Manipulating the stereoselectivity of limonene epoxide hydrolase by directed evolution based on iterative saturation mutagenesis

Article abstract of DOI:10.1021/ja1067542

Limonene epoxide hydrolase from Rhodococcus erythropolis DCL 14 (LEH) is known to be an exceptional epoxide hydrolase (EH) because it has an unusual secondary structure and catalyzes the hydrolysis of epoxides by a rare one-step mechanism in contrast to the usual two-step sequence. From a synthetic organic viewpoint it is unfortunate that LEH shows acceptable stereoselectivity essentially only in the hydrolysis of the natural substrate limonene epoxide, which means that this EH cannot be exploited as a catalyst in asymmetric transformations of other substrates. In the present study, directed evolution using iterative saturation mutagenesis (ISM) has been tested as a means to engineer LEH mutants showing broad substrate scope with high stereoselectivity. By grouping individual residues aligning the binding pocket correctly into randomization sites and performing saturation mutagenesis iteratively using a reduced amino acid alphabet, mutants were obtained which catalyze the desymmetrization of cyclopentene-oxide with stereoselective formation of either the (R,R)- or the (S,S)-diol on an optional basis. The mutants prove to be excellent catalysts for the desymmetrization of other meso-epoxides and for the hydrolytic kinetic resolution of racemic substrates, without performing new mutagenesis experiments. Since less than 5000 tranformants had to be screened for achieving these results, this study contributes to the generalization of ISM as a fast and reliable method for protein engineering. In order to explain some of the stereoselective consequences of the observed mutations, a simple model based on molecular dynamics simulations has been proposed.

Full text of DOI:10.1021/ja1067542

Published on Web 10/19/2010
Manipulating the Stereoselectivity of Limonene Epoxide
Hydrolase by Directed Evolution Based on Iterative Saturation
Mutagenesis
Huabao Zheng and Manfred T. Reetz*
Max-Planck-Institut fu¨r Kohlenforschung, Kaiser-Wilhelm-Platz 1,
45470 Mu¨lheim an der Ruhr, Germany
Received July 29, 2010; E-mail: reetz@mpi-muelheim.mpg.de
Abstract: Limonene epoxide hydrolase from Rhodococcus erythropolis DCL 14 (LEH) is known to be an
exceptional epoxide hydrolase (EH) because it has an unusual secondary structure and catalyzes the
hydrolysis of epoxides by a rare one-step mechanism in contrast to the usual two-step sequence. From a
synthetic organic viewpoint it is unfortunate that LEH shows acceptable stereoselectivity essentially only in
the hydrolysis of the natural substrate limonene epoxide, which means that this EH cannot be exploited as
a catalyst in asymmetric transformations of other substrates. In the present study, directed evolution using
iterative saturation mutagenesis (ISM) has been tested as a means to engineer LEH mutants showing
broad substrate scope with high stereoselectivity. By grouping individual residues aligning the binding pocket
correctly into randomization sites and performing saturation mutagenesis iteratively using a reduced amino
acid alphabet, mutants were obtained which catalyze the desymmetrization of cyclopentene-oxide with
stereoselective formation of either the (R,R)- or the (S,S)-diol on an optional basis. The mutants prove to
be excellent catalysts for the desymmetrization of other meso-epoxides and for the hydrolytic kinetic
resolution of racemic substrates, without performing new mutagenesis experiments. Since less than 5000
tranformants had to be screened for achieving these results, this study contributes to the generalization of
ISM as a fast and reliable method for protein engineering. In order to explain some of the stereoselective
consequences of the observed mutations, a simple model based on molecular dynamics simulations has
been proposed.
Introduction
and stereochemistry-determining step, followed by rapid hy-
drolysis of the enzyme-ester intermediate.1 Limonene epoxide
Epoxide hydrolases (EHs) are enzymes that can be used as
catalysts in the hydrolytic desymmetrization of meso-epoxides
and in the kinetic resolution of racemic substrates with formation
of enantiomerically pure or enriched vicinal diols.1 In some cases
they are attractive alternatives to the powerful Jacobsen salen-
cobalt catalysts.2 Most EHs react by a two-step mechanism in
which a short-lived covalent enzyme-ester is formed by
nucleophilic attack of an aspartate at one of the two C atoms of
the H-bonded and therefore activated epoxide moiety in the rate-
hydrolase (LEH) constitutes one of the rare exceptions,1 because
it catalyzes direct hydrolysis of epoxides by positioning and
activating water at the active site, thereby enabling smooth
nucleophilic attack with inversion of configuration at the epoxide
C atom.3,4 The crystal structure of LEH does not show the usual
R/ꢀ-fold found in most EHs but features a curved 6-stranded
mixed ꢀ-sheet surrounded by four helices, thereby shaping a
deep binding pocket.3b Further insight into the mechanism has
been provided by a recent QM/MM study.5 In nature, LEH
catalyzes the hydrolysis of limonene epoxide (1) present in
Rhodococcus erythropolis DCL14.3
(1) Reviews of epoxide hydrolases: (a) Archelas, A.; Furstoss, R. Curr.
Opin. Chem. Biol. 2001, 5, 112–119. (b) Morisseau, C.; Hammock,
B. D. Annu. ReV. Pharmacol. Toxicol. 2005, 45, 311–333. (c) Faber,
K.; Orru, R. V. A. In Enzyme Catalysis in Organic Synthesis, 2nd
ed.; Drauz, K., Waldmann, H., Eds.; Wiley-VCH: Weinheim, 2002;
pp 579-608. (d) Choi, W. J. Appl. Microbiol. Biotechnol. 2009, 84,
239–247. (e) Hwang, S.; Choi, C. Y.; Lee, E. Y. J. Ind. Eng. Chem.
(Seoul, Repub. Korea) 2010, 16, 1–6. (f) Widersten, M.; Gurell, A.;
Lindberg, D. Biochim. Biophys. Acta 2010, 1800, 316–326.
(2) (a) Jacobsen, E. N. Acc. Chem. Res. 2000, 33 (6), 421–431. (b) Schaus,
S. E.; Brandes, B. D.; Larrow, J. F.; Tokunaga, M.; Hansen, K. B.;
Gould, A. E.; Furrow, M. E.; Jacobsen, E. N. J. Am. Chem. Soc. 2002,
124, 1307–1315. (c) Jacobsen, E. N.; Kakiuchi, F.; Konsler, R. G.;
Larrow, J. F.; Tokunaga, M. Tetrahedron Lett. 1997, 38, 773–776.
(d) Kawthekar, R. B.; Kim, G.-J. HelV. Chim. Acta 2008, 91, 317–
332. (e) Ready, J. M.; Jacobsen, E. N. J. Am. Chem. Soc. 2001, 123,
2687–2688. (f) Ready, J. M.; Jacobsen, E. N. Angew. Chem. 2002,
114, 1432–1435. Angew. Chem., Int. Ed. 2002, 41, 1374-1377. (g)
Private communication of E. N. Jacobsen to M. T. Reetz (Sept 5,
2010).
