Abstract
The publication of the S. coelicolor genome sequence has stimulated interest in the possibility of rationally engineering strains by predicting the effect(s) of altering specific genes on secondary metabolite production. The work described here aims to design a general approach that could be applied to enhance the production of commercially important antibiotics that have similar biosynthetic routes as actinorhodin (ACT). It has been observed that a negative correlation between ACT production and the carbon flux through the pentose phosphate pathway (PPP) exists. As a result, strains with deletions for either one of the two zwf isogenes (encoding the enzyme that catalyses the initial PPP reaction) and for the devB gene (whose product catalyses the subsequent reaction in the pathway) were constructed and assessed. All strains, with the exception of the devB mutant, were found to generate increased levels of ACT compared to the parental strain. In addition to these strains, the above genotypes were accompanied by the pIJ8714 insert, containing an additional copy of the pathway-specific activator gene, actII-0RF4. These strains displayed precocious and significantly increased ACT levels compared to their corresponding controls. In order to compare the carbon flux distribution profiles of the various strains and conditions tested, metabolic flux analysis (MFA) was performed. Correlation analysis was applied to such data to identify reactions that either positively or negatively correlated with ACT synthesis. Both MFA results and in vitro enzyme analysis showed that both zwf stiains had elevated carbon fluxes through the initial reaction of the PPP. This was unexpected but may reflect some co-regulation between the two zwf isogenes. When either one of the two genes was deleted, the presumed repression it exerted was removed, resulting in more activity. Such elevated fluxes allowed more NADPH to fuel the high yielding elementary mode that was identified for ACT synthesis and, presumably amplified to saturation by the pIJ8714 insert. MFA from batch and chemostat cultures showed that greater flexibility at the glucose-6-phosphate principal node was found at the lower growth rates (and in the pIJ8714 strains) which coincided with greater ACT production. Biomass and ACT synthesis compete for the same cofactors, hence at lower growth rates, more e.g. NADPH, was available for ACT production. Furthermore it was shown that microarray analysis could not be used to predict carbon fluxes in S. coelicolor strains, as little correlation between the two techniques was observed. However, it was found to be a useful method for highlighting genes that were more than two-fold up/down-regulated in the strain(s) that generated the most ACT. These findings could be exploited in future metabolic engineering studies to enhance ACT production, as well as for other bioactive compounds that undergo similar biosynthetic pathways.