Bioprocess Biosyst Eng (2004) 26: 165–168 DOI 10.1007/s00449-004-0351-7
O R I GI N A L P A P E R
R. Basse´guy Æ K. De´le´couls-Servat Æ A. Bergel
Glucose oxidase catalysed oxidation of glucose in a dialysis membrane electrochemical reactor (D-MER)
Received: 30 June 2003 / Accepted: 23 January 2004 / Published online: 10 March 2004 Springer-Verlag 2004
Abstract The purpose of this work was to evaluate the effectiveness of a new Membrane Electrochemical Reactor (MER) for the production of gluconic acid by glucose oxidase (GOD) catalysed glucose oxidation. The GOD was confined against the electrode surface with a dialysis membrane. The role of the electrochemical step was to eliminate by oxidation the hydrogen peroxide that appeared as a by-product of the reaction and strongly inhibited and/or inactivated GOD. The dialysis MER gave a transformation ratio of 30% with an initial glucose concentration of around 300 mM. This result is significantly better than the maximum of 10% obtained when hydrogen peroxide was eliminated by addition of a large excess of catalase in solution, as is generally done. The DMER also revealed unexpected properties of the enzyme kinetics, such as an oscillatory behaviour, which were discussed. Keywords Electrochemical engineering Æ Membrane reactor Æ Glucose oxidase Æ Oscillatory reaction
Adsorption, immobilisation, entrapment and covalent attachment have been widely used to deposit the enzyme on electrode surfaces. Recently membrane electrochemical reactors have been designed, where the enzyme is confined on the electrode surface with a dialysis membrane [2, 3] or an ultra-filtration membrane [4], which is closely applied against the electrode surface. Substrates and products of the reaction diffuse through the membrane. A Dialysis Membrane Electrochemical Reactor (D-MER) was used here to process the oxidation of glucose to gluconic acid catalysed by glucose oxidase (GOD): C6 H12 O6 þO2 þH2 O GOD ! C6 H12 O7 þH2 O2 This reaction produces hydrogen peroxide, which strongly inhibits and inactivates GOD. While the inhibition can be reversible or irreversible, the inactivation is irreversible due to irreversible amino acid modifications [5]. Thus hydrogen peroxide must be eliminated to protect the enzyme against degradation. In the D-MER, GOD is confined close to the electrode surface by the membrane, and hydrogen peroxide that is produced by the reaction is eliminated by electrochemical oxidation into oxygen:
Introduction
H2 O2 ! O2 þ 2Hþ þ 2e
Electro-enzymatic systems have been very successful in the field of electrochemical biosensors. Accordingly, the association of one (or more) electrochemical step(s) with one (or more) oxidoreductase-catalysed reaction(s) has now been emerging for synthesis. The electrochemical step is generally used to regenerate a cofactor without introducing any side-reaction [1]. Nevertheless, to develop economically efficient processes, the enzyme must be confined in the strict vicinity of the electrode surface.
Therefore, the GOD inhibition and inactivation by hydrogen peroxide should be avoided or drastically reduced. Generally glucose oxidation has been studied only with low glucose concentrations in the range of a few tens millimolar, where the inhibiting and inactivating effects of hydrogen peroxide remain rather low. On the contrary, using a D-MER allowed studying this reaction with higher glucose concentrations in the range of 250– 300 mM, which is closer to the concentrations used in industrial synthesis [6, 7].
R. Basse´guy (&) Æ K. De´le´couls-Servat Æ A. Bergel Laboratoire de Ge´nie Chimique, CNRS UMR 5503, Basso Cambo, 5 rue Paulin Talabot, BP1301, 31106 Toulouse, France E-mail:
[email protected] Fax: +33-5-34615253
Materials and methods The D-MER has been described elsewhere [3, 4, 8]. The working and auxiliary platinum grid electrodes (196 mesh cm)2, EngelhardCLAL) have 30 cm2 surface area. A dialysis cellulose membrane
166 (6–8 kDa and 12–14 kDa cut-off, Bioblock) was applied against the working electrode. Oxidation of hydrogen peroxide was performed by electrolysis at a constant potential of 0.45 V vs. saturated calomel reference electrode (SCE). A pump ensured the continuous circulation of the solution in a closed loop that involved the working compartment of the D-MER and a storage tank. Glucose concentrations were measured in the storage tank. Each experiment was performed in 0.1 M phosphate buffer pH 7.0. Chemicals, glucose oxidase (GOD, E.C. 1.1.3.4) from Aspergillus niger and catalase (E.C. 1.11.1.6, specific activity 2250 U/mg protein) from bovine liver were purchased from Sigma. GOD activity, which is defined as the number of micromoles of b-D-glucose consumed in one minute at pH 5.1 and 35C, was measured spectrophotometrically according to Sigma’s protocol. Glucose concentration was measured in the storage tank spectrophotochemically using the Sigma diagnostic kit HK [9].