Since this substrate is a fairly bulky trisubstituted epoxide,
one might expect this EH to be a viable catalyst for a wide
variety of structurally different epoxides. Unfortunately, enan-
tioselectivity is generally poor in such cases (E ) 3-11),4a
(3) (a) van der Werf, M. J.; Overkamp, K. M.; de Bont, J. A. M. J.
Bacteriol. 1998, 180, 5052–5057. (b) Arand, M.; Hallberg, B. M.;
Zou, J.; Bergfors, T.; Oesch, F.; van der Werf, M. J.; de Bont, J. A. M.;
Jones, T. A.; Mowbray, S. L. EMBO J. 2003, 22, 2583–2592.
(4) (a) van der Werf, M. J.; Orru, R. V. A.; Overkamp, K. M.; Swarts,
H. J.; Osprian, I.; Steinreiber, A.; de Bont, J. A. M.; Faber, K. Appl.
Microbiol. Biotechnol. 1999, 52, 380–385. (b) van der Werf, M. J.;
De Bont, J. A. M.; Swarts, H. J. Tetrahedron: Asymmetry 1999, 10,
4225–4230.
(5) Hopmann, K. H.; Hallberg, B. M.; Himo, F. J. Am. Chem. Soc. 2005,
127, 14339–14347.
9
15744 J. AM. CHEM. SOC. 2010, 132, 15744–15751
10.1021/ja1067542  2010 American Chemical Society

Stereoselectivity of Limonene Epoxide Hydrolase
A R T I C L E S
throughput ee assays based on standard automated GC or HPLC,
but such libraries need to be characterized by superior quality,
meaning a high frequency of hits showing pronounced catalyst
improvement.10
Our contribution to achieving expediency in this endeavor is
iterative saturation mutagenesis (ISM), which can be applied to
the control of stereoselectivity7,10 and to the enhancement of
thermostability and robustness in hostile organic solvents.10b
Appropriate sites in an enzyme, labeled A, B, C, D, etc., each
comprising one or more amino acid positions, are subjected to
saturation mutagenesis with formation of focused libraries. Sub-
sequently, the hits in the libraries are used as templates to perform
saturation mutagenesis at the respective other sites, and the process
is continued iteratively until the desired degree of catalyst improve-
ment has been achieved (for a 4-site ISM scheme, see Scheme S4
in the Supporting Information). When evolving stereoselectivity
and/or substrate scope (rate), the criterion for choosing appropriate
sites is based on the Combinatorial Active Site-Saturation Test
(CAST),7,10a,c,d according to which residues aligning the binding
pocket are considered. A further tool in “smart” library creation is
the use of reduced amino acid alphabets employing the appropriate
codon degeneracy, because this decimates the degree of oversam-
pling .7,10 In the present study we wanted to test the utility of ISM
in the evolution of stereoselective LEH mutants. At the outset of
the project, stringent requirements were defined by restricting the
size of the mutant libraries to a few thousand transformants, i.e.,
we wanted to know how far one can get by investing a minimum
amount of laboratory work. In contrast to several other EHs,12 LEH
has not been subjected to directed evolution previously.
which means that wild-type (WT) LEH can hardly be exploited
in synthetic organic chemistry. As an example, it catalyzes the
hydrolytic desymmetrization of cyclopentene-oxide (3) with
poor enantioselectivity (ee ) 14% in meager favor of (R,R)-4).
Hoping to obtain LEH mutants which catalyze the formation
of (R,R)-4 and (S,S)-4 on an optional basis with high enanti-
oselectivity, we decided to apply directed evolution.6,7 Since
industrial organic chemists need stereoselective catalysts with
a reasonable substrate scope,8 not just one for the reaction of a
single compound, we also planned to test the evolved mutants
in the reaction of other substrates without performing additional
mutagenesis experiments.
Directed evolution of stereoselective enzymes has emerged
as a prolific source of catalysts for asymmetric transformations
in synthetic organic chemistry.6,7 The bottleneck of laboratory
evolution is the screening step which is particularly challenging
when assaying stereoselectivity.9 Therefore, we and others have
recently stressed “quality, not quantity”.7,10,11 Small mutant
libraries of the order of 100-800 members (transformants) can
be screened for enantioselectivity within 1 day using medium-
Results and Discussion
Initial Saturation Mutagenesis Experiments. LEH is a 149
amino acid (16.5 kDa) enzyme which has been characterized
by X-ray crystallography of a sample containing the inhibitor
valpromide (2-n-propylpentanoic acid amide).3b In order to make
a sound decision regarding proper CAST sites for saturation
mutagenesis, we first performed induced fit docking of the model
substrate 3 using the published crystal structure. This procedure
led to the identification of eight residues for potential random-
ization, namely, Met32, Leu35, Leu74, Met78, Ile80, Val83,
Leu114, and Ile116 (Figure 1a). The next step required a
decision concerning the question of how to group these amino
acid positions into appropriate sites. In principle, one could opt
for eight single residue sites, a strategy that we successfully
applied in the directed evolution of the enoate-reductase
YqjM.10c Using NNK codon degeneracy (N, adenine/cytosine/
guanine/thymine; K, guanine/thymine) encoding all 20 canonical
amino acids, this would entail eight randomization libraries, each
requiring about 100 transformants to be screened for 95% library
coverage assuming the absence of amino acid bias.10
(6) Recent reviews of directed evolution:7(a) Johannes, T. W.; Zhao, H.
Curr. Opin. Microbiol. 2006, 9, 261–267. (b) Lutz, S.; Bornscheuer,
U. T. Protein Engineering Handbook; Wiley-VCH: Weinheim, 2009;
Vols. 1-2. (c) Turner, N. J. Nat. Chem. Biol. 2009, 5, 567–573. (d)
Ja¨ckel, C.; Kast, P.; Hilvert, D. Annu. ReV. Biophys. Biomol. Struct.
2008, 37, 153–173. (e) Bershtein, S.; Tawfik, D. S. Curr. Opin. Chem.
Biol. 2008, 12, 151–158. (f) Romero, P. A.; Arnold, F. H. Nat. ReV.
Mol. Cell Biol. 2009, 10, 866–876. (g) Reetz, M. T. In Asymmetric
Organic Synthesis with Enzymes; Gotor, V., Alfonso, I., Garc´ıa-
Urdiales, E., Eds.; Wiley-VCH: Weinheim, 2008; pp 21-63. (h) Otten,
L. G.; Hollmann, F.; Arends, I. W. C. E. Trends Biotechnol. 2010,
28, 46–54. (i) Shivange, A. V.; Marienhagen, J.; Mundhada, H.;
Schenk, A.; Schwaneberg, U. Curr. Opin. Chem. Biol. 2009, 13, 19–
25.
