General Methodology to Optimize Damping Functions to Account for Charge Penetration Effects in Electrostatic Calculations Using Multicentered Multipolar Expansions

120

146

Classical potentials based on isotropic and additive atomic charges have been widely used to model molecules in computers for the past few decades. The crude approximations in the underlying physics are hindering both their accuracy and transferability across chemical and physical environments. Here we present a new classical potential, AMOEBA+, to capture essential intermolecular forces, including permanent electrostatics, repulsion, dispersion, many-body polarization, short-range charge penetration and charge transfer, by extending the polarizable multipole-based AMOEBA (Atomic Multipole Optimized Energetics for Biomolecular Applications) model. For a set of common organic molecules, we show that AMOEBA+ with general parameters can reproduce both quantum mechanical interactions and energy decompositions according to the Symmetry-Adapted Perturbation Theory (SAPT). Additionally, a new water model developed based on the AMOEBA+ framework captures various liquid phase properties in molecular dynamics simulations while remains consistent with SAPT energy decompositions, utilizing both ab initio data and experimental liquid properties. Our results demonstrate that it is possible to improve the physical basis of classical force fields to advance their accuracy and general applicability.

Section: Introductionmentioning

confidence: 99%

AMOEBA+ Classical Potential for Modeling Molecular Interactions

Liu

Piquemal

Ren

2019

120

146

“…There have been previous attempts to incorporate the charge penetration effect into implicit solvent models, − multipole-based electrostatic models, − charge-density-based (including Gaussian multipole) models, ,− and combined quantum and molecular mechanics (QM/MM) models. ,− Generally, the charge penetration correction involves breaking the atom-centered point charge into an effective core and a valence electron density, as suggested by Gordon et al and Piquemal et al In this way, the electrostatic interaction between two atoms is described as a sum of interactions between core and valence charge densities, which can be modeled with empirical exponential functions. Alternatively, rigorous integration over the two charge densities can be used to model short-range electrostatic interactions, ,, with a significantly greater expenditure of computational effort.…”

Section: Introductionmentioning

confidence: 99%

General Model for Treating Short-Range Electrostatic Penetration in a Molecular Mechanics Force Field

Wang¹,

Rackers

He³

et al. 2015

108

200

Classical molecular mechanics force fields typically model interatomic electrostatic interactions with point charges or multipole expansions, which can fail for atoms in close contact due to the lack of a description of penetration effects between their electron clouds. These short-range penetration effects can be significant and are essential for accurate modeling of intermolecular interactions. In this work we report parametrization of an empirical charge–charge function previously reported (PiquemalJ.-P.26313624J. Phys. Chem. A200310353) to correct for the missing penetration term in standard molecular mechanics force fields. For this purpose, we have developed a database (S101×7) of 101 unique molecular dimers, each at 7 different intermolecular distances. Electrostatic, induction/polarization, repulsion, and dispersion energies, as well as the total interaction energy for each complex in the database are calculated using the SAPT2+ method (ParkerT. M.09410624606352J. Chem. Phys.2014140). This empirical penetration model significantly improves agreement between point multipole and quantum mechanical electrostatic energies across the set of dimers and distances, while using only a limited set of parameters for each chemical element. Given the simplicity and effectiveness of the model, we expect the electrostatic penetration correction will become a standard component of future molecular mechanics force fields.

“…Therefore, we may say that the charge penetration effects are of both short and medium range. Including the charge penetration effect can significantly improve the description of the electrostatics in van der Waals interactions, and various procedures have been proposed to include this effect in molecular modeling. ,− For example, the charge penetration effects may be described by Gaussian-type functions or damping functions, and accurate electron densities, electrostatic potentials, and electrostatic interaction energies may be used for fitting parameters in these functions. The charge penetration effect has already been included in the development of some new force fields. ,, …”

Section: Introductionmentioning

confidence: 99%

Screened Electrostatic Interactions in Molecular Mechanics

Wang

Truhlar

2014

In a typical application of molecular mechanics (MM), the electrostatic interactions are calculated from parametrized partial atomic charges treated as point charges interacting by radial Coulomb potentials. This does not usually yield accurate electrostatic interactions at van der Waals distances, but this is compensated by additional parametrized terms, for example Lennard-Jones potentials. In the present work, we present a scheme involving radial screened Coulomb potentials that reproduces the accurate electrostatics much more accurately. The screening accounts for charge penetration of one subsystem's charge cloud into that of another subsystem, and it is incorporated into the interaction potential in a way similar to what we proposed in a previous article (J. Chem. Theory Comput. 2010, 6, 3330) for combined quantum mechanical and molecular mechanical (QM/MM) simulations, but the screening parameters are reoptimized for MM. The optimization is carried out with electrostatic-potential-fitted partial atomic charges, but the optimized parameters should be useful with any realistic charge model. In the model we employ, the charge density of an atom is approximated as the sum of a point charge representing the nucleus and inner electrons and a smeared charge representing the outermost electrons; in particular, for all atoms except hydrogens, the smeared charge represents the two outermost electrons in the present model. We find that the charge penetration effect can cause very significant deviations from the popular point-charge model, and by comparison to electrostatic interactions calculated by symmetry-adapted perturbation theory, we find that the present results are considerably more accurate than point-charge electrostatic interactions. The mean unsigned error in electrostatics for a large and diverse data set (192 interaction energies) decreases from 9.2 to 3.3 kcal/mol, and the error in the electrostatics for 10 water dimers decreases from 1.7 to 0.5 kcal/mol. We could have decreased the average errors further, but at the cost of sometimes significantly overestimating the screening; instead we chose a more conservative (safer) parametrization that systematically underestimates the screening (which by definition means it improves over point charges) and only occasionally overestimates it. Despite this conservative choice, we find that the screened MM method is even more accurate for the electrostatics than unscreened QM/MM calculations. This new method is easy to implement in any MM program, and it can be used to develop more physical force fields for molecular simulations.