Results and discussion Experiments were first carried out in a beaker containing 43 mL phosphate buffer pH 7.0, 290 mM glucose, and 100 U GOD to quantify the transformation ratio obtained in the absence of elimination of the hydrogen peroxide that is produced by the reaction. Supplying oxygen was achieved by air bubbling. Each test carried out in these conditions gave transformation ratios below 4% as reported in Fig. 1, even after more than 8 h. For comparison it may be noted that typical experiments reported in literature performed with lower glucose concentrations, on the order of 20 mM, gave transformation ratios which increased slowly until 45% in 8 hours for 0.56 U GOD in 2.5 mL solution volume [1]. This suggests that with high glucose concentrations the GOD molecules were immediately surrounded by a very high local concentration of hydrogen peroxide, which strongly reduced its activity. It is common in biocatalysis to eliminate hydrogen peroxide by adding catalase, which catalyses the dismutation of hydrogen peroxide into water and oxygen:
Fig. 1 Dimensionless evolution of glucose concentration as a function of time. In a beaker: without catalase (100 U GOD, 290 mM glucose, volume 43 mL) and · with catalase (4900 U catalase, 100 U GOD, 298 mM glucose, volume 10 mL). In the DMER (cut-off 12–14 kDa): n (112 U GOD, 248 mM glucose, volume 44 mL). All experiments were carried out in 0.1 M phosphate buffer pH 7.0
2H2 O2 catalase O2 þ 2H2 O !
ð3Þ
Adding 4900 U catalase in 10 mL solution containing 298 mM glucose and 100 U GOD, slightly increased the transformation ratios, which however remained lower than 10% (Fig. 1). Even large excesses of catalase did not manage to destroy the hydrogen peroxide fast enough to fully protect GOD against inactivation. These experiments can be compared to those of Hartmeier and Heinrichs [10] who suggested to improve catalase efficiency by co-immobilising catalase and the Gluconobacter oxydans cells containing GOD, in order to have catalase as close as possible to the site of hydrogen peroxide production. The D-MER was used to process 44 mL of 240– 300 mM glucose with 100–160 U GOD confined by the dialysis membrane in the reaction layer. Two representative curves are given in Fig. 2. After 3 h, 15% and 30% transformation were obtained when the reactor was equipped with a 6–8 kDa and a 12–14 kDa cut-off membrane respectively. The influence of the cut-off value showed that mass transfer through the membrane was an important control parameter. Independent experiments described elsewhere [3] were carried out using a rotating disk electrode to determine the diffusion coefficient of hexacyanoferrate (III) in these dialysis membranes according to the procedure given by Gough and Leypoldt [11]. With both membranes, the apparent diffusion coefficient was equal to 1.6·10)10 m2/s, but due to the difference of membrane thickness, the transfer coefficients were 3.5·10)6 m/s and 1.1·10)6 m/s for the 12–14 kDa membrane (46 lm thickness) and 6–8 kDa membrane (144 lm thickness) respectively proving the mass transfer was greater through the 12–14 kDa membrane. At the onset of the reaction, the transformation ratios were similar when catalase or when the electrochemical oxidation was used to eliminate hydrogen peroxide (Fig. 1). For longer reaction times, however, the
Fig. 2 Dimensionless evolution of glucose concentration as a function of time in the D-MER for two membranes of different cut-off. s: 6–8 kDa cut-off (125 U GOD, 250 mM glucose) and n: 12–14 kDa cut-off (112 U GOD, 248 mM glucose). Phosphate buffer 0.1 M pH 7.0, volume 44 mL. Error bars correspond to experimental errors at each point
167 Table 1 Results obtained in the D-MER for different experimental conditions
*Indicates that the C/C ratio presented a minimum, followed by a slight increase, but there was no marked oscillation
Table 2 Initial (Rinit) and maximal (Rmax) rate of glucose oxidation catalysed by GOD with different concentrations of glucose and gluconic acid
Experimental conditions Glucose (mM)
GOD (U)
292 286 246 248 245 244 254 240 274 266 250 380 280
158 100 158 112 158 100 111 111 111 111 125 100 117
Membrane
Maximal transformation for time (hours)
Threshold value for t=24 h
Oscillations
12–14 kDa
25% 24% 14% 30% 16% 14% 15% 14% 20% 14% 15% 18% 17%
30% 16% 11% 23% 10% 10% 12% 13% 16% 6% 10% 18% 10%
no no* no* yes yes yes yes yes yes yes yes yes yes
6–8 kDa
8h 18 h 18 h 3h 2 h, 45 min 2h 3h 4h 2h 8h 3h 4h 1h
Glucose concentration (mM)
Gluconic acid concentration (mM)
Rinit initial rate (mM/min)
Rmax maximal rate (mM/min)
93.5 200 300 200
0 0 0 100
0.051 0.059 0.058 0.054
0.111 0.119 0.113 0.109
