To keep or not to keep? the question of crystallographic waters for enzyme simulations in organic solvent

Abstract

The use of enzymes in non-aqueous solvents expands the use of biocatalysts to hydrophobic substrates, with the ability to tune selectivity of reactions through solvent selection. Non-aqueous enzymology also allows for fundamental studies on the role of water and other solvents in enzyme structure, dynamics, and function. Molecular dynamics simulations serve as a powerful tool in this area, providing detailed atomic information about the effect of solvents on enzyme properties. However, a common protocol for non-aqueous enzyme simulations does not exist. If you want to simulate enzymes in non-aqueous solutions, how many and which crystallographic waters do you keep? In the present work, this question is addressed by determining which crystallographic water molecules lead most quickly to an equilibrated protein structure. Five different methods of selecting and keeping crystallographic waters are used in order to discover which crystallographic waters lead the protein structure to reach an equilibrated structure more rapidly in organic solutions. It is found that buried waters contribute most to rapid equilibration in organic solvent, with slow-diffusing waters giving similar results.

How Does Solvation Layer Mobility Affect Protein Structural Dynamics?

“…Our previous work addressed the question of keeping (or not keeping) crystallographic water molecules in simulations of enzymes, and we found that slowly-diffusing waters in organic solvent should be kept in order for simulations to quickly equilibrate (as described in Methods). We also saw that the location of retained waters affects conformational sampling (Dahanayake et al, 2016 ). We have postulated that both buried and slow-diffusing water molecules are the ones which are retained experimentally when enzymes are used in organic solvent.…”

“…The starting coordinates of CALB enzyme were taken from the X-ray crystallographic structure (PDB ID: 1TCA). Six buried crystallographic water molecules identified by Uppenberg et al ( 1994 ) were kept, as it was concluded in our previous studies (Dahanayake et al, 2016 ) that this leads to fastest equilibration in non-aqueous simulations. All the simulations were performed using GROMACS (version 4.6.3) (Kamatari et al, 1996 ) software package.…”

“…Here, we hypothesize that what really matters in the protein-solvent dynamics relationship can be summarized by that old real estate maxim: location, location, location . Our work evaluating methodologies for simulation of enzymes in organic solvent revealed that the location of crystallographic waters (i.e., which crystallographic waters) kept during simulations influences the conformational dynamics of the enzyme CALB (Dahanayake et al, 2016 ). The present work focuses on the local connection between solvent dynamics and protein dynamics, looking specifically at whether and how a protein region's flexibility or stability may depend on the dynamics of solvent in the solvation layer surrounding different portions of the proteins: for example, solvent in the immediate surroundings of an alpha helix (Figure 1 ).…”

In Silico Studies of Small Molecule Interactions with Enzymes Reveal Aspects of Catalytic Function

“…Simulation results show that the nature of organic solvents influences the extent of stripping of water molecules from the protein surface: non-polar solvents give rise to tightly bound large water clusters on the protein surface, while polar solvents promote loosely bound small clusters of water [ 136 ]. A recent MD simulation study of Candida antarctica lipase B (CALB) and cytochrome c enzymes indicates that retaining buried waters in the presence of different non-aqueous solvents leads to faster-equilibrating MD trajectories [ 138 ]. Furthermore, retaining buried waters affects the conformational sampling of CALB in organic solvents.…”

“…An atomistic MD docking simulation of the P450BM-3 heme domain was performed to investigate preferential binding modes and electron transfer (ET) from an ET mediator, cobalt(II) sepulchrate (Co(II)Sep), to the heme iron in solution at different Co(II)Sep concentrations [ 138 ]. Results provided new insights into the ligand adsorption mechanism, through detailed binding site mapping on the protein surface.…”