necessary requirement for channels to close. A corollary of this is that the voltage dependence of calcium block may contribute to the voltage dependence of deactivation, the closing of activation gates. Second, the release of a calcium ion from the pore may be required for the activation gates to open.

This is a completely novel concept of calcium’s effects on the gating of sodium channels. Although the pore-blocking effects of extracellular calcium are well known, the effects on gating usually have been ascribed to neutralization of a negative surface potential (18, 19), either by screening or binding of the divalent cation (20). Reducing the negative surface potential should shift the voltage dependence of gating by increasing the electric field across the bilayer, thereby stabilizing sodium channels in their closed conformation. It originally was assumed that the negative surface potential, estimated to be ≈−60 mV in vertebrate cells, was caused by negatively charged phospholipids. More recent data suggest, however, that the charge originates primarily on the channel itself (21), either from negatively charged amino acids or from sialic acid residues.

An unfulfilled requirement of standard surface potential theories is that extracellular calcium must shift the voltage dependence of all gating parameters (e.g., activation, deactivation, and inactivation) equally. Many exceptions to this rule have been observed experimentally, beginning with the paper that introduced the surface potential hypothesis (18). To address this complication, modifications of the theory have included the possibility that calcium interacts with specific regions of the channel, such as the negatively charged vestibule near the voltage sensor of the sodium channel (22). The idea that the pore-blocking site is also the modulatory site for the shift of gating was introduced by Armstrong and Cota in 1991 (23). In this paper, they showed a strong correlation between the binding of calcium in the pore and the depolarizing shift of activation gating.

The two new papers from Armstrong’s laboratory both support and extend this idea. First, the rate of deactivation at −80 mV increases linearly with the fraction of channels blocked by calcium (16). This fraction was altered by changing extracellular calcium concentration. Remarkably, extrapolation of this relationship predicts that unblocked sodium channels cannot close; that is, the deactivation rate is zero in the absence of calcium. Unfortunately, a direct test of this hypothesis is not possible with the mammalian cells used in this study because the cells cannot survive the complete removal of extracellular divalent cations.

The second paper examines the effect of extracellular calcium on sodium currents of squid giant axon (17). This preparation has two advantages over the mammalian cells used in the first paper. The axon can tolerate complete removal of calcium, at least for brief periods, and it is possible to measure the movement of the voltage sensors directly as a gating current (24). Armstrong exploited the latter by blocking all ionic current with the pore blocker saxitoxin and examining the effects of extracellular calcium on gating current kinetics. Because saxitoxin prevents access of calcium to its blocking site, this experiment is a direct test of the hypothesis that calcium acts elsewhere, either on the channel itself or on the lipid bilayer, to modulate gating. Calcium not only had small effects on the kinetics of gating currents, but the effects were qualitatively inconsistent with any gate-shifting model. The prediction of such models is that increasing extracellular calcium concentration should decrease the activation (ON) kinetics of the gating current and increase the deactivation (OFF) kinetics. However, there was a small decrease of both rates. Removal of saxitoxin exposes the calcium-binding site in the pore and reveals a large effect of calcium on the deactivation kinetics of ionic currents, as observed for the sodium currents of mammalian cells (16). Clearly, the predominant effect of calcium on gating requires access to the pore. The binding of calcium at other sites can at best play a minor role on the gating of sodium channels.

Complete removal of extracellular calcium had two effects on the sodium currents of squid axon. Over a period of tens of seconds it produced a gradual decrease in the rate of deactivation, presumably because of the gradual loss of calcium from the glial layers surrounding the axon. The second effect was a concomitant disappearance of sodium current. This loss of sodium current was reversible on reintroduction of calcium. The loss of sodium currents when extracellular calcium is removed is, as noted by Armstrong (17), a disputed experimental observation. These data suggest, however, that calcium may be important in maintaining the functional integrity of sodium channels. It remains to be shown whether this preservation of sodium currents is caused by the binding of calcium within the pore or at a separate site.

Does pore block underlie all of the effects of divalent cations on sodium channel gating? Apparently not. The most infamous exception is zinc. Like calcium, extracellular zinc blocks open channels (25) and causes a depolarizing shift in gating (26). However, the genesis of the effect on gating is almost exactly the opposite of that of calcium. Whereas raising calcium concentration enhances the closing rate and has less effect on the opening rate, zinc dramatically decreases the opening rate and has almost no effect on closing (26). (Note that the latter, well documented result is not in accord with the conclusions in ref. 16.) Because these divalent cations appear to block at the same site in the pore (25), and because the pore blocker tetrodotoxin abolishes the action of zinc on gating (26), the disparity in their effects on gating remains a mystery.