Biocatalytic Potential of the Epoxide Hydrolase from Agrobacterium radiobacter AD1 and a Mutant with Enhanced Enantioselectivity

Article abstract of DOI:10.1002/1615-4169(200210)344:9<980::AID-ADSC980>3.0.CO;2-A

Optically pure epoxides are useful synthons for a variety of biologically active compounds. The epoxide hydrolase obtained from Agrobacterium radiobacter AD1 hydrolyses racemic aryl epoxides with moderate and aliphatic epoxides with low enantioselectivity

Full text of DOI:10.1002/1615-4169(200210)344:9<980::AID-ADSC980>3.0.CO;2-A

FULL PAPERS
Biocatalytic Potential of the Epoxide Hydrolase from
Agrobacterium radiobacter AD1 and a Mutant with Enhanced
Enantioselectivity
Jeffrey H. Lutje Spelberg,a Rick Rink,a Alain Archelas,b Roland Furstoss,b Dick B.
Janssena,*
a
Department of Biochemistry, Groningen Biomolecular Sciences & Biotechnology Institute, Nijenborgh 4,
9747 AG, Groningen, Netherlands
Fax: ( ‡ 31)-50-3634165, e-mail: d.b.janssen@chem.rug.nl
b
¬
¬
¬
¬
Groupe Biocatalyse et Chimie Fine, Universite de la Mediterranee, Faculte des Sciences de Luminy, Case 901,
163 Avenue de Luminy, 13288 Marseille Cedex 9, France
Fax: ( ‡ 33)-491-829145, e-mail: furstoss@luminy.univ-mrs.fr
Received: May 18, 2002; Accepted: July 28, 2002
Dedicated to Prof. Roger Sheldon on the occasion of his 60th birthday.
Abstract: Optically pure epoxides are useful synthons resulted in an enzyme with increased enantioselec-
for a variety of biologically active compounds. The tivity towards aryl epoxides. The relatively strong
epoxide hydrolase obtained from Agrobacterium decrease in activity towards the remaining enantiom-
radiobacter AD1 hydrolyses racemic aryl epoxides ers makes this enzyme a much better biocatalyst than
with moderate and aliphatic epoxides with low the wild-type enzyme for the preparation of optically
enantioselectivity. The three-dimensional structure pure (S)-styrene oxide derivatives.
of this enzyme indicates that two tyrosine residues
interact with the epoxide oxygen. Mutating one of Keywords: biotransformation; enzyme catalysis; ep-
these, tyrosine 215, to a phenylalanine (Y215F) oxides; kinetic resolution; protein engineering
Introduction
a covalent intermediate, and the hydrolysis of this ester
by an activated water molecule, yielding the diol
Epoxide hydrolases catalyze the hydrolysis of an (Scheme 1). An initial determination of the enantiose-
epoxide to yield the corresponding vicinal diol. In the lectivity of kinetic resolutions showed that this enzyme
past decade this class of enzymes has been explored for is moderately enantioselective towards styrene oxide
the preparation of optically active epoxides and derivatives.[9] para-Chlorostyrene oxide was hydrolyzed
diols.[1,2,3] The most common application of epoxide with the highest enantioselectivity, followed by ortho-
hydrolases is the preparation of optically pure epoxides chloro- and meta-chloro-substituted styrene oxides.
