Epoxide hydrolase-catalyzed enantioselective conversion of trans-stilbene oxide: Insights into the reaction mechanism from steady-state and pre-steady-state enzyme kinetics

Article abstract of DOI:10.1016/j.abb.2015.12.008

A detailed kinetic study based on steady-state and pre-steady-state measurements is described for the highly enantioselective epoxide hydrolase Kau2. The enzyme, which is a member of the α/β-hydrolase fold family, preferentially reacts with the (S,S)-enantiomer of trans-stilbene oxide (TSO) with an E value of ～200. The enzyme follows a classical two-step catalytic mechanism with formation of an alkyl-enzyme intermediate in the first step and hydrolysis of this intermediate in a rate-limiting second step. Tryptophan fluorescence quenching during TSO conversion appears to correlate with alkylation of the enzyme. The steady-state data are consistent with (S,S) and (R,R)-TSO being two competing substrates with marked differences in k_cat and K_M values. The high enantiopreference of the epoxide hydrolase is best explained by pronounced differences in the second-order alkylation rate constant (k₂/K_S) and the alkyl-enzyme hydrolysis rate k₃ between the (S,S) and (R,R)-enantiomers of TSO. Our data suggest that during conversion of (S,S)-TSO the two active site tyrosines, Tyr¹⁵⁷ and Tyr²⁵⁹, serve mainly as electrophilic catalysts in the alkylation half-reaction, polarizing the oxirane oxygen of the bound epoxide through hydrogen bond formation, however, without fully donating their hydrogens to the forming alkyl-enzyme intermediate.

Full text of DOI:10.1016/j.abb.2015.12.008

Archives of Biochemistry and Biophysics 591 (2016) 66e75
Contents lists available at ScienceDirect
Archives of Biochemistry and Biophysics
journal homepage: www.elsevier.com/locate/yabbi
Epoxide hydrolase-catalyzed enantioselective conversion of trans-
stilbene oxide: Insights into the reaction mechanism from steady-
state and pre-steady-state enzyme kinetics
a
a
a
a
b, *
Alain Archelas , Wei Zhao , Bruno Faure , Gilles Iacazio , Michael Kotik
a
Aix-Marseille Universit ꢀe , Centrale Marseille, CNRS, ISM2 UMR7313, Marseille 13397, France
Laboratory of Biotransformation, Institute of Microbiology of the Czech Academy of Sciences, v. v. i., The Czech Academy of Sciences, Víde nꢁ sk aꢀ 1083, CZ-142
b
2
0 Prague, Czech Republic
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 November 2015
Received in revised form
A detailed kinetic study based on steady-state and pre-steady-state measurements is described for the
highly enantioselective epoxide hydrolase Kau2. The enzyme, which is a member of the a/b-hydrolase
fold family, preferentially reacts with the (S,S)-enantiomer of trans-stilbene oxide (TSO) with an E value
of ~200. The enzyme follows a classical two-step catalytic mechanism with formation of an alkyl-enzyme
intermediate in the first step and hydrolysis of this intermediate in a rate-limiting second step. Tryp-
tophan fluorescence quenching during TSO conversion appears to correlate with alkylation of the
enzyme. The steady-state data are consistent with (S,S) and (R,R)-TSO being two competing substrates
16 December 2015
Accepted 17 December 2015
Available online 20 December 2015
Keywords:
with marked differences in kcat and K
best explained by pronounced differences in the second-order alkylation rate constant (k
alkyl-enzyme hydrolysis rate k between the (S,S) and (R,R)-enantiomers of TSO. Our data suggest that
during conversion of (S,S)-TSO the two active site tyrosines, Tyr
M
values. The high enantiopreference of the epoxide hydrolase is
Catalytic mechanism
Epoxide hydrolase
Electrophilic catalysis
Enantioselectivity
Fluorescence quenching
Tyrosine ionization
2
/K ) and the
S
3
157
259
and Tyr , serve mainly as electro-
philic catalysts in the alkylation half-reaction, polarizing the oxirane oxygen of the bound epoxide
through hydrogen bond formation, however, without fully donating their hydrogens to the forming
alkyl-enzyme intermediate.
©
2015 Elsevier Inc. All rights reserved.
1
. Introduction
in a highly enantioselective manner with an E value of ~200,
resulting in the formation of meso-hydrobenzoin (2) and residual
(R,R)-1 with an ee of >99% in isolated yields close to the theoretical
maximum of 50% (Fig. 1) [5]. (R,R)-1 and 2 are interesting organic
intermediates which have found applications as starting materials
in various syntheses [6e10]. These valuable properties of the Kau2
EH prompted us to analyze its catalytic mechanism of the enan-
tioselective hydrolysis of 1 in more detail. Moreover, kinetic data on
the hydrolysis of 1 have been accumulated for the structurally
related potato EH, whose enantioselectivity was shown to be low
for this compound [11].
Epoxide hydrolases (EHs) catalyze the opening of the epoxide
ring in a substrate by the addition of water, generating a diol as the
final reaction product. Specifically, enantioselective or enantio-
convergent EHs have found their way into various bio-
transformations as valuable biocatalysts [1e3]. Kau2 is
metagenome-derived EH and a member of the -hydrolase fold
family [4]. It turned out to be an interesting catalyst, often exhib-
iting a high enantiopreference (represented by a high E value) and/
or enantioconvergence in conjunction with low product inhibition
a
a/b
[
4,5]. The enzyme has been used in a number of preparative-scale
The catalytic mechanism of a/b-hydrolase fold EHs is known to
ꢀ1
conversions with substrate loads of up to 80 g l [4,5]. As an
example, the enzyme reacted with racemic trans-stilbene oxide (1)
involve the action of a catalytic triad, which consists of a nucleo-
phile (Asp), a general base (His) and an acid (carboxylate function of
Asp or Glu) [12]. After substrate binding to the enzyme, the
nucleophilic carboxylic acid of Asp is polarized by the Hiseacid
charge-relay system; the activated nucleophile then makes an
attack on the epoxide, resulting in the formation of a covalent in-
termediate with an ester bond between the carboxylic acid of the
Abbreviations: A295, absorbance at 295 nm; E, enantiomeric ratio; EH, epoxide
hydrolase; TSO, trans-stilbene oxide.
*
Corresponding author.
E-mail address: kotik@biomed.cas.cz (M. Kotik).
http://dx.doi.org/10.1016/j.abb.2015.12.008
003-9861/© 2015 Elsevier Inc. All rights reserved.
0

A. Archelas et al. / Archives of Biochemistry and Biophysics 591 (2016) 66e75
67
Fig. 1. Kau2 EH-catalyzed hydrolysis of racemic trans-stilbene oxide (1), and other epoxide and diol compounds used in this work. Kinetic resolution of racemic 1 results in the
formation of (R,R)-1 and meso-hydrobenzoin (2). (S)-styrene oxide: (S)-3; (S,S)-trans-1-phenyl-1,2-epoxypropane: (S,S)-4; (R,S)-1-phenylpropane-1,2-diol: (R,S)-5.
nucleophile and a carbon atom of the opened oxirane. This ester is
formed in the so-called alkylation half-reaction and is termed the
alkyl-enzyme intermediate; its formation appears to be associated
with the observed tryptophan fluorescence quenching when
rapidly mixing EH and substrate [13,14]. As a next step in the kinetic
mechanism, the alkyl-enzyme intermediate is hydrolyzed, which is
usually rate-limiting. In this so-called hydrolytic half-reaction the
Hiseacid charge-relay system activates a water molecule which
attacks the carbonyl of the alkyl-enzyme ester. After hydrolysis the
diol product is released with the enzyme being restored to its
original form (Fig. 2). There is substantial evidence that the catalytic
triad is assisted by two conserved active site tyrosines. The orien-
tation of their hydroxyl groups in crystal structures of EHs indicates
an involvement of these residues in the polarization of the epoxide
oxygen prior to the hydrolysis of the CeO bond [12]. Specifically, it
was proposed that one of the two conserved tyrosines would
establish a hydrogen bond with the epoxide oxygen of the bound
substrate and thereby enhance its polarization (Fig. 2); in a
consecutive step a proton transfer from the tyrosine hydroxyl to the
generated alkoxide (alkyl-enzyme intermediate) would occur [12].