(7) Review of directed evolution of stereoselective enzymes:6g Reetz, M. T.
Angew. Chem. 2010, 122, DOI: 10.1002/ange.201000826. Angew.
Chem., Int. Ed. 2010, 49, DOI: 10.1002/anie.201000826.
(8) (a) Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Keꢀeler, M.;
Stu¨rmer, R.; Zelinski, T. Angew. Chem. 2004, 116, 806–843. Angew.
Chem., Int. Ed. 2004, 43, 788-824. (b) Blaser, H.-U.; Schmidt, E.
Asymmetric Catalysis on Industrial Scales; Wiley-VCH: Weinheim,
2004. (c) Berkessel, A.; Gro¨ger, H. Asymmetric Organocatalysis;
Wiley-VCH: Weinheim, 2004. (d) Tao, J.; Lin, G.-Q.; Liese, A.
Biocatalysis for the Pharmaceutical Industry; Wiley-VCH: Weinheim,
2009.
In other studies, however, we discovered that randomization
at sites composed of more than one amino acid position lead to
(9) (a) Reymond, J.-L.; Fluxa`, V. S.; Maillard, N. Chem. Commun.
(Cambridge, U. K.) 2009, 34–46. (b) Reymond, J.-L. Enzyme Assays
- High-throughput Screening, Genetic Selection and Fingerprinting;
Wiley-VCH: Weinheim, 2006. (c) Reetz, M. T. In Enzyme Assays -
High-throughput Screening, Genetic Selection and Fingerprinting;
Reymond, J.-L., Ed.; Wiley-VCH: Weinheim, 2006; pp 41-76.
(10) (a) Reetz, M. T.; Kahakeaw, D.; Lohmer, R. ChemBioChem 2008, 9,
1797–1804. (b) Reetz, M. T.; Carballeira, J. D. Nat. Protoc. 2007, 2,
891–903. (c) Bougioukou, D. J.; Kille, S.; Taglieber, A.; Reetz, M. T.
AdV. Synth. Catal. 2009, 351, 3287–3305. (d) Reetz, M. T.; Prasad,
S.; Carballeira, J. D.; Gumulya, Y.; Bocola, M. J. Am. Chem. Soc.
2010, 132, 9144–9152.
(12) Directed evolution of other EHs:10a(a) Reetz, M. T.; Torre, C.; Eipper,
A.; Lohmer, R.; Hermes, M.; Brunner, B.; Maichele, A.; Bocola, M.;
Arand, M.; Cronin, A.; Genzel, Y.; Archelas, A.; Furstoss, R. Org.
Lett. 2004, 6, 177–180. (b) Reetz, M. T.; Wang, L.-W. Angew. Chem.
2006, 118, 1258–1263, in part M. Bocola. Angew. Chem. Int. Ed. 2006,
45, 1236-1241. Erratum. Angew. Chem. Int. Ed. 2006, 45, 2556.
Erratum. Angew. Chem. Int. Ed. 2006, 45, 2494. (c) van Loo, B.; Lutje
Spelberg, J. H.; Kingma, J.; Sonke, T.; Wubbolts, M. G.; Janssen,
D. B. Chem. Biol. 2004, 11, 981–990. (d) Lee, E. Y.; Shuler, M. L.
Biotechnol. Bioeng. 2007, 98, 318–327. (e) Rui, L.; Cao, L.; Chen,
W.; Reardon, K. F.; Wood, T. K. Appl. EnViron. Microbiol. 2005,
ˇ
71, 3995–4003. (f) Kotik, M.; Steˇpa´nek, V.; Kysl´ık, P.; Maresˇova´, H.
(11) Lutz, S.; Patrick, W. M. Curr. Opin. Biotechnol. 2004, 15, 291–297.
J. Biotechnol. 2007, 132, 8–15.
9
J. AM. CHEM. SOC. VOL. 132, NO. 44, 2010 15745

A R T I C L E S
Zheng and Reetz
Figure 1. (a) Eight amino acid residues aligning the binding pocket of LEH, the analysis being based on induced fit docking of meso-substrate 3 (purple)
using the published crystal structure.3b (b) Grouping of the eight residues into four randomization sites A (Met32/Leu35), B (Leu74/Ile80), C (Leu114/
Ile116), and D (Met78/Val83).
Table 1. Selected Hits from Initial Randomization Libraries at Sites
libraries characterized by particularly high quality,10d,12b which
was traced to the occurrence of cooperative effects operating
between the newly introduced amino acids within a given site
A, B, C, and D
-∆∆G‡
a
mutant
mutations
conv. (%) ee (%) abs. conf. (kJ·mol-1
)
and between sets of mutations arising from the different sites
as the ISM process proceeds. Cooperativity entails epistatic
effects which are more than additive, which means that it is the
ideal form of epistasis in directed evolution.10a,d When choosing
single-residue sites in an ISM scheme, cooperative effects are
by nature not relevant in the initial randomization libraries10c
and become possible only in the subsequent mutagenesis
generations when point mutations begin to interact with one
another. In the present study we decided to group the eight
residues (Figure 1a) into four sites, each composed of two amino
acid positions, namely, A (Met32/Leu35), B (Leu74/Ile80), C
(Leu114/Ile116), and D (Met78/Val83) (Figure 1b).
WT LEH
72
79
78
78
67
75
75
74
72
72
82
80
68
14
24
16
10
53
58
66
50
60
68
29
13
7
(R,R)
(S,S)
(S,S)
(S,S)
(R,R)
(R,R)
(R,R)
(S,S)
(S,S)
(S,S)
(R,R)
(R,R)
(R,R)
0.7
1.2
0.8
0.5
2.9
3.3
3.9
2.7
3.4
4.1
1.5
0.6
0.3
library A H139 Met32Leu/Leu35Phe
H33 Met32Leu/Leu35Cys
H35 Met32Leu/Leu35Val
library B H142 Leu74Val/Ile80Val
H51 Leu74Ile/Ile80Val
H143 Leu74Ile/Ile80Cys
library C H22 Leu114Ile/Ile116Val
H23 Leu114Val/Ile116Val
H126 Leu114Cys/I116Val
library D H175 Met78Phe/Val83Ile
H176 Met78Ile/Val83Ile
H177 Met78Val/Val83Ile
a Energies calculated relative to ee ) 0%.
When choosing NNK codon degeneracy, requiring in each
library the screening of about 3000 transformants for 95%
coverage,10a,b the total screening effort would amount to about
12 000 measurements. Since this does not include any ISM steps
and already exceeds the screening limit set by ourselves, 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 balanced
mixture of building blocks having polar, nonpolar, charged,
noncharged, aromatic, and nonaromatic side chains. As pointed
out previously, NDT requires for 95% library coverage the
screening of only 430 transformants.10a,b We settled for 466,
which means that the total screening effort regarding the initial
four libraries is calculated to involve about 1864 transformants.
