Transmembrane potential of physiologically relevant model membranes: Effects of membrane asymmetry

Abstract

Transmembrane potential difference (V_m) plays important roles in regulating various biological processes. At the macro level, V_m can be experimentally measured or calculated using the Nernst or Goldman-Hodgkin-Katz equation. However, the atomic details responsible for its generation and impact on protein and lipid dynamics still need to be further elucidated. In this work, we performed a series of all-atom molecular dynamics (MD) simulations of symmetric model membranes of various lipid compositions and cation contents to evaluate the relationship between membrane asymmetry and V_m. Specifically, we studied the impact of the asymmetric distribution of POPS (1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-l-serine), PIP2 (phosphatidylinositol 4,5-bisphosphate), as well as Na⁺ and K⁺ on V_m using atomically detailed MD simulations of symmetric model membranes. The results suggest that, for an asymmetric POPC-POPC/POPS bilayer in the presence of NaCl, the presence of the monovalent anionic lipid POPS in the inner leaflet polarizes the membrane (ΔV_m < 0). Intriguingly, replacing a third of the POPS lipids by the polyvalent anionic signaling lipid PIP2 counteracts this effect, resulting in a smaller negative membrane potential. We also found that replacing Na⁺ ions in the inner region by K⁺ depolarizes the membrane (ΔV_m > 0). These divergent effects arise from variations in the strength of cation-lipid interactions and are correlated with changes in lipid chain order and head-group orientation.

SCHEME 1.

Estimating the electrostatic potential (ψ) of an asymmetric membrane using two symmetric membranes. ψ₁ and ψ₂ are the symmetric electrostatic potential profiles obtained from the symmetrized charge distribution profiles of the respective symmetric membranes.

FIG. 1.

Symmetrized charge density distribution (left column) and electrostatic potential (right column) profiles of the five symmetric systems simulated in this work. A bin width of 0.05 nm was used, and error bars (standard deviations) are shown in gray. The non-symmetrized charge density profiles are shown in Fig. S1.

FIG. 2.

Effects of membrane asymmetry on electrostatic potential profiles, ψ(z), and transmembrane potential difference, V_m. (a) Static potential profile ψ(z) across model membranes with asymmetric lipid composition and ion distribution. (b) The transmembrane potential difference V_m in the four asymmetric membrane systems.

FIG. 3.

Contact probability (cutoff = 0.3 nm) between ions (Na⁺ or K⁺) and lipids for the systems PC/PS/PIP2 and PC/PS/PIP2/K. Lipid atoms were colored blue through red based on their contact probability with the ions (the contact portability with all atoms of all lipids is 100%). The same overall trend is obtained using larger cutoff values for contact definition (Fig. S3).

FIG. 4.

Mass density distribution profile of the model membrane systems studied in this work.

FIG. 5.

Lipid chain order parameter of the five simulated symmetric membranes in this work.

FIG. 6.

Effects of Na⁺ and K⁺ on lipid head-group orientation. (a) Schematics of the vectors used for calculating head-group tilt angles. (b) Snapshots of two systems we have simulated in the presence of Na⁺ or K⁺, with POPC in green, POPS in pink, PIP2 in red, Na⁺ and K⁺ in yellow, and Cl⁻ in white. Water molecules are omitted for clarity. (c) Average tilt angles of the three lipids in the presence of different cations. (d) Average lipid height measured by the distance between the glycerol-phosphate phosphorus atom and the terminal carbon atom of the saturated lipid chains. Errors represent the standard deviation of the average over the last three 300 ns trajectories.