Introduction

Due to climate change and fossil fuel deficiency, the demand for renewable and sustainable equivalents instead of fossil-derived commodities is increasing day by day1. In this context, waste plant biomass with its heterogenous composition has a promising potential for utilization in industrial biotechnology2. Lignocellulose accounts for the majority of this plant biomass and thus represents a readily available, sustainable, and renewable source for providing bio-based chemicals3. The environmental production of value-added commodities or building blocks from biomass is quite attractive and has gained increasing attention over the last years. For instance, the synthesis of renewable building blocks for polymeric materials like bioplastics has been intensively studied. FDCA (2,5-furandicarboxylic acid) is a promising precursor for the sustainable synthesis of bioplastics. FDCA can be derived from FFCA (5-formyl-2-furancarboxylic acid), that is produced from HMF (5-hydroxymethylfurfural) either over HMFCA (5-hydroxymethyl-2-furancarboxylic acid) or over DFF (2,5-diformylfuran) (Fig. 1). HMF is produced during the pretreatment of lignocellulosic biomass4. Several studies have dealt with the development of enzymatic routes for the environmentally friendly and sustainable production of FDCA from HMF5,6,7,8,9,10. Fungal enzymes involved in lignin degradation have been thoroughly investigated for this purpose11,12,13.

Figure 1
figure 1

Reaction routes for HMF conversion into FDCA.

Full size image

Several organisms, particularly white-rot fungi, are efficient degraders of lignocellulose and the fungal secretome comprises a variety of enzymes like laccases, peroxidases, and H2O2-generating auxiliary enzymes, that contribute to lignin degradation14. Glyoxal oxidase (GLOX), discovered in secretomes of several fungi, is suggested to be involved in this process by fueling peroxide-depending ligninolytic peroxidases with H2O2. GLOX belongs to the family of copper radical oxidases (CROs) and catalyzes the reduction of O2 to H2O2 while oxidizing several aldehydes to carboxylic acids15. While other members of CROs like galactose oxidase (GalOx) have been extensively studied and some crystal structures of these enzymes are available, much less is known about GLOX enzymes. Similar to all CROs, these enzymes have a conserved active site with a mononuclear copper ion coordinated to an axial tyrosine, two histidines, and a cross-linked cysteine-tyrosyl radical cofactor11,16. Notably, GLOX as well as GalOx is inactivated during purification and requires initial activation by a strong oxidizing agent17,18. Inactivation is caused by the reduction of the active free-radical Cu(II)-state to the catalytically inactive non-radical Cu(I)-state (Fig. 2), as well as by high H2O2 concentrations. Activation of GLOX by regeneration of the oxidized radical form could be achieved by treating GLOX with high-redox potential inorganic oxidants like Na2IrCl618 or other single oxidants generated by lignin peroxidase19. However, the activated enzyme is unstable with a half-life of 4 h for the radical form, and returns to the reduced inactive form18. Up to now, only GLOX from Phanerochaete chrysosporium20,21, from Ustilago maydis22, from Myceliophthora thermophila3, and three GLOX isoenzymes from Pycnoporus cinnabarinus1,23 were characterized. Nevertherless, the usefulness of GLOX for the production of value-added commodities from biomass has been demonstrated. For instance, GLOX was reported to oxidize glycerol and produce glyceraldehyde and glyceric acid23,24. In 2019, Daou et al. investigated the ability of three GLOX isoenzymes for the bioconversion of HMF to yield FDCA, when combined with an aryl-alcohol oxidase (AAO) to prevent HMFCA accumulation1. In this setup, 16% FDCA was formed from HMF along with 84% FFCA. However, further optimization of this promising two-enzyme approach for the complete conversion of HMF to FDCA is still required.

Figure 2
figure 2

Reaction cycle of GLOX and reactivation by ABTS and horseradish peroxidase. I. oxidative half-reaction, II. reductive half-reaction.

