Histidine168 is crucial for ΔpH-dependent gating of the human voltage-gated proton channel, hHV1

Abstract

We recently identified a voltage-gated proton channel gene in the snail Helisoma trivolvis, HtH_V1, and determined its electrophysiological properties. Consistent with early studies of proton currents in snail neurons, HtH_V1 opens rapidly, but it unexpectedly exhibits uniquely defective sensitivity to intracellular pH (pH_i). The H⁺ conductance (g_H)-V relationship in the voltage-gated proton channel (H_V1) from other species shifts 40 mV when either pH_i or pH_o (extracellular pH) is changed by 1 unit. This property, called ΔpH-dependent gating, is crucial to the functions of H_V1 in many species and in numerous human tissues. The HtH_V1 channel exhibits normal pH_o dependence but anomalously weak pH_i dependence. In this study, we show that a single point mutation in human hH_V1-changing His¹⁶⁸ to Gln¹⁶⁸, the corresponding residue in HtH_V1-compromises the pH_i dependence of gating in the human channel so that it recapitulates the HtH_V1 response. This location was previously identified as a contributor to the rapid gating kinetics of H_V1 in Strongylocentrotus purpuratus His¹⁶⁸ mutation in human H_V1 accelerates activation but accounts for only a fraction of the species difference. H168Q, H168S, or H168T mutants exhibit normal pH_o dependence, but changing pH_i shifts the g_H-V relationship on average by <20 mV/unit. Thus, His¹⁶⁸ is critical to pH_i sensing in hH_V1. His¹⁶⁸, located at the inner end of the pore on the S3 transmembrane helix, is the first residue identified in H_V1 that significantly impairs pH sensing when mutated. Because pH_o dependence remains intact, the selective erosion of pH_i dependence supports the idea that there are distinct internal and external pH sensors. Although His¹⁶⁸ may itself be a pH_i sensor, the converse mutation, Q229H, does not normalize the pH_i sensitivity of the HtH_V1 channel. We hypothesize that the imidazole group of His¹⁶⁸ interacts with nearby Phe¹⁶⁵ or other parts of hH_V1 to transduce pH_i into shifts of voltage-dependent gating.

Figure 1.

Alignment of H_V1 S2–S3 linker and nearby regions (E153–D185 in hH_V1). H_V1 from several species were characterized as activating rapidly (green background: HtH_V1, LsH_V1, and SpH_V1) or slowly (pink background: hH_V1, mH_V1, OcH_V1, RrH_V1, and CgH_V1) based on available electrophysiological data. Unshaded species exhibit intermediate kinetics. In hH_V1 numbering, the asterisk (*) indicates the position corresponding to the throttle histidine H168; the caret (^) indicates the position corresponding to F165. The S2–S3 linker in hH_V1 encompasses K157–F165 (Li et al., 2015).

Figure 2.

His¹⁶⁸ mutants are proton selective. Mean ± SEM values for the reversal potential (V_rev) are plotted for 12 H168Q, 9 H168T, and 6 H168S cells or patches. The linear regression slope for each mutant is −52.6, −53.2, and −56.3 mV/unit change in ΔpH, respectively. For comparison, the gray dashed line shows the Nernst potential, E_H, expected for perfect H⁺ selectivity.

Figure 3.

The throttle histidine explains only a fraction of the rapid activation of HtH_V1. Activation time constants (τ_act) were determined by single exponential fits to currents in WT HtH_V1, WT hH_V1, and in four hH_V1 mutants, as indicated, all at pH_o 7 and pH_i 7. Mean ± SEM is plotted for 3–12 WT hH_V1, 2–3 H167K, 3–6 H168S, 3–9 H168Q, 3–6 H168T, and 7–17 WT HtH_V1. At all voltages from −10 to +80 mV, the WT hH_V1 value is significantly larger than that of H168T, H168S, and H168Q combined and larger than HtH_V1 (P < 0.01). The WT hH_V1 data were taken from Cherny et al. (2015) and the HtH_V1 data from Thomas et al. (2018).

Figure 4.

