| A novel approach for structure based ligand design is presented that can account for ligand induced conformational changes in the protein target. The method uses molecular dynamics simulations to sample energetically favorable conformations of the protein and the small functional groups that will form the building blocks of larger ligands. The resulting data are used as input to a Monte Carlo algorithm that forms potential ligands from the functional group positions.; The method designs ligands in a stepwise manner. First, the Multiple Copy Simulatneous Search (MCSS) method is extended to allow for a flexible protein. Many copies of a functional group are randomly placed in the binding site of the protein target. Energetically favorable positions and orientations of small functional groups are determined using quenched molecular dynamics. During the simulation, residues in the binding site are allowed to make conformational changes in response to the field of the functional group copies. Several different quenched molecular dynamical protocols are examined and discussed.; Once the functionality maps are obtained, the new functional group positions are linked together using the Monte Carlo algorithm, Dynamic Ligand Design (DLD). The DLD approach reformulates the ligand design problem into an optimization problem. A pseudo-potential energy function is defined such that finding low energy states on the DLD pseudo-potential surface yields potential ligands with favorable geometry that are complementary to the binding site. In this work, the DLD methodology is extended and a novel simulated annealing protocol is presented for the optimization of functions defined on spaces of high dimensionality. The methods are sufficiently general such that they are expected to be applicable to other optimization problems. The utility of the method is demonstrated by designing novel inhibitors to Endothiapepsin. In a separate set of calculations, the algorithm is used to suggest novel peptide, organic hybrid RGD tripeptides.; The combined molecular dynamics, Monte Carlo approach is applied to the viral protein HIV-1 protease. Novel inhibitors are suggested and the relative binding affinity of each is assessed. |
