Drug-induced ion channel opening tuned by the voltage sensor charge profile

Abstract

Polyunsaturated fatty acids modulate the voltage dependence of several voltage-gated ion channels, thereby being potent modifiers of cellular excitability. Detailed knowledge of this molecular mechanism can be used in designing a new class of small-molecule compounds against hyperexcitability diseases. Here, we show that arginines on one side of the helical K-channel voltage sensor S4 increased the sensitivity to docosahexaenoic acid (DHA), whereas arginines on the opposing side decreased this sensitivity. Glutamates had opposite effects. In addition, a positively charged DHA-like molecule, arachidonyl amine, had opposite effects to the negatively charged DHA. This suggests that S4 rotates to open the channel and that DHA electrostatically affects this rotation. A channel with arginines in positions 356, 359, and 362 was extremely sensitive to DHA: 70 µM DHA at pH 9.0 increased the current >500 times at negative voltages compared with wild type (WT). The small-molecule compound pimaric acid, a novel Shaker channel opener, opened the WT channel. The 356R/359R/362R channel drastically increased this effect, suggesting it to be instrumental in future drug screening.

Figure 1.

Charge distribution and voltage dependence of opening of WT and mutated Shaker K channels. (A and B) Gating charges R1–R4 in S4 denoted as blue sticks in a VSD structure (side view [A] and top view [B]) of an open K channel (the Kv1.2/2.1 chimera structure with Shaker side chains; Long et al., 2007; Henrion et al., 2012). The approximate interaction site for DHA is marked with the red encircled negative sign in B (Börjesson and Elinder, 2011). (C and D) R1 (R362) is highlighted in five superimposed structures of S4 (side view [C] from residue 356 and top view [D] from residue 361), for four closed (C1–C4) and one open (O) model states (Henrion et al., 2012). Backbones are color coded according to opening level, from light gray (C4) to black (O). In D, the approximate interaction site for DHA is marked with the red encircled negative sign (Börjesson and Elinder, 2011). The arrow in D denotes the movement of R1 when S4 moves from C1 to O, the step most sensitive to DHA (Börjesson and Elinder, 2011). (E) Amino acid sequences of the extracellular end of S4 (Shaker channel) from the eight single-residue mutations investigated. The positively charged arginines (R) are marked in blue. The mutated region is shown in gray. R1 = R362. (F) The voltage required to reach 50% of the maximum conductance (V_1/2) plotted against the residue number of the positive charge. The dotted line corresponds to V_1/2 for the channel without charges in the studied region (R362Q). Green symbols denote positive midpoint voltages relative the neutral 356–362 segment (open symbol), and the red symbols denote negative midpoint voltages. (G) Residues 356–362 are denoted as sticks on the open Shaker VSD structure with same color coding as in F.

Figure 2.

DHA sensitivity of the single arginine mutants. (A–C) Representative current traces for voltages corresponding to 10% of G_max in control solution at pH 7.4. Black traces indicate control, and red traces indicate 70 µM DHA. The increments in current amplitudes are R362R (WT), 2.2 times (A); for L361R/R362Q, 0.9 times (B); and A359R/R362Q, 2.7 times (C). (D) DHA-induced G(V) shifts for the single-charge mutants (70 µM DHA at pH 7.4). The black line equals the DHA-induced shift for R362Q. Mean ± SEM (n = 8, 8, 6, 9, 10, 14, 9, and 9). DHA-induced shifts compared with R362Q (one-way ANOVA together with Dunnett’s multiple comparison test: **, P < 0.01; ***, P < 0.001). Green bars denote significantly larger DHA-induced shifts relative to the neutral 356–362 segment (open bar), the red bars denote significantly smaller shifts, and the yellow bars denote no significant differences. (E) Mutated residues are marked on one VSD of the Shaker K channel in states C3 and O. Same color coding as in D. (F) Correlations between the V_1/2 values and the DHA-induced shifts for the channels described in D. Slope is significantly different from zero (Pearson correlation test, and linear regression: P < 0.01 for both). The symbols denote mean ± SEM, the continuous line is the linear regression, and the dashed lines denote the 95% confidence interval.

Figure 3.

