ABSTRACT
The understanding and engineering of complex phenotypes is a critical issue in biotechnology. Conventional approaches for engineering such phenotypes are often resource intensive, marginally effective, and unable to generate the level of biological understanding desired. Here, we report a new approach for rapidly dissecting a complex phenotype that is based upon the combination of genome-scale growth phenotype data, precisely targeted growth selections, and informatic strategies for abstracting and summarizing data onto coherent biological processes. We measured at high resolution (125 NT) and for the entire genome the effect of increased gene copy number on overall biological fitness corresponding to the expression of a complex phenotype (tolerance to 3-hydroxypropionic acid (3-HP) in Escherichia coli). Genetic level fitness data were then mapped according to various definitions of gene-gene interaction in order to generate network-level fitness data. When metabolic pathways were used to define interactions, we observed that genes within the chorismate and threonine super-pathways were disproportionately enriched throughout selections for 3-HP tolerance. Biochemical and genetic studies demonstrated that alleviation of inhibition of either of these super-pathways was sufficient to mitigate 3-HP toxicity. These data enabled the design of combinatorial modifications that almost completely offset 3-HP toxicity in minimal medium resulting in a 20 g/L and 25-fold increase in tolerance and specific growth, respectively.
