Comprehensive in-situ Bioreactor Monitoring and Control Based on a Mid-Infrared Spectroscopic Sensor System

Comprehensive in-situ bioreactor monitoring and control based on a midinfrared spectroscopic sensor system M. Rhiel1, C. Cannizzaro1, S. Valentinotti2, I. Marison1, U. von Stockar1 1 Institute of Chemical Engineering, 2Institute of Automatic Control Swiss Federal Institute of Technology (EPFL), CH- 1015 Lausanne, Switzerland Abstract A mid-infrared spectroscopic sensor system consisting of a spectrometer, a single attenuated total reflectance (ATR) probe immersed in the bioreactor, and a software package for multivariate spectra analysis was used to monitor glucose, ethanol, glycerol, acetic acid, sulfate, phosphate, PPG 2000 antifoam, and biomass in Saccharomyces cerevisiae cultures. The provided ethanol concentration was used to control the feed of a concentrated glucose solution to maintain maximum oxidative growth. Introduction: Bioprocess monitoring and control rely on the use of appropriate sensors. Ideally, these sensors should be in situ and measure simultaneously the concentrations of all major bioprocess metabolites with sufficient accuracy and long-term stability under minimum maintenance. Spectroscopic sensors utilizing the mid-infrared (MIR) electromagnetic range offer many advantages, including simultaneous multi-analyte determinations, in situ sterilizability, low maintenance during operation, and enhanced information about most biologically important species compared to near-infrared (NIR) spectroscopic sensors. bioprocess monitoring. Overlapping absorbance features, however, require the application of advanced chemometric analysis methods. This paper discusses the simultaneous in-situ monitoring of glucose, ethanol, glycerol, acetic acid, biomass, pH, temperature, and bioreactor volume in batch and fed-batch cultures of Saccharomyces cerevisiae with a Therefore, they seem to be good candidates for single-probe mid-infrared spectroscopic sensor (ReactIRTM, ASI Applied Systems). Calibration models for each analyte were established with partial least-squares regression (PLSR) implemented for immediate process analysis (QuantIRTM, ASI Applied Systems). Concentration data was available every two minutes and allowed the implementation of a fed-batch control strategy to maximize biomass production. Materials and Methods: A ReactIRTM 1000 mid-infrared spectrometer (ASI Applied Systems, Millersville, MD) equipped with a diamond ATR immersion probe (DiCompTM , ASI Applied Systems) and QuantIRTM software (ASI Applied Systems) was used for in situ data acquisition and analysis. Saccharomyces cerevisiae CBS426 and S. cerevisiae W303-1A PVD32 cells were grown on defined medium in a 16L stirred tank bioreactor (Bioengineering AG, Wald, Switzerland). A customized bioprocess management and control environment (BioOPT) was developed with LabVIEWTM (National Instruments, Austin, TX). Results: Quantitative spectra analysis was performed by establishing PLSR models for each of the analytes: glucose, ethanol, glycerol, acetic acid, sulfate, phosphate, PPG 2000 antifoam, and biomass. The calibration data set consisted of 60 in situ collected pure component spectra and 40 in situ collected reaction spectra. Maintaining the ethanol concentration reported by the mid-infrared spectroscopic sensor system around 0.7 g/L resulted in an exponential growth rate of 0.255 1/h. Exponential growth was sustained until oxygen, supplied by sparging air, became limiting. Thereafter, the controller adapted automatically and linear growth occurred up to a cell density of 60 g/L.

Identification and control of oxidative metabolism in Saccharomyces cerevisiae during transient growth using calorimetric measurements

