Author

Youbin Tu

Date of Graduation

2008

Document Type

Thesis

Abstract

Cytochromes P450 are a large family of heme-thiolate enzymes which metabolize a wide variety of both endogenous and xenobiotic compounds. Human P450 1A1 and 1A2 are two major members of the P450 1A subfamily. They share 72% sequence identity but display different specificities towards various substrates. In order to further understand the structure-function relationships of these enzymes, we have performed several molecular modeling-based studies that examined various functional domains of P450 1As. In the first part of this project, we have investigated the ligand binding domain by developing a specificity prediction model based on the correlations between MD (Molecular Dynamics)-based descriptors and the specificity towards resorufin substrates of P450 1A2 WT and its five single mutants. The application of this method to multiple mutants of P450 1A2 containing mutations at two to five of these five key positions have led to the prediction that seven out of twenty six mutants would shift specificity from P450 1A2 to 1A1. The results have been verified by site directed mutagenesis and functional analysis of the mutant P450 enzymes. Subsequently, molecular modeling methods were applied for P450-CPR (Cytochrome P450 Reductase) docking, which helped to identify the key residues at the reductase binding sites of P450 1A1 and 1A2. The effect of the substitution of Asn-145 of P450 1A2 to Lys on substrate specificity and reductase binding have been evaluated using site-directed mutagenesis and kinetics assays. Finally, we conducted a preliminary study into the membrane anchoring domain by submerging P450 1A2 into the explicit membrane in two different orientations. Energy minimization and short molecular dynamics simulations have been performed to make it ready for the subsequent rigorous molecular dynamics simulations, which are expected to provide the plausible membrane association model for P450 1A2. For the majority of the modeling studies, we applied the first generation homology models of P450 1As as the starting line. We also developed a second generation homology models of P450 1A1 and 1A2 using a new strategy to optimize the active site using ligand-enzyme interaction information. These models were based on the structures of multiple eukaryotic P450s and NMR-derived T1 relaxation data using phenacetin and acetaminophen as substrates. Taken together, various molecular modeling techniques have been successfully employed to study several functional domains of P450 1As. They helped to investigate the structural basis of the enzyme function and gave insight into their mechanism of action.

Share

COinS