Caltech Library System

About | Browse | Search | Caltech Student Instructions

Type of Document	Dissertation
Author	Joern, John M.
Author's Email Address	joern AT cheme.caltech.edu
URN	etd-02042003-160317
Persistent URL	http://resolver.caltech.edu/CaltechETD:etd-02042003-160317
Title	Engineering dioxygenases by laboratory evolution: a comparison of evolutionary search strategies
Degree	PhD
Option	Chemical Engineering
Advisory Committee	Advisor Name Title Frances Arnold Committee Chair Douglas Rees Committee Member George Gavalas Committee Member Harry Gray Committee Member
Keywords	molecular evolution dioxygenase DNA shuffling directed evolution
Date of Defense	2003-01-23
Availability	unrestricted
Abstract Due to the unique and difficult chemistry they perform, the aromatic ring-hydroxylating dioxygenases are of interest as industrial catalysts. Unfortunately, an application-specific array of problems limits their utility. To address these problems through laboratory evolution, I developed methods for high-throughput screening of tens of thousands of dioxygenase variants. These methods rely on a phenol detection reagent (Gibbs reagent) and can be applied to liquid cultures or to growing bacterial colonies expressing variant enzymes. Recombination of genes encoding homologous enzymes ("family shuffling") has emerged as a promising tool for evolutionary protein engineering. Using the dioxygenases as a model system, I have investigated the value of recombination as a search strategy for laboratory evolution. Chimeric dioxygenase libraries constructed by DNA shuffling are first evaluated for biases that limit sequence diversity using a probe hybridization approach in lieu of sequencing. This analysis shows that crossovers preferentially occur in regions with high sequence identity and that certain parent sequences can be preferred at particular gene positions. High-throughput functional screening allowed characterization of substrate specificity for hundreds of dioxygenase chimeras. These data are coupled with sequence data to reveal sequence-function relationships and demonstrate the utility of recombination as a tool for functional genomics. One region of sequence is shown to be a primary determinant of substrate specificity for the enzymes studied. Furthermore, several sets of variant enzymes with similar functionality are shown to have sequence similarities. Recombination and random mutagenesis are compared as search strategies for generating functionally-diverse dioxygenases. I screened similarly sized libraries of chimeric and mutant dioxygenases for variants with altered substrate specificity or activity toward n-hexylbenzene, which is not accepted by the parent enzymes. Both recombination and random mutagenesis gave rise to enzymes with altered substrate specificity, although such enzymes were more frequent in the chimeric libraries and more distinct specificities were found in the chimeric libraries. Only chimeras were active toward n-hexylbenzene. These results support the view that recombination is an effective search strategy for evolving substrate specificity, and may be more effective than random mutagenesis.
Files	Filename       Size       Approximate Download Time (Hours:Minutes:Seconds)     28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access    1-28-03_thesis.pdf 3.46 Mb 00:16:01 00:08:14 00:07:12 00:03:36 00:00:18