Highly enantioselective kinetic resolution of two ... - Semantic Scholar

Protein Engineering, Design & Selection vol. 20 no. 3 pp. 125– 131, 2007 Published online February 19, 2007 doi:10.1093/protein/gzm003

Highly enantioselective kinetic resolution of two tertiary alcohols using mutants of an esterase from Bacillus subtilis Birgit Heinze1,3, Robert Kourist1,3, Linda Fransson2, Karl Hult2 and Uwe T. Bornscheuer1,4 1

Institute of Biochemistry, Department of Biotechnology and Enzyme Catalysis, Greifswald University, Felix-Hausdorff-Str. 4, D-17487 Greifswald, Germany and 2Royal Institute of Technology (KTH), School of Biotechnology, Department of Biochemistry, AlbaNova University Center, 10691 Stockholm, Sweden

4

To whom correspondence should be addressed. E-mail: [email protected] These authors contributed equally to this work.

3

Enzyme-catalyzed kinetic resolutions of secondary alcohols are a standard procedure today and several lipases and esterases have been described to show high activity and enantioselectivity. In contrast, tertiary alcohols and their esters are accepted only by a few biocatalysts. Only lipases and esterases with a conserved GGG(A)X-motif are active, but show low activity combined with low enantioselectivity in the hydrolysis of tertiary alcohol esters. We show in this work that the problematic autohydrolysis of certain compounds can be overcome by medium and substrate engineering. Thus, 3-phenylbut-1-yn-3-yl acetate was hydrolyzed by the esterase from Bacillus subtilis (BS2, mutant Gly105Ala) with an enantioselectivity of E 5 56 in the presence of 20% (v/v) DMSO compared to E 5 28 without a cosolvent. Molecular modeling was used to study the interactions between BS2 and tertiary alcohol esters in their transition state in the active site of the enzyme. Guided by molecular modeling, enzyme variants with highly increased enantioselectivity were created. For example, a Glu188Asp mutant converted the trifluoromethyl analog of 3-phenylbut-1-yn-3-yl acetate with an excellent enantioselectivity (E > 100) yielding the (S)-alcohol with >99%ee. In summary, protein engineering combined with medium and substrate engineering afforded tertiary alcohols of very high enantiomeric purity. Keywords: tertiary alcohol/enantioselectivity/enzyme catalysis/esterase/rational protein design

esters of tertiary alcohols. These are generally not accepted as substrates by almost any carboxylester hydrolases of commercial interest, probably due to the sterically demanding structure of those compounds (O’Hagan and Zaidi, 1992, 1994; Schlacher et al., 1998). We recently discovered that the only hydrolases which are active towards this class of substrates are characterized by a highly conserved GGG(A)X-motif (Henke et al., 2002), which is located in the active site and contributes to the formation of the so-called oxyanion hole (Pleiss et al., 2000). This binding pocket stabilizes the oxyanion in the tetrahedral intermediate formed during the catalytic cycle of ester hydrolysis (Bryan et al., 1986; Whiting and Peticolas, 1994; Ordentlich et al., 1998). The two groups, GGG(A)X- and GX-hydrolases, differ significantly in the structure of the catalytic site (Pleiss et al., 2000). However, a detailed study using three model compounds showed that the enantioselectivity of the GGG(A)Xhydrolases is very low and did not exceed an E-value of 3 in the kinetic resolution of 3-phenylbut-1-yn-3-yl acetate (1a, Figure 1) using an esterase from Bacillus subtilis (BS2). On the basis of a computer modeling, two mutants were predicted (Gly105Ala, Ala400Leu), of which the Gly105Ala variant led to an increase in enantioselectivity to E ¼ 19 (Henke et al., 2003). Although this represents a 6-fold increase in enantioselectivity, this value is still not synthetically useful to obtain both enantiomers in optically pure form. Alternatively, transesterification of a racemic tertiary alcohol using vinyl acetate as acyl donor in organic solvents was performed using lipase A from Candida antarctica. Autohydrolysis of 3-phenylbut-1-yn-3-yl acetate could be avoided and an E ¼ 65 could be reached, but the reaction suffered from very long reaction times and a maximum conversion of 35% (Krishna et al., 2002).