Recently, it was suggested that one more conserved histidine res-
idue, which is linked to the general base histidine through the
hydrolytic water molecule in the active site of the potato EH, plays a
significant role in EH catalysis [15].
controversial, including the role of the two active site tyrosines.
Questions to be asked in this respect are: Are tyrosinate ions being
transiently formed during catalysis? Do these tyrosines serve as a
general acid catalyst or as an electrophilic catalyst where no com-
plete proton transfer to the transiently formed alkoxide interme-
diate occurs? In this paper we analyzed the kinetics of the
conversion of 1 catalyzed by the Kau2 EH using steady-state and
pre-steady-state measurements with the objective to untangle the
kinetic basis of the very high enantiopreference of this enzyme and
the catalytic role of its active site tyrosines during the enzymatic
hydrolysis.
2. Materials and methods
2.1. Chemicals
0
0
N-cyclohexyl-N -decylurea (CDU) and N-cyclohexyl-N -(4-
iodophenyl)urea (CIU) were synthesized as previously described
16,17]. The racemic compound 1, the meso-diol 2, and (S)-styrene
[
oxide ((S)-3) (Fig. 1) were purchased from SigmaeAldrich. Enan-
tiopure (S,S)-1 and (R,R)-1 were obtained by preparative chiral
HPLC from rac-1 using a Chiralpak IC column (Daicel Corp.). (S,S)-
trans-1-phenyl-1,2-epoxypropane ((S,S)-4) (Fig. 1) was synthesized
as previously described [18]. (R,S)-1-phenylpropane-1,2-diol ((R,S)-
Several issues concerning the EH catalytic mechanism are still
5) (Fig. 1) was obtained by preparative kinetic resolution of rac-4
Fig. 2. Catalytic mechanism of
enzymeesubstrate complex with an equilibrium dissociation constant K
alkyl-enzyme intermediate (E-alkyl); (3) the enzyme intermediate is irreversibly hydrolyzed (k
a
/
b
-hydrolase fold EHs. Usually, catalytic reaction mechanisms of EHs with
; (2) nucleophilic attack of the active-site Asp results in the reversible formation (k
) by a base-activated water molecule, which leads to the release of the diol product.
a
/
b
-hydrolase folds are represented by three steps: (1) formation of an
S
2
, kꢀ2) of a covalent
3

68
A. Archelas et al. / Archives of Biochemistry and Biophysics 591 (2016) 66e75
[4].
by following the conversion of compound 1 to the diol 2 at 228 nm
using a Cary 50 Bio UVevisible spectrophotometer (Agilent Tech-
nologies). The contents of the cuvette (2.0 ml) were continuously
stirred using a tiny magnetic bar. A quadratic equation was found to
2.2. Construction of mutants
The Kau2 EH-encoding gene was PCR-amplified using the
best describe the relationship between absorbance and the (S,S)-1
ꢁ
PfuUltra II Fusion HS DNA polymerase (Agilent Technologies), the
TOPO-cloning primers (Table S1), and the pSEKau2 plasmid as the
template [4]. The PCR products were gel-purified and used for
cloning into the expression vectors pET100/D-TOPO and pET101/D-
TOPO (Invitrogen). Enzyme variants were generated by using either
the Phusion site-directed mutagenesis kit (Thermo Fisher Scientific
Inc.) or the QuickChange Lightning site-directed mutagenesis kit
concentration (0e100
trile (v/v) (Fig. S3). An extinction coefficient of 16,865 M cm
was determined for (S,S)-1 at 23 C in 100 mM phosphate buffer
m
M) at 30 C in the presence of 1% acetoni-
ꢀ1
ꢀ1
ꢁ
(pH 8.0) containing 4% acetonitrile (v/v) (Fig. S4). Under both re-
ꢀ1
ꢀ1
action conditions an extinction coefficient of 650 M cm was
determined for 2. All solutions containing an epoxide were freshly
prepared in flasks made of glass. K
M
and kcat values for (S,S)-1 were
(
Agilent Technologies) according to the manufacturers' in-
calculated from initial velocity measurements using the Michae-
structions. The primers used for mutagenesis are shown in Table S1.
E. coli BL21 Star (DE3) cells (Life Technologies) were transformed
with the generated plasmids whose inserts were verified by
sequencing.
liseMenten equation. The specificity constant for (R,R)-1 (kcat,R
/
KM,R) was calculated using the following relationship [21]:
ꢀ
kcat;S kcat;R
E ¼
(1)
KM;S KM;R
2.3. Enzyme production and purification
The initial velocity data from experiments with reaction mix-
tures containing the two competing substrates (S,S)-1 and (R,R)-1
were analyzed using the following relationships; see [22] for Eq.
(2), and [23,24] for Eq. (3):
The E. coli cells harboring the pET100-based plasmids were
ꢁ
grown at 28 C for 24 h in 5 ml of Luria Bertani medium containing
ꢀ1
2
00 mg ml of ampicillin. These cells were used to inoculate 500 ml
ꢀ
1
of Terrific Broth medium containing 200
mg ml of ampicillin.
v
v
k
cat;S
K
M;R ½Sꢃ
When the OD6 reached a value of ~1, the cultivation temperature
00
S
¼
(2)
(3)
kcat;R KM;S ½Rꢃ
ꢁ
R
was reduced from 37 to 23 C, and IPTG was added (final concen-
tration: 1.0 mM). After 4e5 h the biomass was harvested by
À
ꢁ
Á
À
ꢁ
Á
ꢁ
centrifugation, frozen in liquid nitrogen and stored at ꢀ80 C. The
k
cat;R
K
M;R ½Rꢃ þ kcat;S
K
M;S ½Sꢃ
v
t
¼
ꢁ
ꢁ
cells were resuspended in binding buffer (see below) in the pres-
ence of a protease inhibitor cocktail (SigmaeAldrich) and subse-
quently broken in one shot using a TS Series cell disruptor
1 þ ½Rꢃ K
þ ½Sꢃ K
M;R
M;S
where v is the total initial rate, [S] and [R] are the concentrations of
t
(
Constant Systems Ltd.) at a disruption pressure of 1300 bar. The
(S,S)-1 and (R,R)-1, and v and v are the partial reaction rates for
S
R
soluble fraction of the lysate, which was obtained by centrifugation
(S,S)-1 and (R,R)-1, respectively.
ꢁ
(
15,500 ꢂ g at 4 C, 30 min), was loaded on a Ni Sepharose column
€
(
HisTrap HP, 5 ml; AKTA 900 FPLC system; GE Healthcare Life Sci-
2.6. Pre-steady-state kinetics
ences) equilibrated with binding buffer (20 mM sodium phosphate,
00 mM NaCl, 30 mM imidazole, pH 7.4). The column was washed
with binding buffer, and the bound Kau2 EH was eluted with a
linear gradient of elution buffer (20 mM sodium phosphate,
5
Rapid mixing experiments were conducted on a SFM-300
stopped-flow apparatus (Bio-Logic Science Instruments SAS,
France) equipped with 3 independently controlled syringes and 2
BalleBerger mixers. The Trp residues were excited at 290 nm and
the total fluorescence emission was detected through a 320-nm
cutoff filter. All experiments were performed in 100 mM phos-
phate buffer (pH 8.0) containing either 1 or 4% acetonitrile at 30 or
5
3
00 mM NaCl, 500 mM imidazole, pH 7.4) using a flow rate of
ml min and a gradient time of 30 min (Fig. S1). The purity of the
ꢀ1
enzymes was estimated to be >95% (Fig. S2). A HiTrap Desalting
column (5 ml; GE Healthcare Life Sciences) was used for buffer
exchange of the enzyme sample. Protein concentrations of the
purified wild-type Kau2 EH were determined at 280 nm with a
NanoDrop 2000c spectrophotometer (Thermo Scientific) using a
ꢁ
23 C, respectively. For each ligand concentration in the observation
cell a separate manually prepared stock solution was used without
changing the ligand/enzyme mixing ratio in the apparatus. Usually,
5 traces were recorded and averaged before fitting. The same in-
strument was used for detecting the absorbance signal at 295 nm.
