Premium
Importance of explicit smeared lone‐pairs in anisotropic polarizable molecular mechanics. Torture track angular tests for exchange‐repulsion and charge transfer contributions
Author(s) -
El Khoury Léa,
NaseemKhan Sehr,
Kwapien Karolina,
Hobaika Zeina,
Maroun Richard G.,
Piquemal JeanPhilip,
Gresh Nohad
Publication year - 2017
Publication title -
journal of computational chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.907
H-Index - 188
eISSN - 1096-987X
pISSN - 0192-8651
DOI - 10.1002/jcc.24830
Subject(s) - polarizability , chemistry , lone pair , delocalized electron , context (archaeology) , anisotropy , computational chemistry , molecule , atom (system on chip) , separable space , plane (geometry) , atomic physics , chemical physics , molecular physics , physics , quantum mechanics , geometry , mathematics , mathematical analysis , computer science , paleontology , organic chemistry , biology , embedded system
A correct representation of the short‐range contributions such as exchange‐repulsion ( E rep ) and charge‐transfer ( E ct ) is essential for the soundness of separable, anisotropic polarizable molecular mechanics potentials. Within the context of the SIBFA procedure, this is aimed at by explicit representations of lone pairs in their expressions. It is necessary to account for their anisotropic behaviors upon performing not only in‐plane, but also out‐of‐plane, variations of a probe molecule or cation interacting with a target molecule or molecular fragment. Thus, E rep and E ct have to reproduce satisfactorily the corresponding anisotropies of their quantum chemical (QC) counterparts. A significant improvement of the out‐of‐plane dependencies was enabled when the sp 2 and sp localized lone‐pairs are, even though to a limited extent, delocalized on both sides of the plane, above and below the atom bearer but at the closely similar angles as the in‐plane lone pair. We report calibration and validation tests on a series of monoligated complexes of a probe Zn(II) cation with several biochemically relevant ligands. Validations are then performed on several polyligated Zn(II) complexes found in the recognition sites of Zn‐metalloproteins. Such calibrations and validations are extended to representative monoligated and polyligated complexes of Mg(II) and Ca(II). It is emphasized that the calibration of all three cations was for each Δ E contribution done on a small training set bearing on a limited number of representative N, O, and S monoligated complexes. Owing to the separable nature of Δ E , a secure transferability is enabled to a diversity of polyligated complexes. For these the relative errors with respect to the target Δ E (QC) values are generally < 3%. Overall, the article proposes a full set of benchmarks that could be useful for force field developers. © 2017 Wiley Periodicals, Inc.