Biocatalytic Properties and Structural Analysis of Eugenol Oxidase from Rhodococcus jostii RHA1: A Versatile Oxidative Biocatalyst

Biocatalytic Properties and Structural Analysis of Eugenol Oxidase from Rhodococcus jostii RHA1:AVersatile Oxidative Biocatalyst Quoc-Thai Nguyen+,[a, c] Gonzalo de Gonzalo+,[b] Claudia Binda,[c] Ana Rioz-Martnez,[d] Andrea Mattevi,*[c] andMarco W. Fraaije*[a] Introduction Biocatalysts have becomeavaluable tool for performing selective oxidation reactions that usually require harsh conditions and reagents when performed by conventional chemistry approaches.[1,2] Oxidases make up an emerging biocatalyst class for carrying out biotransformations that involvetwo- or fourelectron oxidations of substrates with use of molecular oxygen as amild oxidant.[3] Among them, the so-called flavoprotein oxidases can be valuablefor synthetic applications, because they can participate in abroad variety of oxidative procedures with high chemo-, regio-, and/orenantioselectivity.[4–6] Catalysis for theseenzymes occurs in two half-reactions:1)flavin cofactor reduction by oxidation of the substrate, and 2) reoxidation of the flavin by molecular oxygen.[7] Awide set of flavoprotein oxidaseshave been extensively employed in biotransformations, with glucose oxidase,[8] d-aminoacid oxidase,[9] and cholesteroloxidase[10] as well-known examples. The vanillyl alcohol oxidase (VAO) familyisagroup of flavoprotein oxidases that share acommonFAD-binding domain and present similar structuralfeatures.[11–13] In addition to alcohol oxidations, leading in mostcases to aldehydes and ketones,the enzymes of this family can also catalyze amine oxidations, thiol oxidations, hydroxylations, and even C¢Cbondformation reactions.[5] Arepresentativefamily member is VAO from Penicillium simplicissimum,anenzyme that acts on 4-hy- droxybenzylic compounds.[14,15] Kinetic studies have shown that the action of VAOinvolves an initial hydride transfer from the substrate to the FAD; this is facilitated by residues of the protein and generates a p-quinonemethide intermediate. Afterwards, the reduced flavin is reoxidized by molecular oxygen, yieldinghydrogen peroxide, whereas the p-quinone methide undergoes hydration in the VAOactive site, yieldingthe corresponding alcoholoraldehyde.[16,17] However, for some substratesthe hydration of the p-quinone methide is inefficient, resultinginthe formation of the corresponding p-alkenylphe- nol:that is, formally,adehydrogenation process.[18] The genome of Rhodococcus jostii RHA1 contains aremarkably large number of genes coding for oxidative enzymes.[19] Amongthem, one gene codes for aeugenoloxidase (EUGO) that displays significant sequence identity with VAO(40%) and asimilar but not fully overlapping pattern of substrate preferences.[20] Unlike VAO, which expressespoorly in Escheri- Eugenoloxidase(EUGO) from Rhodococcus jostii RHA1 had previously been shown to convert only alimited set of phenolic compounds. In this study,wehave explored the biocatalytic potentialofthis flavoprotein oxidase, resultinginabroadened substrate scope and adeeper insight into its structural properties. In addition to the oxidation of vanillyl alcohol and the hydroxylation of eugenol, EUGO can efficiently catalyzethe dehydrogenation of various phenolic ketones and the selective oxidation of aracemic secondary alcohol—4-(1-hydroxyethyl)- 2-methoxyphenol. EUGO was also found to perform the kinetic resolution of aracemic secondary alcohol. Crystal structures of the enzymeincomplexes with isoeugenol, coniferyl alcohol, vanillin, and benzoatehave been determined. The catalytic centerisaremarkable solvent-inaccessible cavity on the si side of the flavin cofactor.Structural comparison with vanillyl alcohol oxidase from Penicillium simplicissimum highlightsafew localized changes that correlate with the selectivity of EUGO for phenolic substrates bearing relativelysmall p-substituents while tolerating o-methoxy substituents. [a] Q.-T.Nguyen,+Prof. M. W. Fraaije Molecular Enzymology,GroningenBiomolecular Sciencesand Biotechnology Institute, UniversityofGroningen Nijenborgh 4, 9747 AG Groningen (NL) E-mail:[email protected] [b] Dr.G.deGonzalo+ Departmento de QumicaOrgnica, Universidad de Sevilla c/Profesor Garca Gonzlez 1, 41012 Sevilla (Spain) [c] Q.