Chemoenzymatic cascade reactions are quite recent, exciting reaction concepts, since they are able to combine benefits of two catalytic fields . Chemoenzymatic cascades have been successfully used e.g. in the coupling of reactions with P450 monooxygenases or hydrolases and water-tolerant metathesis catalysts. The major challenge for the combination of such intrinsically distinct catalysis systems is the mutual compatibility of the catalysts themselves and of their reaction conditions. This compatibility mainly relies on the solvent, as many organometallic catalysts are inactive in water, whereas only a few enzymes show activity in organic solvents. In a one-pot reaction, the choice of compatible catalysts is therefore already strictly limited by the solvent. A pragmatic compromise is to perform a sequential reaction with an extraction of the intermediary formed product. This strategy is often the easiest to perform, but it poses problems with intermediates of low stability. In addition, it typically requires time- and resource- consuming work-up steps. Compartmentalization strategies can make it possible to keep both catalysts at their optimal reaction conditions and at the same time enable a transfer of the intermediate between the reaction compartments [2,3,4].
This study is focused on the production of dihydroxystilbenes as structural analogs of resveratrol. The compound class of resveratrol analogs has received considerable attention due to the diverse biological effects they exert, such as chemoprotection, anti-oxidant, anti-inflammatory, antidiabetic, antithrombotic, cancer prevention, and antiaging properties [5,6]. The natural synthesis of resveratrol is too expensive for in vitro
processes due to its complexity and a high demand of co-substrates such as CoA. Furthermore, resveratrol synthase is specific only for the synthesis of resveratrol and structural analogs as 4,4′-dihydroxy-trans
-stilbene (DHS) are not available with this system. Interestingly, DHS, which shows a four-fold higher antioxidant capacity than resveratrol, has been reported to prevent cancer invasion and metastasis and inhibit growth of human leukemia cells [6,7].
It is therefore a highly interesting target for an effective and selective (bio)synthetic production. Known synthetic routes for the production of 4,4′‑disubstituted stilbene derivatives are complicated multistep approaches and involve Wittig and McMurry coupling reactions. Olefin metathesis is efficiently used for the chemoselective formation of C=C double bonds even within complex precursor molecules. It could thus present an alternative synthesis method for 4,4′‑disubstituted stilbene derivatives, given the efficient production of the required precursors, the 4‑vinylphenols. A chemoenzymatic approach involving an enzymatic production of 4-vinylphenol and subsequent ruthenium-catalyzed olefin metathesis can give rise to symmetric stilbene derivatives in a two-step one-pot reaction.
The enzymatic production of 4-vinylphenol could be achieved by decarboxylation of p‑coumaric acid by a decarboxylase. Due to the abundance of p‑coumaric acid and other hydroxy-cinnamic acid derivatives in lignocellulosic biomass and different agricultural waste residues, these acids impose very attractive resources for bio-based polymers and other materials, such as the here discussed resveratrol analogs . The cofactor-free phenolic acid decarboxylase (bsPAD) from Bacillus subtilis is a robust and suitable biocatalyst for the conversion of different phenolic acids . The enzyme was easily cloned and expressed in E. coli. The purified enzyme converted p-coumaric acid efficiently to produce the desired 4-vinylphenol.
In order to find appropriate catalysts for the subsequent olefin metathesis reaction, different ruthenium complexes were evaluated in different solvents, including e.g. dichloromethane, tetrahydrofuran, and methyl-tert-butyl-ether (MTBE). Grubbs’ second generation catalysts were found to achieve the olefin metathesis in several anhydrous solvents and under anaerobic conditions.
In a next step, the establishment of the chemoenzymatic cascade was pursued. Therefore, the reactions were performed in a sequential two-pot cascade using the free enzyme in aqueous buffer and the metathesis catalyst in anhydrous organic solvents. This straightforward approach enabled both catalysts to work under optimal conditions. As anticipated, the enzymatic decarboxylation was accomplished within 24 h reaction time and 4-vinylphenol was recovered after extraction with MTBE in yields of up to 95%. The full conversion made further purification unnecessary and the metathesis could be directly started by adding the appropriate solvent and ruthenium catalyst. Unfortunately, the isolated intermediates tend to undergo spontaneous polymerization [10,11], thus reducing the effectivity and overall yield of this two-pot method. Although the sequential cascade already showed appreciable yields of DHS, it was considered to use a bi-phasic one-pot system to avoid possible handling issues. In a solvent system consisting of aqueous buffer and isooctane (2,2,4‑trimethylpentane), the enzymatic decarboxylation was feasible. Unfortunately, none of the investigated Grubbs’ catalysts, not even water-and air-stable candidates, exceeded 20% conversion for the metathesis reaction. Apparently, bsPAD is inactive in any of the tested metathesis solvents, so a chemoenzymatic one-pot cascade could only be achieved by increasing the enzyme stability in a non-aqueous environment.
