Simulation of a reactor for glucose isomerization to fructose by immobilized glucose isomerase with continuous enzyme renewal

Transcription

1 Biotechnology Techniques, Vol, No 2, February 997, pp Simulation of a reactor for glucose isomerization to fructose by immobilized glucose isomerase with continuous enzyme renewal A. Converti*, A. Fisichella, A. Riscolo and M. Del Borghi Institute of Chemical and Process Engineering, University of Genoa, Via Opera Pia, 5, 645 Genoa, Italy A. Scaringi and D. Carbone Roquette Italia SpA, Via Serravalle, 26, 5063 Cassano Spinola, AL, Italy Two different dispositions of laboratory-scaled columns have been tested to simulate the isomerization of glucose to fructose in a mobile bed reactor where exhausted immobilized glucose isomerase is continuously renewed. If the simulation columns working at 65 C are arranged in parallel and connected to a section for final enzyme exploitation at 75 C, a syrup with constant composition can be produced, at relatively constant total throughput, by feeding the individual columns at flow rate decreasing according to the enzyme decay profile and following a programmed disphased mode of operation. 24 pts min base to base from Key words to line of text Introduction Fructose and isoglucose are nowadays produced in EU by enzymatic isomerization of syrups with high glucose content coming from starch hydrolysis. This process is performed by feeding the syrup into a column filled with immobilized glucose isomerase (GI), at a residence time consistent with the production of a stream containing no less than 42% fructose (Rechigl, 982). However, the performance of these reactors is largely influenced either by the diffusion resistance (Chen and Chang, 984) or by the thermal inactivation of the enzyme (van den Tweel et al., 993). Several authors demonstrated that GI belongs to that category of enzymes which are protected by substrate, thanks to the stabilization of tertiary structure of the complex between active site and glucose (Chen and Wu, 987). So, Illanes et al. (992) and Houng et al. (993) proposed comprehensive methodologies to take into account this effect in immobilized GI reactor design and modelling. Unfortunately, however, this effect does not seem to be exploitable to increase immobilized reactor productivity in continuous process (Converti and Del Borghi, 996.a,b). To optimize glucose isomerization under industrial conditions, two different laboratory-scaled configurations are compared in this work, each simulating the performance of an immobilized enzyme reactor where the exhausted enzyme is progressively replaced by fresh bio-catalyst. Materials and methods Substrate preparation A corn starch hydrolysate, whose composition is listed in Table, has been used as raw material. It was produced in the laboratories of Roquette Italia, Cassano Spinola, Italy, by enzymatic hydrolysis of corn starch, consisting in a preliminary treatment with a heatresistant -amylase and a subsequent saccharification stage at 60 C with amyloglucosidase. The medium used for continuous isomerization tests was obtained by adding Na 2 SO 3 and MgSO 4 7 H 2 O, each up to 0.4% by weight with respect to dry matter, and adjusting the ph at 8.0 ± 0. with a 0. N NaOH solution. Enzyme The commercial Sweetzyme T, supplied by Novo Nordisk Bioindustriale, Milan, is an active immobilized glucose isomerase (EC D-xylose ketolisomerase) coming from a selected strain of Streptomyces murinus supported on silica. It consisted of dry, brown, cylindershaped granules with a particle size range of 0.3 to.0 mm, which was submitted to re-hydration in distilled water for 24 hours before use. 997 Chapman & Hall Biotechnology Techniques Vol No

2 A. Converti et al. Table Composition and chemical properties of starch hydrolysate. Percentages (w/w) refer to dry matter D-Glucose 92 to 93% Maltose 4 to 5% Polysaccharides 2 to 4% Fructose < 0.% ph 3.9 to 4.3 Dry matter 4 to 50% [ ]20 /D Analytical methods Fructose and glucose were determined by means of a HPLC Waters ALC 20 using an IR-detector. A column Bio-Rad HPX-87C was used with bidistilled water as mobile phase at a flow rate of 0.5 ml/min. Activity assay A solution was prepared by adding 450 g of glucose to litre of a 0.05 M Tris-buffer solution at ph 7.0 containing 20 g of MgSO 4 7 H 2 O and 20 g of Na 2 SO 3. The activity tests were performed in 25 ml-erlenmeyer flasks where 0.6 g of fresh immobilized enzyme were added to 50 ml of the above solution and shaking the flasks in a water bath at 60 C. Samples were periodically analysed for glucose and fructose to estimate the starting reaction rate. The residual activity of the enzyme, which was periodically checked in the same way after withdrawing catalyst samples from the column, was referred to the starting activity of the fresh enzyme. Experimental set-up and operating conditions To simulate a mobile bed non isothermal reactor (Fig. ), three 23 cm-high glass columns, with internal diameter of.97 cm, provided with external jackets for individual temperature regulation, were filled up to about one half of the working volume with 0 g of the immobilized enzyme and connected in series to each other. The columns were thermostated at 55, 65 and 75 C, respectively, and continuously fed with the nutrient. Flow rate was regulated at 3.7 ml/min. When the enzyme activity in the last column fell below 0 20% of the starting value, the immobilized enzyme was renewed and the column shifted from the last place to the first one and thermostated at the lowest temperature (55 C). Therefore, the first column became the second and started to work at the intermediate temperature, and so on. A further plant scheme able to always ensure optimum ph and temperature conditions was also simulated (Fig. 2) arranging in parallel three columns working at 00 Biotechnology Techniques Vol No Figure Simulation scheme of a mobile bed nonisothermal isomerization column. Figure 2 Simulation scheme of an isomerization system provided with a secondary column (75 C) for final enzyme exploitation. 65 C and fed at.3 ml/min. The enzyme utilized in the principal columns at 65 C was passed into the secondary columns kept at 75 C to exploit to the utmost the GI activity. The investigation has been restricted to a portion of the plant constituted by one couple of principal and secondary columns. Results and discussion The values of fructose yield listed in Table 2 have been obtained by filling an immobilized GI column with the nutrient and recirculating the effluent stream until a constant composition was achieved. According to the equilibrium theory (Roels, 983), the fraction of glucose which can be isomerized to fructose by a continuous process, f, can never exceed the ratio of fructose to