Materials and methods

General Introduction The synthesis of optically pure tertiary alcohol is still a challenge, not only in biocatalysis but also in classical stereoselective synthesis (Corey and Guzman-Perez, 1998; Christoffers and Mann, 2001; Christoffers and Baro, 2005). A promising route to an optically pure compound is via hydrolase-catalyzed kinetic resolution of chiral tertiary alcohols. While carboxylester hydrolases are widely used for the synthesis of optically pure secondary alcohols and to a smaller extent also for the resolution of primary alcohols and carboxylic acids (Bornscheuer and Kazlauskas, 2005), there are only few examples of their utilization in the hydrolysis of

All chemicals were purchased from Fluka (Buchs, Switzerland), Sigma (Steinheim, Germany) and Merck (Darmstadt, Germany), unless stated otherwise. Restriction enzymes and polymerases were obtained from New England Biolabs GmbH (Beverly, MA, USA) and Roboklon GmbH (Berlin, Germany). MWG-Biotech (Ebersberg, Germany) provided primers and performed sequence analysis. NMR spectroscopy experiments were performed on an ARX300 device (Bruker, Karlsruhe, Germany).

Synthesis of tertiary alcohols (R, S)-3-Phenylbut-1-yn-2-ol (1b) was purchased from Ubichem plc (Easthleigh, UK). (R, S)-4,4,4-Trifluoro-3phenylbut-1-yn-2-ol (2b) was synthesized as described

# The Author 2007. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: [email protected]

125

B.Heinze et al.

Fig. 1. BS2 (wild-type or mutants) catalyzed kinetic resolution of 3-phenylbut-1-yn-3-yl acetate (1a) and 4,4,4-trifluoro-3-phenylbut-1-yn-3-yl acetate (2a) performed in buffer and/or in the presence of DMSO. Due to the higher priority of the trifluoromethyl group, the assignment of the absolute configuration of 1 and 2 is inverted.

(O’Hagan and Zaidi, 1992). A solution of 2,2,2-trifluoroacetophenone (5.05 g, 29 mmol) in tetrahydrofurane (20 ml) (THF) was added to a refluxing solution of ethynyl magnesium bromide (1.2 eq.) in THF (70 ml). The mixture was heated under reflux until all of the ketone was consumed (control via thin-layer chromatography). Water (20 ml) was added to the cooled reaction mixture and the product was extracted into diethyl ether (3 100 ml), washed with saturated brine (100 ml) and evaporated under reduced pressure. The residue was distilled to give the alcohol as a colorless oil (4.4 g, 22 mmol, 76%). The NMR data matched the literature data (O’Hagan and Zaidi, 1992).

Synthesis of tertiary alcohol acetates Acetyl chloride (1.2 ml, 15 mmol, 3 eq.) was added dropwise to a solution of the appropriate alcohol (5 mmol) in pyridine (6 ml) while the flask was cooled on ice and the mixture was stirred for another 15 min. Then the ice bath was removed and the mixture was stirred at ambient temperature for 16 h. The solution was extracted twice with diethyl ether (100 ml). The organic layer was washed twice with saturated NaHCO3 solution and three times with a 5% aqueous solution of NaHSO4. The organic layer was dried over anhydrous Na2SO4 before the solvent was removed under reduced pressure. (R, S)-3-Phenylbut-1-yn-2-yl acetate (1a): chromatography on silica gel (hexane : dichloromethane 1 : 1) yielded the product as a colorless liquid (510 mg, 2.7 mmol, 54%). The NMR data matched the literature data (O’Hagan and Zaidi, 1992). (R, S)-4,4,4-Trifluoro-3-phenylbut-1-yn2-yl acetate (2a): chromatography on silica gel (hexane: dichloromethane 1:1) yielded the product as a colorless liquid (550 mg, 2.3 mmol, 46%). The NMR data matched the literature data (O’Hagan and Zaidi, 1992).

Expression and protein analysis of BS2 Escherichia coli DH5a and JM109 were used as hosts for the transformation of plasmid DNA. The strains were grown in LB liquid media and on LB agar plates supplemented with 100 mg ml21 ampicillin at 378C (Sambrook et al., 1989). The vector pGaston with a rhamnose-inducible promotor was used for the expression of BS2 (Henke et al., 2003). Cells were grown to an optical density of 0.5 at 600 nm. Then esterase production was induced upon addition of rhamnose (final concentration 0.2% v/v) and cultivation continued for 5 h. Cells were collected by centrifugation (15 min, 48C, 4000 g) and washed twice with sodium phosphate buffer (100 mM, pH 7.5, 48C). Cells were disrupted by sonication on ice for 10 min at 50% pulse and centrifuged to separate soluble from insoluble fractions. In order to produce 126

