Manipulating the Expression Rate and Enantioselectivity of an Epoxide Hydrolase by Using Directed Evolution

Article abstract of DOI:10.1002/cbic.201100078

We describe here a strategy to improve the expression efficiency and enantioselectivity of Aspergillus niger epoxide hydrolase (ANEH) by directed evolution. Based on a blue-colony screening system using the LacZα (β-galactosidase α peptide) complementation solubility reporter, several ANEH variants out of 15000 transformants from a random-mutagenesis library were identified that show improved recombinant expression in E. coli. Among them, Pro221Ser was subsequently used as a template for iterative saturation mutagenesis (ISM) at sites around the ANEH binding pocket. Following four rounds of ISM, a highly enantioselective mutant was identified that catalyzes the hydrolytic kinetic resolution of racemic glycidyl phenyl ether with a selectivity factor of E=160 in favor of the (S)-diol compared to WT ANEH characterized by E=4.6. Expression of this mutant is 50 times higher than that of WT ANEH. It also serves as an excellent stereoselective catalyst in the hydrolytic kinetic resolution and desymmetrization of several other structurally diverse epoxides. Copyright

Full text of DOI:10.1002/cbic.201100078

DOI: 10.1002/cbic.201100078
Manipulating the Expression Rate and Enantioselectivity
of an Epoxide Hydrolase by Using Directed Evolution
Manfred T. Reetz*[a, b] and Huabao Zheng[a]
Dedicated to Carmen Najera on the occassion of her 60th birthday
We describe here a strategy to improve the expression efficien-
cy and enantioselectivity of Aspergillus niger epoxide hydrolase
(ANEH) by directed evolution. Based on a blue-colony screen-
ing system using the LacZa (b-galactosidase a peptide) com-
plementation solubility reporter, several ANEH variants out of
15000 transformants from a random-mutagenesis library were
identified that show improved recombinant expression in
E. coli. Among them, Pro221Ser was subsequently used as a
template for iterative saturation mutagenesis (ISM) at sites
around the ANEH binding pocket. Following four rounds of
ISM, a highly enantioselective mutant was identified that cata-
lyzes the hydrolytic kinetic resolution of racemic glycidyl
phenyl ether with a selectivity factor of E=160 in favor of the
(S)-diol compared to WT ANEH characterized by E=4.6. Expres-
sion of this mutant is 50 times higher than that of WT ANEH. It
also serves as an excellent stereoselective catalyst in the hydro-
lytic kinetic resolution and desymmetrization of several other
structurally diverse epoxides.
Introduction
Directed evolution, or laboratory evolution as it is sometimes
called, permits the manipulation of essentially any property of
enzymes, including substrate acceptance, stereoselectivity and
stability.[1] The most popular gene-mutagensis methods are
error-prone polymerase chain reaction (epPCR), saturation mu-
tagenesis, and DNA shuffling. The application of these molecu-
lar biological techniques can always be expected to lead to
some degree of catalyst improvement, depending upon the
amount of time that is invested in exploring protein sequence
space and in evaluating the respective mutants. Because the
bottleneck of laboratory evolution is the screening (selection)
step,[2] interest in devising better methods and strategies for
generating higher-quality libraries has increased in recent
years.[1,3,4] We have defined “library quality” in terms of the fre-
quency of improved mutants in a given library and the actual
degree of catalyst improvement that these hits induce.[1a,4]
High quality thus means less screening effort. Our contribution
regarding methodology development in directed evolution is
iterative saturation mutagenesis (ISM).[1a,4,5] It is a knowledge-
driven approach according to which sites comprising one or
more amino acid positions are randomized by using saturation
mutagenesis followed by a screening step in which the respec-
tive libraries are assayed for a given enzyme property. A hit in
one library is then used as a template to perform saturation
mutagenesis at another site, and the process is continued until
the desired degree of catalyst improvement has been ach-
ieved. In a given ISM Scheme many pathways can be chosen,
but it is not necessary to explore all of them.[1a] However, the
choice of appropriate sites is crucial. For this purpose struc-
ture-based guidelines have been proposed that depend upon
the protein property to be manipulated, namely stereoselectiv-
ity,[5] substrate acceptance (rate),[5] or thermostability.[6] In the
case of substrate scope and/or stereoselectivity, we have pro-
posed the combinatorial active-site saturation test (CAST), ac-
cording to which sites around the binding pocket are chosen
for ISM.[1a] The first example of ISM based on CASTing was the
enhancement of enantioselectivity of the epoxide hydrolase
from Aspergillus niger (ANEH) as a catalyst in the hydrolytic ki-
netic resolution of glycidyl phenyl ether (rac-1), the selectivity
factor favoring (S)-2 increasing from E=4.6 (wild-type, WT) to
E=115 in five ISM steps (Scheme 1).[5a] Only small libraries
Scheme 1.
were necessary; the total number of transformants screened
was less than 20000. No tradeoff regarding thermostability
was observed. Subsequently, the underlying reason for the ob-
served efficacy was traced to the occurrence of pronounced
cooperative (synergistic) effects operating between sets of mu-
tations at each evolutionary step, as demonstrated by the ex-
[a] Prof. Dr. M. T. Reetz, Dr. H. Zheng
Department of Synthetic Organic Chemistry
Max-Planck-Institut fꢀr Kohlenforschung
Kaiser-Wilhelm-Platz 1, 45470 Mꢀlheim an der Ruhr (Germany)
Fax: (+49)208-306-2985
E-mail: reetz@mpi-muelheim.mpg.de
[b] Prof. Dr. M. T. Reetz
Fachbereich Chemie, Philipps-Universitꢁt Marburg
Hans-Meerwein-Strasse, 35032 Marburg (Germany)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201100078.