Charge-dependent DHA sensitivity. (A) Normalized representative current traces for voltages corresponding to 10% of G_max in control solution at pH 7.4. Black and gray traces indicate control solution for A359R/R362Q and A359E/R362Q, respectively (normalized to 1), blue trace denotes 70 µM DHA on A359R/R362Q, and red trace denotes 70 µM DHA on A359E/R362Q. (B) DHA-induced G(V) shifts for the channels in A. 70 µM DHA at pH 7.4. R362Q is included for comparison. Mean ± SEM (n = 7, 12, and 9). DHA-induced shifts are compared by one-way ANOVA together with Bonferroni post-hoc test: *, P < 0.05; ***, P < 0.001. (C) As in A. Red trace for L361E/R362Q and blue trace for L361R/R362Q. (D) As in B (n = 9, 12, and 14). (E) As in B (n = 7, 12, and 10; **, P < 0.01). (F) Cartoon to qualitatively explain data in A–E. Figures denote approximate positions of specific residues in the open state. Arrows denote the movements of the respective residues from the C1 state to the O state. The minus sign denotes a DHA position consistent with the experimental data.

Figure 4.

Combinations of arginines make channels more PUFA sensitive. DHA-induced shifts for multicharge channels (70 µM DHA at pH 7.4). Mean ± SEM (n = 9, 12, 10, 15, 11, 8, and 8). Horizontal lines denote shifts of R362Q (black), I360R/R362Q (light blue), and A359R/R362Q (dark blue). DHA-induced shifts compared with R362Q (one-way ANOVA together with Dunnett’s multiple comparison test: *, P < 0.05; **, P < 0.01; ***, P < 0.001).

Figure 5.

Characteristics of the 3R channel. (A and B) Current families for R362Q (A) and 3R channel (B) when stepping the membrane voltage from −80 mV to voltages between −80 and 40 mV (−60 and 60 mV for 3R channel) in 5-mV increments in control solution, pH 7.4, at a frequency of 0.2 Hz. (C) Representative current traces at pH 7.4 for 3R in control solution (black) and in 70 µM DHA (red) for a voltage corresponding to 10% of G_max in control solution (i.e., 10 mV). The increment in current amplitude is >10-fold. (D) Representative G(V) curves. Same cell as in C (control, black symbols; DHA, red symbols). ΔG(V)_DHA = −22.1 mV in this example.

Figure 6.

pH-dependent DHA sensitivity. (A) Representative current sweeps in control (black), 2.1 µM DHA (red), 7 µM DHA (orange), 21 µM DHA (yellow), 70 µM DHA (green), and 210 µM DHA (blue) at pH 9.0 and −15 mV (the voltage corresponding to 10% of G_max in control solution minus 10 mV). Current amplitude is increased >40-fold at 70 µM DHA. (B) Representative G(V) curves. Same cell, color coding, and order as in A. The shifts are −4.9, −15.7, −26.7, −50.0, and −54.6 mV in this example. (C) Dose–response curve for R362Q (gray) and the 3R channel (red) at pH 7.4 (light colored) and at pH 9.0 (dark colored). Error bars indicate SEM (n = 4–15). (D) VSD of the open 3R Shaker structure (top view to the left and side view to the right) with arginines in S4 shown as blue sticks.

Figure 7.

Drug sensitivity is increased for 3R. (A) Structure of PiMA, here in the uncharged form with the carboxylic acid group protonated. (B) G(V) shifts induced by 70 µM PiMA at pH 7.4 and 9.0 for WT and 3R. Mean ± SEM (n = 12, 6, 4, and 6; one-way ANOVA together with Bonferroni’s pairwise comparison test: **, P < 0.01; ***, P < 0.001). (C) Representative current traces at pH 9.0 for WT in control solution (black) and in 70 µM PiMA (red) for a voltage corresponding to 10% of G_max in control solution (i.e., −35 mV). The current amplitude is increased fourfold. (D) Representative G(V) curves. Same cell as in C (control, black symbols; PiMA, red symbols). ΔG(V)_PiMA = −9.0 mV in this example. (E) Representative current traces at pH 9.0 for 3R in control solution (black) and in 70 µM PiMA (red) for a voltage corresponding to 10% of G_max in control solution (i.e., 5 mV). The current amplitude is increased >10 times. (F) Representative G(V) curves. Same cell as in F (control, black symbols; PiMA, red symbols). ΔG(V)_PiMA = −28.5 mV in this example.