Full size image

In the present study, we produced a new glyoxal oxidase from Trametes versicolor (TvGLOX) in Pichia pastoris and investigated this enzyme for the production of FDCA from HMF in a two-enzyme reaction cascade. A suitable redox activator for continuous GLOX reactivation in the course of the reaction was identified and applied in our setup, giving almost complete conversion of 8 mM HMF to FDCA within 24 h.

Results and discussion

TvGLOX production and enzyme purification

TvGLOX was identified by protein BLAST search, using GLOX from P. chrysosporium (UniProtKB: Q01772.1) as a query. The tvglox encoding DNA sequence with the native signal sequence for secretion (NCBI accession number XM_008037054) was codon optimized using the online tool JCat25 and cloned into the pPICZA plasmid under the control of the AOX1 promoter. The resulting plasmid was integrated into the P. pastoris X-33 genome by homologous recombination. Among 48 transformants, four P. pastoris colonies showing the most intensive green halo formation on methylglyoxal/HRP/ABTS-containing agar plates were selected for expression in shaking flasks. The transformant with the highest volumetric activity towards methylglyoxal of 0.23 U/ml was chosen for fed-batch cultivation in a 7.5 L bioreactor. After 9 days of fermentation, a volumetric activity of 19,000 U/l and a protein concentration of 1.6 g/l were obtained. Similar volumetric activity of 22.7 U/ml21 and protein concertation of 1–2 g/l15 were reported for expression of P. chrysosporium GLOX in P. pastoris.

TvGLOX was concentrated from the supernatant and purified by hydrophobic interaction chromatography. The purified enzyme displayed a specific activity towards methylglyoxal of 4.0 U/mg. During purification, TvGLOX changed its color from light green, attributed to the oxidized form of the enzyme, to purple color, which was observed for GLOX from P. cinnabarinus as well23.

Purified TvGLOX had a molecular mass of around 80 kDa (Fig. S1), while the calculated molecular mass of native TvGLOX is 58 kDa. Four potential N-glycosylation sites were found in TvGLOX (Fig. S2). N-deglycosylation of TvGLOX shifted the band to around 58 kDa (Fig. S1), giving about 29% N-glycosylation of TvGLOX which is higher than ~ 18% N-glycosylation of GLOX from P. chrysosporium, heterologously expressed in P. pastoris26.

Biochemical characterization of TvGLOX

TvGLOX showed the highest activity towards methylglyoxal at pH 6.5 (Fig. S3) which is similar to other GLOX enzymes1,3,21,23 and retained about 100% of its initial activity after 5 h incubation at 25 °C and pH 6.5 (Fig. S4). After 3 h incubation of TvGLOX between 40 °C and 60 °C, about 90% of its initial activity remained, whereas at 70 °C, it lost its activity immediately (Fig. S4). The thermostability of other GLOX enzymes like PciGLOX1 and PciGLOX2 at 40 °C and 50 °C was comparable to that of TvGLOX, while at 60 °C PciGLOX1 and PciGLOX2 retained only 50% of their initial activity after 2 h of incubation23.

The substrate spectrum of TvGLOX was estimated in a coupled assay, including HRP and ABTS, to measure H2O2 generated by TvGLOX during substrate oxidation. A total of 15 compounds including HMF, HMFCA, DFF, FFCA, furfural, glycerol, formaldehyde, glyoxal, methylglyoxal, glyoxylic acid, D-galactose, D-glucose, benzyl alcohol, veratryl alcohol, and veratral were tested. As shown in Fig. 3, TvGLOX possessed the highest activity towards glyoxylic acid followed by methylglyoxal, glyoxal, formaldehyde and DFF. Activity towards glycerol, HMF, FFCA and furfural was also detected, while no activity was found towards d-glucose, d-galactose, benzyl alcohol, veratryl alcohol, veratral and HMFCA. These results confirm that TvGLOX is a typical glyoxal oxidase, that mainly accepts aldehyde groups and hardly oxidizes alcohol groups. In comparison, PciGLOX1 also demonstrated the highest activity towards glyoxylic acid, whereas other GLOX enzymes such as PciGLOX3 or GLOX from P. chrysosporium preferred methylglyoxal over glyoxylic acid1,19,23. Homology modeling confirmed that the active site and overall structure of TvGLOX is quite similar to other GLOX enzymes such as GLOX from P. chrysosporium (Fig. 4). Like in other CROs, the active site of TvGLOX contains an axial tyrosine (Tyr399), two histidines (His400 and His493), and a cross-linked cysteine-tyrosyl radical cofactor formed by Cys93 and Tyr158.

