Pairwise energies for polypeptide coarse-grained models derived from atomic force fields

被引:26
作者
Betancourt, Marcos R. [1 ]
Omovie, Sheyore J. [1 ]
机构
[1] Indiana Univ Purdue Univ, Dept Phys, Indianapolis, IN 46202 USA
关键词
biochemistry; biology computing; Boltzmann equation; knowledge based systems; molecular biophysics; molecular dynamics method; polymers; proteins; PROTEIN-STRUCTURE SIMULATIONS; QUASI-CHEMICAL APPROXIMATION; AMINO-ACIDS; STATISTICAL POTENTIALS; BIOMOLECULAR SYSTEMS; STRUCTURE PREDICTION; GLOBULAR-PROTEINS; CONTACT ENERGIES; CONFORMATIONS; REPRESENTATION;
D O I
10.1063/1.3137045
中图分类号
O64 [物理化学(理论化学)、化学物理学];
学科分类号
070304 ; 081704 ;
摘要
The energy parametrization of geometrically simplified versions of polypeptides, better known as polypeptide or protein coarse-grained models, is obtained from molecular dynamics and statistical methods. Residue pairwise interactions are derived by performing atomic-level simulations in explicit water for all 210 pairs of amino acids, where the amino acids are modified to closer match their structure and charges in polypeptides. Radial density functions are computed from equilibrium simulations for each pair of residues, from which statistical energies are extracted using the Boltzmann inversion method. The resulting models are compared to similar potentials obtained by knowledge based methods and to hydrophobic scales, resulting in significant similarities in spite of the model simplicity. However, it was found that glutamine, asparagine, lysine, and arginine are more attractive to other residues than anticipated, in part, due to their amphiphilic nature. In addition, equally charged residues appear more repulsive than expected. Difficulties in the calculation of knowledge based potentials and hydrophobicity scale for these cases, as well as sensitivity of the force field to polarization effects are suspected to cause this discrepancy. It is also shown that the coarse-grained model can identify native structures in decoy databases nearly as well as more elaborate knowledge based methods, in spite of its resolution limitations. In a test conducted with several proteins and corresponding decoys, the coarse-grained potential was able to identify the native state structure but not the original atomic force field.
引用
收藏
页数:11
相关论文
共 59 条
[41]   Decoys 'R' Us: A database of incorrect conformations to improve protein structure prediction [J].
Samudrala, R ;
Levitt, M .
PROTEIN SCIENCE, 2000, 9 (07) :1399-1401
[42]   Predicting interresidue contacts using templates and pathways [J].
Shao, Y ;
Bystroff, C .
PROTEINS-STRUCTURE FUNCTION AND BIOINFORMATICS, 2003, 53 :497-502
[43]   Statistical potential for assessment and prediction of protein structures [J].
Shen, Min-Yi ;
Sali, Andrej .
PROTEIN SCIENCE, 2006, 15 (11) :2507-2524
[44]   Multiscale methods for macromolecular simulations [J].
Sherwood, Paul ;
Brooks, Bernard R. ;
Sansom, Mark S. P. .
CURRENT OPINION IN STRUCTURAL BIOLOGY, 2008, 18 (05) :630-640
[45]   CALCULATION OF CONFORMATIONAL ENSEMBLES FROM POTENTIALS OF MEAN FORCE - AN APPROACH TO THE KNOWLEDGE-BASED PREDICTION OF LOCAL STRUCTURES IN GLOBULAR-PROTEINS [J].
SIPPL, MJ .
JOURNAL OF MOLECULAR BIOLOGY, 1990, 213 (04) :859-883
[46]   BOLTZMANN PRINCIPLE, KNOWLEDGE-BASED MEAN FIELDS AND PROTEIN-FOLDING - AN APPROACH TO THE COMPUTATIONAL DETERMINATION OF PROTEIN STRUCTURES [J].
SIPPL, MJ .
JOURNAL OF COMPUTER-AIDED MOLECULAR DESIGN, 1993, 7 (04) :473-501
[47]  
Skolnick J, 1997, PROTEIN SCI, V6, P676
[48]   Prediction of protein-protein interactions by docking methods [J].
Smith, GR ;
Sternberg, MJE .
CURRENT OPINION IN STRUCTURAL BIOLOGY, 2002, 12 (01) :28-35
[49]   Medium- and long-range interaction parameters between amino acids for predicting three-dimensional structures of proteins. [J].
TANAKA, S ;
SCHERAGA, HA .
MACROMOLECULES, 1976, 9 (06) :945-950
[50]   Statistical potentials extracted from protein structures: How accurate are they? [J].
Thomas, PD ;
Dill, KA .
JOURNAL OF MOLECULAR BIOLOGY, 1996, 257 (02) :457-469