Résumé
BACKGROUND: Nonribosomal peptide synthases (NRPS) can synthesize functionally diverse bioactive peptides by incorporating nonproteinogenic amino acids, offering a rich source of new drug leads. The bacterium Escherichia coli is a well-characterized production host and a promising candidate for the synthesis of nonribosomal peptides, but only limited bioprocess engineering has been reported for such molecules. We therefore developed a medium and optimized process parameters using the design of experiments (DoE) approach. RESULTS: We found that glycerol is not suitable as a carbon source for rhabdopeptide production, at least for the NRPS used for this study. Alternative carbon sources from the tricarboxylic acid cycle achieved much higher yields. DoE was used to optimize the pH and temperature in a stirred-tank reactor, revealing that optimal growth and optimal production required substantially different conditions. CONCLUSIONS: We developed a chemically defined adapted M9 medium matching the performance of complex medium (lysogeny broth) in terms of product concentration. The maximum yield in the reactor under optimized conditions was 126 mg L-1, representing a 31-fold increase compared to the first shaking-flask experiments with M9 medium and glycerol as the carbon source. Conditions that promoted cell growth tended to inhibit NRPS productivity. The challenge was therefore to find a compromise between these factors as the basis for further process development.