Applications of Atomistic Modeling in Drug Design

Structure based computer aided drug design has been proven as a potent tool in identifying lead molecules for drug targets, with the increasing number of X-ray structures, validated drug targets and the ever increasing computer power. In my thesis, I have applied the structure based drug design for multiple target families including antibiotics targets (enzymes), transcription factors and epigenetic readers. The computational techniques I have used to study these systems are mainly docking, molecular dynamics simulation, MMGBSA binding free energy calculation and homology modeling. These techniques are fast, cost efficient and can provide mechanisms of binding.

My major project is to develop a blueprint for antimicrobial hit discovery targeting metabolic networks. To illustrate this methodology, E. coli & S. aureus were used as examples for which multiple inhibitors have been discovered. System-level identification of drug targets was achieved through metabolic network construction of E. coli & S. aureus. This is followed by atomistic modeling of small molecules capable of modulating the targets' activity and in vitro enzyme and cell viability assays to test the activities of the predicted inhibitors. This project combines genome analysis, network biology and computational chemistry, which provides a general strategy for strain-specific anti-infective therapy. My contribution to this project includes both computational prediction of potential inhibitors and experimental validation of their activities in both in vitro and bacterial cell viability.

My second project involves designing peptide analogs to disrupt the protein-protein interactions of the NOTCH transactivation complex, the loose or gain-of-function of mutations of which are related with various diseases, including cancer. I have identified binding hot-spots of MAML & CSL/ANK interface through MMGBSA energy decomposition. Stapled peptide analogs of natural MAML are then computationally designed with mutations around the hot-spots to other residues or non-natural amino acids using MD simulations and MMGBSA rescoring, which have been subsequently validated experimentally by our collaborators to have stronger potency.

My third project is studying a small molecule inhibitor Ì¢‰âÂ' (+)-JQ1 targeting bromodomains for potential cancer therapy. Using docking and molecular dynamics simulation, I explained the selectivity of BRD4(1) over JQ1 enantiomers and the different binding modes of BRD3(1) and BRD3(2) with (+)-JQ1.