by enantioselective hydrolysis of a racemic epoxide. With styrene oxide derivatives, the remaining epoxide
Recently, the usefulness of these enzymes in organic was generally of (S)-configuration, except for a-methyl-
synthesis has been demonstrated by various prepara- styrene oxide where, surprisingly, the result was oppo-
tive-scale conversions.[4,5] Avariety of biologically active site.[10]
products were synthesized by implementing such
The X-ray structure shows that the enzyme consists of
two domains, an a/b-hydrolase fold main domain and a
enzymatic or chemo-enzymatic reaction steps.[2]
Previously, the epoxide hydrolase from Agrobacteri- cap domain.[11] The active site with the catalytic triad is
um radiobacter AD1, an organism able to grow on located in a hydrophobic cavity between the two
epichlorohydrin, was investigated.[6] The enzyme was domains. Two tyrosine residues (Tyr215 and Tyr152)
purified and overexpressed in E. coli, which made it are positioned in the cap domain in such a manner that
available in large amounts.[7,8] X-ray crystallography and their phenolic hydroxy groups point towards the
site-directed mutagenesis studies showed that the expected binding position of the oxygen atom of the
residues Asp107, His275 and Asp246 function as a epoxide ring (Figure 1). These residues were proposed
catalytic triad. The two-step catalytic mechanism to stabilize the transition state towards the formation of
involved nucleophilic attack on the primary carbon of the intermediate by protonation of the leaving oxygen
the epoxide ring by Asp107, resulting in the formation of group. On basis of the structure, mutants were con-
980
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1615-4150/02/34409-980 985 $ 17.50-.50/0
Adv. Synth. Catal. 2002, 344, No. 9

Biocatalytic Potential of the Epoxide Hydrolase from Agrobacterium radiobacter
FULL PAPERS
was investigated for both enantiomers of styrene oxide
and para-nitrostyrene oxide.[12] Another study showed
that the enantioselectivity with substituted styrene
oxides was higher for the Y215F mutant.[13] These
results prompted us to investigate the enantioselecti-
vity towards a wider variety of aromatic and aliphatic
substrates in order to obtain insight in the potential use
the mutant enzyme for the preparation of optically pure
epoxides.
A
O
O
H
H
O
−
H
O
−
O
O
N
N
H
O
O
H
H
N
Results and Discussion
Enantioselectivity of Wild-Type and Y215F Epoxide
Hydrolase
O
B
The wild-type and Y215F epoxide hydrolase were
brought to overexpression in E. coli and purified to
homogeneity. To investigate the enzymatic conversions,
3 mM solutions of racemic epoxides were subjected to
the wild-type and the Y215F epoxide hydrolase. The
concentration of both enantiomers was followed in time
by withdrawing samples from the reaction mixture and
analyzing them using chiral GC and chiral HPLC. The
enantioselectivity (E value) of the conversions was
determined by progress curve analysis as described
before.[9]
−
O
O
H
OH
H
O
−
O
O
N
O
N
H
H
N
O
H
Kinetic resolutions with the racemic epoxides, styrene
oxide, phenyl glycidyl ether, epichlorohydrin and 1,2-
epoxyhexane, which represent different classes of
epoxides, were chosen to investigate the enantioselec-
tivity of the hydrolysis catalyzed by the wild-type and
Y215F epoxide hydrolase. The hydrolysis of the ali-
phatic epoxides epichlorohydrin and 1,2-epoxyhexane
catalyzed by the wild-type and mutant enzyme occurred
with no difference in enantioselectivity (E < 2). A
significant increase in enantioselectivity was observed
with phenyl glycidyl ether. The E value increased from
11 using the wild-type to 16 with mutant enzyme. The
hydrolysis of styrene oxide showed the most significant
increase in enantioselectivity, the E value increased
from 16with the wild-type to32 with the mutant enzyme.
To further explore this class of substrates, a range of
racemic styrene oxides, substituted with electron-
donating and -withdrawing groups, were subjected to
the wild-type and Y215F epoxide hydrolase (Table 1).
The mutant enzyme showed an increased enantioselec-
tivity towards all substituted styrene oxides. Kinetic
resolutions with both enzymes resulted in optically pure
epoxides with an (S)-configuration. The kinetic reso-
lution of styrene oxide using wild-type enzyme showed a
remarkable progress curve. After conversion of the (R)-
enantiomer, the (S)-enantiomer was converted at a
O
Scheme 1. Reaction mechanism of the epoxide hydrolase
from A. radiobacter AD1. (A) alkylation half reaction and
(B) hydrolysis half reaction. Which tyrosine residue donates a
proton is not known.
Figure 1. Active site of the epoxide hydrolase from A.
radiobacter AD1 as observed in the X-ray structure.
structed in which either tyrosine residue was replaced by much higher rate (Figure 2A). A previous study of the
a phenylalanine (Y215F and Y152F). The influence of kinetics showed that Km and kcat had an opposite effect
these mutations on the kinetic parameters of hydrolysis on the enantioselectivity.[14] The 46-fold lower Km for
Adv. Synth. Catal. 2002, 344, 980 985
981

Jeffrey H. Lutje Spelberg et al.
