(22a) Inducible Directed Evolution of an Anticancer Terpene Biosynthetic Pathway in E. coli

A central challenge in metabolic engineering lies in identifying the correct expression levels and activities of enzymes that lead to the production of a desired product. While a plethora of methods enable efficient exploration of expression space for entire metabolic pathways, enzymatic activity improvements are primarily achieved through directed evolution of single enzymes. We wished to understand whether directed evolution of more complete metabolic pathways would lead to greater productivities than evolution of single enzymes alone. To do this, we took production of (S)-(-)-perillyl alcohol (POH), a putative anticancer molecule, as a case study. Studies have shown that POH and its derivatives can pass the blood-brain barrier and treat glioblastoma with few side effects. The low amounts that can be isolated from plant sources, as well as the complexity and high cost of chemical synthesis, has made microbial production an ideal platform for POH synthesis. We have developed a minimal, highly efficient enzymatic pathway for POH biosynthesis in E. coli and a transcription factor-based biosensor to detect POH production in situ. However, further improvement in productivity has been hampered by the difficulty of evolving the large POH biosynthetic pathway (~ 10 kb). To solve this issue, we used Inducible Directed Evolution (IDE), which is capable of evolving DNA sequences up to 36 kb in length. Specifically, we used a mutagenesis plasmid to introduce random mutations into the POH biosynthetic pathway, followed by transfer into a biosensor-containing strain via a temperate bacteriophage. Variants with high cellular fluorescence, corresponding to high POH productivity, were recovered by multiple rounds of fluorescence-activated cell sorting spanning over 10^6 variants, and the spectrum of mutations present during each round were determined by next generation sequencing. By synthetically introducing combinations of key mutations, we investigated whether synergies were occurring between mutations in different enzymes. These results demonstrate the utility of inducible directed evolution for biosensor-enabled screens and point to a pathway-wide framework for performing directed evolution on multigene metabolic pathways.