Distinct Mechanistic Behaviour of Tomato CYP74C3 and Maize CYP74A19 Allene Oxide Synthases: Insights from Trapping Experiments and Allene Oxide Isolation

The product specificity and mechanistic peculiarities of two allene oxide synthases, tomato LeAOS3 (CYP74C3) and maize ZmAOS (CYP74A19), were studied. Enzymes were vortexed with linoleic acid 9-hydroperoxide in a hexane–water biphasic system (20–60 s, 0 °C). Synthesized allene oxide (9,10-epoxy-10,12-octadecadienoic acid; 9,10-EOD) was trapped with ethanol. Incubations with ZmAOS produced predominantly 9,10-EOD, which was converted into an ethanolysis product, (12Z)-9-ethoxy-10-oxo-12-octadecenoic acid. LeAOS3 produced the same trapping product and 9(R)–α–ketol at nearly equimolar yields. Thus, both α–ketol and 9,10-EOD appeared to be kinetically controlled LeAOS3 products. NMR data for 9,10-EOD (Me) preparations revealed that ZmAOS specifically synthesized 10(E)-9,10-EOD, whereas LeAOS3 produced a roughly 4:1 mixture of 10(E) and 10(Z) isomers. The cyclopentenone cis-10-oxo-11-phytoenoic acid (10-oxo-PEA) and the Favorskii-type product yields were appreciable with LeAOS3, but dramatically lower with ZmAOS. The 9,10-EOD (free acid) kept in hexane transformed into macrolactones but did not cyclize. LeAOS3 catalysis is supposed to produce a higher proportion of oxyallyl diradical (a valence tautomer of allene oxide), which is a direct precursor of both cyclopentenone and cyclopropanone. This may explain the substantial yields of cis-10-oxo-PEA and the Favorskii-type product (via cyclopropanone) with LeAOS3. Furthermore, 10(Z)-9,10-EOD may be produced via the reverse formation of allene oxide from oxyallyl diradical.


Introduction
Allene oxide synthase (AOS, syn. hydroperoxide dehydratase, EC 4.2.1.92) is an enzyme controlling the dehydration of fatty acid hydroperoxides to the short-lived allene oxides [1,2]. AOSs of the CYP74 family (P450 superfamily) are widespread in plants and play a key role in biosynthesis of phytohormone jasmonates [3,4]. AOSs of stony corals are also P450s of the CYP74 clan but not the CYP74 family [5]. AOSs detected in some fungi are distinct P450s, having no relation to the CYP74 family and clan [6]. Another kind of AOS present in some soft corals is the catalase-like haemoproteins of the peroxidase superfamily, not the P450s [2].
Although the primary products of AOSs are short-lived allene oxides, the total conversion is rather complex. Firstly, the allene oxides themselves are known to co-exist with their valence tautomers oxyallyl and cyclopropanone [7]. Secondly, the exo-double bond at the oxirane of allene oxide may have either (E) or (Z) configuration depending on the AOS specificity [8][9][10]. Thirdly, the short-lived allene oxides undergo rapid conversions into ketols (hydrolysis) [11,12] or cyclopentenones (cyclization) [13][14][15]. Finally, some AOSs, such as those of the CYP74C subfamily, possess distinct product patterns compared to the most common CYP74A subfamily AOSs. For instance, the CYP74C AOSs produce

Ethanol Trapping Experiments with 9-HPOD. Major Products
To minimize the hydrolysis of allene oxide, the incubations (described in full detail in Materials and Methods) were carried out in a biphasic system via a modified procedure by Brash et al. [10]. An ice-cold solution of linoleic acid 9(S)-hydroperoxide (9-HPOD) in hexane was extensively vortexed with LeAOS3 or ZmAOS for 20 s-5 min (as specified below). The water was quickly frozen at −79 • C, the hexane phase was decanted, and an excess of ice-cold ethanol was directly added to the hexane solution for allene oxide trapping. The predominant part of hexane was evaporated in vacuo and carboxylic acids were esterified with ethereal diazomethane. The resulting Me esters were trimethylsilylated and, thus, free alcohol functions present in some products were converted into TMS derivatives. The products were subjected to GC-MS analyses. The structural formulae of the major products described below are presented in Figure 1.
