Applying metabolic pathway analysis to make good use of methanol

Abstract

Metabolic engineering requires powerful theoretical tools because the networks of enzymes performing biosyntheses are too complex to be understood intuitively. Metabolic pathway analysis and, in particular, the concept of elementary flux modes were developed for this purpose because they allow one to make important predictions even if little is known about kinetic parameters.So far, the number of theoretical papers on pathway analysis exceeds the number of applied ones. Van Dien and Lidstrom now present a comprehensive theoretical and experimental study on a proposed metabolic network comprising 67 reactions in the facultative methylotroph bacterium, Methylobacterium extorquens AM1 [1xStoichiometric model for evaluating the metabolic capabilities of the facultative methylotroph Methylobacterium extorquens AM1, with application to reconstruction of C3 and C4 metabolism. Van Dien, S.J. and Lidstrom, M.E. Biotechnol. Bioeng. 2002; 78: 296–312CrossRef | PubMed | Scopus (79)See all References][1]. One of the reactions is an overall biomass synthesis reaction using 20 key precursor metabolites in a fixed proportion. On compiling this network, the available genome data on the bacterium were included. For nearly all enzymes not previously reported for M. extorquens AM1, gene candidates were identified and several of these were amplified by PCR. In a theoretical analysis, the flux space of the reaction network was explored by determining and analysing the elementary modes (using METATOOL) separately for three different substrates; methanol, pyruvate and succinate. For example, for growth on methanol the number of elementary modes is 467, several of which represent futile cycles. The optimal biosynthetic mode implies a yield of 0.498 g biomass carbon per mole of methanol consumed. This is close to the experimentally measured value (0.375−0.5) therefore corroborating the proposed network stoichiometry. For each substrate, the elementary modes having a biomass yield of at least 90% of the maximum yield were analysed and compared with each other.Van Dien and Lidstrom also studied mutants of M. extorquens AM1, both experimentally and theoretically. Null mutants for citrate synthase or succinate dehydrogenase did not grow on any substrate. When either of the corresponding reaction equations was cancelled from the network, the remaining elementary modes did not produce biomass. By contrast, the fact that the null mutants for putative phosphoenolpyruvate (PEP) carboxykinase or malic enzyme genes did grow normally, whereas a mutant lacking both enzymes did not grow on succinate, was also shown by modelling. The only discrepancy between the experiment and the modelling occurred for PEP synthase mutants with pyruvate as the substrate. Thus, the theoretically predicted alternative pyruvate-consuming pathway is likely to be downregulated in vivo. The study was finalized with a comparison of a metabolic network proposed for Methylobacillus flagellatum. It uses the ribulose monophosphate cycle instead of the serine cycle for transforming one-carbon units into three-carbon units, and results in a higher biomass yield (0.661 g biomass carbon per mole of methanol).This excellent study combines theoretical and experimental tools, and shows that the concept of elementary flux modes can successfully be applied to study the metabolic capabilities of biotechnologically relevant organisms and to reconstruct bacterial metabolisms. This has practical implications because the bacteria being studied can serve to convert methanol, which is cheap and easy to handle, into products of commercial interest.