Expression of Heterologous Sigma Factors in Escherichia coli to Explore the Heterologous Genomic Space for Building, Stepwise, Complex, Multicomponent Phenotypes

A key limitation in using heterologous genomic or metagenomic libraries in functional genomics and genome engineering is the low-level expression of heterologous genes in screening hosts, such as Escherichia coli.  To overcome this limitation, we constructed E. coli strains capable of recognizing heterologous promoters by expressing sigma factors from the phylogenetically distant Lactobacillus plantarum and Bacillus subtilis.  Such strains were employed for screening heterologous DNA libraries using the promoter GFP-trap concept. We show greatly increased transcription from single and combined genomic libraries of L. plantarum, B. subtilis, Clostridium pasterianum, C. acetobutylicum and Deinococcus radiodurans, thus enlarging the genomic space that can be functionally sampled in E. coli.  We show two applications. In one, we show that screening fosmid-based L. plantarum genomic libraries in an E. coli strain with a chromosomally integrated L. plantarum rpoD (coding for the major sigma factor) allowed the identification of L. plantarum genetic determinants imparting 14-fold increased survivability to 7% v/v ethanol compared to the control in E. coli. In the second, we demonstrate the concept for a sequential, iterative assembly strategy for building multigenic traits by exploring the synergistic effects of genetic determinants from broader genomic spaces. Specifically, building upon the success of our recently reported semi-synthetic stress response system expressed off plasmid pHSP, we probed the genomic space of the solvent tolerant L. plantarum to identify genetic determinants that impart solvent tolerance in combination with pHSP. Using two targeted enrichments, one for superior viability and one for better growth under ethanol stress, we identified several beneficial heterologous DNA determinants that act synergistically with pHSP. In separate strains, a 209% improvement in survival and an 83% improvement in growth over previously engineered strains based on pHSP were thus generated. We then developed a composite phenotype of improved growth and survival by combining the identified L. plantarum genetic fragments. The best performing strain produced a 3.7-fold improved survival under 8% ethanol stress, as well as a 32% increase in growth under 4% ethanol. We also show that this strain can significantly improve ethanol productivity in a practically significant Melle-Boinot like fermentation process.