The mass spectrum of the predominant product 1 (Me) is presented in Figure 2C. The spectrum possessed M + at m/z 354 (0.1%) and [M -MeO] + at m/z 323 (1%). Fragment [M -Me(CH 2 ) 7 CH=O] + at m/z 215 (100%) indicated the presence of 9-oxo-10-ethoxy function. The mass spectrum of the product of catalytic hydrogenation of compound 1 (Me) and the corresponding mass fragmentation scheme are presented in Supplementary Figure S2A. The spectrum possessed M + at m/z 356 (0.1%) and [M -MeO] + at m/z 325 (2%), thus, indicating the addition of two hydrogens upon hydrogenation. The rest of the spectrum was nearly identical to that of compound 1. The NaBH 4 reduction of compound 1 followed by the sequential methylation/trimetylsilylation resulted in a product whose mass spectrum (Supplementary Figure S2B    The mass spectrum of the predominant product 1 (Me) is presented in Figure 2C ] + at m/z 325 (2%), thus, indicating the addition of two hydrogens upon hydrogenation. The rest of the spectrum was nearly identical to that of compound 1. The NaBH4 reduction of compound 1 followed by the sequential methylation/trimetylsilylation resulted in a product whose mass spectrum (Supplementary Figure S2B Figure S2B, inset) signified the presence of vic-diol function at C9/C10. The fragment at m/z 317 and the losses thereof indicated the TMSO and EtO substituents at C10 and C9, respectively. Overall, the spectrum indicated the structure of (12Z)-9-ethoxy-10-hydroxy-12-octadecenoic acid (Me/TMS) for NaBH4 reduced compound 1. In turn, the obtained data allow one to ascribe the structure of (12Z)- The α-ketol 2 formed upon the brief (60 s) incubations of LeAOS3 with 9-HPOD followed by ethanol trapping was isolated and purified by NP-HPLC. Purified α-ketol was subjected to the analysis of its enantiomeric composition by chiral-phase HPLC. The obtained results showed that α-ketol was composed of 92% pure (9R) enantiomer. The isomer was not detectable. In contrast, the same α-ketol 2 formed after the identical incubation with ZmAOS followed by ethanol trapping was composed of ca. 70% (9R) and 30% (9S) enantiomers.
A minor product (retention time 16.08 min) detected after the ethanol trapping experiments with LeAOS3 was eluted in front of the cis-10-oxo-PEA (4) peak ( Figure 2A). The mass spectral patterns of product 3 (Et/Me), shown in Supplementary Figure S3A, closely matched those of the dimethyl ester of the Favorskii-type products previously described [18]. The spectrum possessed [M -MeO] + at m/z 323 (2%), [M -EtOH] + at m/z 308 (4%) and other distinctive fragments (see the Supplementary Figure S3A). The mass fragmentation indicated the structure of (2 Z)-2-(2 -octenyl)-decane-1,10-dioic acid (Et/Me ester) formed via the ethanolysis of short-lived cyclopropanone. For further struc-tural confirmation, product 3 was hydrogenated over PtO 2 . The mass spectrum of hydrogenated product (Supplementary Figure S3B) showed characteristic fragmentation patterns (inset in the same figure). Most of the fragments were formed due to the cleavage of bonds at tertiary carbon (C2). The results enabled us to assign the structure of 2-(octyl)decane-1,10-dioic acid (Et/Me ester) to hydrogenated compound 2, thereby confirming the parent compound 3 structure as (2 Z)-2-(2 -octenyl)-decane-1,10-dioic acid (Et/Me ester). Compound 3 was also detected after the ethanol trapping experiments with ZmAOS, but it was ca. 100-times less abundant compared to the LeAOS3 ethanol trapping products.

