Premium
Critical assessment of side‐chain conformational space sampling procedures designed for quantifying the effect of side‐chain environment
Author(s) -
Gautier R.,
Tufféry P.
Publication year - 2003
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.10334
Subject(s) - side chain , dihedral angle , conformational entropy , molecular dynamics , monte carlo method , entropy (arrow of time) , statistical physics , chemistry , computer science , computational chemistry , physics , mathematics , thermodynamics , molecule , hydrogen bond , statistics , organic chemistry , polymer
We introduce a family of procedures designed to sample side‐chain conformational space at particular locations in protein structures. These procedures (CRSP) use intensive cycles of random assignment of side‐chain conformations followed by minimization to determine all the conformations that a group of side‐chains can adopt simultaneously. First, we consider a procedure evolving in the dihedral space (dCRSP). Our results suggest that it can accurately map low‐energy conformations adopted by clusters of side‐chains of a protein. dCRSP is relatively insensitive to various important parameters, and it is sufficiently accurate to capture efficiently the constraint induced by the environment on the conformations a particular side‐chain can adopt. Our results show that dCRSP, compared with molecular dynamics (MD), can overcome the problem of the limited set of conformations reached in a reasonable amount of simulations. Next, we introduce procedures (vCRSP) in which valence angles are relaxed, and we assess how efficiently they quantify the conformational entropy of side‐chains in the protein native state. For simple peptides, entropies obtained with vCRSP are fully compatible with those obtained with a Monte Carlo procedure. For side‐chains in a protein environment, however, vCRSP appears of limited use. Finally, we consider a two‐step procedure that combines dCRSP and vCRSP. Our tests suggest that it is able to overcome the limitations of vCRSP. We also note that dCRSP provides a reasonable initial approximation. This family of procedures offers promise in quantifying the contribution of conformational entropy to the energetics of protein structures. © 2003 Wiley Periodicals, Inc. J Comput Chem 15: 1950–1961, 2003