3 Simulation of a reactor for glucose isomerization to fructose by immobilized glucose isomerase with continuous enzyme renewal Table 2 Temperature dependence of the fraction (f) of glucose isomerized in a batch immobilized GI reactor T ( C) f glucose concentrations at the equilibrium, corresponding to the equilibrium constant: f f max = F/G = K. So, the related equilibrium curve, not depending on the process kinetics, has been assumed to be the maximum threshold to which the plant performance may tend. Simulation of a mobile-bed non-isothermal column A mobile-bed non-isothermal isomerization column (Fig. ) was simulated by an in-series disposition which allowed to separately evaluate the performance of each plant section working at different temperature, namely 55, 65 and 75 C. Fig. 3 shows the results of continuous isomerization runs in terms of fructose fraction in the streams leaving each section and entering the successive one, with respect to the glucose content of the feed. As can be seen, the isomerization yield not only rapidly reaches relatively constant values in each column, but also increases from about 0.8 to 0.30 and 0.42 g/g passing from the st to the 2nd and 3rd columns, respectively, due to the simultaneous increases in both temperature (from 55 to 75 C) and overall residence time (from 0.57 to h). Comparing these results with those of Table 2, it is evident that the proposed system does not allow fructose yields in the product exceeding 75% of the theoretical equilibrium value. Moreover, a consistent acidification of the product was observed, which is expected to remarkably contribute to enzyme inactivation. In fact, the relative stability of the GI used in this study, which achieves its optimum at ph 7.3, is so sensitive to acidity as to decrease up to about 50% at ph 6.5. This result suggests that the system did not operate under optimum conditions, since the st section worked at satisfactory ph (8.25) but at rather low temperature (55 C), the 2nd section at a temperature near the optimum compromise between activity and stability (65 C) but at ph never exceeding 7.5, and the 3rd one Figure 3 Performance of a simulated mobile bed nonisotermal isomerization column. ( ) st section at 55 C; ( ) 2nd section at 65 C; ( ) 3rd section at 75 C. Arrows indicate enzyme renewals. at suboptimal values of both parameters (ph 5.8 and T = 75 C). Simulation of an isothermal column with final section at 75 C To overcome the above problems related to the strong acidification in the st stage, an alternative simulation scheme has been investigated, which differs from the preceding one because of the simultaneous presence of a set of 3 columns all working under optimum ph and temperature for the isomerization as well as a further set of 3 columns where the residual activity of the exhausted enzyme can be exploited as much as possible at 75 C (see Fig. 2). Figure 4 shows the fructose yield during continuous operation carried out with only one enzyme charge and one pair of columns working at 60 and 75 C, respectively. While the overall enzyme activity progressively decreases during the run because of the enzyme inactivation, different behaviours can be observed for the principal and the secondary columns. In particular, it is evident that the fructose yield of the principal column working at suitable temperature (T = 65 C) kept almost constant at the starting maximum value (about 0.40 g/g) for no less than 25 days, while, as expected from a more remarkable thermal inactivation, that of the secondary column working at the higher temperature decreased since the beginning. Biotechnology Techniques Vol No

