Prediction of non-intuitive metabolic targets with bayesian metabolic control analysis to improve 3-hydroxypropionic acid production in Aspergillus niger

Development of efficient bioconversion processes is limited by the ability to predictably improve metabolic flux. Here we deployed Bayesian Metabolic Control Analysis as a platform to integrate multi-omics data with metabolic modeling and evaluated its ability to predict genetic interventions that improve metabolic flux. Global Metabolomics and proteomics data was collected from 17 <i>Aspergillus niger</i> strains engineered to produce the platform biochemical 3-hydroxypropionic acid from which seven actional genetic interventions were predicted from significant flux control coefficients. Of the suggested genetic interventions, two were present within the intuitively designed strains used for training (malonic semialdehyde dehydrogenase and pyruvate carboxylase) while five predicted targets were present within non-intuitive areas of the metabolic network including 5-formyltetrahydrofolate deformylase and four mitochondrial enzymes, alcohol dehydrogenase, succinyl-CoA ligase, aspartate aminotransferase, and malate dehydrogenase. Six of the targets were validated in the highest performing 3-HP strain used for multi-omics data generation which contained a prior disruption of the highest scoring target malonic semialdehyde dehydrogenase. Predicted directional perturbation of five of the six tested targets significantly improved titer and rate of 3-HP production and two significantly improved yield. The greatest improvements were observed following disruption of the non-intuitive target succinyl-CoA ligase which increased titer by 39% and yield by 29% (to 20.4 g/L 3-HP and 0.31 g 3-HP/g glucose) over the strains used for training. This study demonstrates the utility of Bayesian Metabolic Control Analysis and highlights the ability to predict meaningful genetic targets in unexpected areas of metabolism to improve engineered strains for bioconversion.