Figure 3
figure 3

Substrate spectrum of TvGLOX (activity towards glyoxylic acid was set to 100%).

Full size image
Figure 4
figure 4

Structural model of TvGLOX and GLOX from P. chrysosporium. Homology models of both enzymes were built using GalOx (PDB: 2EIE) as template. Catalytic site of (a) TvGLOX and (b) GLOX from P. chrysosporium. The crosslink between the cysteine-tyrosine cofactor is missing due to the method of homology model building. Cartoon and surface representation of (c) TvGLOX and (d) GLOX from P. chrysosporium. The copper ion is depicted as a blue sphere.

Full size image

Oxidation of HMF and derivatives thereof by TvGLOX

Although GLOX typically oxidizes aldehydes to the corresponding acids and thus converts HMF to HMFCA, this enzyme is also capable of oxidizing alcohols to aldehydes, like glycerol to glyceraldehyde. While PciGLOX1, PciGLOX2, and PciGLOX3 have been reported to predominantly oxidize the aldehyde group of HMF and to produce HMFCA1, MtGLOx from Myceliophthora thermophila M77 has been shown to oxidize the alcohol group of HMF leading to DFF, but failed to further oxidize DFF3. We investigated the activity of TvGLOX towards HMF, DFF, HMFCA, and FFCA in more detail and determined the corresponding product profiles in these reactions (Table 1). With HMF, TvGLOX was mainly active on the aldehyde group and converted HMF to 88% HMFCA and about 12% FDCA after 24 h. Conversion of HMF by PciGLOX2 and PciGLOX3 reached only 39% and 41%, respectively after 24 h1. In contrast to MtGLOX, which failed to oxidize DFF3, TvGLOX showed the highest activity towards DFF among the investigated furan derivatives and completely converted DFF via FFCA to FDCA within 24 h.

Table 1 Conversion of furan derivatives catalyzed by TvGLOX.
Full size table

With HMFCA no product formation was detected after 24 h, which is consistent with the preference of TvGLOX for the aldehyde group (as in HMF) and low acceptance of alcohol groups (as in HMFCA). According to this, the aldehyde group of FFCA was oxidized very well, leading to 100% FDCA after 24 h. As HMFCA was not accepted at all, the formation of small amounts of FDCA in the conversion of HMF with TvGLOX might be explained by very slow oxidation of HMF to DFF (not observed after 2 h of reaction), followed by a fast conversion of DFF to FFCA and further to FDCA within 24 h of reaction. In contrast, with PciGLOX1-3 only a low conversion of HMFCA to FFCA and FDCA within 24 h has been reported1. The absence of oxidation products in the reaction of TvGLOX with HMFCA may also be related to stability issues caused by this furan derivative. Therefore, we investigated the influence of the furan derivatives on the stability of TvGLOX. Only a slight decrease in activity was observed when TvGLOX was incubated for 24 h with HMFCA or FFCA, while no loss in activity was observed with HMF and DFF (Fig. S5). Other GLOX enzymes are more sensitive to furan derivatives, e.g. all three GLOX enzymes from P. cinnabarinus almost completely lost their activity after 24 h incubation with HMFCA1. From our results, it can be concluded, that the lack of product formation in the reaction of TvGLOX with HMFCA can be attributed to the inability of this enzyme to oxidize the alcohol group of HMFCA.

