Voltage-gated proton channel in a dinoflagellate

Abstract

Fogel and Hastings first hypothesized the existence of voltage-gated proton channels in 1972 in bioluminescent dinoflagellates, where they were thought to trigger the flash by activating luciferase. Proton channel genes were subsequently identified in human, mouse, and Ciona intestinalis, but their existence in dinoflagellates remained unconfirmed. We identified a candidate proton channel gene from a Karlodinium veneficum cDNA library based on homology with known proton channel genes. K. veneficum is a predatory, nonbioluminescent dinoflagellate that produces toxins responsible for fish kills worldwide. Patch clamp studies on the heterologously expressed gene confirm that it codes for a genuine voltage-gated proton channel, kH(V)1: it is proton-specific and activated by depolarization, its g(H)-V relationship shifts with changes in external or internal pH, and mutation of the selectivity filter (which we identify as Asp(51)) results in loss of proton-specific conduction. Indirect evidence suggests that kH(V)1 is monomeric, unlike other proton channels. Furthermore, kH(V)1 differs from all known proton channels in activating well negative to the Nernst potential for protons, E(H). This unique voltage dependence makes the dinoflagellate proton channel ideally suited to mediate the proton influx postulated to trigger bioluminescence. In contrast to vertebrate proton channels, whose main function is acid extrusion, we propose that proton channels in dinoflagellates have fundamentally different functions of signaling and excitability.

Fig. 1.

The primary sequence of kH_V1 (line 1) compared with sequence logos of TM regions S1 to S4 of families of homologs that were created as described in Materials and Methods. The height of each letter in a stack indicates its relative representation at that location. The total stack height at each position indicates its information content, which for proteins, has a theoretical maximum of 4.3 bits (25) and depends both on the number of sequences in the alignment and the number of substitutions observed at a position. Numbers on the left and right sides indicate the length of N and C termini, respectively, (mean ± SD) of the sequences included in the alignment from which the logos were created (H_V1 n = 37, C15orf27 n = 15, VSP n = 11) or in the case of K_V, from a subset of 13 sequences drawn at random from the 38 sequences included in the alignment. All sequences used for the logos are listed in Table S1. One-way ANOVA followed by Tukey's test indicates that the length of the H_V1 C terminus differs significantly from the length of the C terminus of each other family (P < 0.001). Numbering of TM residues is for hH_V1, c15orf27, CiVSP, and Shaker (K_V). Although kH_V1 displays some significant differences from the most common H_V1 sequence, 30 of 87 TM residues match the predominant H_V1 pattern.

Fig. 2.

The kH_V1 gene product is a voltage-gated proton channel. (A) Voltage-gated proton currents generated by kH_V1 in an inside-out patch from a HEK-293 cell at pH_o = pH_i = 7.0 during voltage pulses applied from a holding potential of −60 mV in 5-mV increments from −55 to +15 mV. (B) The current–voltage relationship from this family illustrates that inward currents occur over a wide voltage range negative to the Nernst potential for protons, E_H. (C) The measured V_rev is extremely close to E_H. V_rev was determined directly from the reversal of current during depolarizing pulses. Data from 79 whole-cell and excised patch measurements are included. The dashed line shows E_H.

Fig. 3.

Effects of pH_i on proton currents in an inside-out patch of membrane at pH_o = 7.0 (A–E) and regulation of the g_H–V relationship in kH_V1 by the pH gradient ΔpH (F). Pulses were applied in 10-mV increments up to the voltage shown at the indicated pH_i from a holding potential of −90, −60, or −40 mV for A to C, respectively. (D) Steady state current–voltage relationship for the families in A–C. The current amplitude was obtained from single exponential fits. (E) Corresponding g_H–V relationships. (F) The voltage at which H⁺ current was first detectable (typically ∼1% of the maximum g_H), V_threshold, is plotted against the V_rev measured in each solution in each cell or patch studied. The dashed line indicates equality between the parameters. The solid line is the least squares fit to the data given by V_threshold = 0.79 V_rev − 37 mV. The dotted line shows the relationship found previously for 15 cell types (27) described by V_threshold = 0.79 V_rev + 23 mV.

Fig. 4.

Mutation of Asp⁵¹ alters the selectivity of kH_V1. (A) In symmetrical pH 5.5 TMA⁺ CH₃SO₃⁻ solutions, the whole-cell D51A conductance is activated well below 0 mV, although the inward currents are small. Pulses were applied in 10-mV increments from a holding potential of −40 mV. V_rev is between −10 and 0 mV (arrow). (B) Replacing 130 mM CH₃SO₃⁻ with 130 mM Cl^– in the bath greatly increased outward currents during a family of pulses identical to the family in A. (C) The tail current V_rev in the pH 5.5 Cl^– solution was just negative to −40 mV. (D) Shifts of V_rev when CH₃SO₃⁻ was replaced by Cl^–, all at symmetrical pH 5.5, reveal that D51A, D51S, and D51H all conduct Cl^–. Values for WT and D112E do not differ significantly from 0 mV. Error bars are SEM (n = 3, 4, 4, 4, and 4, respectively).