Engineered high-affinity zinc binding site reveals gating configurations of a human proton channel

Abstract

The voltage-gated proton channel (HV1) is a voltage sensor that also conducts protons. The singular ability of protons to penetrate proteins complicates distinguishing closed and open channels. When we replaced valine with histidine at position 116 in the external vestibule of hHV1, current was potently inhibited by externally applied Zn2+ in a construct lacking the two His that bind Zn2+ in WT channels. High-affinity binding with profound effects at 10 nM Zn2+ at pHo 7 suggests additional groups contribute. We hypothesized that Asp185, which faces position 116 in our closed-state model, contributes to Zn2+ chelation. Confirming this prediction, V116H/D185N abolished Zn2+ binding. Studied in a C-terminal truncated monomeric construct, V116H channels activated rapidly. Anomalously, Zn2+ slowed activation, producing a time constant independent of both voltage and Zn2+ concentration. We hypothesized that slow turn-on of H+ current in the presence of Zn2+ reflects the rate of Zn2+ unbinding from the channel, analogous to drug-receptor dissociation reactions. This behavior in turn suggests that the affinity for Zn2+ is greater in the closed state of hHV1. Supporting this hypothesis, pulse pairs revealed a rapid component of activation whose amplitude decreased after longer intervals at negative voltages as closed channels bound Zn2+. The lower affinity of Zn2+ in open channels is consistent with the idea that structural rearrangements within the transmembrane region bring Arg205 near position 116, electrostatically expelling Zn2+. This phenomenon provides direct evidence that Asp185 opposes position 116 in closed channels and that Arg205 moves between them when the channel opens.

Figure 1.

Location of key amino acids in an hH_V1 dimer model. Homology model of the closed hH_V1 dimer (Li et al., 2015) is shown in side view (left) and top view (right), where top corresponds to the extracytoplasmic side. Transmembrane helices are shown as pipes; connecting loops are shown as ribbons. The S1, S2, S3, and S4 helices are indicated in red, yellow, green, and blue, respectively. The hydrophobic gasket residues V109 (on S1), F150 (on S2), and V177 and V178 (on S3) are indicated in brown ball-and-sticks. Note that for clarity the side view is truncated above the strongest coiled-coil interaction of the C termini of the protomers.

Figure 2.

Zn²⁺ binds very weakly to hH_V1 with His¹⁴⁰ and His¹⁹³ removed, but with extraordinarily high affinity when His is introduced at position 116 (V116H). (A) The H140A/H193A background for dimer studies lacks the two His critical for Zn²⁺ binding and is insensitive to Zn²⁺ up to 100 µM. Families of currents in 10-mV increments up to +40 mV are shown for the indicated concentrations of Zn²⁺, all at pH_o 7, pH_i 6.5. (B) Extreme Zn²⁺ sensitivity of V116H mutant in the H140A/H193A background. Families of currents at pH_o 7, pH_i 7 showing that 10 nM Zn²⁺ produces distinct inhibition. It is evident that in addition, τ_act is slowed and the g_H-V relationship is shifted positively.

Figure 3.

The kinetics of activation is not changed in the Zn²⁺-insensitive H140A/H193A mutant. Turn-on of current was fitted with a single exponential. Direct comparison at pH_o 7 and pH_i 6.5 is shown, along with pH 7//7 WT data (from Cherny et al., 2015) for reference. Mean ± SEM is plotted for numbers of cells from negative to positive voltages: WT 5, 10, 12, 11, 11, 9, 10, and 3; and H140A/H193A 4, 6, 9, 9, 9, 6, 6, and 4. Although the values at −20 mV are just “significantly” different at P = 0.026, this is obviously a statistical aberration.

Figure 4.

Comparison of Zn²⁺ inhibition of proton current in V116H, D185N, and WT hH_V1 and in the Zn²⁺-insensitive H140A/H193A background. Inhibition of H⁺ current by Zn²⁺ at +60 mV in the V116H/H140A/H193A mutant (●), WT hH_V1 (▪), V116H/H140A/D185N/H193A (♦), and background construct H140A/H193A (▴). The mean ± SEM ratio of test pulse current in the presence of Zn²⁺ to its absence is plotted. Solid symbols and lines show measurements at pH_o 7, open symbols and dashed lines at pH_o 6. For concentrations <1 µM, Zn²⁺ was buffered with ADA. Numbers of cells, from low to high [Zn²⁺], for V116H are 6, 3, and 6 at pH_o 7; 3, 3, and 8 at pH_o 6; for WT 13, 10 and 7 at pH_o 7; 4, 6, 5, 4, and 1 at pH_o 6; for V116H/D185N 3, 4, and 4 at pH_o 7; 4, 4, 4, and 3 at pH_o 6; and for H140A/H193A 2, 5, 6, 4, and 1 at pH_o 7; 3, 3, 3, 3, and 3 at pH_o 6.

