Comparative Analysis of the Conversion of Mandelonitrile and 2

molecules

Review Comparative Analysis of the Conversion of Mandelonitrile and 2-Phenylpropionitrile by a Large Set of Variants Generated from a Nitrilase Originating from Pseudomonas ﬂuorescens EBC191

Andreas Stolz 1,*, Erik Eppinger 1, Olga Sosedov 1,2 and Christoph Kiziak 1,3 1 Institut für Mikrobiologie, Universität Stuttgart, Allmandring 31, 70569 Stuttgart, Germany; [email protected] (E.E.); [email protected] (O.S.); [email protected] (C.K.) 2 BioChem Labor für biologische und chemische Analytik GmbH, Daimlerstr. 5B, 76185 Karlsruhe, Germany 3 Lonza AG, Rottenstr. 6, 3930 Visp, Switzerland * Correspondence: [email protected]; Tel.: +49-711-685-65489; Fax: +49-711-685-65725

Academic Editors: Ludmila Martínková and Margit Winkler Received: 12 August 2019; Accepted: 18 November 2019; Published: 21 November 2019

Abstract: The arylacetonitrilase from the bacterium Pseudomonas fluorescens EBC191 has been intensively studied as a model to understand the molecular basis for the substrate-, reaction-, and enantioselectivity of nitrilases. The nitrilase converts various aromatic and aliphatic nitriles to the corresponding acids and varying amounts of the corresponding amides. The enzyme has been analysed by site-specific mutagenesis and more than 50 different variants have been generated and analysed for the conversion of (R,S)-mandelonitrile and (R,S)-2-phenylpropionitrile. These comparative analyses demonstrated that single point mutations are sufficient to generate enzyme variants which hydrolyse (R,S)-mandelonitrile to (R)-mandelic acid with an enantiomeric excess (ee) of 91% or to (S)-mandelic acid with an ee-value of 47%. The conversion of (R,S)-2-phenylpropionitrile by different nitrilase variants resulted in the formation of either (S)- or (R)-2-phenylpropionic acid with ee-values up to about 80%. Furthermore, the amounts of amides that are produced from (R,S)-mandelonitrile and (R,S)-2-phenylpropionitrile could be changed by single point mutations between 2%–94% and <0.2%–73%, respectively. The present study attempted to collect and compare the results obtained during our previous work, and to obtain additional general information about the relationship of the amide forming capacity of nitrilases and the enantiomeric composition of the products.

Keywords: nitrilase; reaction speciﬁcity; enantioselectivity; structure-function relationship

1. Introduction Nitrilases are hydrolytic enzymes found in many bacteria, fungi, and plants that convert nitriles to the corresponding carboxylic acids and ammonia. They are traditionally assigned according to their substrate preference as aliphatic nitrilases, benzonitrilases, or arylacetonitrilases, although recent results challenge this classification [1,2]. There is a considerable interest in chemistry in the utilization of nitrilases for chemo-, regio- or enantioselective biotransformation reactions [3–6]. The application of nitrilases for enantioselective reactions has been intensively studied for the production of (R)-mandelic acid from (R,S)-mandelonitrile or benzaldehyde plus cyanide [7]. This reaction is catalysed by different arylacetonitrilases which are often obtained from the species Alcaligenes faecalis or organisms which are affiliated to the genus Pseudomonas and is already industrially realized in a multi-ton scale [8–14]. These and other arylacetonitrilases allow in principle also the enantioselective synthesis of other alpha-hydroxycarboxylic acids [15–17]. Furthermore, enantioselective reactions have

Molecules 2019, 24, 4232; doi:10.3390/molecules24234232 www.mdpi.com/journal/molecules Molecules 2019, 24, 4232 2 of 21 Molecules 2019, 24, 4232 2 of 21 Molecules 2019, 24, 4232 2 of 21 have also been observed with arylacetonitrilases carrying various other substituents in the alpha- have also been observed with arylacetonitrilases carrying various other substituents in the alpha- positionalso been (e.g., observed 2-amino-, with 2-halo-, arylacetonitrilases or 2-alkylnitriles) carrying [18–21]. various other substituents in the alpha-position position (e.g., 2-amino-, 2-halo-, or 2-alkylnitriles) [18–21]. (e.g.,Nitrilases 2-amino-, possess 2-halo-, a or catalytical 2-alkylnitriles) triad, [which18–21]. is composed of a cysteine, a glutamate, and a lysine Nitrilases possess a catalytical triad, which is composed of a cysteine, a glutamate, and a lysine residueNitrilases [22]. Moreover, possess a a catalytical second glutamate triad, which residu is composede seems also of a to cysteine, be important a glutamate, for the and enzymatic a lysine residue [22]. Moreover, a second glutamate residue seems also to be important for the enzymatic reactionresidue [[23].22]. In Moreover, the proposed a second catalytical glutamate triad the residue importance seems alsoof the to cysteine be important residue for has the been enzymatic proven reaction [23]. In the proposed catalytical triad the importance of the cysteine residue has been proven byreaction the detection [23]. In theof a proposed covalently catalytical bound reaction triad the in importancetermediate, of which the cysteine was identified residue hasas a beenthioimidate proven by the detection of a covalently bound reaction intermediate, which was identified as a thioimidate orby acylenzyme the detection [24], of a covalentlyand by the bound generation reaction of intermediate,inactive enzyme which variants was identiﬁed after the as exchange a thioimidate of the or or acylenzyme [24], and by the generation of inactive enzyme variants after the exchange of the relevantacylenzyme cysteine [24], andagainst by the other generation amino acid of inactive residues enzyme [25]. It variants is therefore after thegenerally exchange assumed of the relevantthat the relevant cysteine against other amino acid residues [25]. It is therefore generally assumed that the nitrilescysteine are against initially other bound amino to acid nitrilases residues via [25 a]. cysteine It is therefore residue generally in the form assumed of a that thioimidate, the nitriles then are nitriles are initially bound to nitrilases via a cysteine residue in the form of a thioimidate, then hydratedinitially bound to a tetrahedral to nitrilases intermediate via a cysteine and residue subsequently in the form deaminated of a thioimidate, to an acylenzyme, then hydrated which to is a hydrated to a tetrahedral intermediate and subsequently deaminated to an acylenzyme, which is finallytetrahedral hydrolysed intermediate to the and carboxylic subsequently acid (Figure deaminated 1). It to has an been acylenzyme, proposed which that isin ﬁnally this reaction hydrolysed the finally hydrolysed to the carboxylic acid (Figure 1). It has been proposed that in this reaction the glutamateto the carboxylic residue acidof the (Figure catalytical1). It triad has beenacts as proposed general base that catalyst in this reactionand that the lysine glutamate stabilizes residue the glutamate residue of the catalytical triad acts as general base catalyst and that the lysine stabilizes the tetrahedralof the catalytical intermediate triad acts [26]. as general base catalyst and that the lysine stabilizes the tetrahedral tetrahedral intermediate [26]. intermediate [26]. NH EnzSH NH2 3 NH H2O NH H2O O EnzSH H O NH2 3 O H O NH 2 O 2 O R C N R C R C OH R C R C R C N R C R C OH R C R C SEnz SEnz OH OH SEnz SEnz SEnz EnzSH SEnz EnzSH Figure 1. Proposed general reaction mechanism for nitrilases.nitrilases Figure 1. Proposed general reaction mechanism for nitrilases It has been repeatedly reported that some nitrilases form with certain substrates in addition to the acidsIt has also been the repeatedly correspondingcorresponding reported amides that [[27–33].some27–33 nitrilases]. Th Therefore,erefore, form an with extended certain reaction substrates scheme in addition has been to suggestedthe acids also to explain the corresponding this observation amides (Figure [27–33].2 2))[ [21,34].21 Th,34erefore,]. an extended reaction scheme has been suggested to explain this observation (Figure 2) [21,34].

Lys Lys Lys NLys+ NLys+ NLys H + H H + H H NH NH H NH H H H H H H O H O H H O R CN OH H –O Glu RN –O Glu R H O Glu R CN H O N H H H –O Glu RN –O Glu R N H O Glu H S S H OH O S O O H O S CysS O H CysS H O H Cys Cys H Cys H H Cys

I H2O I H+2 O~ H+ ~ Lys Lys A B Lys +NLys NLys A B +NLys H + H H H + H NH H NH H NH H H O – NH3 H O H H O H OHH H HN H H R O –O – NH3 R + H –O R –O Glu Glu O NH Glu N OH H H R O –O Glu R +N H –O Glu R O –O Glu S S O O S H O O H O O S H H CysS H O H CysS H O H Cys O Cys H H Cys H H Cys H H III IIa IIb III IIa IIb

– + R COO NH4 R CONH2 R COO– NH + R CONH 4 2 Figure 2. ProposedProposed reaction scheme for th thee conversion of nitriles to acids or amides by nitrilases [21]. [21]. Figure 2. Proposed reaction scheme for the conversion of nitriles to acids or amides by nitrilases [21]. This hypothetical reaction scheme suggests that the acids or the amides can be alternatively formedThis fromfrom hypothetical the the tetrahedral tetrahedral reaction intermediate intermediatescheme suggests in which in wh thethatich substrate the the acids substrate is or covalently the isamides covalently bound can to be thebound alternatively catalytically to the catalyticallyformedactive cysteine from active residuethe cysteinetetrahedral of the residue enzyme intermediate of (IIa, the IIb enzyme inin Figure wh (IIa,ich2). IIb Inthe thein substrate Figure proposed 2). isIn mechanism covalentlythe proposed the bound formation mechanism to the of thecatalytically amidesformation formally active of the cysteine occursamides by residueformally a thiol of elimination occurs the enzyme by a of th the(IIa,iol active elimination IIb in cysteine Figure of residue 2).the In active the and proposed cysteine the acids residuemechanism are formed and the formationacids are formed of the amides after the formally release occursof ammonia by a th fromiol elimination the tetrahedral of the intermediate. active cysteine The residue release and of the acids are formed after the release of ammonia from the tetrahedral intermediate. The release of