The stopped-flow fluorescence signal was fitted to a single expo-
nential using
ꢀ
1
ꢀ1
cm
calculated molar extinction coefficient of 58,900 M
determined using the ProtParam tool of ExPASY, http://www.
expasy.org/). One unit of Kau2 EH activity is defined as the
amount of enzyme necessary to hydrolyze 1 mol of (S,S)-1 per min
(
m
in 100 mM phosphate buffer (pH 8.0) containing 4% acetonitrile (v/
v) at 23 C.
ꢁ
F ¼ A eꢀk þ C
obs
t
(4)
2
.4. Difference spectra and determination of global pK
a
for Tyr
where A is the amplitude, kobs is the observed rate constant, and C is
the floating end point.
Difference spectra were recorded between 250 and 350 nm
using UV-1700 spectrophotometer (Shimadzu Corp.). The
following buffers (100 mM) were used: phosphate in the pH range
.0e8.0; borate for the pH range 8.0e9.0, and glycine for the pH
range 9.0e10.5. The concentration of tyrosinate was determined at
a
2.7. Computational ligand docking and molecular dynamics
simulations
7
Computational docking experiments were performed using a
modified version of the original Autodock program, implemented
in YASARA (dock_run.mcr macro). A previously constructed ho-
mology model of the Kau2 EH was used [17]. The structures of the
substrate molecules were generated in ChemDraw Ultra 9.0 and
energy-minimized in Chem3D Pro 9.0. Free rotation of the
ꢀ1
ꢀ1
2
95 nm using Dε of 2400 M cm [19].
2.5. Steady-state kinetics
Initial velocity measurements were performed according to [20]

A. Archelas et al. / Archives of Biochemistry and Biophysics 591 (2016) 66e75
69
hydrogen-bonding OH groups in both oxirane-polarizing tyrosines
was allowed during substrate docking. The selection of appropri-
ately bound epoxide conformers was based on the presence of
reaction temperature resulted in a higher kcat value for (S,S)-1,
whereas a lower K value was determined under the high-
temperature conditions (Table 1).
Further, we performed initial velocity measurements with
mixtures of (R,R)-1 and (S,S)-1 and analyzed the data using Eq. (2).
M
157
hydrogen bonds with the oxirane-polarizing tyrosines (Tyr and/
or Tyr2 ) and a proper orientation of the oxirane ring, allowing
59
109
attack of the nucleophilic aspartate (Asp ) at either carbon atom
of the epoxide moiety. Molecular dynamics simulations were per-
formed in a water box using YASARA (version 14.7.17) and the
YAMBER3 force field. The epoxide ring of the docked (S,S)-1
enantiomer was manually opened at the C1eO bond; then, a new
covalent bond between the C1 carbon atom and the oxygen atom
As shown in Fig. S7 the (R,R)-1-specific reaction rate v
to be low compared to the (S,S)-1-specific reaction rate v
only 1% of v under conditions of a two-fold excess of (R,R)-1 over
(S,S)-1. The low v is a consequence of the high E value for the Kau2
EH reacting with rac-1. These initial velocity data were also
analyzed using Eq. (3) with the determined kcat and K values for
(S,S)-1, and the estimated kcat/K value for (R,R)-1 (Table 1). In these
simulations a good agreement between experimental velocity data
and simulated curves was obtained with K for (R,R)-1 being equal
to 150 M, when allowing for parameter variation within the
R
was found
S
, being
S
R
M
3
16
109
(
located near His ) of the catalytic Asp
was generated. The
M
ꢁ
md_refine.mcr macro was run in the 500-ps simulations at 25
and pH 7.4 using a water density of 0.997 g l ; one pdb file of the
ligandeprotein complex was saved every 25 ps.
C
ꢀ
1
M
m
margins of the experimental error (Fig. 3). Hence the initial velocity
data determined in the presence of (R,R)-1 and (S,S)-1 are consis-
tent with (R,R)-1 and (S,S)-1 being competing substrates for the
3
. Results
active site of the Kau2 EH with a three-fold difference in K
between (R,R)-1 and (S,S)-1.
M
values
3.1. Solubility and photostability of compound 1
We started our kinetic experiments using published reaction
conditions (1% acetonitrile and 30 C) [25]. A quadratic equation
was found to best describe the non-linear relationship between
ꢁ
3.3. Rapid kinetics of (S,S)-1 conversion
Multiple-turnover experiments were performed with 10e90 mM
of (S,S)-1 under conditions used for the determination of the
steady-state parameters. The recorded fluorescence traces followed
absorbance at 228 nm and the (S,S)-1 concentration (0e100
suggesting limited solubility of this compound (Fig. S3). Moreover,
samples of 1 with concentrations of >100 M appeared to be
mM),
m
a single exponential curve with the apparent rate constants (kobs
being linearly dependent on the (S,S)-1 concentration (Fig. 4). Eq.
5) describes the non-linear substrate concentration ([S]) depen-
)
opalescent, confirming its limited solubility (Fig. S4). To improve
the solubility of 1 we modified the experimental conditions and
increased the acetonitrile concentration to 4%; in addition, we
(
ꢁ
dence of kobs of a three-step reaction model (Fig. 2):
lowered the reaction temperature to 23 C to improve the long-
term stability of the Kau2 EH. Under these conditions, a perfect
linear relationship between the absorbance and the concentration
of 1 within 0e100 mM was observed (Fig. S5).
k ½Sꢃ
2
kobs
¼
þ kꢀ2 þ k3
(5)
½
Sꢃ þ KS
The photostability of 1 was assessed by following the absor-
bance at 228 nm in the absence of enzyme using the stopped-flow
apparatus (Fig. S6). It turned out that the absorbance signal
decreased with time as a function of the light intensity, suggesting
photolability of this compound. However, no time-dependence of
the absorbance signal at 228 nm was detected in the spectropho-
tometer used for steady-state measurements (data not shown).
The observed linear relationship of kobs with (S,S)-1 implies that
enzyme saturation occurred at substrate concentrations of [
90
mM, indicating a high K
S
value. Therefore, the data of kobs versus
(S,S)-1 concentration was fitted to the simplified Eq. (6):
3.2. Steady-state kinetics
The Kau2 EH converted rac-1 enantioselectively with (S,S)-1
being the preferred substrate; we determined in the spectroscopic
activity test a >300 times higher enzyme activity in the presence of
(
S,S)-1 compared with (R,R)-1. The hardly detectable enzyme ac-
tivity with (R,R)-1 precluded a direct determination of any steady-
state parameters for this compound. Nevertheless, the kcat/K value
for (R,R)-1 could be estimated using the kcat/K value obtained for
S,S)-1 and the previously determined E value for the kinetic res-
M
M
(
olution of 1 (Eq. (1); Table 1) [5]. Not surprisingly, an increase in
Table 1
Steady-state kinetic parameters for the hydrolysis of (S,S)-1 or (R,R)-1 catalyzed by
the Kau2 EH.
cat (sꢀ1)
k
cat/K
(
m
M
ꢀ1
s
ꢀ1
)
Conditions
Substrate
K
M
(m
M)
k
M
ꢁ
2
3
0
C, 4% CH
C, 1% CH
3
CN
CN
(S,S)-TSO
45 ± 4
150
15 ± 1
n.d.
12 ± 1
0.21
14 ± 1
n.d.
