Methodology-independent free energies of changing charges

During the last decades, molecular simulation has established itself as a powerful tool to study molecular structure, thermodynamics and dynamics at a resolution often inaccessible to experiment. However, because of limitations in the computing power, simulated systems are of microscopic sizes and non-bonded interactions have to be treated in an approximate fashion. Especially the artifacts arising from the use of an approximate electrostatic interaction scheme can cause the outcome of a simulation to be dependent on the employed simulation methodology. For instance, it is well-known that computed free energies of charging are crucially affected by this issue. However, in the case of monoatomic ions, it is nowadays an established procedure to correct the raw simulation results (charging free energies) ex post so that methodological independence is achieved. On the contrary, the corresponding correction scheme has never been formulated for and applied in a consistent fashion to the solvation of polyatomic ions, thus preventing the obtension of reliable results in simulations involving the charging of polyatomic compounds. The two main consequences of this are that: (i) current (bio-)molecular force fields lack a sound thermodynamics-based parameterization of amino acid side chains carrying charged groups; (ii) it is currently not possible to perform accurate atomistic computer simulation studies of receptor-ligand binding reactions involving the comparison of ligands of different net charge. We plan to address these issues by providing and validating a scheme for the calculation of methodology-independent charging free energies for polyatomic ions in water and for charging processes occurring in a more complex environment (protein-ligand binding).