Redox activators for TvGLOX

As known from the literature, GLOX easily undergoes inactivation by reduction of the active free-radical Cu(II)-state to the catalytically inactive non-radical Cu(I)-state requiring reactivation in the course of substrate oxidation reaction17,19,24. High-redox potential inorganic oxidants like potassium octacyanomolybdate (K3Mo(CN)8), sodium hexachloroiridate (Na2IrCl6) or Mn3+ can reactivate GLOX18,24. In ligninolytic cultures, GLOX from P. chrysosporium was fully activated in the presence of lignin peroxidase combined with its substrate veratryl alcohol19. In this system, lignin peroxidase consumed H2O2 produced by GLOX. HRP can also be used for GLOX activation when applied together with a suitable HRP substrate such as ABTS or veratryl alcohol18,27. Although the activating effect has not been completely understood, it has been suggested that a single-electron oxidant generated by HRP or other peroxidases, is used to regenerate the oxidized radical form of GLOX28. Recently, this hypothesis has been supported by the finding that ABTS cation radicals, generated through the oxidation by HRP in HRP/GLOX-coupled reactions, were in fact the species responsible for GLOX activation during methylglyoxal oxidation (Fig. 2)17. In order to find the most suitable redox activator for TvGLOX, we tested different HRP substrates—ABTS, 2,6-dimethoxyphenol, catechol, methyl syringate, syringaldazine, p-hydroquinone, guaiacol, veratryl alcohol and Mn2+, respectively—as redox activators of TvGLOX during oxidation of FFCA (Table 2). Mn2+ is not a common HRP substrate, but Mn3+-oxalate complexes were previously shown to strongly activate GLOX when HRP was included in the reaction, due to the HRP-catalyzed oxidation of Mn2+ to Mn3+24.

Table 2 Formation of FDCA (%) from 2 mM FFCA by TvGLOX in the presence of HRP and different redox activators after 24 h.
Full size table

In the absence of both, HRP and a redox activator, a slight formation of FDCA by TvGLOX was detected after 24 h, which may indicate that TvGLOX was not completely inactive after purification. MtGLOx from M. thermophila M77 was found to be fully active even without the addition of peroxidases or oxidizing agents3. Addition of commercial HRP Type II improved FDCA formation to a certain extent (18% FDCA with HRP vs. 5% FDCA w/o HRP). Whether TvGLOX was activated by any compounds contained in the commercially available HRP preparation or activated by the enzyme itself, as it was reported for GLOX from P. chryosporium, which was moderately activated when incubated with lignin peroxidase19, remains elusive. The HRP/ABTS-system gave the best results with complete FFCA conversion within 24 h, followed by the HRP/methyl syringate system (94% FDCA). When FFCA conversion was analyzed after 2 h, it became apparent that HRP/ABTS overperformed HRP/methyl syringate in activating TvGLOX (85% FDCA vs. 62% FDCA, respectively). HRP/Mn2+ was also able to activate TvGLOX, but only half the FFCA conversion was achieved compared to the reaction with ABTS. Veratryl alcohol only moderately activated TvGLOX as was also reported by Kersten et al.27, and resulted only in 25% conversion of FFCA. Syringaldazine did not improve FFCA conversion compared to the reaction without any redox activator, whereas 2,6-dimethoxyphenol, guaiacol, p-hydroquinone and catechol reduced FFCA conversion. Inhibition of GLOX by guaiacol and catechol when peroxidase was added to the reaction has been described before27. Consequently, the best HRP substrate, ABTS, with the strongest activating effect on TvGLOX was used in further experiments.

