Protein Folding with Coarse-Grained Off-Lattice Models of the Polypeptide Chain
A hierarchical approach, together with the United Residue (UNRES) model of the polypeptide chain, is used to study protein structure prediction. First, an efficient method has been developed as an extension of the hierarchical approach for packing alpha-helices in proteins. The results for 42 proteins show that the approach reproduces native-like folds of alpha-helical proteins as low-energy local minima. Moreover, this technique successfully predicted the structure of the largest protein obtained so far with the UNRES force field in the sixth Critical Assessment of Techniques for Protein Structure Prediction (CASP6). Next, two popular methods of global optimization are coupled, and the performance of the resulting method is compared with that of its components and with other global optimization techniques. The Replica-Exchange Method together with Monte Carlo-Minimization (REMCM) was applied to search the conformational space of coarse-grained protein systems described by the UNRES force field. In summary, REMCM located global minima for four proteins faster and more consistently than two of three other global optimization methods, while being comparable to the third method used for comparison. Finally, efficient methods for calculating thermodynamic averages were implemented with the UNRES force field, namely a Replica Exchange method (REM), a Replica Exchange Multicanonical method (REMUCA), and Replica Exchange Multicanonical with Replica Exchange (REMUCAREM), in both Monte Carlo (MC) and Molecular Dynamics (MD) versions. The algorithms were applied to one peptide and two small proteins (with alpha-helical and alpha+beta topologies). To compare the different methods, thermodynamic averages are calculated, and it is found that REM MD has the best performance. Consequently, free energy maps are computed with REM MD, to evaluate the folding behavior for all test systems.
This work was supported by National Science Foundation (NSF) and National Institutes of Health (NIH). Support was also received from the National Foundation for Cancer Research. This research was carried out by using the resources of our 392-processor Beowulf cluster at the Baker Laboratory of Chemistry and Chemical Biology, Cornell University, the National Science Foundation Terascale Computing System at the Pittsburgh Supercomputer Center, and the National Center for Supercomputing Applications System at the University of Illinois at Urbana-Champaign.
1. National Academy of Sciences, U.S.A.; 2. John Wiley & Sons Inc,
Global Optimization; Free Energy; Multiple-Minima Problem; Generalized Ensemble Methods; Secondary Structure Packing; Structure Prediction
Previously Published As
1. Nanias, M.; Chinchio, M.; Pillardy, J.; Ripoll, D.R.; Scheraga, H.A., Proc. Natl. Acad.Sci. USA, 2003, 100, 1706. 2. Nanias, M.; Chinchio, M.; Oldziej, S.; Czaplewski, C.; Scheraga, H.A., J. Comp. Chem., 2005, 26, 1472.
dissertation or thesis