Mutant cycle analysis with modified saxitoxins reveals specific interactions critical to attaining high-affinity inhibition of hNaV1.7

Abstract

Improper function of voltage-gated sodium channels (NaVs), obligatory membrane proteins for bioelectrical signaling, has been linked to a number of human pathologies. Small-molecule agents that target NaVs hold considerable promise for treatment of chronic disease. Absent a comprehensive understanding of channel structure, the challenge of designing selective agents to modulate the activity of NaV subtypes is formidable. We have endeavored to gain insight into the 3D architecture of the outer vestibule of NaV through a systematic structure-activity relationship (SAR) study involving the bis-guanidinium toxin saxitoxin (STX), modified saxitoxins, and protein mutagenesis. Mutant cycle analysis has led to the identification of an acetylated variant of STX with unprecedented, low-nanomolar affinity for human NaV1.7 (hNaV1.7), a channel subtype that has been implicated in pain perception. A revised toxin-receptor binding model is presented, which is consistent with the large body of SAR data that we have obtained. This new model is expected to facilitate subsequent efforts to design isoform-selective NaV inhibitors.

Keywords: guanidinium toxin; mutant cycle analysis; sodium channel.

Fig. 1.

(A) Schematic drawing of 1 bound in the Na_V outer pore as suggested by previous electrophysiology and mutagenesis experiments. Each of the four domains (I, orange; II, red; III, gray; and IV, teal) is represented by a separate panel. (B) Schematic representation of double-mutant cycle analysis and mathematical definition of coupling energy (ΔΔE_Ω). X¹ = IC₅₀(WT⋅STX)/IC₅₀(MutNa_V⋅STX), X² = IC₅₀(WT⋅MeSTX)/IC₅₀(MutNa_V⋅MeSTX), Y¹ = IC₅₀(MutNa_V⋅STX)/IC₅₀(MutNa_V⋅MeSTX), and Y² = IC₅₀(WT⋅STX)/IC₅₀(WT⋅MeSTX).

Fig. S1.

Mutant cycle analysis definition and examples. (A) Schematic of a single mutant cycle with mathematical expressions for coupling energy ΔΔE_Ω. R is the ideal gas constant and T is temperature. Each IC₅₀ is the half maximal inhibition concentration determined by whole-cell voltage-clamp electrophysiology. When the separation between IC₅₀ values for the reference compound and the modified compound is different with a mutant than with the WT protein, a nonzero value for ΔΔE_Ω is obtained (B), but when the separation is the same (C), ΔΔE_Ω is equal to 0. In B, the difference in the relative affinity of 1 and 4 with Y401A is smaller than the difference with the WT channel, indicating a positive coupling (ΔΔE_Ω > 0). In C, the relative affinities of 1 and 8 against WT rNa_V1.4 and Y401A are similar, and ΔΔE_Ω ∼0 kcal/mol.

Fig. S2.

Dose–response curves for current inhibition of rNa_V1.4 and mutants thereof by compounds 1–8 determined by whole-cell voltage-clamp electrophysiology. Recordings were made on Na_V channels recombinantly expressed in CHO cells. Data were fit to Langmuir isotherms to produce IC₅₀ values and each data point represents the average of n ≥ 3 cells ± SD.

Fig. 2.

Mutant cycle analysis with compounds 2–8 and Na_V single-point mutants. ΔΔE_Ω absolute values are shown in units of kilocalories per mole. Dose–response curves, relative inhibition constants (IC₅₀s), and SEM for calculations are given in Fig. S2 and Table S1.

Fig. S3.

Representative current recordings elicited by a 10-ms voltage step from −100 to 0 mV before (black) and following (red) application of STX to CHO cells expressing single-point and double-point rNa_V1.4 mutants. Unless otherwise noted, traces in red represent current following application of 100 nM STX.

Fig. 3.

Additional mutagenesis of M1240 and D1241 reveals interactions responsible for coupling of C13-substituents to DIII. (A) Absolute values of ΔΔE_Ω (kilocalories per mole) for compounds 7 and 8 with M1240 and D1241 single-point mutants. (B) Absolute values of ΔΔE_Ω (kilocalories per mole) for compounds 4, 6, 7, and 8 with the rNa_V1.4 M1240T/D1241I double-point mutant. (C) Concentration response curves for current inhibition of rNa_V1.4 (solid line) and hNa_V1.7 (dotted line) by 8 determined by whole-cell voltage-clamp electrophysiology. (D) Current recordings elicited by a 10-ms voltage step from −100 to 0 mV before (black) and following (red) application of 25 nM 8 to CHO cells expressing rNa_V1.4 (Top) and HEK cells expressing hNa_V1.7 (Bottom). WT rNa_V1.4 and STX were used as references in ΔΔE_Ω calculations.

Fig. S4.

Dose–response curves for current inhibition of rNa_V1.4 domain III mutants by compounds 1, 6, 7, 8, and 9 determined by whole-cell voltage-clamp electrophysiology. Recordings were made on Na_V channels recombinantly expressed in CHO cells. Data were fit to Langmuir isotherms to produce IC₅₀ values and each data point represents the average of n ≥ 3 cells ± SD.

Fig. S5.

Overlaid dose–response curves and fit parameters for current inhibition of rNa_V1.2, rNa_V1.4, hNa_V1.5, and hNa_V1.7 by compound 8 determined by whole-cell voltage-clamp electrophysiology. Recordings were made on rNa_V1.4 and hNa_V1.5 channels recombinantly expressed in CHO cells, rNa_V1.2 stably expressed in CHO cells, and hNa_V1.7 stably expressed in HEK cells. Data were fit to Langmuir isotherms to produce IC₅₀ values and each data point represents the average of n ≥ 3 cells ± SD.

Fig. 4.

Docking of STX at site 1 of rNa_V1.4 in the canonical (A) and proposed (B) orientations. The four domains are shown in cartoon representations and colored orange (DI), red (DII), gray (DIII), and teal (DIV). Images highlighting the differences between electrostatic potential surfaces of STX (C) and C13-OAc STX 8 (D) docked to the rNa_V1.4 outer pore in the proposed orientation. Equivalent images showing STX (E) and C13-OAc STX 8 (F) in the binding pocket of the M1240T/D1241I double mutant. Depicted potential range is −20 (red) to +5 kT (blue). Toxin structures generated using OMEGA version 2.5.1; docking performed with OEDocking version 3.0.1 (OpenEye Scientific Software, www.eyesopen.com).