a larger amount of enzyme for use in preparative scale, cells were disrupted using a French press (French Pressure cell press, Thermo Spectronic, USA). The supernatant was aliquoted and stored at 2208C. His-tag purification was performed following the standard protocol by using BD TalonTM IMAC resin from Clontech, Laboratories Inc., Mountain View, USA. Elution was done by pH-shift to pH 5. Protein content was determined using Bradford reagent with bovine serum albumin as standard. Esterase activity was determined spectrophotometrically by hydrolysis of p-nitrophenol acetate ( pNPA, 1 mM) in sodium phosphate buffer (10 mM, pH 7.5, 10% (v/v) DMSO). p-Nitrophenol released was quantified at 410 nm (1 ¼ 15 103 M21 cm21). One unit (U) of activity was defined as the amount of enzyme releasing 1 mmol p-nitrophenol per min under assay conditions (Krebsfa¨nger et al., 1998). Proteins were also analyzed by a 12% separating and 4% stacking sodium dodecyl sulphate polyacrylamide gel (Sambrook et al., 1989). After electrophoresis, the gel was first activity-stained (Krebsfa¨nger et al., 1998) with a-naphthyl acetate and Fast RedTM followed by Coomassie brilliant blue staining.

Site-directed mutagenesis For site-directed mutagenesis, complementary primers bearing the nucleotides to be changed and the vector pGaston encoding for the wild-type gene of BS2 were used for PCR with Pfu-DNA-polymerase and the following reaction conditions: (i) 958C, 30 s and (ii) 18 cycles: 958C, 50 s; 558C, 60 s; 688C, 10 min. The PCR mixture was treated afterwards with DpnI to digest methylated template DNA. The mutated plasmids were transformed into competent E. coli cells prepared by the transformation and storage solution method, and transformed according to a standard protocol (Sambrook et al., 1989). The mutants were confirmed by sequencing. The following primers were used: Ala107Trp: 50 GTGGATTCACGGAGGCTGGTTTTATCTAGGAGCGGG-30 Ala400Trp: 50 GCCGCCGTACAATAAAGCGTTTCACTGGTTAGAGCTTCC-30 , Glu188Asp: 50 GTAACAGTATTTGGAGATTCCGCCGGCGGGATGAG-30 , Glu188Phe: 50 CGTAACAGTATTTGGATTCTCCGCCGGCGGGATGAG-30 , Glu188Gln: 50 CGTAACAGTATTTGGACAGTCCGCCGGCGGGATGAG-30 .

General method for esterase-catalyzed small-scale resolutions To a stirred solution of substrate (25 mM) in phosphate buffer (100 mM, pH 7.5) and the appropriate amount of cosolvent, the esterase solution was added. The reaction mixture was stirred in a thermoshaker (Eppendorf, Hamburg, Germany) at 378C for 20 min. The reaction mixture was then extracted twice with 400 ml dichloromethane. The combined organic

Esterase design for a tertiary alcohol

layers were dried over anhydrous sodium sulphate and the organic solvent was removed under nitrogen. The samples were analyzed by gas chromatography. 1b was analyzed using a chiral Hydrodexw-b-3B (Heptakis-(2,6-di-O-methyl3-O-pentyl)-b-cyclodextrin; 25 m, 0.25 mm) column, and 1a was analyzed using a chiral heptakis-(2,3-diO-acetyl-6-O-t-butyldimethylsilyl)-cyclodextrin column ( provided by Prof. Ko¨nig, University of Hamburg, Germany), both installed in a GC-14A gas chromatograph (Shimadzu, Tokyo, Japan) with hydrogen as the carrier gas. 2a and 2b were analyzed using the Hydrodexw-b-3B column in a GC-MS (model QP-2010, Shimadzu, Tokyo, Japan). Enantioselectivity and conversion were calculated according to Chen et al., (1982).

Esterase-catalyzed kinetic resolution in preparative scale Esterase solution (12 ml, 108 U) was added to a stirred solution of 2a (90 mg, 0.37 mmol) in DMSO (4 ml) and phosphate buffer (4 ml, 100 mM, pH 7.5). The reaction mixture was stirred for 15 min at 308C. The reaction mixture was then extracted twice with 10 ml dichloromethane. The organic layers were combined and washed with brine (2 10 ml), water (2 10 ml) and dried over anhydrous NaHSO4. The solvent was removed under reduced pressure. Chromatography on silica gel (hexane:dichloromethane 2:1) yielded 2b (28 mg, 0.14 mmol, 38%, ee ¼ 94%) and 2a (20 mg,) as colorless liquids (20 mg, 0.08 mmol, 22%, ee . 99%).

