Stabilization of Glucose Dehydrogenase with Polyethyleneimine in an Electrochemical Reactor with NAD(P) + Regeneration Jose ´ Marı ´a Obo ´ n, † Pastora Casanova, † Arturo Manjo ´ n, † Vı ´ctor M. Ferna ´ ndez, ‡ and Jose ´ Luis Iborra* ,†,§ Departamento de Bioquı ´mica y Biologı ´a Molecular B e Inmunologı ´a, Facultad de Quı ´mica, Universidad de Murcia, Apdo. correos 4021, 30100 Murcia, Spain, and Instituto de Cata ´ lisis, CSIC, Campus de la Universidad Auto ´noma, Cantoblanco, 28049 Madrid, Spain The stability of the enzyme glucose dehydrogenase (GDH) has been studied under turnover conditions in an electrochemical reactor with NAD(P) + regeneration on a preparative scale. The enzyme showed first-order deactivation patterns closely related to imposed potential. An increase in the applied potential caused a decrease of the half-life deactivation time of the enzyme (t 1/2 ). However, this detrimental effect was compensated with an enhancement of the substrate consumption rate (r s ) attained as a consequence of the higher cofactor regeneration rates observed at more positive potentials. A 0.7 V potential (vs Ag|AgCl) was selected as a compromise between the activity and the stability of the enzyme (t 1/2 ) 4.2 h; r s ) 32 µmol min -1 ). The protective effect on the activity of glucose dehydrogenase of well-known stabilizing agents such as NaCl, sorbitol, bovine serum albumin (BSA) or polyethyleneimine (PEI) has been studied. PEI (50 000 MW) at concentrations between 0.3 and 0.5 mM showed the highest protection of the enzyme activity in the electrochemical reactor as well as the highest substrate consumption rates (t 1/2 ) 24.5 h; r s ) 59 µmol min -1 ). This beneficial effect of PEI is explained in terms of an electrode, cofactor, and enzyme modification that induces an increase of the concentrations of NAD(P)H and glucose dehydrogenase in the vicinity of the electrode and minimizes the adsorption of the enzyme on the electrode contact. Introduction Enzyme technology has experienced a considerable interest for commercial scale-up during the past few years. However, in cases in which enzymes require a coenzyme, there are difficulties in its application. En- zymatic electrocatalysis is an interesting approach able to be applied for the efficient use of NAD(P)(H) dependent enzymes in biocatalysis (Laval et al., 1984; Biade et al., 1992; Bourdillon, 1992; Lortie et al., 1992). It combines the enzymatic catalysis and the electrochemical regen- eration of the expensive coenzyme in such a way that solves the main limitation mentioned of these enzymes for their economical feasibility in organic synthesis: the stoichiometric coenzyme consumption during the reac- tion. Furthermore, it avoids the downstream separation of the substrate and product required when the widely applied enzymatical regeneration methods are used (Lee and Whitesides, 1985). The direct electrochemical oxidation of NAD(P)H gives an almost fully active coenzyme, offering in compact batch electrochemical reactors high coenzyme turnover numbers (>10 000) and reasonably high regeneration rates of 200 cycles h -1 . Thus, it can be considered as a real alternative to enzymatical regeneration methods (Laval et al., 1987; Bonnefoy et al., 1988). Although direct electrochemical oxidation of NAD(P)H can be performed, reduction of NAD(P) + has not given valuable results for its application in organic synthesis until now. However, the use of a mediator and enzymes which catalyze the reduction of NAD(P) + (Di Cosimo et al., 1981; de Lacey et al., 1995), the direct electrochemical reduction of NAD + by hydrogenases (Cantet et al., 1996), and the formation of enzymatically active NADH by a direct electrochemical procedure (Yun et al., 1994) appear to be promising approaches. Preparative scale applications in electrochemistry need high current densities, and the use of electrodes with high specific surface area is the key point. Carbon felts are the most appropriate supports since graphite carbon is a conductor making excellent electrodes, and its packing into fibers of around 10 µm in diameter allows high specific surface areas of about 2600 cm 2 g -1 . Laboratory- scale electroenzymatic reactors with these graphite-felt electrodes have shown NADH oxidation yields higher than 99.95%. These studies also have shown the close relationship between a decrease in catalytic current, and thus in conversion, and the decrease in the active enzyme concentration by deactivation (Fassouane et al., 1990). Consequently, the use of appropriate conditions to main- tain enzyme stability will enhance the applicability of such a reactor for chemical synthesis. To the best of our knowledge, there are no works related to the loss of activity under working conditions when a controlled electrical potential is established in order to regenerate electrochemically a coenzyme. In this way, the aim of this paper is the optimization of the conditions for NAD(P) + regeneration in an electrochemi- cal reactor, from the point of view of enzyme stability and coenzyme regeneration rate. Preparative scale experiments with a graphite-felt electrode were done with glucose dehydrogenase (GDH) as a model enzyme be- cause it is specific for NAD(P) + . The enzyme converts with both coenzymes -D-glucose to D-glucono-δ-lactone * To whom all correspondence should be addressed. † Universidad de Murcia. ‡ CSIC. § E-mail address: jliborra@fcu.um.es. 557 Biotechnol. Prog. 1997, 13, 557-561 S8756-7938(97)00063-5 CCC: $14.00 © 1997 American Chemical Society and American Institute of Chemical Engineers