-T.Nguyen,+Dr.C.Binda, Prof. A. Mattevi DepartmentofBiology and Biotechnology,UniversityofPavia Via Ferrata 1, 27100 Pavia(Italy) E-mail:[email protected] [d] Dr.A.Rioz-Martnez StratinghInstitute for Chemistry,University of Groningen Nijenborgh 4, 9747 AG Groningen (NL) [+ +]These authors contributed equally to this work. Supportinginformation for this article can be found under http:// dx.doi.org/10.1002/cbic.201600148. Ó2016 The Authors. Published by Wiley-VCH Verlag GmbH&Co. KGaA. This is an open access article under the terms of the Creative Commons AttributionLicense, whichpermits use, distributionand reproduction in any medium, provided the original work is properly cited. ChemBioChem 2016,17,1359 –1366 Ó2016 The Authors. PublishedbyWiley-VCH Verlag GmbH &Co. KGaA, Weinheim1359 Full Papers DOI:10.1002/cbic.201600148 chia coli,recombinant EUGO can be produced in large quantities (a rather impressive 160 mg of pure protein from 1L of E. coli culture), which makes it an attractive target for largescale oxidative transformations. In this report, the EUGO-catalyzed oxidation of eugenol and vanillyl alcohol, the dehydrogenation of various a,b-unsaturat- ed ketones, and the kinetic resolution of aracemic secondary alcohol were examined. Optimal conditions for these biocatalytic conversions were determined, revealingthat the enzyme is quite toleranttowards organic solvents. Valuable insights into the substrate selectivity of EUGO were obtained by resolving enzyme crystal structures in complexes with isoeugenol, coniferyl alcohol, vanillin, and benzoate at 1.7–2.6 æresolution. Inspection of the active-site cavity structure reveals aconserved substrate-binding mode with respecttothat of VAObut also highlightsspecific features that fine-tunesubstrate specificity. Results and Discussion Substrate specificity of eugenol oxidase Previous experiments had shown that EUGO is active on phenolic substrates, with the identification of eugenol, vanillyl alcohol, vanillylamine, and indan-5-ol as good substrates (kcat = 0.26–12 s¢1), whereas 4-ethylguaiacol and 4-(methoxymethyl)- phenol were poorlyc onverted (kcat =0.004–0.026s ¢1).[20] Althoughitwas shown that EUGO displays oxidaseactivity towards thesearomatic compounds, no product analysis had been carried out for any of the above substrates. Therefore, as afirst step in our project, we performed conversionsand subsequent product analyses to establish whichphenolic products can be prepared with the aid of EUGO (Scheme 1). Incubation of eugenol(1)with the enzymeinTris·HCl buffer (pH 8.0) containing 1% (v/v)DMSO afforded coniferyl alcohol(2)with complete conversion after 4hat room temperature. If the reaction was performed in the presence of 10%DMSO (v/v)ascosolvent, an increase in the enzymatic activity was observed, with only 2hbeing required to reach completeconversion.Vanillyl alcohol (3)was very effectively converted by EUGO. After 4h, vanillin (4)was recovered with complete conversion when the oxidation was performed at pH 8.0 either in aqueous medium or in buffer containing 10%(v/v)DMSO. To explore the potential of EUGO we also tested other substrates. An interesting result was observed when zingerone (5), a ginger-derived naturalproduct,[21] was incubated with EUGO in Tris·HCl buffer (pH 8.0). After 4h,58% conversion was observed, with dehydrozingerone (6)being obtained as the single product. When the reaction wascarried out for12h, complete conversion was achieved. Again, the presence of 10%(v/v)DMSO in the reactionmedium increased the enzymatic activity,with the reaction being complete after 8h.This shows that EUGO is also able to catalyze dehydrogenation processes.Interestingly,dehydrozingerone is also presentin ginger and exhibits interesting medicinal properties.[22] Analogously,raspberry ketone (7)was incubated with EUGO at room temperature. This led to the formation of 4-phenyl- but-3-en-2-one (8)asthe sole dehydrogenation product. The enzymepresented significantlylower activity with this compoundthan with zingerone, with only 48%conversion being observed after 24 hatpH8.0. Previous experimentshave shown that indan-5-ol (9)isa EUGO substrate,[20] which is interesting in view of the cyclic substituent of this compound. After 24 hincubation of this substrate at room temperature and pH 8.0, 34%conversion was achieved. The resulting mixture contained two different products in a7:3 proportion. Analysis by MS/GC and NMR revealed that the major product was 1H-inden-6-ol(10), the dehydrogenation product, whereas 5-hydroxyindan-1-one (11) was obtained as the minor product. Again, when the oxidation was performed in the presenceof10% (v/v)DMSO, the enzymatic activity was doubled, leadingto63% conversion after 24 h. Other potential substrates, such as 4-(1,3-dithiolan-2-yl)phenol, capsaicin, (4-hydroxyphenyl)acetonitrile, 