Molecules 2019, 24, 4232 3 of 21 Molecules 2019, 24, 4232 3 of 21 afterammonia the release (and ofthus ammonia the formation from the of tetrahedral the acids) intermediate. would chemically The release require of ammonia a protonation (and thus of the formationnitrogen atom of the originating acids) would from chemically the nitrile require group in a protonationorder to obtain of the a good nitrogen leaving atom group. originating This model from thegives nitrile a plausible group inexplanation order to obtain for the a goodobservation leaving that group. the Thisamounts model of givesamide a formed plausible from explanation a certain forsubstrate the observation strongly thatvary the among amounts different of amide nitrilases formed and from that a certain for a substrate single nitrilase strongly an vary increased among diformationﬀerent nitrilases of amides and thatcan forbe acaused single nitrilaseeither by an th increasede structure formation of the of substrate amides can (e.g., be causedby electron- either bywithdrawing the structure substituents of the substrate which destabilize (e.g., by electron-withdrawing the positive charge on substituents the nitrogen which atom) destabilize or by sterical the positivefactors which charge might on the hamper nitrogen the atom) transfer or by to sterical or the factorsstabilization which of might the positive hamper charge the transfer on the to amino or the stabilizationgroup [21]. of the positive charge on the amino group [21]. The explanationexplanation of theof reactionthe reaction mechanism mechanism and the and elucidation the elucidation of structure /offunction structure/function relationships ofrelationships nitrilases have of nitrilases been hampered have been for hampered several yearsfor several by the years inability by the to inability crystallize to crystallize native nitrilases. native Presumably,nitrilases. Presumably, this problem this is problem caused byis caused the peculiar by the characteristic peculiar characteristic of enzymatically of enzymatically active nitrilases active tonitrilases form asymmetric to form asymmetric spiral structures spiral structures [35]. To [35] solve. To these solve questions, these questions, two strategies two strategies have been have used.been Zhangused. Zhang et al. [36 et] obtainedal. [36] obtained crystals fromcrystals a C-terminally-deleted from a C-terminally-deleted variant of the variant nitrilase of fromthe nitrilaseSynechocystis from sp.Synechocystis PCC6803 andsp. PCC6803 subsequently and subsequently solved the structure. solved the In a stru secondcture. approach, In a second Sewell approach, et al. used Sewell electron et al. microscopyused electron to microscopy analyse the to structures analyse ofthe di stﬀructureserent nitrilases of different [26,37 nitrilases]. [26,37].

2. The Nitrilase from Pseudomonas fluorescensfluorescens EBC191EBC191 Pseudomonas fluorescensfluorescensEBC191 EBC191 was isolatedwas afterisolated selective after enrichment selective with enrichment 2-ethylbenzylcyanide with 2- (EBC)ethylbenzylcyanide as a sole source (EBC) of nitrogen as a sole andsource shown of nitrog to synthesizeen and shown a nitrilase to synthesize converting a nitrilase a broad converting range of nitrilesa broad [range38]. The of nitriles gene coding [38]. The for thegene nitrilase coding hadfor the been nitrilase cloned had and been a highly cloned effi andcient a recombinanthighly efficientE. colirecombinantstrain constructed. E. coli strain The constructed. amino acid The residues amino involvedacid residu ines catalysis involved were in catalysis identified were by identified sequence comparisonsby sequence comparisons with other members with othe ofr themembers nitrilase of superfamilythe nitrilase andsuperfamily mutagenesis and mutagenesis [20,23,39] and [20,23,39] a model ofand the a activemodel centreof the generatedactive centre by generated homology by modelling homology (Figure modelling3). (Figure 3).

Figure 3. ModelModel for the active centre of the nitrilase from P. ﬂuorescensfluorescens EBC191.EBC191.

The residuesresidues participating participating in thein catalyticthe catalytic triad (Glu48,triad (Glu48, Lys130, Cys164)Lys130, andCys164) the second and catalyticallythe second importantcatalytically glutamate important residue glutamate (Glu137) residue are shown(Glu137) in orange.are shown Amino in orange. acid residues Amino whichacid residues are located which at aare distance located of at lessa distance than 10 of Å less from than Cys164 10 Å arefrom shown Cys164 in accordanceare shown in with accordance their distance with their to Cys164 distance in decreasingto Cys164 in shades decreasing of blue. shades The model of blue. was The prepared model forwas the prepared present for study the based present on study the structure based on of the nitrilasestructure from of theSynechocystis nitrilase fromsp. Synechocystis PCC6803, PDB sp. PCC6803, 3WUY, using PDB the 3WUY, program using Yasara the program [40]. The Yasara resulting [40]. homologyThe resulting model homology was refined model by was the refined “md-refine by the macro” “md-refine using themacro” default using parameters. the default parameters. The purifiedpurified nitrilase nitrilase from fromP. fluorescensP. fluorescensEBC191 EBC191 forms forms like otherlike other nitrilases nitrilases spiral spiral structures structures which canwhich be visualizedcan be visualized by electron by el microscopyectron microscopy (Figure (Figure4). 4).

MoleculesMolecules 20192019,, 2424,, 4232 4232 4 4of of 21 21 Molecules 2019, 24, 4232 4 of 21

Figure 4. Electron microscopic image of the purified nitrilase from P. fluorescens EBC191 [41]. FigureFigure 4. 4. ElectronElectron microscopic microscopic image image of of the the purified purified nitrilase nitrilase from from P. fluorescensfluorescens EBC191 [[41].41]. Therefore, it will be necessary to include the structural constraints caused by the helix formation into theTherefore, structural it will model be necessary as soon to as include detailed the (cryo) st structuralructural EM dataconstraints for the caused nitrilase by fromthe helix P. fluorescensformation EBC191into thethe arestructuralstructural available. model model as as soon soon as detailedas detailed (cryo) (cryo) EM dataEM data for the for nitrilase the nitrilase from P. from fluorescens P. fluorescensEBC191 EBC191are available.The are wild-type available. nitrilase hydrolyses various phenylacetonitriles (e.g., mandelonitrile, 2- phenylpropionitrile,The wild-type wild-type nitrilasephenylglycinonitrile, nitrilase hydrolyses hydrolyses variousO-ac variousetylmandelonitrile) phenylacetonitriles phenylacetonitriles and (e.g., aliphatic (e.g., mandelonitrile, mandelonitrile, mono- and 2- dinitrilesphenylpropionitrile,2-phenylpropionitrile, to the corresponding phenylglycinonitrile, carboxylic acids O-acetylmandelonitrile)O-ac (Figureetylmandelonitrile) 5). and and aliphatic aliphatic mono- andmono- dinitriles and dinitrilesto the corresponding to the corresponding carboxylic carboxylic acids (Figure acids5). (Figure 5).

NH2 OH CH3

NH2 OH CH3 CN CN CN

CN CN CN

Mandelonitrile (MN) 2-Phenylpropionitrile (PPN) Phenylglycinonitrile Mandelonitrile (MN) 2-Phenylpropionitrile (PPN) Phenylglycinonitrile

OCO-CH3 OH

OCO-CH3 OH CN CN

CN CN

O-Acetylmandelonitrile 2-Hydroxy-2-phenylpropanenitrile O-Acetylmandelonitrile 2-Hydroxy-2-phenylpropanenitrile CN CN CN CN CN Cl CN NC Cl Valeronitrile 2-Chloropropionitrile NC Fumarodinitrile Valeronitrile 2-Chloropropionitrile Fumarodinitrile Figure 5. Examples for nitriles which are converted by the nitrilase from Pseudomonas ﬂuorescens EBC191. Figure 5. Examples for nitriles which are converted by the nitrilase from Pseudomonas fluorescens EBC191.FigureIn addition 5. Examples to the for acids, nitriles with which some are substrates converted also by the signiﬁcant nitrilase from amounts Pseudomonas of the correspondingfluorescens amidesEBC191. are formed [2,20,21,38,42–44]. Most of the tested chiral alpha-substituted nitriles were In addition to the acids, with some substrates also significant amounts of the corresponding converted by the wild-type nitrilase with a low degree of enantioselectivity. Thus, (R,S)-mandelonitrile amides are formed [2,20,21,38,42–44]. Most of the tested chiral alpha-substituted nitriles were (MN)In wasaddition hydrolysed to the acids, preferentially with some to substrates (R)-mandelic also acidsignificant ((R)-MA). amounts In contrast, of the corresponding the nitrilase convertedamides are by formed the wild-type[2,20,21,38,42–44]. nitrilase Mostwith ofa thlowe testeddegree chiral of enantioselectivity.alpha-substituted nitrilesThus, (R,Swere)- mandelonitrileconverted by (MN)the wild-type was hydrolysed nitrilase preferentially with a low to (degreeR)-mandelic of enantioselectivity. acid ((R)-MA). In Thus,contrast, (R,S the)- mandelonitrile (MN) was hydrolysed preferentially to (R)-mandelic acid ((R)-MA). In contrast, the

Molecules 2019, 24, 4232 5 of 21 preferentially formed (S)-2-phenylpropionic acid ((S)-2-PPA) from (R,S)-2-phenylpropionitrile ((R,S)-2-PPN) [20]. The enantioselectivity of the reactions has also been analysed for the hydrolysis of O-acetylmandelonitrile, phenylglycinonitrile (2-aminophenylacetonitrile), 2-chlorophenylacetonitrile, and acetophenone cyanohydrine and it was found that the enzyme formed preferentially the (R)-acids from mandelonitrile, O-acetylmandelonitrile, 2-chlorophenylacetonitrile, and acetophenone cyanohydrine. In contrast, the (S)-acids were produced in excess during the conversion of (R,S)-2-PPN and (R,S)-phenylglycinonitrile [20,21,38,44]. This suggested that electronegative substituents (-OH, -OCOCH3, -Cl) induce the preferred formation of the corresponding (R)-acids. In contrast, nitriles which carry in the α-position less electronegative or protonable substituents (-CH3, -NH2) are preferentially converted to the (S)-acids. The low degrees of enantioselectivity which were found during the conversion of (R,S)-MN and (R,S)-2-PPN (ee-values of 31% for (R)-mandelic acid and 65% for (S)-2-PPA at about 30% substrate conversion) and the ability of the enzyme to convert—although with a more pronounced enantioselectivity—both enantiomers of O-acetylmandelonitrile proved that the active site of the nitrilase from P. ﬂuorescens EBC 191 is in principle able to bind the (R)- and (S)-enantiomers of α-substituted nitriles (even with rather bulky substituents) [20,38]. The ability of the nitrilase to bind a wide range of substrates with diﬀerent stereochemistries at the α-position was also substantiated by the observation that the enzyme was able to convert sterically demanding α-,α-disubstituted phenylacetonitriles, such as 2-methyl-2-phenylpropionitrile or 2-hydroxyphenylpropionitrile (acetophenone cyanhydrine), which contain a quaternary carbon atom in the α-position towards the nitrile group [45].