Allene Oxide Isolation and NMR Study
The allene oxide 9,10-EOD (Me) was prepared after biphasic incubations of ZmAOS and LeAOS3 with 9-HPOD followed by methylation with diazomethane and purification by NP-HPLC on a LiChrosphere CN (5 µm) 30 × 4 mm column maintained at -15 • C. The NMR spectra ( 2 H 14 -hexane, 253 K) were recorded. The yield achieved after the ZmAOS incubation allowed us to record the 1 H-NMR, 1 H-1 H-COSY, 1 H-1 H-TOCSY, 1 H-1 H-NOESY, 1 H-13 C-HSQC and 1 H-13 C-HMBC spectral data, presented in Table 1. The 1 H-13 C-HSQC and 1 H-13 C-HMBC are also depicted in Supplementary Figure S6, and the 1 H-1 H-COSY spectrum is shown in Supplementary Figure S7. All signal attributions were verified by COSY, TOCSY and HMBC correlations. The spectrum matched that previously described for allene oxides produced by CYP74A AOSs [9,10]. The chemical shifts in all three olefinic protons, H11, H12 and H13 (Table 1, Figure 4), were in agreement with the (E)-configuration of the exo double bond at the oxirane (C10) [10]. Overall, the NMR data (Table 1) confirmed the structure of (10E,12Z)-9,10-epoxy-10,12-octadecadienoic acid (Me) for 9,10-EOD biosynthesized by ZmAOS. LeAOS3 produced 9,10-EOD at a much poorer yield. Nevertheless, the 9,10-EOD (Me) was also isolated, purified by HPLC and the 1 H-NMR, 1 H-1 H-COSY and 1 H-1 H-TOCSY spectra were recorded ( Table 2). The 1 H-1 H-NOESY, 1 H-13 C-HSQC and 1 H-13 C-HMBC spectra were not recorded due to the too-low 9,10-EOD concentration. The 1 H-NMR spectrum ( Figure 4B) revealed the presence of two geometric isomers of 9,10-EOD at a ratio of ca. 4:1, as judged by the signal integration data. The major isomer signals matched those of the above-described 10(E)-isomer synthesised by ZmAOS. Two isomers were distinguished by their 1 H-1 H-COSY and 1 H-1 H-TOCSY correlations. The minor isomer had very similar shapes of olefinic proton multiplets to those of the 10(E)-isomer, but their chemical shifts were significantly different ( Figure 4). For instance, the H12 signal was downshifted from 5.87 ppm (E-isomer) to 6.11 ppm (Z-isomer) due to the deshielding effect of epoxide oxygen. Contrariwise, the H11 signal was upfield shifted from 5.66 ppm (E-isomer) to 5.49 ppm, since H11 of the minor Z-isomer is anti-configured towards the oxygen atom of oxirane and is, thus, less affected by oxygen anisotropy. The chemical shifts in the minor isomer matched those of the (11Z)-isomer described before [10]. Overall, the recorded NMR data demonstrated that ZmAOS specifically synthesized (10E)-9,10-EOD, while LeAOS3 produced two geometric isomers of allene oxide, namely (10E)-9,10-EOD and (10Z)-9,10-EOD, at a ratio ca. 4:1. spectrum is shown in Supplementary Figure S7. All signal attributions were verified by COSY, TOCSY and HMBC correlations. The spectrum matched that previously described for allene oxides produced by CYP74A AOSs [9,10]. The chemical shifts in all three olefinic protons, H11, H12 and H13 (Table 1, Figure 4), were in agreement with the (E)-configuration of the exo double bond at the oxirane (C10) [10]. Overall, the NMR data (Table 1) confirmed the structure of (10E,12Z)-9,10-epoxy-10,12-octadecadienoic acid (Me) for 9,10-EOD biosynthesized by ZmAOS.   