Molecular modeling Molecular modeling was performed using the SYBYL molecular modeling package version 7.0 (Tripos Inc.) on an SGI Octane UNIX workstation. The enzyme-substrate system was described by using the Kollman All Atom force field (Weiner et al., 1984, 1986) for the purpose of energyminimization calculations and MD simulations. The Kollman All Atom types for the tetrahedral intermediate were assigned as follows: aromatic carbon, CA; tetrahedral carbon, CT; hydrogen, HC; oxyanion oxygen, OH; and nonoxyanion oxygen, OS. A new Kollman All Atom type for the carbon atoms of the ethinyl group was defined as CX. Non-standard partial charges were calculated by using the Pullman method (Berthod and Pullman, 1965; Berthod et al., 1967) with a formal charge of 21 for the substrate oxyanion. Non-standard Kollman All Atom force field parameters were assigned in analogy to the existing parameters for vdW radius, bond length, angle and torsion angle bending as ˚ , mass 12.01 g mole21; follows: (i) CX, vdW radius 1.7 A (ii) OH – CT – OS; (iii) CA– CT – OS; (iv) CX – CT– OS; (v) CX – CX – HC; (vi) OS – CT –OS, for (ii) – (vi): 109.58, force constant 50 kJ mole21 degree22; (vii) CA – CT– CX 109.58; (viii) CT – CT – CX, for (vii) – (viii): 109.58, force constant 63 kJ mole21 degree22; (ix) CT – CX – CX 1808, force con˚ , force constant 125 kJ mole21 degree22; (x) CX – CX 1.2 A 21 ˚ 22 ˚ stant 650 kJ mole A ; (xi) CT– CX 1.5 A, force constant ˚ 22; ˚, 300 kJ mole21 A (xii) 1.08 A force constant 21 ˚ 22 34 kJ mole A ; (xiii) CA – CT– CX –CX; (xiv) OS – CT – CX – CX, (xv) CT – CT – CX – CX, (xvi) CT – CX – CX – HC, for (xiii) – (xvi): 08, periodicity 0, force constant 0 kJ mole21 degree22. ˚ and a distanceA non-bonded cut off distance of 8 A dependent dielectric function with a scaling factor of 1 were

used in all calculations. An NTV ensemble (i.e., constant number of atoms, temperature and volume), a temperature of 300 K after a warm-up phase and a time step of 1 fs were used in all MD simulations. Energy minimizations were performed by using the Powell method (Powell, 1977). All crystal structure coordinates were obtained from the Protein Data Bank (Berman et al., 2000).

Preparation of the free enzyme The preparation of the free enzyme was done similar to the procedure described for C. antarctica Lipase B (Raza et al., 2001). The crystal structure of the highly homologous B. subtilis p-Nitrobenzyl Esterase (Schmidt et al., 2007) (PDB code 1QE3, Spiller et al., 1999) was used as the starting point for the modeled tetrahedral intermediate. Hydrogen atoms were added to the enzyme. The positions of the water hydrogen atoms and then the enzyme hydrogen atoms were optimized by using a consecutive series of short (1 ps) MD runs and energy minimizations. This series of optimization steps was repeated until the energy of the system was stable. Thereafter, an iterative series of energy minimizations were performed on the water hydrogens, enzyme hydrogens and full water molecules. Finally, the whole system was energy minimized.

Preparation of the tetrahedral intermediate The catalytic histidine, His399, was defined as protonated. The general position of the tetrahedral intermediate of 1a and the orientation of its acyl moiety were modeled by superimposing the catalytic triad of BS2 with the catalytic triad of C. rugosa lipase in complex with a tetrahedral inhibitor (PDB code 1LPM, Cygler et al., 1994). The alcohol moiety was modeled in both (R) and (S) configurations. For each enantiomeric configuration of the alcohol moiety, three binding modes fitted sterically into the active site. To avoid local minima, for each binding mode, the Ser 189-bound substrate was submitted to an MD simulation of 1 ps. Finally, energy minimization of the whole system was performed to produce the starting structure for the subsequent MD simulations.

