DOI: 10.1002/cbic.200500546
Directed Evolution of an Esterase from
Pseudomonas fluorescens Yields a Mutant with
Excellent Enantioselectivity and Activity for the
Kinetic Resolution of a Chiral Building Block
Marlen Schmidt,[a]
Daniel Hasenpusch,[b]
Markus Kähler,[c]
Ulrike Kirchner,[c]
Kerstin Wiggenhorn,[c]
Walter Langel,[b]
and Uwe T. Bornscheuer*[a]
Introduction
Directed evolution has emerged in the past decade as the
most powerful method for improving biocatalysts.[1]
For instance,
the enantioselectivity of a d-hydantoinase could be reversed,
thus leading to a substantially improved process for
the synthesis of optically pure l-amino acids.[2]
Reetz and
Jaeger were able to increase the enantioselectivity of a lipase
from Pseudomonas aeruginosa towards a chiral carboxylic acid
(2-methyl decanoate), initially from E=1.1 (wild-type or WT) to
E=11.[3]
Using a combination of a broad range of molecular
biology methods they were finally able to identify a variant
with practically useful selectivity (E=51).[4]
Similarly, we have
been able to increase the enantioselectivity of an esterase
from Pseudomonas fluorescens (PFE) towards 3-phenylbutyric
acid from E=3.5 to E=12 by combining error-prone PCR
(epPCR) with saturation mutagenesis.[5]
Successful directed-evolution experiments depend on several
aspects. The identification of desired variants that exhibit an
increased enantioselectivity strongly depends on the highthroughput
screening (HTS) method used—the selectivity determined
with a surrogate substrate (i.e., a p-nitrophenyl ester)
can differ significantly from the true substrate (i.e., a methyl
ester), which can lead to false-positive variants. In addition, it is
assumed that the random introduction of mutations does not
significantly affect other properties of the biocatalyst and its
production in the microbial host. Another aspect is the location
of productive mutations. Directed-evolution experiments
have often led to variants in which effective mutations were
far from the active-site region, although they affected substrate
specificity or enantioselectivity. Consequently, the question
of whether closer mutations are better has been already
addressed in the literature.[6]
The hydrolase-catalyzed resolution of the acetate of 1a
(Scheme 1) is very challenging as this secondary alcohol has
only small differences in the size of its substituents. In accordance
with “Kazlauskas’ rule”,[7]
it is converted by lipases or esterases
only with low to modest E values.
The (R)-alcohol can be used for the synthesis of (À)-akolacACHTUNGTRENNUNGtone
A a cyctotoxic butenolide,[8]
pancrastatin an antitumour
alkaloid,[9]
chiral cyclopropane-based ligands[10]
and for the
large-scale production of (R)-benzyl 4-hydroxylpent-2-ynoate.[11]
Both pure enantiomers of 1a have been employed in the
preparation of enantio- and diastereomerically pure allylboronic
ester by Johnson rearrangement, which led to enantiopure
homoallyl alcohols.[12]
Other examples include the synthesis of
3-bromopyrrolines from a-amino allenes,[13]
optically active bicyclic
ligands used for the synthesis of HIV protease inhibi-
tors[14]
or key intermediates of avb3 antagonist (useful for the
treatment of osteoporosis).[15]
One example for an application
of the (S)-alcohol is the preparation of chiral cyclic carbo-
nates.[16]
Previously, we investigated more than 100 hydrolases for the
kinetic resolution of 1b[17]
and the best enzyme was the ester[a]
Dr. M. Schmidt, Prof. Dr. U. T. Bornscheuer
Institute of Biochemistry, Department of Biotechnology & Enzyme Catalysis
Greifswald University, Soldmannstrasse 16, 17487 Greifswald (Germany)
Fax: (+49)3834-86-80066
E-mail: uwe.bornscheuer@uni-greifswald.de
[b] D. Hasenpusch, Prof. Dr. W. Langel
Institute of Biochemistry, Department of Biophysical Chemistry
Greifswald University, Greifswald (Germany)
[c] Dr. M. Kähler, Dr. U. Kirchner, K. Wiggenhorn
Dr. Rieks Healthcare GmbH
Deichstrasse 25a, 25436 Uetersen (Germany)
A triple mutant of an esterase from Pseudomonas fluorescens
ACHTUNGTRENNUNG(PFE) that was created by directed evolution exhibited high enantioselectivity
(E=89) in a kinetic resolution and yielded the building
block (S)-but-3-yn-2-ol. Surprisingly, a mutation close to the
active site caused the formation of inclusion bodies, but remote
mutations were found to be responsible for the high selectivity.