3-(4-hydroxyphe- nyl)propanoic acid, 2-methoxy-4-methylphenol, and 4-(hept-1- enyl)-2-methoxyphenol, did not show any significant oxidase activity nor yield any product after incubation with EUGOfor 24 h. These data were complemented by measurements of the initial oxidase rates for someofthe EUGO substrates (see Table S1 in the Supporting Information). Next to vanillyl alcohol, eugenol (1)and its analogue 4-(1-hydroxyethyl)-2-meth- oxyphenol(12)were found to be the best substrates for this biocatalyst (2.1 and4.3 s¢1,respectively), whereas raspberry ketone(7)was by far the most slowly reacting substrate (0.02 s¢1)ofall tested. In essence, EUGO prefers phenolic substrateswith ortho-methoxy substituents and arelatively small group at the para position. Scheme1.Reactions catalyzed by EUGO from R. jostii RHA1. a) EUGO, Tris·HCl (50 mm,pH7.5), DMSO, RT,200 rpm. ChemBioChem 2016,17,1359 –1366 www.chembiochem.org Ó2016 The Authors. PublishedbyWiley-VCH Verlag GmbH &Co. KGaA, Weinheim1360 Full Papers Enzymatic dehydrogenationofraspberry ketone (7): Effect of organic cosolvent, temperature, and pH on EUGO activity The data above show that raspberry ketone (7)was only slowly converted by EUGO. For this reason, we set out to optimize the reaction conditions using 7as asubstrate. Because we had observed that EUGO is quite tolerant towards DMSO [10%(v/v)ofDMSO resulted in higherlevelsofconversion for all testedsubstrates],severalsolvents were tested in the enzymatic dehydrogenation of this ketone. The reaction carried out in the presence of 10%(v/v)DMSO afforded the final product 8with aslightly higher level of conversion than achieved in the aqueous medium (60%conversion after 24 h; Table 1, entry 2). The use of other cosolvents, with different physicochemicalproperties, led to much lower levels of conversion (entries 3–7). With DMSO selected as the best cosolvent for the enzyme, the effect of its concentration on the enzymaticactivity was analyzed (Figure 1). The conversion of 7catalyzed by EUGO is optimal at 10%(v/v)cosolvent. Higher DMSO concentrations led to progressive deactivation of the biocatalyst, although it is noteworthy that amoderate level of conversion (30%) was still obtained at 50%(v/v)DMSO after 24 h. This shows that EUGO is fairly solvent-tolerant. The effect of temperature on the EUGO-catalyzed bioconversion of 7in Tris·HCl (pH 8.0)/10%(v/v)DMSO medium was also analyzed (Table 1, entries 8–12). At 178C, only 15%conversion was obtained after 24 h. Higher temperaturesled to increased levels of conversion,with60% of 8being achieved at 258C. A slightly lower level of conversion was observed on working at higher temperatures but EUGOremained active at 458C, with 33%conversion after 24 h. In contrast, no reaction was observed at 608C. Dehydrogenation of raspberry ketone was also performed in Tris·HCl buffer with 10%(v/v)DMSOatdifferent pH values (from pH 7.0 to 10.0;entries 13–16). The enzyme showed higher activity at higherpHvalues (from 8.0 to 10.0), with levels of conversion around 50–60%after 24 h. Lower pH values led to arapid decrease in conversion,with only 23% conversion at pH 7.0. Collectively,this set of experiments indicated that EUGO is highly solvent-tolerant and maximally active at room temperature, whereas it is rathersensitivetovariation of the pH of the medium. Eugenol oxidase stability EUGO is arobustenzyme,being activeover awide range of pH values and tolerating the use of several cosolvents. We determined its thermostability by determining the apparent melting temperatures(Tmvalues) under different conditions by the ThermoFAD method.[23] At pH valuesfrom 5.0 to 8.0 the enzyme is exceptionally thermostable, exhibiting Tmvalues around658C(Table S2). Moreover,the presence of 10%(v/v) DMSO or ethyl acetate does not significantly affect the enzyme’s thermostability.Only extreme pH conditions or the addition of some organic cosolvents decrease the Tmbelow 608C. This high thermostability is in line with the observation that EUGO can be employedinreaction mixtures including an organic cosolvent. As would be expected, enzyme-stabilizing agents such as glyceroland salts such as NaCl have apositive effect on the thermostability of EUGO. Kinetic resolution of 4-(1-hydroxyethyl)-2-methoxyphenol [()-12] catalyzed by EUGO Because vanillyl alcohol [3]isanexcellent substrate for EUGO, its racemic analogue 4-(1-hydroxyethyl)-2-methoxyphenol [()- 12]was studied in asecondary alcoholkinetic resolution procedure (Scheme 2). In this process, one enantiomer of the substrate is oxidized to the ketone while the