Computational Enzyme Design

Author: Sosa Padilla Araujo, Bernardo

Year: 2014

Degree: Dissertation (Ph.D.)

Advisors: Mayo, Stephen L.; Miller, Thomas F.

Committee Members: Shan, Shu-ou; Mayo, Stephen L.; Miller, Thomas F.; Clemons, William M.

Option: Biochemistry and Molecular Biophysics

DOI: 10.7907/Z9KP805N

Abstract

Computational protein design (CPD) is the automated identification of amino acid sequences that will fold into a specified three-dimensional structure. This method has emerged as a promising tool for engineering enzymes. In this thesis, I describe my efforts at improving and applying CPD methods to the design of enzymes.

Chapter II describes the development and benchmark results for a molecular dynamics (MD) protocol to prescreen enzyme designs. Results indicate that the MD protocol is successful in screening enzymes for enzymatic activity. The protocol is general, reproducible, and excels at discarding false positive designs while predicting the most active enzymes correctly.

Conformational changes are part of the repertoire that natural enzymes use to catalyze reactions. Chapter III comprises the computational design and experimental characterization of triosephosphate isomerase (TIM), a model enzyme for the study of catalytic activity in the context of conformational changes. Using a novel multi-state design (MSD) method that can consider multiple states in a single protein sequence optimization calculation, we designed the flexible hinges in TIM’s active site loop.

A central challenge in the CPD field is the reliable engineering of an enzyme for any desired reaction. In Chapter IV, I describe the conceptualization and application of a high-throughput computational framework for the de novo design of enzymatic activity into inert scaffolds. TIM is used as a model system.

Files