Single or double? Think zinc!

The voltage-gated proton channels (Hv channels) are extremely selective and are tightly regulated by pH and voltage (DeCoursey, 2003). Recently identified molecular correlates, VSOP (Sasaki et al. 2006) or Hv1 (Ramsey et al. 2006) of the mammalian Hv channel, contain four-transmembrane proteins (S1–S4) with significant homology to the voltage-sensing domain shared by conventional voltage-gated ion channels and voltage-sensing phosphatase (Murata et al. 2005). Hv channels form dimers (Koch et al. 2008; Tombola et al. 2008; Lee et al. 2008) and each subunit has an independent proton-selective permeation pathway and intrinsic voltage sensor (Koch et al. 2008; Tombola et al. 2008). Recent studies of knock-out mice of the HVCN1 gene encoding VSOP/Hv1 protein have revealed that the Hv channel plays a critical role in multiple physiological functions.

Activities of the mammalian Hv channel are suppressed by zinc in the micromolar range. However, this inhibitory effect of zinc is not as a result of blocking ion permeation, but is rather by shifting the voltage threshold and slowing the speed of channel opening. These effects are much more notable than non-specific binding or screening effects by divalent cations, suggesting the existence of a molecular structure specifically recognizing zinc ions. A previous pioneering study of the electrophysiological characterization of native Hv channels (Cherny & DeCoursey, 1999) showed that this effect of zinc is strongly dependent on the extracellular pH, and that zinc competes with two or three protons with a pKa of 6–7. They predicted that high-affinity zinc binding requires two or more histidines. Later this was proven at the molecular level: mutagenesis experiments showed that loss of H140 and H193 made the Hv1 channel insensitive to zinc (Ramsey et al. 2006). However, detailed mechanisms by which zinc binds to Hv channels and affects their gating remained elusive.

In a recent issue of The Journal of Physiology, DeCoursey and his colleagues have tried to address where in dimeric Hv channels zinc binds, motivated by an homology model of the Hv channel structure, based on crystal structures of the voltage-sensor domain of voltage-gated potassium channels (Musset et al. 2010). This model revealed that the two externally accessible histidine residues, H140 and H193, were 14 Å apart, a distance too great for the binding site to be on the same subunit. They suspected that zinc must bind to two histidines from two individual subunits that form dimers. To test this, zinc sensitivity was compared between the dimer and monomer Hv1 channels heterologously expressed in mammalian cells. The C-terminus of the Hv channel contains a conserved coiled-coil motif and truncation of the whole C-terminus significantly reduces dimerization (Koch et al. 2008; Tombola et al. 2008). This C-terminus truncated (ΔC) Hv1 channel was used as the monomer channel and exhibited lower sensitivity to zinc than the wild-type (dimer) channels. However, similar observations would be obtained if zinc binds two sites on the same subunit, and the distance between two histidines or spatial profile of zinc binding could differ between monomer channels and dimer channels. To exclude this possibility, the authors used tandem dimer constructs in which two subunits were linked by a short polypeptide. This strategy enabled them to target each of four histidines in the dimers. If zinc binds two histidines on the same subunit, the tandem dimer consisting of the wild-type and H140A/H193A mutant should exhibit modification by zinc to some extent. The tandem dimer consisting of a histidine mutation on each unit (H193A-H140A) should exhibit low sensitivity to zinc. However, the results were inconsistent with this. The WT–H140A/H193A tandem dimer showed resistance to zinc, and the H193A–H140A tandem showed moderate sensitivity to zinc. Based on these results, the authors reached the conclusion that zinc binds to histidines from two subunits in a dimer.

What can we learn from these findings? Two recent papers have shown that gating becomes cooperative by dimerization (Tombola et al. 2010; Gonzalez et al. 2010), and the paper by Musset et al. (2010) adds additional evidence for this cooperativity. Q₁₀ of channel gating was shown to be significantly larger in dimer channels than monomer channels, consistent with the idea that dimer channels undergo more complicated changes of protein conformation during activation than monomer channels, although this still requires further study. The idea that zinc binds across two subunits raises a possibility that zinc may alter gating through modifying cooperative interaction between dimers.

Furthermore, the dimer stoichiometry of the Hv channel has been demonstrated only in an heterologous expression system. Zinc sensitivity of native Hv channels falls within a range (DeCoursey, 2003) seen with dimerized Hv channels in a heterologous expression system (Sasaki et al. 2006; Ramsey et al. 2006). The requirement of dimer organization for high-affinity binding of zinc confirms that the native Hv channel is a dimer. However, it should be noted that a significant population of Hv channels are present in intracellular membranes such as phagosomes (Okochi et al. 2009) at which zinc sensitivity cannot be accessed by conventional methods.

This study raises specific questions for future studies. First, the molecular model supported in this study indicates that the S1 and S4 domains are located at the outer region of the dimer complex. However, in the crosslinking study of the human proton channel, S1 was cross-linked between two subunits in dimers, suggesting that the S1domains are opposed next to each other in a dimer (Lee et al. 2008). Even if dynamic motion of the voltage-sensor domain is taken into consideration, this model requires the spatial relationship between two subunits in the dimer to be highly flexible. In addition, in the proposed molecular model, the C-terminal ends of S4 of the two subunits might not be close enough to position the downstream coiled-coil motif to form a pair along the full length of the C terminus.

Another question is how do the two accessible histidine residues, H140 and H193, play a role in zinc-dependent regulation of gating. They exhibited distinct changes in zinc modulation of kinetics and the I–V relation, suggesting that the two sites do not have an equivalent role. H140 is highly conserved in Hv channels from many species, suggesting that it has a primary role in gating. On the other hand, H193 is only conserved among avian and mammals but not in ascidian or teleost, suggesting that this site was created relatively late during chordate evolution. Zinc action on Hv channels is of physiological significance. Human sperm robustly expresses Hv channel (Lishko et al. 2010) and millimolar levels of zinc are contained in seminal fluid. Zinc-dependent regulation of the Hv channel in human sperm might underlie the processes of sperm maturation (capacitation).

Both in the structure–function relationship and physiological roles of mammalian Hv channels, zinc modulation of Hv channel gating will continue to be a hot topic.