To shed more light on the mechanistic specificities of two enzymes, the kinetics and concentration dependencies were studied. The results are described in the next section. Table 3 summarises rough quantitative data on the relative yields of fatty acid hydroperoxide conversion products by LeAOS3 and ZmAOS. The most obvious difference is the much higher yield of α-ketols upon LeAOS3 incubations with all substrates. The Favorskii-type product 3 from 9-HPOD was formed with LeAOS3 only, not ZmAOS. The ratio of the Favorskii-type product 3 to the allene oxide trapping product 1 after LeAOS3 incubations was nearly constant, irrespective of enzyme concentration and incubation time (results not presented). The conversion of 13-HPOT by both enzymes yielded the allene oxide trapping product 5 as well as the detectable Favorskii-type product 7. The cyclopentenone cis-12-oxo-PDA (8) was formed, but at a lower yield compared to that of 10-oxo-PEA formed upon the LeAOS3 incubations with 9-HPOD. The relative yield of cyclopentenone cis-10-oxo-PEA (4) in relation to α-ketol (2) was nearly constant, about 1:20 (as estimated by integration of TIC and selected ion chromatograms), irrespective of LeAOS3 concentration (25 µg, 50 µg, 100 µg, 150 µg or 200 µg). Furthermore, when the allene oxide (free acid) was kept for hours in hexane solution, the ratio of cis-10-oxo-PEA (4) in relation to α-ketol (2) remained about the same. Incubations of LeAOS3 with 9-HPOD for 20, 40 or 75 s in a biphasic system followed by EtOH trapping resulted in product profiles possessing nearly the same constant proportion of cis-10-oxo-PEA (4) to α-ketol (2).

Allene Oxide Conversions in Aprotic Solvent
A special series of experiments was conducted to elucidate the allene oxide conversions in an aprotic solvent. Allene oxide, 9,10-EOD (free carboxylic acid), prepared by the biphasic incubations of LeAOS3 and ZmAOS with 9-HPOD, was allowed to stay in the hexane solution. Then, the products were treated with diazomethane, trimethylsilylated and subjected to GC-MS analyses. The resulting profiles of products are shown in Figure 5. Incubations with both LeAOS3 and ZmAOS resulted in the formation of two nonpolar products 11 and 12 ( Figure 5A,B). The profile of LeAOS3 incubation products possessed α-ketol as a major constituent ( Figure 5A). In contrast, analogous ZmAOS incubation yielded mainly the products 11 and 12 ( Figure 5B), while α-ketol was only a minority. Compounds 11 and 12 were formed at a ratio of about 2:1 (by total ion current) with LeAOS3 and 2:3 with ZmAOS. The mass spectra and fragmentation schemes for products 11 and 12 presented in Figure 6 corresponded to those previously described for macrolactones (12Z)-10-oxo-12-octadecen-11-olide and (12Z)-10-oxo-12-octadecen-9-olide, respectively [16]. Catalytic hydrogenation over PtO 2 converted products 11 and 12 into the saturated analogues, 10-oxooctadecan-11olide and 10-oxooctadecan-9-olide (Supplementary Figure S5). Compounds 11 and 12 are the products of intramolecular nucleophilic substitution, specifically, the carboxylic group attack on the oxirane of allene oxide.

