Enzymes typically depend on either NAD(P)H or FADH 2 as hydride source for reduction purposes. In contrast, organometallic catalysts most often rely on isopropanol or formate to generate the reactive hydride moiety. Here we show that incorporation of a Cp*Ir cofactor possessing a biotin moiety and 4,7-dihydroxy-1,10-phenanthroline into streptavidin (Sav) yields an NAD(P)H-dependent artificial transfer hydrogenase (ATHase). This ATHase (0.1 mol%) catalyzes imine reduction with 1 mM NADPH (2 mol%), which can be concurrently regenerated by a glucose dehydrogenase (GDH) using only 1.2 equivalents glucose. A four enzyme cascade consisting of the ATHase, the GDH, a monoamine oxidase and a catalase leads to the production of enantiopure amines.The introduction of synthetic catalysts into a biological context is at the focus of current efforts in both synthetic-and chemical biology.1,2 Goals include i) the supplementation of existing-or engineered metabolic pathways, ii) decaging inactive forms of enzymes to trigger enzymatic cascades, 2,3 iii) shifting the redox equilibrium in cancer cells to induce apoptosis 4 or iv) produce fuels with the help of biological redox equivalents.3-5 Achieving high productivity of synthetic organometallic catalysts inside a living system remains challenging and progress is likely to be incremental. 2,6 In contrast, the combination of isolated enzymes and transition metal catalysts in carefully designed in vitro systems of modest complexity experiences increasing success.
7-9Transition metal-mediated formal hydride transfer occupies a prominent role in many of these initiatives. Remarkably, synthetic catalysts and enzymes have gained common ground for the conversion of ketones to alcohols, imines to amines (and vice versa), the reduction of activated double bonds and the racemisation of secondary alcohols and amines.10 A few isolated studies have shown that transition metal complexes can accept NAD(P)H as a hydride source. 5,11,12 To the best of our knowledge however, their concurrent use coupled with enzymatic processes, where the hydride is utilized in a productive fashion, have not been disclosed yet (Scheme 1a). We hypothesize that this may be traced back to the mutual deactivation of both transition metal catalysts and natural enzymes. To overcome this challenge, spatial separation of both catalytic partners has proven most effective. [13][14][15] In this context, artificial metalloenzymes (AMEs) have received increasing attention as alternative to both homogeneous catalysts and enzymes. 16 To test the versatility of AMEs for the implementation of enzymatic cascades, we reported on the compartmentalization of a biotinylated Cp*Ir-based transfer hydrogenation catalyst into streptavidin (Sav) variants (Scheme 1d).15 To drive the hydride transfer reactions at reasonable rates however, concentrations of formate in the molar range were required. 7,15,17 Such concentrations of formate may lead to inactivation of natural enzymes and are thus incompatible with in vivo application...