Determination of Amino Acids' pKa: Importance of Cavity Scaling within Implicit Solvation Models and Choice of DFT Functionals

Protonation states of molecules significantly influence the thermodynamics and kinetics of chemical reactions. This is especially important in biochemical processes, where appropriate protonation states of amino acids control the exo/endoergicity of practically all biochemical cycles. This paper is focused on appraisal of the impact of DFT functionals and PCM solvation models on the accuracy of pKa evaluations for all proteinogenic amino acids. Eight functionals (B3LYP, PBE0, revPBE0, M06-2X, M11, M11-L, TPSSh, and ωB97X-D) and four basis sets are considered, together with four kinds of implicit solvation models when additional attention is paid to a cavity construction. An influence of nonelectrostatic contributions and Wertz's corrections on Gibbs free energy is investigated together with accuracy of provided proton solvation energy. The best model is based on the M06-2X/6-311++G**/D-PCM/UAKS computational level. The fitting procedure is utilized to improve the accuracy of the evaluated models. All of these results are also compared with values obtained from the COSMOtherm program and CCSD(T) calculations. Results for cysteine and histidine are discussed individually, as they can be found in different protonation states at neutral pH.