Some properties of enzymes obtained from natural sources limit their wider applicability in lab or industry. Examples are low or no activity towards non-natural substrates, lack of regiospecificity, substrate or product inhibition. In order to improve such disadvantageous properties, we use structure-based optimization methods such as rational re-design or homology modelling based methods for predicting advantageous mutations (see Fig. 2, work together with RG Computational Chemistry). E.g., optimizations of cofactor-forming enzymes improve access and range of these cost-limiting substances important for many enzyme classes to work properly (methyl transferases, coenzyme A ligases, hydroxylases).

Our research primarily focuses on multi-enzyme cascades, i.e. sequential enzyme-catalyzed reactions that produces a certain natural bioactive compound or intermediate from simple precursor molecules. Purified enzymes are used, for example, in an immobilized form for catalysis. In this in vitro method toxic compounds can be used and produced, e.g. antibiotics, and it can be combined with chemical steps and run in standard chemical reactors (drop-in technology). Usually also concentrations of compounds are higher and purification of the products is simpler than in organism based systems.

An alternative, are transformations of whole (artificial) pathways into host microorganisms (whole-cell biocatalysts), i.e. the enzyme cascade is produced in vivo ("synthetic biology" SynBio). The starting substance is added to the cell suspension, taken up by the organism, and ideally transformed into the product. Two of several advantages of an in vivo production system are that cofactors and standard building blocks are provided by the host organism, and that sensitive intermediates and enzymes are easier to deal with.

The products can be obtained by extraction. For the production of the bitter-masker (taste modifier) homeriodictyol shown in Fig. 3 we could provide both alternatives to an industrial partner.