2.1. Mutational Studies with the Nitrilase from P. fluorescens EBC191 Several variants of the nitrilase from P. fluorescens EBC191 have been generated in order to analyse the factors that determine the substrate-, reaction-, and enantiospecifity of nitrilases. In the course of these investigations, variants have been generated by introducing modifications close to the catalytical active cysteine residue, in the proposed substrate binding site, or by the formation of C-terminally deleted mutants. In addition, experiments were performed in order to increase the amides forming capacity of the enzyme by random and site-directed mutagenesis. Subsequently, these nitrilase variants were used for biotransformation reactions with mandelonitrile and 2-phenylpropionitrile as substrates and analysed for the enantioselectivities of the reactions and amide formation. These two substrates were converted by the purified wild-type enzyme with specific activities of 33 and 4.1 U/mg of protein, respectively [23,39,46–48]. Mandelonitrile and 2-phenylpropionitrile were used as model substrates, because they were converted by the wild-type enzyme with a different stereopreference (see above). These compounds show in aqueous solutions at neutral pH-values a rather different behaviour. While 2-phenylpropionitrile is completely stable under these conditions, mandelonitrile exists in equilibrium with benzaldehyde and HCN. This results in a pronounced tendency of mandelonitrile to racemize in neutral or alkaline aqueous solutions which leads in the presence of an enantioselective nitrilase to a dynamic kinetic resolution. This is industrially exploited for the almost stoichiometrical conversion of benzaldehyde plus cyanide to (R)-mandelic acid [8,9,12]. The analysis of the large set of nitrilase variants generated throughout the years, showed that the catalytical properties of the enzyme could be significantly modified by single point mutations [2,23, 39,45–48]. Thus, enzyme variants were found which hydrolysed (R,S)-mandelonitrile preferentially either to (R)- or (S)-mandelic acid [23,39,46–48]. In addition, nitrilase variants were identified which either formed almost exclusively mandelic acid or mandelamide [23,46,47]. In the present study, it was attempted to perform a comparative analysis of the variants to give an overview about the catalytical potential of a single nitrilase and its variants and to obtain additional information which might allow a better understanding of the features which govern the reaction- and enantiospecificity of nitrilases. Molecules 2019, 24, 4232 6 of 21

2.2. Mutations Which Result in the Increased Formation of (R)-Acids In order to compare the inﬂuence of diﬀerent mutations on the enantiomeric composition of the formed acids, the results were analysed that had been obtained previously for the conversion of (R,S)-MN and (R,S)-2-PPN. Therefore, for each enzyme variant, the ee-values that were obtained during the conversion of (R,S)-mandelonitrile were plotted against those values that were observed during the conversion of (R,S)-2-PPN (each after about 30% substrate conversion). These comparisons established that by single point mutations nitrilase variants could be generated which formed either (R)-mandelic acid with ee-values up to 91% or (S)-mandelic acid with an ee-value of 47%. Similarly, during the conversion of (R,S)-2-PPN, ee-values between 80% for (S)-2-PPA and 91% for (R)-2-PPA wereMolecules observed 2019, 24, 4232 (Figure 6). 7 of 21

Figure 6. Correlation of the enantiomeric composition of the acids formed from (R,S)-mandelonitrile Figureand (R, S6.)-2-phenylpropionitrile Correlation of the enantiomeric by diﬀerent comp nitrilaseosition variants. of the Theacids indicated formed valuesfrom (R,S were)-mandelonitrile obtained from andthe experimental(R,S)-2-phenylpropionitrile data presented inby thedifferent articles nitrilase of Kiziak variants. et al. (2007) The [indicated39], Kiziak values and Stolz were (2009) obtained [46], fromSosedov the etexperimental al. (2010) [47 data], Baum presented et al. (2012) in the [articles45], Sosedov of Kiziak and et Stolz al. (2007) (2015) [39], [23], Kiziak and includes and Stolz also (2009) some [46],unpublished Sosedov results. et al. (2010) The respective[47], Baum reaction et al. (2012) mixtures [45], containedSosedov and initially Stolz 10 (2015) mM of[23], (R ,andS)-mandelonitrile includes also someor (R, Sunpublished)-2-PPN. The results. ee-values The were respective determined reaction after aboutmixtures 30% contained substrate turn-over.initially 10 mM of (R,S)- mandelonitrile or (R,S)-2-PPN. The ee-values were determined after about 30% substrate turn-over.

A bulky tryptophan residue at the position corresponding to Trp188 seems to be highly conserved among nitrilases and seems generally to be required to impede the binding of (S)- mandelonitrile (or promote the binding of (R)-mandelonitrile). This can also be deduced from a previous study by Robertson et al. (2004) [19] who described the existence of nitrilases that show some preference for the formation of (S)-mandelic acid from (R,S)-mandelonitrile. In this study it was shown that among 48 different mandelonitrile hydrolyzing nitrilases obtained by a metagenomic approach only four were (S)-selective and formed (S)-mandelic acid with ee-values up to 30%. Sequence alignments constituted evidence that in these nitrilases the usually highly conserved relevant tryptophan residue is replaced by a tyrosine or a threonine residue [23].

2.4. Correlation Between the Relative Proportions of Amides Formed from (R,S)-Mandelonitrile or (R,S)-2- Phenylpropionitrile

Molecules 2019, 24, 4232 7 of 21

The comparisons suggested that four groups of nitrilase variants could be defined. One group encompasses nitrilase variants which form (R)-MA and (R)-2-PPA with extraordinary high enantioselectivities (Figure6, group A). These comparisons confirmed that only those nitrilase variants that formed (R)-mandelic acid with ee-values >75% formed preferentially (R)-2-PPA (Figure6). This attribute inevitably correlated with the presence of comparably large amino acid residue in the amino acid residue 165 in the direct neighbourhood of the catalytical active cysteine residue (towards the C-terminus). This resulted in the generation of enzyme variants that formed (R)-mandelic acid with ee-values >90% (compared to an ee-value of about 31% for the wild-type enzyme under the same conditions). The observed interdependence of a pronounced selectivity for the formation of (R)-MA from (R,S)-MN and the presence of a large amino acid residue in direct neighbourhood to the catalytical active cysteine residue nicely correlates to results obtained with other nitrilases. Thus, in the nitrilase from Alcaligenes faecalis ATCC8750 (which forms (R)-MA with a high enantiomeric surplus [8,9]) a tryptophan residue is present at the homologous position. The second group of nitrilase variants (Figure6, group B) showed compared to the wild-type enzyme a significantly decreased enantioselectivity for the formation of (S)-2-PPA and a (slight) tendency for an increased formation of (R)-mandelic acid. In this group mainly variants cluster which carry mutations in the C-terminal region or in Tyr 54. The largest group of variants (Figure6, group C) behaved similar to the wild-type enzyme and hydrolysed (R,S)-mandelonitrile with a low degree of enantioselectivity preferentially to (R)-mandelic acid and (R,S)-2-PPN preferentially to (S)-2-PPA. These variants showed only a low degree of variance regarding the formation of (S)-2-PPA (ee-values of about 50%–80%), but more pronounced differences for the formation of (R)-MA (ee-values of about 10%–70%). The nitrilase variants belonging to these three groups generally exhibited with (R,S)-mandelonitrile a more pronounced tendency towards the formation of the (R)-acids than with (R,S)-2-PPN. Thus, almost all variants belonging to the group A, produced (R)-MA with a higher ee-value than (R)-2-PPA. This effect was even more pronounced for the variants clustering in groups B and C which formed preferentially (R)-MA, but (S)-2-PPA (Figure6).

2.3. Mutations Which Result in the Increased Formation of (S)-Mandelic Acid In the course of the mutational studies also a group of variants had been generated which converted (R,S)-MN preferentially to (S)-MA (Figure6, group D). In all these variants, the tryptophan residue Trp188 was exchanged against smaller amino acid residue. This group of variants diﬀered from the other enzyme variants also in other aspects and most of these enzymes formed from (R,S)-mandelonitrile more mandelamide than mandelic acid (see below). A bulky tryptophan residue at the position corresponding to Trp188 seems to be highly conserved among nitrilases and seems generally to be required to impede the binding of (S)-mandelonitrile (or promote the binding of (R)-mandelonitrile). This can also be deduced from a previous study by Robertson et al. (2004) [19] who described the existence of nitrilases that show some preference for the formation of (S)-mandelic acid from (R,S)-mandelonitrile. In this study it was shown that among 48 diﬀerent mandelonitrile hydrolyzing nitrilases obtained by a metagenomic approach only four were (S)-selective and formed (S)-mandelic acid with ee-values up to 30%. Sequence alignments constituted evidence that in these nitrilases the usually highly conserved relevant tryptophan residue is replaced by a tyrosine or a threonine residue [23].