0.27 ± 0.05
0.0014
a
a
b
Fig. 3. Initial velocity experiments with reaction mixtures containing (S,S)-1 and (R,R)-
. The concentration of (R,R)-1 was varied between 0 and 90 M in the presence of
). Reaction conditions: 100 mM phosphate buffer
(
R,R)-TSO
(S,S)-TSO
R,R)-TSO
1
m
ꢁ
3
3
0.93 ± 0.15
b
either 10
mM (B) or 20
m
M (S,S)-1 (
▫
(
0.0047
ꢁ
(
pH 8.0) with 4% acetonitrile, 23 C; 60 nM Kau2 EH. The curves are based on simu-
ꢀ
1 ꢀ1
s
n.d.: not determined.
lations using Eq. (3) with the following input parameters: 0.29 and 0.0014
cat,S/KM,S and kcat,R/KM,R, 47 and 150 M for the (S,S)-1 and (R,R)-1-based K
respectively.
mM
for
a
Estimated from simulations using Eq. (3).
Estimated value using E ¼ 200 and Eq. (1).
k
m
M
values,
b

70
A. Archelas et al. / Archives of Biochemistry and Biophysics 591 (2016) 66e75
Scheme 1. Two-step reaction scheme for the Kau2 EH under the conditions used for
the stopped-flow experiment.
kcat ¼ k3
(7)
Hence k
3
and kꢀ2 could be calculated. Only lower limits for k
2
and K
S
could be assessed (Table 2).
Additional stopped-flow experiments were performed to gain
more insight into the factors responsible for the high enantiopre-
ference of the enzyme towards (S,S)-1. In these experiments
amplitude and kobs values were found to be essentially indepen-
dent of the (R,R)-1 concentration at a fixed (S,S)-1 concentration
(
(
Fig. S8), which suggests that the binding affinity of the enzyme for
S,S)-1 is significantly higher than for (R,R)-1.
We found that the Kau2 EH was substantially photodegraded on
longer time scales (1e2 s) of irradiation at 290 nm independently of
the (S,S)-1 concentration (Fig. S9). We also observed that the
bleaching amplitude decreased by reducing the cosolvent (aceto-
nitrile or ethanol) concentration. However, we did not find exper-
imental conditions where the photodegradation would be
eliminated with the substrate solubility being sufficient. We were
unable to determine correct k
3
values from single-turnover ex-
periments due to this rapid photodegrading effect and the inability
to saturate the substrate with the enzyme exhibiting such a high K
S
value [28].
3.4. Protein fluorescence quenching and enzyme mutants
The protein fluorescence quenching was found to be transient
during conversion of (S,S)-1, as shown in stopped-flow experiments
with the wild-type enzyme under conditions approaching single-
turnovers (Fig. 5). The cause of fluorescence quenching in the
wild-type Kau2 EH was investigated by constructing several
1
09
enzyme variants. Replacing the nucleophile Asp
by Asn or the
316
general base His by Ala or Gln resulted in inactive enzymes with
no change in their intrinsic protein fluorescence on a 400-ms time
scale upon mixing them with (S,S)-1. These data indicate that the
presence and activation of the nucleophilic carboxylic acid of
1
09
Asp
is required for enzyme activity and detectable quenching of
the Trp fluorescence during enzymatic reaction of (S,S)-1.
Interestingly, mixing the wild-type Kau2 EH with either CDU or
CIU, which are strong competitive disubstituted cyclohexyl urea
i
inhibitors (K of 19 or 50 nM, respectively) [16,17], did not result in
Fig. 4. Multiple-turnover experiments with different concentrations of (S,S)-1. (A) The
any detectable fluorescence quenching. As an alkyl-enzyme inter-
mediate cannot be formed with CDU or CIU, fluorescence quench-
ing is not observed. CDU is such a potent inhibitor that it prevents
(S,S)-1 from forming detectable amounts of an alkyl-enzyme in-
termediate; i.e. no fluorescence quenching was observed with
fluorescence traces were recorded in the presence of 1.2
m
M enzyme with 10, 20, 50,
ꢁ
and 80
mM of (S,S)-1 at 23 C (pH 8.0). (B, C) Dependence of kobs on the (S,S)-1 con-
ꢁ
centration; the fits to the linear Eq. (6) are shown; (B) 23 C, pH 8.0, with 4% aceto-
nitrile, the data are averaged from three independent experiments; (C) 30 C, pH 8.0,
ꢁ
with 1% acetonitrile.
1.5
m
M of wild-type Kau2 EH in the presence of 126 nM of CDU and
M of (S,S)-1 (data not shown).
Computational docking of (S,S)-1 into the active site of the Kau2
EH revealed three Trp residues as possible fluorescence quenching
16 m
k ½Sꢃ
2
kobs
¼
þ kꢀ2 þ k3
(6)
K
S
41
probes: Trp , which is part of the conserved HGW/FP motif of a/b-
1
10
109
This enabled us to determine the values for kꢀ2 þ k
intercept and k /K
that saturation of the Michaelis complex did not occur; hence a
two-step process with k /K as the second-order rate constant for
the formation of the alkyl-enzyme intermediate adequately de-
scribes the enzymatic reaction under the experimental conditions
used (Scheme 1) [14,22]. In this two-step process kcat is defined as
3
from the
hydrolase fold EHs; Trp , which is next to the nucleophilic Asp
;
317
2
S
from the slope. The linear relationship implies
and Trp being in close proximity (<3 Å) to one of the phenyl rings
of bound (S,S)-1 (Fig. 6).
for His or Ala dramatically decreased the
specific activity of the enzyme from 11.5 to 0.6 or 0.1 U mg
respectively; this indicates that mutations next to Asp can distort
the orientation of the nucleophilic carboxylic acid of this residue,
resulting in a profound decrease in enzyme activity.
110
2
S
Exchanging Trp
ꢀ
1
,
109
[
26,27]:

A. Archelas et al. / Archives of Biochemistry and Biophysics 591 (2016) 66e75
71
Table 2
Determined and estimated rate constants for hydrolysis of (S,S)-1 or (R,R)-1 catalyzed by the Kau2 EH.
ꢀ
1
ꢀ1
ꢀ1
ꢀ1
ꢀ2 (s
ꢀ1
(s )
Conditions
Substrate
k
/K
2 S
(
m
M
s
)
K
S
(
mM)
k
2
(s
)
k
)
k
3
ꢁ
a
a
2
3
0
C, 4% CH
C, 1% CH
3
CN
CN
(S,S)-TSO
0.28 ± 0.01
0.0014
1.02 ± 0.05
>1600
e
>450
e
3.8 ± 1.0
<0.01d
8.3 ± 1.8
12.0
b
c
c
0.21d
14.0
(R,R)-TSO
ꢁ
a
a
3
3
(S,S)-TSO
>1600
>1600
a
b
c
Estimated value considering that K
S
þ [S] z K
S
if K
S
> 20 ꢂ [S] (see Eqs. (5) and (6); [26]).
Assuming kcat/K
Not assessable.
M
¼ k
2 S
/K .
d
Estimated values using Eqs. (7) and (8).
Fig. 5. Stopped-flow experiments detecting protein fluorescence. The fluorescence
ꢁ
traces were recorded at 23 C (pH 8.0) with 8
in the presence of 1% acetonitrile.
mM Kau2 EH and 10, 30 and 40 mM (S,S)-1
3.5. Detection of tyrosinates
First, we intended to estimate the global pK
a
value of the 9
tyrosine residues in the Kau2 EH. UV absorption spectra were
recorded at different pH values between pH 7.0 and 10.5; sub-
tracting the spectrum recorded at pH 7.0 from spectra recorded at
higher pH values revealed a peak at 295 nm, which increased with
pH above a pH value of 9.5. This is indicatory of tyrosinate forma-
tion [29]. From these data a global pK
data not shown).
