Mutation of tyrosine residues involved in the alkylation half reaction of epoxide hydrolase from Agrobacterium radiobacter AD1 results in improved enantioselectivity [7]

Full text of DOI:10.1021/ja990501o

J. Am. Chem. Soc. 1999, 121, 7417-7418
Mutation of Tyrosine Residues Involved in the
7417
Alkylation Half Reaction of Epoxide Hydrolase from
Agrobacterium radiobacter AD1 Results in Improved
Enantioselectivity
†
†,‡
†,‡
Rick Rink, Jeffrey H. Lutje Spelberg, Roland J. Pieters,
†
§
‡
Jaap Kingma, Marco Nardini, Richard M. Kellogg,
Bauke W. Dijkstra, and Dick B. Janssen*
§
,†
Groningen Biomolecular Sciences and
Biotechnology Institute, UniVersity of Groningen
Nijenborgh 4, 9747 AG, Groningen, The Netherlands
Figure 1. Schematic representation of the first step of the reaction
mechanism of epoxide hydrolase.
ReceiVed February 17, 1999
Enantiomerically pure epoxides (oxiranes) are uniquely suited
bacter AD112 that Tyr152 and Tyr215 are positioned close to the
nucleophilic Asp107 in a manner such that their phenolic hydroxyl
groups could be proton donor. No backbone amides or other acid
groups are present that can serve as oxyanion hole or as proton
donor during ring opening. A mechanistical role for Tyr215 was
supported by a sequence alignment of known epoxide hydrolase
sequences, which revealed that this tyrosine residue is absolutely
conserved in the C-terminal part of the cap domain. This is
remarkable considering that the overall similarity between various
epoxide hydrolase sequences is often less than 20%, and indicates
an important role for this residue. The tyrosine residue is
conserved within a short stretch of sequence that is different for
soluble and microsomal epoxide hydrolases, namely N-W/Y-Y-R
and R-F/Y-Y-K, respectively. Sequence alignments were par-
ticulary poor in the N-terminal part of the cap domain where
Tyr152 is located, and only alignments done by hand indicated
that a second tyrosine might be present in the other soluble
epoxide hydrolases.
1
building blocks for synthetic purposes. Such epoxides are often
2
prepared by means of remarkably effective synthetic catalysts.
3
There are, however, few enzymatic routes in the repertoire. We
have studied the enzyme mediated kinetic resolution of readily
available racemic epoxides by selective hydrolysis of one enan-
tiomer to the 1,2-diol, a process for which a synthetic catalyst
has recently also been developed.4 Epoxide hydrolases that
perform this conversion have been found in various organisms.5
An attractive enzyme is the recombinant epoxide hydrolase from
6
Agrobacterium radiobacter AD1 that can be produced in large
amounts and which has good potential for the kinetic resolution
7
of styrene oxides. Here we report novel aspects of the catalytic
mechanism of the enzyme and a mechanism-based approach that
has led to the first site-specific mutant of an epoxide hydrolase
that has improved characteristics in kinetic resolutions.
8
The epoxide hydrolase from A. radiobacter AD1 belongs to
the R/â-hydrolase fold family and contains a catalytic triad in
9
the active site. The catalytic mechanism involves two discrete
To investigate the role of Tyr215 and Tyr152, we constructed
mutant enzymes in which the tyrosine was replaced by a
phenylalanine13 and the resulting mutant enzymes were expressed
N
chemical steps. The first is an S 2 nucleophilic attack by an
Asp107 carboxylate oxygen on the least-hindered carbon atom
of the epoxide, resulting in a covalent ester intermediate (Figure
14
and purified to homogeneity. The Tyr215Phe mutant showed a
100- to 1000-fold increase of the Km for both enantiomers of
styrene oxide (SO) and p-nitrostyrene oxide (pNSO), a strong
decrease of the kcat for the (S)-enantiomers, and a small decrease
of the kcat for the (R)-enantiomers (Table 1).15 Mutation of Tyr152
1
). In the second step, the ester intermediate is hydrolyzed by a
8
water molecule that is activated by the Asp246-His275 pair.