Back mutations gave a variant (double mutant PFE Ile76Val/
Val175Ala) that showed excellent selectivity (E=96) and activity
(20 min for 50% conversion, which corresponds to 1.25 U per mg
of protein).
ChemBioChem 2006, 7, 805 – 809  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 805
ase from P. fluorescens expressed in E. coli.[18]
However, detailed
analysis of the reaction time course revealed that the enantioselectivity
dropped considerably during hydrolysis of the acetate
and complete conversion was observed (Figure 1A;
Table 1), which led to racemic alcohol 1a. A similar effect was
described in a patent for an esterase from Burkholderia glumae
(previously designated Pseudomonas glumae).[19]
In the hydrolysis
of the corresponding butyrate, the initial value was E~30,
but also dropped at higher conversion which made a highyield
resolution impractical. Only Nakamura and co-workers
have reported acceptable enantioselectivity, but the smallscale
resolution suffered considerably from the use of large
amounts of lipase Amano AH and very long reaction times
(2 days).[20]
The alternative asymmetric enzymatic reduction of
the ketone was hampered by the very low enantiomeric
excess, as shown for a NADPH-dependent alcohol dehydrogenase
from Lactobacillus brevis, which afforded the (R)-alcohol
with 60% ee or a NADH-dependent Candida parapsilosis carbonyl
reductase that yielded the (S)-alcohol with 49% ee.[21]
In
addition, but-3-yn-2-one is rather unstable and has high risk of
thermal decomposition, which would restrict its large-scale reduction
even if a highly selective reductase were available.
In this work, we used methods of directed evolution to improve
the enantioselectivity of an esterase by using the secACHTUNGTRENNUNGondary
alcohol rac-but-3-yn-2-ol as target compound (1a;
Scheme 1). As detailed below, the number and positions of
amino-acid substitutions gave rather unexpected results.
Results and Discussion
To create an enantioselective variant for resolving 1b, epPCR libraries
with a mutation rate of one to two mutations per gene
of PFE were created, followed by HTS by using the “acetic-acid
assay” previously developed in our laboratory.[22]
Acetic acid released
in the esterase-catalyzed hydrolysis of acetate 1b was
stoichiometrically converted in an enzyme-cascade reaction
Table 1. Specific activities of wild-type PFE and its variants towards pNPA, their E values at 50% conversion and their enantioselectivity after prolonged
ACHTUNGTRENNUNGreaction times (Emax).
Name PFE variant Activity [U per mg protein] E~50%[b]
t [min] Emax
[c]
t [min]
lyophilisate IB[a]
WT wild-type 77 À 63 5 3 (96%) 1440
V2A Ile76Val/Gly98Ala/Val175Ala 0.006 + 89 5700 89 (54%) 5700
2A Gly98Ala/Val175Ala 0.2 + >100[d]
180 >100 (40%) 180
VA1 Ile76Val/Gly98Ala 0.6 + 80[e]
10 80 (25%) 10
A1 Gly98Ala 9 + >100 5 90 (57%) 1500
VEA2 Ile76Val/Asp99Glu/Val175Ala 37 À 92 420 92 (53%) 420
V Ile76Val 49 À >100 1 16 (83%) 1500
VA2 Ile76Val/Val175Ala 57 À 96 20 96 (53%) 20
A2 Val175Ala 67 À >100 1 26 (74%) 1500
[a] IB: inclusion body; [b] calculated at 50% conversion; [c] calculated at maximal conversion given in brackets (%); [d] calculated at 40% conversion;
[e] calculated at 25% conversion.
Scheme 1. Principle of esterase-catalyzed kinetic resolution.
