A new ECEPP/3 package has been developed for protein modeling and docking. The goal was to define an arbitrary level of molecular flexibility for the ligand and the receptor while maintaining reasonable computational cost. This was accomplished by defining the protein system in dihedral angle space and dividing it into flexible and rigid protein segments. The computational time is further reduced by incorporating the energetic field of the rigid part into a grid. A new interpolation scheme based on Bezier splines has also been developed to allow energy-minimization using this grid. A biased Monte Carlo-Energy-Minimization procedure using information on the normal modes of vibration of the protein complex has been implemented as a global optimization tool. This method has been shown to be successful for docking a small ligand into its receptor. Application of this program to the calculation of the structure of the complex of residues 7-20 of the fibrinogen A-alpha chain with bovine alpha-thrombin using NMR data has been carried out. Calculation of the structure of the bound ligand has been realized in a single step by sampling the ligand conformation in the presence of the receptor using a chain-growing procedure. The fibrinogen as well as the active site residues were defined as flexible. The NMR constraints were optimized along with the ECEPP/3 energy function. The calculated set of structures shows structural variability that results from the broad specificity cavity of thrombin beyond the cleavage region. This broad specificity cavity provides new possibilities for antagonist-receptor interactions. The structure of residues 7-20 of fibrinogen A-alpha also reveals some insights about the docking of the complete N-terminal portion of fibrinogen onto the thrombin receptor. Further improvement of this project is the incorporation of new global optimization tools that have been developed recently in our laboratory. The goal is to achieve high reliability (around 80%) of finding the global energy minimum when docking a flexible drug molecule into its receptor. Among these methods are """"""""Conformational Space Annealing"""""""" based on a Genetic Algorithm which has been shown to provide the global minimum of Leu-enkephalin in 100% of the cases in less than 2700 energy minimizations compared to 3000 with 50% chance of success in our study. The second is the """"""""Distance Scaling Method"""""""" that sucessfully finds the global minimum of the 20-residues polyalanine. The incorporation of the MERK force field is also underway in order to apply this package to a drug-design oriented project such as the characterization of new antagonists molecules of Ribonuclease.
Showing the most recent 10 out of 20 publications
