Fast Simulation Protocol for Protein Structural Transitions: Modeling of the Relationship of Structure and Function
- PDF / 1,150,244 Bytes
- 13 Pages / 612 x 792 pts (letter) Page_size
- 87 Downloads / 224 Views
V3.5.1
Fast Simulation Protocol for Protein Structural Transitions: Modeling of the Relationship of Structure and Function Arun K. Setty and D. M. Zuckerman Center for Computational Biology and Bioinformatics and Department of Environmental and Occupational Health, Graduate School of Public Health, University of Pittsburgh
ABSTRACT An approach is developed to study the dynamics of protein conformational transitions in depth. A computational (Monte Carlo) approach based on a united residue model is used. Unbiased transitions between the Apo and Holo conformations/states of calmodulin are observed at the rate of 1 per day per processor. A series of models of increasing complexity is studied, accounting for hydrophobic interactions and calcium binding. Details of the transitional region and structural information about intermediate states are obtained. Statistically converged ensembles of transitions are obtained in a reasonable real time period. I. INTRODUCTION Proteins function by means of conformational transitions between two or more native configurations. Complete knowledge of the pathways traversed is essential to understand the dynamics, the mechanism of the process under consideration, and to identify long-lived intermediates (which could potentially be verified experimentally in the future). Further, the selectivity (or in the case of calmodulin, versatility) of the protein with regard to the number of targets that it can bind to can be understood from an examination of the pathways explored by it. Due to the enormous computational cost involved, most protein studies have determined either a single, prominent pathway [1-3] or, by using guiding forces [4-5] a few plausible pathways. In Ref.[6] a simple model was presented for which the entire ensemble could be recovered in a reasonable time period. This model, involving a generalization of the Go model [7-8], deliberately creates energy minima in the phase space corresponding to the native states. In this article, the Go potential is largely replaced by a physically motivated hydrophobic potential (an implicit solvent model) [9-10] and by calcium-binding interactions. As shall be shown later in this article, the results still show good general agreement with experiment (in a few cases, there is a significant improvement). Further, there is no dramatic increase in the computational time requirements as a result of the refinement in the potentials used. The results from the three models will be presented and compared. The system under study in this article is calmodulin [11-13], an extremely common cellular protein, which is phylogenetically highly conserved and involved in calcium signaling; it has multiple roles in various functions, including neuronal gene expression, smooth muscle contraction, and mitosis. Calmodulin is a regulatory role in cell proliferation and hormonal secretion [14-15]. Monitoring
V3.5.2
Ca2++ localization in living cells has been done using a bioengineered form of calmodulin. Calmodulin is a small (148 residue; Fig. 1) protein with
Data Loading...