Figure 1. Hydrolysis of 1b with A) wild-type PFE, and B) the PFE triple
mutant (Ile76Val/Gly98Ala/Val175Ala); (*) conversion; (&) enantioselectivity.
806 www.chembiochem.org  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2006, 7, 805 – 809
U. T. Bornscheuer et al.
into NADH, which was quantified spectrophotometrically at
340 nm. As enantiopure (R)- and (S)-acetates were used in separate
wells of a microtiter plate, the apparent enantioselectivity
(Eapp) of each esterase variant could be determined from these
initial rate measurements. The E values were then confirmed
by kinetic resolution after shake-flask production of positive
variants in small-scale experiments by GC analysis (Etrue). After
screening ~7000 mutants, a PFE mutant was identified that exhibited
an Etrue =89 at 54% conversion. Unfortunately, the reaction
time was extremely long (>24 h; Figure 1B) compared to
only a few minutes required for similar conversion values
when using the wild-type PFE (Figure 1A).
Sequencing of this variant identified three point mutations
(Ile76Val/Gly98Ala/Val175Ala). Surprisingly, cell fractionation
and analysis of the pellet by SDS-PAGE showed that this triple
mutant—in contrast to the WT—was produced as inclusion
bodies (IBs) and only a minor fraction of soluble protein was
formed. This had a specific activity of only 0.006 U per mg protein
(Vmax/KM =5.3”10À5
minÀ1
) compared to 77 U per mg protein
(Vmax/KM =0.57 minÀ1
) for the wild-type PFE when using pnitrophenyl
acetate (pNPA) as assay substance. IB formation
could be slightly decreased by incubating host cells at 258C,
which yielded a specific activity of 0.5 U per mg protein.
To locate the mutations, a model of the triple mutant was
created based on the known 3D structure of PFE[23]
(PDB ID:
1VA4; Figure 2). From this structure, it was initially assumed
that the Gly98Ala mutation close to the catalytic triad of the
esterase (Ser94, His251, Asp222) led to the substantial increase
in enantioselectivity (“closer is better”),[6]
and that the other
two remote mutations (Ile76Val, Val175Ala) were responsible
for the formation of IBs.
To verify this assumption, all corresponding double and
single mutants were created by QuikChange site-directed mutagenesis
(Table 1). Beside the planned variants, an additional
mutation was found at position 99 (Asp99Glu).
Four variants (VEA2, V, VA2 and A2) that lacked the Gly98Ala
mutation exhibited specific activities similar to the WT enzyme
without IB formation. It is worth noting that variant VEA2
(Ile76Val/Asp99Glu/Val175Ala) formed no IBs (compared to
V2A) and had a specific activity similar to the WT. From this we
concluded that the Gly98Ala mutation (A1) next to the catalytic
triad must be responsible for IB formation.
To gain a deeper insight into the structural reasons for our
experimental findings a molecular-dynamic simulation of both
enzyme variants, wild-type PFE and the triple mutant V2A,
were performed with an AMBER94 force field implemented in
the AMBER7 package, which included the TIP3P model for
water. An alignment of both resulting structures was performed
with PyMOL (http://pymol.sourceforge.net). In the
triple mutant V2A a helix extended by one loop was identified
at position 98 (Figure 3). Here, the helix breaker, glycine,[24]
was
replaced by the helix creator, alanine, at the end of the a-helix.
This new conformation appears to destroy the original tertiary
structure and might result in the formation of IBs. No significant
alterations at the other two mutated positions were ob-
served.
Next, the enantioselectivity of all variants was investigated
by small-scale resolutions and we were pleased to find, that
two mutants exhibited satisfying activity and also enantioselectivity
in the resolution of 1b without further conversion and
reduction in E values: the double mutant VA2 (Ile76Val/
Val175Ala) with E=96 (53% conversion, 20 min, 1.25 U per mg
protein) and a specific activity towards the standard substrate
pNPA of 57 U per mg protein (Vmax/KM =0.51 minÀ1
), and VEA2
Figure 2. 3D homology model of PFE. The catalytic triad is shown in grey
(Ser94, His 251, Asp222), mutation sites are highlighted in black (Val76,
Ala98 and Ala175). The model was created by using PyMOL and amino-acid
exchanges were introduced with the “Wizard/Mutagenesis” feature (http://
pymol.sourceforge.net).
