The utility and performance of near-infra red spectroscopy in simultaneous monitoring of multiple components in a high cell density recombinant Escherichiacoli production process

Abstract Bioprocess optimisation is often limited by an inability to measure biomass, nutrient and by-product concentrations in a time frame which allows process adjustments. Near-infrared (near-IR) spectroscopy can potentially be used to measure each of these components within 2 minutes of sampling, using an unprocessed whole broth sample. In the present study the use of near-IR spectroscopy for at-line (rapid off-line) monitoring of biomass, glycerol, ammonium, and acetate in a recombinant Escherichia coli fed batch process was investigated. The following robust correlation models were developed for these analytes using multiple least squares linear regression (MLR): [Glycerol],gl−1 =15.957 − 2219.270*A(λ2274)−1705.041*A(λ2172); [Acetate],gl−1=27.683 −1757.258*A(λ2254)+296.903*A(λ2340)−21.325*A (λ620); [Ammonium],gl−1=−1310.502−47912.960*A (λ2148)−135149.300*A(λ1782)−27636.200*A (λ830); and [Biomass],gl−1=14.034−3.548*A(λ602/λ1134) −4286.050*A(λ928). Using these models permitted rapid simultaneous analysis of all four analytes. This improved monitoring capability was used to develop a high cell density recombinant E. coli fed-batch process in which ammonium and acetate accumulation were minimised leading to higher cell densities. By manipulation of the C : N ratio in the complex feed, the toxic effects of ammonium accumulation upon the organism were minimised, thereby facilitating the application of a carbon limited feeding strategy. The effect of these C : N ratio medium changes, upon the near-IR measurement capability, was investigated. In this process, near-IR spectroscopy has been shown to be a powerful, accurate and precise method for simultaneously measuring several key process variables. Its accuracy, precision and utility for at-line measurement and control are evaluated, particularly in reference to processes where the initial medium composition may vary, leading to changes in the chemical matrix. The potential of near-infra red for online analysis and control is discussed.