Following the generation of the four initial libraries at sites A,
B, C, and D, we screened for activity using an adapted form13a
of Reymond’s adrenaline test13b,c and subsequently applied
automated GC in order to obtain the respective ee values.
The performance of the three best mutants in each library is
summarized in Table 1, which reveals some remarkable trends.
In particular, mutants which improve (R,R)-selectivity of WT-
LEH beyond 14% ee are found only in libraries B and D,
whereas mutants displaying reversed enantioselectivity in favor
of (S,S)-4 occur solely in libraries A and C. There was no
compelling reason to expect this. Indeed, in our previous study
concerning the enhancement as well as the reversal of enanti-
oselectivity of the enoate-reductase YqjM as a catalyst in the
asymmetric reduction of 3-methylcyclohexenone, both (R)- and
(S)-selective mutants were discovered in one and the same
randomization library.10c The best hits showing the highest
degree of enantioselectivity in the present study proved to be
double mutants, namely, H139 (Met32Leu/Leu35Phe) from
library A, H143 (Leu74Ile/Ile80Cys) from library B, H126
(Leu114Cys/Ile116Val) from library C, and H175 (Met78Phe/
Val83Ile) from library D. The biggest change in terms of net
∆∆G‡ relative to WT occurs in the case of mutant H126 with
reversed enantioselectivity (0.7 + 4.1 ) 4.8 kJ ·mol-1).
Deconvolution Experiments. In order to identify the nature
of epistatic interactions operating between the respective point
mutations in each double mutant originating in the initial
libraries, systematic deconvolution experiments were performed
by generating and testing the respective two single mutants. By
comparing the relevant ee values and the respective ∆∆G‡ data
(Table 2), it can be seen that cooperative effects are operating
in two of the four mutants, namely, H139 from library A and
H126 from library C. For example, in the case of the double-
mutant H126 showing an ee value of 68% (S,S) corresponding
to ∆∆G‡ ) -4.1 kJ ·mol-1, the single mutants Leu114Cys and
Ile116Val lead to enantioselectivities of only 4% ee and 38%
ee, corresponding to ∆∆G‡ values of -0.2 and -2.0 kJ ·mol-1
,
respectively. Noteworthy are also the results of deconvoluting
the (S,S)-selective double-mutant H139 (Met32Leu/Leu35Phe).
The respective single mutants Met32Leu and Leu35Phe are both
(R,R)-selective! This stands in sharp contrast to the traditional
expectation of protein engineers. Going the other way, i.e.,
combining two point mutations each correlating with a certain
improved enzyme property is generally believed to result in an
enhancement of that particular catalytic parameter. Indeed, in
(13) (a) Kahakeaw, D.; Reetz, M. T. Chem.-Asian J. 2008, 3, 233–238.
(b) Wahler, D.; Reymond, J.-L. Angew. Chem. 2002, 114, 1277–1280.
Angew. Chem., Int. Ed. 2002, 41, 1229-1232. (c) Wahler, D.; Boujard,
O.; Lefe`vre, F.; Reymond, J.-L. Tetrahedron 2004, 60, 703–710.
9
15746 J. AM. CHEM. SOC. VOL. 132, NO. 44, 2010

Stereoselectivity of Limonene Epoxide Hydrolase
A R T I C L E S
Table 2. Results of Deconvoluting the Double Mutants H139,
H143, H126, and H175, the Model Reaction Being the Hydrolytic
Desymmetrization of epoxide 3
-∆∆G‡
a
mutant
mutations
conv. (%) ee (%) abs. conf (kJ·mol-1
)
H139 Met32Leu/Leu35Phe
H1 Met32Leu
H130 Leu35Phe
79
76
75
75
80
75
72
82
63
82
78
78
24
14
2
66
28
50
68
4
38
29
25
18
(S,S)
(R,R)
(R,R)
(R,R)
(R,R)
(R,R)
(S,S)
(S,S)
(S,S)
(R,R)
(R,R)
(R,R)
1.2
0.7
0.1
3.9
1.4
2.7
4.1
0.2
2.0
1.5
1.3
0.9
H143 Leu74Ile/Ile80Cys
H6
Leu74Ile
H136 Ile80Cys
H126 Leu114Cys/Ile116Val
H17
H10
Leu114Cys
I1e116Val
H175 Met78Phe/Val83Ile
H137 Met78Phe
H138 Val83Ile
a Energies calculated relative to ee ) 0%.
an earlier study regarding substrate acceptance (rate) of a lipase
in the hydrolysis of bulky esters, we combined point mutations
introduced earlier, which resulted in a further enhancement of
rate.6g Relevant is the report by Bornscheuer and co-workers,
who showed that upon deconvoluting an evolved double mutant
of an esterase with enhanced (S)-selectivity, one of the respective
single mutants displayed the expected (S)-selectivity but the
other proved to be (R)-selective.14 Our case is even more drastic,
which shows that such cooperative effects can hardly be
predicted by present theory. In the case of mutant H143 from
library B, essentially full additivity pertains, whereas the
interaction between the two point mutations in the case of mutant
H175 from library D proved to be somewhat less than additive
but not antagonistic (Table 2).
Iterative Rounds of Saturation Mutagenesis. We then had to
make decisions regarding further evolutionary optimization
based on ISM. A complete 4-site ISM scheme leading to diol
4 with a given absolute configuration would entail 24 pathways
and 64 randomization libraries (S4, Supporting Information).
Thus, for both (R,R)- and (S,S)-selectivity a total of 48 pathways
and 128 randomization libraries are relevant. In the present study
the goal was not to test all theoretically possible pathways but
to see how far one gets by exploring a severely limited number
of options. We therefore decided to restrict ISM to three sites
and to explore only two pathways, one directed toward (R,R)-
selectivity and the other toward (S,S)-selectivity.
At this stage we were particularly interested in improving
reversed enantioselectiviy of mutant H126 (68% ee in favor of
(S,S)-4) originating from library C. Its gene was used as a
template for saturation mutagenesis at two other sites. We opted
for sites A and B, although the initial randomization round at
site B had provided only (R,R)-selective mutants. Consequently,
two pathways C f B f A and C f A f B were explored
experimentally (Figure 2).
