enzyme is based on the ease of removal of by-products an on equilibrium considerations.
Dehydrogenases (alcohol, formate, glucose, lactate etc.) have been systematically used as
regenerating enzymes.
In the second approach (Figure 6.16b), the enzyme that synthesises the product of
interest also catalyses the regeneration of the cofactor by using another substrate (Kulbe
et al., 1990; Leuchtenberger et al., 1984). For instances, in the synthesis of pheromone
sulcatol catalysed by alcohol dehydrogenase, NADP+ was regenerated by the same
enzyme at the expense of isopropanol as a secondary substrate (Röthig et al., 1990).
The membrane reactor, apart from immobilising the enzyme(s), should retain the
cofactor, or at least increase its residence time in the system. This retention can be
accomplished through size exclusion (Maeda et al., 1990), electrostatic repulsion (Kulbe
et al., 1990; Röthig et al., 1990; Drioli et al., 1993; Nidetzky et al., 1996a), or size
exclusion via enlargement by covalent binding to polymers (e.g. PEG) (Röthig et al.,
1990; Leuchtenberger et al., 1984; Hayakawa et al., 1985; Wandrey et al., 1984;
Wichmann et al,, 1984; Ohshima et al., 1985; Berke et al., 1984, 1988; Kragl et al.,
1996b). The cofactor can also be used as a permeable solute by an immobilised enzyme
(Ishikawa et al., 1989a, 1989b; Fujii et al., 1991; Miyawaki et al., 1990).
These cofactor regenerating membrane reactors were used for the production of
several different compounds: NADPH (Peters & Kula, 1991), amino acids (Fujii et al.,
1991; Leuchtenberger et al., 1984; Wandrey et al., 1984; van Eikeren et al., 1990;
Ohshima et al., 1985; Kragl et al., 1996), hydroxyacids (Wichmann et al., 1984; van
Eikeren et al., 1990), alcohols (Kulbe et al., 1990; Rôthig et al., 1990; Miyawaki et al.,
1990; van Eikeren et al., 1990), acids (Leuchtenberger et al., 1984; Carrea et al., 1991;
Maeda et al., 1990), glucose-6 phosphate (Ishikawa et al., 1989a, 1989b; Berke et al.,
1984), 6-phosphogluconate (Miyawaki et al., 1990), lactate (Hayakawa et al., 1985;
Miyawaki et al., 1990) and aldehydes and lactones (van Eikeren et al., 1990). A large
scale process was developed at DEGUSSA AG for the production of L-alanine from
pyruvic acid catalysed by L-alanine dehydrogenase with NAD+ being regenerated back to
NADH by using formate dehydrogenase (Leuchtenberger et al., 1984). In this case
NADH was maintained in the system by increasing its molecular weight upon binding to
polyethylene glycol. This use of membrane reactors as a process strategy for regenerating
cofactors has enabled cycle numbers (defined as the number of product molecules per
cofactor molecule) as high as 500,000 (reported by Wandrey, 1987).
Many of the membrane reactor systems developed for cofactor regeneration use two
(e.g. Ishikawa et al., 1989a; Maeda et al., 1990; Fujii et al., 1991) or even three (e.g.
Wandrey et al., 1984; Carrea et al., 1991) enzymes acting synergistically. This possibility
of easily immobilising different enzyme molecules is one of the particularly attractive
features of membrane reactors that enable them to conduct enzyme catalysed sequential
reactions. Some authors have described further applications (other than cofactor
regenerating) of membrane reactors using the conjugated action of two enzymes. Kragl
and co-workers (Kragl et al., 1990b, 1990c) used an epimerase for the isomerisation of
N-acetylglucosamine to N-acetylmannosamine that was then further converted to the
final product, N-acetylneuraminic acid, by the addition of pyruvic acid catalysed by a
lyase. Other bi-enzyme membrane reactors investigated were: β-amylase/isoamylase
(Hausser et al., 1983) and α-amylase/glucoamylase (Houng et al., 1992) for the
Enzymatic membrane reactors 171