Molecular dynamics simulations The energy-minimized free enzyme and each energyminimized enzyme substrate system were run through a MD warm-up phase to a temperature of 300 K in a series of 30 short (1 ps) steps of 10 K intervals. Thereafter, an MD simulation of 200 ps was performed at 300 K for each system. A sample structure was extracted at every 0.2 ps from each simulation. Results and discussion In this study, we applied two complementary strategies to achieve a highly enantioselective kinetic resolution: Medium-engineering via addition of cosolvents was used to suppress the autohydrolysis of acetate 1a; rational protein design of BS2 was further investigated, to suggest alternative mutants with stronger influence on the enantioselectivity. In addition, the more stable trifluoromethyl analog 2a was synthesized and included in these studies (Fig. 1). These strategies resulted in a highly enantioselective kinetic resolution of both acetates with E-values exceeding 100. 127

B.Heinze et al.

Medium engineering The addition of water-miscible organic cosolvents such as dimethyl sulfoxide (DMSO) is a promising and quite frequently used method to improve the enantioselectivity of hydrolytic enzymes (Faber, 2004). The influence of DMSO on the enzymatic hydrolysis of 1a was investigated for wildtype BS2, and the mutants Gly105Ala and Glu188Asp (Table I). Best results were achieved by the addition of 20% (v/v) DMSO resulting in substantially higher ee-values for the Gly105Ala mutant (corresponding to observed enantioselectivity equal to E0 ¼ 56 vs. E0 ¼ 28 (E0 stands for observed enantioselectivity, which is a result of enzyme hydrolysis and autohydrolysis)) and the Glu188Asp mutant (E0 ¼ 41 vs. E0 ¼ 8). Of course, this effect was much less pronounced for the wild-type esterase (E0 ¼5 vs. E0 ¼ 3) as the enzyme-catalyzed hydrolysis had a low true enantioselectivity. Higher concentrations of DMSO led to enzyme inactivation (data not shown). We mostly attribute the increase in enantioselectivity to the suppression of the autohydrolysis. We used the program KRESH developed by Faber et al. (Faber,K., Mischitz,M. and Kleewein,A., Institute of Organic Chemistry, Graz University of Technology, Stremayergasse 16, A-8010 Graz, Austria) to calculate true E-values. In all cases, we could see that the mutants had very high enantioselectivity with E-values exceeding 100. With these high E-values, we could not detect any effect of DMSO on the enantioselectivity. Therefore, we ascribe the effect of DMSO to repression of autohydrolysis, leaving the increase of enantioselectivity to the mutations only. To avoid competing autohydrolysis in the resolution of 1a, the trifluoromethyl analog 2a (Fig. 1) was used.1 Due to the electron-withdrawing nature of the CF3-group, 2a is known to be resistant against autohydrolysis (O’Hagan and Zaidi, 1992). Thus, the wild-type of BS2 preferably converted (R)-1a and (S)-2a. No autohydrolysis was observed when using 2a. Hence, the enantioselectivity of wild-type BS2 and the mutant Gly105Ala increased considerably (Table I).

Table I. Influence of substrate and medium engineering on the kinetic resolution of 1a and 2a BS2

Substrate

DMSO [%v/v]

Enantiomeric excess

Conversion [%]b

E

[%eeS]a [%eeP]a Wild-type Wild-type Wild-type Gly105Ala Gly105Ala Gly105Ala Glu188Asp Glu188Asp

1a 1a 2a 1a 1a 2a 1a 1a

0 20 20 0 20 20 0 20

15 47 46 78 91 94 20 62

47 52 93 85 89 96 74 91

25c 48 33 48 50 58 22 40

3 5 42 28 56 .100 8 46

a

Calculated from the peak areas obtained by chiral GC analysis. Calculated from the eeS, eeP-values. c Autohydrolysis was 24% in the absence and 5% in the presence of 20% (v/v) DMSO. b

The slow reacting enantiomer of the substrate 1a positioned its phenyl group between Gly105 and Ala400. Inspired by the earlier mutation Gly105Ala, which afforded an increase in enantioselectivity (from E¼3 to E ¼ 19 (Henke et al., 2003)), the space was further decreased by mutating either Ala107 or Ala400 to Trp. The large substrate binding site could presumably accommodate these bulky insertions, but was followed by a decrease of reaction rate to 5% of that of the wild-type as observed in the hydrolysis of

