*2.1. Purification and Identification of the Alkene Cleavage Activity*

Within a screening of 17 basidiomycetes for alkene cleavage activity using the substrate *trans*-anethole (Table S1) *P. sapidus* turned out to be a promising candidate for the production of the desired activity. The lyophilized mycelium as well as the culture supernatant was examined for the ability to cleave *trans*-anethole after submerged cultivation. The culture supernatant showed no activity, whereas the incubation in the presence of the mycelium resulted in formation of 5.36 mM *p*-anisaldehyde (molar yield of 79.03%; Figure 1).

**Figure 1.** Bioconversion of *trans*-anethole by *P. sapidus*. (**a**) Alkene cleavage of *trans*-anethole resulted in the formation of *p*-anisaldehyde. (**b**) GC-FID chromatogram of an *n*-hexane extract of the conversion of the blank sample (blue) and after incubation with lyophilized mycelium of *P. sapidus* (orange). Retention times: *trans*-anethole (11.53 min) and *p*-anisaldehyde (13.45 min).

For the identification of the enzyme catalyzing the *trans*-anethole cleavage, it was semi-purified from the rehydrated mycelium by hydrophobic interaction and anion exchange chromatography (IEX). During the purification, a high activity loss occurred, which resulted in low product concentrations after conversion (Table S2). This was most likely a result of protein loss, enzyme degradation or denaturation as described for other enzymes [22–24]. Another possibility is the loss of cofactors or –substrates, such as metal ions or peroxides during the purification steps [6]. Addition of Mn2<sup>+</sup> led to a 15-fold increase in *p*-anisaldehyde concentration, which was further increased by addition of hydrogen peroxide, indicating a cosubstrate dependency (Table S2). Chemical conversion by Mn2<sup>+</sup> alone was excluded, while product formation was observed with hydrogen peroxide (18 μM), but with a yield around 30-fold lower than the one for the enzymatic reaction (Table S2). Thus, the improved bioconversion in the presence of Mn2<sup>+</sup> and hydrogen peroxide was verified to be the result of an increased enzyme activity.

A class of fungal enzymes that requires hydrogen peroxide and some of which need Mn2<sup>+</sup> for catalysis are peroxidases [12,25,26]. The anion exchange fractions, which showed alkene cleavage activity, also exhibited peroxidase activity, thus verifying the presence of a peroxidase. For visualization of the activity, a semi-native PAGE was performed and stained with ABTS in the presence of hydrogen peroxide (Figure 2). Two peroxidases running at 45 and 52 kDa were detected. The respective protein bands stained with Coomassie Brilliant Blue were excised for electrospray ionization tandem mass spectrometry. Due to the low protein concentration of the 52 kDa band, no meaningful peptides were found (Figure 2, lane 1). This paper presents the data obtained for the protein running at 45 kDa (Figure 2, arrow).

Three tryptic peptides (EGSELLGAR, DGSFLTFR, and SGAPIEITPLKDDPK) were identified by ESI-MS/MS. Homology searches against the public database NCBI using the mascot search engine (Matrix Science, London, UK) identified a DyP-type peroxidase of *P. ostreatus* PC15 (*Pleos*-DyP4 [27]; GenBank accession no. KDQ22873.1), a close relative of *P. sapidus*, as the best hit.

**Figure 2.** Semi-native PAGE of the active fraction after purification of the alkene cleavage enzyme from *P. sapidus* by IEX. 1: gel stained with Coomassie Brilliant Blue; 2: gel stained with ABTS in the presence of hydrogen peroxide, M: pre-stained molecular mass marker. An arrow and a box marks the protein that was successfully identified by sequencing.

