Reply to “Complementarities and Convergence of Results in Bacteriorhodopsin Trimer Simulations”

Abstract

z The results published by Kandt, Schlitter, and Gerwert in a recent issue of Biophysical Journal (Kandt et al., 2004) on molecular dynamics (MD) simulations of bacteriorhodopsin (bR) trimer in a water/lipid environment represent a new and valuable step in the field of bacteriorhodopsin modeling and, more generally, in the developing field of simulations of fully integrated and functional membrane biosystems. In their work, Kandt et al. report on water dynamics in and around bacteriorhodopsin trimer as a function of time and as a function of the protonation state of the retinal moiety. Their findings complement our previously published results on simulation of bR in monomers as well as of the purple membrane (PM), comprising bacteriorhodopsin trimers explicitly hydrated in their complete native functional lipid environment (Baudry et al., 2001). The two studies exhibit differences in technical and structural respects in the type of lipids (POPC bilayer in Kandt et al., native PM lipids in Baudry et al., including squalene molecules, the removal of which leads to modified photocycle kinetics), in the force field (GROMACS in Kandt et al., CHARMM in Baudry et al.), in simulation time (5 ns in Kandt et al., 1 ns in Baudry et al.), in the MD engine used (GROMACS in Kandt et al., NAMD2 for MD simulations and CHARMM for free energy calculations in Baudry et al.), as well as in the starting structures for ground-state bR. Despite these differences, we find it extremely interesting that several of the findings previously published in Baudry et al., in particular the role of Asp-96, Arg-82, and retinal isomerization on internal water movement and water exchange with the bulk, are very similar to those reported in Kandt et al. As the results of Baudry et al. were not cited nor commented on in Kandt et al., we believe it is of interest to discuss the similarities of the two articles in this letter to exemplify the conver-gence, reliability, and maturity of recent advances in the field of bR molecular modeling CONFORMATIONAL CHANGES OF ARG-82 AND ITS EFFECT ON WATER MOVEMENT The hydration pattern around retinal is similar in both Kandt et al. and Baudry et al., in particular, with water molecules 401, 402, 406, and 407 present in both models in the retinal/ Asp-85/Asp-212/Arg-82 region. The results published in Baudry et al. on bR monomer simulations indicated that water rearrangement could take place when the retinal changes its isomeric state. Water rearrangement, connecting the Asp-85/Asp-212/Arg-82 re-gion with the extracellular channel in monomeric bR, was observed when several conditions were met: a), photo-isomerization of retinal, b), protein flexibility, and c), Arg-82 side-chain down/up conformational change. As was noted in Baudry et al., and confirmed in subsequent quantum mechanical investigations of the photoevent, the nature of the potential, as well as the initial placement of water molecules and Arg-82, could influence the detailed timing of these results. Nevertheless, it was shown that water molecules could possibly move from a level below Arg-82 to the retinal binding site depending on the down-to-up movement of Arg-82 and the isomerization state of retinal. Another possibility suggested in Baudry et al. was the displacement of a water molecule (W4 in Baudry et al., corresponding to region III in Kandt et al.) to establish H-bond contact with Asp-85. In agreement with these findings, Kandt and co-workers report that water densities are highly sensitive to the conformation of Arg-82 and Schiff base protonation. When modeling the Schiff base deprotonation by neutralizing the retinal charge, Kandt et al. find that an upward movement of Arg-82 takes place, allowing a water rearrangement that connect densities IV (above Arg-82) and V (below Arg-82), linking the retinal to the extracellular bulk through a Grotthus-like proton pathway. ASP-96 AS AN INTRACELLULAR CHANNEL GATE