Development and Applications of a Four-Channel Enzyme Thermistor System for Bioprocess Control

Development and Applications of a Four-Channel Enzyme Thermistor System for Bioprocess Control H.-G. Hundeck, U. Hubner, A. Lübbert, T. Scheper, J. Schmidt, M. Weiß Institut für Technische Chemie, Universität Hannover, Callinstr. 3, D-3000 Hannover 1 F. Schubert Zentralinstitut für Mikrobiologie, Robert-Rössle-Str., D-1115 Berlin-Buch Summary A large number of papers on biosensors have been published in the last few years. However, few of these sophisticated analysis systems have been used to monitor real bioprocesses. In this paper, a newly developed four-channel enzyme thermistor system and its application for biotechnological process monitoring is presented and discussed. Different sugars were detected simultaneously and online during the cultivation of Spodoptera frugiperda and Bacillus licheniformis in technical media. Immobilized enzymes and entrapped microorganisms were used as biological compound in this biosensor. In addition, enantioselective analysis was performed by two enzyme reactions. For example, the detection of D,L-racemates of aminoacid esters was presented in an aqueous system. Futhermore the possibility of using this detection system in organic solvents was shown. Introduction On-line monitoring techniques are important for optimal bioprocess control and automation. This type of measurement places special demands on the types of sensors involved (e.g., long termstability, analytical accuracy, and automated analytical procedure). A major focus of biotechnological analysis is the cultivation medium, since substrates and products are dissolved in it. The environment of the cells to be cultured influences the activity and the state of the biomass and thus of the whole bioprocess. Often, the analysis of the cultivation medium is performed by complex off-line methods. Biosensors provide the possibility of measuring such compounds with on-line systems, and greatly increase the power of automated bioprocess monitoring and control. The literature on biosensors is overwhelming (1-6), but there is still a lack of applications in real industrial processes (7). The four-channel enzyme thermistor system is run as a stand-alone device, controlled by a computer. The simultaneous analysis of different monoand disaccharides such as glucose, maltose, sucrose, and lactose with this version was simply 322 H.-G. Hundeck et al. performed at cultivation processes over periods of up to 300 hours and can thus be used as a basis for process optimization. Material and Methods Materials Glucose oxidase (EC 1.1.3.4) and catalase (EC 1.11.1.6) were used for glucose analysis. Invertase (EC 3.2.1.26) was used for sucrose analysis, and a-glucosidase (EC 3.2.1.20) was employed for maltose analysis. The enzymes were covalently bound on oxirane acrylic supports (VA Epoxy Biosynth, Riedel de Haen AG, W. Germany) (8). This kind of immobilization is more efficient than the usual CPG immobilization (9). The organisms used in the sensor (e.g., cells of Saccharomyces cerevisiae) were immobilized in calcium alginate (10). The Enzyme Thermistor The analytical system is based on a simple isoperibol flow calorimeter, in which the heat produced is measured as a temperature change by thermistors (11). The temperature resolution is about 10-5 K. A schematic diagram of an enzyme thermistor is shown in Figure 1. reference enzyme column gold tube A— LEA ELL LE „a a Y A A / ; J Y 5 YZ ’ 7 4} 7 | 7 7 thermistor i) ] a heat exchange 4 A A y A rt aluminium cylinder 5 A Vv 1 [A insulation VLAZZLZ. LAL LZ] a © 1 4 sample injection buffer Ki ee buffer Fig. 1 Schematic diagram of an enzyme thermistor. A Four-Channel Enzyme Thermistor System for Bioprocess Control 323 Buffer is pumped continuously through the thermostated aluminium cylinder. Before entering the columns, the carrier flow passes through a heat exchanger coil. The temperature at the inlet and outlet of the column is registered by thermistors placed on gold tubes. One of the columns contains the immobilized biological material (e.g., enzymes), while the other column is filled with inactivated resin material. In this reference column, nonspecific heat effects are measured. Buffer is pumped continuously through the analysis system. A defined sample volume is injected by an automated valve into the buffer stream according to the principles of flow injection analysis (12). The carrier flow transports the sample through the columns; the reaction occurs in the enzyme column. The heat produced here is measured by the thermistors. analog amplifier 68000 microprocessor Fig. 2 Computer control system of the four-channel enzyme thermistor. A four-channel version of an enzyme thermistor was designed and built in the Institut für Technische Chemie as a flow calorimeter, based on the experiences with Lund-type enzyme thermistors (13-14). It can analyze four different substances simultaneously. After changing an enzyme column, only 15 minutes are necessary before the analysis can be continued. Special columns have been designed for the detection of dissolved enzyme activities. In 324 H.