Figure 3. Alignment of WT (light grey, Gly98 labelled) and triple mutant V2A
(dark grey, Ala98 labelled); the extended loop of the helix is highlighted.
ChemBioChem 2006, 7, 805 – 809  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chembiochem.org 807
Directed Evolution of an Esterase from P. fluorescens
(Ile76Val/Asp99Glu/Val175Ala) with E=92 (53% conversion,
7 h, 0.06 U per mg protein) and a specific activity of 37 U per
mg protein (Vmax/KM =0.31 minÀ1
). Both variants have the same
temperature stability and pH optimum as wild-type PFE (data
not shown).
Thus, by using directed evolution and subsequent site-directed
mutagenesis, we were able to create several esterase
variants that showed excellent enantioselectivity and kinetics
in the resolution of the acetate of but-3-yn-2-ol to yield the
(S)-enantiomer. Due to their high selectivity and activity with
these variants, the (R)-enantiomer was also available from the
remaining acetate.
Beside the creation of a biocatalyst that is useful for efficient
kinetic resolution, this study has also shown that a mutation
close to the active site can have a substantial impact on protein
folding. This aspect has not been described so far in the
literature and should be considered further in protein engineering,
especially when using directed-evolution methods. In
addition, mutations near the active site do not necessarily lead
to altered enantioselectivity and in this case remote mutations
were shown to be most effective in creating a highly enantioselective
enzyme.
Furthermore, this is the first example of the successful creation
of an esterase with substantially improved enantioselectivity
towards secondary alcohols, obtained by using methods of
directed evolution.
Experimental Section
General: All chemicals were purchased from Fluka (Buchs, Switzerland),
Sigma (Steinheim, Germany) and Merck (Darmstadt, Germany)
unless stated otherwise. Restriction enzymes, ligase, DNAseI
and polymerases were obtained from Promega (Madison, WI, USA)
and New England BioLabs GmbH (Beverly, MA, USA). MWG-Biotech
(Ebersberg, Germany) provided primers and performed the sequencing
experiments.
Bacterial strains, plasmid, growth conditions and protein analysis:
E. coli DH5a or JM109 were used as hosts for transformation of
plasmid DNA. The strains were grown at 378C in Luria–Bertani (LB)
liquid media or on LB agar plates supplemented with ampicillin
(100 mgmLÀ1
).[25]
The vector pJOE2792.1 with a rhamnose-inducible
promoter was used for expression of PFE. Esterase production was
induced upon addition of rhamnose (final concentration 0.2%, v/v)
and the culture was continued for 5 h. Cells were collected by centrifugation
(15 min, 48C, 3939 g) and washed twice with sodium
phosphate buffer (10 mm, pH 7.4, 48C). Cells were disrupted by
sonication on ice for 5 min at 50% pulse and centrifuged to separate
soluble from insoluble fractions, the latter contained the IBs.
The supernatant was lyophilized and stored at 48C. Protein content
was determined by using Bradford reagent with bovine serum albumin
as standard.
Esterase activity was determined spectrophotometrically by hydrolysis
of pNPA (10 mm in DMSO) in sodium phosphate buffer
(10 mm, pH 7.4). Released p-nitrophenol was quantified at 410 nm
(e=15”103
mÀ1
cmÀ1
). One unit (U) of activity was defined as the
amount of enzyme that released 1 mmol p-nitrophenol per min
under assay conditions.[18]
Proteins from soluble and insoluble fractions
were also analyzed by separating (12%) and stacking (4%)
SDS polyacrylamide gels.[25]
After electrophoresis the gels were first
activity stained with a-naphthylacetate and Fast Red[18]
followed
by Coomassie brilliant blue staining.