other enantiomer hence becomes enantiomerically enriched.EUGO is atrue oxidase, so molecular oxygen consumption leads to the formation of hydrogen peroxide. For this reason, the reactions were Table 1. Effect of organiccosolvents,temperature, and pH on EUGO activity in the catalysis of the dehydrogenationofraspberry ketone (7) under different reaction conditionsafter 24 h. Cosolvent (v/v)pHT[8C] c[%][a] 1none 8.0 25 48 210%DMSO 8.0 25 60 310%1,4-dioxane 8.0 25 10 410%CH 3CN 8.0 25 3 510%EtOAc 8.0 25 17 610%CH 2Cl28.0 25 3 710% tBuOMe 8.0 25 7 810%DMSO 8.0 17 15 910%DMSO 8.0 30 43 10 10%DMSO 8.0 37 38 11 10%DMSO 8.0 45 33 12 10%DMSO 8.0 60 3 13 10%DMSO 7.0 25 23 14 10%DMSO 7.5 25 38 15 10%DMSO 9.0 25 53 16 10%DMSO 10.025 57 [a] Conversion determined by GC. Figure 1. Effect of DMSOconcentration on the degree of conversion( *)in EUGO-catalyzed oxidation of raspberry ketone (7,2.0 mm)after 24 h. ChemBioChem 2016,17,1359 –1366 www.chembiochem.org Ó2016 The Authors. PublishedbyWiley-VCH Verlag GmbH &Co. KGaA, Weinheim1361 Full Papers performed in the presence of catalase, which catalyzes the decomposition of hydrogen peroxide, thereby regenerating molecular oxygen. It was observed that the presence of small amountsofcatalase indeed led to slightly higherlevels of conversion with no effect on the selectivity (data not shown). Racemic ()-12 was found to be selectively oxidized by EUGO into ketone 13,asshown in Table 2. Initial experiments performed with aTris·HCl buffer (50 mm,pH9.0) showed that the kinetic resolution occurred rapidly and with very low selectivity (Table 2, entry 1). Use of lower pH values (see pH 7.5, entry 2) led to significantly higherenantioselectivity (E=35), resulting in the enantioenrichment of (S)-12.Conversely,the presence of 10%(v/v)DMSO in the reaction mediumdecreased the level of conversion andthe selectivity of the reaction (entry 3). For this reason,the amount of this organic cosolventwas reduced to 5% (v/v), resultinginahigherlevel of conversion, as shown in entry 4. The effects of alternative organic cosolvents (5%) in the EUGO-catalyzed kinetic resolution of ()-12 were studied (entries 5–11). The use of ethyl acetate (Table 2, entry 5) led to similar resultsinterms of activity and selectivity,whereas the best result was afforded with iPr2O, in aprocess with apromising selectivity (E=45) and a39% level of conversion after 4h (entry 7). Kinetic resolutions in reactionmediacontaining 5% (v/v)tBuOMe, hexane, or octan-2-ol were much faster,with up to 50%conversion after 1.5 h, but the selectivity was from moderate to poor.Oxidationswith cosolvents such as toluene or CH2Cl2occurred with very low conversion efficiencies (entries 9and 10). The overall three-dimensional structure of EUGO The structure of EUGO in complexwith the inhibitor isoeugenol was solved at 1.7 æresolutionbymolecular replacement with use of the VAOmonomer (PDB ID:2VAO)[11] as asearch model.The asymmetricunit contains two enzyme subunits forming acompactdimer that is also observed in solution (Figure 2A andB). The dimer interface area is quite large, approximately 3500 æ2,accountingfor 16.5%ofthe monomer’s surface as calculated by PISA.[25] The overall structure of EUGO is highly similar to that of VAO, as indicated by a1.1 ærmsd for 490 Caatom pairs (45%sequence identity).[11] As predicted, EUGO also shares with VAOthe presence of acharacteristiccovalent bond between His390(equivalent to His422 of VAO) and the C8M atom of the flavin ring (Figure 3). Nevertheless, the oligomeric organization in the two enzymes is partly different, because, though both form essentially the same functional dimer (1.3 ærmsd for 999 Caatom pairs), EUGOlacks the dimer–dimer interacting loop that in VAOmediates the octameric structure of the enzyme. Table 2. EUGO-catalyzed kinetic resolution of the racemic alcohol 4-(1-hy- droxyethyl)-2-methoxyphenol [()-12].[a] Cosolvent t[h] c[%][b] ee [%][c] E[d] 1[e] none 449 23 2.0 2none 433 45 35 310%DMSO 429 17 3.0 45% DMSO230 35 14 55% EtOAc431 49 38 65% tBuOMe1.5 40 55 18 75% iPr2O439 59 45 85% hexane 448 13 1.5 95% CH2Cl243–– 10 5% toluene 4612 3.7 11 5% octan-2-ol1.5 50 65 9.1 [a] For reaction conditions, see the Experimental Section. [b] Determined by GC. [c] Determined by HPLC.[d] Enantiomericratio, E=ln[(1¢c)/ (1¢ees)]/ln[(1¢c)/(1++ees)].