"Chemical engineers have long sought enzymes as alternatives to traditional chemocatalytic routes as they are highly selective and have evolved to function under mild conditions (physiological temperature, neutral pH, and atmospheric pressure). Enzymes, the workhorses of biological chemistry, represent a vast catalogue of chemical transformations. This feature lends their use in a variety of industrial applications including food processing, biofuels, engineered biosynthetic pathways, and as biocatalysts for preparing specialty chemicals (e.g. pharmaceutical building blocks). The totality of an enzymatic bioprocess is a function of its catalytic efficiency (specificity and turnover), product profile (i.e. regio- and enantio-selectivity), and thermodynamic and kinetic stability. For native enzymes, these parameters are seldom optimal. Importantly, they can be modified using protein engineering techniques, which generally involves introducing mutation(s) to a protein sequence and screening for beneficial effects. However, robust enzyme engineering and design based on first principles is extremely challenging, as mutations that improve one parameter often yield undesired tradeoffs with one or more other parameters. In this thesis, deep mutational scanning - the testing of all possible single-amino acid substitutions of a protein sequence using high-throughput screens/selections and DNA counting via deep sequencing - was used to address two fundamental constraints on functional enzyme evolution. First, how do enzymes encode substrate specificity? To address this question, deep mutational scanning of an amidase on multiple substrates was performed using growth-based selections. Comparison of the resulting datasets revealed that mutations benefiting function on a given substrate were globally distributed in both protein sequence and structure. Additionally, our massive datasets permitted the most rigorous testing to date of theoretical models of adaptive molecular evolution. These results have implications for both design of biocatalysts and in understanding how natural enzymes function and evolve. Another fundamental constraint of enzyme engineering is that mutations improving stability (folding probability) of an enzyme are often inactivating for catalytic function, and vice versa. Towards overcoming this activity-stability constraint, I sought to improve the heterologous expression and maintain the catalytic function of a Type III polyketide synthase from Atropa belladonna. This was accomplished using deep mutational scanning and high-throughput GFP-fusion stability screening, followed by novel filtering methods to only accept beneficial mutations with high probability for maintaining function. Lastly, deep mutational scanning relies on the construction of user-defined DNA libraries, however current available techniques are limited by accessibility or poor coverage. To address these limitations, I will present the development of Nicking Mutagenesis, a new method for the construction of comprehensive single-site saturation mutagenesis libraries that requires only double-stranded plasmid DNA as input substrate. This method has been validated on several gene targets and plasmids and is currently being used in academic, government, and industry laboratories worldwide."--Pages ii-iii.
Read