Creation of epPCR library: Plasmids, isolated with the QIAprep kit
(Qiagen, Hilden, Germany) were used in epPCR experiments at a
final concentration of 0.1 ngmLÀ1
. The reaction mixture consisted
of dNTP Mix (10 mL; 2 mm ATP, 2 mm GTP, 10 mm CTP, 10 mm TTP),
mutation buffer (10 mL; 70 mm MgCl2, 500 mm KCl, 100 mm Tris
pH 8; 0.1% (w/v) gelatine), primers (100 pmol of each; forward
primer: GACTGGTCGTAATGAACAATTC; reverse primer: AATGATGATGATGATGGCATC),
Taq polymerase (1 mL, 5 UmLÀ1
), MnCl2 (3 mL,
10 mm) and plasmid DNA in a final volume of 100 mL. The reaction
conditions were: 1) 958C 60 s, 2) 25 cycles: 958C 30 s, 508C 30 s,
728C 45 s, 3) 728C 150 s. PCR products were purified with the QIA-
quickTM
PCR Purification Kit, digested with BamHI and NdeI to generate
cohesive ends, ligated in the empty vector pJOE2792.1 and
transformed into competent E. coli cells that had been prepared by
using the rubidium chloride method. Transformants were transferred
by replica plating to LB/Amp (100 mg) agar plates containing lrhamnose
(final concentration 0.2% v/v) to induce esterase production
and to enable the activity towards a-naphthylacetate to be
analysed. Active clones showing a red colour with Fast Red were
transferred to microtiter plates (MTPs) containing LB/Amp (200 mL)
in each well (master plate). The plates were incubated for 16 h at
378C and 50 rpm, glycerol or DMSO (end concentration 10%, v/v)
were added and the MTP was stored at À808C.
Enzyme production in MTPs with “steady-state growth method”:
Master plates were duplicated by transferring colonies with a
96 pin head onto a new MTP containing LB/Amp (200 mL per well;
production plate). These new plates were incubated for 24 h at
378C and 50 rpm; the cultures were then at the stationary growth
phase. A sample of the culture (100 mL) was transferred with a
pipet robot (Miniprep 75, Tecan, Crailsheim, Germany) from each
well to a new MTP well containing fresh LB/Amp (100 mL) and incubated
for a further 3 h to reach the exponential phase. Next, lrhamnose
was added to induce esterase production for another
5 h. Cells were harvested by centrifugation (1750 g, 15 min, 48C)
and supernatants were discarded. Lysis buffer (200 mL; 300 mm
NaCl, 50 mm Na2HPO4, pH 8) containing DNAseI (final concentration
1 UmLÀ1
) and lysozyme (final concentration 0.1%, w/v) were
added and cells were destroyed by one freeze-thaw-cycle. The supernatant
containing the esterase was either lyophilized or directly
stored at À208C. As the soluble protein contained mostly the recombinant
esterase (about 50% of total protein), the crude lyophilisate
was not purified before the activity and kinetic data were
ACHTUNGTRENNUNGdetermined.
Synthesis of corresponding racemic and enantiopure acetates:
Acetates were enzymatically synthesized in a transesterification reaction
from rac-1a (1 equiv) with vinyl acetate (1.5 equiv) in nhexane
(5 mL). Reaction mixtures contained molecular sieves and
Candida antarctica lipase B (5 mg; CAL-B, Novozym, Denmark) and
were stirred at 378C, overnight. Enzyme and molecular sieves were
removed by centrifugation, solvent and excess vinyl acetate were
evaporated. As CAL-B exhibited no enantioselectivity towards 1a,
complete conversion to the racemic acetate was possible.
Screening with the acetic acid assay: The test kit for the determination
of released acetic acid was from R-Biopharm GmbH (Darmstadt,
Germany) and applied according to the manufacture’s protocol.
Enzyme solution (20 mL) from the production plate and substrate
solution (20 mL; 7.5 mgmLÀ1
) of enantiopure (R)- or (S)-2-acetoxybut-3-yn
(1b) were added to a mixture (150 mL) of the test kit
components. The increase of NADH was monitored at 340 nm by
808 www.chembiochem.org  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioChem 2006, 7, 805 – 809
U. T. Bornscheuer et al.
using the Fluostar Galaxy or Fluostar Optima (BMG, Offenburg, Germany).