[24] [e] Reactionperformed at pH 9.0. Figure 2. Crystalstructure of eugenol oxidase from R. jostii RHA1 in complex with the inhibitor isoeugenol. A) Ribbon diagram of the EUGO dimerwith the two monomers in ice blue (monomer A) and gold (monomer B), respectively.B)Close-up sliced view of the EUGO surface in the vicinity of the catalytic pocket. The inhibitor isoeugenolisburied in apocketinclose contact with adeep invagination originating at the top of the dimer and being formed by residues at the dimer interface, which might representthe pathway for ligand accesstothe active site. The orientation and atom coloringis the sameasthat of the structure shownin(A). The FADcarbon atomsare showninyellow,the ligand carbon atoms in lawn green,oxygen atoms in red, nitrogen atomsinblue, and phosphorus atoms in magenta. Scheme2.EUGO-catalyzed kinetic resolution of racemic 4-(1-hydroxyethyl)- 2-methoxyphenol [()-12]. a) EUGO,Tris·HCl (50 mm,pH7.5), cosolvent, 308C, 200 rpm. ChemBioChem 2016,17,1359 –1366 www.chembiochem.org Ó2016 The Authors. PublishedbyWiley-VCH Verlag GmbH &Co. KGaA, Weinheim1362 Full Papers The active site of EUGO The active site of EUGOconsists of around-shaped cavity located in front of the flavin cofactor (Figure 2B). As in VAO, this substrate-binding site is completely solvent-inaccessible.However,inspectionofthe dimericstructure of EUGO reveals aputative admission path for the ligands. The active-site cavity of each monomer is in close contact with afunnel-shaped large chamberthat runs along the dimer interface to the top of the structureand might represent apassagewayfor the diffusion of the substrate into the binding site (in the orientation of Figure 2B). The structural bases of substrate and inhibitor binding by the cavity were explored by solving the crystal structures of EUGO in complexes with isoeugenol(inhibitor), coniferyl alcohol (2,reaction product and inhibitor), and vanillin(4,reaction product);these were obtained by crystal soaking, resulting in highly defined electron density maps (Figure 3A–C). All these ligands bind with their aromatic moiety stacking against the pyrimidine ring of the flavin and the Caatom of the aliphatic substituent lying right above the flavin N5 atom at adistance of 3.2 æ. In this orientation, the ligand 4-hydroxy group is H-bonded to Tyr91 and Tyr471 and is also in close contact with Arg472(3.5 æ). These three residues are involved in the stabilization of the phenolate form of the substrate and are strictly conserved in VAO. The fact that such aTyr-Tyr-Arg cluster formsananion-binding site is supported by the structure obtained from native crystals(i.e.,not soaked in aligand solution),which exhibited aclear electron density peak close to the flavin rings of both monomers presentinthe asymmetric unit. This residual density was tentatively assignedtoabenzoate ion bound with its carboxylate group directly interacting with the Tyr-Tyr-Arg residues (Figure 3D). Indeed,wefound benzoate to be aweak EUGO inhibitor,probablycaptured by the enzyme in the cell and retained during purification. Collectively,these features are fully consistent with the substrate being bound in the phenolic form andoxidized through areaction mechanism involving ahydride transfer from the Cato the flavin N5, coupled to the stabilization of a p-quinone methide intermediate as demonstrated forVAO.[16] However,comparativeanalysis of the three-dimensional structures also revealed striking differences between EUGO and VAO. Aboveall, inhibitor/product binding differsina1808 flipped orientation. This can be visualized by superimposing the isoeugenol complexesofthe two enzymes. As shown in Figure 4A,the ligand methoxy groups are in opposed orientations and occupy different niches in the active sites of the two enzymes.InEUGO, the methoxy moiety is in close contact with Gly392, which replaces the Phe424 residueofVAO. This bulkieraromatic side chain of VAOlikely createssteric hindrance, impedingthe binding orientation found in EUGO. Con- Figure 3. The EUGO active site in complex with A) isoeugenol, B) coniferyl alcohol, C) vanillin, and D) benzoate. Unbiased 2Fo¢Fcomit electrondensitymaps calculated at 1.7–2.6æresolution and contouredat1.0sare drawn in chicken-wire style. Hydrogen bondcontactsare shownasdashed lines.Residuesin direct contact with the bound ligands are labeled. Protein, ligand, and FADcarbonatoms are colored in gray,lawn green,and yellow,respectively,oxygen atoms are in red, nitrogen atomsare in blue, and phosphorusatoms are in magenta. The orientation is the sameasthat of monomer AinFigure2. ChemBioChem 2016,17,1359 –1366 www.chembiochem.org Ó2016 The Authors. PublishedbyWiley-VCH Verlag GmbH &Co. KGaA, Weinheim1363 Full Papers versely, the flipped isoeugenol bound to VAOpositions the methoxy group in direct contactwith avaline residue thatis conserved in EUGO (Val166, Figure 4A). It is gratifying that these findings correlate with the nonoverlapping patterns of substrate preferences and selectivities displayed by the two enzymes.Indeed, inspired by these structural data, we tested 4-allyl-2,6-dimethoxyphenol. This dimethoxy compound is not asubstrate of VAO.