Abstract
Streptomyces clavuligerus produces the potent beta-lactamase inhibitor clavulanic acid, which is used in conjunction with other beta-lactam antibiotics. Although the genome of S. davuligerus has not yet been sequenced, tools were developed to study the metabolic flux distribution (fluxomics) and gene expression (transcriptomics) at genome scale. A Streptomyces clavuligerus-specific genome-scale metabolic network was derived from the network reconstructed for the close relative S. coelicolor (Borodina et al., 2005), removing reactions unique to S. coelicolor and adding the ones known to take place in S. clavuligerus. This metabolic network was then used to perform Flux Balance Analysis. Genome-scale transcription analysis was also made possible by customising S. coelicolor DNA microarrays thanks to the addition of 80 S. clavuligerus-specific probes, corresponding to most of the sequenced S. clavuligerus genes, to the 8,000 S. coelicolor probes, including probes for the genes of the clavulanic acid cluster. These genome- scale methodologies were applied to S. clavuligerus grown in chemostat and batch cultures. Macromolecular compositions at two dilution rates (0.02h-1 and 0.05h-1) and in different growth phases were analysed in order to obtain reliable flux data. Flux distribution was simulated accurately and predicted supplements such as threonine, malate, pyruvate or citrate that might lead to increased antibiotic production. The customised DNA microarrays demonstrated that the genes of the clavulanic acid cluster are regulated differently in batch and in chemostat culture. The analysis of the transcription data also predicted that the use of supplements such as glycine or thymidine could lead to increased clavulanic acid production. Finally, the gene expression patterns showed that S. clavuligerus might be using an alternative pathway to metabolise glycerol during clavulanic acid production. In batch culture, the fluxomic and transcriptomic data showed no correlation, while some correlation was observed in chemostat culture. This demonstrated that chemostat culture is more useful than batch culture when performing genome-scale analysis. However, batch culture remains a valuable tool to assess gene co-expression or to study a switch between two phases of growth.