It can be seen that pathway C f B f A ends with the
evolution of mutant H178 having an ee value of 93% (S,S),
whereas C f A f B provides mutant H150 with 92% ee (S,S)
(Figure 2). Although the two ee values are essentially identical,
the point mutations accumulated in the respective second and
third generations are quite different: Mutant H178 (Leu114Cys/
Ile116Val/Met32Cys/Ile80Phe) versus mutant H150 (Leu114Cys/
Ile116Val/Met32Leu/Leu35Phe/Leu74Phe/Ile80Val). It is also
interesting to note that in pathway C f A f B leading to mutant
Figure 2. Multiple pathways explored for inverting and improving the
enantioselectivity of LEH as a catalyst in the desymmetrization of
cyclopentene oxide (3). Red arrows denote the results of generating and
screening the initial libraries at sites A, B, C, and D; green arrows highlight
the results of a restricted set of ISM experiments.
H150, saturation mutagenesis at site A in the second mutagenesis
step results in the simultaneous introduction of two point
mutations, Met32Leu/Leu35Phe, which are identical to those
introduced in the initial randomization library at site A. There
is no reason to expect this, although in the initial rounds of
saturation mutagenesis this was one of the two cases displaying
cooperative effects. In contrast, the last randomization experi-
ment in pathway C f B f A at site A provides a single new
point mutation Met32Cys. Taken together, the two pathways
leading to 92-93% ee in favor of (S,S)-4 required a total
screening effort of only 2300 transformants.
In the quest to improve the (R,R)-selectivity of WT LEH,
only a select number of ISM experiments were performed
(Figure 2). The gene of improved mutant H143 obtained in the
initial saturation mutagenesis library at site B was used as a
template to perform randomization at site A. This provided
mutant H167 characterized by ee ) 73%, which was improved
to 80% ee by continuing ISM at site D (Figure 2). It is
noteworthy that in the final ISM round at site D of this (R,R)-
selective pathway B f A f D, the same two point mutations
(Met78Phe/Val83Ile) occur as in the initial library D. Appar-
ently, no other combination is more effective. The (R,R)-
selective pathway B f A f D required the screening of about
1400 transformants, which means that the total screening effort
in Figure 2 amounted to less than 4700 transformants.
Thermostability Studies. The introduction of point mutations
in an enzyme, for whatever purpose, may lead to a decrease in
thermostability.6,7 In order to test this in the case of the best
mutants H150, H173, and H178, the respective T5015 values were
measured, the temperatures at which 50% residual activity
15
occurs following a heat treatment of 15 min.10b,15 Whereas T50
for WT LEH was found to be 49.5 °C, which is in line with
literature reports,4 the mutants showed slightly lower values:
H150 (47.0 °C), H173 (47.0 °C), H178 (48.0 °C). The thermal
(14) Bartsch, S.; Kourist, R.; Bornscheuer, U. T. Angew. Chem. 2008, 120,
1531–1534. Angew. Chem., Int. Ed. 2008, 47, 1508-1511.
9
J. AM. CHEM. SOC. VOL. 132, NO. 44, 2010 15747

A R T I C L E S
Zheng and Reetz
Table 3. Hydrolytic Desymmetrization of meso-Epoxides Using LEH Mutants Specifically Evolved for the Reaction of Substrate 3
11
specific activitya
3
5
7
9
mutant
% ee
specific activitya
% ee
specific activitya
% ee
specific activitya
% ee
specific activitya
% ee
WT
14 (R,R)
92 (S,S)
93 (S,S)
80 (R,R)
23
13
20
24
0
74
67
197
123
19 (S,S)
93 (S,S)
98 (S,S)
77 (R,R)
30
26
65
73
5 (S,S)
146
ndb
127
144
93 (R,R)
95 (R,R)
63 (R,R)
99 (R,R)
264
ndb
187
297
H150
H178
H173
96 (S,S)
97 (S,S)
90 (R,R)
86 (S,S)
93 (S,S)
83 (R,R)
a Specific activity: mmol/min/mg protein. b nd: not determined.
Table 4. Hydrolytic Kinetic Resolution of Epoxides rac-13 and rac-15 Using WT LEH and Mutants H173 and H178 as Catalysts
rac-13 rac-15
E value E value
entry
mutant
eep (%)
conv. (%)
specific activitya
eep (%)
conv. (%)
specific activitya
1
2
3
WT
H173
H178
37
35
92
33
26
31
2.6 (R)
2.3 (R)
32 (R)
267
292
223
40
92
91
36
30
43
2.8 (R)
36 (R)
44 (S)
32
45
137
a Specific activity: (mmol/min/mg protein).
stability profiles are basically consistent with those of circular
dichroism (see Supporting Information). Thus, there is some
but no significant trade off in going from WT LEH to the
evolved stereoselective mutants.
98% ee; 5, 90% ee; 9, 90% ee).16 Other epoxide hydrolases
identified by gene mining show slightly higher enantioselectivity
than the mutants of the present study but again only in favor of
the (R,R)-products.17 Single-site saturation mutagenesis of the
EH from Agrobacterium radiobacter resulted in the improve-
ment of enantioselectivity in the desymmetrization of 9 favoring
(R,R)-10 (ee ) 99%), but reversal of stereoselectivity was not
reported.18 Synthetic Jacobsen-type salen-cobalt catalysts in ring-
opening reactions of meso-epoxides have the enormous advan-
tage that different nucleophiles such as azide (>95% ee)2a or
benzoate (65-93% ee)2c can be employed with formation of
compounds having high enantiomeric enrichment, the absolute
configuration of the catalyst ligand determining the absolute
configuration of the products.2 In the original Jacobsen system
involving monomeric chiral Co catalysts, water was used as
the nucleophile in the hydrolytic kinetic resolution of mono-
substituted epoxides with very high selectivity factors often
exceeding E ) 100,2b but desymmetrization of meso-epoxides
under such conditions was not reported. Subsequently, the
analogous oligomeric complexes were used in the hydrolytic
desymmetrization of meso-epoxide 5 (ee ) 94%) being the only
example.2e,f Later it was found that such complexes catalyze
the hydrolytic desymmetrization of epoxides 3 (98% ee) and 9
(87% ee).2g
Investigation of Substrate Scope and Activity. In an effort to
see how the three best mutants H150, H173, and H178 perform
as catalysts in the hydrolytic desymmetrization of other meso-
configurated epoxides, compounds 5, 7, 9, and 11 were tested
as substrates. Table 3 shows that substrate scope of these mutants
is remarkably broad, because activity, improved enantioselec-
tivity, and inverted enantioselectivity are in most cases even
higher than in the reaction of the model compound 3 which
was actually used in the evolution experiments. With the
exception of substrate 11, WT-LEH is consistently a poor
catalyst (Table 3). Cases of higher enantioselectivity than in
the model reaction are mutant H173 as a catalyst in the
desymmetrization of 5 (97% ee in favor of (S,S)-6) and variant
H178 as a catalyst for inverting stereoselectivity (90% ee in
favor of (R,R)-6). Mutant H173 leads to an inversion of
configuration, except for the case of epoxide 11, which is
currently difficult to rationalize. WT LEH and the mutants
generally display similar activity. All of the intermediate mutants
along the ISM scheme (Figure 2) were likewise tested as
catalysts in the hydrolytic desymmetrization reactions, likewise
resulting in high enantioselectivity with respect to improvement
and inversion as seen by the respective “ISM schemes”, although
no additional mutagenesis experiments were performed (see
Supporting Information).