-G. Hundeck et al. these columns, substrate and enzymes are mixed together . The heat measured is proportional to the enzyme activity.Integrated prethermostating creates a very stable temperature at the measurement center. A microprocessor control unit is utilized for the signal detection (Figure 2) . The user can define the analysis channels with the software. He can choose a reference column, a measurement between two thermistors of the same column, or other special measurement applications. The control system checks the measurement accuracy, calibration, and data analysis and processing. The 68000 microprocessor system with a multi-tasking operating system allows the implementation of contro! algorithms for sophisticated process control; e.g., the regulator controls the substrate feed for bioprocesses via the pump flow. A special filtration sampling probe was integrated into the whole analytical system. A cell-free sample can be withdrawn from the fermentor continuously for injection into the enzyme thermistor. Dynamic or static dilution steps are initiated by the computer if necessary. The sample is injected into the carrier buffer stream via a computer-controlled injection valve. Results and Discussion Cultivation of Bacillus licheniformis The cultivation of Bacillus licheniformis was performed for protease production. Batch and fed-batch cultivations in complex technical media were investigated. The fermentation media contained hydrolyzed corn starch, soy meal, casein, and corn steep liquor .The information obtained in the batch cultivations provided the basis for the fed-batch investigations. In the beginning of the process, biomass is produced in a pure growth phase. The protease production phase starts when the glucose concentration falls below a certain concentration and the microorganisms excrete proteases into the medium to assimilate protein. Sucrose or maltose can be added during the production phase. This substrate feed must be controlled for optimal protease yield without any lag phase. If the sugar concentration is too high, biomass will be produced and the protease production will decrease. On the other hand, protease production will decrease when the substrate concentration is below a critical level that results in decreasing cell activity. Figure 3 shows the simultaneous on-line analysis of maltose, sucrose and glucose with the four-channel enzyme thermistor for a batch cultivation of Bacillus licheniformis. A comparison of the online data with different off-line methods was presented before (15). An interesting relation between the sugar data and the oxygen transfer A Four-Channel Enzyme Thermistor System for Bioprocess Control 325 aD T T T! 3 | | - | & Pur a Ba I ||| ES | A: pated = | | 1 | ane = a 5 35[628 Pr ® 3 742 2 3 15 E D> . 2 o 3 a 1 10 2 iq 45 5 { | 8 0 05 10 1 3 fermentation time (h) Fig. 3 Glucose( o)-, maltose( ¢ )- und sucrose concentration (=) during a batch cultivation of B. licheniformis compared with the oxygen transfer rate (OTR). rate (OTR) can be observed. In a first phase the sucrose consumption was not correlated with a change in the OTR amount. After the complete consumption of sucrose, the consumption of glucose with a linear increase of the OTR start. After a lag phase caused by the complete glucose consumption, the OTR data increase is longer. A fedbatch cultivation monitored with the four-channel enzyme thermistor is presented in Figure 4. In comparison between maltose concentration and culture fluorescence signal demonstrates an important difference between the batch and the fed-batch process. In the fed-batch process the culture fluorescence signal increases earlier, and reaches higher level. The culture fluorescence Signal is correlated with the amount of living cells and thus is an indicator for ceil growth. The results of this Study indicate that maltose as C-source is responsible for the cell growth, because the culture fluorescene increases with the maltose consumption. A small peak in the culture fluorescence signal can be observed after the complete consumption of glucose, and indicates the change from glucose to maltose consumption. This peak appears earlier in the fed-batch process, causing a smaller amount of glucose in the start medium and indicates the rapid consumption of glucose. During earlier experiments it has become obvious that the multiple addition of maltose has no positive effect on the production of proteases. However, a detailed analysis with the enzyme thermistor was successfully performed for sucrose analysis. The results of this 326 H.