Mixtures of test kit components with buffer or cell lysate
from induced E. coli that contained a vector without an esterase
gene, served as controls. The ratio of the initial rates thus determined
for each enantiomer was defined as the apparent enantioselectivity
(Eapp).[17]
General method for esterase-catalyzed small-scale resolutions:
Esterase solution was added to a stirred solution of substrate 1b
(25 mm) in phosphate buffer (10 mm, pH 7.4). The reaction mixture
was stirred in a thermoshaker (Eppendorf, Hamburg, Germany) at
378C. At different time points, samples (100 mL) were taken and extracted
twice with dichloromethane (100 mL). The combined organic
layers were dried over anhydrous sodium sulphate and the organic
solvent was removed under nitrogen. The samples were analyzed
by gas chromatography (GC-14A gas chromatograph, Shimadzu,
Japan) by using a chiral column Hydrodex-b-3P (heptakis(2,6-di-O-methyl-3-O-pentyl)-b-cyclodextrin;
0.25 mm) with hydrogen
as carrier gas. Enantioselectivity (Etrue) and conversion were calculated
according to Chen et al.[26]
QuikChange: The QuikChangeTM
Site-Directed Mutagenesis Kit
(Stratagene, La Jolla, USA) was used according to the manufacturer’s
instructions with complementary primers: G229A exchange: 5’CTTCGCCGACGACATCGCCCAGTTGATC-3’,
C296G exchange: 5’CATGGGCGGCGGCGATGTGGCCCG-3’
and C527T exchange: 5’TCTCCCAAGGCGTGCAGACCCAGACC-3’.
The plasmid encoding for
the triple mutant (0.1 ngmLÀ1
as template) and the following reaction
conditions: 1) 958C 45 s, 2) 20 cycles: 958C 45 s, 558C 60 s,
688C 10 min. The mutated plasmids were transformed into competent
E. coli cells (4 mL reaction in 50 mL competent cells).
Molecular-dynamic simulation: The effect of mutations on the
structure of the PFE enzyme was investigated by a molecular-dynamics
simulation according to the following procedure. One
chain of the hexamer (PBD ID: 1VA4) was used as model for the
wild-type PFE and a mutant containing three mutations (Ile76Val,
Gly98Ala, Val175Ala) was generated from this strand. Both proteins
were embedded into orthorombic cells containing around 9000
water molecules that provided a layer of at least 10 Š in thickness.
The AMBER94 forcefield as implemented in the AMBER7 package
was used including the TIP3P model for water. Geometry optimization
was performed in three stages (12000 steps each) which provided
the optimization of the water and protein molecule alone
and of the total system. A molecular dynamics run of 20000 steps
corresponding to 20 ps at constant pressure and temperature
(NpT) resulted in a slight reduction of the cell size to 68.77”
64.32”68.83 Š3
, thus attaining a standard water density. After this
relaxation a production run of 200 ps without temperature and
pressure control (NVE) was added.
The alignment of the resulting structures for wild-type and mutant
enzymes was performed with PyMOL (http://pymol.sourceforge.
net).
Acknowledgements
Financial support by the Deutsche Bundesstiftung Umwelt
(Osnabrück, Germany, grant no. AZ13062) is gratefully acknowl-
edged.
Keywords: directed evolution · enantioselectivity · enzyme
catalysis · esterases · protein engineering
[1] a) H. F. Arnold, A. A. Volkov, Curr. Opin. Chem. Biol. 1999, 3, 54–59;
b) U. T. Bornscheuer, Angew. Chem. 1998, 110, 3285–3288; Angew.
Chem. Int. Ed. 1998, 37, 3105–3108; c) K.-E. Jaeger, T. Eggert, Curr. Opin.
Biotechnol. 2004, 15, 305–513; d) M. T. Reetz, Proc. Natl. Acad. Sci. USA
2004, 101, 5716–5722.
[2] O. May, P. T. Nguyen, F. H. Arnold, Nat. Biotechnol. 2000, 18, 317–320.
[3] M. T. Reetz, A. Zonta, K. Schimossek, K. Liebeton, K.-E. Jaeger, Angew.
