www.acsprf.org

The goal of this PRF funded project is to develop and apply a new method based  on molecular dynamics simulations to enable theoretical studies of pH-dependent properties of surfactant assemblies. Toward this goal, in the first grant year we  parameterized a generalized Born (GB) implicit-solvent model for efficient calculation of solvation energetics of surfactant molecules.  In the second grant year, we developed the continuous constant pH molecular dynamics (CpHMD) technique based on a hybrid-solvent scheme. In this scheme, we make use of the efficiency of the Generalized-Born (GB) implicit-solvent model for estimating the free energy of protein solvation, while propagating conformational dynamics using the more accurate explicit-solvent model. Furthermore, we employ a pH-based replica exchange algorithm to significantly enhance both protonation and conformational state sampling. Benchmark titration simulations for five proteins of various sizes yield an average absolute deviation of 0.53 and a root mean squared deviation of
0.74 from experimental pKa's. This level of accuracy is obtained with 1-ns simulations per replica. This work has led to a publication with a graduate student as the first author (Wallace and Shen, J Chem Theory Comput 2011, 7, 2617-2629).

We then applied the hybrid-solvent CpHMD simulations to predict and dissect the microscopic origins of the pKa's of a single surfactant solubilized in ionic and nonionic micelles, which is of interest to detergent industry and oil refinery. Calculation of surfactant pKa's in micelles is a challenging task using
traditional electrostatic methods due to the lack of structure data and knowledge of an effective dielectric constant. We test the implicit- and explicit-solvent based continuous constant pH molecular dynamics (CPHMD) methods for predicting the pKa shift of a lauric acid solubilized in three micelles: dodecyl sulfate (DS), dodecyl trimethylammonium (DTA), and dodecyl triethylene glycol ether (DE3). Both types of simulations were able to reproduce the positive pKa shifts for the anionic DS and nonionic DE3 micelles. However, for the cationic DTA micelle, the implicit-solvent simulation failed to predict the direction of the pKa shift, while the explicit-solvent result is consistent with experiment, although the specific-ion effects remain to be accurately determined. Comparison between the implicit- and explicit-solvent data shows that the latter gives a more realistic description of the conformational environment of the titrating probe. Surprisingly, in the DTA micelle, surfactants are only slightly attracted to the laurate ion, which diminishes the magnitude of electrostatic stabilization, giving rise to a positive pKa shift that can not be explained by chemical intuition or other theoretical models. Our data underscores the importance of microscopic models and ionization-coupled conformational dynamics in quantitative prediction of pKa shifts in micelles. This work has led to a manuscript with two graduate students and a postdoc.