D-MER protected the enzyme more effectively from inhibition and inactivation by hydrogen peroxide and led to transformation ratios of 30% in 3 h, while only 10% and 4% were reached with and without catalase respectively. Even the largest GOD activity (10 U/mL) used in the beaker in the presence of catalase did not allow reaching the transformation rate obtained in the D-MER with only 2.5 U/mL. A lot of different experimental conditions were tested in the D-MER and the results are reported in Table 1. Results are fairly dissimilar: transformation rates obtained with the 12–14 kDa membrane varied from 14% to 30% with 100– 110 U GOD and from 14% to 25% with 158 U GOD. It seemed that the GOD activity was not a key point in the reactor. The rather significant differences in the results may be due to a poor control of some other parameters, such as the inner geometrical arrangement of the reaction layer. The strict application of the membrane on the grid electrode and the pressure put on the rubber joint, for instance, are parameters that could modify the reaction layer thickness, the oxygen gradient in this layer, and the coupling conditions between the transfer rates through the membrane and the transfer rates from the electrode surface. The synthesis process in this reactor is a very complex combination of heterogeneous electron reaction, which occurs on a grid surface, the homogeneous enzyme-catalysed reaction and the transfer of reactants and products through the membrane. Variation of a few local geometrical parameters may significantly influence the final transformation ratio. For instance with the 6–8 kDa membrane the maximum ratio only varied from 15% to 18%. In this case the mass transfer through the membrane seems to become
rate-limiting and decreases the influence of the other parameters giving more reproducible results. However in most experiments beyond a few hours, oscillations and a threshold value were obtained as shown in Fig. 2. The oscillations meant that the concentration of glucose in the storage tank increased during some periods with a magnitude which was significantly higher than the experimental error. A threshold value meant that the transformation ratio remained constant when the experiment was pursued for a very long time. Several tracks were explored to explain these two phenomena. The possible inhibition of GOD by glucose was evaluated by measuring the maximal rate of glucose oxidation with different glucose concentrations. The procedure was identical to the measurement of GOD activity. The initial and maximal rates that were calculated for glucose concentrations varying from 93.5 mM to 300 mM were approximately identical (Table 2). Therefore no inhibition by glucose affected the process. Inhibition by gluconic acid was also estimated, following the same method. In order to come close to the experimental conditions, and since the inhibition by the glucose was rejected, the rates were determined with a solution containing simultaneously 100 mM gluconic acid and 200 mM glucose. The rates obtained with and without gluconic acid (Table 2) were also similar; consequently, inhibition by gluconic acid can also be neglected. Analysing the classic ‘initial kinetic’ equations [12] that are commonly used for low glucose concentrations and short reaction times showed that the oscillations observed on the glucose concentration–time curves cannot be explained with these classic equations, even when inhibition and inactivation of GOD by hydrogen
168
peroxide were introduced. Nevertheless, it has been experimentally demonstrated that oxidation of glucose catalysed by GOD in capillary spaces led to ‘‘anomalous’’ spatially resolved structures [13] that cannot be modelled with classic initial kinetics. pH changes could be evoked as another possible explanation of the oscillations [14]. But even introducing local pH changes in the kinetic expression cannot explain how glucose concentration could increase during the process. Very high concentrations of enzyme were also believed to provoke oscillations [15], but the range of GOD quantity used here are too low to evoke this possibility. As a conclusion, it may be thought that classic kinetic laws might no longer be valid for specific conditions, as here with high glucose concentrations and long reaction times. On the other hand GOD inactivation can be evoked to explain the threshold values. This inactivation involves two different processes: inactivation by hydrogen peroxide [5], and auto-inactivation, which is favoured by high concentrations of the two substrates glucose and oxygen [16, 17, 18]. It is difficult to estimate auto-inactivation, but it was demonstrated that GOD immobilised as a monolayer on glassy carbon, led to a maximal enzyme turnover of 107, irrespective of the concentration of both substrates [16]. The maximal transformation ratio obtained here was 30% with an initial 248 mM glucose concentration and with 112 U GOD in the D-MER. This gave a turnover of 1.7·106 (molar mass of GOD=186,000; specific activity=250 U/mg solid; 80% protein; D-MER volume=44 mL). This maximal turnover was close to the maximal value of 107estimated for adsorbed GOD. Autoinactivation of GOD could consequently be responsible for the threshold phenomenon. The D-MER was quite efficient to protect GOD against inhibition/inactivation by hydrogen peroxide, and the main limit of the transformation seems to be only linked to the auto-inactivation of the enzyme. The D-MER allowed reaching very fast the limits imposed by the intrinsic properties of the enzyme. To overcome this obstacle the enzyme kinetics itself should be modified. For instance Fortier and Be´langer [19] showed that the use of polypyrrole as immobilisation matrix gives a two-fold decrease of the auto-inactivation rate, as compared to the soluble GOD. Bourdillon et al. [18] utilized also an artificial electron acceptor to avoid oxygen entirely, improving the GOD stability and increasing the production yield from 12 to over 200 g gluconic acid produced per mg GOD used.
