Natural Deep Eutectic Solvents as Performance Additives for Peroxygenase Catalysis

Natural deep eutectic solvents (NADES) are proposed as alternative solvents for peroxygenase‐catalysed oxyfunctionalization reactions. Choline chloride‐based NADES are of particular interest as they can serve as solvent, enzyme‐stabiliser and sacrificial electron donor for the in situ H2O2 generation. This report provides the first proof‐of‐concept and basic characterisation of this new reaction system. Highly promising turnover numbers for the biocatalysts of up to 200,000 have been achieved.

Natural deep eutectic solvents (NADES) are proposed as alternative solvents for peroxygenase-catalysed oxyfunctionalization reactions. Choline chloride-based NADES are of particular interest as they can serve as solvent, enzyme-stabiliser and sacrificial electron donor for the in situ H 2 O 2 generation. This report provides the first proof-of-concept and basic characterisation of this new reaction system. Highly promising turnover numbers for the biocatalysts of up to 200,000 have been achieved.
Water is the solvent of biocatalysis; the vast majority of biocatalytic reactions reported today take place in aqueous reaction media. Apparently, this is a direct consequence of the importance of water for life and the fact that most enzymes (the catalysts of life) are readily soluble in aqueous media.
In many aspects, however, water is not an ideal solvent for chemical transformations utilising enzymes. Most reagents of interest, for example, are rather hydrophobic and therefore only sparingly soluble in aqueous reaction media. Also, it should be noted that the stability of enzymes in non-aqueous media can significantly be higher than in aqueous media. [1] Therefore, the quest for alternative solvents enabling higher reagent loadings and improved biocatalyst performance, is ongoing and gaining momentum. [2] In 2003 Abbot and co-workers first reported a novel class of eutectic mixtures, termed Deep Eutectic Solvents (DES), [3] which ever since have found wide interest as simpler and environmentally more acceptable alternatives to ionic liquids. [4] Shortly afterwards some of us discovered that DES occur naturally (termed Natural Deep Eutectic Solvents, NADES). [5] Not very astonishingly, (NA)DES have also been evaluated as solvents for biocatalytic reactions. [2d] The majority of studies has been focusing on hydrolase-catalysed (trans) esterification reactions while the application of NADES on further enzyme classes is still in its infancy. [2d] We therefore set out to explore the potential of NADES as 'performance solvents' for peroxygenase-catalysed oxyfunctionalisation reactions. Peroxygenases are a new class of hemethiolate dependent enzymes catalysing P450-monooxygenaselike oxyfunctionalisation reactions [6] but, unlike P450 monooxygenases, do not rely on complicated electron transport chains. Instead, peroxygenases utilise H 2 O 2 as stoichiometric oxidant. However, if applied in surplus H 2 O 2 leads to oxidative inactivation of the biocatalysts. Therefore, a range of in situ H 2 O 2 generation approaches have been developed to balance the H 2 O 2 concentration to levels where the desired peroxygenase reaction rate is maximised while the undesired oxidative inactivation rate is minimised. [7] Here, we propose an in situ H 2 O 2 generation system based on Choline oxidase (ChOx)-catalysed conversion of choline into betaine and the concomitant generation of two equivalents of H 2 O 2 to drive peroxygenase-catalysed oxyfunctionalisation (Scheme 1). This way, we envision utilising ChCl-based NADES both as solvent and as sacrificial electron donor at the same time. [8] As the biocatalysts, we chose the choline oxidase from Arthrobacter nicotianae (AnChOx), [9] and the recombinant, evolved peroxygenase from Agrocybe aegerita (rAaeUPO). [10] Both enzymes were produced by heterologous expression in Escherichia coli or Pichia pastoris, respectively, following estab-lished protocols. [9][10] The choline chloride (ChCl)-based NADES used in this study (Table S1) were synthesised as described previously. [8] In a first set of experiments we evaluated the effect of the ChCl-based NADES on the bienzymatic hydroxylation of ethyl benzene and the epoxidation of cis-ß-methylstyrene ( Figure 1). It should be mentioned here that no product formation was observed in negative controls leaving out either rAaeUPO or AnChOx. Furthermore, the optical purities of the products were always higher than 95 % ee. We therefore concluded that the product formation reported in this contribution was indeed a result of the cascade shown in Scheme 1. Various dilutions of the individual NADES were evaluated.