ChemBioChem 2011, 12, 1529 – 1535
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1529

M. T. Reetz et al.
perimentally generated fitness landscape and the quantitative
analysis of epistatic effects along a given pathway.[4b] Compari-
son with traditional approaches based on epPCR as a “shot-
gun” method in this particular case[7] underscored the superior-
ity of ISM, as in several other studies.[1a,5b–d]
In spite of the success of ISM in improving the stereoselec-
tivity of ANEH, the system proved to have limited practical
application due to the relatively low expression rate of the
enzyme in E. coli.[5a,7] The efficiency of protein expression in
heterologous hosts depends upon many factors, which are
indeed complex.[8] It is therefore not surprising that numerous
approaches have been described in the quest to boost the ex-
pression efficiency of proteins, including directed evolution.[1,9]
In many cases of observed expression improvement, protein
engineering actually focused on a different parameter, for ex-
ample, rate or stability, the induced mutations leading (seren-
dipitously?) to higher expression.[9] Indeed, when focusing on
rate using some kind of analytical technique during the screen-
ing of supernatants in the wells of microtiter plates, the appa-
rently higher rate may in fact be due partially or solely to in-
creased amounts of protein, brought about by the particular
mutations.[1] Careful analysis of the results is necessary to pin-
point all effects, as shown by several studies.[9] Therefore, if di-
rected evolution is to be targeted toward enhancing expres-
sion rate, then a screening process is necessary which is specif-
ic for identifying this particular parameter (“you get what you
screen for”).[1,2,9] In the present study we take this approach in
an effort to improve the expression efficiency of ANEH in
E. coli. In a second step, one of the improved mutants is em-
ployed as a starting point for manipulating stereoselectivity in
the test reaction of rac-1 using CASTing-based ISM.
Figure 2. A blue-colony screening system was used to identify ANEH mu-
tants leading to higher expression. Arrow shows a hit in LB agar plate con-
taining X-gal.
By using this screening system (Supporting Information), var-
iants with improved expression can be identified on LB agar
plates containing X-gal. Blue colonies indicate mutants leading
to high expression (Figure 2). This efficient pre-screen was sub-
sequently checked by performing expression on a large scale
and by quantifying efficiency.
epPCR at a low mutation rate was chosen as the mutagene-
sis method,[1] which addresses the whole enzyme ANEH ran-
domly (although some bias can be expected). A total of 15000
transformants were screened by using the blue-colony screen-
ing system. Fourteen variants having one or more mutations
were identified by picking blue colonies, and these were evalu-
ated quantitatively for b-gal activity (Table 1). This data was ob-
tained with pMALC4X; the fused tags were removed when
ANEH and its mutants were expressed in the pET system.
During the expression screening process, the b-galactosidase
activities were measured when the ANEH gene was fused with
the lacZ fragment. As can be seen, several hits were identified
Results and Discussion
Enhancing expression efficiency of ANEH
Because it is not obvious how to apply ISM when wanting to
enhance expression efficiency, we turned to epPCR,[1] a prereq-
uisite being the establishment of an appropriate screening
system.[2] Several high-throughput screening methods based
on fusion reporter systems are known, such as the green fluo-
rescent protein (GFP) folding report method,[10] the N-terminal
fusion system with chloramphenicol acetyltransferase[11] and
the LacZa (b-galactosidase a peptide) complementation solu-
bility reporter assay.[12] We opted for a screening protocol
based on structural complementation between the a- and w-
fragments of b-galactosidase (b-gal; Figure 1). It should be
pointed out that the deleted lacZ is provided by the E. coli
host. In this study E. coli DH5a was used.
Table 1. ANEH mutants detected in the initial epPCR library showing en-
hanced expression rate.
Mutants[a] Mutation
b-Gal activity [U/OD600]
1
2
3
4
5
6
7
8
9
EH87
EH94
Arg219Trp/Ala233Thr
Arg219Gly
Pro43His/Arg219Met
Ala220Val/Phe340/Leu/Val415Ala
Ala220Thr/Ser226Pro
Ala220Val
29.1
31
33
EH101
EH103
EH107
EH116
EH117
EH126
EH135
8.0
18.5
17.6
25.2
10.2
22.3
16.7
33.6
47.1
45.3
30.6
5.8
Arg219Trp
Pro222Thr/Ile262Val
Asp161Gly/Arg219Gly
Thr33Asn/Val197Ile/Pro222Ser
Val211Ile/Arg219Trp
Pro221Ser
10 EH145
11 EH146
12 EH147
13 EH150
14 EH153
15 WT
Arg219Lys
Pro222Leu
Figure 1. The complementary construct of fused ANEH-linker-a fragment in
the pMALC4X plasmid.