The chemical opening of an epoxide is facilitated by an acidic
functional group that interacts with the ring oxygen. Such an
activation likely also takes place in epoxide hydrolases. 10 Earlier
speculations were made that the proton donor could be a lysine
m
to Phe resulted in an enzyme that had an even higher K value
residue, but evidence in support of this is scant.8
,9c,11
We observed
from crystallographic data for epoxide hydrolase from A. radio-
for (R)-SO than the Tyr215Phe mutant whereas the kcat value again
remained in the same order of magnitude as the value for wild-
type enzyme (Table 1). The similar changes in the steady-state
kinetics of the Tyr215Phe and the Tyr152Phe mutant compared
to wild-type enzyme indicate that both tyrosines perform similar
roles in the kinetic mechanism of epoxide hydrolase. A mutant
*
Address correspondence to this author.
†
Department of Biochemistry.
‡
§
Department of Organic and Molecular Inorganic Chemistry.
BIOSON Research Institute and Laboratory of Biophysical Chemistry.
(
1) Besse, P.; Veschambre, H. Tetrahedron 1994, 50, 8885.
(
2) (a) Jacobsen, E. N. In ComprehensiVe Organometallic Chemistry II;
(11) Beetham, J. K.; Grant, D.; Arand, M.; Garbarino, J.; Kiyosue, T.;
Pinot, F.; Oesch, F.; Belknap, W. R.; Shinozaki, K.; Hammock, B. D. DNA
Cell Biol. 1995, 14, 61.
Wilkinson, G., Stone, F. G. A., Abel, E. W., Hegedus, L. S., Eds.; Pergamon:
New York, 1995; Vol. 12, Chapter 11.1. (b) Johnson, R. A.; Shareless, K. B.
Catalytic Asymmetric Synthesis; Ojima I., Ed.; VCH Publishers: New York,
(12) Nardini, M.; Ridder, I. S.; Rozeboom, H. J.; Kalk, K. H.; Rink, R.;
Janssen, D. B.; Dijkstra, B. W. J. Biol. Chem. 1999, 274, 14579.
(13) The Tyr215Phe and the Tyr152Phe mutants of epoxide hydrolase were
1
993; p 103. (c) Jacobsen, E. N. Catalytic Asymmetric Synthesis; Ojima, I.,
Ed.; VCH Publishers: New York, 1993; p 159.
8
(
(
3) Archelas, A.; Furstoss, R. Annu. ReV. Microbiol. 1997, 51, 491.
4) For an example with a transition metal catalyst see: Tokunaga, M.;
constructed as described before. The primers 5′-caactacttccgtgccaac-3′ and
5′-gagtcgtggttctcgcaattcc-3′ (mutated codons are underlined) were used for
constructing the Tyr215Phe mutant and the Tyr152Phe mutant, respectively.
Subsequently, the mutated epoxide hydrolase genes were sequenced.
(14) The mutant and wild-type enzyme were overexpressed in E. coli
Larrow, J. F.; Kakiuchi, F.; Jacobsen, E. N. Science 1997, 277, 936.
(5) (a) Archer, I. V. J. Tetrahedron 1997, 53, 15617. (b) Faber, K.; Mischitz,
M.; Kroutil, W. Acta Chem. Scand. 1996, 50, 249. (c) Archelas, A.; Furstoss,
R. Tibtech 1998, 16, 108.
8
BL21(DE3) and purified as described before. The enzymes were stored in
(
6) Jacobs, M. H. J.; van den Wijngaard, A. J.; Pentenga, M.; Janssen, D.
TEMAG buffer at 4 °C and remained fully active for at least two month.
B. Eur. J. Biochem. 1991, 202, 1217.
7) Lutje Spelberg, J. H.; Rink, R.; Kellogg, R. M.; Janssen, D. B.
Tetrahedron: Asymmetry 1998, 9, 459.
8) Rink, R.; Fennema, M.; Smids, M.; Dehmel, U.; Janssen, D. B. J.