Noteworthy, neat NADES as solvents were unfavourable in all cases as no product formation was observed under these conditions, which at least partially can be attributed to the water-demand of the reaction (Scheme 1). For almost all NADES screened, at least one dilution with buffer was identified that enabled a higher product concentration than operating in aqueous buffer alone (the sole exceptions being the epoxidation reactions in ChCl-Sor and ChCl-EG). In case of the hydroxylation reaction, the highest product titers were observed in the presence of 25 % NADES (Figure 1a, pink) while in case of the epoxidation reaction 50 % was found more favourable (Figure 1b, blue). Apparently, for every reaction or substrate an optimal solvent composition exists. Currently, we are lacking a satisfactory explanation for this observation. Possibly, solubility issues play a role; a preliminary MD simulation suggested that some of the NADES can penetrate rAaeUPO's active site ( Figure S4 and S5). Hence, different NADES may influence the orientation of the substrates relative to compound I and thereby may influence rAaeUPO's catalytic efficiency. More systematic studies will be necessary to fully elucidate this phenomenon and predict the optimal solvent for a given reaction.
Also, AnChOx activity was somewhat influenced by different NADES (Figure 2). Interestingly, AnChOx activity was generally reduced compared to the aqueous reaction medium.
Next, we investigated the influence of some NADES on the stability of rAaeUPO by incubating the biocatalyst in the respective NADES or in buffer under reaction conditions (Fig-Scheme 1. Envisioned bienzymatic cascade for the selective oxyfunctionalisation of ethyl benzene or cis-ß-methylstyrene comprising the evolved peroxygenase from Agrocybe aegerita (rAaeUPO) and choline oxidase from Arthrobacter nicotianae (AnChOx) to provide rAaeUPO with H 2 O 2 . Based on these observations, we chose ChCl-Urea-Gly as solvent for the hydroxylation of ethyl benzene and ChCl-Pro-H 2 O for the epoxidation of cis-ß-methylstyrene, respectively. The influence of some additional reaction parameters such as the concentration of choline chloride (ChCl), the concentration of biocatalysts (rAaeUPO and AnChOx), reaction pH, reaction temperature in buffer system were investigated in some more detail (Figure 4). Again, some interesting differences between both reaction systems became apparent. In case of the hydroxylation reaction (in buffer) the product formation steadily increased over the concentration of ChCl range investigated. Contrarily, the epoxidation reaction (in buffer) had an apparent optimum at 100 mM ChCl (Figure 4a and 4b). One possible explanation for this difference may lie in the lower specific activity of rAaeUPO for the epoxidation reaction as compared to the hydroxylation reaction. It may be assumed that higher ChCl concentrations increased the H 2 O 2 generation rate. Hence, the optimal H 2 O 2 formation rate for the slower epoxidation reaction is reached at lower ChCl concentrations as compared to the faster hydroxylation reaction. This assumption is confirmed by experiments systematically increasing the concentration of rAaeUPO (Figure 4g and 4h) and AnChOx (Figure 4i and 4j). It is interesting to note that the NADES systems supported higher ChCl and AnChOx concentrations as compared to the buffer systems. This may indicate a protective effect of NADES on the enzymes against H 2 O 2 . Similar observations had been previously reported. [11] Also, the temperature-dependence of both reactions was very different. As shown in Figures 4c and 4d, the apparent activity of the hydroxylation reaction in NADES decreased by only 50 % while raising the reaction temperature from 30°C to 50°C (the same reaction in buffer medium decreased by more than 75 %). However, the apparent activity of the epoxidation rate decreased by more than 80 % in NADES. Possibly, epoxide hydrolysis or isomerisation at elevated temperatures may account for this observation. A close inspection of the gas chromatograms gave no indication of other products than the desired epoxide. Currently, we are lacking a plausible explanation for this observation.
In the present study we have demonstrated that Choline oxidase is a very promising catalyst for the in situ generation of H 2 O 2 to drive peroxygenase-catalysed hydroxylation and epoxidation reactions. This system is particularly interesting when combining with ChCl-based NADES as here the solvent serves two purposes at the same time, i. e. as solvent (and performance additive stabilising the biocatalysts) and as cosubstrate providing the reducing equivalents needed for the reductive activation of O 2 to H 2 O 2 .  Some of the results obtained in this study still need further investigation. Particularly the question remains why the type of reaction (or substrate) has an influence on the optimal reaction conditions (choice of NADES). Also the effect of viscosity on the reaction kinetics deserve further, in-depth investigation.