1530
www.chembiochem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1529 – 1535

Manipulating an Epoxide Hydrolase
by this process, the best one was mutant EH147, which is char-
acterized by a single amino acid exchange Pro221Ser.
canonical amino acids, this would entail nine randomization li-
braries, each requiring about 100 transformants to be screened
for 95% library coverage assuming the absence of amino acid
bias.[1a,4] In more recent studies, however, we discovered that
randomization at sites composed of more than one amino acid
position constitutes the better strategy,[5c–e] which was traced
to the occurrence of cooperative effects operating between
the newly introduced amino acids within a given site and
between sets of mutations in the ISM process.[4b] To keep the
screening effort as low as possible, we decided to delete one
of the residues (Ala217) from further consideration and fo-
cused on the remaining eight, which were grouped into four
sites, each composed of two amino acid positions: A (Leu215/
Arg219), B (Leu349/Cys350), C (Thr317/Thr318), and D (Phe244/
Leu249; Figure 3B). These sites are very similar, but not quite
identical to the previously utilized CAST sites (note that the
numbering system of the sites is different from the previous
study).[5a]
The results are encouraging, but the expression of the ANEH
mutants in the original plasmid pMALC4XEH is still not ideal,
as indicated by SDS-PAGE analyses. To boost the expression
further, pET22b was used under the action of promoter T7. All
variants were cloned into pET22b and induced with 0.1 mm
IPTG. Pronounced differences in SDS-PAGE were observed.
Unfortunately, many of these variants expressed in pET22b
proved to be insoluble. Therefore, to increase the solubility of
the target enzyme, chaperone plasmids were applied, in these
cases by using mutants EH117 (Arg219Trp) and EH147
(Pro221Ser). Of the five kinds of chaperone plasmids tested,
the plasmid of pKJE7 expressing chaperone dnaK-dnaJ-grpE
significantly improved the solubility of variant EH147 (Support-
ing Information), which was then chosen for further studies.
Mutation Pro221Ser is in a loop far removed from the catalytic
triad Asp192/Tyr251/Tyr314. In the hydrolytic kinetic resolution
of rac-1, mutant EH147 was found to be only slightly S-selec-
tive (E=5), which is essentially identical to the performance of
WT ANEH (E=4.6). This set the stage for improving the stereo-
selectivity of mutant EH147.
When choosing NNK codon degeneracy, requiring in each li-
brary the screening of about 3000 transformants for 95% cov-
erage,[4b,6] the total screening effort would amount to about
12000 measurements. Because this does not include any ISM
steps and already exceeds the screening limit arbitrarily set by
us, we opted for NDT codon degeneracy (D: adenine/guanine/
thymine; T: thymine), encoding 12 amino acids (Phe, Leu, Ile,
Val, Tyr, His, Asn, Asp, Cys, Arg, Ser, Gly). They constitute a bal-
anced mixture of building blocks having polar, nonpolar,
charged, uncharged, aromatic, and nonaromatic side chains.
NDT requires for 95% library coverage the screening of only
430 transformants.[4b,6] We settled for 480, which means that
the total screening effort involves only 1920 transformants. As
a first step, we focused on site D and performed randomiza-
tion by using the QuikChange protocol[14] but using NDT
codon degeneracy.[4a,6] After screening 480 transformants for
enantioselectivity of the kinetic resolution of rac-1, an im-
proved mutant EH181 (E=7) was identified that was character-
ized by two new mutations Phe244Cys/Leu249Phe. At this
stage a decision had to be made regarding further mutagene-
sis experiments, that is, to continue with EH181 in an ISM pro-
cess or to first generate the other three possible randomization
libraries at sites A, B, and C. Because the libraries are quite
small, it really did not matter. We opted for ISM exploration by
starting from EH181. When using the gene of this variant as a
template, either site A, B, or C can be chosen in the next ran-
domization experiment. We arbitrarily chose pathway D!A!
C!B. As Figure 4 reveals, this ISM pathway provides a highly
enantioselective mutant (EH222) displaying an E value of 160
in favor of (S)-2. To our surprise, saturation mutagenesis at site
C led to the insertion of extra residues (Pro-Thr-Ala-Ser-Ala-Pro-
His-Thr-Tyr-Arg-Glu-Phe-Ile). Unintended insertions may well
occur during the PCR process as part of the QuikChange proto-
col, but are hardly ever reported.[1,14] Indeed, in most cases the
likelihood of observing this phenomenon is low because it
usually has an adverse effect and will therefore not be detect-
ed in the screening process. Rather than discarding the result
and turning to other saturation mutagenesis libraries, we
prefer to report this observation. In the same library the
Enhancing enantioselectivity of ANEH
To choose optimal CAST sites for saturation mutagenesis, we
first performed induced fit docking of the model substrate rac-
1
by using the “mutated” structure of variant EH147
(Pro221Ser) based on the crystal structure of WT ANEH.[13] Nine
residues were identified for potential randomization, namely
Leu215, Ala217, Arg219, Phe244, Leu249, Thr317, Thr318,
Leu349, and Cys350 (Figure 3A). The next step required a deci-
Figure 3. A) CAST analysis of ANEH mutant Pro221Ser based on the crystal
structure of WT ANEH,[13] into which mutation Pro221Ser was “added”; in-
duced fit docking of rac-1 (purple) then led to the identification of nine res-
idues for possible saturation mutagenesis. B) Grouping eight of the nine
identified residues into four potential CAST sites, A (Leu215/Arg219), B
(Leu349/Cys350), C (Thr317/Thr318), and D (Phe244/Leu249).
