FOx News: Towards Methanol‐driven Biocatalytic Oxyfunctionalisation Reactions

The novel formate oxidase from Aspergillus oryzae (AoFOx) is a useful catalyst to promote H2O2‐dependent oxyfunctionalisation reactions. In this contribution we exploit the substrate promiscuity of AoFOx to fully oxidise methanol and formaldehyde to CO2 and drive peroxygenase‐catalysed stereoselective oxyfunctionalisation reactions. The highly atom efficient H2O2 generation system also enabled high catalytic turnover of the peroxygenase production enzyme.


FOx News: Towards Methanol-driven Biocatalytic Oxyfunctionalisation Reactions
Sébastien J.-P. Willot + , [a] Manh Dat Hoang + , [b] Caroline E. Paul, [a] Miguel Alcalde, [c] Isabel W. C. E. Arends, [d] Andreas S. Bommarius, [e] Bettina Bommarius, [e] and Frank Hollmann* [a] The novel formate oxidase from Aspergillus oryzae (AoFOx) is a useful catalyst to promote H 2 O 2 -dependent oxyfunctionalisation reactions. In this contribution we exploit the substrate promiscuity of AoFOx to fully oxidise methanol and formaldehyde to CO 2 and drive peroxygenase-catalysed stereoselective oxyfunctionalisation reactions. The highly atom efficient H 2 O 2 generation system also enabled high catalytic turnover of the peroxygenase production enzyme.
Selective oxyfunctionalisation of CÀ H-bonds is one of the most challenging reactions in organic chemistry. In recent years, peroxygenases have emerged as promising catalysts for this reaction enabling a broad range of regio-and stereoselective oxyfunctionalisation reactions. [1] Like the well-known cytochrome P450 monooxygenases [2] peroxygenases utilise a highly oxidised heme Fe-oxo complex (Compound I) to electrophilically insert an oxygen atom into the starting material. While P450 monooxygenases depend on molecular oxygen and rather complicated electron transport chains to attain the reductive activation of O 2 (yielding Compound I), peroxygenases use already reduced H 2 O 2 for the same goal. Hence, peroxygenases appear as simpler, more easily applicable alternatives to P450 monooxygenases. However, such as all heme enzymes, peroxygenases are rapidly inactivated by H 2 O 2 , necessitating controlled provision with H 2 O 2 to balance the H 2 O 2 -dependent catalytic activity and the (likewise H 2 O 2 -dependent) inactivation reaction. [3] In situ generation of H 2 O 2 through catalytic reduction of O 2 is one of the most promising approaches, which, however, also necessitates a sacrificial co-substrate to provide the reducing equivalents needed for the reductive activation of O 2 . Envisioning large-scale preparative applications, H 2 O 2 generation systems producing no or innocuous wastes are preferred. Electrochemical generation of H 2 O 2 , for example, is an attractive means to drive peroxygenase reactions. [4] Also, simple reductants such as H 2 [5] or H 2 O [6] appear promising. Methanol would be a very suitable sacrificial electron donor as it is readily available, easy to handle and, in principle, can be fully oxidised to CO 2 to provide reducing equivalents for 3 equiv. H 2 O 2 per equiv. MeOH. [7] Similarly, complete oxidation of methanol to regenerate reduced nicotinamide cofactors has been reported. [8] Recently, we reported that the formate oxidase from Aspergillus oryzae (AoFOx) is an efficient catalyst for the in situ generation of H 2 O 2 to drive peroxygenase-catalysed oxyfunctionalisation reactions. [9] Encouraged by the very promising results, we originally aimed at a further characterisation of AoFOx. To our surprise, while evaluating methanol as potential co-solvent, we found that AoFOx also exhibited a methanol oxidase activity. We therefore further investigated AoFOx to drive peroxygenase catalysed oxyfunctionalisation reactions using either methanol, formaldehyde or formic acid as sacrificial electron donor ( Figure 1). As production enzyme we chose the recombinant, evolved peroxygenase from Agrocybe aegerita (rAaeUPO). [10] Using 250 mM of either MeOH, HCHO or HCO 2 H as sacrificial reductant, hydroxylation of ethyl benzene or cyclohexane as well as epoxidation of cis-ß-methyl styrene to (R)-1-phenyl ethanol, cyclohexanol and (1R,2S)-cis-ß-methyl styrene oxide, respectively, was observed ( Figure 1). AoFOx exhibited a low, but detectable methanol oxidase activity.
Interestingly, AoFOx appeared to be promiscuous with respect to the oxidation state of the C1-substrate but accepted no other starting material such as ethanol or propanol (Supporting Information, Figure S8).
We suspected a poor affinity of wild-type AoFOx for MeOH to account for the low product formation observed in these preliminary experiments (Figure 1). We therefore determined the kinetic parameters of AoFOx for the oxidation of methanol and formaldehyde (Table 1).
Indeed, the catalytic performance of AoFOx on formaldehyde and especially methanol falls back behind its natural formate oxidation activity. Instead of interpreting this as a disadvantage, we rather see the advantage of this relative activity as it, in principle, results in low in situ concentrations of harmful (HCHO) or acidifying (HCO 2 H) intermediate oxidation states of the sacrificial cosubstrate. However, even the comparably low methanol oxidase activity under saturation conditions corresponded to a specific activity of 0.43 U mg À 1 .
Robust and efficient reaction schemes involving peroxygenases need a balancing of the H 2 O 2 generation reaction rate and the (H 2 O 2 -consuming) peroxygenase reaction rate. On the one hand, high H 2 O 2 -generation rates are desirable to attain high space-time-yield. On the other hand, accumulation of H 2 O 2 must be avoided in order to minimise the oxidative inactivation of the prosthetic heme group. We therefore investigated the influence of the methanol concentration on the overall reaction system ( Figure 2). As expected from the kinetic measurements (Table 1), increasing the methanol concentration increased the overall product formation (Figure 2A). A methanol concentration around 10 % (vol/vol, approx. 2.5 M) appeared optimal. We attribute the decreasing product formation at higher methanol concentrations to a decreasing long-term stability of either AoFOx or rAaeUPO (or both) in aqueous methanol solutions. We therefore used a methanol concentration of 2.5 M for further experiments. Under these conditions, AoFOx-catalysed H 2 O 2 generation appeared to be overall rate-limiting, as up to at least 1000 nM AoFOx (equimolar to rAaeUPO) the productivity of the system increased linearly with the increasing AoFOx concentration ( Figure 2B, Figure S12). Higher productivities may be accessible upon further increase of the AoFOx concentration. However, to avoid possible H 2 O 2 accumulation, we continued with the more conservative equimolar ratio of AoFOx and rAaeUPO.
As shown in Figure 3, robust oxyfunctionalisation of all three model starting materials over at least 5 days could be achieved using methanol as sacrificial electron donor. Very promising turnover numbers for rAaeUPO and AoFOx of up to 49000 were observed (equimolar concentrations of biocatalysts used). The differences in the overall rates of cyclohexane hydroxylation, ethyl benzene hydroxylation and cis-β-methyl styrene epoxidation (0.05 mMh À 1 , 0.23 mMh À 1 and 0.32 mMh À 1 ,   respectively) can be attributed to the relative activity of the peroxygenases for these substrates. [12] Overall, with the current contribution we have demonstrated that methanol-driven peroxygenase reactions are principally possible using just one biocatalyst (AoFOx), in addition to rAaeUPO. At present, the major limitation of this approach is the comparably high K M value of the wild-type AoFOx for MeOH, which may be addressable through protein engineering. Envisioning preparative scale applications, further issues are likely to occur such as the poor solubility of O 2 in aqueous media, which we will address by either using nonaqueous reaction media [13] and/or innovative reactor concepts such as the bubble column [14] or flow-chemistry [15] to attain higher k L values for O 2 .
Nevertheless, already at present stage, the catalytic performance of the biocatalysts with turnover numbers as high as 49000 for rAaeUPO and AoFOx, respectively represent a good starting point for further characterisation and optimisation of the reaction system.

