Permeation mechanism in voltage-activated proton channels: a new glimpse.

E ver since Hodgkin and Huxley’s work in 1952 first described the permeation changes for sodium and potassium in the squid giant axon (1), scientists in the membrane biophysics field have been fascinated as they unravel the structure and function relationship of ion channels. Decades of cleaver electrophysiology experiments (2), the first cloning of voltage-activated Na (Nav) and K (Kv) channels (3, 4) allowing the use of molecular biology techniques, and the recent crystal structures of Kv channels (5), have all led to the idea that a voltagegated ion channel is composed of separate domains: one ion-permeating pore domain and four voltage-sensing domains (VSDs). Most members in the super family of voltage-gated ion channels (i.e., Nav and Kv channels) are composed of four subunits or domains, each with six transmembrane segments. The first four transmembrane segments of each subunit (in Kv channels) or domain (in Nav channels) form a separate VSD, and the fifth and sixth transmembrane segments from all four subunits come together to form one common, centrally located pore region. However, in 2006, two laboratories published the first clones of voltageactivated proton channels (Hv) (6, 7). Hv channels have only four transmembrane segments, homologous to the VSD of Kv channels, and lack the classic pore domain of other voltage-gated ion channels. What, then, constitutes the proton permeation pathway in Hv channels? The work by Sakata et al. in this issue of PNAS furthers our understanding of how a single VSD can function both as a voltage-sensing domain and a proton permeation domain (8). Before the cloning of Hv channels, mutagenesis experiments had shown that replacing certain key residues in the fourth transmembrane domain S4 in the VSD of Kv channels could produce a proton or cation permeation pathway in the VSD of Kv channels, separate from the K permeation pathway through the pore domain of Kv channels (9, 10). This current has been called the “omega current” (10) to separate it from the “alpha current” generated by K ions flowing through the central pore domain of Kv channels. All voltage-gated cation channels have a motif in S4, with every third amino acid being a positively charged arginine or lysine (Fig. 1A). The first report of an omega current came from Bezanilla’s laboratory, showing that replacing either the first or the fourth arginine in S4 with a histidine produced a proton channel in the VSD of Kv channels (9). If they replaced the first arginines (R1) with a histidine, they produced a hyperpolarization-activated proton channel. If they replaced the fourth arginines (R4) with a histidine, they produced a depolarization-activated proton channel. Later work from the Isacoff’s laboratory showed that replacing the first arginine with a smaller amino acid could even produce a hyperpolarizationactivated potassium channel through the VSD of Kv channels (10). These results were interpreted to mean that there must be a narrow constriction in the VSD that normally prevents cation and proton flow through the VSD (Fig. 1B). Replacing the S4 residue located at this constriction with a histidine allows for protons to flow through this restricted space, or replacing this residue with a smaller side chain would allow for K to flow through this restriction in the VSD of Kv channels (Fig. 1B). To explain the opposite voltage dependence of the R1 and R4 mutations, it was assumed that the voltage-dependent S4 movement would move R1 into this restricted space at hyperpolarized potentials and R4 at depolarized potentials (Fig. 1B). Therefore, when Hv channels were shown to be composed of only a VSD, it was natural to assume that Hv channels conduct protons by a similar omega current mechanism. So what does the S4 of Hv channels look like? It turns out that S4 in Hv channels only have three positively charged arginines and that there is a neutral asparagine residue at the position equivalent to R4 in Kv channels (Fig. 1A). Therefore, it was logical to hypothesize that this asparagine (N210 in the mouse Hv channel and N214 in the human Hv channel) is at the constriction in the VSD at depolarized potentials, thus allowing protons to flow by this constriction at depolarized potentials. In support of this hypothesis, mutating N214 in the human Hv channel to a cysteine and modifying this cysteine with a large positively charged cysteine reagent abolished proton currents (11). In their present work in this issue of PNAS (8), Sakata et al. further tested this hypothesis experimentally by deleting different regions of S4 by truncating the protein at different position around and before N210 in S4 of the mouse Hv channel. Truncations that delete the region around N210 were expected to make the channels nonfunctional or, at least, nonselective. Truncations at I209, G211, and I213, unexpectedly, still generated channels that were proton-selective and voltage-dependent. Even a truncation between R2 and R3 at position A206 showed proton selective currents. A truncation at residue L200 completely abolished the function of mouse Hv channels, even though the number of channel proteins in the membrane was the same as for fulllength Hv channels. These results show that the region around N210 is not necessary for proton selectivity in Hv channels. What about the earlier results showing that putting a charge at N214 in the human Hv channel abolished the proton currents? Does this not show that N214 is necessary for the proton permeation? Sakata et al. (8) found that the mouse Hv channel is still a voltage-activated proton channel after replacing N210 with a positively charged amino acid (arginine or lysine). However, the current density was greatly reduced by this mutation. The authors conclude that N210 does not determine the proton selective. So why does the N210R mutation reduce the proton currents? One possibility is that the N210R mutation partly blocks the proton currents, but does not completely abolish the currents. Therefore, the results on the N214R mutation do not exclude a role for N210 in the permeation pathway, they only mean that N210 is not solely determining the proton selectivity of Hv channels. Finally, Sakata et al. determine which part of the S4 transmembrane segment is facing the aqueous environment and which part of S4 is inaccessible to the aqueous environment (8). Using a PEGylationprotection assay, Sakata et al. show that residues 192 to 205 in mouse Hv channels are inaccessible or partly inaccessible from the aqueous solution. In the A206 truncation, a similar region is inaccessible, suggesting that the S4 topology and insertion are similar in the A206 truncation