Premium
A comparison of quantum chemical models for calculating NMR shielding parameters in peptides: Mixed basis set and ONIOM methods combined with a complete basis set extrapolation
Author(s) -
Moon Seongho,
Case David A.
Publication year - 2006
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.20388
Subject(s) - oniom , basis set , extrapolation , isotropy , basis (linear algebra) , chemistry , limit (mathematics) , computational chemistry , electromagnetic shielding , set (abstract data type) , anisotropy , statistical physics , computational physics , mathematics , molecule , quantum mechanics , physics , density functional theory , mathematical analysis , computer science , geometry , organic chemistry , programming language
This article compares several quantum mechanical approaches to the computation of chemical shielding tensors in peptide fragments. First, we describe the effects of basis set quality up to the complete basis set (CBS) limit and level of theory (HF, MP2, and DFT) for four different atoms in trans N ‐methylacetamide. For both isotropic shielding and shielding anisotropy, the MP2 results in the CBS limit show the best agreement with experiment. The HF values show quite a different tendency to MP2, and even in the CBS limit they are far from experiment for not only the isotropic shielding of carbonyl carbon but also most shielding anisotropies. In most cases, the DFT values differ systematically from MP2, and small basis‐set (double‐ or triple‐zeta) results are often fortuitously in better agreement with the experiment than the CBS ones. Second, we compare the mixed basis set and ONIOM methods, combined with CBS extrapolation, for chemical shielding calculations at a DFT level using various model peptides. From the results, it is shown that the mixed basis set method provides better results than ONIOM, compared to CBS calculations using the nonpartitioned full systems. The information studied here will be useful in guiding the selection of proper quantum chemical models, which are in a tradeoff between accuracy and cost, for shielding studies of peptides and proteins. © 2006 Wiley Periodicals, Inc. J Comput Chem 27: 825–836, 2006