Search

Enantioselective epoxide hydrolases are useful biocatalysts for the preparation of enantiopure epoxides and diols. The kinetic resolution of racemic epoxides can be carried out in an organic/aqueous biphasic system to allow use of high epoxide concentrations. Enzyme inactivation in such a system, however, may occur by contact with the interface. In this study, we investigated the factors which influence the interfacial inactivation of Agrobacterium radiobacter epoxide hydrolase in an octane/water biphasic system. Rates of interfacial inactivation were measured both in a stirred-cell, which has a planar interface, and in an emulsion reactor. Interfacial inactivation rates measured in the stirred-cell at a fixed interfacial area increased with mixing intensity. Interfacial inactivation rates per unit area were lower in the emulsion reactor than in the stirred-cell and increased with bulk aqueous enzyme concentration. Circular dichroism measurements showed that during biphasic incubation all unadsorbed soluble enzyme existed in the native conformation. Activity assays showed that the dissolved enzyme was also fully active, indicating that inactivated enzyme precipitated from solution. Using an inactive epoxide hydrolase mutant structurally similar to the wild-type enzyme in order to avoid the conversion of the epoxide, it was found that high concentrations of epoxide in the organic phase increased the rate of interfacial inactivation.

Optically active epoxides can be obtained by kinetic resolution of racemic mixtures using enantiose- lective epoxide hydrolases. To increase the productivity of the conversion of sparingly aqueous soluble epoxides, we investigated the use of a two-phase aqueous/organic system. A kinetic model which takes into account inter- phase mass transfer, enzymatic reaction, and enzyme inactivation was developed to describe epoxide conver- sion in the system by the epoxide hydrolase from Agro- bacterium radiobacter. A Lewis cell was used to deter- mine model parameters and results from resolutions car- ried out in the Lewis cell were compared to model predictions to validate the model. It was found that n- octane is a biocompatible immiscible solvent suitable for use as the organic phase. Good agreement between the model predictions and experimental data was found when the enzyme inactivation rate was fitted. Simula- tions showed that mass transfer limitations have to be avoided in order to maximize the yield of enantiomeri- cally pure epoxide. Resolution of a 39 g/L solution of racemic styrene oxide in octane was successfully carried out in an emulsion batch reactor to obtain (S)-styrene oxide in high enantiomeric excess (>95% e.e.) with a yield of 30%. © 2001 John Wiley & Sons, Inc. Biotechnol Bio- eng 73: 44-54, 2001.