Reaction cascade for the production of FDCA from HMF

In order to achieve the complete conversion of HMF to FDCA and avoid the accumulation of the dead-end product HMFCA, a cascade involving a second enzyme, the aryl-alcohol oxidase MaAAO from Moesziomyces antarcticus, was established (Fig. S6). In our previous work, MaAAO was found to convert HMF to FFCA within 24 h, but was not able to convert FFCA to FDCA29. We proposed that by combining the activities and substrate preferences of MaAAO and TvGLOX, complete conversion of HMF to FDCA can be achieved. Several experimental setups were tested. TvGLOX was added after 2 h and 24 h of reaction with MaAAO, respectively, and in one setup both enzymes were added from the beginning of the reaction. When TvGLOX and MaAAO were added at the same time as the substrate HMF, complete conversion of 2 mM HMF to FDCA was achieved after 24 h (Fig. 5). In a similar approach, Daou et al. combined UmAAO from Ustilago maydis and PciGLOX3 from P. cinnabarinus in a cascade for HMF conversion to FDCA, but detected only 16% FDCA along with 84% FFCA1. In that study, catalase was also added to decompose high concentrations of H2O2 but conversion could be only marginally improved. Comparison with our results allows us to suggest that continuous reactivation of GLOX has a strong effect on the cascade outcome and GLOX inactivation by H2O2 only plays a marginal role.

Figure 5
figure 5

Dependence of FDCA production in the TvGLOX/MaAAO reaction system on increasing HMF concentrations.

Full size image

In the next step, we increased the HMF concentration while maintaining the enzyme load, and observed complete conversion of up to 6 mM HMF to FDCA. With 8 mM HMF 7.8 mM FDCA and 0.2 mM FFCA were measured and with 10 mM HMF 7.8 mM FDCA and 2.2 mM FFCA (Figs. 5 and S7). For comparison, in a three-enzyme system consisting of a galactose oxidase, an unspecific peroxygenase, and an aryl-alcohol oxidase, 7.9 mM FDCA was achieved after 24 h starting from 9.7 mM HMF7. Our results show that TvGLOX is a promising biocatalyst for the enzymatic synthesis of bioplastics precursors, and the continuous reactivation of this Cu-radical oxidase represents a critical factor during process optimization.

Conclusion

In this study, a new glyoxal oxidase from T. versicolor—TvGLOX—was heterologously expressed at high levels in P. pastoris and characterized regarding its substrate spectrum, stability and reactivation by different redox activators. TvGLOX was most efficiently reactivated by ABTS in the presence of horseradish peroxidase (HRP) and preferred aldehydes over alcohols, leading to HMFCA formation from HMF. HMFCA was not accepted by this enzyme preventing further oxidation to yield FDCA. To achieve complete conversion to FDCA, a biocatalytic system consisting of MaAAO and TvGLOX, the latter one continuously reactivated through HRP and ABTS, was developed. With this system, up to 8 mM HMF was almost completely converted to FDCA within 24 h, demonstrating that TvGLOX is a promising biocatalyst for biotechnological applications.

Methods and material

Construction of recombinant Pichia pastoris strain

TvGLOX from Trametes versicolor FP-101664 SS1 (NCBI accession number XM_008037054) was identified by protein BLAST search using GLOX from T. cinnabarina (GenBank: ANJ20632.1) as query (88.91% identity, E-value 0.0, query coverage 100%). The tvglox encoding sequence with its native signal sequence for secretion was synthesized by BioCat GmbH (Heidelberg, Germany) after codon optimization for expression in Saccharomyces cerevisiae using the online tool JCat25. The gene was cloned into the Pichia pastoris expression vector pPICZA between EcoRI and NotI restriction sites (pPICZA_tvglox). Chemically competent E. coli DH5α cells were used for plasmid propagation. Transformants were selected on low salt lysogeny broth agar plates including 25 μg/ml Zeocin. Plasmid isolation was performed by using the ZR Plasmid Miniprep Kit (Zymo Research, USA). Electrocompetent P. pastoris X-33 cells were transformed by electroporation with 10 µg pPICZA_tvglox plasmid after linearization with MssI. Positive transformants were selected on yeast extract peptone dextrose sorbitol (YPDS) agar plates supplemented with 100 μg/ml Zeocin after incubation at 30 °C for 3 days.