sion concerning how to group these amino acid positions into
appropriate sites, keeping in mind the screening effort. One
could opt for nine single-residue sites, a strategy that we had
successfully applied in the directed evolution of the enoate-re-
ductase YqjM.[5b] Using NNK codon degeneracy (N: adenine/cy-
tosine/guanine/thymine; K: guanine/thymine) encoding all 20
ChemBioChem 2011, 12, 1529 – 1535
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1531

M. T. Reetz et al.
stereochemical results on a molecular level with any degree of
confidence, extensive molecular dynamics (MD) simulations
need to be carried out, an important topic for future work.
Moreover, the unexpected loop insertion in library C and its
influence on stereoselectivity also have to be considered.
To check whether the best mutant EH222 (E=160) maintains
the improved expression found in the initial mutant EH147,
the respective control experiments were performed. We ob-
tained 9.1 mg enzyme per gram of wet pellet by following af-
finity purification, which is 50-times higher than in the case of
WT ANEH.
Investigation of substrate scope
In an effort to see how the best mutant EH222 performs as a
catalyst in the hydrolytic desymmetrization of meso-configurat-
ed epoxides,[5d] compounds 3a, 3b, and 5 were tested as sub-
strates. Table 3 reveals that the substrate scope of mutant
Table 3. Catalytic performance of mutant EH222 compared to WT ANEH
under standard conditions (4 h).
Figure 4. Multiple pathways explored in the quest to boost the expression
rate of WT ANEH by using epPCR (red pathway) and enantioselectivity of the
respective ANEH mutant as a catalyst in the hydrolytic kinetic resolution of
rac-1 by using ISM (green pathways). Extra residues (ProThrAlaSerAlaPro-
HisThrTyrArgGluPheIle) appear in EH204 and EH222, indicating an inserting
sequence produced by QuikChange.[14]
WT ANEH
Conv. [%]
EH222
Subtrates
ee [%]
Conv. [%]
ee [%]
3a
3b
5
0
82
0
39
8
70
94
72
46
15
94 (S,S)
97 (S,S)
95 (S,S)
56 (1S,2S)
46 (1R,2R)
79 (S,S)
rac-7
rac-9
0
0
second best variant in terms of stereoselectivity was identified,
namely EH202 characterized by E=51, which is also a notable
degree of enantioselectivity. However, upon continuing ISM by
subsequently visiting site B, no significant improvement was
detected. Table 2 and Figure 4 summarize the results, which re-
quired the screening of a total of only 2880 transformants. Fur-
ther exploration of ISM pathways is possible, but because the
goal of acceptable stereoselectivity had been reached, we de-
cided at this point to leave this to future work.
EH222 is quite broad, and the activity and improved enantiose-
lectivity are generally distinctly higher than WT ANEH. More-
over, in hydrolytic kinetic resolution of rac-7 and rac-9
(Scheme 2),[5d] mutant EH222 again shows much better perfor-
mance than WT ANEH, which behaves stereorandomly in these
cases.
As delineated earlier, induced-fit docking was performed to
identify appropriate CAST residues. However, to interpret the
Table 2. Typical ANEH mutants arising from an arbitrarily chosen ISM process D!A!C!B starting from mutant EH147 (Pro221Ser) produced by epPCR
for higher expression, their performance as catalysts was tested in the hydrolytic kinetic resolution of rac-1. In all cases NDT codon degeneracy was em-
ployed as part of the QuikChange protocol14 encoding an amino acid alphabet of 12 amino acids (see text).
Site Mutant Mutations
ee [%]
(S)
Conv.
