Real-Time Metabolite Monitoring Of Glucose-Fed Clostridium Acetobutylicum Fermentations Using Raman Assisted Metabolomics

Data obtained from in situ Raman spectroscopy probes and high-performance liquid chromatography (HPLC) analysis were applied together with chemometrics to build partial least squares models of metabolite concentrations for the industrially relevant organism Clostridium acetobutylicum. Models were built for predominant products (acetic acid, butyric acid, and butanol) of C.acetobutylicum cultures grown on glucose as a substrate. The partial least squares models were then applied to a 3-day C.acetobutylicum culture for real-time, quantitative metabolite analysis. The predicted outcomes of new fermentation cultures were validated by analyzing HPLC data from corresponding experiments from these new fermentation cultures. Model predictions showed good correlation with measured data (goodness of fit [R2Y] values of 0.99, and goodness of prediction [Q2Y] values of 0.98 from agitated cultures. Predictive models based upon Raman spectral data are promising tools for characterization of synthetic organisms, guiding process control, and facilitating optimization of culture conditions.