2.4. Correlation Between the Relative Proportions of Amides Formed from (R,S)-Mandelonitrile or (R,S)-2-Phenylpropionitrile Previously, it had been described that different nitrilases demonstrate various degrees of amide forming capacity and it was found for the nitrilase from P. fluorescens EBC 191 and also for a nitrilase Molecules 2019, 24, 4232 8 of 21 from the plant Arabidopsis thaliana that in general larger amounts of amides were formed from nitriles that carried in the α-position electronegative substituents, e.g., halide- or hydroxy-groups [21,31]. The wild-type nitrilase from P. fluorescens EBC191 formed from (R,S)-MN about 17% mandelamide (MAA), but from (R,S)-2-PPN only about 0.2% of the corresponding amide 2-phenylpropionamide (2-PPAA) [20]. To further analyse the correlation between amide formation from (R,S)-MN and (R,S)-2-PPN, the available variants of the nitrilase from P. fluorescens EBC191 were compared for the respective reactions. This proved that by single point mutations the amounts of mandelamide formed from (R,S)-MN could be changed between <1% and 94%. The respective values for the conversion of (R,S)-2-PPN varied between <0.1% and 73% 2-PPAA (Figure7). The individual enzyme variants formed generally more mandelamide from (R,S)-mandelonitrile than 2-PPAA from (R,S)-2-PPN. This confirmed the assumption that electron-withdrawing substituents (as the hydroxyl-group in mandelonitrile) induceMolecules an increased2019, 24, 4232formation of amides. 9 of 21

100 65 64 69 90

59, 63 68 60 80 58 61 62 71

67 72 70

70 39 60 36 10 MAA/(MAA+MA) [%] MAA/(MAA+MA) 15 9 50 14 12 35 34 11 40 40 16 73 13 51 75 30 74 49 31 76 47 20 50 48 53 4654 41 10 43 42 32 29 333828 30 37 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

2-PPAA/(2-PPAA+2-PPA) [%]

(9)DelC67 (10)DelC60 (11)DelC55 (12)DelC47 (13)H296K (14)H296A (15)H296F (16)H296R (20)w.t. (28)A165W (29)A165Y (30)A165F (31)A165G (32)A165H (33)A165E (34)A165R (35)C163N (36)C163Q (37)C163A (38)C163S (39)C163N + T110I (40)T110F (41)l111F (42)l113F (43)A116F (44)A116C (45)I117F (46)S190A (47)F191A (48)S192A (49)Y194A (50)A49T (51)P47G (52)R195A (53)E166A (54)H167A (58)W188A (59)W188l (60)w188L (61)W188V (62)W188F (63)W188Y (64)W188R (65)W188K (67)W188S (68)W188T (69)W188P (70)W188G (71)W188C (72)W188D (73)W188E (74)W188M (75)W188H (76)W188N (77)W188Q FigureFigure 7. 7. CorrelationCorrelation between between the theamount amountss of amides of amidesformed from formed (R,S)-mandelonitrile from (R,S)-mandelonitrile and (R,S)-2- and (R,Sphenylpropionitrile)-2-phenylpropionitrile by different by di nitrilaseﬀerent variants. nitrilase The variants. data were The taken data from were previous taken publications from previous publications(see Figure (see 6). Figure6).

Molecules 2019, 24, 4232 9 of 21

Almost all enzyme variants that produced between 0%–20% mandelamide from (R,S)-mandelonitrile resembled the wild-type enzyme and formed only rather low amounts (<1%) of 2-PPAA from (R,S)-2-PPN (Figure7). The comparison of the individual variants also showed that in those variants which formed more than 50% of mandelamide from (R,S)-mandelonitrile a significant change between the degrees of amide formation can occur. Thus, the wild-type nitrilase of P. fluorescens formed about 17% of mandelamide from (R,S)-mandelonitrile, but only about 0.2% of 2-PPAA from (R,S)-2-PPN. This corresponds to an 85-fold higher rate for the formation of mandelamide than 2-PPAA. The most efficient amide forming mutant derived from this enzyme by a single mutagenic event (Trp188Lys) formed 94% of mandelamide from (R,S)-mandelonitrile and about 60% of 2-PPAA from (R,S)-2-PPN which corresponds only to a 1.6-fold higher rate for the formation of mandelamide compared to 2-PPAA. Several types of mutations could be identified which resulted in an increased tendency of the variants to form amides (Figure7). The most pronounced increases in amide formation were observed when the tryptophan residue Trp188 was exchanged against a smaller residue and it was found that three of these mutants (Trp188Arg, Trp188Lys, Trp188Pro) formed more than 90% of mandelamide from (R,S)-MN groups [23]. An increased formation of mandelamide was also found with nitrilase variants that carried deletions of approximately 40–60 amino acids at the C-terminus or in the histidine residue His296, which is located close to the C-terminus. (The full-length enzyme consists of 350 amino acid residues.) Up to 50% of mandelamide were also formed by variants which possessed NH2-functions (as present in asparagine, glutamine, or arginine) close to the catalytical centre (e.g., the enzyme variants Cys163Gln or Cys163Asn) [39,46]. Furthermore, also the exchange of a threonine residue in position 110 against an isoleucine or phenylalanine or the replacement of an asparagine residue in position 206 by isoleucine-, lysine-, tyrosine-, or aspartate-residues increased the amide formation [39,46,47]. In addition, also the presence of a tyrosine residue in position 54 (as present in the wild-type nitrilase from P. fluorescens EBC191) results in a stronger tendency to form amides compared to the presence of almost all other possible amino acids at this position [45].

2.5. Correlation Between the Relative Amounts of Amides Formed from (R,S)-Mandelonitrile or (R,S)-2-Phenylpropionitrile and the Enantiomeric Composition of the Amides The wild-type nitrilase from P. fluorescens converted (R,S)-mandelonitrile and (R,S)-2-PPN each to opposite enantiomers of the respective acids and amides. Thus, (R)-mandelic acid and (S)-mandelamide and (S)-2-PPA and (R)-2-PPAA were preferentially formed from the racemic nitriles [20,21]. Furthermore, it was observed that the nitrilase transformed the two enantiomers of mandelonitrile (and also of O-acetylmandelonitrile) to considerably different amounts of the corresponding amides as side-products. Thus, (S)-mandelonitrile was converted to almost equimolar concentrations of (S)-mandelamide and (S)-mandelic acid, but (R)-mandelonitrile was transformed to (R)-mandelic acid and (R)-mandelamide in a ratio of about 85:15 [21,43,44]. To gain additional information, two plots were produced which related the relative amounts of amides formed with the enantiomeric composition of the amides (Figures8 and9). The graphical representation of the data for the formation of mandelamide from (R,S)-MN demonstrated that the enantiomeric composition of the amides was largely independent from the relative amounts of amides formed. Thus, variants were compared which produced between 2%–94% of mandelamide. Nevertheless, almost consistently (S)-mandelamide was preferentially formed with ee-values of 40%–85% (Figure8). These experiments established that the mutations resulting in the formation of increased amounts of mandelamide from (R,S)-mandelonitrile did only marginally influence the enantioselectivity. Molecules 2019, 24, 4232 10 of 21

To gain additional information, two plots were produced which related the relative amounts of amides formed with the enantiomeric composition of the amides (Figures 8 and 9). The graphical representation of the data for the formation of mandelamide from (R,S)-MN demonstrated that the enantiomeric composition of the amides was largely independent from the relative amounts of amides formed. Thus, variants were compared which produced between 2%–94% of mandelamide. Nevertheless, almost consistently (S)-mandelamide was preferentially formed with ee-values of 40%– 85% (Figure 8). These experiments established that the mutations resulting in the formation of increased amounts of mandelamide from (R,S)-mandelonitrile did only marginally influence the enantioselectivity. Molecules 2019, 24, 4232 10 of 21

32 0 33 )-amide R

-25

53 39 62 63 34 59 77 74 -50 41 51 37 60 65 73 9 72 71 -61 68 64 75 76 10 67 38 4035 70 58 48 47 69 43 16 11 12 14 )-amide ee[%] ( ee[%] )-amide 54 49 50 13 15 36

S -75

( 46

31 52 29 42 w.t.

-100 0 102030405060708090100 MAA/(MA+MAA)[%]

(9)DelC67 (10)DelC60 (11)DelC55 (12)DelC47 (13)H296K (14)H296A (15)H296F (16)H296R (20)w.t. (29)A165Y (31)A165G (32)A165H (33)A165E (34)A165R (35)C163N (36)C163Q (37)C163A (38)C163S (39)C163N + T110I (40)T110F (41)l111F (42)l113F (43)A116F (44)A116C (45)I117F (46)S190A (47)F191A (48)S192A (49)Y194A (50)A49T (51)P47G (52)R195A (53)E166A (54)H167A (76)W188N (58)W188A (60)w188L (59)W188l (61)W188V (62)W188F (63)W188Y (64)W188R (65)W188K (67)W188S (69)W188P (68)W188T (70)W188G (71)W188C (72)W188D (73)W188E (74)W188M (75)W188H (76)W188N (77)W188Q Figure 8. Relationship between the amounts of mandelamide formed from (R,S)-mandelonitrile and Figure 8. Relationship between the amounts of mandelamide formed from (R,S)-mandelonitrile and the enantiomeric composition of the formed amide. The data were taken from previous publications the enantiomeric composition of the formed amide. The data were taken from previous publications (see Figure6). In order to exclude errors from inaccurate measurements only those reactions were (see Figure 6). In order to exclude errors from inaccurate measurements only those reactions were analysed in which more than 2% of mandelamide were formed from (R,S)-mandelonitrile. analysed in which more than 2% of mandelamide were formed from (R,S)-mandelonitrile. In contrast to the conversion of mandelonitrile by the variants, significant differences in the enantiomericIn contrast composition to the conversion of the productsof mandelonitrile formed were by the observed variants, when significant the formation differences of 2-PPAA in the enantiomericfrom (R,S)-2-PPN composition was compared. of the products Thus, those formed variants were which observed formed when as the the wild-type formation enzyme of 2-PPAA only fromminor (R amounts,S)-2-PPN of was 2-PPAA compared. (<3%) Thus, preferentially those variants synthesized which formed (R)-2-PPAA as the with wild-type different enzyme degrees only of minorenantioselectivity amounts of (Figure2-PPAA9, group(<3%) A).preferentially synthesized (R)-2-PPAA with different degrees of enantioselectivityThe second group (Figure (Figure 9, group9, group A). B) encompassed variants which formed 3%–17% of 2-PPAA and showed compared to the wild-type enzyme a slightly enhanced enantioselectivity for the formation of (R)-2-PPAA with ee-values of about 80%. The third group (Figure9, group C) assembled those variants in which the tryptophan residue in position 188 had been replaced by a smaller amino acid residue. These variants all formed preferentially

(S)-2-PPAA. Furthermore, it was found that the large majority of these variants synthesized (S)-2-PPAA with ee-values of about 80%, but that the amounts of 2-PPAA formed diﬀered largely among the variants.

2.6. Comparison of the Enantiomeric Composition of the Acids and Amides Formed During the Turn-Over of (R,S)-Mandelonitrile or (R,S)-2Phenylpropionitrile As indicated above, the wild-type nitrilase from P. ﬂuorescens EBC191 preferentially converted (R,S)-mandelonitrile to (R)-mandelic acid and (S)-mandelamide and (R,S)-2-PPN to (S)-2-PPA and (R)-2-PPAA [20,21]. In the following, it was analysed if this was a general trait of the available nitrilase Molecules 2019, 24, 4232 11 of 21 variants. Therefore, for the nitrilase variants the enantiomeric excesses of the acids and amides producedMolecules 2019 were, 24, correlated4232 to each other (Figures 10 and 11). 11 of 21

Figure 9. Relationship between the amounts of 2-phenylpropionamide formed from Figure 9. Relationship between the amounts of 2-phenylpropionamide formed from (R,S)-2- (R,S)-2-phenylpropionitrile and the enantiomeric composition of the amide. The data were taken from phenylpropionitrile and the enantiomeric composition of the amide. The data were taken from previous publications (see Figure6). previous publications (see Figure 6).