It was also of interest to test structurally very different
substrates such as monosubstituted epoxides rac-13 and rac-
15, this time in hydrolytic kinetic resolution. As in the previous
examples, WT LEH proved to be a poor catalyst, the selectivity
factor E amounting to only 2.6 and 2.8, respectively, both in
slight favor of the respective (R)-diols. Table 4 shows that
notable improvements in (R)-selectivity are observed for both
substrates (E ) 32 and 36, respectively) and that in the case of
(15) Reviews covering protein engineering of thermostability: (a) Oshima,
´
T. Curr. Opin. Struct. Biol. 1994, 4, 623–628. (b) O’Fa´ga´in, C. Enzyme
Microb. Technol. 2003, 33, 137–149. (c) Eijsink, V. G. H.; Gåseidnes,
S.; Borchert, T. V.; van den Burg, B. Biomol. Eng. 2005, 22, 21–30.
(d) Bommarius, A. S.; Broering, J. M. Biocatal. Biotransform. 2005,
23, 125–139. (e) Amin, N.; Liu, A. D.; Ramer, S.; Aehle, W.; Meijer,
D.; Metin, M.; Wong, S.; Gualfetti, P.; Schellenberger, V. Protein
Eng., Des. Sel. 2004, 17, 787–793.
(16) Weijers, C. A. G. M. Tetrahedron: Asymmetry 1997, 8, 639–647.
(17) Zhao, L.; Han, B.; Huang, Z.; Miller, M.; Huang, H.; Malashock, D. S.;
Zhu, Z.; Milan, A.; Robertson, D. E.; Weiner, D. P.; Burk, M. J. J. Am.
Chem. Soc. 2004, 126, 11156–11157.
It is informative to compare these results with the catalytic
profiles of other EHs as catalysts in the same hydrolytic
reactions. In the desymmetrization of meso-substrates 3, 5, and11
using the EH from Rhodotorula glutinis, high enantioselectivity
can be achieved but only in favor of the (R,R)-products (3, g
(18) van Loo, B.; Kingma, J.; Heyman, G.; Wittenaar, A.; Lutje Spelberg,
J. H.; Sonke, T.; Janssen, D. B. Enzyme Microb. Technol. 2009, 44,
145–153.
9
15748 J. AM. CHEM. SOC. VOL. 132, NO. 44, 2010

Stereoselectivity of Limonene Epoxide Hydrolase
A R T I C L E S
Table 5. Test for Possible Stereoconvergent Catalysis Using WT
LEH and Mutants in the Hydrolytic Reaction of rac-17 and rac-19
Leading to Diols 18 and 20, respectively
rac-15 reversal of enantioselectivity is possible on an optional
basis by using mutant H178 (E ) 44 in favor of (S)-8). These
results are of theoretical interest, but from a synthetic point of
view they are not particularly significant because Jacobsen’s
catalysts are clearly superior in such asymmetric
transformations,2b as are some WT EHs1 or mutants.12b,19 For
example, although the EH from Aspergillus niger (ANEH) is a
poor catalyst in the hydrolytic kinetic resolution of rac-13 (E
) 4.6 (S)),12b a mutant evolved by ISM proved to be highly
enantioselective (E ) 115 (S)),12b which was also efficient in
the reaction of other racemic monosubstituted epoxides.19
Directed evolution of the EH from Agrobacterium radiobacter
as a catalyst in the hydrolytic kinetic resolution of rac-15
improved enantioselectivity from E ) 16 (WT) to E ) 38 (best
mutant favoring (R)-16).20
18
20
specific
activityb conv. (%)a
entry
mutant
conv. (%)a
% ee
% ee
1
2
3
4
WT LEH
H150
H173
>99
>99
>99
>99
19 (1S,2S) 2573
63 (1S,2S) ndc
50 (1R,2R) 2291
55 (1S,2S) 2197
99
98
98
99
2 (1S,2S)
28 (1S,2S)
25 (1R,2R)
51 (1S,2S)
H178
a Conversion at reaction time of 12 h. b Specific activity (mmol/min/
mg protein). c Specific activity not determined.
stereoconvergent transformation of epoxides rac-17 and rac-
19. Table 5 reveals poor catalytic behavior of WT LEH in these
reactions. At complete conversion of rac-17, very limited
stereoconvergency was observed in favor of (1S,2S)-18 (ee )
19%), while in the reaction of rac-19 essentially racemic product
resulted. In contrast, some of the LEH mutants proved to be
markedly improved catalysts, although in no case truly practical
results were achieved. Even the best ee value did not exceed
63% (Table 5, entries 2-4). However, both (1R,2R)- and
(1S,2S)-selective mutants are now available for partially ste-
reoconvergent transformations of both substrates rac-17 and rac-
19, setting the stage for future optimization using the tools of
directed evolution. Some of the intermediate mutants evolved
for the hydrolytic desymmetrization of 3 display similar catalytic
profiles (see Supporting Information).
Testing Stereoconvergency of WT LEH and Evolved Mutants
in Hydrolytic Reactions of Other Racemic Substrates. It has been
reported that WT LEH displays stereoconvergent behavior in
the hydrolytic reaction of a mixture of limonene oxide diaster-
eomers,(1S,2R,4R)and(1R,2S,4R),bothleadingtothe(1S,2S,4R)-
limonene diol in a sequential manner.4 From a synthetic organic
viewpoint, enantioconvergency entails interesting perspectives,
especially when a racemate leads to a single enantiomer at >95%
conversion. This has been observed for some EHs21 and for a
mutant evolved by ISM for this purpose.22 In the present study
we tested WT LEH and the mutants, specifically evolved for
the desymmetrization of meso-substrate 3, as catalysts in the
(19) Reetz, M. T.; Bocola, M.; Wang, L.-W.; Sanchis, J.; Cronin, A.; Arand,
M.; Zou, J.; Archelas, A.; Bottalla, A.-L.; Naworyta, A.; Mowbray,
S. L. J. Am. Chem. Soc. 2009, 131, 7334–7343.
(20) (a) Lutje Spelberg, J. H.; Rink, R.; A., A.; Furstoss, R.; Janssen, D. B.
AdV. Synth. Catal. , 344, 980–985.