FULL PAPERS
Table 1. Enantioselectivity of kinetic resolutions of para-
substituted styrene oxides with the epoxide hydrolase from
A. radiobacter AD1 and the Y215F mutant.
enzyme. After conversion of (R)-styrene oxide, con-
version of the (S)-styrene oxide also occurred, but at a
lower rate (Figure 2A). The progress curve of the
hydrolysis of para-methylstyrene oxide using wild-type
epoxide hydrolase showed that both enantiomers were
hydrolyzed at an almost equal rate (Figure 2B). With the
mutant enzyme, the hydrolysis rate of the remaining (S)-
enantiomer was substantially lower than that of the (R)-
enantiomer.
Epoxide
E value
E value
Y215F[a]
wild-type[a]
para-methylstyrene oxide
styrene oxide
para-chlorostyrene oxide
para-cyanostyrene oxide
para-nitrostyrene oxide
26 (S)
16 (S)
35 (S)
90 (S)
65 (S)
63 (S)
32 (S)
130 (S)
130 (S)
>200 (S)
[a]
Between parentheses is the absolute configuration of the
remaining enantiomer.
Kinetics of Wild-Type and Mutant Enzyme with
Substituted Styrene Oxide Enantiomers
To determine initial rates, optically pure enantiomers
were subjected to wild-type and Y215F enzyme. With
the wild-type enzyme, the initial rates at high substrate
concentrations were equal to Vmax. However, with the
exception of styrene oxide, no saturation of the reaction
velocity was observed with Y215F epoxide hydrolase.
The low solubility of the epoxides limits the measure-
ment of initial activities to concentrations lower than
2.5 mM.
Previous work elucidated the kinetic mechanism of
hydrolysis of both enantiomers of styrene oxide by the
wild-type enzyme using steady state and pre-steady state
techniques. The first kinetic step involves an attack on
the b-carbon atom of the epoxide, resulting in the
covalentesterintermediate(alkylation half reaction). In
the second catalytic step the covalent intermediate is
hydrolyzed by a water molecule that attacks the
carbonyl carbon of Asp107, yielding the diol (hydrol-
ysis half reaction). With both enantiomers of styrene
oxide the rate-limiting step is the hydrolysis of the
covalent intermediate, which has a rate constant
approximately equal to the overall kcat. An electron-
donating or -withdrawing substituent on the aromatic
ring can have a large influence on the alkylation rate, but
since this step is more than 100-fold faster than the rate-
limiting hydrolysis step, an effect on the overall rate
(kcat) cannot be observed. The influence of a substituent
on the aromatic ring on the rate of hydrolysis of the
covalent intermediate will be less significant. Thus, the
reaction rate is expected to be only slightly dependent
on the nature of the substituent. This hypothesis
coincides with Vmax values obtained from the initial
rate measurements using the substituted optically pure
(S)-enantiomers (Table 2). Increasing the electron-
withdrawing nature of the substituent resulted in only
a minor increase in initial rate. These results are
Figure 2. Kinetic resolution of styrene oxide (A) and para-
* *
~ ~
methylstyrene oxide (B) by wild-type ( , ) and Y215F ( , )
epoxide hydrolase. Closed symbols: (R)-enantiomer; open
symbols (S)-enantiomer.
(R)-styrene oxide caused a preferred conversion of this visualized in Figure 3, in which the logarithm of the
enantiomer. However, an almost 3-fold higher kcat for relative Vmax for each substituent (VmaxX/VmaxH) is
the (S)-enantiomer caused its higher hydrolysis rate.
plotted against the Hammett constant sp. For the
Styrene oxide was hydrolyzed by mutant enzyme with hydrolysis of the (S)-enantiomers, the same plot
a 2-fold higher enantioselectivity. The kinetic resolu- showed a slightly positive slope 1Vmax ( ‡ 0.12), confirm-
tions catalyzed by the Y215F mutant showed more ing that either a nucleophilic mechanism is unlikely to
standard progress curves than those with wild-type occur in the rate-determining step, or that the nature of
982
Adv. Synth. Catal. 2002, 344, 980 985

Biocatalytic Potential of the Epoxide Hydrolase from Agrobacterium radiobacter
FULL PAPERS
Table 2. Activities of wild-type and Y215F epoxide hydrolase towards optically pure para-substituted styrene oxides at 2.5 mM
substrate concentration.