The second group (Figure 9, group B) encompassed variants which formed 3%–17% of 2-PPAA and showed compared to the wild-type enzyme a slightly enhanced enantioselectivity for the formation of (R)-2-PPAA with ee-values of about 80%. The third group (Figure 9, group C) assembled those variants in which the tryptophan residue in position 188 had been replaced by a smaller amino acid residue. These variants all formed preferentially (S)-2-PPAA. Furthermore, it was found that the large majority of these variants synthesized (S)-2-PPAA with ee-values of about 80%, but that the amounts of 2-PPAA formed differed largely among the variants.

Molecules 2019, 24, 4232 12 of 21 variants. Therefore, for the nitrilase variants the enantiomeric excesses of the acids and amides produced were correlated to each other (Figures 10 and 11). The plot for the turn-over of (R,S)-mandelonitrile indicated that the acids and amides forming reactions did not show any direct correlation in their enantioselectivities. Almost all enzyme variants that formed (R)-mandelic acid with ee-value of 10%–90% or (S)-mandelic acid with ee-values up to 50% formed (S)-mandelamide with ee-values between 40%–95%, but without any recognizable correlation between the enantioselectivities of both reactions (Figure 10). The only exception from this rule were two mutants (Ala165His and Ala165Glu) which exhibited an increase in the enantioselectivity towards the formation of (R)-mandelic acid in combination with a slightly decreased formation of (S)-mandelamide (Figure 10). Molecules 2019, 24, 4232 12 of 21

25 )-amide R

0 32 33

-25

62 53 39 63 59 77 60 51 34 65 61 -50 37 72 64 41 9 76 68 10

73 71 )-amide ee[%] ( 38 35 47 40 69 S 48 70 ( 54 43 49 16 12 14 50 1315 -75 46 31 42 20 29 30 -100 -75 -50 -25 0 25 50 75 100 (S)-MA ee [%] (R)-MA

(9)DelC67 (10)DelC60 (11)DelC55 (12)DelC47 (13)H296K (14)H296A (15)H296F (16)H296R (20)w.t. (29)A165Y (31)A165G (32)A165H (33)A165E (34)A165R (35)C163N (36)C163Q (37)C163A (38)C163S (39)C163N + T110I (40)T110F (41)l111F (42)l113F (43)A116F (44)A116C (45)I117F (46)S190A (47)F191A (48)S192A (49)Y194A (50)A49T (51)P47G (52)R195A (53)E166A (54)H167A (58)W188A (59)W188l (60)w188L (61)W188V (62)W188F (63)W188Y (64)W188R (65)W188K (67)W188S (68)W188T (69)W188P (70)W188G (71)W188C (72)W188D (73)W188E (74)W188M (75)W188H (76)W188N (77)W188Q Figure 10. Correlation of the enantiomeric composition of the amides and acids formed from Figure 10. Correlation of the enantiomeric composition of the amides and acids formed from (R,S)- (R,S)-mandelonitrile by various nitrilase variants. The data were taken from previous publications mandelonitrile by various nitrilase variants. The data were taken from previous publications (see (see Figure6). But those variants were excluded from the analysis which formed less than 2% of Figure 6). But those variants were excluded from the analysis which formed less than 2% of mandelamide (in relation to the amounts of mandelic acid formed). mandelamide (in relation to the amounts of mandelic acid formed). The plot for the turn-over of (R,S)-mandelonitrile indicated that the acids and amides forming The corresponding plot for the conversion of 2-PPN indicated that also with this substrate the reactions did not show any direct correlation in their enantioselectivities. Almost all enzyme variants enantiomeric composition of the formed amide was to a large extent independent of the that formed (R)-mandelic acid with ee-value of 10%–90% or (S)-mandelic acid with ee-values up enantioselectivity of the acid formation. Furthermore, the plot suggested that the variants could be to 50% formed (S)-mandelamide with ee-values between 40%–95%, but without any recognizable separated into two groups (Figure 11). correlation between the enantioselectivities of both reactions (Figure 10). The only exception from this rule were two mutants (Ala165His and Ala165Glu) which exhibited an increase in the enantioselectivity towards the formation of (R)-mandelic acid in combination with a slightly decreased formation of (S)-mandelamide (Figure 10). The corresponding plot for the conversion of 2-PPN indicated that also with this substrate the enantiomeric composition of the formed amide was to a large extent independent of the enantioselectivity of the acid formation. Furthermore, the plot suggested that the variants could be separated into two groups (Figure 11). One group (group A) consisted of the variants which preferentially formed (R)-2-PPA and (R)-2-PPAA with a high degree of selectivity. These variants formed (R)-2-PPAA with ee-values >75%. The second group (group B) of variants (which also encompassed the wild-type enzyme) formed (S)-2-PPA with ee-values of about 50%–75%, but formed (S)- or (R)-(2)-PPAA with ee-values varying between about 80% for (S)- or (R)-2-PPAA. Thus, it appears that with (R,S)-2-PPN the enantiopreference for the formation of the amide could be easily modiﬁed by mutations and that the enantiopreference for the formation of the acid is rather conserved. This is in contrast to the situation observed during the conversion of (R,S)-mandelonitrile (see Figure 10). Molecules 2019, 24, 4232 13 of 21 Molecules 2019, 24, 4232 13 of 21

Figure 11. Correlation of the ee-values of the amides and acids formed from (R,S)-2-phenylpropionitrile byFigure various 11. nitrilase Correlation variant. of The the data ee-values were taken of fromthe previousamides publicationsand acids (seeformed Figure from6), but (R,S those)-2- variantsphenylpropionitrile were excluded by various from the nitrila analysisse variant. which The formed data lesswere than taken 1% from of 2-PPAA previous (in publications relation to (see the amountsFigure 6), of but 2-PPA those formed). variants were excluded from the analysis which formed less than 1% of 2-PPAA (in relation to the amounts of 2-PPA formed). 2.7. Comparison of the Relative Amounts of Mandelamide Formed and the Enantiomeric Excess of the Mandelic Acid ProducedOne group During (group the A) Turn-Over consisted of (R,S)-Mandelonitrileof the variants which preferentially formed (R)-2-PPA and (R)- 2-PPAAThe wild-typewith a high nitrilases degree fromof selectivity.P. fluorescens TheseEBC191 variants and formedA. faecalis (R)-2-PPAAATCC8750 with demonstrated ee-values >75%. with (TheR,S)-mandelonitrile second group (group as a substrate B) of variants very clear(which di ffalsoerences encompassed with respect theto wild-type enantioselectivity enzyme) formed and amide (S)- formation.2-PPA with Thus,ee-values the nitrilaseof about from50%–75%,A. faecalis but formedshowed (S a)- high or (R degree)-(2)-PPAA of enantioselectivity with ee-values varying for the formationbetween about of (R)-mandelic 80% for acid(S)- andor produced(R)-2-PPAA. almost Thus, no mandelamide.it appears that In contrast, with (R for,S)-2-PPN the nitrilase the fromenantiopreferenceP. fluorescens afor low the degree formation of enantioselectivity of the amide coul andd be a easily pronounced modified formation by mutations of mandelamide and that the wereenantiopreference reported [39]. for It was the thereforeformation analysed of the acid if this is rather correlation conserved. was also This valid is in for contrast the di fftoerent the situation nitrilase variantsobserved from duringP. fluorescens the conversionEBC191. of ( Therefore,R,S)-mandelonitrile a diagram (see was Figure generated 10). which correlated the relative amounts of mandelamide with the enantiomeric composition of (R)-mandelic acid formed. Surprisingly, 2.7. Comparison of the Relative Amounts of Mandelamide Formed and the Enantiomeric Excess of the this representation of the data suggested that for most enzyme variants increased enantioselectivity Mandelic Acid Produced During the Turn-Over of (R,S)-Mandelonitrile for the formation of (R)-mandelic acid correlated with an increase in the relative amounts of the producedThe wild-type mandelamide nitrilases (Figure from 12 P., groupfluorescens A). InEBC191 addition, and A. a smallfaecalis group ATCC8750 of enzyme demonstrated variants with was (R,S)-mandelonitrile as a substrate very clear differences with respect to enantioselectivity and amide formation. Thus, the nitrilase from A. faecalis showed a high degree of enantioselectivity for the

Molecules 2019, 24, 4232 14 of 21

formation of (R)-mandelic acid and produced almost no mandelamide. In contrast, for the nitrilase from P. fluorescens a low degree of enantioselectivity and a pronounced formation of mandelamide were reported [39]. It was therefore analysed if this correlation was also valid for the different nitrilase variants from P. fluorescens EBC191. Therefore, a diagram was generated which correlated the relative amounts of mandelamide with the enantiomeric composition of (R)-mandelic acid formed. Surprisingly, this representation of the data suggested that for most enzyme variants increased enantioselectivity for the formation of (R)-mandelic acid correlated with an increase in the relative amounts of the produced mandelamide (Figure 12, group A). In addition, a small group of enzyme variants was identified (Figure 12, group B) which were highly enantioselective for the formation of (R)-mandelic acid, but which formed only very low amounts of mandelamide. The nitrilase from A. faecalis is the “archetypical” representative of this small group of nitrilases. These variants all carry (suchMolecules as2019 the, 24nitrilase, 4232 from A. faecalis) a large substituent in direct neighbourhood to the catalytical14 of 21 active cysteine residue and were already identified in the plot shown in Figure 6 because of their extraordinary enantioselectivity for the formation of (R)-mandelic acid. identiﬁed (Figure 12, group B) which were highly enantioselective for the formation of (R)-mandelic The graphic also suggested that the variants carrying an exchange in Trp188 (and thus forming acid, but which formed only very low amounts of mandelamide. The nitrilase from A. faecalis is the preferentially (S)-mandelic acid and (S)-mandelamide) could be separated into two groups with “archetypical” representative of this small group of nitrilases. These variants all carry (such as the different tendencies to form mandelamide. Thus, many of the variants carrying polar amino acids in nitrilase from A. faecalis) a large substituent in direct neighbourhood to the catalytical active cysteine position 188 (e.g., Glu, Gln, Asn, His) synthesized only low amounts of mandelamide (Figure 12, residue and were already identiﬁed in the plot shown in Figure6 because of their extraordinary group E) compared to all other Trp188X variants (Figure 12, group D). enantioselectivity for the formation of (R)-mandelic acid.