Heterologous expression of TvGLOX in shaking flask and bioreactor

P. pastoris transformants with TvGLOX activity were selected on buffered minimal methanol (BMM) screening agar plates (1.34% yeast nitrogen base, 100 mM potassium phosphate buffer pH 6.0, 4∙10–5 biotin, 0.5% methanol, 2% agar, 0.5 mM 2,2′-azino-di-(3-ethylbenzthiazoline sulfonic acid (ABTS), 0.006 mg/ml HRP, 2 mM methylglyoxal). Colonies with the strongest green halo formation after 72 h incubation at 30 °C were selected for cultivation in shaking flasks. For this purpose, a colony from the agar plate was grown in 10 ml BMGY (buffered complex glycerol) medium overnight at 30 °C and 200 rpm. The overnight-culture was used for the inoculation of 200 ml BMMY (buffered complex methanol) medium to an OD600 of 1 and incubated for 3–4 days at 25 °C and 200 rpm. Every 24 h, the cultures were fed with 0.5% (v/v) methanol. The OD600 value and volumetric activity of the supernatant were measured daily using methylglyoxal as substrate.

The most active P. pastoris transformant was used for fed-batch fermentation. Fed-batch fermentation was performed in a 7.5 L bioreactor (Infors, Switzerland) as previously described30. The temperature was set to 25 °C after induction of TvLGOX expression and the fermentation proceeded for 9 days with daily sampling to follow OD600 and volumetric activity towards methylglyoxal.

Enzyme purification

P. pastoris cells were separated from fermentation broth by centrifugation at 4 °C and 11,325×g for 15 min, and the collected supernatant containing secreted TvGLOX was concentrated and rebuffered with 25 mM sodium phosphate pH 6.0 by tangential flow filtration (TFF, 10 kDa molecular cut-off; Pall, Port Washington, USA). TvGLOX was purified via hydrophobic interaction chromatography (HIC) on an ÄKTA purifier FPLC system (GE Healthcare, USA). At first, 10 ml concentrated supernatant was incubated in 25 mM sodium phosphate buffer pH 6.0 containing 1.5 M ammonium sulfate at 10 °C overnight. Then, the sample was centrifuged at 4 °C and 18,000×g for 30 min, filtered and applied on the column with butyl sepharose HP medium. Elution was performed by decreasing ammonium sulfate concentration by adding 25 mM sodium phosphate buffer pH 6.0. Fractions with activity towards methylglyoxal were concentrated and desalted via an ultrafiltration membrane filter (Vivaspin Turbo 15, 10 kDa molecular cut-off). The purified TvGLOX enzyme was kept at 4 °C until use.

Biochemical characterization

The activity of TvGLOX was routinely measured in 96 well microtiter plates in a coupled assay using horseradish peroxidase (HRP, Type II; purchased from Sigma-Aldrich (Schnelldorf, Germany)) and ABTS for initial activation of GLOX and monitoring of substrate oxidation through H2O2 formation. Reaction mixtures contained 20 μl TvGLOX solution, 20 μl 100 mM methylglyoxal, 20 μl 0.06 mg/ml HRP, 20 μl 5 mM ABTS, and 120 μl 50 mM sodium phosphate buffer pH 6.5, including 0.5 μM H2O2. ABTS oxidation was followed at 420 nm using a Infinite M200 Pro plate reader (Tecan, Switzerland).

Protein concentration was measured using the Bradford method31. Deglycosylation of purified TvGLOX (20 μg) was carried out under native and denaturing conditions (for up to 24 h) by using Peptide N-Glycosidase F (PNGase F, New England Biolabs, Germany) according to the manufacturer’s instructions. Purified and deglycosylated protein was visualized by SDS-PAGE32. Potential N-glycosylation sites were predicted by using the NetNGlyc 1.0 server prediction tool33.

pH optimum of purified TvGLOX was measured in 40 mM Britton-Robinson buffer at several pH values between pH 2.0–11.5 by using methylglyoxal as substrate as described above. The stability of TvGLOX at pH 6.5 was investigated by incubating the enzyme in 50 mM sodium phosphate buffer pH 6.5 at 25 °C for 5 h. Thermal stability was determined by incubating TvGLOX in 50 mM sodium phosphate buffer pH 6.5 at different temperatures (40 °C, 50 °C, 60 °C, and 70 °C) for 3 h. After incubation, samples were taken, incubated on ice and the remaining activity was measured at room temperature using methylglyoxal as substrate.