In an attempt to assess the possible transient formation of
a
value of >10.5 was estimated
(
tyrosinates during the catalytic turnover of (S,S)-1, we recorded the
absorbance at 295 nm in stopped-flow experiments and correlated
this signal with fluorescence quenching data obtained under the
same conditions (Fig. 7). A small build-up of the A295 signal was
observed, which was followed by a decrease starting at ~40 ms after
the mixing. The absorbance traces roughly mirrored the course of
the fluorescence quenching signals. The observed decrease in the
Fig. 6. View into the substrate binding pocket of the Kau2 EH homology model. The
model was created by SWISS-MODEL based on the X-ray structure of murine soluble
EH (1CQZ-A) [17]. The substrate (S,S)-1 was docked into the active site (upper section)
using AutoDock implemented in YASARA. The alkyl-enzyme intermediate (lower
section) was constructed as described in Materials and methods and is shown after
A
295 signal appears to be determined by the concentration differ-
25 ps of molecular dynamics simulation. The oxygen atoms of the epoxide and the
ence between the epoxide substrate and the diol product, which
changes during the enzymatic reaction. This change in A295
alkyl-enzyme intermediate are positioned a hydrogen-bond distance away from both
157
259
oxirane-polarizing tyrosines Tyr
enzymeeligand complex or the alkyl-enzyme intermediate are depicted as green lines
for hydrophobic interactions and red lines for interactions (the interaction
and Tyr . Favorable interactions stabilizing the
(
beyond ~40 ms) has to be explained by the depletion of the
epoxide during the catalytic turnover, which per se leads to a
decrease in absorbance at 295 nm because εepoxide >εdiol. This
sep/pep
strength is represented by shades of red, from intense red for optimal interactions to
gray for minimal interactions). Only residues involved in these interactions, and the
substrateeproduct-related effect interferes with
a possible
1
09
316
157
259
catalytically important residues Asp , His , Tyr and Tyr , and the fluorescence-
enzyme-based change in the A295 signal. At least part of the
detected small transient increase in A295 has to be explained by
mixing artifacts; we detected such artifacts with compound 1
within the first 20e25 ms after mixing and with compound 4
within the first 10e15 ms (Figs. S10, S11, and S12). Additional
control experiments were performed in the presence and absence
generating residues Trp , Trp110, and Trp
41
317
are shown in stick representation. (For
interpretation of the references to colour in this figure legend, the reader is referred to
the web version of this article.)
of the Kau2 EH with the expectation that a fundamental difference

72
A. Archelas et al. / Archives of Biochemistry and Biophysics 591 (2016) 66e75
Fig. 7. Superposition of absorbance and fluorescence traces after rapid mixing of the Kau2 EH with (S,S)-1. Superposition of absorbance (Abs) and fluorescence (Fluo) traces
ꢁ
recorded in the presence of 10, 30 and 40
mM (S,S)-1 and 8
mM Kau2 EH at 1% CH
3
CN and 23 C (pH 8.0).
in the A295 signal would be observed between experiments with
fast ((S,S)-1 or (S,S)-4) and slow substrate enantiomers ((R,R)-1 or
unimolecular isomerization of the enzymeesubstrate complex had
to be incorporated in the reaction scheme to adequately describe
the kinetic behavior of the EH from Agrobacterium radiobacter AD1
with (R)-3 as the substrate [26]. In all these EHs the hydrolysis of
the alkyl-enzyme intermediate, which is characterized by k ,
3
limited the overall enzymatic reaction.
(R,R)-4, Fig. S13) or between fast substrate ((S,S)-1 or (S,S)-4) and
product-based (2 or (R,S)-5) mixing experiments. These experi-
ments compared the A295 traces of the first ~30 ms after mixing
where significant product formation did not yet occur in order to
eliminate the substrateeproduct-related effect of
D
A
295. These
The formation of the covalent alkyl-enzyme intermediate in the
Kau2 EH was found to be reversible, as with the other characterized
controls (comparing experiments performed with and without
enzyme), however, revealed that the differences in A295 within the
first ~30 ms were small, representing not more than approximately
EHs of the
a
/
b
-hydrolase fold family, with kꢀ2 being smaller than k
3
ꢁ
by only a factor of about three (at 23 C, 4% acetonitrile). This dif-
ꢁ
0
.2 tyrosinate equivalents per enzyme molecule being transiently
ference between kꢀ2 and k
3
was found to be even smaller at 30 C in
generated during conversion of (S,S)-1 or (S,S)-4 (assuming tyro-
sinate formation as the sole source of an increase in A295).
the presence of 1% acetonitrile. When comparing the values for k
of both reaction conditions, it appears that
2
and the sum of kꢀ2 þ k
3
accumulation of the alkyl-enzyme intermediate is more pro-
ꢁ
nounced at 30 C and 1% acetonitrile (Table 2). This accumulation of
4
. Discussion
the intermediate is seen in rapid protein fluorescence quenching
experiments under conditions approaching single-turnovers
4
.1. Kinetic mechanism and enantioselectivity
(
>
Fig. 5). The accumulation of the alkyl-enzyme intermediate (k
) during conversion of (S,S)-1 is the reason for the K value
being markedly lower than the K value (45 versus >1600 M).
2
k
3
M
The reaction conditions which were previously described in a
S
m
kinetic study of the potato EH and compound 1 served as a starting
point for our experiments [25]. We then improved the solubility of
1
Concerning the overall kinetic mechanism with (S,S)-1 as the sub-
strate, the Kau2 enzyme is a standard EH with accumulation of the
alkyl-enzyme intermediate and a rate-limiting hydrolysis step
using modified experimental conditions, which resulted in
simplified data analysis. A complication was encountered in the
photodecomposition of compound 1, which was found to interfere
with the enzyme-catalyzed depletion of 1 when following the
absorbance signal of 1 at 228 nm in the stopped-flow apparatus
with its intense light source. As a result the enzymatic conversion of
[
13,25,26]. Interestingly, in the soybean EH, which also exhibits an
-hydrolase fold, the rate-limiting step of 9,10-epoxystearate
a/b
conversion was deduced to be the formation of the covalent
alkyl-enzyme intermediate and not its hydrolysis [32].
An E value of ~200 has been previously determined for the Kau2
EH-mediated kinetic resolution of 1 [5]. The factors responsible for
this high enantiopreference of the enzyme were investigated by
steady-state and pre-steady-state measurements. A good agree-
1
cannot be monitored at 228 nm using the above-mentioned set-
up of the stopped-flow apparatus. In addition, we were confronted
with photodegradation of the Kau2 EH in conjunction with a high
K value of this enzyme for (S,S)-1, preventing a reliable analysis of
S
ment between the kcat/K
found (Tables 1 and 2), which indicates that the free enzyme and
S,S)-1 were in thermodynamic equilibrium with the enzy-
M 2 S
and k /K values for (S,S)-1 has been
single-turnover experiments. Fast irreversible photobleaching
phenomena were reported for other proteins as well [30,31].
The overall kinetic mechanism of the Kau2 EH is represented by
three steps (Fig. 2); it can be simplified e under the conditions used
for the stopped-flow experiments with no saturation of the (S,S)-1-
containing Michaelis complex e to a two-step mechanism char-
acterized by k /K , kꢀ2, and k (Scheme 1). Two-step or three-step
2 S 3
reaction mechanisms have been established for the rat liver
microsomal and potato EHs depending on the substrates used
(
meesubstrate complex [22]. Hence the association and dissociation
rates which describe the kinetics of formation of the Michaelis
complex were [k
to an estimation of k
2
. Assuming a similar relationship for (R,R)-1 led
/K for this substrate (Table 2). Further, k and
M
ꢀ2 for (R,R)-1 could be assessed using the kcat and K values of
2
S
3
k
M
Table 1 and Eq. (7) in combination with Eq. (8), which defines K for
a two-step process (Scheme 1) [26,27]:
[13,11]. On the other hand, an additional step representing a

A. Archelas et al. / Archives of Biochemistry and Biophysics 591 (2016) 66e75
73
intermediate can be deduced from the absence of detectable fluo-
rescence quenching when mixing the enzyme with the competitive
inhibitors CDU or CIU, which cannot be attacked by the nucleophile
KSðk
ꢀ2
þ k3Þ
KM
¼
(8)
k2
109
Asp
due to the presence of their urea NH groups [16].