Rational protein design Molecular dynamics was used to study the interaction between the enantiomeric substrates and the enzyme in transition state. The tetrahedral intermediate was used as a model of the transition state as often done when studying serine hydrolases (Haeffner et al., 1998; Henke et al., 2003) (Fig. 2A). The position of the acetyl moiety of 1a and the oxyanion hole was defined by superimposing the active site of BS2 on the highly similar structure of C. rugosa lipase with a bound transition state inhibitor (Cygler et al., 1994). The positions of the acetyl and the oxyanion hole were also in agreement with the same interactions in C. antarctica lipase B (Haeffner et al., 1998). The large active site of BS2 allows the alcohol part of the substrate to adopt several possible orientations. Thus, by manual docking three binding modes for each enantiomer were found. By comparing the binding modes and studying the interactions between the enzyme and the enantiomeric alcohols moieties, we looked for mutations that either would obstruct the reaction for the slow enantiomer or facilitate it for the fast one. 1

It should be noted that the trifluoromethyl group has a higher priority in the Cahn– Ingold–Prelog system and therefore the absolute configuration of 2 is inverted compared to 1 (Cahn et al., 1956).

128

Fig. 2. Schematic representation of the tetrahedral intermediate of 1a in the active site of BS2. The catalytically important hydrogen bonds of the catalytic triad (between Glu310 and His399), the oxyanion hole (between the oxyanion and Gly106, Ala107 and Ala190) and between His399 and the oxygen atoms of the alcohol moiety and Ser189 are indicated (A). In the MD simulations, a further hydrogen bond between His399 and Glu188 was observed (B).


Fig. 3. Snap shots from MD simulations showing the tetrahedral intermediate of 1a and the catalytically important amino acids His399 and Glu310 in the active site of BS2. The upper panel shows how His399 bends down and disrupts the hydrogen bonds to the tetrahedral intermediate to form one to Glu188 instead. The lower panel shows how this is avoided by the mutation Glu188Asp, which shifts the carboxyl away from His399 rendering it impossible to form the unproductive hydrogen bond. Substrate 1a is seen in green except for the tetrahedral carbon and its oxygens, which are shown in white and red, respectively. The functional groups of His399, Glu310 and Glu188 are color coded by type of atom (blue for nitrogen and light blue for hydrogen). The oxyanion hole is formed from the backbone amide hydrogens of Ala107, Gly 106 and Gly105.

129

B.Heinze et al.

p-nitrophenyl acetate ( pNPA). Further, no improvement of the enantioselectivity was seen. Thus, the strategy of introducing steric repulsion into the active site of BS2 based on manual docking did not prove to be a promising approach. The binding modes were studied by molecular dynamics simulations in order to identify stable productive ones. The hydrogen bonding network of a productive binding mode should have the typical H-bond pattern (Fig. 2A) of the tetrahedral intermediate of serine hydrolases (Uppenberg et al., 1994). The hydrogen bonding pattern was monitored during several molecular dynamics runs of 100 ps in order to investigate which of the six identified binding modes could lead to a reaction. The H-bonds of the oxyanion hole and between His399 and Glu310 were intact during all MD simulations. However, the two hydrogen bonds between His399 and the tetrahedral intermediate were only kept in one binding mode of the (S)-1a configuration, suggesting a stable hydrogen bonding network, while no catalytically competent binding mode could be found for (R)-1a. According to the modeling, the imidazol group of His399 moved downwards and formed a hydrogen bond with Glu188, the neighboring amino acid of the catalytically active Ser189 (Figs 2B and 3). Fig. 4 shows a section of 14 ps of the MD simulation of the productive binding mode of the (S)-enantiomer. At 22 ps, both H-bonds from His399 to Ser189 and to the alcohol were formed, allowing a reaction. However, at 16 ps a stable hydrogen bond existed between His399 and Glu188, while the distances between His399 and the alcohol, and His399 and Ser189 were too long to allow hydrogen bonding. Thus, the bending of His399 appears to hinder the formation of the hydrogen bonds between His399 and the tetrahedral intermediates. To study the role of Glu188 on the bending of His399, a model of the mutant Glu188Asp was investigated. The residue 188 in Glu188Asp would still have the negative charge while the distance from the carboxylate to His399