General Aspects of Product Specificities of LeAOS3 and ZmAOS
Allene oxides (Me) produced by both ZmAOS (CYP74A19) and LeAOS3 (CYP74C3) were isolated and purified by NP-HPLC. ZmAOS provided a sufficient yield of allene oxide. The results of the NMR study demonstrated that ZmAOS (CYP74A19) produced the 10(E) isomer of allene oxide (9,10-EOD) quite specifically. The isolation of allene oxide produced by LeAOS3 was more complicated, largely due to the competitive formation of α-ketol (2). Thus, the yield of allene oxide was significantly lower compared to ZmAOS. Nevertheless, the allene oxide produced by LeAOS3 was isolated. The NMR study revealed the presence of two geometrical isomers of allene oxide, in full agreement with previous work [10]. The ratio of the 10(E) and 10(Z) isomers of 9,10-EOD detected was approximately 4:1. Thus, two enzymes possessed different geometrical specificities. A summarized scheme of allene oxide (9,10-EOD) synthesis and conversions by LeAOS3 and ZmAOS is presented in Figure 7. In the present work, we used the approach of biphasic incubations of AOSs with substrates originally employed by Alan Brash [23], increasing the allene oxide yield due to the improved protection from hydrolysis. The combination of biphasic incubations with direct ethanol trapping of allene oxide without hexane evaporation enabled the efficient ethanolysis of short-lived allene oxide. Ethanol was used instead of methanol, which was In the present work, we used the approach of biphasic incubations of AOSs with substrates originally employed by Alan Brash [23], increasing the allene oxide yield due to the improved protection from hydrolysis. The combination of biphasic incubations with direct ethanol trapping of allene oxide without hexane evaporation enabled the efficient ethanolysis of short-lived allene oxide. Ethanol was used instead of methanol, which was originally used by Mats Hamberg [11], due to its free mixability with hexane. The obtained results showed that biphasic incubations followed by direct ethanol trapping are quite useful for allene oxide detection in reaction mixtures.
The ethanol trapping experiments with LeAOS3 and ZmAOS revealed distinct product specificities for both enzymes. Trapping with both LeAOS3 and ZmAOS expectedly yielded the products of allene oxide ethanolysis. On the other hand, two enzymes behaved differently in several respects. Firstly, LeAOS3, unlike ZmAOS, stereospecifically produced (9R)-α-ketol at a high yield during the brief incubations. Secondly, LeAOS3 produced a detectable yield of cis-10-oxo-PEA (4), which was essentially absent with ZmAOS. Thirdly, a perceptible peak of the Favorskii-type product 3 (Et/Me) was present after the LeAOS3 trapping experiments but nearly absent after analogous ZmAOS experiments. Detection of the Favorskii-type products indicated the presence of the short-lived cyclopropanones along with allene oxides. The Favorskii-type products (Et/Me) were formed via the ethanolysis of cyclopropanones (Figure 7), the valence tautomers of allene oxide.
The ethanol trapping performed after both LeAOS3 and ZmAOS incubations with 13-HPOT resulted in detectable yields of the Favorskii-type product 7 (Et/Me), in addition to the allene oxide trapping product 5 (Supplementary Figure S8). Products 7 and 5 were formed through the ethanolysis of cyclopropanone and allene oxide, respectively. This indicated the co-existence of cyclopropanone with allene oxide (Supplementary Figure S8). In addition to the Favorskii-type product 7, cis-12-oxo-PDA (8) was detectable at a low yield after incubations with 13-HPOT. Thus, only a small part of the cyclizable allene oxide 12,13-EOT undergoes cyclization under these conditions. In contrast, in experiments with LeAOS3 (but not ZmAOS), the 9,10-EOD was quickly cyclized.
The expected allene oxide ethanolysis product 9 was also detected after similar experiments with 13-HPOD. At the same time, only traces of the Favorskii-type product, analogous to compound 7, and no cyclopentenone were detectable after the trapping experiments after LeAOS3 or ZmAOS incubations with 13-HPOD (Supplementary Figure S9). Thus, the formation of cyclopropanone showed a dependence on the β,γ double bond at the oxygenbearing carbon of hydroperoxide, in agreement with the previously published data [21].