Using these experimental and assessed data, the high enantio-
A clear connection between protein fluorescence quenching and
the formation of the alkyl-enzyme intermediate has been previ-
ously established for the EH from Agrobacterium radiobacter AD1
preference of the Kau2 EH is best explained by pronounced dif-
ferences in kinetic rate constants between the two competing
substrates (S,S)-1 and (R,R)-1. It appears that (R,R)-1 is a very slow
substrate with markedly lower values for the second-order alkyl-
[
26]. Drastically reduced alkylation and hydrolysis rates were re-
3
00
ported for the potato EH variant H N [25]. This is in line with our
2 S 3
ation rate constant (k /K ) and k when compared to the corre-
findings, which led to the conclusion that the mutation of the
sponding values for (S,S)-1. This is in agreement with additional
stopped-flow data: (1) (R,R)-1 altered neither the amplitude nor
the kobs values during the conversion of (S,S)-1 when following
enzyme fluorescence, suggesting that binding of (R,R)-1 and its
conversion to the alkyl-enzyme intermediate is severely restricted
compared to (S,S)-1; and (2) no quenching of the enzyme fluores-
cence was observed in the presence of (R,R)-1 alone, indicating that
the formation of the alkyl-enzyme intermediate did not occur on a
detectable level. In conclusion, the high stereospecificity of the
109
general base in the Kau2 EH hindered the nucleophile Asp
from
being properly activated, resulting in the loss of enzyme activity
and the lack of detectable fluorescence quenching.
As our data suggest that significant formation of tyrosinates did
not occur during the catalytic turnover of (S,S)-1 or (S,S)-4 (see
below), the observed quenching could be explained by the build-up
of the oxyanion in the alkyl-enzyme intermediate in close prox-
41
110
imity to Trp and Trp (Fig. 6), since tryptophan fluorescence in
proteins is highly dependent on the polarity of the micro-
environment of this fluorophore [33]. However, we were sur-
prised to note that (S)-3, which is a substrate of the Kau2 EH [4], did
not generate any detectable protein fluorescence quenching (at (S)-
Kau2 EH appears to be based on pronounced differences in K
cat values, which are best explained by marked differences in k
and k between (S,S)-1 and (R,R)-1.
M
and
/K
k
2
S
3
The factors that determine the enantioselectivity have been
analyzed in other EHs; however, these EHs exhibited much lower E
values than the one of the Kau2 EH with rac-1 as the substrate. The
rat liver microsomal EH was shown to have a very modest E value of
for glycidyl-4-nitrobenzoate, the enantioselectivity being chiefly
controlled by kꢀ2 and k , whose values were 10 times lower for the
S)-enantiomer than for the preferentially hydrolyzed (R)-enan-
tiomer [13]. On the other hand, when analyzing the potato EH at pH
.0 the determined modest selectivity for (S,S)-1 (E value of 2.9)
was shown to be primarily based on a slower decay reaction (kꢀ2) of
the alkyl-enzyme intermediate back to the Michaelis complex [11].
Further, Rink and Janssen showed that the kinetic basis for the
enantiopreference in the EH from A. radiobacter AD1 is mainly a
3
concentrations of up to 640
enzyme). This could be a special case explained by a change in the
rate-limiting step from k to k , which would result in no accu-
mM in the presence of 3.6 mM
3
2
mulation of the alkyl-enzyme intermediate. Alternatively, the
smaller structure of (S)-3 (compared to (S,S)-1) would be inade-
quate to generate detectable fluorescence quenching upon alkyl-
ation of the enzyme. Concerning the much bulkier epoxide (S,S)-1,
we propose that the observed changes in Trp fluorescence are
2
3
(
8
4
1
associated with changes in the micro-environment of mainly Trp
and Trp
110
caused by the transient build-up of the alkyl-enzyme
intermediate generated during the enzymatic reaction. This inter-
mediate may have its phenyl groups displaced (compared to the
Michaelis complex), leading to a modification in the interaction
2 S
large difference in alkylation rates (k /K ) between (S)- and (R)-3,
41
110
between Trp eTrp and the bound substrate as it is converted to
resulting in an E value of 16 [26]. In a rather similar manner, the
very high enantiopreference of the Kau2 EH towards (S,S)-1, which
resulted in hardly detectable enzyme activity towards (R,R)-1, ap-
pears to be a consequence of extensive differences in k /K and k
2 S 3
values.
the transient alkyl-enzyme intermediate. See Fig. 6 for a simulation
of the hydrophobic and
sep/pep interactions between the
enzyme and (S,S)-1 before and after the attack of the nucleophilic
109
Asp . Molecular dynamics simulations of the alkyl-enzyme in-
termediate suggest a rotation of the phenyl groups after the
opening of the epoxide ring, resulting in modified interaction
41
110
4
.2. Protein fluorescence quenching and enzyme mutants
patterns around Trp and Trp (Fig. 6).
The origin of the observed protein fluorescence quenching in
4.3. The role of the active site Tyr157 and Tyr259
the Kau2 EH during conversion of (S,S)-1 appears to be linked to the
transient formation of the alkyl-enzyme intermediate. An analysis
The full role of the active site tyrosines during the EH-catalyzed
conversion of the epoxide substrate is still a matter of some debate.
One uncontroversial function is their involvement in the substrate
binding event. As shown in several papers which analyzed active
site Tyr/Phe mutants and EHeinhibitor structures [14,16,34], at
least one of the active site tyrosines in EHs appears to facilitate the
proper orientation of the bound epoxide by forming a hydrogen
bond between the oxirane oxygen and the Tyr hydroxyl, thereby
preparing it for the attack of the nucleophilic Asp.
A possible function of the active site tyrosines as proton donors
has been proposed, but so far it was experimentally investigated
only for the potato EH [12,35]. If these tyrosine residues were
involved in a proton transfer to the transiently generated alkoxide
109
316
110
of active site mutants (Asp
/ Asn, His
/ Ala/Gln, Trp
/
1
09
His/Ala) supports the concept that the nucleophile Asp has to be
present, properly oriented, and properly activated by the general
base His3 for a successful attack on the epoxide ring to occur.
Mutations at position 110 can have a strong impact on the enzyme
activity, presumably by distorting the correct position of the
nucleophile Asp1 . It is worthwhile to note that the deleterious Ala
mutation at position 110 can be overcome by introducing a second
mutation at the position 113 (Gly, Arg, or Glu instead of Phe),
resulting in active enzyme variants [17]. A correlation between
protein fluorescence quenching and the formation of the alkyl-
enzyme intermediate was also deduced from the fact that no
quenching was detected within the dead-time of the instrument
when mixing the Kau2 EH with (S,S)-1, indicating that substrate
binding and formation of the Michaelis complex e taking place on a
very fast time scale within the dead-time e did not cause
quenching. Furthermore, a causal connection between quenching
of the protein fluorescence and the formation of the alkyl-enzyme
16
09
(alkyl-enzyme intermediate), a perturbed more acidic pK
a
of the
Tyr-phenol group (compared to a pK value of 10.1 in the free amino
a
acid; [22]) would be expected in order to facilitate the proton
transfer. However, our titration data do not support the presence of
a
tyrosine residues with unusually low pK values in the free enzyme;
similar results were reported for the potato EH [35].

74
A. Archelas et al. / Archives of Biochemistry and Biophysics 591 (2016) 66e75
Still the possibility of a transient formation of tyrosinates during
the catalytic turnover of (S,S)-1 could not be excluded; a decrease in
the pK of the active site Tyr residues might be brought about by
deprotonated during the catalytic cycle [42]. For the time being, it is
unclear to what extent these theoretical studies and experimental
data represent general or substrateeenzyme-specific results of EH-
catalyzed reactions.
a
transient catalytic events during conversion of the substrate. In
order to examine this possibility the A295 was recorded in stopped-
flow experiments and correlated with fluorescence quenching data.