[%]
E
value
WT
D
EH147 Pro221Ser
EH181 Pro221Ser/Phe244Cys/Leu249Phe
EH182 Pro221Ser/Leu249Ser
EH183 Pro221Ser/Phe244Tyr/Leu249Arg
EH185 Pro221Ser/Phe244His/Leu249His
EH191 Pro221Ser/Phe244Cys/Leu249Phe/Leu215Phe
EH202 Pro221Ser/Phe244Cys/Leu249Phe/Leu215Phe/Thr317His/Thr318Thr
EH203 Pro221Ser/Phe244Cys/Leu249Phe/Leu215Phe/Thr317Ile/Thr318Ser
EH204 Pro221Ser/Phe244Cys/Leu249Phe/Leu215Phe/Thr317/Thr318(PheValProThrAlaSerAlaProHisThrTyrArgGluPheIle)
EH205 Pro221Ser/Phe244Cys/Leu249Phe/Leu215Phe/Thr317Val/Thr318Ile
EH207 Pro221Ser/Phe244Cys/Leu249Phe/Leu215Phe/Thr317Val/Thr318Ser
EH222 Pro221Ser/Phe244Cys/Leu249Phe/Leu215Phe/Thr317Phe/Thr318Val(ProThrAlaSerAlaProHisThrTyrArgGluPheIle)-
Leu349Val
54
69
55
70
66
86
93
92
96
85
92
97
40
23
53
23
24
27
40
35
41
43
18
45
5
7
6
7
6
18
51
41
91
24
29
160
A
C
B
1532
www.chembiochem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1529 – 1535

Manipulating an Epoxide Hydrolase
Qiagen. Chaperone plasmid set was bought from TaKaRa. All com-
mercial chemicals were purchased from Sigma–Aldrich. Commer-
cially unavailable chemicals were synthesized and characterized by
GC-MS and NMR spectroscopy. Terrific Broth (TB; per liter) contain-
ing: tryptone (12 g), yeast extract (24 g), glycerol (4 mL), KH2PO4
(2.31 g), K2HPO4 (12.54 g). Lysis buffer was prepared by using po-
tassium phosphate buffer (50 mm, pH 7.4) containing lysozyme
(1 mgmLꢀ1, AppliChem) and DNaseI (0.1 mgmLꢀ1, 4066 Umgꢀ1
AppliChem).
,
Recombinant plasmid construction: The primers of UpEH_linker_
pMALC4X and DownEH_linker_pMALC4X were used to amplify the
gene of epoxide hydrolase of Aspergillus niger (ANEH). The “GSAG-
SAAGSGAS” peptide, which is encoded by 5’-ggatccgctggctccg-
ctgctggttctggcgcaagc-3’ was used as a linker between ANEH and
the a-fragment of b-galactosidase, the PCR was performed for 26
cycles at 958C, 30 s; 528C, 30 s and 728C for 1.5 min. The primers
of upVector_pMALC4X and DownVector_pMALC4X were used to
amplify the vector fragment of pMALC4X. The PCR was performed
for 26 cycles at 958C, 30 s; 528C, 30 s and 728C for 6 min. The orig-
inal templates were digested by DpnI twice within 6 h. The purified
PCR products of the ANEH gene and vector fragment of pMALC4X
were ligated according to enzymatic assembly procedures.[15] The
assembly products were then transformed into chemically compe-
tent DH5a cells. The recombinant plasmid was confirmed by se-
quencing.
Scheme 2.
Conclusions
Construction and screening of epPCR-based random mutagene-
sis library for expression efficiency: Random mutagenesis library
was constructed by GeneMorph II Random Mutagenesis Kit. The
primers of upEPgenepMALC4X and downEPgenepMALC4X were
used to amplify the ANEH gene. A low range of mutation frequen-
cy (0 ~4.5 mutations per kb) was utilized. Twenty-five cycles were
run by PCR at 958C, 1 min, 608C, 1 min, and 728C for 2 min. The
primers of upEPpMALC4X and downEPpMALC4X were used to
amplify the vector fragment of pMALC4X by thirty cycles at 958C,
1 min, 608C, 1 min, and 728C for 6 min. The original templates
were digested by DpnI twice within 6 h. The purified PCR products
of the ANEH gene and vector pMALC4X were ligated based on en-
zymatic assembly procedures. The assembly product was trans-
formed to the electroporation-competent DH5a cells. The trans-
formed product was plated on LB agar plates containing
50 mgmLꢀ1 carbenicillin, 20 mgmLꢀ1 X-gal, and 0.1 mm IPTG. The
plates were kept at 378C in an incubator overnight and moved
into a 308C incubator for about 15 h to allow the blue colonies to
appear.
We have demonstrated that it is possible to exploit the tech-
niques of directed evolution[1] to achieve two very different
goals, namely increasing the expression rate and enhancing
the stereoselectivity of an enzyme. Both parameters are crucial
in biotechnology. In principle, it is conceivable that either one
of the two properties can be optimized first followed by the
second optimization. In the present study we focused on ex-
pression first, and then turned to the improvement of enantio-
selectivity. As a test case the epoxide hydrolase from Aspergil-
lus niger (ANEH) was used as a catalyst in the hydrolytic kinetic
resolution of glycidyl phenyl ether (rac-1), WT leading to E=
4.6 (S).[5a,7] Expression in E. coli was notably improved by
screening a 15000-membered epPCR-based library, providing
mutant Pro221Ser. The latter was subsequently used in a limit-
ed number of iterative saturation mutagenesis experiments
(less than 3000 transformants screened), leading to several
enantioselective mutants characterized by selectivity factors
ranging between 51 and 160 in favor of the (S)-diol. The ex-
pression efficiency of the best mutant was found to be 50
times higher than that of WT ANEH. Thus, our strategy proved
to be successful, but it remains to be seen if the opposite
order of events delivers similar results, or whether it is less or
perhaps even more efficient. In future studies the reasons for
catalyst improvement also need to be illuminated on a molec-
ular level.