À
Activity [mmol minÀ1 mg ]
% relative activity[b]
1
Substituent
Configuration
wild-type[a]
Y215F
methyl
H
R
S
R
S
R
S
R
S
9.2
5.4
6.7
18.9
11.9
17.4
27.2
22.7
21.0
21.8
2.9
0.1
32%
2%
69%
3%
70%
<1%
80%
<1%
70%
1%
4.6[c]
0.6
chloro
cyano
nitro
8.3
0.1
21.7
0.2
14.6
0.3
R
S
[a]
Activity is equal to Vmax, with the exception of (S)-para-methyl styrene oxide.
Relative activity of Y215F enzyme compared to the wild-type enzyme (100%).
[b]
[c]
Activity is equal to Vmax
.
conformational change needed to set up the covalent
intermediate for hydrolysis or c) an intramolecular
rearrangement of the covalent intermediate due to an
acyl transfer. The occurrence of such an intramolecular
migration was considered previously with rat micro-
somal epoxidehydrolase.[15] Besides this, acylmigrations
have been reported during reactions catalyzed by
lipases, baker×s yeast and antibodies.[16,17,18] In such a
type of rearrangement, the covalently bound Asp107
migrates from the b-position of the substrate to the a-
position. Assuming that this migration step would be
rapid (thus having no effect on the overall rate of the
process), the subsequent hydrolysis of this intermediate
is expected to give a major and positive Hammett effect.
The chemical base-catalyzed hydrolysis of para-substi-
tuted benzyl acetate derivatives, which can be regarded
as structural analogues of to the covalent intermediate,
showed a 1 value of ‡ 0.743.[19] This kind of acyl
migration will not cause an inversion at the chiral center
and does not influence the configuration of the diol. The
products with or without the occurrence of the acyl
Figure 3. Hammett plot representing log (VmaxX/VmaxH) of the
hydrolysis of (R)-para-substituted styrene oxides ( ) and (S)-
para-substituted styrene (~) oxides versus sp, catalyzed by
wild-type enzyme.
*
the substituent does not have an effect on this step. transfer are identical, making the proof for such a
Surprisingly, Vmax values with the (R)-enantiomers were mechanism difficult. Although the kinetically observed
significantly influenced by the nature of the substituent. unimolecular isomerization step can also correlate with
The rates increased when the ring substituent was more a conformational change or a protonation of a residue, a
electron-withdrawing. The high value of 1Vmax ( ‡ 0.55) significant influence of the nature of the ring-substitu-
suggests that a nucleophilic step is rate determining ents on the subsequent ester-cleavage step would then
during the conversion of the (R)-enantiomer.
no be expected, because of the substantial distance to
What could be the explanation for the difference the aromatic group. It can also not be excluded that the
between the two enantiomers? Rink et al. described the larger substituent effect on the rate of conversion of (R)-
conversion of the (S)-styrene oxide by a three-step and enantiomers is caused by a different positioning of the
the (R)-styrene oxide by a four-step mechanism to covalent intermediates in the active site, or by a
explain the observed transient kinetics.[14] The extra step differential effect of the protonation state of the active
involved a unimolecular isomerization of the covalent site cavity.
enzyme-substrate complex. Three possible mechanistic
In other studies, the kinetic mechanisms of epoxide
steps were considered: a) a (re)protonation step of a hydrolases from fungi were investigated using the same
tyrosine residue or the substrate oxygen, b) a slow range of substrates.[20,21] With the epoxide hydrolase
Adv. Synth. Catal. 2002, 344, 980 985
983

Jeffrey H. Lutje Spelberg et al.
