Comparison of Selectively Polarizable Force Fields for Ion−Water−Peptide Interactions: Ion Translocation in a Gramicidin-like Channel

Free energy simulations of single and double ion occupancy in gramicidin A

Simultaneous occupancy of the two binding sites in gramicidin A by monovalent cations is a well known property of this channel, but the energetic feasibility of this process in molecular dynamics simulations has not been established so far. Here the authors study the energetics of single and double ion occupancy in gramicidin A by constructing the potential of mean force for single and pair of cations. As representatives of small and large ions, they consider both Na+ and K+ ions in the calculations. Binding constants of ions are estimated from the free energy profiles. Comparisons with the experimental results indicate 3-4 kT discrepancy in the binding energies. They also study the coordination of the ions in their respective binding sites and the dynamic behavior of the channel water during the double ion binding process.

Energetics of ion permeation, rejection, binding, and block in gramicidin A from free energy simulations

The rigid force fields currently used in molecular dynamics (MD) simulations of biomolecules are optimized for globular proteins. Whether they can also be used in MD simulations of membrane proteins is an important issue that needs to be resolved. Here we address this issue using the gramicidin A channel, which provides an ideal test case because of the simplicity of its structure and the availability of a wealth of functional data. Permeation properties of gramicidin A can be summarized as "it conducts monovalent cations, rejects anions, and binds divalent cations." Hence, a comprehensive test should consider the energetics of permeation for all three types of ions. To that end, we construct the potential of mean force for K(+), Cl(-), and Ca(2+) ions along the channel axis. For an independent check of the potential-of-mean-force results, we also calculate the free energy differences for these ions at the channel center and binding sites relative to bulk. We find that "rejection of anions" is satisfied but there are difficulties in accommodating the other two properties using the current MD force fields.

Test of molecular dynamics force fields in gramicidin A

The force fields commonly used in molecular dynamics simulations of proteins are optimized under bulk conditions. Whether the same force fields can be used in simulations of membrane proteins is not well established, although they are increasingly being used for such purposes. Here we consider ion permeation in the gramicidin A channel as a test of the AMBER force field in a membrane environment. The potentials of mean force for potassium ions are calculated along the channel axis and compared with the one deduced from the experimental conductance data. The calculated result indicates a rather large central barrier similar to those obtained from other force fields, which are incompatible with the conductance data. We suggest that lack of polarizability is the most likely cause of this problem, and, therefore, urge development of polarizable force fields for simulations of membrane proteins.

Gramicidin A channel as a test ground for molecular dynamics force fields

We use the well-known structural and functional properties of the gramicidin A channel to test the appropriateness of force fields commonly used in molecular dynamics (MD) simulations of ion channels. For this purpose, the high-resolution structure of the gramicidin A dimer is embedded in a dimyristoylphosphatidylcholine bilayer, and the potential of mean force of a K(+) ion is calculated along the channel axis using the umbrella sampling method. Calculations are performed using two of the most common force fields in MD simulations: CHARMM and GROMACS. Both force fields lead to large central barriers for K(+) ion permeation, that are substantially higher than those deduced from the physiological data by inverse methods. In long MD simulations lasting over 60 ns, several ions are observed to enter the binding site but none of them crossed the channel despite the presence of a large driving field. The present results, taken together with many earlier studies, highlights the shortcomings of the standard force fields used in MD simulations of ion channels and calls for construction of more appropriate force fields for this purpose.

Voltage-gated proton channels and other proton transfer pathways

Proton channels exist in a wide variety of membrane proteins where they transport protons rapidly and efficiently. Usually the proton pathway is formed mainly by water molecules present in the protein, but its function is regulated by titratable groups on critical amino acid residues in the pathway. All proton channels conduct protons by a hydrogen-bonded chain mechanism in which the proton hops from one water or titratable group to the next. Voltage-gated proton channels represent a specific subset of proton channels that have voltage- and time-dependent gating like other ion channels. However, they differ from most ion channels in their extraordinarily high selectivity, tiny conductance, strong temperature and deuterium isotope effects on conductance and gating kinetics, and insensitivity to block by steric occlusion. Gating of H(+) channels is regulated tightly by pH and voltage, ensuring that they open only when the electrochemical gradient is outward. Thus they function to extrude acid from cells. H(+) channels are expressed in many cells. During the respiratory burst in phagocytes, H(+) current compensates for electron extrusion by NADPH oxidase. Most evidence indicates that the H(+) channel is not part of the NADPH oxidase complex, but rather is a distinct and as yet unidentified molecule.