4 A. Converti et al. Figure 4 Continuous isomerization runs in one pair of principal and secondary columns. Fructose yield: ( ) principal column; ( ) secondary column; ( ) overall system. Figure 5 Behaviours of the specific feed flow rate and the enzyme relative activity in the principal column and in the overall system. However, as constancy of product composition is one of the most important requirements for industrial application, the investigation on this simulation scheme, which showed higher overall yields with respect to the in-series arrangement, was addressed to the search of an operating way capable of ensuring a constant fructose yield. Assuming first order kinetics for the isomerization, such an aim can be reached, as suggested by Illanes et al. (992), by progressively lowering the influent flow rate according to the same law describing the enzyme inactivation. It was demonstrated that also immobilized GI inactivation can be described by first-order kinetics and that the relative inactivation constant, k f, can easily be calculated by plotting, versus time, the experimental data of the activity coefficient,, defined as the ratio of the active enzyme concentration at a given time to that at the beginning (Chen and Wu, 987). Following this procedure, k f values of and d have been calculated for the principal column and for the system with final recovery at 75 C, respectively. Related enzyme half lifes ( 0 ) of 36 and 44 days have then been calculated considering that 0 = ln2/k f. It is evident from these values that the enzyme is much more stable in the principal column working under optimum conditions than in the secondary column, which is used at the higher temperature to exploit the enzyme activity to the utmost. This result is in accordance with the yield behaviours of Figure Biotechnology Techniques Vol No According to Illanes et al. (992), a constant product yield can be ensured by progressively lowering the feed flow rate, F, so as to follow the first-order enzyme decay profile. Figure 5 shows the behaviour of the feed flow rate, referred to.0 kg of fresh enzyme, which would be required to obtain syrups with constant composition either by the principal column (42% fructose) or by the overall system (48% fructose). As shown, due to the enzyme inactivation, the ratio between the specific flow rates of these systems varies from, at the beginning, to about 2.5 after 300 days. To evaluate which configuration is more profitable and then to estimate the optimal number of reactors necessary to obtain syrups with the above constant compositions, the specific productivities of both systems have been estimated for a number of reactors varying from to 8. As shown in Figure 6, the system consisting only in the principal column, ensuring a fructose yield of 42%, was not able to reach a specific productivity of 4 tons/y per kg of fresh enzyme even when using a set of 8 columns arranged in parallel. The same unsatisfactory result was obtained both replacing the exhausted enzyme with a fresh charge whenever the working time achieved the enzyme half life (36 days) and by replacing it more frequently (44 days). For the system constituted by both principal and secondary sections, the number of columns refers to column couples ensuring a fructose yield of 48%. As

5 Simulation of a reactor for glucose isomerization to fructose by immobilized glucose isomerase with continuous enzyme renewal Figure 6 Influence of the number of columns on fructose productivity. ( ) Only principal columns; ( ) system with principal and secondary columns. Renewal frequency: (open) 44 d; (full) 36 d. shown in the same figure, both enzyme renewals after 44 or 36 days reach specific productivities up to 7% higher than the system lacking in the secondary column. The curve exhibiting the highest productivity, among those shown in Fig. 6, is the one corresponding to the in-parallel arrangement of couples of principal and secondary columns, replacing the exhaust enzyme with fresh enzyme every 44 days. Moreover, in both cases, the lower the column number, the more the performances of these two operations differ. Taking into account the low significance of the enzyme cost to the overall costs, it is now possible to maximize the productivity of the system by selecting the most suitable configuration on this curve to the detriment of the total consumption of the enzyme. For example, a productivity of 5. tons/y.kg, that is only 6.6% less than that attainable with 8 column pairs, can be obtained employing a system constituted by 3 parallel principal columns, working at 65 C and where the enzyme is renewed every 44 days, connected in series to a set of 3 secondary columns, working at 75 C, where the enzyme coming from the principal columns is moved. If one considers a production of 20,000 tons/year of a syrup containing 48% fructose, that is a value which approximates the Italian production quota (Riscolo and Fisichella, 996), a consumption only of 300 kg of immobilized GI would be sufficient, according to the proposed plant scheme, to ensure this prodution threshold. References Chen, K.-C., Chang, C.M Operational stability of immobilized D-glucose isomerase in a continuous feed stirred tank reactor. Enzyme Microb. Technol. 6: Chen, K.-C., Wu, J.-Y Substrate protection of immobilized glucose isomerase. Biotechnol. Bioeng. 30: Converti, A., Del Borghi, M. 996.a. Thermodynamic study of glucose isomerization to fructose by immobilized glucose isomerase [in Italian]. Proc. 5th Scientific Congress of the Italian Society of General Microbiology and Microbial Biotechnology. Abbadia San Salvatore, SI, 8 October 996, Poster session n. 4. Converti, A., Del Borghi, M. 996.b. Simultaneous effects of immobilization and substrate protection on the thermodynamics of glucose isomerase activity and inactivation. Enzyme Microb. Technol. Submitted. Economic European Community 98. Regulament n. 785/8. Bruxelles, 30 June 98. Houng, J.-Y., Yu, H.-J., Chen, K.-C Analysis of substrate protection of an immobilized glucose isomerase reactor. Biotechnol. Bioeng. 4: Illanes, A., Zuñiga, M.E., Contreras, S., Guerrero, A Reactor design for the enzymatic isomerization of glucose to fructose. Bioproc. Eng. 7: Rechigl, M Handbook of nutritive value of processed foods. CRC Press, Boca Raton, pp Riscolo, A., Fisichella, A Optimization of the continuous process of glucose isomerization to fructose by immobilized isomerase [in Italian]. Thesis and Dissertation, Genoa University, Italy. Roels, J.A Energetics in biotechnology. Elsevier, Amsterdam. van den Tweel, W.J., Harder, A., Buitelaar, R.M Stability and stabilization of enzymes. Elsevier, Amsterdam. Received as Revised 2 January 997 Biotechnology Techniques Vol No