Near Infrared Spectroscopic Monitoring and Control of a Whole-Cell Biocatalytic Process

Accurate and robust monitoring of product and reactants in a complex bioconversion stream is essential for the development of effective process control strategies. To monitor a microbially-catalysed Baeyer-Villiger bioconversion of a cyclic ketone to an optically pure lactone, a near infrared (NIR) spectroscopic method has been developed. The reaction, catalysed by cyclohexanone monooxygenase from Acinetobacter calcoaceticus (expressed in Escherichia coli) is characterised by substrate (ketone) and product (lactone) inhibition at relatively low concentrations. Quantitative multivariate calibration of a NIR spectrophotometer for ketone and lactone resulted in a standard error of prediction (SEP) at-line of 0.088 and 0.110 g.l"1 and on-line of 0.130 and 0.180 g.l'1, respectively. The directed modification of quantitative models, by the inclusion of spiked process samples improved the SEP for lactone prediction where bioprocess development meant existing NIR models were not relevant. The monitoring and control of the BaeyerVilliger bioconversion by NIR has allowed intermittent feeding of ketone such that the concentration of substrate does not rise above 0.6 g.l'1. Using this feeding strategy 5.7 g.l'1 of lactone product has been produced. This represents a 2-fold increase in productivity. The application of a technique to monitor analytes at low concentration demonstrates the utility of NIR for control of biotransformation processes.