90 59 63 68 60 61 80 62 71

67 72 70

70 39 D 60 36 10 15 9 50 12 14 35 34 11 40 16 40 73 13 MAA/(MAA+MA) [%] MAA/(MAA+MA) 75 51 74 30 49 31 A 76 47 20 50 48 53 E 54 46 77 10 41 43 C 42 32 B29 38 28 33 30 0 37 -50 -25 0 25 50 75 100 S)-MA ee [%] (R)-MA (9)DelC67 (10)DelC60 (11)DelC55 (12)DelC47 (13)H296K (14)H296A (15)H296F (16)H296R (20)w.t. (28)A165W (29)A165Y (30)A165F (31)A165G (32)A165H (33)A165E (34)A165R (35)C163N (36)C163Q (37)C163A (38)C163S (39)C163N + T110I (40)T110F (41)l111F (42)l113F (43)A116F (44)A116C (45)I117F (46)S190A (47)F191A (48)S192A (49)Y194A (50)A49T (51)P47G (52)R195A (53)E166A (54)H167A (59)W188l (60)w188L (61)W188V (63)W188Y (62)W188F (63)W188Y (64)W188R (64)W188R (65)W188K (67)W188S (68)W188T (69)W188P (70)W188G (71)W188C (72)W188D (73)W188E (74)W188M (75)W188H (76)W188N (77)W188Q Figure 12. Correlation between the relative amounts of mandelamide formed and the ee-values of the mandelic acid produced during the conversion of (R,S)-mandelonitrile. The data were taken from previous publications (see Figure6).

The graphic also suggested that the variants carrying an exchange in Trp188 (and thus forming preferentially (S)-mandelic acid and (S)-mandelamide) could be separated into two groups with diﬀerent tendencies to form mandelamide. Thus, many of the variants carrying polar amino acids in position 188 (e.g., Glu, Gln, Asn, His) synthesized only low amounts of mandelamide (Figure 12, group E) compared to all other Trp188X variants (Figure 12, group D). Molecules 2019, 24, 4232 15 of 21

2.8. Comparison of the Reaction- and Enantiospecificity of the Nitrilase Variants During the Conversion of (R,S)-Mandelonitrile The nitrilase variants converted (R,S)-mandelonitrile to different amounts of (R)- or (S)-mandelic acid and (R)- or (S)-mandelamide. In the course of our investigations no indications for any enantioconversion were observed during the formation of the acids or amides. Therefore, for the final comparisons it was attempted to define a parameter, which reflected the total enantiomeric composition of the products formed in a single numerical term. The concentrations of the (S)-enantiomers of mandelic acid and mandelamide at the relevant point of conversion (about 30% substrate turn-over, see above) were added up and the respective (R)-enantiomers were treated in the same way, as exemplified in Table1. These values should in principle reflect the enantiodiscrimination of the nitrilase variants during the initial binding of the nitriles.

Table 1. Examples for the calculation of the “total enantiomeric excess” observed during the conversion of (R,S)-mandelonitrile to mandelamide and mandelic acid.

“Total MAA ee (R)-MA (S)-MA (R)-MAA (S)-MAA Nitrilase ee MA (R)/(S) Enantiomeric (%) MAA (%) (%) (%) (%) Excess” Wild 16 84 (S) 31 (R) 55 29 1 15 56/44 12 (R) type Del C60 52 59 (S) 73 (R) 41 7 11 41 52/48 4 (R) A165H 6 1 (R) 91 (R) 90 4 3 3 93/7 86 (R) W188K 94 57 (S) 47 (S) 2 4 20 74 22/78 56 (S) MA: Mandelic acid, MAA: Mandelamide. The data for the amounts of mandelamide and mandelic acid formed and the enantiocomposition of the products were taken from Kiziak et al. (2007) [39], Kiziak and Stolz (2009) [46], and Sosedov and Stolz (2015) [23]. The relative concentrations of the respective enantiomers were calculated from these data.

Subsequently, the relative amounts of mandelamide formed from (R,S)-mandelonitrile were plotted against the enantiomeric excesses of the products (=sum of the concentrations of the (S)-enantiomers subtracted from the sum of the concentrations of the (R)-enantiomers divided by the sum of the concentrations of the (S)- plus (R)-enantiomers). This type of plot confirmed the results that the nitrilase variants could be differentiated in three major groups (Figure 13). The comparisons shown in Figures 12 and 13 established that the integration of the enantiomeric composition of the amides into the calculations resulted for most enzyme variants (those combined in group A in Figure 12) in significantly lower values for the (R)-selectivity. This was caused by the fact that for these enzyme variants the increased ee-values for (R)-mandelic acid correlated with an increased amide formation and that the preferentially formed enantiomer of the amide possessed an opposite stereochemistry compared to the acid. This resulted in the chosen graphical presentation of the data in the clustering of most enzyme variants in a group of enzymes which exhibited in-summa during the turn-over of (R,S)-mandelonitrile only a low preference for the formation of the (R)-enantiomers of the two relevant products (acid and amide). In most of these enzyme variants the alanine residue in position 165 of the nitrilase from P. fluorescens was conserved (with the exceptions of Nit(A165G) and Nit(A165R)). Thus, it appears that this group of enzymes accepts (S)- and (R)-mandelonitrile as substrates without a significant preference for one of the substrate enantiomers. In combination with the observations that these variants preferentially form the (R)-acid together with the (S)-amide and the previous demonstration that nitrilases intermediately form a covalent adduct between the C-atom originating from the nitrile moiety and the catalytical active cysteine residue (Figure1) this suggests that there is a difference in the probability of the way how the tetrahedral intermediate is resolved. Thus, it appears that the enzyme bound (R)-form can be easier deaminated than the enzyme bound (S)-form. This results finally in the observed preference for the formation of (R)-mandelic acid and (S)-mandelamide. The graphical presentation shown in Figure 13 did not significantly change the position of those enzymes that were already identified in Figure 12 as outstanding (groups B and C), Molecules 2019, 24, 4232 16 of 21 because the members of these groups combined the enantioselective formation of (R)-mandelic acid Moleculeswith a rather2019, 24, low 4232 degree of amide formation or formed (S)-mandelic acid plus (S)-mandelamide.16 of 21

100 65 69 64 90 68 59 63 58 61 71 62 80

72 70

70 39 60 10 15 50 9 12 34 11 40 40 16 73 13 75 51 74 30 49 31 20 MAA/(MAA+MA) (%) MAA/(MAA+MA) 76 20 50 47 52 48 77 53 41 43 42 54 10 32 37 38 28 303329 0 -80 -60 -40 -20 0 20 40 60 80 100

(S)-enantiomers "total enantiomeric excess" (R)-enantiomers of MA plus MAA

(9)DelC67 (10)DelC60 (11)DelC55 (12)DelC47 (13)H296K (14)H296A

(15)H296F (16)H296R (20)w.t. (28)A165W (29)A165Y (30)A165F

(31)A165G (32)A165H (33)A165E (34)A165R (53)E166A (54)H167A

(37)C163A (38)C163S (39)C163N + T110I (40)T110F (41)l111F (42)l113F

(43)A116F (44)A116C (45)I117F (46)S190A (47)F191A (48)S192A

(49)Y194A (52)R195A (50)A49T (51)P47G (57)Y54F (58)W188A

(59)W188l (60)w188L (61)W188V (62)W188F (63)W188Y (64)W188R

(65)W188K (67)W188S (68)W188T (69)W188P (70)W188G (71)W188C

(72)W188D (73)W188E (74)W188M (75)W188H (76)W188N (77)W188Q

Figure 13. Correlation of the relative amounts of mandelamide formed and the “total enantiomeric excess” of the products formed during the turn-over of (R,S)-mandelonitrile. The data were taken Figurefrom previous13. Correlation publications of the (seerelative Figure amounts6) and theof mand calculationselamide performed formed and as describedthe “total inenantiomeric the text and excess”exempliﬁed of the in products Table1. formed during the turn-over of (R,S)-mandelonitrile. The data were taken from previous publications (see Figure 6) and the calculations performed as described in the text and exemplifiedThe group in of Table enzymes, 1. which combined the highly enantioselective formation of (R)-mandelic acid with a low degree of amide formation (group B in Figure 12) comprised the enzyme variants Nit(A165W),The comparisons Nit(A165Y), shown Nit(A165F), in Figures Nit(A165H), 12 and 13 establ and Nit(A165E).ished that the These integration enzymes of possessed the enantiomeric aromatic compositionamino acids of or otherthe amides bulky into substituents the calculations in the position, resulted which for most is (C-terminally) enzyme variants in direct (those neighbourhood combined into group the catalytical A in Figure active 12) cysteinein significantly residue. lower This values type of for substituents the (R)-selectivity. occurs toThis accomplish was caused two by types the factof functions,that for these an enhancementenzyme variants in enantioselectivity the increased ee-values and a signiﬁcantfor (R)-mandelic reduction acid incorrelated amide formation. with an increasedFurthermore, amide it canformation be deduced and that that the this preferential type of enzymely formed binds enantiomer (R)-mandelonitrile of the amide with possessed a pronounced an oppositepreference stereochemistry and that the binding compared or conversionto the acid. ofThis (S )-mandelonitrileresulted in the chosen is impeded graphical by the presentation presence ofof a thebulky data substituent in the clustering in direct of neighbourhoodmost enzyme va toriants the catalytically in a group of active enzymes cysteine which residue. exhibited This in-summa could also duringexplain the the turn-over increased amountsof (R,S)-mandelonitrile of the (R)-enantiomers only a oflow the preference amide formed for the and formation the decreased of the amounts (R)- enantiomers of the two relevant products (acid and amide). In most of these enzyme variants the alanine residue in position 165 of the nitrilase from P. fluorescens was conserved (with the exceptions of Nit(A165G) and Nit(A165R)). Thus, it appears that this group of enzymes accepts (S)- and (R)- mandelonitrile as substrates without a significant preference for one of the substrate enantiomers. In

Molecules 2019, 24, 4232 17 of 21 of the totally formed amides (as the lower amounts of enzyme bound (S)-enantiomers will also decrease amide formation because of the reduced tendency of the enzyme to convert (R)-mandelonitrile to (R)-mandelamide). The group of variants carrying a Trp188X mutation was also in the plot shown in Figure 13 clearly separated from all other variants (group C). Furthermore, a comparison of Figure 12 with Figure 13 proved that these variants indeed were the only group that clearly showed a preference for the conversion of (S)-mandelonitrile, as both the (S)-enantiomers of mandelamide and mandelic acid were formed.