The stopped-flow traces revealed a small transient increase in A295
within the first 30 ms. Taking into account all the associated control
experiments, we came to the conclusion that a maximum of ~0.2
tyrosinate equivalents per enzyme molecule were transiently
generated during conversion of the substrate, assuming that tyro-
sinate formation is the sole source of an increase in A295. It is
important to note at this point that a transient increase in A295
cannot be unequivocally correlated to an increase in ionization
level of tyrosines; hydrogen bonding of tryptophans or perturba-
tions in the environment of these residues such as their transfer to a
more hydrophobic medium can also result in an increase in
absorbance at this wavelength [29,36e38]. As the Kau2 EH contains
tryptophans and tyrosines, a proper interpretation of the small
transient increase in absorbance signal upon substrate conversion
has to include the possibility that it may represent not only changes
in ionization level of tyrosines but also perturbations in the envi-
ronment of tryptophans. Due to this ambiguity a clear-cut conclu-
sion as to the transient formation of tyrosinates cannot be drawn;
nevertheless, the data rather suggest that tyrosinate formation is
not a major event in the Kau2 EH catalytic mechanism of (S,S)-1 and
5. Conclusions
Monitoring the protein fluorescence quenching in the Kau2 EH
enabled us to follow the transient formation of the alkyl-enzyme
intermediate during conversion of 1. We have shown that the
Kau2 enzyme in the presence of substrate 1 reacts as a standard EH
with accumulation of the alkyl-enzyme intermediate followed by a
rate-limiting hydrolysis step. The kinetic data suggest that the high
enantiopreference of the Kau2 EH towards 1 (E value of ~200) is
based on marked differences in kcat and KM values between the
competing substrates (S,S)-1 and (R,R)-1. Further, the extensive
enantioselectivity of the EH is best explained by profound differ-
ences in the second-order alkylation rate constant (k /K ) and k for
2
S
3
the two enantiomers of 1. Our data do not support the presence of
significant tyrosinate equivalents during the enzymatic hydrolysis
of epoxides 1 and 4. Thus it appears that the active site
1
57
259
Tyr eTyr
pair does not serve as a proton donor to the tran-
siently formed alkoxide enzyme intermediate during the alkylation
half-reaction; the active site tyrosines rather serve as electrophilic
catalysts, polarizing the oxygen of the bound epoxide through (a)
hydrogen bond(s), thereby facilitating the opening of the epoxide
ring.
(S,S)-4 hydrolysis. We are in favor of a mechanism where complete
proton transfer from the Tyr to the enzyme-bound alkoxide does
not occur and transient tyrosinates are not being formed to a sig-
nificant extent. The above-mentioned experiments suggest that the
two conserved active site tyrosines Tyr and Tyr in the Kau2 EH
serve as electrophilic catalysts [22], forming hydrogen bonds with
the negatively charged alkyl-enzyme intermediate, however,
without completely donating their protons to the intermediate.
This is basically in line with the conclusions drawn by Elfstr o€ m and
Acknowledgments
157
259
We are grateful to Jalila Simaan (Aix-Marseille University) for
her help with the stopped-flow apparatus. In addition, we thank
Marion Jean and Nicolas Vanthuyne (Aix-Marseille University) for
their preparative chiral HPLC experiments. In addition, we are
indebted to Elise Courvoisier-Dezord (Platform AVB, Aix-Marseille
University) for her help in the bioreactor manipulations. This
work was partially supported by the Academy of Sciences of the
Czech Republic (grant F-13-14-05) and the CNRS (France). We also
acknowledge support from the Institute of Microbiology of the
Academy of Sciences of the Czech Republic (RVO61388971).
Widersten for the potato EH (also containing tryptophans and ty-
rosines) [35], although we find their interpretation of the recorded
decrease in A295 during the conversion of 1 problematic. They failed
to mention that
DA295 is strongly dependent on the concentration
ratio of epoxide: diol, which is not constant during the reaction
ꢀ
1
ꢀ1
(
7
D
ε ¼ εepoxide e εdiol ¼ 546 or 571 M cm for compound 1 at pH
.2 or 8.5, respectively). The above-mentioned mechanism implies
Appendix A. Supplementary data
that a water molecule e instead of the active site tyrosines e acts as
the proton donor for the anionic alkyl-enzyme intermediate prior
to its hydrolysis (as suggested in [35]). As an alternative, proton-
ation of the alkyl-enzyme intermediate would be managed by one
of the active site tyrosines, the immediate re-protonation of Tyr
would occur through the action of a near-by water molecule [39].
Still another mechanism of Tyr re-protonation was proposed by
Schiøtt and Bruice [40]; in this scenario a proton shuttle would
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.abb.2015.12.008.
References
[1] A. Archelas, R. Furstoss, Synthetic applications of epoxide hydrolases, Curr.
Opin. Chem. Biol. 5 (2001) 112e119.
[2] A. Steinreiber, K. Faber, Microbial epoxide hydrolases for preparative bio-
transformations, Curr. Opin. Biotechnol. 12 (2001) 552e558.
316
transfer the proton of the active site base (His ) to the active site
Tyr and thereby prepare it for the next catalytic cycle.
[3] M. Kotik, A. Archelas, R. Wohlgemuth, Epoxide hydrolases and their applica-
tion in organic synthesis, Curr. Org. Chem. 16 (2012) 451e482.
The likely absence of significant tyrosinate equivalents during
the EH-catalyzed conversion of compounds 1 and 4 appears not to
be in contradiction with molecular dynamics simulations per-
formed by Schiøtt and Bruice using the murine soluble EH as a
model [40]. However, Hopmann and Himo concluded in their study
of a small quantum-chemical model of human soluble EH using a
density functional theory method that the formed alkyl-
intermediate was protonated through a proton transfer from one
of the two conserved active site Tyr during the conversion of 4 [41].
On the other hand, in a quantum mechanics/molecular mechanics
modeling study Lonsdale and co-workers suggested that their data
rather support a mechanism in which the active site tyrosines in
murine soluble EH act as hydrogen bond donors and are not
ꢁ
ꢁ ꢀ
[
4] M. Kotik, V. Stepanek, M. Grulich, P. Kyslík, A. Archelas, Access to enantiopure
aromatic epoxides and diols using epoxide hydrolases derived from total
biofilter DNA, J. Mol. Catal. B Enzym. 65 (2010) 41e48.
[5] W. Zhao, M. Kotik, G. Iacazio, A. Archelas, Enantioselective bio-hydrolysis of
various racemic and meso aromatic epoxides using the recombinant epoxide
hydrolase Kau2, Adv. Synth. Catal. 357 (2015) 1895e1908.
[6] Nugent W.A. (2001), Catalysts for asymmetric addition of organozinc reagents
to aldehydes and method for preparation. US Patent 6187918 B1.
[7] S. Florio, V. Aggarwal, A. Salomone, Lithiation and reactions of stilbene oxides:
synthetic utility, Org. Lett. 6 (2004) 4191e4194.
[8] K. Bica, P. Gaertner, Application of meso-hydrobenzoin-derived chiral auxil-
iaries for the stereoselective synthesis of highly substituted pyrrolidines by
1,3-dipolar cycloaddition of azomethine ylides, Tetrahedron Asymmetry 21
(2010) 641e646.
[9] T. Arai, K. Sakagami, Asymmetric benzoylation of meso-hydrobenzoin using a
reusable tripodal imidazolineepyridineeCu catalyst, Eur. J. Org. Chem. 2012

A. Archelas et al. / Archives of Biochemistry and Biophysics 591 (2016) 66e75
75
(
2012) 1097e1100.
styrene oxide by epoxide hydrolase from Agrobacterium radiobacter AD1,
Biochemistry 37 (1998) 18119e18127.