Some experiments regarding the hydrolytic kinetic resolution
were also performed in the case of rac-17. WT LEH is only
slightly (1S,2S) selective (E ) 4; specific activity, 2573 mmol/
min/mg protein), mutant H178 increasing this to E ) 13 (specific
activity,2197 mmol/min/mg protein), and H173 inducing inver-
sion of enantioselctivity (E ) 8) in favor of (1R,2R)-18. This
suggests that ISM should be applied directly to substrates of
this kind. In the case of trisubstituted epoxides, Jacobsen
catalysts fail.2
Scale-Up in the Hydrolytic Desymmetrization of meso-Ep-
oxides. Having in hand several enantioselective LEH mutants
as catalysts in the hydrolytic desymmetrization of meso-epoxides
and kinetic resolution of racemic substrates, we considered scale
up for practical applications. For illustrative purposes, mutants
(21) Examples of enantioconvergent EH-catalyzed reactions of chiral
epoxides: (a) Pedragosa-Moreau, S.; Morisseau, C.; Zylber, J.;
Archelas, A.; Baratti, J.; Furstoss, R. J. Org. Chem. 1996, 61, 7402–
7407. (b) Bellucci, G.; Chiappe, C.; Cordoni, A. Tetrahedron:
Asymmetry 1996, 7, 197–202. (c) Kroutil, W.; Mischitz, M.; Faber,
K. J. Chem. Soc., Perkin Trans. 1 1997, 3629–3636. (d) Mayer, S. F.;
Steinreiber, A.; Orru, R. V. A.; Faber, K. Eur. J. Org. Chem. 2001,
4537–4542. (e) Karboune, S.; Archelas, A.; Furstoss, R.; Baratti, J. C.
Biocatal. Biotransform. 2005, 23, 397–405. (f) Cao, L.; Lee, J.; Chen,
W.; Wood, T. K. Biotechnol. Bioeng. 2006, 94, 522–529. (g) Xu, W.;
Xu, J.-H.; Pan, J.; Gu, Q.; Wu, X.-Y. Org. Lett. 2006, 8, 1737–1740.
(h) Hwang, S.; Choi, C. Y.; Lee, E. Y. Biotechnol. Lett. 2008, 30,
1219–1225. (i) Lin, S.; Horsman, G. P.; Chen, Y.; Li, W.; Shen, B.
J. Am. Chem. Soc. 2009, 131, 16410–16417.
(22) Zheng, H.; Kahakeaw, D.; Acevedo, J. P.; Reetz, M. T. ChemCatChem
2010, 2, 958–961.
9
J. AM. CHEM. SOC. VOL. 132, NO. 44, 2010 15749

A R T I C L E S
Zheng and Reetz
Table 6. Scale Up of the Hydrolytic Desymmetrization of Selected
meso-epoxides Using Mutants H173 and H178 as Catalysts
Table 7. Results of MD Simulations and Docking Experiments
Using Epoxide 3a
substrate
loading (mg)
conv.
(%)a
% ee of
diol
distance (Angstroms) distance (Angstroms) angle (deg) epoxy angle (deg) epoxy
from water O atom from water O atom O atom/CR/water O atom/CS/water
entry
mutants
substrate
yield (%)
mutant
to CR of 3
to CS of 3
O atom
O atom
1
2
3
H178
H178
H173
5
7
11
300
150
200
>99
>99
15
97 (S,S)
98 (S,S)
>99 (R,R)
90
86
91
WT
H173
H178
3.4
3.5
3.6
3.2
3.1
5.2
116.4
112.9
146.6
133.9
140.9
107.5
a Reaction time: 15 h.
a Note that attack at enantiotopic C-atom CR occurs with inversion of
configuration and leads to the (S,S)-product, while attack at CS provides
(R,R)-4.
H178 and H173 were chosen as catalysts in the desymmetri-
zation of meso-substrates 5, 7, and 9. Accordingly, 0.5 g (wet
cell weight) of fresh cells was used in each 10 mL reaction
mixture containing 150-300 mg of epoxide (Table 6). Although
there is room for further activity improvement in the case of
the sterically demanding diphenyl-substituted epoxide 11 (Table
6, entry 3), the present results document the synthetic utility of
these LEH mutants.
reshaping of the binding pocket, featuring the residues that are
mutated. In the case of WT LEH, the two distances are not
much different, 3.2 versus 3.4 Å (Figure 3c), in line with the
observed direction and low degree of enantioselectivity in favor
of (R,R)-4. In sharp contrast, the mutational changes in variant
H178 cause the repositioning of 3 so that now the other
enantiotopic C atom is considerably closer to the attacking water,
3.6 versus 5.2 Å (Figure 3d). Close inspection of the results
shows that the larger phenyl group in mutation Ile80Phe
“pushes” the substrate down so that now the other C atom of
the epoxide moiety is closer to the activated water, a process
that is spatially made possible by the other three point mutations
Met32Cys/Leu114Cys/Ile116Val. The respective volume change
is in line with this model (Supporting Information). Distance
alone is not sufficient to describe the ideal positioning of the
activated water, because the trajectory of nucleophilic attack
needs to be considered. As Table 7 shows, angles describing
the trajectory of the favored nucleophilic attack is in fact larger
than that of the disfavored mode, as expected. This analysis
can be related to a near-attack-conformation as postulated by
Bruice in an MD study of another EH.25
Unveiling the Source of Inverted Enantioselectivity. The
mechanism of WT LEH involves activation of the substrate by
H-bonding of Asp101 to the epoxide O atom, followed by
nucleophilic attack of water which is activated and properly
positioned by Tyr53, Asp132, and Asn55.3,4 It was of particular
interest to uncover the source of inverted enantioselectivity
induced by mutant H178 in the hydrolytic desymmetrization
of cyclopentene-oxide (3), because this entails the largest degree
of stereoselectivity change observed in the present study. As
an initial step in this direction, induced fit docking23 and
molecular dynamics (MD) simulations using the Schro¨dinger
package24 were performed (6 ns) for WT LEH and mutant H178,
respectively. As the crucial parameter we considered the
distance, d, between the attacking O atom of activated water
and the two possible enantiotopic C atoms of the bound substrate
undergoing nucleophilic substitution with inversion of config-
uration and formation of the trans-diol 4. Preferred attack is to
be expected at the C atom closest to the activated water. As a
corollary, the larger the difference in the two distances, the
higher the enantioselection. Figure 3a and 3b reveals the
Conclusions and Perspectives
Limonene epoxide hydrolase (LEH) is an unusual EH because
it does not exert its catalytic effect by the usual two-step
mechanism in which an aspartate undergoes nucleophilic
Figure 3. Repositioning of substrate 3 in the reshaped binding pocket of mutants relative to WT LEH as indicated by MD simulations. (a) Binding pocket
of (R,R)-selective WT LEH featuring the residues which undergo mutational change when going to (S,S)-selective mutant H178. (b) Reshaped binding
pocket of mutant H178. (c) Close-up view of the binding pocket of WT LEH featuring the residues which undergo mutational change when going to mutant
H178 and the distances d1 and d2 (in Angstroms) to the activated water. (d) Close-up view of the binding pocket of mutant H178 showing mutational
changes and the respective distances d3 and d4 (in Angstroms) to the activated water.
