The results presented in this thesis detail efforts to apply computational chemistry techniques to investigate larger organic and biomolecular systems. Due to recent advances in computational hardware, as well as computational chemistry software, highly accurate quantum chemical methods can now be applied successfully to the treatment of larger and larger systems, up to hundreds of atoms in size. In conjunction with more computationally efficient molecular mechanics methods, it is now possible to perform detailed computational investigations of molecular conformations and reaction mechanims in silico.
The reactions of a variety of halogenated benzenes with a model system of the cytochrome P450 active site, known as compound I, were examined using DFT methods. Reaction pathways were identified which lead to experimentally observed products, and transition states were characterized which were consistent with the experimentally observed oxidation products. Further, the relative transition state barriers for oxidation of mixed fluorinated and chlorinated rings suggested that the former were more reactive, with possible implications in pharmeutical design.
Similar mechanistic studies using DFT were performed to analyze dendridic and protodendritic systems. The proposed catalytic cycle for the yttrium-catalyzed acyl transfer reaction, a reaction which has been demonstrated to proceed with moderate-to-high degrees of enantioselectivity, was validated using DFT methods. The reaction pathway was examined, and diastereomeric transition states were characterized which matched well with experimental stereoselectivities.
In the biological milieu, simulations were performed to analyze human paraoxonase I (huPON1), with an aim toward understanding its structure, and interactions of the protein with OP substrates. Initially, molecular dynamics simulations (MD) were performed to ‘relax’ the geometries of the protein from an available X-ray crystal structures, prior to computational binding studies to analyze orientations of ligands in the active sites of the proteins. Refinement of these binding modes was performed using MD simulations, to obtain free energies of binding for these substrates with the protein; these were compared with experiment and showed strong correlations with observed KM values. Analysis of these poses suggested a consistent binding mode for OP substrates which was consistent with mutagenesis studies, although the mechanism of the protein remained elusive.
Additional work on huPON1 involved a more detailed structure-activity relationship study, performed in collaboration with the Magliery group, of the interactions of the protein with a variety of analogs of the substrate paraoxon. This was performed to examine the limits of substrate tolerance in the enzyme active site, and attempt to gain additional information regarding the catalytic mechanism. The results of this study suggest that the active site possesses significant steric limitations, and molecular modeling results have produced some possible rationalizations for these results with implications for the design of protein mutants with enhanced activities against a broad range of substrates. Additional work related to the stereochemical preference of the enzyme, and possible avenues to modulate thereof, are an ongoing area of study.
An additional project involved the use of a combination of ground and excited state calculations, which were used to examine azobenzene oligomers which demonstrated photomodulated chiral transfer in solution. A detailed conformational study of these molecules, using both Monte Carlo and Replica Exchange Molecular Dynamics (REMD) was able to identify the minimum energy conformations of a variety of oligomers, as well as a potential avenue for the low-energy interconversion of helical sense. Following characterization of the solution phase conformation of these oligomers, the UV-vis and circular dichroism (CD) spectra of the molecules were predicted using excited state calculations, and employed to validate the solution-phase conformation of the molecules. The good agreement between experimental and theoretical spectra was additional support for the notion that amplificiation of chiral bias was occuring in the azobenzene oligomers as a function of size, a key hypothesis of the study.
Overall, the work described in this dissertation represents several approaches to the study of larger organic and biological systems using computational methods, with an overall aim of obtaining structural and/or mechanistic information from which we may verify theoretical methods, and predict future results.