FULL PAPERS
chlorostyrene oxide, para-cyanostyrene oxide and para-nitro-
styrene oxide were prepared as described previously.[9,22]
Optically pure enantiomers (ee > 98%) of para-methylstyr-
ene oxide, para-chlorostyrene oxide, and para-nitrostyrene
oxide and para-cyanostyrene oxide were prepared as described
before.[21]
from Aspergillus niger, the small magnitude of 1 ( ‡ 0.3)
suggested that the rate-limiting step was the ring
opening of the epoxide. Similar experiments with a
crude enzymatic extract prepared from Syncephalas-
trum racemosum suggested that the enzymatic mecha-
nism was a concerted process implying a general acid
activation of the oxirane ring, together with a nucleo-
philic attack.
Kinetic Resolution Experiments
The initial activities of the Y215F mutant towards (R)-
styrene oxide derivatives (2.5 mM) were lower than
those measured with the wild-type enzyme. With the
exception of para-methylstyrene oxide, the remaining
activity was around 70 to 80% of that of wild-type
activity. The mutant enzyme showed a drastic decrease
in activity towards the (S)-enantiomers. The effect was
most pronounced with styrene oxides bearing an
electron-withdrawing group, resulting in a residual
activity of less than 1%. The low initial rates towards
the remaining (S)-enantiomer improve the applicability
of this enzyme for kinetic resolutions.
A closed reaction vessel containing 20 mLTris buffer (50 mM,
pH 9.0) was incubated in a water bath at 30 8C. The epoxide
was added to a final concentration of 3 mM. para-Substituted
styrene oxides were added from stock solution in DMSO to a
final concentration of 0.5% DMSO. The enzyme was added to
such a concentration that allowed the first reacting enantiomer
to be converted within 50 min. The reaction was monitored by
periodically taking samples from the reaction mixture. The
samples were extracted with diethyl ether containing mesity-
lene as an internal standard and analyzed by chiral GC or chiral
HPLC.
Initial Rate Experiments
Conclusion
A closed reaction vessel containing 20 mLTris buffer (50 mM,
pH 9.0) was incubated in a water bath at 30 8C. The epoxide
was added from a stock solution in DMSO to a final
concentration of 1.5 mM to 2.5 mM and the enzyme was
added immediately afterwards. The reaction was monitored by
periodically taking samples from the reaction mixture,
followed by GC or HPLC analysis.
The wild-type epoxide hydrolase from A. radiobacter
AD1 is a moderately enantioselective biocatalyst for the
preparation of optically pure epoxides. The enantiose-
lectivity towards aliphatic epoxides is low. Aromatic
epoxides were hydrolyzed with moderate enantioselec-
tivity. Mutation of tyrosine 215 to a phenylalanine
resulted in an increased enantioselectivity and in this
study we showed that this improved performance holds
for a range of substituted styrene oxides. The increased
enantioselectivity and strong decrease in activity to-
wards the remaining (S)-enantiomer make the Y215F
enzyme very useful for biocatalytic applications. Indeed,
it allows circumventing the non-classical two phase
kinetic behavior of the wild-type enzyme that seriously
hampers its applicability. With the Y215F enzyme this
problem no longer exists, which in combination with the
increased enantioselectivity, makes the mutant a much
better biocatalyst for the preparation of optically pure
styrene oxide derivatives.
Chiral Analysis of Epoxides
The enantiomeric excess of the epoxides was determined using
the following columns: Chiraldex G-TA (Astec): styrene oxide,
epichlorohydrin and 1,2-epoxyhexane; Chiralsil Dex CB
(Chrompack): para-cyanostyrene oxide, para-chlorostyrene
oxide and para-nitrostyrene oxide; Chiralcel OD (Daicel):
phenyl glycidyl ether.
Calculation of E Values
The E values of the kinetic resolutions were calculated from
the Michaelis Menten parameters for both enantiomers.[9] To
À
estimate the kinetic parameters, the equations describing
À
competitive Michaelis Menten kinetics were fitted by numer-
ical integration to the data obtained from the kinetic resolution
experiments. With the epoxides styrene oxide, para-methyl-
styrene oxide and para-chlorostyrene oxide the first order
chemical hydrolysis rate constants of respectively 8.6 Â
Experimental Section
À
À
À
À
1
10À6 s , 1.2 Â 10 5 s and 6.2 Â 10À6 s were taken into
General
1
1
account.