9
15750 J. AM. CHEM. SOC. VOL. 132, NO. 44, 2010

Stereoselectivity of Limonene Epoxide Hydrolase
A R T I C L E S
reaction at the H-bond-activated substrate in the rate- and
stereochemistry-determining step but rather via direct nucleo-
philic attack by properly positioned and activated water. The
natural role of LEH, which has been characterized by X-ray
crystallography,3b is the hydrolytic conversion of limonene oxide
present in Rhodococcus erythropolis DCL14 with formation of
the respective limonene diol.3 Unfortunately, WT LEH is not
suited for application in synthetic organic chemistry or bio-
technology, because substrate scope in terms of acceptable levels
of stereoselectivity is highly restricted.4 We have therefore
undertaken the first directed evolution study of LEH. The
purpose was not only to manipulate stereoselectivity for a model
compound (cyclopentene-oxide) and for a defined set of
substrates in hydrolytic desymmetrization and kinetic resolution
for practical applications. We also wanted to test the previously
developed strategy of protein engineering based on iterative
saturation mutagenesis (ISM) in a difficult case. Efficacy in
directed evolution means maximizing the degree of catalyst
improvement while minimizing the screening effort by generat-
ing high-quality libraries.7,10,11 Keeping these points in mind,
we specifically restricted the number of screened transformants
in the Darwinian process to less than 5000. A crucial additional
tool proved to be the use of a reduced amino acid alphabet based
on NDT codon degeneracy encoding 12 instead of the usual 20
canonical amino acids, because this reduces the screening effort
drastically. The results of the present study show once more
that ISM is indeed a highly efficient tool in accelerated directed
evolution. We also provide a guideline for grouping single amino
acid residues into optimal randomization sites. As a general
recommendation, sites comprising two amino acid positions are
to be preferred over single-residue sites.
Using cyclopentene-oxide (3) as the model substrate for
hydrolytic desymmetrization, with WT LEH displaying an ee
value of only 14% (R,R), two out of 48 (2 × 24) theoretically
possible ISM pathways were arbitrarily explored for achieving
both improved (R,R)-selectivity and inverted (S,S)-selectivity.
One pathway favors the formation of the corresponding (S,S)-
diol (4) (ee ) 93%) with inverted stereoselectivity, the other
optionally providing access to the enantiomeric product (R,R)-4
with improved enantioselectivity (ee ) 80%). Moreover, the
evolved mutants were subsequently tested successfully in the
hydrolytic desymmetrization of other meso-epoxides and in
the kinetic resolution of structurally very different racemic
substrates, enantioselectivity again being substantial in contrast
to WT LEH (ee up to 99%). These efforts did not require any
additional mutagenesis experiments, which means once more
that in directed evolution “you may get more than what you
screen for”.10c Thermostability of the best mutants proved to
be similar to that of WT LEH, which means that no notable
compromise regarding this important property was made in the
process of genetic optimization. Another noteworthy result of
the present study concerns the surprising finding that the point
mutations of two single mutants (Met32Leu and Leu35Phe),
each being (R,R)-selective, when combined in the form of the
respective double mutant (Met32Leu/Leu35Phe), induces (S,S)-
selectivity. This shows that the traditional way of viewing
mutational effects as being additive in a given direction should
be treated with caution.
Unveiling the source of enhanced stereoselectivity of an
evolved mutant enzyme requires investigations which include
kinetics, inhibition experiments, crystal structures, and extensive
MD simulations as well as induced docking experiments.19 Such
a detailed study still needs to be performed for the LEH mutants.
Reported herein are only induced docking experiments and MD
simulations of the reaction of cyclopentene-oxide, which indicate
that the distances, d, between activated water and the two
possible C atoms of meso-epoxides moiety being attacked
nucleophilically are similar when using WT LEH having low
enantioselectivity but quite different in the case of the best (S,S)-
selective mutant. The predicted (preferred) attack corresponds
to the shorter distance and explains the observed switch in the
direction and degree of enantioselectivity. However, this is a
crude model, and other factors may well be involved as well.
ISM constitutes a rational approach to directed evolution,
because it is a combination of knowledge-based design and
randomization, requiring only small mutant libraries. Recently,
it has been applied in the quest to evolve such protein properties
as enhanced stereoselectivity and activity, and altered cofactor
binding of a number of different enzymes, including enzyme
promiscuity and metabolic pathway engineering.26
Acknowledgment. We thank Professor Michael Arand (ETH
Zu¨rich) for providing the plasmid of pGEF-LEH.
Supporting Information Available: Experimental details and
data regarding MD simulations. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA1067542
(26) Recent examples of ISM: (a) Engstro¨m, K.; Nyhlen, J.; Sandstro¨m,
A. G.; Ba¨ckvall, J.-E. J. Am. Chem. Soc. 2010, 132, 7038–7042. (b)
Okrasa, K.; Levy, C.; Wilding, M.; Goodall, M.; Baudendistel, N.;
Hauer, B.; Leys, D.; Micklefield, J. Angew. Chem. 2009, 121, 7827–
7830. Angew. Chem., Int. Ed. 2009, 48, 7691-7694. (c) Hawwa, R.;
Larsen, S. D.; Ratia, K.; Mesecar, A. D. J. Mol. Biol. 2009, 393, 36–
57. (d) Liang, L.; Zhang, J.; Lin, Z. Microb. Cell Fact. 2007, 6, 36.
(e) Ivancic, M.; Valinger, G.; Gruber, K.; Schwab, H. J. Biotechnol.
2007, 129, 109–122. (f) De Groeve, M. R. M.; De Baere, M.; Hoflack,
L.; Desmet, T.; Vandamme, E. J.; Soetaert, W. Protein Eng., Des.
Sel. 2009, 22, 393–399. (g) Bhuiya, M.-W.; Liu, C.-J. J. Biol. Chem.
2010, 285, 277–285. (h) Kelly, R. M.; Leemhuis, H.; Dijkhuizen, L.
Biochemistry 2007, 46, 11216–11222. (i) Tang, W. L.; Li, Z.; Zhao,
H. Chem. Commun. 2010, 45, 5461–5463.
(23) Sherman, W.; Day, T.; Jacobson, M. P.; Friesner, R. A.; Farid, R.
J. Med. Chem. 2006, 49, 534–553.
(24) Schro¨dinger, LLC Maestro 8.5 User Manual; Schro¨dinger Press: New
York, 2008.
(25) Schiøtt, B.; Bruice, T. C. J. Am. Chem. Soc. 2002, 124, 14558–14570.
9
J. AM. CHEM. SOC. VOL. 132, NO. 44, 2010 15751