Multi-omic data fusion reveals the in vivo enzyme kinetics of Vibrio natriegens at the genome-scale

Vibrio natriegens is a halophilic, Gram-negative marine bacterium that is increasingly used in metabolic engineering applications due to its fast growth rate. In sparse minimal medium the organism has a doubling time of 25 minutes, which is about twice as fast as Escherichia coli under similar conditions. Given that its protein density is similarly constrained to that of E. coli, this necessitates that its metabolic enzymes are able to catalyze flux at a higher rate to sustain its metabolism. In this work, we measure the apparent turnover numbers of metabolically active enzymes in V. natriegens under a variety of growth conditions. The apparent turnover numbers of V. natriegens enzymes were measured in vivo by conducting coupled quantitative proteomics and 13C metabolic flux analysis experiments under seven different carbon source conditions in sparse minimal medium. A high quality genome-scale metabolic model was constructed and curated using additional experimental data. This model was extended with enzyme constraints, and subsequently used to find kinetic parameters that minimize the difference between model predictions and experimental observations. This model guided data fusion approach enabled the estimation of 357 apparent turnover numbers for metabolically active enzymes in V. natriegens. Our results reveal that the metabolic enzymes of V. natriegens are in median 14-fold faster than those of E. coli under similar conditions. Moreover, we show that machine learning generated turnover number estimates substantially underestimate the kinetics of V. natriegens. Our turnover number estimates were used to parameterize multiple condition dependent enzyme constrained flux balance analysis models of V. natriegens, which improved their predictive accuracy compared to the machine learning parameterisation. The combined experimental-computational approach employed here sheds light on the mechanism V. natriegens uses to accelerate its growth. This approach can also be extended to other bacteria, increasing the availability of in vivo measured enzyme turnover numbers, and improving the predictive accuracy of enzyme constrained metabolic models of other microbes.