Brownian dynamics simulations of ion transport through the VDAC

Abstract

It is important to gain a physical understanding of ion transport through the voltage-dependent anion channel (VDAC) because this channel provides primary permeation pathways for metabolites and electrolytes between the cytosol and mitochondria. We performed grand canonical Monte Carlo/Brownian dynamics (GCMC/BD) simulations to explore the ion transport properties of human VDAC isoform 1 (hVDAC1; PDB:2K4T) embedded in an implicit membrane. When the MD-derived, space-dependent diffusion constant was used in the GCMC/BD simulations, the current-voltage characteristics and ion number profiles inside the pore showed excellent agreement with those calculated from all-atom molecular-dynamics (MD) simulations, thereby validating the GCMC/BD approach. Of the 20 NMR models of hVDAC1 currently available, the third one (NMR03) best reproduces both experimental single-channel conductance and ion selectivity (i.e., the reversal potential). In addition, detailed analyses of the ion trajectories, one-dimensional multi-ion potential of mean force, and protein charge distribution reveal that electrostatic interactions play an important role in the channel structure and ion transport relationship. Finally, the GCMC/BD simulations of various mutants based on NMR03 show good agreement with experimental ion selectivity. The difference in ion selectivity between the wild-type and the mutants is the result of altered potential of mean force profiles that are dominated by the electrostatic interactions.

Figure 1

BD simulation system, taken from a snapshot in a symmetric 0.1 M KCl solution. The hVDAC1 structure is aligned parallel to the Z direction and placed at the center of an implicit lipid membrane. The top trans bath (Z > 0) represents the intermembrane region (inside mitochondria), and the bottom cis one (Z < 0) represents the cytosol. K⁺ (magenta) and Cl⁻ (green) are represented as spheres.

Figure 2

I/V curves of K⁺ (magenta), Cl⁻ (green), and total (black) currents from the BD (solid) and MD (dotted) simulations of NMR01 in a symmetric 1.0 M KCl solution. Standard errors are also shown for the BD results.

Figure 3

I/V curves of K⁺ (magenta), Cl⁻ (green), and total (black) currents from the BD simulations of NMR03 in an asymmetric 1.0 (cis): 0.1 (trans) M KCl solution. A linearly fitted line (black dashed) is used to determine the average reversal potential where the total current is zero.

Figure 4

Single-channel conductance (black circle) at V_mp = 50 mV in symmetric 0.1 M KCl solution, and permeability ratio (red square) in an asymmetric 1.0:0.1 M KCl solution of the 20 hVDAC1 NMR models from the BD simulations. Error bars represent the 95% confidence interval assuming a t distribution of the samples. In the selection window (blue shaded) taken from the experimental measurements, the ion conductance ranges from 0.4 nS to 0.6 nS, and the permeability ratio ranges from 1.5 to 2.5. The permeability ratios of NMR09, NMR14, NMR15, NMR17, and NMR19 are not shown, because their values (5.55 < p < 7.82) are too large for the current scale. The conductance and permeability ratio values (with standard error) are listed in Table S2.

Figure 5

Normalized distribution of (A) single-channel conductance and (B) reversal potential for the 20 hVDAC1 NMR models. The conductance distribution profile is obtained from 10 conductance data points, and the reversal potential distribution profile is obtained from 100 reversal potential data points (100 combinations resulted from connecting two bias points near the reversal potential) of each NMR model. The ensemble average of single-channel conductance is 0.36 ± 0.04 nS, and that of the reversal potential is 23.7 ± 2.2 mV. The shaded area corresponds to the selection window for conductance (0.4 nS < G < 0.6 nS) and permeability ratio (1.5 < p < 2.5) as shown in Fig. 4. The solid squares in the plot correspond to the ensemble average, and open circles indicate the results from NMR03.

Figure 6

1D multi-ion PMF for K⁺ (magenta) and Cl⁻ (green) calculated from the BD simulations of (A) NMR02, (B) NMR03, and (C) NMR11 in 0.1 M KCl solution under the equilibrium condition (V_mp = 0 mV). To show the contribution of the protein charge, the PMFs without the protein charge are also drawn as dotted lines.

Figure 7

(A) Top view (from Z < 0) and (B) side view of NMR03 with the selected mutation sites of Asp-16, Lys-20, Asp-30, Lys-61, Lys-96, Gly-220, and Lys-274. (C) Comparison of permeability ratio changes of the selected mutants between the BD simulation and experimental data. The permeability ratio change (R_p) is defined as p_mutant/p_WT.