Biol. Chem. 1997, 272, 14650.
m
(15) The steady-state parameters kcat and K for styrene oxide (SO) and
(
p-nitrostyrene oxide (pNSO) were obtained from progress curves, using an
8
amount of enzyme sufficient to complete the reaction within 20 min. SO
8
(
was analyzed by gas chromatography. Substrate depletion curves for pNSO
were recorded in TE buffer at 30 °C on a Kontron Uvikon 930 UV/VIS
spectrophotometer. The reaction was started by the addition of a stock solution
of pNSO in acetonitrile to the cuvette with the enzyme solution to a final
concentration of 1% acetonitrile. By using the extinction coefficients for pNSO
(
9) (a) Lacourciere, G. M.; Armstrong, R. N. J. Am. Chem. Soc. 1993,
1
1
15, 10466. (b) Tzeng, H.-F.; Laughlin, L. T.; Armstrong, R. N. Biochemistry
998, 37, 2905. (c) Laughlin, L. T.; Tzeng, H.-F.; Lin, S.; Armstrong, R. N.
-
1
-1
-1
-1
Biochemistry 1998, 37, 2897.
(ꢀ310 ) 4289 M cm ) and the corresponding diol (ꢀ310 ) 3304 M cm ),
(
10) Moussou, P.; Archelas, A.; Baratti, J.; Furstoss, R. J. Org. Chem.
the recorded traces were directly fitted with the Michaelis-Menten equation
1
998, 63, 3532.
m
to obtain kcat and K values.
1
0.1021/ja990501o CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/30/1999

7418 J. Am. Chem. Soc., Vol. 121, No. 32, 1999
Communications to the Editor
Table 1. Steady State Parameters of Wild-Type and Mutant Epoxide Hydrolase for Both Enantiomers of PNSO and SO Determined at pH 9
and 30 ˚C
wild-type enzyme
(mM)
0.008
>0.5
0.0005
0.021
Tyr215Phe mutant enzyme
Tyr152Phe mutant enzyme
(mM) )
(mM-1 s-1
substrate
k
cat (s-1)
K
m
k
cat/K
m
(mM-1 s-1
)
k
cat (s-1
)
K
m
(mM)
>0.8
>1
0.6
k
cat/K
m
(mM-1 s-1
)
k
cat (s-1
)
K
m
k
cat/K
m
a
0.8a
0.005
0.3
(
R)-pNSO
S)-pNSO
R)-SO
7.8
>7.7
3.8
975
5
>3.4
>0.02
2.5
4.2
>1.2
>1.5
>1
a
0.02a
4.2
a
a
(
(
(
>0.005
8000
500
1.1
3.5
>15
S)-SO
10.5
0.7
5
0.14
>0.12
0.008
a
Due to high K
m
values and maximum solubility of pNSO (1 mM) and SO (10 mM) only the kcat/K
values.
m
values could be determined, setting lower
limits for the kcat and K
m
Table 2. Enantioselectivity of Wild-Type and Tyr215Phe Epoxide
Hydrolase
E value
racemic substrate
wild-type
Tyr215Phe
a
pNSO
SO
100
16
6.5
32
>200
30
12
130
b
b
mClSO
b
pClSO
a
15
Determined by measuring kcat/K
m
with the pure enantiomers.
b
17
Determined from a kinetic resolution experiment.
in which both tyrosines were replaced by phenylalanines showed
no detectable catalytic activity. These observations suggest that
Tyr152 and Tyr215 are involved in hydrogen bonding and/or
proton donation to the ring oxygen.
Figure 2. Kinetic resolution of 5 mM p-chlorostyrene oxide with 1 µM
of wild-type epoxide hydrolase (b, 9) and with 3 µM of Tyr215Phe
mutant enzyme (O, 0). The round symbols represent the (R)-enantiomer
and the square symbols represent the (S)-enantiomer.
Besides remaining catalytically active, the Tyr215Phe mutant
enzyme showed an increased stereoselectivity with styrene oxide
16
and substituted variants thereof (Table 2). The E values of the
mutant enzyme for SO, m-chlorostyrene oxide (mClSO), and
pNSO were increased approximately 2-fold compared to the E
values of wild-type enzyme.17 A spectacular improvement of the
E value was found for p-chlorostyrene oxide (pClSO), which
increased 4-fold from 32 for wild-type enzyme to over 130 for
the Tyr215Phe mutant (Figure 2). This corresponded to 96% of
the theoretical yield of (S)-pClSO, with an enantiomeric excess
of 99%. The kcat values with the (R)-enantiomers are hardly
affected, and the Tyr215Phe mutant enzyme remains an excellent
catalyst, considering that a low concentration of the Tyr215Phe
mutant is still sufficient to convert (R)-pClSO in a reasonable
amount of time (Figure 2).