#### *2.2. Amplification and Expression of PsaPOX*

Specific primers successfully amplified the 1512 bp coding region of the gene from *P. sapidus*. The translated amino acid sequence of 504 aa contained the peptide fragments that were obtained by ESI-MS/MS and showed highest identity (94%) to the sequence of *Pleos*-DyP4 (Figure 3), which has not been investigated for its alkene cleavage activity against *trans*-anethole or other substrates before [27,28]. In our hands, *P. ostreatus* showed a weaker *trans*-anethole cleavage activity, too (Table S1). Identity of the *P. sapidus* peroxidase (PsaPOX) to other DyPs and proteins was lower than 60%. A sequence alignment with other DyPs (Figure 3) confirmed that PsaPOX exhibited the typical GXXGD motif and all important residues known for the catalytic activity of DyPs, such as the proximal histidine (His-334) (fifth ligand of heme iron) and distal Asp-196 and Arg-360 involved in the activation of the enzyme (formation of compound I) by H2O2 cleavage [14,15]. Furthermore, Trp-405 was identified as a homolog to the surface exposed Trp-377 of AauDyP of *Auricularia auricula-judae*, which serves as an oxidation site for bulky substrates such as Reactive blue 19 (RBB19) using a long-range electron transfer [29]. Comparison of the PsaPOX and *Pleos*-DyP4 sequence indicated that PsaPOX exhibited a non-canonical Mn2<sup>+</sup>-oxidation site on its surface (Asp-215, Glu-345, Asp-352 and Asp-354; Trp-339 participates in the electron transport from the oxidation site to the heme) like *Pleos*-DyP4 [28] and can oxidize Mn2<sup>+</sup> to Mn3<sup>+</sup>, which is known for a few fungal DyPs only [24,27,30,31].

A structural homology model of PsaPOX (Figure S1), which was generated using the X-ray crystal structure of *Pleos*-DyP4 (PDB-ID 6fsk) on the SWISS-MODEL server, possessed typical characteristics of the DyP-type peroxidase family (N- and C-terminal ferredoxin-like domain, each formed by four-stranded antiparallel β-sheets and several α-helices) [15] and supported the classification of PsaPOX as DyP. Furthermore, the analysis of the amino acid sequence using PeroxiBase [32] related the *P. sapidus* peroxidase (PsaPOX) to the class "DyP-type peroxidase D". As known from literature, DyPs differ significantly in amino acid sequence, tertiary structure, and catalytic residues from other representatives of the heme peroxidases, such as HRP, human myeloperoxidase, lignin peroxidase, or *Coprinus cinereus* peroxidase [12,15,33], all of which are able to cleave *trans*-anethole or structurally related alkenes [10,11]. So far no DyP is known to catalyze the mentioned reaction.


**Figure 3.** Alignment of alkene cleaving peroxidase from *P. sapidus* (PsaPOX) with the *Pleos*-DyP4 of *P. ostreatus* (PosDyP; KDQ22873.1) and other characterized DyPs. AgaDyP: *Armillaria gallica* (PBK80505.1), VvoDyP: *Volvariella volvacea* (AKU04643.1), CtrDyP: *Coriolopsis trogii* (AUW34346.1), GluDyP: *Ganoderma lucidum* (ADN05763.1), and TveDyP: *Trametes versicolor* (XP\_008039377.1). Inverted triangles show amino acids important for heme binding (histidine (magenta) functions as ligand for heme and the four other amino acid residues form a hydrogen peroxide binding pocket). Aspartic acid, which forms a hydrogen bond with histidine to stabilize compound I (oxidized heme after transfer of two electrons to H2O2) is shown in grey. The black box indicates the GXXDG motif containing the catalytic aspartic acid residue (yellow), which cleaves H2O2 heterolytically with the help of the neighboring arginine (green) to form compound I, and the circle presents an exposed tryptophan potentially involved in an LRET (long range electron transfer). Important amino acids for Mn2<sup>+</sup>-oxidation are highlighted in cyan; asterisks indicate conserved residues, colons equivalent residues and dots partial residue conservation. Peptides identified by protein sequencing are underlined. Alignment was performed with Clustal Omega (European Bioinformatics Institute, Hinxton, UK).

## *2.3. Production and Purification of the Recombinant PsaPOX*

The *PsaPOX* gene was amplified and cloned into the *K. pfa*ffi*i* expression vector pPIC9. The initial expression of the gene yielded average peroxidase activities of 65 U/L after 72 h of cultivation. The best performing colonies produced activities up to 142 U/L, indicating a multiple insertion of the expression construct [34]. Similar results were obtained for the heterologous production of a DyP from *Funalia trogii* in *K. pfa*ffi*i* previously [24]. Further experiments were performed using the clone with the highest peroxidase activity for maximum protein production.

The recombinant peroxidase was purified by Ni-NTA affinity. Using SDS-PAGE, a molecular mass of around 61 kDa was determined (Figure 4a), which is slightly higher than the calculated molecular mass of 54.9 kDa (ExPASy). In addition, the native recombinant enzyme was detected at 52 kDa after semi-native PAGE, while the native wild-type enzyme showed a band at 45 kDa (Figures 2 and 4b). Deglycosylation by endoglycosidase H (EndoH) showed that the higher molecular mass was attributed to post-translational modifications by *K. pfa*ffi*i*, as has been described for other proteins [35,36]. The wild-type peroxidase, on the contrary, was not glycosylated (see Figures 2 and 4b). That is uncommon for DyPs, which usually exhibit a carbohydrate content of 9 to 30% [12]. The molecular mass of the monomeric PsaPOX was similar to other DyP-type peroxidases [12].