Cyclopentenone Formation
Allene oxides are able to cyclize spontaneously. However, the natural allene oxides, generated enzymatically, exist in the presence of water. Thus, their cyclization competes with hydrolysis. Most allene oxides, e.g., those synthesized from the linoleic acid hydroperoxides, are quickly hydrolysed to α-ketols and afford only traces of cyclopentenones. There are two exceptions. The first is exemplified by allene oxides, such as 12,13-EOT, which have an additional double bond in the β,γ position towards the oxirane [20,24]. Their cyclization capability was attributed to the anchimeric assistance of this double bond, which promotes oxirane opening and the formation of oxyallyl species, which are essential for cyclization [20,24]. The second exception is represented by CYP74C AOSs. The capability of biosynthesizing cis-10-oxo-PEA is a unique property of CYP74C AOSs such as CYP74C3 [10,[16][17][18]. The most common AOSs of the CYP74A subfamily, such as ZmAOS, do not afford any noticeable yields of cyclopentenones from linoleic acid hydroperoxides [19,20]. Previously, the capability of LeAOS3 to produce 10-oxo-PEA was attributed to either (a) the formation of presumably cyclizable allene oxide (10Z)-9,10-EOD [10] or (b) control of allene oxide cyclization by LeAOS3 itself [18].
Observations from the present work allow one to presume that CYP74C AOSs such as LeAOS3 produce a tautomeric pair, allene oxide-oxyallyl, where the balance is partly shifted in favour of the latter. Both cyclopropanone and cyclopentenone ring closures are accomplished via the oxyallyl diradical (Figure 7). Both the Favorskii-type product (formed through the cyclopropanone) and 10-oxo-PEA production from 9-HPOD in the presence of LeAOS3 (but not ZmAOS) were detected in the present work. Their formation is a distinctive property of LeAOS3 compared to ZmAOS.
The cyclization of allene oxides to cyclopentenones was originally supposed to proceed via the conrotatory electrocyclization of a pericyclic pentadienyl cation [13,25,26]. However, recent theoretical studies have revealed that the cyclization occurs via the pericyclic pentadienyl diradical (oxyallyl diradical) [24,[27][28][29][30][31]. The Favorskii-type product detected after incubating LeAOS3 with 9-HPOD is evidently formed via cyclopropanone ( Figure 7). In turn, the closure of the cyclopropanone 3-membered ring proceeds through the oxyallyl diradical [24]. Therefore, these results indicate the occurrence of oxyallyl diradical in 9,10-EOD conversions in the presence of LeAOS3. The elevated cyclopentenone (cis-10-PEA) formation with LeAOS3 is presumably due to the occurrence of an oxyallyl diradical, a precursor of both cyclopropanone and cyclopentenone. In contrast, the yields of both the Favorskii-type product and cis-10-PEA are dramatically lower with ZmAOS. These findings allow us to propose that CYP74C AOSs such as LeAOS3 produce a tautomeric pair of allene oxide and oxyallyl, with a higher proportion of the latter. Equilibrium conversion of allene oxide and oxyallyl might also explain the appearance of the (10Z) isomer of allene oxide along with the ordinary (10E)-isomer. Back conversion of oxyallyl to allene oxide may be accompanied by a partial inversion of double-bond geometry.
These thoughts are supported by the following observations. Firstly, 10-oxo-PEA is quickly formed in parallel with allene oxide, 9,10-EOD and (9R)-α-ketol, as seen from the results of trapping experiments with LeAOS3 but not ZmAOS. In contrast, conversion of 13-HPOT via 12,13-EOT leads to α-ketol, trapping product, while the 12-oxo-PDA yield was relatively low. Secondly, keeping allene oxide 9,10-EOD (free acid) in hexane results in macrolactone formation via intramolecular nucleophilic substitution with the carboxylic group but no observable cyclization into 10-oxo-PEA. This indicates that the cyclization, if it ever occurred to a small extent, could not compete with the outrunning macrolactone formation.

