Computational Chemistry

My post-doctoral reserach at sanofi-aventis involves understanding the mechanism behind how the the body metabolizes small molecules. The cytochrome P450 family of enzymes are mainly responsible for the processing of drug-like molecules. They are a complex and promiscious family - one molecule can be modified by one or many of these enzymes. The need to understand and predict how these enzymes modify potential drugs is vital to the pharmaceutical industry.

My research in the Jorgensen Laboratory involved the use of free energy perturbations and linear response techniques to understand several organic and biomolecular systems. My main focus of research was the computational design of novel HIV-1 non-nucleoside reverse transcriptase inhibitors (NNRTIs). Research projects I was involved in are:

Many single point mutations degrade the efficacy of NNRTIs however etravirine maintains excellent potency against eighteen common single and double mutants. The relative fold change of etravirine and three related analogs with varying activity against HIV-1 Reverse Transcriptase (HIVRT) are examined via the Monte Carlo/Free Energy Perturbation (MC/FEP) methodology to determine the influence of the L100I and L100I + K103N mutation in the binding site. Anaysis of the structural results provide insight into the structural reasoning for the superior potency of etravirine. The design of novel inhibitors incorporating these structural aspects are then synthesised and assayed.

Malaria is a serious, world-wide health concern. Inhibitors of a potential non-active site on the Plasmodium falciparum Thymidylate Synthase - Dihydrofolate Reductase (TS-DHFR) has been pursued. A library of ca. 16,000 drug-like Maybridge compounds was used in a virtual screen of this potential non-active site. Four moderate non-active site inhibitors of TS-DHFR have been found utilizing in silico screening coupled with a steady-state spectroscopic assay.

Free Energy Perturbations (FEPs), a rigorous technique in computational chemistry, was used to determine the relative pKa values of several organic molecules. The basicity of a series of substituted pyridine molecules in a 50% aqueous ethanol solution was examined using Density Functional Theory (DFT) and MC/FEP calculations. Gas-phase basicities and relative pKa’s agree excellently with the experimental results and predict that ditert-butylpyridine(DTBP) is less basic than tert-butylpyridine and methylpyridine. By including an explicit representation of solvent molecules in the calculations, steric effects on pKa could be reproduced and analyzed. Hydrogen bond analysis, energy pair distributions, and gas phase optimizations confirm that DTBP can not participate in traditional linear hydrogen bonds due to the large steric bulk of the tert-butyl groups rather they form hydrogen bonds with the pyridine p cloud. The interaction of DTBPH+ with water is reduced when compared with other protonated pyridines.

Monte Carlo statistical mechanics simulations were used in combination with the extended linear response (ELR) approach to develop a model to predict the activities of neurotrophic inhibitors of the FK506 binding protein. The individual cores of the inhibitors are analyzed first and an ELR model using a minimal number of descriptors were obtained for each core. The two datasets will be combined to yield a general ELR model for FKBP inhibition.