2.9. Comparison of the Reaction- and Enantiospecificity of the Nitrilase Variants During the Conversion of (R,S)-2-Phenylpropionitrile In the previous paragraph, the conversion of mandelonitrile by the enzyme variants was analysed. In the following, it was attempted to perform the same comparisons for the turn-over of (R,S)-2-PPN. The interpretation of the relevant data obtained during the conversion of (R,S)-2-PPN was much more difficult than the previous analysis because with 2-PPN as substrate in general much lower amounts of amides were formed. This complicated especially the chiral analysis and thus the interpretation of the experimental results. Furthermore, the nitrilase variants differed much stronger in the enantioselectivity of 2-PPA-formation than observed for the formation of mandelic acid from (R,S)-mandelonitrile. The investigated nitrilase variants could be differentiated in four groups in respect to the relationship of the “total enantiomeric excess” and the amounts of 2-PPAA formed (Figure 14). The largest group of variants exhibited similar “total enantiomeric excesses” for the formation of the (S)-enantiomers as the wild-type nitrilase from P. fluorescens EBC191 (ee-values of about 50%–80%) and formed only low amounts of 2-PPAA. The enzymes of the second group showed almost no enantioselectivity. Members of this group were i.a. the four C-terminally deleted enzyme variants studied, the variants carrying the point mutations Nit(His296Lys) und Nit(His296Arg) within the C-terminus and a mutant carrying an exchange in the amino acid residue threonine- 110 (Nit(Thr110Phe)). In the third group all enzyme variants were combined that showed a strong bias for the formation of (R)-2-PPA and formed only very low amounts of amide. This group consisted of the variants Nit(Ala165Trp), Nit(Ala165Tyr), Nit(Alas165Phe), Nit(Ala165His), and Nit(Ala165Glu), and Nit(Ala165Arg). Thus, this group was identical to the group of enzyme variants that preferred (R)-mandelonitrile as substrate and did form only rudimentary amounts of mandelamide. This suggested that the enlargement of the amino acid residue in position 165 had a similar effect on the hydrolysis of (R,S)-2-PPN as in the hydrolysis of (R,S)-mandelonitrile and reduced for both substrates the binding or conversion of the (S)-enantiomers of the nitriles. The fourth group assembled those variants in which Trp188 was exchanged against a smaller amino acid residue. Surprisingly, some of these variants showed during the conversion of (R,S)-2-PPN compared to the wild-type a somehow reduced “total enantiomeric excess” for the formation of the (S)-enantiomers. This differed significantly from the situation found for the conversion of (R,S)-mandelonitrile (Figure 13). Molecules 2019, 24, 4232 18 of 21 mandelonitrile as substrate and did form only rudimentary amounts of mandelamide. This suggested that the enlargement of the amino acid residue in position 165 had a similar effect on the hydrolysis of (R,S)-2-PPN as in the hydrolysis of (R,S)-mandelonitrile and reduced for both substrates the binding or conversion of the (S)-enantiomers of the nitriles. The fourth group assembled those variants in which Trp188 was exchanged against a smaller amino acid residue. Surprisingly, some of these variants showed during the conversion of (R,S)-2-PPN compared to the wild-type a somehow reduced “total enantiomeric excess” for the formation of the (S)-enantiomers. This differed Molecules 2019, 24, 4232 18 of 21 significantly from the situation found for the conversion of (R,S)-mandelonitrile (Figure 13).

Figure 14. Correlation between the relative amounts of 2-phenylpropionamide formed and the “total Figureenantiomeric 14. Correlation excess” of between the products the relative formed amounts during theof 2-phenylpropionamide conversion of (R,S)-2-phenylpropionitrile. formed and the “total The enantiomericdata were taken excess” from of the the previous products publications formed during (see Figure the conversion6). of (R,S)-2-phenylpropionitrile. The data were taken from the previous publications (see Figure 6). 3. Conclusions 3. ConclusionsThe comparative analysis of the numerous variants generated from the nitrilase of Pseudomonas fluorescens EBC191 throughout the years gives an impressive example of the pronounced effects which The comparative analysis of the numerous variants generated from the nitrilase of Pseudomonas single amino acid exchanges can have on the reaction specificity and enantioselectivity of enzymes. fluorescens EBC191 throughout the years gives an impressive example of the pronounced effects The mutational studies allowed the identification of two “hot spots” which determine the reaction- and which single amino acid exchanges can have on the reaction specificity and enantioselectivity of enantiospecificity of the nitrilase from P. fluorescens EBC191 (and probably also from other organisms) to enzymes. The mutational studies allowed the identification of two “hot spots” which determine the a large extent. Thus, the presence of a large aromatic residue in direct neighbourhood to the catalytical active cysteine residue (towards the C-terminus) determines nitrilases which show a pronounced reaction specificity for the formation of the acids and are highly (R)-selective for the formation of products such as (R)-mandelic acid or (R)-phenylglycine. There is good evidence that this substitution pattern strongly decreases the ability of the relevant (S)-nitriles to bind to the enzymes. Nitrilases with this amino acid arrangement are highly relevant for enantioselective biotransformation reactions targeting optical active α-substituted acids. Molecules 2019, 24, 4232 19 of 21

The second important amino acid residue which was identified in the course of our studies was Trp188. Homologous tryptophan residues are highly conserved among nitrilases. The exchange of this tryptophan residue against smaller amino acid residues resulted in nitrilase variants which showed a preference for the conversion of the (S)-enantiomers of mandelonitrile and 2-phenylpropionitrile combined with the formation of significant amounts of amides. This type of nitrilase variants might be useful for the (enantioselective) formation of α-substituted amides and could offer an alternative to true nitrile hydratases which often show limitation in respect to substrate specificity, enantioselectivity, and cyanide resistance. In summary, our studies resulted in two catalytically very different forms of a nitrilase, which either produced from racemic mandelonitrile almost stoichiometrically the (R)-acids or the (S)-amides in a large surplus. It is tempting to speculate that these two “extremes” might indicate the limits for nitrilase-catalyzed biotransformations.

Author Contributions: All authors contributed to the text and the figures. Funding: This research received no external funding. Acknowledgments: Parts of the described experimental work were supported by the DFG (Deutsche Forschungsgemeinschaft). Conflicts of Interest: There are no conflicts of interest.

References

1. Nagasawa, T.; Mauger, J.; Yamada, H. A novel nitrilase, arylacetonitrilase, of Alcaligenes faecalis JM3. Eur. J. Biochem. 1990, 194, 765–772. [CrossRef][PubMed] 2. Brunner, S.; Eppinger, E.; Fischer, S.; Gröning, J.; Stolz, A. Conversion of aliphatic nitriles by the arylacetonitrilase from Pseudomonas fluorescens EBC191. World J. Microbiol. Biotechnol. 2018, 34, 91. [CrossRef] 3. Martinková, L.; Kˇren, V. Nitrile- and amide-converting microbial enzymes: Stereo-, regio- and chemoselectivity. Biocatal. Biotrans. 2002, 20, 79–93. [CrossRef] 4. Gong, J.S.; Shi, J.S.; Lu, Z.M.; Li, H.; Zhou, Z.M.; Xu, Z.H. Nitrile-converting enzymes as a tool to improve biocatalysis in organic synthesis: Recent insights and promises. Crit. Rev. Biotechnol. 2017, 37, 69–81. [CrossRef][PubMed] 5. Martinková, L.; Rucká, L.; Nešvera, J.; Pátek, M. Recent advances and challenges in the heterologous production of microbial nitrilases for biocatalytic applications. World J. Microbiol. Biotechnol. 2017, 33, 8. [CrossRef][PubMed] 6. Bhalla, T.C.; Kumar, V.; Kumar, V.; Thakur, N.; Savitri. Nitrile metabolizing enzymes in biocatalysis and biotransformation. Appl. Biochem. Biotechnol. 2018, 185, 925–946. [CrossRef][PubMed] 7. Martinková, L.; Kˇren,V. Biocataytic production of mandelic acid and analogues: A review and comparison with chemical processes. Appl. Microbiol. Biotechnol. 2018, 102, 3893–3900. [CrossRef][PubMed] 8. Yamamoto, K.; Oishi, K.; Fujimatsu, I.; Komatsu, K.-I. Production of R-(-)-mandelic acid from mandelonitrile by Alcaligenes faecalis ATCC 8750. Appl. Environ. Microbiol. 1991, 57, 3028–3032. 9. Yamamoto, K.; Fujimatsu, I.; Komatsu, K.-I. Purification and characterization of the nitrilase from Alcaligenes faecalis ATCC 8750 responsible for enantioselective hydrolysis of mandelonitrile. J. Ferment. Bioeng. 1992, 73, 425–430. [CrossRef] 10. Ress-Löschke, M.; Friedrich, T.; Hauer, B.; Mattes, R. Verfahren zur Herstellung chiraler Carbonsäuren aus Nitrilen mit Hilfe einer Nitrilase oder Mikroorganismen, die ein Gen für eine Nitrilase enthalten. German Patent Application DE 19848129A1, 19 October 1998. 11. Kaul, P.; Banerjee, A.; Mayilraj, S.; Banerjee, U.C. Screening for enantioselective nitrilases: Kinetic resolution of racemic mandelonitrile to (R)-mandelic acid by new bacterial isolates. Tetrahedron Asymmetry 2004, 15, 207–211. [CrossRef] 12. Liese, A.; Seelbach, K.; Wandrey, C. Industrial Biotransformations, 2nd ed.; Wiley-VCH: Weinheim, Germany, 2006. 13. Banerjee, A.; Kaul, P.; Banerjee, U.C. Purification and characterization of an enantioselective arylacetonitrilase from Pseudomonas putida. Arch. Microbiol. 2006, 184, 407–418. [CrossRef][PubMed] Molecules 2019, 24, 4232 20 of 21