[15] Conversely,we found that it is converted by EUGOwith moderate efficiency (kcat =0.49 s¢1,Km=0.8 mm). Thus, the presence of Gly392(instead of Phe424 in VAO) is akey feature that enables EUGO to act on molecules bearing two o-methoxy substituents. This is in contrast with VAO, which does not accept dimethoxy-substi- tuted molecules as substrates.[15] The other relevant element in EUGO that partially differs from the situation in VAOinvolves the section of the activesite cavity that is involved in the binding of the substrate alkyl chain.Here, despite aset of conserved residues (Asp151, Tyr168, Arg278, and Glu378, corresponding to Asp170,Tyr187, Arg312, and Glu410 in VAO), there are afew major differing amino acids between the two enzymes:Thr457, Trp413, and Leu316 in VAOare replaced by Gln425, Leu381, andMet282, respectively,inEUGO (Figure 4B). In particular, these replacementscollectively reshape this part of the substrate site, which becomes narrower in EUGOthan in VAO. As aresult, it is more difficult for EUGO to accept long aliphatic chains and hydrophobic substituents than it is for VAO(Figure 4B). Indeed, whereas VAOisreactive towards alkylphenols and can accommodate aliphatic substituents containing up to seven carbon atoms,[11,26] we found no activity of EUGO with such compounds. Conclusion The synthetic potentialofeugenol oxidase(EUGO) in oxidative procedures has been demonstrated. The data show that EUGO is arobust biocatalyst, tolerating awide range of organic solvents and high temperatures.Interestingly,the addition of DMSO as cosolventincreased the enzymatic activity in most of the catalytic processes. EUGO could also be employedinthe kineticresolution of aracemic secondary alcohol with moderate to good selectivity.Structuralanalysis of the EUGOcatalytic centershowedthat the enzyme shares agenerally conserved substrate binding mode with VAObut with key different residues finely controlling the substrate specificity.With an effective expression system, detailed structuralinformation, and generic oxidase screening methods to hand, EUGO appears a perfect candidate for engineering tailored oxidasevariants. Experimental Section General materials and methods:Recombinant eugenol oxidase from R. jostii RHA1 was overexpressed and purified according to previously described methods.[20] All other chemicals and solvents were obtained from Acros Organics, Alfa Aesar,ABCR, Sigma–Aldrich, TCI Europe, and Roche Diagnostics. Flash chromatography was performed with Merck silica gel 60 (230–400 mesh). 1HNMR, 13CNMR, and DEPT spectra were recorded with TMS (tetramethylsilane) as the internal standard, with aBruker AC-300-DPX (1H: 300.13 MHz and 13C: 75.4 MHz) spectrometer.Mass spectra were recorded by ESI+with aHP1100 chromatograph mass detector or by EI with aFinnigan MAT95spectrometer.GC/MS analysis for compounds 1–11 was performed with aGCHewlett Packard 6890 Series II and aHewlett Packard 5973 chromatograph MS (Agilent Technologies) equipped with aHP-1 cross-linked methyl siloxane column (30 mÕ0.32 mmÕ0.25 mm, 1.0 bar N2). The injector temperature was 2258C, and the flame ionization detector (FID) temperature was 2508C. The following temperature program was employed:1008C(5min), 108Cmin¢1to 2508C(2min). tR1:11.0 min. tR2:15.6 min. tR3:12.2 min. tR4:11.3 min, tR5:14.5 min. tR6: 16.4 min. tR7:11.8 min. tR8:12.9 min, tR9:10.6 min. tR10: 10.8 min. tR11:9.6 min. GC analysis for measuring the conversion of ()-12 was performed with aRestek RtbDEXse (30 mÕ Figure 4. Comparison between EUGO and VAOactive sites. A) The substratebinding siteofEUGO in complex with isoeugenol(lawngreen) is superposed onto that of VAOincomplex with the same ligand (light green;PDB ID:2VAO). Unconserved residues in VAOare drawn with carbon atoms in white.The orientation and atom coloring is the same as that of the structure shown in Figure2.For the sake of clarity,Tyr168, Arg278, and Glu378 were omittedinthis figure. The gcarbon atom of isoeugenolinthe VAOstructure was disordered and was not included in the refined model.