How well does microsolvation represent macrosolvation? A test case: Dynamics of decarboxylation of 4-pyridylacetic acid zwitterion

We present a systematic comparison of a cluster model to a continuum solvation model for the rate of decarboxylation of 4-pyridylacetic acid zwitterion. Geometries of the zwitterion and the transition state were optimized for the gas-phase reaction, for one model of reaction in aqueous solution in a previous study, and for five additional models of reaction in aqueous solution in the present study. The aqueous calculations include models in which one or two waters are treated explicitly, either without or with a surrounding continuum of bulk solvent; the latter is called a mixed discrete-continuum model. Bulk solvation effects are modeled both by solvation model SM5.42 and by a COSMO united atom solvation model; both of these models contain bulk electrostatics and short-range terms. Kinetic isotope effects were calculated by conventional transition state theory. We find that a discrete model of solvation with two explicit water molecules underestimates the free energy of activation for the decarboxylation reaction by at least 17 kcal/mol.