The role of Trp side chains in tuning single proton conduction through gramicidin channels

We present an extensive set of measurements of proton conduction through gramicidin A (gA), B (gB), and M (gM) homodimer channels which have 4, 3, or 0 Trp residues at each end of the channel, respectively. In gA we find a shoulder separating two domains of conductance increasing with concentration, confirming the results of Eisenman, G., B. Enos, J. Hagglund, and J. Sandblom. 1980. Ann. NY. Acad. Sci. 339:8-20. In gB, the shoulder is shifted by approximately 1/2 pH unit to higher H(+) concentrations and is very sharply defined. No shoulder appears in the gM data, but an associated transition from sublinear to superlinear I-V values occurs at a 100-fold higher [H(+)] in gM than in gA. The data in the low concentration domain are analyzed using a configuration space model of single-proton conduction, assuming that the difference in the proton potential of mean force (PMF) between gA and its analogs is constant, similar to the results of Anderson, D., R. B. Shirts, T. A. Cross, and D. D. Busath. 2001. Biophys. J. 81:1255-1264. Our results suggest that the average amplitudes of the calculated proton PMFs are nearly correct, but that the water reorientation barrier calculated for gA by molecular dynamics using the PM6 water model (Pomès, R., and B. Roux. 1997. Biophys. J. 72:246a) must be reduced in amplitude by 1.5 kcal/mol or more, and is not rate-limiting for gA.

Noncontact dipole effects on channel permeation. V. Computed potentials for fluorinated gramicidin

Experimental and theoretical calculations indicate that the dipole moment of the four Trp side chains in gramicidin A (gA) channels modify channel conductance through long-range electrostatic interactions. Electrostatic ion/side-chain interaction energies along the channel were computed with CHARMM using ab initio atom charges for native and 4-, 5-, or 6-fluorinated Trp side chains. The bulk water reaction to the polar side chains was included using the method of images as implemented by, and channel waters in idealized structures were included. Ion/Trp interaction energies were approximately -0.6 kcal/mol throughout the channel for all four of the native Trp pairs. Channel waters produced a modest reduction in the magnitude of interactions, essentially offsetting images representing the bulk water outside the channel. The effects of side-chain fluorination depended on ring position and, to a lesser extent, residue number. Compared with native Trp, 5-fluorination reduces the translocation barrier with minor effects on the exit barrier. In contrast, 6-fluorination primarily reduces exit barrier. 4-Fluorination produces a more complex double-well energy profile. Effects of measured side-chain movements resulting from fluorination or change in lipid bilayer were negligible whereas thermal side chain librations cause large effects, especially in the region of the ion-binding sites.

Ionic interactions in multiply occupied channels

A significant number of physiologically important ion channels function via multi-ion mechanisms where repulsion between ions at slightly separated locations is believed to be critical for permeation. We apply the semi-microscopic Monte Carlo approach and analyse how multiple occupancy affects permeation energetics and ion-water-peptide correlations. We consider double occupancy in idealized models of two systems: gramicidin A and the KcsA K+ channel. We focus on the excess repulsion energy due to ion-water and ion-peptide correlations (repulsion energy adjusted for direct ion-ion interaction). Gramicidin, where multiple occupancy is marginally important functionally, is ideal for correlating structure and ion interactions. Pair occupancy is stabilized by interaction with bulk solvent, destabilized by interaction with both the channel water and, as binding sites are far apart, the peptide backbone. In the KcsA K+ channel, double occupancy is promoted by the uneven spacing and the large ion-water separations in the selectivity filter. The carbonyls forming the binding cavities are equally important for pair stabilization. Due to the binding pocket's design, net ionic repulsion is approximately 25-30% of what it would be in a gramicidin-like structure with the same interionic spacing.