Year: 2016
Type: dissertation
Abstract: Proteins propagate information within a cell by interacting with other proteins, DNA, RNA, and small molecules, enabling a cell to adapt to its environment and to maintain normal cellular functionality. Direct manipulation of a cell's protein interaction network can be achieved by using synthetically designed proteins to replace existing interactions or to create new interactions with other cellular components. A major challenge to engineering new synthetic proteins is the inability to search over the massive space of possible protein variants to identify candidates that bind the biological target of interest. In this thesis, I evaluate existing search strategies and propose a new search strategy for the computational design of proteins, which restricts the search space to one that is enriched for protein variants that tightly bind a protein target. I first evaluated current computational protein design strategies on two biological systems. Using the Fyn SH3 domain bound to different peptide targets, I assessed the design strategies' ability to recover the amino acid variability of the domain core as experimentally determined by phage display. Next, I evaluated the design methods by their ability to engineer the ubiquitin interface for tightly bind towards the USP21 protein. For this assessment, a new parallel validation strategy is introduced. In both biological systems, I found that ensemble protein design strategies outperformed other protein design methods. A limiting factor in current protein design strategies was found to be the accurate creation of alternative protein backbones when a single protein structure was used as a template. I present a flexible backbone design strategy, GP backbone design, that incorporates information from multiple protein structures. I show that using multiple structures to learn a space of possible protein movements increases the number of protein variants that tightly bind a protein target.
Source: TSpace (University of Toronto)
Author: Mark Sun
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W2784781481