would increase. In two investigated binding modes of 1a in Glu188Asp (one (R) and one (S) configuration), the entire hydrogen bonding network stayed intact over 80% of the simulation period, while the hydrogen bond between His399 and the carboxylate at 188 was not formed. Based on these observations, three mutations were introduced at position Glu188 by the site-directed mutagenesis and analyzed experimentally. Glu188Asp was chosen, as the side-chain is still negatively charged, but has a smaller size. Substituting Glu188Gln, the charge was removed without altering the size of the functional group. Glu188Phe was created in order to introduce steric hindrance very close to the catalytically active serine. It turned out that Glu188Gln had a reduced activity in the hydrolysis of pNPA and a lower enantioselectivity towards 1a and 2a than the wild-type enzyme (Table II). The substitution of Glu188 with phenylalanine resulted in a reduced activity towards pNPA (40%) and, surprisingly, in an inversion of the enantiopreference in the conversions of 1a (ES ¼ 2) and 2a (ER ¼ 4). Inversion of enantioselectivity of lipases and esterases towards tertiary aclohol acetates is still a challenge. A successful example of an enantiopreferenceswitch towards a secondary alcohol by rational protein design has been reported recently: A single amino acid exchange in the active site of C. antarctica lipase B of a tryptophane by an alanine resulted in an altered enantiopreference towards 1phenylethanol from (R) to (S) (Magnusson et al., 2005). Additionally, the inversion of enantioselectivity towards esters of sec-alcohols was also achieved by directed evolution (Koga et al., 2003; Zha et al., 2001). An inversion of the enantioselectivity of BS2 towards the tertiary alcohol linalool has been reported for the mutant Gly105Ala (Henke et al., 2003). We were very pleased to find that the Glu188Asp mutant converted 2a with an excellent enantioselectivity (ES . 100), which was also confirmed by a preparative scale experiment yielding (R)-2a at 94%ee and (S)-2b at .99%ee with

Fig. 4. Distances from the nitrogen atom of His399 to the oxygen atom of Ser189 (squares), to the oxygen atom of the alcohol moiety of 2a (open triangles) and to Glu188 (open circles) in a section of 14 ps of a MD simulation of (S)-2a in BS2. The maximal distance that allows the formation of hydrogen bonds ˚ ) is indicated in gray shade. (2.8 A

130


Table II. Esterase mutants and their properties E a,b and preferred enantiomer substrate 1a Wild-type Gly105Ala Glu188Gln Glu188Phe Glu188Asp

5 56 2 2 46

Relative activityc pNPA %

2a R R R S R

42 .100 2 4 .100d

S S S R S

100 9 16 40 15

a

Calculated from the eeS, eeP-value. 20%(v/v) DMSO. Activity of purified enzyme in the hydrolysis of p-nitrophenyl acetate relative to the wild-type (100% ¼ 151.4 U/mg protein). d Preparative scale. b c

satisfactory isolated yields. The high enantioselectivity of Glu188Asp towards 2a was also confirmed using the enzyme purified via C-terminal His-tag (data not shown). In the conversion of 1a, the enantioselectivity of Glu188Asp (ER ¼ 46) was in the range of the Gly105Ala variant (ER ¼ 56). The alteration of the position of a carboxyl group directly in the catalytic center of the enzyme led to a considerable reduction of activity towards pNPA (15%), but also to a substantial increase in enantioselectivity. The reason for this is difficult to elucidate, but an effect on the bending of His399 appears to be most likely. Conclusions We could demonstrate that the enantioselectivity of the esterase BS2 towards esters of the tertiary alcohols 1a and 2a could be substantially increased by the suppression of autohydrolysis for reactions performed in the presence of the water-miscible cosolvent DMSO, use of the more stable trifluoromethyl analog of 1a and by rational protein design. The last strategy led to a variant with inversed enantioselectivity, but especially to one mutant with synthetically useful enantioselectivity. The effective combination of both strategies facilitated the highly enantioselective kinetic resolution of two tertiary alcohols.