Gating of His¹⁶⁸ mutants has normal pH_o dependence. (A–C) Families of currents in a cell expressing H168Q with pH_i 6 and pH_o 5, 6, and 7, with pulses in 10-mV increments up to the voltage indicated. Holding potential, V_hold was −40 mV (A and B) or −60 mV (C). (D and E) Proton current–voltage curves (D) and g_H-V relationships (E) from the families in A–C exhibit a normal 40-mV shift/unit change in pH_o.

Figure 5.

Mutation of His¹⁶⁸ in hH_V1 weakens the pH_i dependence of gating. (A–E) Families of currents at several pH_i values in an inside-out patch from a cell transfected with H168T, with pH_o 7 in the pipette. Pulses were applied in 10-mV increments up to the voltage indicated from V_hold = −80 mV (A), −60 mV (B), or −40 mV (C–E). (D) Shorter pulses were applied to 60 mV and above, and the tail currents have been removed. (E) Pulses above 60 mV were omitted for clarity. (F) Current–voltage curves. (G) g_H-V relationships from this experiment.

Figure 6.

The His¹⁶⁸ mutation recapitulates the anomalous ΔpH dependence of HtH_V1. (A) The effect of pH_o and pH_i on the position of the g_H-V relationship in the three H168x mutants combined (mean ± SEM) is plotted. The position of the g_H-V relationship was defined in terms of V(g_H,max/10). The dashed green line shows a slope of 40 mV/unit for reference. The dependence of these mutants on pH_o is normal, whereas their pH_i response is greatly attenuated. Linear regression on all points (ignoring the obvious nonlinearity) gives a slope of 38.9 mV/unit change in pH_o and 16.1 mV/unit change in pH_i. (B) The same data are replotted (shaded symbols) along with the analogous measurements for HtH_V1 (open symbols) taken from Fig. 9 of the companion article (Thomas et al., 2018). In whole-cell measurements, pH_o was varied with pH_i 7 (blue symbols). When pH_i was varied by using inside-out patches with pH_o 7, there was very little shift of the g_H-V relationship (red symbols). Numbers of cells for increasing ΔpH in H168X mutants for pH_i 7 are 3, 10, 11, and 5 and for pH_o 7 are 6, 10, 14, 11, and 4.

Figure 7.

The H167K mutant of hH_V1 has normal pH_i dependence. (A–C) Families of currents are shown in an inside-out patch with pH_o 7 and pH_i 6, 7, or 8, as labeled. Pulses were applied in 10-mV increments up to the voltage indicated, from V_hold = −60 mV (A) or -40 mV (B and C). (D and E) Current–voltage curves (D) and g_H-V relationships from this patch are illustrated (E). At pH_i 8, g_H was estimated from tail current amplitudes and scaled according to g_H calculated from the outward current at 90 mV.

Figure 8.

The F166L and H167K mutations do not impair pH_i sensing of hH_V1. The pH_i responses of the H168X mutants are replotted from Fig. 6. The pH_i responses of WT hH_V1, the F166L mutant of hH_V1, and the H167K mutant of hH_V1 are indistinguishable from each other. The F165X data are from three F165A and three F165H patches combined, and 13.8 mV has been added to all values to facilitate comparison with WT. All data (mean ± SEM) come from inside-out patches studied at pH_o 7, with n = 2–9 (H168X), 1–4 (F166L), 3 (H167K), and 6 (F165X).

Figure 9.

The Q229H mutation does not restore human-like activation kinetics or pH_i sensitivity to HtH_V1. (A–D) Families of currents in an inside-out patch with pH_o 7 in the pipette solution in 10-mV increments up to the voltage shown from V_hold = −80, −60, −40, and −40 mV, at pH_i 5, 6, 7, and 8, respectively. (E) Currents from this patch converted to g_H produce g_H-V relationships that shift much less than 40 mV/unit. Corresponding measurements in WT HtH_V1 are plotted in Fig. 5 G.

Figure 10.

The Q229H mutation does not restore human-like pH_i sensitivity to HtH_V1. The dependence of the position of the g_H-V relationship on both pH_o and pH_i is similar in WT HtH_V1 (open symbols and dotted lines) and in the Q229H mutant (shaded symbols and solid lines). The WT data are replotted from Fig. 6 B. Mean ± SEM are shown for n = 2–5 cells with pH_i 7 and three inside-out patches with pH_o 7.