[11] B) The superposition of the EUGO·isoeugenolstructure onto VAOincomplex with 4- (hept-1-enyl)phenol (light green; PDB ID:1AHZ)shows that EUGOfeatures aslightly narrower cavity due to the presence of Gln425 (Thr457 in VAO), which wouldclash with ligands with longer alkyl chains. In this figurethe structure is rotated by about 458with respect to (A). For the sake of clarity, the side chains of I427/T459aswell as the labels of L381/W413 andH390/ H422were omitted. ChemBioChem 2016,17,1359 –1366 www.chembiochem.org Ó2016 The Authors. PublishedbyWiley-VCH Verlag GmbH &Co. KGaA, Weinheim1364 Full Papers 0.25 mmÕ0.25 mm, 1.0 bar N2)column. The following temperature program was used:708C(5min), at 38Cmin¢1to 1608C(2min), 208Cmin¢1to 1808C(7min). tR12:41.4 min. tR13:40.8 min. To monitor levels of conversion, substrates and products were quantified by use of calibration curves. HPLC analyses were developed with aHewlett Packard 1100 LC liquid chromatograph. The following conditions were used for the determination of the enantiomeric excess of alcohol 12:Chiralcel OJ-H column (0.46Õ25cm), isocratic eluent: n-hexane/EtOH (95:5), 408C, flow 1mLmin¢1.tR(S)- 12:31.4 min. tR(R)-12:35.1 min. General procedure for the enzymatic oxidations catalyzed by isolated EUGO:Unless stated otherwise, starting compounds 1,3, 5,7,and 9(2–10 mm)were dissolved in Tris·HCl buffer (50 mm, pH 8.0, 1.0 mL) containing, when indicated, organic cosolvent (10%, v/v)and EUGO (0.33 mm). Reaction mixtures were shaken at 250 rpm and room temperature for the times indicated. Once reaction was complete, the crude mixtures were extracted with EtOAc (3Õ500 mL). The organic phases were dried onto Na2SO4and analyzed directly by GC/MS. Kinetic resolution of racemic 4-(1-hydroxyethyl)-2-methoxyphe- nol catalyzed by EUGO:Unless stated otherwise, the starting racemic alcohol ()-12 (10 mm,Scheme 2) was dissolved in Tris·HCl buffer (50 mm,pH7.5) containing, when stated, organic cosolvent (5–10%, v/v,1.0 mL), catalase (2 mL, six unitsmL¢1), and EUGO (1.0 mm). Reaction mixtures were shaken at 250 rpm and 308Cin arotatory shaker for the times indicated. Once reaction was complete, the crude mixtures were extracted with EtOAc (2Õ500 mL). The organic phases were dried onto Na2SO4and analyzed directly by GC and HPLC in order to determine the levels of conversion to 13 and the enantiomeric excesses of the alcohol (S)-12. Measurement of initial rates:Conversion rates were determined by monitoring the consumption of molecular oxygen. Oxygen concentrations were monitored by use of aREDFLASH sensor spot in a3mL cuvette in combination with aFiresting O2detector and light source (Pyroscience, Aachen, Germany). Measurements were performed with substrates (2.0 mm)inKPi buffer (50 mm,pH7.5) containing DMSO (10%, v/v). Determination of kinetic parameters for 4-allyl-2,6-dimethoxy- phenol:The reaction was monitored by following the absorbance of the product at 270 nm (e=14.1 mm¢1cm¢1at pH 7.5) in Tris·HCl (50 mm,pH7.5) at 258C. Analysis of EUGO thermostability:EUGO thermostability was assayed by use of the ThermoFAD protocol,[23] with aMiniOpticon real-time PCR detection system and 48-well RT-PCR plates (Biorad Laboratories). Samples were prepared by mixing EUGO (50 mm, 2.0 mL) in Tris·HCl (10 mm,pH7.5) together with the buffers and/or organic solvents to afinal volume of 20 mLand in duplicate for each condition tested. Protein crystallization, X-ray data collection, and structure determination:Native EUGO was crystallized by use of the sittingdrop vapor diffusion technique at 208Cbymixing equal volumes of EUGO (15 mgmL¢1)inTris·HCl (10 mm,pH7.5) and mother liquor containing PEG6000 (24%, w/v)and Tris·HCl (0.1m,pH8.0). All EUGO·ligand complexes were prepared by soaking (0.5–2 h) the crystals in cryoprotectant solutions consisting of PEG6000 (26%), Tris·HCl (0.1m,pH8.0), glycerol (20%), and the compound of interest (5.0 mm), followed by flash-freezing in liquid nitrogen. X-ray diffraction data were collected at the PXI and PXIII beamlines of the Swiss Light Synchrotron in Villigen, Switzerland (SLS) and at the ID23–1 beamline of the European Synchrotron Radiation Facility in Grenoble, France (ESRF). The images were integrated by MOSFLM,[27] whereas data scaling was performed by use of programs of the CCP4 suite.[28] The detailed data processing statistics of the collected datasets are shown in Table S3. EUGO’s structure was solved by molecular replacement with use of MOLREP[29] and the coordinates of VAO(PDB ID:2VAO)[11] as the search model devoid of all ligand and water molecules. COOT[30] and REFMAC5[31] programs were employed to carry out alternating cycles of model building and refinement (data shown in the Supporting Information). Figures were created by use of CCP4mg.