Measurement of b-galactosidase activity: WT and mutants ob-
tained from the blue-colony screening system were cultured in LB
medium containing 50 mgmLꢀ1 carbenicillin at 378C. The expres-
sion of target protein was induced with 0.1 mm IPTG for 4 h at
308C when the OD of the culture reached 1.0. The activity of b-gal-
actosidase was measured by an adapted method by using ONPG
(O-nitrophenyl-b-d-galactopyranoside) as a substrate.[12] An aliquot
of expression culture (1 mL) was centrifuged and washed twice by
using potassium phosphate buffer (50 mm, pH 7.0) and resuspend-
ed in potassium phosphate buffer (1 mL). An aliquot of resuspend-
ed cells (0.5 mL) was taken for reaction. Reactions were initiated by
adding O-nitrophenyl-b-d-galactopyranoside (ONPG) solution
(0.25 mL; 4.0 mgmLꢀ1 dissolved in potassium phosphate buffer)
and incubated at 378C for 10 min. Reactions were quenched by
adding 1m Na2CO3 (0.25 mL). After centrifugation, the superna-
tant’s absorption at 420 nm was measured. O-Nitrophenol (ONP)
was used to make standard curve. One unit of enzyme was equiva-
Experimental Section
General: The restriction enzymes were purchased from New Eng-
land Biolabs. KOD Hot Start DNA Polymerase was obtained from
Novagen. The oligonucleotides were synthesized by Invitrogen.
Plasmid preparation kit and PCR purification kit were bought from
ChemBioChem 2011, 12, 1529 – 1535
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1533

M. T. Reetz et al.
lent to 1 mmol ONP liberated from ONPG per minute per OD600 of
culture.
washed twice by using potassium phosphate buffer (50 mm,
pH 7.4). The pellets were resuspended with the same buffer at the
ratio of 8 mL potassium phosphate buffer per gram of cells. The
cells were disrupted by sonication (Bandelin, 120 seconds, 40%
pulse, on ice broth), and a clear supernatant was obtained by cen-
trifugation at 10 000 rpm and 48C for 1 h. Recombinant epoxide
hydrolases were purified by affinity chromatography (HisTrap HP,
5 mL, GE Healthcare). The impurities were removed by ten column
volumes of washing buffer (50 mm potassium phosphate buffer,
pH 7.4, 500 mm NaCl and 40 mm imidazole) at 2 mLminꢀ1 flow
rate. His-tagged epoxide hydrolases were obtained by using three
column volumes of elution buffer (50 mm potassium phosphate
buffer, pH 7.4, 500 mm NaCl and 500 mm imidazole). The elution
fractions were concentrated by Amicon Ultra-15 (10 K Nominal Mo-
lecular Weight Limit, Minipore). PD-10 desalting column (5 mL, GE
Healthcare) was then utilized to remove imidazole from concen-
trated enzyme solution.
Construction and screening of iterative saturation mutagenesis
libraries for enantioselectivity: The mutant gene of ANEH from
random mutagenesis library was cloned into pET22b to construct
and screen ISM libraries. The primers of up_22bEH_ANEH and
down_22bEH_ANEH were used to amplify the mutated ANEH
gene. The primers of up_22bEH_Vector and down_22bEH_Vector
were used to amplify vector pET22b. The cloning steps were car-
ried out based on enzymatic assembly procedures described as
above. The constructed plasmids were confirmed by sequencing.
Electroporation-competent cells containing five different chaper-
one plasmids (chap1, pG-KJE8; chap2, pGro7; chap3, pKJE7; chap4,
pG-Tf2; chap5, pTf16) were prepared. Mutant ANEH cloned into
pET22b were applied to transform different competent cells of
BL21Gold (DE3) containing each different chaperone plasmid. The
expression of mutant ANEH together with chaperones was per-
formed according to TaKaRa protocols.
Acknowledgements
Iterative saturation mutagenesis libraries were constructed by
QuikChange. Twenty-two cycles were run by PCR at 958C, 30 s,
608C, 30 s, and 728C for 3 min. The parent template was digested
by DpnI twice within 6 h. The purified products were transformed
electroporation-competent cells of BL21Gold (DE3).
This work was supported by the Deutsche Forschungsgemein-
schaft as part of SPP1170 “Directed Evolution to Optimize and
Understand Molecular Biocatalysis” (Project Re 359/13–1). We
thank Despina Bougioukou for helpful discussions.
Transformants grown on LB agar plates were picked and cultured
overnight in 96-deep-well plates containing LB (800 mL) and
50 mgmLꢀ1 carbenicillin, 20 mgmLꢀ1 chloramphenicol, and
2 mgmLꢀ1 l-arabinose at 378C with shaking (800 rpm). An aliquot
of culture (100 mL) was transferred into 96-deep-well plates con-
taining TB (700 mL), 0.1 mm IPTG, 50 mgmLꢀ1 carbenicillin,
20 mgmLꢀ1 chloramphenicol, and 2 mgmLꢀ1 l-arabinose. The ex-
pression of transformants was performed for 6 h and at 308C.
Keywords: casting · directed evolution · enantioselectivity ·
epoxide hydrolase · expression rate
[1] Reviews of directed evolution: a) M. T. Reetz, Angew. Chem. 2011, 123,
144–182; Angew. Chem. Int. Ed. 2011, 50, 138–174; b) T. W. Johannes,
H. Zhao, Curr. Opin. Microbiol. 2006, 9, 261–267; c) S. Lutz, U. T. Born-
scheuer, Protein Engineering Handbook, Vols. 1–2, Wiley-VCH, Weinheim,
2009; d) N. J. Turner, Nat. Chem. Biol. 2009, 5, 567–573; e) C. Jꢂckel, P.
