Metabolic Engineering of Methylobacterium extorquens AM1
Steven Van Dien
University of Washington

A stoichiometric model of central metabolism was developed based on new information regarding metabolism in this bacterium to evaluate the steady-state growth capabilities of the serine cycle facultative methylotroph Methylobacterium extorquens AM1 during growth on methanol, succinate, and pyruvate.  The model incorporates 20 reversible and 47 irreversible reactions, 65 intracellular metabolites, and experimentally-determined biomass composition.  The flux space for this underdetermined system of equations was defined by finding the elementary modes, and constraints based on experimental observations were applied to determine which of these elementary modes give a reasonable description of the flux distribution for each growth substrate.  The predicted biomass yield, on a carbon atom basis, is 49.8%, which agrees well with the range of published experimental yield measurements (37-50%).  The model predicts the cell to be limited by reduced pyridine nucleotide availability during methylotrophic growth, but energy-limited when growing on multicarbon substrates.

Mutation and phenotypic analysis was used to test model predictions regarding key enzymes for growth on C3 and C4 compounds.  Three enzymes involved in C3-C4 interconversion pathways were predicted to be mutually redundant: malic enzyme, phosphoenolpyruvate carboxykinase, and phosphoenolpyruvate synthase. Insertion mutations in the genes from the genome sequence that are predicted to encode these enzymes were made, and these mutants were capable of growing on all substrates tested, confirming the model predictions.  Likewise, citrate synthase and succinate dehydrogenase were predicted by the simulations to be essential for all growth substrates.  In keeping with these predictions, null mutants could not be obtained in these genes.  In addition, a random approach using transposon mutagenesis was used to generate mutants with impaired growth on succinate or pyruvate.  A mutant in a gene predicted to encode a subunit of the NADH-quinone oxidoreductase was obtained, and was unable to grow on succinate or pyruvate but grew normally on methanol.  Since this function is necessary for the entry of NADH into the electron transport chain, this finding supports the model prediction that NADH must be oxidized to ultimately yield ATP during multicarbon growth, but not with methanol as the carbon source.  A transposon mutant in a putative a-ketoglutarate dehydrogenase gene was also unable to grow on succinate or pyruvate.  However, the model does not predict this enzyme activity to be required for growth on any substrate.  In situations such as this in which the phenotype does not agree with predictions, the model has helped to identify errors in the current understanding of Methylobacterium extorquens AM1 central metabolism.

Return to Table of Contents