Figure 5.

Map of key amino acids in closed and open hH_V1 models. (A and B) Alpha carbons of S2 helices (yellow ribbon) of closed (A, left) and open (B, right) homology models of human H_V1 (Li et al., 2015) were superimposed using the Matchmaker program in Chimera (resource for Biocomputing, Visualization, and Informatics, University of California, San Francisco, San Francisco, CA; supported by NIGMS P41-GM103311; Pettersen et al., 2004) and are shown at the same viewing angle. S1, S2, S3, and S4 helices are indicated as red, yellow, green, and blue ribbons, respectively. The hydrophobic gasket residues V109 (on S1), F150 (on S2), and V177 and V178 (on S3) are indicated in brown.

Figure 6.

Addition of the D185N mutation eliminates the Zn²⁺ sensitivity of V116H in the H140A/H193A background. (A–C) A cell at pH_o 7, pH_i 7 is distinctly inhibited by 100 µM Zn²⁺. Families were generated by pulses from V_hold = −40 mV in 10-mV increments from −10 mV up to the voltage indicated. (D–F) A different cell at pH_o 6, pH_i 6 is inhibited only at 1 mM Zn²⁺. Families were generated by pulses from V_hold = −40 mV in 10-mV increments from 0 mV up to +80 mV.

Figure 7.

The monomeric construct is insensitive to Zn²⁺. Zn²⁺ exerted minimal effects on the background construct used in this study, in which the two His that bind Zn²⁺ in WT channels were replaced with Ala, and the C terminus was truncated at position T222 (so that the final amino acid was K221).

Figure 8.

Activation kinetics in the absence of Zn²⁺ in V116H/H140A/H193A constructs used here. Values for τ_act (mean ± SEM) are plotted for dimeric (blue) and monomeric (red) V116H mutants at pH_o 7 (solid symbols and lines) and pH_o 6 (open symbols and dashed lines), all with pH_i 6. For comparison, τ_act from WT hH_V1 is plotted (green) at pH_o 7, pH_i 7 (from Cherny et al., 2015), as well as the single mutant V116H at pH pH_o 7 and pH_i 7 (purple hexagons). It should be noted that τ_act with pH_i 7 is substantially slower than with pH_i 6. Numbers of cells starting at the most negative voltage are WT 3, 3, 6, 9, 11, 12, 12, 12, 12, 8, and 6; pH_o 7 dimer 3, 4, 4, 4, 4, 4, and 3; pH_o 6 dimer 4, 4, 4, 4, and 4; pH_o 7 monomer 5, 8, 9, 10, 10, 10, 10, 7, and 3; and pH_o 6 monomer 3, 3, 14, 15, 15, 15, 13, 4, and 3.

Figure 9.

In the V116H monomer, the turn-on of current in the presence of Zn²⁺ is practically independent of both voltage and Zn²⁺ concentration. (A and B) Families of currents are shown at pH_o 6 (A) and pH_o 7 (B), both with pH_i 6, in the indicated concentrations of Zn²⁺ with pulses applied in 10-mV increments up to the voltage shown. (C and D) The τ_act values (mean ± SEM) from these cells are shown in the absence (●) or presence of Zn²⁺ in concentrations indicated in the inset to D. Every τ_act value at every voltage is significantly (P < 0.05) higher in Zn²⁺ than in control in C and D except two values at +60 mV in D. Numbers of cells at pH_o 6 are control 3–10, 100 nM Zn²⁺ 2–7, 1 µM Zn²⁺ 2–8, 10 µM Zn²⁺ 3–6, and 100 µM Zn²⁺ 3–5; numbers of cells at pH_o 7 are control 3–7, 100 nM Zn²⁺ 2 or 3, 1 µM Zn²⁺ 2–4, and 10 µM Zn²⁺ 3. The control values are a subset of those plotted in Fig. 8; here, we include only cells in which we also had τ_act data in Zn²⁺.

(Scheme 1)

Figure 10.

Families of currents in the V116H monomer at pH 6//6. (A and B) Pulses are in 10-mV increments from V_hold = −40 mV to +90 mV. Note that there is a distinct rapidly rising phase in A at 100 nM Zn²⁺ but not in B at 10 µM Zn²⁺. (C) In the same cell, currents at +90 mV are superimposed in the presence of the indicated Zn²⁺. The fast component is only detectable at low [Zn²⁺].

Figure 11.

Pulse pair experiments in the V116H monomer (V116H/H140A/H193A/T222stop) reveal two components of “activation.” (A and B) Panels differ only in the interpulse voltages of −60 mV or 0 mV, respectively. (C) The first pulse duration was varied. All measurements were from a cell at pH_o 7, pH_i 6 with 1 µM Zn²⁺.