Theoretical Determination of Activation Free Energies for Alkaline Hydrolysis of Cyclic and Acyclic Phosphodiesters in Aqueous Solution

Abstract

First-principles electronic structure calculations were performed in this study to examine the reaction pathway and corresponding activation free energies for alkaline hydrolysis of representative phosphodiesters, including dimethyl phosphate (DMP), trimethylene phosphate (TMP), ethylene phosphate (EP), and a simplified model (cAMPm) of adenosine 3', 5'-phosphate (cAMP). Reaction coordinate calculations show that for all of these phosphodiesters the alkaline hydrolysis follows a one-step bimolecular mechanism initialized by the attack of hydroxide ion at the phosphorus atom of the ester. Five self-consistent reaction field (SCRF) methods were used to calculate the activation free energies and the calculated results were compared with available experimental data. It has been shown that the results calculated using a recently developed SCRF method, known as the surface and volume polarization for electrostatics (SVPE) or fully polarizable continuum model (FPCM), which accurately determines both surface and volume polarization, are rather insensitive to the used solute charge isodensity contour value that determines the solute cavity size. The SVPE calculations plus nonelectrostatic interaction corrections led to activation free energies 32.6, 31.6, 24.8, and 29.4 kcal/mol for DMP, TMP, EP, and cAMPm, respectively. The calculated activation free energies are all in good agreement with available experimentally estimated activation free energies ∼32, ∼32, ∼21-24, and ∼29 kcal/mol for DMP, TMP, EP, and cAMP, respectively. The SVPE results show that the solvation dramatically decreases the activation free energies for the alkaline hydrolysis of phosphodiesters and strongly support the conclusion that the remarkable difference in the hydrolysis rate between DMP and EP is mainly due to the solvation, rather than the ring-strain. Compared to the SVPE results and available experimental data, an SCRF method that completely ignores volume polarization systematically overestimated the activation free energies, but the relative values of the calculated activation free energies are still in qualitative agreement with those of the SVPE results and available experimental data. The other three SCRF methods using a certain charge renormalization scheme also overestimated the activation free energies, and the relative values of the calculated activation free energies are all significantly different from those of the SVPE results and available experimental data.