Quantifying variation within the bacterial species E. coli

Escherichia coli strains are widely used in academic and applied research as well as in biotechnology for production of various compounds. Despite its status as a model organism, strain-specific differences and their underlying contributing factors are still not well characterized. These differences have a major impact on cell physiology and for the applied purposes of synthetic biology, metabolic engineering, and process design. In this study, strain-specific differences are quantified in seven widely-used applied biotechnology strains of E. coli (BL21, C, Crooks, DH5a, K-12 MG1655, K-12 W3110, and W) using genomics, phenomics, transcriptomics and genome-scale modelling to guide the choice of strain for a given product. Even given the genetic similarity of the strains, mMetabolic physiology and gene expression varied widely with downstream implications for productivity, product yield, and titre. Further, these differences can be linked to differential regulatory structure. Analysing high flux reactions and the expression levels of their encoding genes revealed a quantitative link between these sets and show that often, these sets are correlated with strain-specific caveats. Integrated modelling also revealed that certain strains are better suited to produce a given compound or express a desired construct considering native expression states of pathways that enable high-production phenotypes. The result of this study is a resource comparing strains in an important model species and a general strategy for choosing a host strain or chassis selection for applied biotechnology. 7 different strains of E. coli were grown in glucose M9 minimal media aerobically and anaerobically. Their gene expression was measured using RNAseq on an Illumina HiSeq2000