Kast, D. Hilvert, Annu. Rev. Biophys. Biomol. Struct. 2008, 37, 153–173;
f) S. Bershtein; D. S. Tawfik, Curr. Opin. Chem. Biol. 2008, 12, 151–158;
D. S. Tawfik, Curr. Opin. Chem. Biol. 2008, 12, 151–158; g) P. A. Romero,
F. H. Arnold, Nat. Rev. Mol. Cell Biol. 2009, 10, 866–876; h) M. T. Reetz in
Asymmetric Organic Synthesis with Enzymes (Eds.: V. Gotor, I. Alfonso, E.
Garcꢃa-Urdiales), Wiley-VCH, Weinheim, 2008; pp. 21–63; i) L. G. Otten,
F. Hollmann, I. W. C. E. Arends, Trends Biotechnol. 2010, 28, 46–54;
j) A. V. Shivange, J. Marienhagen, H. Mundhada, A. Schenk, U. Schwane-
berg, Curr. Opin. Chem. Biol. 2009, 13, 19–25.
An aliquot of expression culture (150 mL) was transferred into a 96-
deep-well plate, then phosphate buffer mixed with 13.6 mm rac-1
(containing 8.1% v/v MeCN; 600 mL) was added. The reaction was
performed at 308C with shaking at 500 rpm for 1 h. The plates
were centrifuged for 20 min and supernatant (150 mL) was trans-
ferred into a new 96-deep-well plate filled with EtOH (150 mL). The
plates were again centrifuged at 48C for 1 h. An aliquot of super-
natant (150 mL) was transferred into a microtiter plate. These plates
were sealed with aluminum stick seal and subjected to a medium-
throughput chiral HPLC assay.
[2] a) J.-L. Reymond, V. S. Fluxꢄ, N. Maillard, Chem. Commun. 2009, 34–46;
b) J.-L. Reymond, Enzyme Assays, High-Throughput Screening, Genetic
Selection and Fingerprinting, Wiley-VCH, Weinheim, 2006.
Investigation of substrate scope: The typical reaction procedure
for enzymatic hydrolysis of epoxide is as follows: WT and ANEH
mutants were expressed in TB at 308C for 4 h with 0.1m IPTG and
2 mgmLꢀ1 l-arabinose as inducers. The pellets of each 1.0 mL of
culture were harvested by centrifuging. The cells were broken by
the addition of 400 mL lysis buffer. Then 5.0 mm of epoxides
(200 mm stock solution in acetonitrile) were applied to each re-
action for 10 h at 308C with shaking (400 rpm). The products and
residual substrates were extracted with decane (1 mm) in EtOAc
(400 mL). For conversion measurement, an aliquot of solvent layer
(1 mL) was analyzed by GC. Then, the extraction was carried out by
using EtOAc (3ꢁ400 mL). The EtOAc layer (ca. 1200 mL) was collect-
ed and concentrated by speed vacuum to a final volume of
150 mL. An aliquot of EtOAc (1 mL) was analyzed by chiral GC
column.
[3] a) S. Lutz, W. M. Patrick, Curr. Opin. Biotechnol. 2004, 15, 291–297;
b) R. J. Fox, G. W. Huisman, Trends Biotechnol. 2008, 26, 132–138; c) T. S.
Wong, D. Roccatano, M. Zacharias, U. Schwaneberg, J. Mol. Biol. 2006,
355, 858–871; d) J. D. Bloom, M. M. Meyer, P. Meinhold, C. R. Otey, D.
MacMillan, F. H. Arnold, Curr. Opin. Struct. Biol. 2005, 15, 447–452.
[4] a) M. T. Reetz, D. Kahakeaw, R. Lohmer, ChemBioChem 2008, 9, 1797–
1804; b) M. T. Reetz, J. Sanchis, ChemBioChem 2008, 9, 2260–2267.
[5] a) M. T. Reetz, L.-W. Wang, M. Bocola, Angew. Chem. 2006, 118, 1258–
1263; Erratum: p. 2556; Angew. Chem. Int. Ed. 2006, 45, 1236–1241;
Erratum: p. 2494; b) D. J. Bougioukou, S. Kille, A. Taglieber, M. T. Reetz,
Adv. Synth. Catal. 2009, 351, 3287–3305; c) M. T. Reetz, S. Prasad, J. D.
Carballeira, Y. Gumulya, M. Bocola, J. Am. Chem. Soc. 2010, 132, 9144–
9152; d) H. Zheng, M. T. Reetz, J. Am. Chem. Soc. 2010, 132, 15744–
15751; e) M. T. Reetz, S. Wu, J. Am. Chem. Soc. 2009, 131, 15424–15432.
[6] M. T. Reetz, J. D. Carballeira, Nat. Protoc. 2007, 2, 891–903.
