Voltage-regulated water flux through aquaporin channels in silico

Abstract

Aquaporins (AQPs) facilitate the passive flux of water across biological membranes in response to an osmotic pressure. A number of AQPs, for instance in plants and yeast, have been proposed to be regulated by phosphorylation, cation concentration, pH change, or membrane-mediated mechanical stress. Here we report an extensive set of molecular dynamics simulations of AQP1 and AQP4 subject to large membrane potentials in the range of ±1.5 V, suggesting that AQPs may in addition be regulated by an electrostatic potential. As the regulatory mechanism we identified the relative population of two different states of the conserved arginine in the aromatic/arginine constriction region. A positive membrane potential was found to stabilize the arginine in an up-state, which allows rapid water flux, whereas a negative potential favors a down-state, which reduces the single-channel water permeability.

Figure 1

(A) Simulation box of two stacked aquaporin-1 tetramers (cartoon representation) embedded in a phospholipid membrane (gray sticks), and solvated in water (not shown) and 150 mM sodium chloride (red and blue spheres). The electrostatic membrane potential was generated by additional cations to the central compartment (red +) and additional anions to the outer compartment (two blue –). (B) Electrostatic potential Φ(z) along the membrane normal z during 25 simulations with increasing additional charges in the two water compartments. The membrane potential ΔΦ is indicated by a black arrow.

Figure 2

(A) Single-channel permeability p_f of AQP1 versus membrane potential ΔΦ. (B and C) Conserved Arg¹⁹⁵ in AQP1 (B) in the up-state and (C) in the down-state. (D) Wide distribution of Arg¹⁹⁵-His¹⁸⁰ distance d_R-H taken from all simulations. d_R-H in x-ray structure is indicated by a shaded bar (6). (E) Probability for an open channel versus ΔΦ. Linear fit (shading) and P_open derived from a two-state model (dashed).

Figure 3

Voltage-sensitive openness of the aromatic/arginine (ar/R) region of aquaporin-4 (AQP4), as measured from the distance d_R-H between Arg²¹⁶ and His²⁰¹. (A) Distribution of d_R-H, taken from 13 AQP4 simulations at membrane voltages between –1.4 and +1.4 V (shaded histogram), revealing two distinct states. d_R-H in the AQP4 crystal structure (7) is indicated by a shaded bar. (B) Probability for an open channel P_open versus ΔΦ. Linear fit (gray) and P_open derived from a two-state model (dashed).