Urs von Stockar

The objective of this study was to characterize the dynamic adaptation of the oxidative capacity of Saccharomyces cerevisiae to an increase in the glucose supply rate and its implications for the control of a continuous culture designed to produce biomass without allowing glucose to be diverted into the reductive metab. Continuous cultures subjected to a sudden shift-up in the diln. rate showed that the glucose uptake rate increased immediately to the new feeding rate but that the oxygen consumption could not follow fast enough to ensure a completely oxidative metab. Thus, part of the glucose assimilated was degraded by the reductive metab., resulting in a temporary decrease of biomass concn., even if the final diln. rate was below Dcrit. The dynamic increase of the specific oxygen consumption rate, qo2, was characterized by an initial immediate jump followed by a first-order increase to the max. value. It could be modeled using three parameters denoted qo2jump, qo2max, and a time const. t. The values for the first two of the parameters varied considerably from one shift to another, even when they were performed under identical conditions. On the basis of this model, a time-dependent feed flow rate function was derived that should permit an increase in the diln. rate from one value to another without provoking the appearance of reductive metab. The idea was to increase the glucose supply in parallel with the dynamic increase of the oxidative capacity of the culture, so that all of the assimilated glucose could always be oxidized. Nevertheless, corresponding feed-profile expts. showed that deviations in the reductive metab. could not be completely suppressed due to variability in the model parameters. Therefore, a proportional feedback controller using heat evolution rate measurements was implemented. Calorimetry provides an excellent and rapid est. of the metabolic activity. Satisfactory control was achieved and led to const. biomass yields. Ethanol accumulated only up to 0.49 g L-1 as compared to an accumulation of 1.82 g L-1 without online control in the shift-up expt. to the same final diln. rate. [on SciFinder (R)]

1998

Methodology for real-time, multianalyte monitoring of fermentations using an in-situ mid-infrared sensor

Christopher Cannizzaro, Henri Kornmann, Ian William Marison, Martin Rhiel, Urs von Stockar

An in-situ, mid-IR sensor was used to monitor the major analyte concns. involved in the cultivation of Gluconacetobacter xylinus and the prodn. of gluconacetan, a food-grade exopolysaccharide. To predict the analyte concns., three different sets of std. spectra were used to develop calibration models, applying partial least-squares regression. It was possible to build a valid calibration model to predict the 700 spectra collected during the complete time course of the cultivation, using only 12 spectra collected every 10 h as stds. This model was used to reprocess the concn. profiles from 0 to 15 g/L of nine different analytes with a mean std. error of validation of 0.23 g/L. However, this calibration model was not suitable for real-time monitoring as it was probably based on non-specific spectral features, which were correlated only with the measured analyte concns. Valid calibration models capable of real-time monitoring could be established by supplementing the set of 12 fermn. spectra with 42 stds. of measured analytes. A pulse of 5 g/L ethanol showed the robustness of the model to sudden disturbances. The prediction of the models drifted, however, toward the end of the fermn. The most robust calibration model was finally obtained by the addn. of 34 std. spectra of non-measured analytes. Although the spectra did not contain analyte-specific information, it was believed that this addn. would increase the variability space of the calibration model. Therefore, an expanded calibration model contg. 88 spectra was used to monitor, in real time, the concn. profiles of fructose, acetic acid, ethanol and gluconacetan and allowed std. errors of prediction of 1.11, 0.37, 0.22, and 0.79 g/L, resp. [on SciFinder (R)]

2003

Influence of residual ethanol concentration on the growth of Gluconacetobacter xylinus I 2281

Henri Kornmann, Ian William Marison, Urs von Stockar

The influence of residual ethanol on metab. of food grade Gluconacetobacter xylinum I 2281 was investigated during controlled cultivations on 35 g/l glucose and 5 g/l ethanol. Bacterial growth was strongly reduced in the presence of ethanol, which is unusual for acetic acid bacteria. Biomass accumulated only after complete oxidn. of ethanol to acetate and carbon dioxide. In contrast, bacterial growth initiated without delay on 35 g/l glucose and 5 g/l acetate. It was found that acetyl CoA was activated by the acetyl CoA synthetase (Acs) pathway in parallel with the phosphotransacetylase (Pta)-acetate kinase (Ack) pathway. The presence of ethanol in the culture medium strongly reduced Pta activity while Acs and Ack remained active. A carbon balance calcn. showed that the overall catabolism could be divided into two independent parts: upper glycolysis linked to glucose catabolism and lower glycolysis linked to ethanol catabolism. This calcn. showed that the carbon flux through the tricarboxylic cycle is lower on ethanol than on acetate. This corroborated the diminution of carbon flux through the Pta-Ack pathway due to the inhibition of Pta activity on ethanol. [on SciFinder (R)]

Springer-Verlag2003

Identification and control of oxidative metabolism in Saccharomyces cerevisiae during transient growth using calorimetric measurements

Urs von Stockar

1998