**Figure 4.** Purification of the recombinant PsaPOX by Ni-IMAC. (**a**) SDS-PAGE stained with Coomassie Brilliant Blue. 1: flow through, 2: elution fraction, 3: elution fraction incubated with EndoH, 4: EndoH, M1: molecular mass marker. (**b**) Semi-native PAGE stained with ABTS in the presence of hydrogen peroxide. 5: elution fraction, 6: elution fraction incubated with EndoH, M2: pre-stained molecular mass marker. (**c**) Isoelectric focusing gel. 7: elution fraction stained with phenylendiamine in the presence of urea peroxide. M3: standard protein marker for isoelectric focusing stained with Coomassie Brilliant Blue.

Analysis of the purified recombinant peroxidase by isoelectric focusing indicated an isoelectric point around pH 6.7 (Figure 4c), which differs slightly from the calculated value of 6.28 (ExPASy), but was similar to the isoelectric point of another DyP-type peroxidase from *P. sapidus* [19]. Most other proteins belonging to the DyP-type peroxidase family showed lower values (pI 3.5-4.3, [12]).

#### *2.4. Biochemical Characterization of PsaPOX*

The influence of pH and temperature on PsaPOX activity and stability was determined using ABTS in the presence of hydrogen peroxide as substrate (Figure 5). The enzyme showed a pH optimum of 3.5 while more than 50% of activity was conserved between pH 3 and 5 (Figure 5a). At lower or higher pH values of ≤ 25% of activity remained, most likely due to conformational changes of the enzyme. The results were consistent with the findings for other fungal DyPs, which had pH optima in the range between pH 2 and 5 [13,19,27]. PsaPOX showed the highest pH stability with a residual peroxidase activity of ≥90% between pH 2.0 and 5.5 after 1 h of incubation (Figure 5c). At pH values higher than six, near the isoelectric point, the stability decreased drastically, probably due to a reduced solubility and changes of the protein structure, which may have resulted in protein aggregation.

**Figure 5.** Influence of pH and temperature on activity and stability of PsaPOX. The pH optimum (**a**) was determined to be 3.5 and the temperature optimum (**b**) 40 ◦C. Relative peroxidase activity [%] was defined as the percentage of activity detected with respect to the highest activity in each experiment. pH stability (**c**) was determined after incubation of PsaPOX in Britton Robinson buffer ranging from pH 2.0 to 9.5 for 1 h at RT and temperature stability (**d**) after incubation at 20 to 90 ◦C and pH 3.5 for 1 h. Residual activities were determined at pH 3.5 and 40 ◦C. Values are the average of triplicate experiments with standard deviations shown as error bars.

Peroxidase activity of PsaPOX increased with rising temperature, reaching its maximum at 40 ◦C (Figure 5b), which was similar to the optimum (30–40 ◦C) of a recombinant DyP from *P. ostreatus* [13], but higher than the optimum (RT) of another DyP-type peroxidase of *P. sapidus* produced heterologously in *Escherichia coli* [37]. With further temperature increase, the peroxidase activity of PsaPOX decreased continuously. The temperature stability of PsaPOX was determined after an incubation for 1 h at different temperatures (Figure 5d). The enzyme was relatively stable at temperatures from 20 to 60 ◦C with a residual activity ≥80%. At higher temperatures, a high loss of activity was observed due to protein denaturation, resulting in residual activities <5%. The temperature stability of PsaPOX was higher than the stabilities of DyPs from *Bjerkandara adusta* and *Auricularia auricular-judae*, which were produced heterologously in *E. coli* (residual activity ≥80% and <5% after 1 h at 20–50 ◦C and 60 ◦C, respectively [13]), and of a DyP from *P. sapidus* produced in *Trichoderma reesei* (residual activity ≥80% and <65% after 5 min at 15-45 ◦C and 50 ◦C, respectively [19]).

