| This dissertation represents an effort in the expanding field of metabolic engineering. Designing cellular phenotypes using metabolic engineering principles for the production of natural products is becoming an increasingly critical step in the next generation of commercial bioprocesses. To meet the engineering challenges involved with such endeavors, new tools need to be developed and evolved before efficient strains can be realized, and it is in this vein that I present the development of two new engineering methodologies.;First, as a result of the complexity and scope of cellular networks, novel approaches must focus on analyzing and interpreting the systemic properties of the network behavior. This has led me to develop a computational tool that takes advantage of an optimized genetic algorithm to create a framework for predicting gene deletions. This routine overcomes many of the challenges imposed by stoichiometric modeling approaches and is termed Cipher of Evolutionary Design (CiED). The development of this novel computational algorithm is discussed in Chapter 2 and is followed by its application in Chapter 3. Specifically, CiED is used to investigate the impact of gene deletions and other network modifications on the metabolite profile of the recombinant microorganisms Escherichia coli and Saccharomyces cerevisiae in an effort to find optimal phenotypes for the production of high-value end products, such as recombinant natural products. In this work, I have shown the model-predicted metabolic behaviors are consistent experimentally and led to strain designs (genotypes) not identifiable from simple network inspection. Overall, this study provides additional insights to the continuing effort to better understand cellular physiology and to codify such understanding in mathematical models whose ability to predict behavior under a variety of conditions is increasingly challenging.;Second, while protein engineering has led to the development of novel therapies and biosynthetic platforms, complexities not elucidated by rational protein design can be identified using libraries of mutant proteins. Directed evolution experiments can 'train' proteins to achieve novel and improved activities as long as the screening of evolved mutant libraries is efficient and focuses towards the engineering objective. Chapter 4 discusses the ability of both fungal and bacterial laccases, a class of phenoloxygenases, to use flavonoids as substrates. Knowledge gained in this study is then applied in Chapter 5 toward a novel screening method used for the evolution of a stilbene synthase, the enzyme responsible for biosynthesis of resveratrol. This work illustrates that a detailed understanding of the protein structure and function can lead to more focused directed evolution experiments. Furthermore, the novel screening method developed allowed the identification of mutant enzymes with random point residue changes resulting in altered biochemistry and leading to improved natural product production.;As a whole, this research will strengthen the intellectual foundations of metabolic engineering, thus increasing the confidence with which microbial pathways can be modulated for the purpose of product overproduction. This dissertation also adds to the array of engineering methods available for the production of valuable natural products with pharmaceutical and nutraceutical potential. |
