AmericanChemicalSociety.com

The goal of this PRF funded project is to develop and apply a new approach based on molecular dynamics simulations to enable theoretical studies of pH-dependent surfactant micelle properties. This approach will bridge the gap between all-atom and coarse-grained simulations. While all-atom simulations are not feasible for studying conformational processes of surfactant micelles due to high computational demand, coarse-grained simulations can not offer detailed information such as the pka of surfactant molecules embedded in micelles. Over the course of last grant year, we have made significant progress towards the proposal goal. In order to accelerate simulations and yet retain atomistic features of micelles, we have parameterized a generalized Born (GB) implicit-solvent model for treatment of solvation effects while keeping the all-atom potential energy function for representation of the surfactant molecules. We applied the new approach in continuous constant pH molecular dynamics simulations (CPHMD) to predict the pka's of surfactant molecules embedded in micelles, which is of interest in detergent industry and oil refinery processes. This work has led to a manuscript with two graduate students as the joint first authors (Wang, Wallace, and Shen, to be submitted). A brief description of the results is given below. While several variants of GB implicit-solvent models have been successfully applied to protein simulations, it remains unknown whether they can be used for simulations of micellar assemblies. Our focus here is on neutral and charged surfactant molecules, lauric acid (LAU), sodium dodecyl sulfate (SDS), dodecyl trimethyl ammonium chloride (DTA), and dodecyl triethylene glycol ether (DE3). Following a previously published procedure we calculated the radial distribution functions of charge density to obtain a set of GB input radii that define the solute-solvent dielectric boundary. These radii were then adjusted by targeting the head-to-head and side-by-side interaction free energies calculated from explicit-solvent simulations. Finally, the radii were tested by comparing various conformational properties of SDS, DTA, and DE3 micelles in GB simulations with those in explicit-solvent simulations. We applied the newly parameterized GB models together with the existing surface-area based nonpolar model in replica-exchange CPHMD titration simulations to predict pka values of a single lauric acid embedded in SDS, DTA and DE3 micelles. The resulting pka's are in good agreement with experimental data. Nevertheless, these simulations indicate that the surface-area based model is insufficient for an accurate description of hydrophobic effects of surfactant micelles.