of one hydrogen bond in the Tyr152Phe and the Tyr215Phe
mutant would affect the binding of the substrate in the Michaelis
complex and destabilize the transition state of the alkyl-enzyme
formation leading to lower alkylation rates. How the dealkylation
rate, the back reaction, is influenced by the mutations depends
on the stability of the covalent intermediate. If a reduction of the
alkylation rate is also the main effect of the mutation for (S)-SO
conversion, the rate of formation of the alkyl enzyme must have
become rate-limiting in agreement with a reduced kcat. Thus, the
increase of the K
reduction in the kcat for the (R)-enantiomer, which caused a smaller
decrease of kcat/K for the (R)-enantiomer than for the (S)-
m
with SO is accompanied by a relatively small
m
enantiomer and hence an increase in enantioselectivity. The same
changes may explain the observed increase in enantioselectivity
with the other substrates. Indeed, the kcat for the hydrolysis of
the (R)-enantiomer of pClSO is close to that of wild-type enzyme,
but the kcat for the (S)-enantiomer has decreased virtually to zero
in the mutant (Figure 2).
The above considerations suggest a general principle for the
observed increase in enantioselectivity for the Tyr215Phe mutant.
Owing to the strongly reduced stabilization of the transition state
and/or the Michaelis complex, the alkylation rate of the enzyme
has become rate-limiting for (S)-enantiomers, which are already
the poorer substrates for wild-type enzyme. The resulting reduc-
What is the cause of this behavior? During the conversion of
SO by the wild-type epoxide hydrolase, the hydrolysis of the
covalent intermediate to product is the rate-limiting step and its
18
rate is close to the overall kcat
type enzyme are due to very high alkylation rates which are over
00-fold faster than the hydrolysis rates. This results in a high
m
. The low K values of the wild-
1
accumulation of the alkyl-enzyme during turnover as was also
observed for rat microsomal epoxide hydrolase.19 The Tyr215Phe
and the Tyr152Phe mutant enzymes showed a large increase of
m
the K for (R)-SO whereas the kcat values were only slightly
reduced. This suggests that for (R)-SO only the alkylation half
reaction, which comprises the Michaelis complex and the revers-
ible formation of the alkyl-enzyme, was affected by the mutation.
The X-ray structure of epoxide hydrolase indicates that only the
hydroxyl groups of Tyr152 and Tyr215 are in a position to form
hydrogen bonds with the oxirane oxygen. Consequently, the loss
m
tion of the kcat/K with (S)-enantiomers for the mutant enzyme
enhances the enantioselectivity. The hypothesis that Tyr215 and
Tyr152 act as proton donor in the reaction mechanism of epoxide
hydrolase, which is indicated by the X-ray structure, is supported
by the steady-state kinetics of the mutant enzymes.
In conclusion, we have observed that mutation of a tyrosine,
which interacts with the epoxide during ring opening, yields an
enzyme with reduced turnover number for the nonpreferred
enantiomer and with increased enantioselectivity. Since Tyr215
is conserved among epoxide hydrolase sequences, an activating
role of this residue is suggested for other epoxide hydrolases as
well. Replacing this tyrosine residue, an operation in principle
deleterious to the catalytic machinery, might also result in mutants
with improved enantioselectivity for these enzymes.
(
16) The E value is defined as the ratio (kcat,R/Km,R)/(kcat,S/Km,S), with R
and S representing the enantiomers. For a discussion on E values and
enantiomeric ratios, see: Straathof, A. J. J.; Jongejan, J. A. Enzyme Microb.
Technol. 1997, 21, 559.
8
(17) Kinetic resolution experiments were done as described before. For
p- and m-chlorostyrene oxide, DMSO was used as a cosolvent (10%) to reach
a substrate concentration of 5 mM. The conversion of both enantiomers with
time was described by competitive Michaelis-Menten kinetics, including a
8
term for chemical hydrolysis. In most cases, this procedure did not result in
1
6
a solution for all steady-state parameters, but it gave unique E values.
(
18) Rink, R.; Janssen, D. B. Biochemistry 1998, 37, 18119.
19) Tzeng, H.-F.; Laughlin, T.; Lin, S.; Armstrong, R. N. J. Am. Chem.
(
Soc. 1996, 118, 9436.
JA990501O