-G. Hundeck et al. study indicated that the addition of sucrose should improve the ane | | ® oO o 5 | 3 @ | Ss = | © 3 3 = 55,68 5 2 8 20 8 }5 2 < 42 3 15 E mo era: . +10 r2 r5 4 0 + - o0zfr0S Ten: fr fed-batch-cultivation 20 3 5 3 iss [43 E a °, 0: I?» r 2 5 b 4 "0 0 0 5 10 15 fermentation time (h) Fig. 4 Glucose( no )-, maltose( © )- und sucrose concentration ( =) during a batch and a fed-batch cultivation of B. licheniformis compared with the culture fluorescence signal. protease production. Two fed-batch fermentations with control of sucrose addition with the four-channel enzyme thermistor are presented in Figure 5. If the sucrose concentration is too high, biomass will be produced while the protease production decreases. The protease yield was measured on-line by a stopped-flow-FIA (16). The data of the protease concentration were fitted and differtentiated to get the protease production rate. Immobilized whole cells of Saccharomyces cerevisiae were used instead of the enzyme columns to monitor the concentration of assimilable sugars during a cultivation. The cells were immobilized in calcium alginate. The heat produced by the immobilized cells after addition of different monoand disaccharides is shown in Figure 6 as temperature change versus concentration. The immobilized cells produce heat during the assimilation of different sugars in the medium. Thus, this microbial sensor doesn’t detect a single compound but the A Four-Channel Enzyme Thermistor System for Bioprocess Control 327 amount of all assimilable substances in the medium. This parameter is ie 80 7 : 2 a © fi «604 gh : oe ae io | & er - y } a Y wre ne % =; a ° ne 000° u ES FIRE rf Ea oa @) Sate En 516 RR one 5) en nl : 10 20 a fermentation time (h) Fig. 5 Comparison of on-line data for sucrose determination using the enzyme thermistor with protease production rate during two fedbatch cultivations of B. licheniformis. < 10 E © oe Ro} = 4 6 4 = © 4 glucose 7 44 * sucrose = J = fructose = 23 maltose 4 oF 2 3 4 concentration (g/l) Fig. 6 Heat evolution of different sugars measured with immobilized cells of S. cerevisiae. highly interesting for use in technical media, in which the concentration distribution of different substrates is often nearly undefined. It would be preferable to use the same organisms in the microbial sensor as in the cultivation to be monitored. However, the use of yeast cells demonstrates the principle and potential of this analytical method. The 328 H.-G. Hundeck et al. data of the microbial sensor during batch cultivations were shown before (15). Enantioselective analysis The principle for detection of enantiomeric excess is based on two enzyme reactions (Figure 7). One of the used enzymes (achymotrypsin) can only react with one enantiomer, preferably the Lstereospecific compound. The other used enzyme reacts with both stereospecific compounds. Both enzymes react with the sample. D,L-aminoacid ester + 0 ES RCOO’+H"+ ROH a-chymotrypsin L-aminoacid ester + H,O > RCOO’+ H'+ ROH Fig. 7 Enzyme reactions for enantiomeric analysis The heat generated during the reaction is measured via the enzyme thermistor. The concentration of both enantiomeric compounds was obtained by this reaction and the enantiomeric excess was calculated from this data. The temperature signal increases in a linear way by a reaction of a-chymotrypsin with different concentrations of a 50:50 racemate of D,L-esters, the reaction with the esterase results in a twice as high temperature signal (Figure 8). Ss 40 E SC 5 B 304 © a © a wee § 3 104 04 - T 0 10 20 phenylalaninemethylester (mM) Fig. 8 Calibration function for reaction of an aminoacidester 1.)L- (=) und D,L-esters (+) with a-chymotrypsin 2.) L- (a) und D,L-esters ( + ) with esterase A Four-Channel Enzyme Thermistor System for Bioprocess Control 329 By using the L-ester both enzyme reactions give the same temperature signal. By variation of the D-compound in a mixture of both enantiomeres it could impressively be shown that the temperature signal of the a-chymotrypsin reaction depends on the amount of the Dcompound and the temperature signal of the esterase reaction was constant caused by the same concentration as a whole, of both compounds (Figure 9). By combining the whole concentration of D,Lcompounds, obtained by the esterase reaction and the concentration of the L-compound, available from the a-chymotrypsin reaction, it is possible to calculate the enantiomeric excess. This procedure is a new principle to analyze enantiomeric mixtures. The same reaction functioned well in solutions that contained dimethyl formamide. The results of this study indicated that enantiospecific detection is possible also in organic solvents. 40 y = 31,405 - 4,2857e-4x R*2 = 0,000 a a z 7 a y = 27,362 - 0,27457x R*2 = 0,996 temperature signal (mK) wo oO 1 10 J 0 20 40 60 80 100 ec (%) Fig. 9 Calibration function for a reaction of an aminoacidester racemate by variation of the D-compond with a-Chymo trypsin ( « ) und Esterase ( = ) Summary A biosensor system was developed and sucessfully applied for Cultivation monitoring in biotechnology. To achieve high reliability and long term stability, the four-channel enzyme thermistor system was integrated into a FIA system. The entire biosensor system included cellfree sampling, sample conditioning, and analysis. The system was automated and could be run as a stand-alone version for on-line monitoring. The system offers the possibility to obtain a more detailed insight into the whole bioprocesses of Spodoptera frugiperda and Bacillus licheniformis and can serve as a basis for process control.