The enzyme activity towards 15 compounds (HMF, HMFCA, DFF, FFCA, furfural, glycerol, formaldehyde, glyoxal, methylglyoxal, glyoxylic acid, d-galactose, d-glucose, benzyl alcohol, veratryl alcohol, and veratral) was estimated at a substrate concentration of 10 mM in a coupled assay as described above.

The enzyme stability towards furan derivatives was assessed by incubating TvGLOX in 50 mM sodium phosphate buffer pH 6.5 with 2 mM substrate (HMF, DFF, HMFCA, and FFCA, respectively). Samples were taken after 4 h, 24 h, and 48 h incubation, and the remaining activity towards methylglyoxal was measured as described above.

Oxidation of furan derivatives

Reactions were performed in a total volume of 200 µl in 50 mM sodium phosphate buffer pH 6.5 with 2 mM substrate (HMF, DFF, HMFCA and FFCA, respectively), 2 μM TvGLOX, 0.5 μM H2O2, 0.1 mg/ml HRP and 0.5 mM ABTS. Reactions were incubated at 25 °C and 500 rpm for up to 24 h. Fifty μl samples were taken after 15 min, 2 h and 24 h, acidified with 10 µl 6 M HCl and 2-furoic acid (20 mM) was added as an internal standard. Methyl tert-butyl ether (MTBE) was used for sample extraction, each sample was dried over MgSO4, evaporated to dryness and resuspended in 50 µl of N,O-bis (trimethylsilyl) trifluoroacetamide (BSTFA), and incubated at 60 °C for 30 min before GC–MS analysis.

Investigation of redox activators

Reactions were performed in a total volume of 200 µl in 50 mM sodium phosphate buffer pH 6.5 with 2 mM FFCA, 2 μM TvGLOX, 0.5 μM H2O2, 0.1 mg/ml HRP and 2 mM redox activator (2,6-dimethoxyphenol, catechol, methyl syringate, syringaldazine, p-hydroquinone, guaiacol, veratryl alcohol and MnSO4, respectively). Reactions with HRP and Mn2+ were conducted in 50 mM sodium malonate buffer pH 6.5. Reactions were incubated at 25 °C and 500 rpm for up to 24 h. Fifty μl samples were taken in the course of the reaction and prepared for analysis via GC–MS measurements as described above.

Optimization of HMF conversion

For optimization of the reaction cascade Moesziomyces antarcticus aryl alcohol oxidase (MaAAO) was used29. Reactions were performed in a total volume of 200 µl in 50 mM sodium phosphate buffer pH 6.5 containing 2–10 mM HMF, 0.5 μM H2O2, 2 μM MaAAO, 0.1 mg/ml HRP, and 0.5 mM ABTS. All reactions were incubated at 25 °C, 500 rpm for up to 24 h. Samples were taken in the course of the reaction and prepared for analysis via GC–MS measurements as described above.

GC–MS analysis

Oxidation of furan derivatives was analyzed using a GC–MS-QP-2010 (Shimadzu, Japan) on an FS-Supreme-5 ms column (CS Chromatographie Service GmbH, Germany). The injection, interface, and ion source temperatures were set to 250 °C, 285 °C, and 200 °C, respectively. The temperature for the column was adjusted to 110 °C, maintained for 2 min, then increased to 300 °C gradually with 20 °C/min. Substrates and products were identified by comparing the measured mass spectra with authentic standards. Substrate conversion and product formation were calculated from substrate depletion (control set to 100%), using 2-furoic acid as the internal standard and from external calibration curves of furan derivatives (0–2 mM) and 2-furoic acid as internal standard.