Purified wild-type and the Y215F mutant enzyme were
prepared as described before.[9] The enantiomeric excess (ee)
and the yields of the epoxides were determined using a gas
chromatograph with an FID-detector or an HPLC both
equipped with chiral columns. NMR spectra were recorded
in CDCl3. The racemic epoxides styrene oxide, phenyl glycidyl
ether, epichlorohydrin and 1,2-epoxyhexane are commercially
available (Aldrich). Racemic para-methylstyrene oxide, para-
Acknowledgements
This research was financially supported by Innovation Oriented
Research Program (IOP) on Catalysis (no. 94007a)of the Dutch
984
Adv. Synth. Catal. 2002, 344, 980 985

Biocatalytic Potential of the Epoxide Hydrolase from Agrobacterium radiobacter
FULL PAPERS
Ministry of Economic Affairs. Part of this work has been
achieved in the context of the EC BIOC4-CT95-0005 and QLK-
CT-2000-00426 contracts.
[11] M. Nardini, I. S. Ridder, H. J. Rozeboom, K. H. Kalk, R.
Rink, D. B. Janssen, B. W. Dijkstra, J. Biol. Chem. 1999,
274, 14579.
[12] R. Rink, J. Kingma, J. H. Lutje Spelberg, D. B. Janssen,
Biochemistry 2000, 39, 5600.
[13] R. Rink, J. H. Lutje Spelberg, R. J. Pieters, J. Kingma, M.
Nardini, R. M. Kellogg, B. W. Dijkstra, D. B. Janssen, J.
Am. Chem. Soc. 1999, 121, 7417.
[14] R. Rink, D. B. Janssen, Biochemistry 1998, 37, 18119.
[15] H. -F. Tzeng, T. Laughlin, R. N. Armstrong, Biochemistry
1998, 37, 2905.
[16] N. Murakami, T. Morimoto, H. Imamura, A. Nagatsu, J.
Sakakibara Tetrahedron 1994, 50, 1993.
[17] R. Hayakawa, M. Shimizu, T. Fujisawa, Tetrahedron
Asymmetry 1997, 8, 3201.
References
[1] I. V. J. Archer, Tetrahedron 1997, 53, 15617.
[2] A. Archelas, R. Furstoss, Curr. Opin. Chem. Biol. 2001,
5, 112.
[3] A. Archelas, R. Furstoss, in Biocatalysis. From discovery
to application, (Ed.: W.-D. Fessner), Topics in Current
Chemistry, 200, Springer, Berlin, 1998, p. 159.
[4] K. M. Manoj, A. Archelas, J. Baratti, R. Furstoss,
Tetrahedron 2001, 57, 695.
[5] R. V. A. Orru, I. Osprian, W. Kroutil, K. Faber, Synthesis
1998, 1259.
[6] A. J. van den Wijngaard, D. B. Janssen, B. Witholt, J.
Gen. Microbiol. 1989, 135, 2199.
[7] M. H. J. Jacobs, A. J. van den Wijngaard, M. Pentenga,
D. B. Janssen, Eur. J. Biochem. 1991, 202, 1217.
[8] R. Rink, M. Fennema, M. Smids, U. Dehmel, D. B.
Janssen, J. Biol. Chem. 1997, 272, 14650.
[9] J. H. Lutje Spelberg, R. Rink, R. M. Kellogg, D. B.
Janssen, Tetrahedron: Asymmetry 1998, 9, 459.
[10] A. Archelas, R. Furstoss, J. Org. Chem. 1999, 64, 6112.
[18] E. Fernholz, D. Schloeder, K. K. C. Liu, C. W. Bradshaw,
H. Huang, K. Janda, R. A. Lerner, C. H. Wong, J. Org.
Chem. 1992, 57, 4757.
[19] H. H. Jaffe, Chem. Rev. 1953, 53, 191.
[20] A. Pedragosa-Moreau, C. Morisseau, J. Zylber, A.
Archelas, J. Baratti, R. Furstoss, J. Org. Chem. 1996,
61, 7402.
[21] P. Moussou, A. Archelas, J. Baratti, R. Furstoss, J. Org.
Chem. 1998, 63, 3532.
[22] R. B. Westkaemper, R. P. Hanzlik, Arch. Biochem.
Biophys. 1981, 208, 195.
Adv. Synth. Catal. 2002, 344, 980 985
985