Chen,C.S., Fujimoto,Y., Girdaukas,G. and Sih,C.J. (1982) J. Am. Chem. Soc., 104, 7294–7299. Christoffers,J. and Mann,A. (2001) Angew. Chem. Int. Ed., 40, 4591– 4597. Christoffers,J. and Baro,A. (2005) Quaternary Stereocenters. Wiley-VCH, Weinheim. Corey,E.J. and Guzman-Perez,A. (1998) Angew. Chem. Int. Ed., 37, 388–401. Cygler,M., Grochulski,P., Kazlauskas,R.J., Schrag,J.D., Bouthillier,F., Rubin,B., Serreqi,A.N. and Gupta,A.K. (1994) J. Am. Chem. Soc., 116, 3180– 3186. Faber,K. (2004) Biotransformations in Organic Chemistry. Springer, Berlin. Haeffner,F., Norin,T. and Hult,K. (1998) Biophys. J., 74, 1251–1262. Henke,E., Pleiss,J. and Bornscheuer,U.T. (2002) Angew. Chem. Int. Ed., 41, 3211– 3213. Henke,E., Bornscheuer,U.T., Schmid,R. D. and Pleiss,J. (2003) Chem. Biol. Chem., 4, 485 –493. Koga,Y., Kato,K., Nakano,H. and Yamane,T. (2003) J. Mol. Biol., 331, 585–592. Krebsfa¨nger,N., Zocher,F., Altenbuchner,J. and Bornscheuer,U.T. (1998) Enzyme Microb. Technol., 22, 641–646. Krishna,S.H., Persson,M. and Bornscheuer,U.T. (2002) Tetrahedron: Asymmetry, 13, 2693–2696. Magnusson,A.O., Takwa,M., Hamberg,A. and Hult,K. (2005) Angew. Chem. Int. Ed., 44, 4582– 4585. O’Hagan,D. and Zaidi,N.A. (1992) J. Chem. Soc. Perkin Trans. I, 947–949. O’Hagan,D. and Zaidi,N.A. (1994) Tetrahedron: Asymmetry, 5, 1111–1118. Ordentlich,A., Barak,D., Kronman,C., Ariel,N., Segall,Y., Velan,B. and Shafferman,A. (1998) J. Biol. Chem., 273, 19509–19517. Pleiss,J., Scheib,H. and Schmid,R.D. (2000) Biochimie, 82, 1043–1052. Powell,M.J.D. (1977) Math. Program., 12, 241– 254. Raza,S., Fransson,L. and Hult,K. (2001) Protein Sci., 10, 329 –338. Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Habor, NY. Schlacher,A., Stanzer,T., Osprian,J., Mischitz,M., Klingsbichel,E., Faber,K. and Schwab,H. (1998) Biotechnol., 62, 47–54. Schmidt,M., Henke,E., Heinze,B., Kourist,R., Hidalgo,A. and Bornscheuer,U.T. (2007) Biotechnol. J., 2, 249–253. Spiller,B., Gershenson,A., Arnold,F.H. and Stevens,R.C. (1999) Proc. Natl Acad. Sci. USA, 96, 12305– 12310. Uppenberg,J., Hansen,M.T., Patkar,S. and Jones,T.A. (1994) Structure, 2, 293–308. Weiner,S.J., Kollman,P.A., Case,D.A., Singh,U.C., Ghio,C., Alagona,G., Profeta,S. and Weiner,P. (1984) J. Am. Chem. Soc., 106, 765– 784. Weiner,S.J., Kollman,P.A., Nguyen,D.T. and Case,D.A. (1986), J. Comp. Chem., 7, 230 –252. Whiting,A.K. and Peticolas,W.L. (1994) Biochemistry, 33, 552–561. Zha,D.X., Wilensek,S., Hermes,M., Jaeger,K.E. and Reetz,M.T. (2001) Chem. Commun., 2664–2665

Acknowledgements The authors gratefully acknowledge Matthias Ho¨hne and Anett Kirschner for helpful discussions and Aurelio Hidalgo for providing the altered pG-BS2 construct. This work was supported by the Deutscher Akademischer Austauschdienst (DAAD, Germany) and The Swedish Foundation for International Cooperation in Research and Higher education (STINT, Sweden).

References Berman,H.M., Westbrook,J., Feng,Z., Gilliland,G., Bhat,T.N., Weissig,H., Shindyalov,I.N. and Bourne,P.E. (2000) Nucleic Acids Res., 28, 235– 242. Berthod,H. and Pullman,A. (1965) J. Chim. Phys., 62, 942–946. Berthod,H., Giessner-Prettre,C.L. and Pullman,A. (1967) Theor. Chim. Acta, 8, 212– 222. Bornscheuer,U.T. and Kazlauskas,R.J. (2005) Hydrolases in Organic Synthesis. Wiley-VCH, Weinheim. Bryan,P., Pantoliano,M.W., Quill,S.G., Hsiao,H.Y. and Poulos,T. (1986) Proc. Natl Acad. Sci. USA, 83, 3743–3745. Cahn,R.S., Ingold,C.K. and Prelog,V. (1956) Experientia, 12, 81– 94

131