[7] M. T. Reetz, C. Torre, A. Eipper, R. Lohmer, M. Hermes, B. Brunner, A. Mai-
chele, M. Bocola, M. Arand, A. Cronin, Y. Genzel, A. Archelas, R. Furstoss,
Org. Lett. 2004, 6, 177–180.
[8] See for example: a) S. C. Makrides, Microbiol. Rev. 1996, 60, 512–538;
b) G. Marsischky, J. LaBaer, Genome Res. 2004, 14, 2020–2028; c) E. Anti-
Expression and purification: The expression of mutant and WT
was induced by using 0.1 mm IPTG in 300 mL TB containing
50 mgmLꢀ1 carbenicillin at 308C, 20 mgmLꢀ1 chloramphenicol and
2 mgmLꢀ1 l-arabinose for 6 h. The cells were harvested and
1534
www.chembiochem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2011, 12, 1529 – 1535

Manipulating an Epoxide Hydrolase
pov, A. E. Cho, K. D. Wittrup, A. M. Klibanov, Proc. Natl. Acad. Sci. USA
viron. Sci. 2007, 20, 113–118; n) P. K. R. Kambam, D. J. Sayut, Y. Niu, D. T.
Eriksen, L. Sun, Biotechnol. Bioeng. 2008, 101, 263–272; o) C. A. Sarkar, I.
Dodevski, M. Kenig, S. Dudli, A. Mohr, E. Hermanns, A. Plꢆckthun, Proc.
Natl. Acad. Sci. USA 2008, 105, 14804–14813.
ˇ
2008, 105, 17694–17699; d) A. Kern, E. Tilley, I. S. Hunter, M. Legisa, A.
Glieder, J. Biotechnol. 2006, 129, 6–29.
[9] Examples of manipulating the expression rate of enzymes by protein
engineering: a) N. J. Silman, M. A. Carver, C. W. Jones, J. Gen. Microbiol.
1991, 137, 169–178; b) C. Bessler, J. Schmitt, K.-H. Maurer, R. D. Schmid,
Protein Sci. 2003, 12, 2141–2149; c) T. Bulter, M. Alcalde, V. Sieber, P.
Meinhold, C. Schlachtbauer, F. H. Arnold, Appl. Environ. Microbiol. 2003,
69, 987–995; d) S. R. Leong, J. C. C. Chang, R. Ong, G. Dawes, W. P. C.
Stemmer, J. Punnonen, Proc. Natl. Acad. Sci. USA 2003, 100, 1163–1168;
e) G. S. Waldo, Curr. Opin. Chem. Biol. 2003, 7, 33–38; f) V. G. H. Eijsink,
A. Bjørk, S. Gꢅseidnes, R. Sirevꢅg, B. Synstad, B. van den Burg, G. Vriend,
J. Biotechnol. 2004, 113, 105–120; g) F. Cedrone, S. Niel, S. Roca, T. Bhat-
nagar, N. Ait-Abdelkader, C. Torre, H. Krumm, A. Maichele, M. T. Reetz,
J. C. Baratti, Biocatal. Biotransform. 2003, 21, 357–364; h) I. Matsumura,
L. A. Rowe, Biomol. Eng. 2005, 22, 73–79; i) S. Y. McLoughlin, C. Jackson,
J.-W. Liu, D. Ollis, Protein Expr. Purif. 2005, 41, 433–440; j) C. Roodveldt,
A. Aharoni, D. S. Tawfik, Curr. Opin. Struct. Biol. 2005, 15, 50–56; k) R.
Woodyer, W. A. van der Donk, H. Zhao, Comb. Chem. High Throughput
Screening 2006, 9, 237–245; l) Y.-S. Kim, R. Bhandari, J. R. Cochran, J.
Kuriyan, K. D. Wittrup, Proteins Struct. Funct. Bioinf. 2006, 62, 1026–
1035; m) J.-W. Li, J.-L. Zheng, X.-W. Wang, M. Jin, F.-H. Chao, Biomed. En-
[10] G. S. Waldo, B. M. Standish, J. Berendzen, T. C. Terwilliger, Nat. Biotech-
nol. 1999, 17, 691–695.
[11] K. L. Maxwell, A. K. Mittermaier, J. D. Forman-Kay, A. R. Davidson, Protein
Sci. 1999, 8, 1908–1911.
[12] W. C. Wigley, R. D. Stidham, N. M. Smith, J. F. Hunt, P. J. Thomas, Nat. Bio-
technol. 2001, 19, 131–136.
[13] a) J. Y. Zou, B. M. Hallberg, T. Bergfors, F. Oesch, M. Arand, S. L. Mowbray,
T. A. Jones, Structure 2000, 8, 111–122; b) http://www.pdb.ccdc.cam.
ac.uk/pdb.
[14] Details of QuikChange Protocol of Stratagene: H. H. Hogrefe, J. Cline,
G. L. Youngblood, R. M. Allen, BioTechniques 2002, 33, 1158–1165.
[15] D. G. Gibson, L. Young, R.-Y. Chuang, J. C. Venter, C. A. Hutchison, H. O.
Smith, Nat. Methods 2009, 6, 343–345.
Received: February 1, 2011
Published online on May 12, 2011
ChemBioChem 2011, 12, 1529 – 1535
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
1535