Parameterization of the Hamiltonian Dielectric Solvent (HADES) Reaction‐Field Method for the Solvation Free Energies of Amino Acid Side‐Chain Analogs

Abstract

Optimization of the Hamiltonian dielectric solvent (HADES) method for biomolecular simulations in a dielectric continuum is presented with the goal of calculating accurate absolute solvation free energies while retaining the model's accuracy in predicting conformational free-energy differences. The solvation free energies of neutral and polar amino acid side-chain analogs calculated by using HADES, which may optionally include nonpolar contributions, were optimized against experimental data to reach a chemical accuracy of about 0.5 kcal mol(-1). The new parameters were evaluated for charged side-chain analogs. The HADES results were compared with explicit-solvent, generalized Born, Poisson-Boltzmann, and QM-based methods. The potentials of mean force (PMFs) between pairs of side-chain analogs obtained by using HADES and explicit-solvent simulations were used to evaluate the effects of the improved parameters optimized for solvation free energies on intermolecular potentials.

Pitfall in Free‐Energy Simulations on Simplest Systems

“…Nonetheless, the conventional free‐energy profile in the literature uses the probability density function without dividing the Dirac delta function by a Jacobian scale factor [e. g., Eq. (25)], which equivalently considers contributions from configuration space only (Table ) . We call the related quantities as unconstrained free‐energy profile , unconstrained probability density function , and unconstrained absolute partition function , because all can be directly acquired via unconstrained (and restrained / biased ) free‐energy simulations, e. g., through umbrella samplings with histogram binning.…”

“…(83)]. We determine that we should use the (mass‐scaling dependent ) constrained for TST [in contrast to the conventional practice that uses the (mass‐scaling independent ) unconstrained ,−; and thus regardless of the definition of the reaction coordinate, neither energy barrier nor reactant and transition states exist for a free‐body system. This identical conclusion that we should use the constrained ( not the unconstrained ) for TST has also been made by Ref …”

“…Note that using the definition of the gradient operator in Eq. (7), which explicitly involves individual Jacobian scale factors, we can derive other standard differentiation operators, e. g., divergence and Laplacian operators.…”

“…A compelling ability of molecular simulations [e. g., molecular dynamics (MD) and Monte Carlo (MC) simulations] that attracts more and more researchers across many different disciplines in science and engineering to use is simulating a free‐energy profile (which is sometimes also known as “potential of mean force”, PMF) that quantitatively describes a molecular event along a reaction coordinate (i. e., the coordinate of interest; Figure ) . This kind of free‐energy simulations provides a powerful computational approach for complementing experiments in understanding the mechanisms behind molecular processes, e. g., drug bindings, aggregation of nanoparticles driven by lipid, chemical reactions that happen on material surfaces and in protein or RNA enzymes …”

“…Nonetheless, we feel that according to the literature, people (including us) conventionally think unconstrained free‐energy profiles are “superior” to the constrained ones ,. For example, people often use the Fixman term for “correcting” constrained free‐energy profiles to unconstrained ones; not the other way around .…”