14. Zhang, X.-H.; Liu, Z.-Q.; Xue, Y.-P.; Zhang, Y.-G. Nitrilase catalyzed conversion of (R,S)-mandelonitrile by immobilized recombinant Escherichia coli cells harboring nitrilase. Biotechnol. Appl. Biochem. 2015.[CrossRef] 15. Zhang, C.-S.; Zhang, Z.J.; Li, C.-X.; Yu, H.-L.; Zheng, G.-W.; Xu, J.-H. Efficient production of (R)-o-chloromandelic acid by deracemization of o-chloromandelonitrile with a new nitrilase mined from Labrenzia aggregata. Appl. Microbiol. Biotechnol. 2012, 95, 91–99. [CrossRef] 16. Wang, H.; Sun, H.; Gao, W.; Wei, D. Efficient production of (r)-o-chloromandelic acid by recombinant Escherichia coli cells harboring nitrilase from Burkholderia cenocepacia J2315. Org. Process. Res. Dev. 2014, 18, 767–773. [CrossRef] 17. Sun, H.; Wang, H.; Gao, W.; Chen, L.; We, K.; Wei, D. Directed evolution of nitrilase PpL19 from Pseudomonas psychrotolerans and identification of enantiocomplementary mutants toward mandelonitrile. Biochem. Biophys. Res. Commun. 2015, 468, 820–825. [CrossRef][PubMed] 18. DeSantis, G.; Zhu, Z.; Greenberg, W.A.; Wong, K.; Chaplin, J.; Hanson, S.R.; Farwell, B.; Nicholson, L.W.; Rand, C.L.; Weiner, D.P.; et al. An enzyme library approach to biocatalysis: Development of nitrilases for enantioselective production of carboxylic acid derivatives. J. Am. Chem. Soc. 2002, 124, 9024–9025. [CrossRef] 19. Robertson, D.E.; Chaplin, J.A.; DeSantis, G.; Podar, M.; Madden, M.; Chi, E.; Richardson, T.; Milan, A.; Miller, M.; Weiner, D.P.; et al. Exploring nitrilase sequence space for enantioselective catalysis. Appl. Environ. Microbiol. 2004, 70, 2429–2436. [CrossRef] 20. Kiziak, C.; Conradt, D.; Stolz, A.; Mattes, R.; Klein, J. Nitrilase from Pseudomonas fluorescens EBC191: Cloning and heterologous expression of the gene and biochemical characterization of the recombinant enzyme. Microbiology 2005, 151, 3639–3648. [CrossRef] 21. Fernandes, B.C.M.; Mateo, C.; Kiziak, C.; Wacker, J.; Chmura, A.; van Rantwijk, F.; Stolz, A.; Sheldon, R.A. Nitrile hydratase activity of a recombinant nitrilase. Adv. Synth. Catal. 2006, 348, 2597–2603. [CrossRef] 22. Brenner, C. Catalysis in the nitrilase superfamily. Curr. Opinion Struct. Biol. 2002, 12, 775–782. [CrossRef] 23. Sosedov, O.; Stolz, A. Improvement of the amides forming capacity of the arylacetonitrilase from Pseudomonas fluorescens EBC191 by site-directed mutagenesis. Appl. Microbiol. Biotechnol. 2015, 99, 2623–2635. [CrossRef] [PubMed] 24. Stevenson, D.E.; Feng, R.; Storer, A.C. Detection of covalent enzyme-substrate complexes of nitrilases by ion-spray mass spectroscopy. FEBS Lett. 1990, 277, 112–114. [CrossRef] 25. Kobayashi, M.; Komeda, H.; Yanaka, N.; Nagasawa, T.; Yamada, H. Nitrilase from Rhodococcus rhodochrous J1. Sequencing and overexpression of the gene and identification of an essential cysteine residue. J. Biol. Chem. 1992, 267, 20746–20751. [PubMed] 26. Thuku, R.N.; Brady, D.; Benedik, M.J.; Sewell, B.T. Microbial nitrilases: Versatile, spiral forming, industrial enzymes. J. Appl. Microbiol. 2008, 106, 703–727. [CrossRef][PubMed] 27. Hook, R.H.; Robinson, W.G. Ricinine nitrilase: II. Purification and properties. J. Biol. Chem. 1964, 239, 4263–4267. [PubMed] 28. Robinson, W.G.; Hook, R.H. Ricinine nitrilase: I. Reaction product and substrate specificity. J. Biol. Chem. 1964, 239, 4257–4262. 29. Goldlust, A.; Bohak, Z. Induction, purification, and characterization of the nitrilase of Fusarium oxysporum f. sp. melonis. Biotechnol. Appl. Biochem. 1989, 11, 581–601. 30. Stevenson, D.E.; Feng, R.; Dumas, F.; Groleau, D.; Mihoc, A.; Storer, A.C. Mechanistic and structural studies on Rhodococcus ATCC39484 nitrilase. Biotechnol. Appl. Biochem. 1992, 15, 283–302. 31. Effenberger, F.; Osswald, S. Enantioselective hydrolysis of (RS)-2-fluoroarylacetonitriles using nitrilase from Arabidopsis thaliana. Tetrahedron Asymmetry 2001, 12, 279–285. [CrossRef] 32. Osswald, S.; Wajant, H.; Effenberger, F. Characterization and synthetic applications of recombinant AtNIT1 from Arabidopsis thaliana. Eur. J. Biochem. 2002, 269, 680–687. [CrossRef] 33. Piotrowski, M.; Schönfelder, S.; Weiler, E.W. The Arabidopsis thaliana isogene NIT4 and its orthologs in tobacco encode β-cyano-L-alanine hydratase/nitrilase. J. Biol. Chem. 2001, 276, 2616–2621. [CrossRef][PubMed] 34. Kobayashi, M.; Goda, M.; Shimizu, S. Nitrilase catalyzes amide hydrolysis as well as nitrile hydrolysis. Biochem. Biophys. Res. Comm. 1998, 253, 662–666. [CrossRef][PubMed] 35. Thuku, R.N.; Weber, B.W.; Varsani, A.; Sewell, B.T. Post-translational cleavage of recombinantly expressed nitrilase from Rhodococcus rhodochrous J1 yields a stable, active helical form. FEBS J. 2007, 274, 2099–2108. [CrossRef][PubMed] Molecules 2019, 24, 4232 21 of 21

36. Zhang, L.; Bin, Y.; Wang, C.; Jiang, S.; Wang, H.; Yuan, Y.A.; Wei, D. Structural insights into enzymatic activity and substrate specificity by a single amino acid in nitrilase from Synechocystis sp. PCC6803. J. Struct. Biol. 2014, 188, 93–101. [CrossRef][PubMed] 37. Woodward, J.D.; Trompetter, I.; Sewell, B.T.; Piotrowski, M. Substrate specificity of plant nitrilases complexes is affected by their helical twist. Commun. Biol. 2018, 1, 186. [CrossRef][PubMed] 38. Layh, N.; Stolz, A.; Förster, S.; Effenberger, F.; Knackmuss, H.-J. Enantioselective hydrolysis of O-acetylmandelonitrile to O-acetylmandelic acid by bacterial nitrilases. Arch. Microbiol. 1992, 158, 405–411. [CrossRef] 39. Kiziak, C.; Klein, J.; Stolz, A. Influence of different carboxyterminal mutations on the substrate-, reaction-, and enantiospecificity of the arylacetonitrilase from Pseudomonas fluorescens EBC191. Prot. Eng. Design Sel. 2007, 20, 385–396. [CrossRef] 40. Krieger, E.; Koraimann, G.; Vriend, G. Increasing the precision of comparative models with YASARA NOVA- a self-parameterizing force field. Proteins 2002, 47, 393–402. [CrossRef] 41. Sosedov, O. Steigerung der Nitril-Hydratase-Aktivität der Nitrilase aus Pseudomonas fluorescens EBC191. Ph.D. Thesis, Universität Stuttgart, Stuttgart, Germany, 2012. 42. Heinemann, U.; Kiziak, C.; Zibek, S.; Layh, N.; Schmidt, M.; Griengl, H.; Stolz, A. Conversion of aliphatic 2-acetoxynitriles by nitriles hydrolysing bacteria. Appl. Microbiol. Biotechnol. 2003, 63, 274–281. [CrossRef] 43. Mateo, C.; Chmura, A.; Rustler, S.; van Rantwijk, F.; Stolz, A.; Sheldon, R.A. Synthesis of enantiomerically pure (S)-mandelic acid using an oxynitrilase-nitrilase bienzymatic cascade. A nitrilase surprsisingly shows nitrile hydratase activity. Tetrahedron Asymmetry 2006, 17, 320–323. [CrossRef] 44. Rustler, S.; Müller, A.; Windeisen, V.; Chmura, A.; Fernandes, B.C.M.; Kiziak, C.; Stolz, A. Conversion of mandelonitrile and phenylglycinenitrile by recombinant E. coli cells synthesizing a nitrilase from Pseudomonas fluorescens EBC191. Enzyme Microb. Technol. 2007, 40, 598–606. [CrossRef] 45. Baum, S.; Williamson, D.S.; Sewell, T.; Stolz, A. Conversion of sterically demanding α-α-disubstituted phenylacetonitriles by the arylacetonitrilase from Pseudomonas fluorescens EBC191. Appl. Environ. Microbiol. 2012, 78, 48–57. [CrossRef][PubMed] 46. Kiziak, C.; Stolz, A. Identification of amino acid residues which are responsible for the enantioselectivity and amide formation capacity of the arylacetonitrilase from Pseudomonas fluorescens EBC 191. Appl. Environ. Microbiol. 2009, 75, 5592–5599. [CrossRef][PubMed] 47. Sosedov, O.; Baum, S.; Bürger, S.; Matzer, K.; Kiziak, C.; Stolz, A. Construction and application of variants of the arylacetonitrilase from Pseudomonas fluorescens EBC191 which form increased amounts of acids or amides. Appl. Environ. Microbiol. 2010, 76, 3668–3674. [CrossRef][PubMed] 48. Sosedov, O.; Stolz, A. Random mutagenesis of the arylacetonitrilase from Pseudomonas fluorescens EBC191 and identification of variants which form increased amounts of mandelamide from mandelonitrile. Appl. Microbiol. Biotechnol. 2014, 98, 1595–1607. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).