As mentioned above, the addition of hydrogen peroxide as well as Mn2<sup>+</sup> led to an increase of the product concentration for the biotransformation of *trans*-anethole using the lyophilized mycelium of *P. sapidus* containing the wild-type PsaPOX (Table S2). For this reason, the hydrogen peroxide and Mn2<sup>+</sup> dependencies were examined for the recombinant enzyme using ABTS as substrate at optimal pH and temperature (Figure 6). As expected, no peroxidase activity was detectable without hydrogen peroxide. The activity rose with increasing peroxide concentration and reached its optimum in the presence of 100 μM H2O2 (Figure 6a). An increase of the hydrogen peroxide concentration led to a continuous activity decrease. Suicide inhibition in the presence of excess hydrogen peroxide is well known for classical peroxidases as a result of the formation of an inactive oxidative state (Compound III) by reaction of H2O2 and Compound II [20,21], even if the existence of Compound II has not been confirmed for DyP-type peroxidases universally [17]. However, inhibition of other DyPs in the presence of higher hydrogen peroxide concentrations has been reported [18,19].

**Figure 6.** Effect of hydrogen peroxide (**a**) and Mn2<sup>+</sup> concentration (**b**) on the activity of PsaPOX. Relative peroxidase activity [%] was defined as the percentage of activity detected with respect to the highest activity obtained in each experiment. Values are the average of triplicate experiments with standard deviations shown as error bars.

Investigation of the Mn2<sup>+</sup> dependency (Figure 6b) showed that PsaPOX activity rose with increasing Mn2<sup>+</sup> concentration, but was not completely dependent on the addition of Mn2+. 30% of peroxidase activity were detected without addition of Mn2<sup>+</sup>. PsaPOX reached the maximal activity in the presence of 25 mM Mn2<sup>+</sup>. Evaluation of Mn3<sup>+</sup> formation by Mn2<sup>+</sup> oxidation revealed a manganese peroxidase activity of 0.4 U compared to 1 U of peroxidase activity using ABTS as substrate. This result fits the prediction of a Mn2<sup>+</sup> oxidation site. Only a few fungal DyPs are known to catalyze the oxidation of Mn2<sup>+</sup> [24,27,30,31]. Calculation of kinetic constants (Table 1) showed that the catalytic efficiency of PsaPOX towards Mn2<sup>+</sup> was similar to the one of *Pleos*-DyP1 from *P. ostreatus* and Ftr-DyP from *Funalia trogii* [24,27]. However, the catalytic efficiency of *Pleos*-DyP4 was higher [27].

**Table 1.** Michaelis constants (*K*m), catalytic constants (*k*cat), and catalytic efficiencies (*k*cat/*K*m) for PsaPOX using ABTS, Mn2+, Reactive blue 19 (RB19), and Reactive black 5 (RB5) as substrate. Values are the average of triplicate experiments with indication of standard deviations.


n. d.: no activity was detected.

Kinetic parameters were also calculated for the oxidation of ABTS at optimal conditions (Table 1). The affinity of PsaPOX to ABTS (37 μM) was similar to a DyP from *Irpex lacteus* (28 μM), but higher in comparison to the FtrDyP from *F. trogii* (182 μM) and the *Pleos*-DyP2 from *P. ostreatus* (787 μM) [18,24,27]. In contrast, the catalytic efficiency of PsaPOX (184 s−<sup>1</sup> mM<sup>−</sup>1) was lower than the efficiency of the DyP from *Irpex lacteus* (8000 s−<sup>1</sup> mM<sup>−</sup>1) and *Pleos*-DyP4 (352 s−<sup>1</sup> mM<sup>−</sup>1), but higher than the efficiency of the FtrDyP (54 s−<sup>1</sup> mM<sup>−</sup>1).

It is known that DyP-type peroxidases typically oxidize anthraquinones and other dyes. Exemplary, decolorization of Reactive blue 19 (anthraquinone dye) and Reactive black 5 (recalcitrant azo dye) by recombinant PsaPOX (1 U/L) was tested. Unexpectedly, PsaPOX showed activity for neither of the substrates (Table 1), although the protein sequence and tertiary structure as well as the presence of typical catalytic residues and the GXXDG motif identified the enzyme as a DyP-type peroxidase. However, *Pleos*-DyP1 from *P. ostreatus* and *Tv*DyP1 from *Trametes versicolor* also did not oxidize the

high redox-potential Reactive black 5, although they degraded Reactive blue 19 [27,30]. A missing activity against Reactive blue 19 or another anthraquinone has not been described for a fungal or class D type DyP before, but one bacterial DyP of *Pseudomonas fluorescens* (DyP2B, DyP typ class B) is known not to oxidize the anthraquinone dye Reactive blue 4 [38].