Conclusion The dialysis membrane electrochemical reactor (DMER) was revealed to be very efficient in protecting GOD against inhibition/inactivation by hydrogen peroxide. It should consequently open new promising opportunities for electro-enzymatic synthesis. It also revealed a new behaviour of the GOD catalysed oxidation of glucose. The threshold value may be explained by auto-inactivation of the enzyme, but the reproducible
oscillations that were observed suggest that classical enzymatic kinetic laws are no longer valid with high glucose concentrations. Further insight into the enzyme kinetics would be necessary to explain the oscillation phenomenon, and the D-MER could also be a powerful analytical tool to pursue this study.
References 1. Devaux-Basseguy R, Gros P, Bergel A (1997) Electroenzymatic processes: a clean technology alternative for highly selective synthesis? J Chem Tech Biotechnol 68:389–396 2. Grimes MT, Drueckhammer DG (1993) Membrane-enclosed electroenzymatic catalysis with a low molecular weight electron-transfer mediator. J Org Chem 58:6148–6150 3. De´le´couls-Servat K, Bergel A, Basse´guy R (2004) Designing and modelling a dialysis membrane electrochemical reactor (D-MER) for oxidoreductase-catalysed synthesis. J Appl Electrochem (2004, in press) 4. De´le´couls-Servat K, Basse´guy R, Bergel A (2002) Membrane Electrochemical Reactor (MER): application to NADH regeneration for ADH-catalysed synthesis. Chem Eng Sci 57:4633–4642 5. Kleppe K (1966) The effect of hydrogen peroxide on glucose oxidase from Aspergillus niger. Biochemistry 5:139-143 6. Buzzini P, Gobbetti M, Rossi J, Haznedari S (1993) Gluconic acid production in batch fermentations from concentrated rectified grape must by Gluconobacter oxydans associated with Aspergillus niger. Ann Microb Enzimol 43:195–205 7. Sankpal NV, Kulkarni BD (2002) Optimization of fermentation conditions for gluconic acid production using Aspergillus niger immobilized on cellulose microfibrils. Process Biochem 37:1343–1350 8. De´le´couls-Servat K, Bergel A, Basse´guy R (2003) Membrane electrochemical reactors (MER) for NADH regeneration in HLADH-catalysed synthesis: comparison of effectiveness. Bioprocess Biosyst Eng (submitted) 9. Kunst A et al (1984) Methods of enzymatic analysis, vol 6, 3rd edn. VCH, Basel, pp 163–172 10. Hartmeier W, Heinrichs A (1986) Membrane enclosed alginate beads containing gluconobacter cells and molecular dispersed catalase. Biotechnol Lett 8:565–570 11. Gough DA, Leypoldt JK (1980) A novel rotated disc electrode and time lag method for characterizing mass transport in liquid-membrane systems. Am Inst Chem Eng J 26:1013–1019 12. Dixon M, Webb EC (1979) Enzymes, 3rd edn. Longman, New York 13. Lin AL, Feldman MS, Kopelman R (1997) Spatially resolved anomalous kinetics of a catalytic reaction: enzymatic glucose oxidation in capillary spaces. J Phys Chem 101:7881–7884 14. Ricard J, Crasnier M (1984) pH modulation of transient state kinetics of enzymes, Part I: simple theoretical models. Biophys Chem 19:85–91 15. Ryde-Pettersson U (1992) Oscillations in coupled enzymic reactions at high concentration of enzyme. Biochem Biophys Acta 1119:118–122 16. Bourdillon C, Thomas V, Thomas D (1982) Electrochemical study of D-glucose oxidase autoinactivation. Enzyme Microb Technol 4:175–180 17. Tse PHS, Gough DA (1987) Time-dependent inactivation of immobilized glucose oxidase and catalase. Biotechnol Bioeng 29:705–713 18. Bourdillon C, Lortie R, Laval JM (1988) Gluconic acid production by an immobilized glucose oxidase reactor with electrochemical regeneration of an artificial electron acceptor. Biotechnol Bioeng 31:553–558 19. Fortier G, Be´langer D (1991) Characterization of the biochemical behavior of glucose oxidase entrapped in a polypyrrole film. Biotechnol Bioeng 37:854–958
