In this thesis, we describe the discovery of novel mechanisms for improving the production of the antibiotic erythromycin in Escherichia coli . The synthesis of erythromycin in E. coli requires the heterologous expression of polyketide synthases and over a dozen other enzymes that occur naturally in Saccharopolyspora erythraea, an actinomycete bacterium. Classical methods for strain improvement often involve multiple rounds of random mutagenesis and screening, leaving unknown mutations. To study the effects of random mutations, we screened a library of transposon mutants of E. coli for increased production of erythromycin; the use of a transposon as the mutagen allowed the identification of the causative mutations. From this screen, two new mutations were discovered. Each identified a strain improvement that doubles the yield of erythromycin from E. coli.

One improved mutant contains a transposon inserted between the −10 and −35 regions of a promoter for rpoE, the gene encoding the sigma factor for the extracytoplasmic stress response, σ ^E. Rather than inactivating the promoter, the transposon insertion in this mutant increases the transcription of rpoE while simultaneously reducing σ^E autoinduction. Other mutations that increase the σ^E response also improve the production of erythromycin, yet trigger an increase in cell lysis not seen in cultures of the transposon mutant. The transposon mutant contains higher concentrations of the soluble enzymes for the synthesis of erythromycin, likely resulting from the modified transcription of the σ^E regulon.

A second mutant contains a transposon inserted within metY, one of four identical genes encoding the initiator tRNA of E. coli . The deletion of metY or the identical gene metV both improve the production of erythromycin, but the simultaneous deletion of multiple genes that encode initiator tRNA does not further increase production. The initiator tRNA mutants also contain higher concentrations of the soluble enzymes for the synthesis of erythromycin, perhaps resulting from modified translational initiation within these strains.

While the rational engineering of production strains often targets specific metabolic pathways, the new mutants revealed by this screen involve mechanisms that impact global transcription or translation within the cells. Additionally, these mutants demonstrate the success of transposon mutagenesis in improving production strains, where the transposon insertions, rather than inactivating pathways, alter the expression of key components of normal cellular function. Combining the new mutations fails to further increase the yield of erythromycin, highlighting the importance of relevant characteristics of the initial strain on the types of mutations identified by a screen. Consequently, for engineered microorganisms, multiple rounds of mutagenesis are likely required to achieve high productivity.