Concluding Remarks
1. (9R)-α-ketol is a kinetically controlled LeAOS3 product in addition to the 9,10-EOD, as judged by the data of biphasic incubations followed by ethanol trapping. The α-ketol was detected at nearly equal yields with allene oxide ethanolysis product upon the very brief biphasic incubations. In contrast, ZmAOS (CYP74A19) produced predominantly the 9,10-EOD (detected as the ethanolysis product) upon the identical incubations, while α-ketol was only a minority.
6. Oxyallyl diradical, a common intermediate of cyclopropanone and cyclopentenone formation, is supposed to be produced by LeAOS3 as a tautomeric form of allene oxide (9,10-EOD). This might explain the uncommon capability of LeAOS3 to produce appreciable yields of cis-10-oxo-PEA and the Favorskii-type product in contrast with ZmAOS. The appearance of a distinctive 10(Z) isomer of allene oxide with LeAOS3 may also be caused by the double-bond geometry inversion via the oxyallyl-allene oxide tautomeric equilibrium.

Expression and Purification of Recombinant Enzyme
The open reading frame (ORF) of the maize AOS (ZmAOS, CYP74A19) was earlier cloned into the pET32 Ek/LIC vector (Novagen, Madison, WI, USA) to yield the target recombinant proteins with His-tags at N and C termini [21]. Recombinant plasmid pET-23a containing the ORF of the tomato AOS (LeAOS3, CYP74C3) was generously gifted by Prof. G. Howe (Michigan University, USA). The resulting constructions were transformed into the Escherichia coli host strain BL21 (DE3)pLysS (Novagen, Madison, WI, USA). The recombinant gene was expressed in host cells, as described before [32]. Purification of the His-tagged recombinant protein was performed using a Bio-Scale Mini Profinity IMAC (immobilized metal affinity chromatography) cartridge in the Bio-Rad NGC Discover 10 Chromatography System (Bio-Rad Laboratories, Moscow, Russia). The recombinant enzyme was eluted from the cartridges using 50 mM histidine. The homogeneity in the purified protein was confirmed by SDS-PAGE. The haemoprotein concentration was estimated using the pyridine haemochromogen assay [33].

The Allene Oxide Conversions in an Aprotic Solvent
LeAOS3 and ZmAOS preparations (100 µg) dissolved in 100 mM phosphate buffer (100 µL), pH 7.0, were extensively vortexed with 9-HPOD (200 µg) in hexane (4 mL) at 0 • C for 60 s. The water was quickly frozen at −79 • C, and the hexane phase was decanted. The resulting allene oxide (free acid) preparation was allowed to stay in hexane solution for 20 h. Then, the products were treated with diazomethane, trimethylsilylated and subjected to GC-MS analyses.

Methods of Spectral Analyses
Products were analysed as Me/TMS derivatives by GC-MS. The GC-MS analyses were performed using a Shimadzu QP5050A mass spectrometer connected to a Shimadzu GC-17A gas chromatograph equipped with a Supelco MDN-5S (5% phenyl, 95% methylpolysiloxane)-fused capillary column (length, 30 m; ID 0.25 mm; film thickness, 0.25 µm). Helium at a flow rate of 30 cm/s was used as the carrier gas. Injections were made in split mode using an initial column temperature of 120 • C and injector temperature 230 • C. The column temperature was raised at 10 • C/min until 240 • C. The electron impact ionization (70 eV) was used. Most GC-MS analyses were carried out in full-scan mode. The NMR 1 H, 1 H-1 H-COSY, 1 H-13 C-HSQC, 1 H-13 C-HMBC, 1 H-1 H-NOESY and 1 H-1 H-TOCSY spectra ([ 2 H 14 ]n-hexane) were recorded on a Bruker Avance III 600 spectrometer at 253 K. Funding: The authors are thankful for the financial support from the government assignment for the FRC Kazan Scientific Center of the Russian Academy of Sciences (preparation of the recombinant enzyme). The study's catalytic activities of the recombinant enzymes, as well as the structural and mechanistic studies, were carried out with funding from the Russian Science Foundation [21-14-00397].