[32] Acknowledgements This research was supported the European Union (EU) project ROBOX (grant agreementno. 635734) under the EU’s Horizon 2020 Programme Research and Innovationactions H2020-LEIT BIO-2014–1. G.d.G. (RamûnyCajal Program) thanks MINECO for personal funding. We thank the European Synchrotron Radiation Facility (ESRF) and the Swiss Light Source (SLS) for providing beamtime and assistance. We thank Mohamed Mohamed Habib for his experimental support. We also thank the BioStruct-X program (project no. 10205) for funding synchrotron trips. Keywords: biocatalysis ·dehydrogenation ·enzyme structures ·kinetic resolution ·oxidases [1] S. Schulz, M. Girhard, V. B. Urlacher, ChemCatChem 2012,4,1889–1895. [2] R. D. Schmid, V. B. Urlacher, Modern Biooxidation:Enzymes,Reactions and Applications,Wiley-VCH, Weinheim, 2007. [3] N. J. Turner, Chem. Rev. 2011,111,4073–4087. [4] M. J. H. Moonen, M. W. Fraaije, I. M. C. M. Rietjens, C. Laane, W. J. H. van Berkel, Adv.Synth. Catal. 2002,344,1023–1035. [5] W. P. Dijkman, G. de Gonzalo, A. Mattevi,M.W.Fraaije, Appl. Microbiol. Biotechnol. 2013,97,5177–5188. [6] R. T. Winter,M.W.Fraaije, Curr.Org. Chem. 2012,16,2542–2550. [7] P. Chaiyen,M.W.Fraaije, A. Mattevi, TrendsBiochem.Sci. 2012,37,373– 380. [8] S. B. Bankar, M. V. Bule, R. S. Singhal, L. Ananthanaryan, Biotechnol. Adv. 2009,27,489–501. [9] X. Ma, S. Deng, E. Su, D. Wei, Biochem. Eng. J. 2015,95,1–8. [10] S. Dieth, D. Tritsch, J. Biellmann, Tetrahedron Lett. 1995,36,2243–2246. [11] A. Mattevi, M. W. Fraaije, A. Mozzarelli,L.Olivi, A. Coda, W. J. van Berkel, Structure 1997,5,907–920. [12] M. W. Fraaije, W. J. van Berkel, J. A. Benen, J. Visser,A.Mattevi, Trends Biochem.Sci. 1998,23,206–207. [13] N. G. Leferink, D. P. Heuts, M. W. Fraaije, W. J. van Berkel, Arch. Biochem. Biophys. 2008,474,292–301. [14] E. de Jong, W. J. van Berkel, R. P. van der Zwan, J. A. de Bont, Eur.J.Biochem. 1992,208,651–657. [15] M. W. Fraaije, C. Veeger,W.J.van Berkel, Eur.J.Biochem. 1995,234, 271–277. [16] M. W. Fraaije, W. J. van Berkel, J. Biol. Chem. 1997,272,18111 –18116. [17] M. W. Fraaije, R. H. van den Heuvel, J. C. Roelofs, W. J. van Berkel, Eur.J. Biochem. 1998,253,712–719. [18] R. H. H. van den Heuvel,M.W.Fraaije, W. J. H. van Berkel, FEBS Lett. 2000,481,109–112. [19] M. P. McLeod, R. L. Warren, W. W. Hsiao, N. Araki, M. Myhre, C. Fernandes, D. Miyazawa, W. Wong, A. L. Lillquist, D. Wang, M. Dosanjh, H. Hara, A. Petrescu, R. D. Morin, G. Yang, J. M. Stott, J. E. Schein, H. Shin, D. Smailus, A. S. Siddiqui, M. A. Marra, S. J. Jones, R. Holt, F. S. Brinkman, K. Miyauchi, M. Fukuda, J. E. Davies, W. W. Mohn, L. D. Eltis, Proc. Natl. Acad. Sci. USA 2006,103,15582–15587. [20] J. Jin, H. Mazon, R. H. van den Heuvel, D. B. Janssen,M.W.Fraaije, FEBS J. 2007,274,2311–2321. [21] Y. Shukla,M.Singh, Food Chem. Toxicol. 2007,45,683–690. ChemBioChem 2016,17,1359 –1366 www.chembiochem.org Ó2016 The Authors. PublishedbyWiley-VCH Verlag GmbH &Co. KGaA, Weinheim1365 Full Papers [22] S. Yogosawa, Y. Yamada, S. Yasuda, Q. Sun, K. Takizawa,T.Sakai, J. Nat. Prod. 2012,75,2088–2093. [23] F. Forneris, R. Orru, D. Bonivento, L. R. Chiarelli,A.Mattevi, FEBS J. 2009, 276,2833–2840. [24] J. L. L. Rakels, A. J. J. Straathof, J. J. Heijnen, EnzymeMicrob.Technol. 1993,15,1051–1056. [25] E. Krissinel, K. Henrick, J. Mol. Biol. 2007,372,774–797. [26] R. H. H. van den Heuvel, M. W. Fraaije, C. Laane, W. J. H. van Berkel, J. Bacteriol. 1998,180,5646–5651. [27] T. G. Battye,L.Kontogiannis, O. Johnson,H.R.Powell, A. G. Leslie, Acta Crystallogr.Sect. DBiol. Crystallogr. 2011,67,271–281. [28] M. D. Winn, C. C. Ballard,K.D.Cowtan, E. J. Dodson,P.Emsley, P. R. Evans, R. M. Keegan, E. B. Krissinel, A. G. Leslie, A. McCoy, S. J. McNicholas, G. N. Murshudov,N.S.Pannu,E.A.Potterton,H.R.Powell, R. J. Read, A. Vagin,K.S.Wilson, Acta Crystallogr.Sect. DBiol. Crystallogr. 2011,67,235–242. [29] A. Vagin, A. Teplyakov, Acta Crystallogr.Sect. DBiol. Crystallogr. 2010,66, 22–25. [30] P. Emsley,K.Cowtan, Acta Crystallogr.Sect. DBiol. Crystallogr. 2004,60, 2126–2132. [31] G. N. Murshudov,A .A .V agin, E. J. Dodson, Acta Crystallogr.S ect. DB iol. Crystallogr. 1997,53,240–255. [32] S. McNicholas, E. Potterton, K. S. Wilson, M. E. Noble, Acta Crystallogr. Sect. DBiol. Crystallogr. 2011,67,386–394. Manuscript received:March 9, 2016 Accepted article published:April 28, 2016 Final article published:June 7, 2016 ChemBioChem 2016,17,1359 –1366 www.chembiochem.org Ó2016 The Authors. PublishedbyWiley-VCH Verlag GmbH &Co. KGaA, Weinheim1366 Full Papers