Chem. 1997, 109, 2961–2963; Angew. Chem. Int. Ed. Engl. 1997, 36,
2830–2832.
[4] a) K. Liebeton, A. Zonta, K. Schimossek, M. Nardini, D. Lang, B. W. Dijkstra,
M. T. Reetz, K.-E. Jaeger, Chem. Biol. 2000, 7, 709–718; b) M. T.
Reetz, Tetrahedron 2002, 58, 6595–6602.
[5] E. Henke, U. T. Bornscheuer, Biol. Chem. 1999, 380, 1029–1033.
[6] K. L. Morley, R. J. Kazlauskas, Trends Biotechnol. 2005, 23, 231–237.
[7] R. J. Kazlauskas, A. N. E. Weissfloch, A. T. Rappaport, L. A. Cuccia, J. Org.
Chem. 1991, 56, 2656–2665.
[8] W. P. Gallagher, R. E. J. Maleczka, J. Org. Chem. 2003, 68, 6775–6779.
[9] H. Ko, E. Kim, J. E. Park, D. Kim, S. Kim, J. Org. Chem. 2004, 69, 112–121.
[10] G. A. Molander, J. P. Burke, P. J. Carroll, J. Org. Chem. 2004, 69, 8062–
8069.
[11] X. Fu, J. Yin, Thiruvengadam, K. T. McAllister, L. T. C.-H. Tann, C. Colon,
Org. Process Res. Dev. 2002, 6, 308–310.
[12] J. Pietruszka, N. Schçne, Eur. J. Org. Chem. 2004, 24, 5011–5019.
[13] A. Horvµth, J. Benner, J. E. Baeckvall, Eur. J. Org. Chem. 2004, 15, 3240–
3243.
[14] R. Yanada, Y. Koh, N. Nishimori, A. Matsumura, S. Obika, H. Mitsuya, N.
Fujii, Y. Takemoto, J. Org. Chem. 2004, 69, 2417–2422.
[15] F. W. Hartner, Y. Hsiao, K. K. Eng, N. R. Rivera, M. Palucki, L. Tan, N.
Yasuda, D. L. Hughes, S. Weismann, D. Zewge, T. King, D. Tschaen, R. P.
Voante, J. Org. Chem. 2004, 69, 8723–8730.
[16] M. Yoshida, Y. Ohsawa, M. Ihara, J. Org. Chem. 2004, 69, 1590–1597.
[17] M. Baumann, B. H. Hauer, U. T. Bornscheuer, Tetrahedron Asymmetry
2000, 11, 4781–4790.
[18] N. Krebsfänger, F. Zocher, J. Altenbuchner, U. T. Bornscheuer, Enzyme
Microb. Technol. 1998, 22, 641–646.
[19] B. Hauer, T. Friedrich, C. Nübling, R. Stürmer, 2002. German patent application,
DE10042892A1, 31.08.2000.
[20] K. Nakamura, K. Takenaka, Tetrahedron Asymmetry 2002, 13, 415–422.
[21] T. Schubert, W. Hummel, M. R. Kula, M. Müller, Eur. J. Org. Chem. 2001,
22, 4181–4187.
[22] M. Baumann, R. Stürmer, U. T. Bornscheuer, Angew. Chem. 2001, 113,
4329–4333; Angew. Chem. Int. Ed. 2001, 40, 4201–4204.
[23] J. D. Cheeseman, A. Tocilj, S. Park, J. D. Schrag, R. J. Kazlauskas, Acta
Crystallogr. Sect. D Biol. Crystallogr. 2004, 60, 1237–1243.
[24] P. Y. Chou, G. D. Fasman, Biochemistry 1974, 13, 211–222.
[25] J. Sambrook, E. F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory
Manual, 2nd ed., Cold Spring Harbor Laboratory Press, New York, 1989.
[26] C. S. Chen, Y. Fujimoto, C. J. Sih, J. Am. Chem. Soc. 1982, 104, 7294–
7299.
Received: December 19, 2005
Published online on March 31, 2006
ChemBioChem 2006, 7, 805 – 809  2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chembiochem.org 809
Directed Evolution of an Esterase from P. fluorescens