Experimental Section
Enzyme preparation. Recombinant expression and purification of the evolved unspecific peroxygenase mutant from A. aegerita in P. pastoris was performed following a previously described procedure. [10a] Formate oxidase from Aspergillus oryzae RIB40 (AoFOx) was produced recombinantly in E. coli BL21(DE3) as reported before with slight modifications. [9] During desalting step with HiTrap, an additional buffer exchange was applied by using a phosphate potassium buffer (25 mM, pH 6.0) for elution of the target enzyme of the column. A final protein concentration of 2.16 � 0.06 mg mL À 1 was measured by BCA assay. AoFOx purity of approximately 60 % was determined by SDS-PAGE. With consideration of the extinction absorption at 472 nm, [16] a molar protein concentration of 21 μM was calculated.
Enzyme activity assay. AoFOx kinetics were determined indirectly by (2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) ABTS assay. ABTS assay was performed with 50 mM of acetate buffer at pH 4.5 or phosphate buffer at pH 6.0, respectively. A final concentration of 1 mM of ABTS, 10 U mL À 1 of HRP (horseradish peroxidase) and different concentration of methanol or formaldehyde were used. ABTS was either dissolved in 50 mM acetate buffer (pH 4.5) or 50 mM phosphate buffer (pH 6.0). HRP was stored in 25 mM phosphate buffer (pH 7.5). The reactions were measured with an UV-Vis spectrometer at 420 nm for 1 min at a controlled temperature of 25°C. For calculation, an extinction coefficient of 36 000 M À 1 cm À 1 was used.
Enzymatic reaction. Reactions of 200 μL volume were performed in GC glass vials of 1.5 mL. Thermoshakers were used for temperature and stirring control. All components were mixed and the reactions started by the addition of either methanol, formaldehyde or formate. For each time point, a full reaction is extracted with ethyl acetate (equivolume) containing 5 mM n-octanol as internal standard. The extraction is then dried over MgSO 4 and analysed by gas chromatography (temperature profiles in supporting information). Further information for the negative control can be found in supporting information.