Reconciling in vivo and in silico key biological parameters of Pseudomonas putida KT2440 during growth on glucose under carbon-limited condition

BACKGROUND: Genome scale metabolic reconstructions are developed to efficiently engineer biocatalysts and bioprocesses based on a rational approach. However, in most reconstructions, due to the lack of appropriate measurements, experimentally determined growth parameters are simply taken from literature including other organisms, which reduces the usefulness and suitability of these models. Pseudomonas putida KT2440 is an outstanding biocatalyst given its versatile metabolism, its ability to generate sufficient energy and turnover of NADH and NAD. To apply this strain optimally in industrial production, a previously developed genome-scale metabolic model (iJP815) was experimentally assessed and streamlined to enable accurate predictions of the outcome of metabolic engineering approaches. RESULTS: To substantially improve the accuracy of the genome scale model (iJP815), continuous bioreactor cultures on a mineral medium with glucose as a sole carbon source were carried out at different dilution rates, which covered pulling analysis of the macromolecular composition of the biomass. Besides, the maximum biomass yield (on substrate) of 0.397 g(DCW) · g(glc)(-1), the maintenance coefficient of 0.037 g(glc) · g(DCW)(-1) · h(-1) and the maximum specific growth rate of 0.59 h(-1) were determined. Only the DNA fraction increased with the specific growth rate. This resulted in reliable estimation for the Growth-Associated Maintenance (GAM) of 85 mmol(ATP) · g(DCW)(-1) and the Non Growth-Associated Maintenance (NGAM) of 3.96 mmol(ATP) · g(DCW)(-1) · h(-1). Both values were found significantly different from previous assignment as a consequence of a lower yield and higher maintenance coefficient than originally assumed. Contrasting already published (13)C flux measurements and the improved model allowed for constraining the solution space, by eliminating futile cycles. Furthermore, the model predictions were compared with transcriptomic data at overall good consistency, which helped to identify missing links. CONCLUSIONS: By careful interpretation of growth stoichiometry and kinetics when grown in the presence of glucose, this work reports on an accurate genome scale metabolic model of Pseudomonas putida, providing a solid basis for its use in designing superior strains for biocatalysis. By consideration of substrate specific variation in stoichiometry and kinetics, it can be extended to other substrates and new mutants.