[27] A. Cornish-Bowden, Fundamentals of Enzyme Kinetics, third ed., Portland
Press Ltd., London, 2004, p. 48.
[
[
[
[
10] M. Torres-Werl eꢀ , A. Nano, A. Maisse-François, S. Bellemin-Laponnaz, Asym-
metric benzoylation and Henry reaction using reusable polytopic bis(oxazo-
line) ligands and copper(II), New J. Chem. 38 (2014) 4748e4753.
11] D. Lindberg, A. Gogoll, M. Widersten, Substrate-dependent hysteretic
behavior in StEH1-catalyzed hydrolysis of styrene oxide derivatives, FEBS J.
[28] K.A. Johnson, Enzyme Kinetics: Transient Phase. in: ELS, John Wiley & Sons
Ltd, Chichester, 2005. http://www.els.net.
[29] S.N. Timasheff, Some physical probes of enzyme structure in solution, in:
P.D. Boyer (Ed.), The Enzymes, Volume II, third ed., Academic Press, New York
and London, 1970, pp. 371e442.
275 (2008) 6309e6320.
12] C. Morisseau, B.D. Hammock, Epoxide hydrolyses: mechanisms, inhibitor
designs, and biological roles, Annu. Rev. Pharmacol. Toxicol. 45 (2005)
311e333.
[30] H.-H. Kim, Y.M. Lee, J.-K. Suh, N.W. Song, Photodegradation mechanism and
13] H.-F. Tzeng, L.T. Laughlin, S. Lin, R.N. Armstrong, The catalytic mechanism of
microsomal epoxide hydrolase involves reversible formation and rate-
limiting hydrolysis of the alkyleenzyme intermediate, J. Am. Chem. Soc. 118
reaction kinetics of recombinant human interferon-a2a, Photochem. Photo-
biol. Sci. 6 (2007) 171e180.
[31] F. Rivardo, T.G.H. Leach, C.S. Chan, T.M.L. Winstone, C.L. Ladner, K.J. Sarfo,
R.J. Turner, Unique photobleaching phenomena of the twin-arginine trans-
locase respiratory enzyme chaperone DmsD, Open Biochem. J. 8 (2014) 1e10.
[32] E. Bl eꢀ e, S. Summerer, M. Flenet, H. Rogniaux, A. Van Dorsselaer, F. Schuber,
Soybean epoxide hydrolase, J. Biol. Chem. 280 (2005) 6479e6487.
[33] A.B.T. Ghisaidoobe, S.J. Chung, Intrinsic tryptophan fluorescence in the
detection and analysis of proteins: a focus on F o€ rster resonance energy
(
1996) 9436e9437.
[
[
[
14] R. Rink, J. Kingma, J.H. Lutje Spelberg, D.B. Janssen, Tyrosine residues serve as
proton donor in the catalytic mechanism of epoxide hydrolase from Agro-
bacterium radiobacter, Biochemistry 39 (2000) 5600e5613.
15] B.A. Amrein, A. Bauer, F. Duarte, Naworyta A. Carlsson Å.J., S.L. Mowbray,
M. Widersten, S.C.L. Kamerlin, Expanding the catalytic triad in epoxide hy-
drolases and related enzymes, ACS Catal. 5 (2015) 5702e5713.
16] M.A. Argiriadi, C. Morisseau, M.H. Goodrow, D.L. Dowdy, B.D. Hammock,
D.W. Christianson, Binding of alkylurea inhibitors to epoxide hydrolase im-
plicates active site tyrosines in substrate activation, J. Biol. Chem. 275 (2000)
transfer techniques, Int. J. Mol. Sci. 15 (2014) 22518e22538.
[34] T. Yamada, C. Morisseau, J.E. Maxwell, M.A. Argiriadi, D.W. Christianson,
B.D. Hammock, Biochemical evidence for the involvement of tyrosine in
epoxide activation during the catalytic cycle of epoxide hydrolase, J. Biol.
Chem. 275 (2000) 23082e23088.
15265e15270.
[
17] M. Kotik, W. Zhao, G. Iacazio, A. Archelas, Directed evolution of metagenome-
derived epoxide hydrolase for improved enantioselectivity and enantio-
convergence, J. Mol. Catal. B Enzym. 91 (2013) 44e51.
[35] L.T. Elfstr o€ m, M. Widersten, Implications for an ionized alkyl-enzyme inter-
mediate during StEH1-catalyzed trans-stilbene oxide hydrolysis, Biochemistry
45 (2006) 205e212.
[
18] S. Pedragosa-Moreau, A. Archelas, R. Furstoss, Microbiological transformations
[36] D. Freifelder, Physical Biochemistry, Applications to Biochemistry and Mo-
lecular Biology, second ed., W.H. Freeman and Company, San Francisco, USA,
1982.
32: use of epoxide hydrolase mediated biohydrolysis as a way to enantiopure
epoxides and vicinal diols: application to substituted styrene oxide de-
rivatives, Tetrahedron 52 (1996) 4593e4606.
[37] E.H. Strickland, C. Billups, E. Kay, Effects of hydrogen bonding and solvents
[
[
19] A.N. Glazer, The chemical modification of proteins by group-specific and site-
specific reagents, in: H. Neurath, R.L. Hill (Eds.), The Proteins, Volume 2, third
ed., Academic Press, New York, 1976, 1e103.
20] L.S. Hasegawa, B.D. Hammock, Spectrophotometric assay for mammalian
cytosolic epoxide hydrolase using trans-stilbene oxide as the substrate, Bio-
chem. Pharmacol. 31 (1982) 1979e1984.
upon the tryptophanyl 1
a
L absorption band. Studies using 2,3-dimethylindole,
Biochemistry 11 (1972) 3657e3662.
[38] K.J. Rothschild, P. Roepe, P.L. Ahl, T.N. Earnest, R.A. Bogomolni, S.K. Das Gupta,
C.M. Mulliken, J. Herzfeld, Evidence for a tyrosine protonation change during
the primary phototransition of bacteriorhodopsin at low temperature, Proc.
Natl. Acad. Sci. U. S. A. 83 (1986) 347e351.
[
[
[
[
[
[
21] A.J.J. Straathof, J.A. Jongejan, The enantiomeric ratio: origin, determination
and prediction, Enzyme Microb. Technol. 21 (1997) 559e571.
22] A. Fersht, Enzyme Structure and Mechanism, second ed., W.H. Freeman and
Company, New York, USA, 1985.
23] M. Dixon, E.C. Webb, Enzymes, second ed., Longmans, Green and Co. Ltd.,
London, UK, 1964.
24] T. Pocklington, J. Jeffery, Competition of two substrates for a single enzyme,
Biochem. J. 112 (1969) 331e334.
25] L.T. Elfstr o€ m, M. Widersten, Catalysis of potato epoxide hydrolase, StEH1,
[39] M. Widersten, A. Gurell, D. Lindberg, Structureefunction relationships of
epoxide hydrolases and their potential use in biocatalysis, Biochim. Biophys.
Acta Gen. Subj. 1800 (2010) 316e326.
[40] B. Schiøtt, T.C. Bruice, Reaction mechanism of soluble epoxide hydrolase: in-
sights from molecular dynamics simulations, J. Am. Chem. Soc. 124 (2002)
14558e14570.
[41] K.H. Hopmann, F. Himo, Theoretical study of the full reaction mechanism of
human soluble epoxide hydrolase, Chem. Eur. J. 12 (2006) 6898e6909.
[42] R. Lonsdale, S. Hoyle, D.T. Grey, L. Ridder, A.J. Mulholland, Determinants of
reactivity and selectivity in soluble epoxide hydrolase from quantum me-
chanics/molecular mechanics modeling, Biochemistry 51 (2012) 1774e1786.
Biochem. J. 390 (2005) 633e640.
26] R. Rink, D.B. Janssen, Kinetic mechanism of the enantioselective conversion of