The crucial idea was to encapsulate the enzyme in PVA/PEG (polyvinyl alcohol/polyethylene glycol) cryogels that entrap the enzyme with its aqueous environment . This enabled the decarboxylation “in water” with pure organic solvent present outside the capsules. Encapsulated cell-free extract of bsPAD showed 39% activity compared to an equivalent amount of cell-free extract. After recycling the cryogel beads for 5 cycles the enzymes still retained 69% of its initial activity, thus compensating for the activity loss compared to the free catalyst format and proving the re-usability of the cryogel beads. For the metathesis in presence of the beads, non-polar solvents performed better than partially water-miscible ones, so that MTBE was chosen as solvent for the cascade reaction.
Under cascade conditions, the enzymatic decarboxylation afforded >99% substrate conversion with bsPAD in cryogel beads after 24 hours. The macroscopic size of the beads (ca. 2-3 mm diameter, fig. 2) made them easy to remove, leaving behind the intermediate 4-vinylphenol in MTBE. After removing residual water with anhydrous MgSO4, the metathesis was initiated simply by addition of the ruthenium catalyst. A conversion of 95% for 4-vinylphenol proved the high activity of the Grubbs catalyst under cascade conditions. Overall, an isolated yield of 90% (58 mg) DHS could be achieved from 100 mg p-coumaric acid, and also ferulic acid and caffeic acid were converted to the corresponding stilbenes in 36% resp. 37% overall isolated yields.
Overall, it could be shown that compartmentalization is a successful strategy to optimize reaction conditions for the application of enzymes and metathesis catalysts with low water tolerance in one-pot approaches. In order to obtain enzyme activity at non-aqueous conditions required for metathesis, enzyme-polyvinyl alcohol cryogels proved to be a simple but highly efficient tool, thus resulting in a “compatibility window” for two very different catalysts. Since the method is applicable to cell-free extracts of enzymes with low tolerance towards organic solvents and the cryogel beads can easily be reused, it is expected that this elegant approach will be applicable to a large number of future cascade reactions.
Álvaro Gómez Baraibar1, Florian Busch1, Dennis Reichert1, Carolin Mügge1, and Robert Kourist1,2
1 Junior Research Group for Microbial Biotechnology, Ruhr-University Bochum, Bochum, Germany
2 Institute of Molecular Biotechnology, Graz University of Technology, Graz, Austria
Prof. Dr. Robert Kourist
Graz University of Technology
Institute of Molecular Biotechnology
 H. Gröger et al., Curr. Opin. Chem. Biol. 2014, 19, 171–179. doi: 10.1016/j.cbpa.2014.03.002
 M. T. Mwangi et al., Chem. Eur. J. 2008, 14, 6780–6788. doi: 10.1002/chem.200800094
 H. Sato et al., Angew. Chem. Int. Ed. 2015, 54, 4488–4492. doi: 10.1002/anie.201409590
 T. Hischer et al., Biocatal. Biotransform. 2009, 24, 437–442. doi: 10.1080/10242420601040261
 M. Li et al., Metab. Eng. 2015, 32, 1–11. doi; 10.1016/j.ymben.2015.08.007
 M. Savio et al., Sci. Rep. 2016, 6, 19973. doi: 10.1038/srep19973
 G.-J. Fan, et al., Bioorg. Med. Chem. 2009, 17, 2360–2365. doi: 10.1016/j.bmc.2009.02.014
 I. Barbara et al., Eur. Polym. J. 2015, 62, 236–243. doi: 10.1016/j.eurpolymj.2014.11.035
 J.-F. Cavin et al., Appl. Environ. Microbiol. 1998, 64, 1466–1471.
 K. Kodaira et al., Makromol. Chem., Rapid Commun. 1980, 1, 427–431. doi: 10.1002/marc.1980.030010704
 R. Sovish, J. Org. Chem. 1959, 24, 1345–1347. doi: 10.1021/jo01091a606