Exploration of the biocatalytic potential of a halohydrin dehalogenase using chromogenic substrates

Article abstract of DOI:10.1016/S0957-4166(02)00222-7

Halohydrin dehalogenases are bacterial enzymes that catalyse the reversible formation of epoxides from vicinal halohydrins. A spectrophotometric assay for halohydrin dehalogenases based on the absorption difference between the halohydrin para-nitro-2-bromo-1-phenylethanol and the epoxide para-nitrostyrene oxide was developed. The enantioselectivity of ring-closure reactions catalysed by three different halohydrin dehalogenases could be estimated from the shape of progress curves. Evaluation of ring-opening reactions catalysed by halohydrin dehalogenase from Agrobacterium radiobacter AD1 established that, in addition to Cl^- and Br^-, nucleophiles such as N₃^-, CN^- and NO₂^- are also accepted for the ring opening of para-nitrostyrene oxide. The ring-opening reactions with these nucleophiles resulted in highly enantioselective kinetic resolutions, which expands the scope of synthetically valuable conversions catalysed by a halohydrin dehalogenase.

Full text of DOI:10.1016/S0957-4166(02)00222-7

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 13 (2002) 1083–1089
Exploration of the biocatalytic potential of a halohydrin
dehalogenase using chromogenic substrates
Jeffrey H. Lutje Spelberg,a,b Lixia Tang,a Marc van Gelder,b Richard M. Kelloggb and
Dick B. Janssena,*
aDepartment of Biochemistry, Groningen Biomolecular Sciences & Biotechnology Institute, University of Groningen, Nijenborgh 4,
9747 AG, Groningen, The Netherlands
bDepartment of Organic & Molecular Inorganic Chemistry, University of Groningen, Nijenborgh 4, 9747 AG,
Groningen, The Netherlands
Received 24 April 2002; accepted 4 May 2002
Abstract—Halohydrin dehalogenases are bacterial enzymes that catalyse the reversible formation of epoxides from vicinal
halohydrins. A spectrophotometric assay for halohydrin dehalogenases based on the absorption difference between the halohydrin
para-nitro-2-bromo-1-phenylethanol and the epoxide para-nitrostyrene oxide was developed. The enantioselectivity of ring-closure
reactions catalysed by three different halohydrin dehalogenases could be estimated from the shape of progress curves. Evaluation
of ring-opening reactions catalysed by halohydrin dehalogenase from Agrobacterium radiobacter AD1 established that, in addition
to Cl− and Br−, nucleophiles such as N3−, CN− and NO2 are also accepted for the ring opening of para-nitrostyrene oxide. The
−
ring-opening reactions with these nucleophiles resulted in highly enantioselective kinetic resolutions, which expands the scope of
synthetically valuable conversions catalysed by a halohydrin dehalogenase. © 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction
available in abundant quantities.4,5 Sequence similari-
ties and substrate specificities suggest that halohydrin
dehalogenases obtained from Agrobacterium radiobac-
ter AD1, Arthrobacter sp. AD2 and Mycobacterium sp.
GP1 belong to three different groups.4 A general mech-
anism for enzymatic dehalogenation of halohydrins was
proposed based on significant sequence similarity to
short chain dehydrogenases/reductases and confirmed
by site-directed mutagenesis of the conserved catalytic
residues. The mechanism is fundamentally different
from that of hydrolytic dehalogenases since it does not
include a covalent enzyme–substrate intermediate.6
Chiral non-racemic epoxides are important building
blocks in the synthesis of various pharmaceutical prod-
ucts. Halohydrins can be considered as direct precur-
sors of epoxides, since base-catalysed ring closure of an
enantiomerically pure halohydrin generally yields an
enantiomerically pure epoxide. Halohydrin dehaloge-
nases catalyse the ring closure of a vicinal halohydrin
and the reverse reaction, the ring opening of an epoxide
by a halide (Scheme 1).
An enzyme-catalysed (enantioselective) ring-closure or
ring-opening reaction can be monitored by extracting
the substrate or product from the aqueous phase using
an organic solvent and analysing the yield and enan-
tiomeric purity by (chiral) GC or HPLC. This method
is tedious and laborious and cannot be applied easily
for an examination of numerous reactions or reaction
conditions. Colorimetric assays for measuring dehalo-
genase activity are based on the detection of chloride
ions or the use of pH indicators.7,8 The method based
on measuring chloride concentrations requires continu-
ous sampling from the reaction mixture. Use of a pH
Scheme 1.
This class of enzymes can be used to prepare various
optically active epoxides and halohydrins.1–3 Several
genes encoding halohydrin dehalogenases have been
cloned and overexpressed making these enzymes
* Corresponding author. Tel.: 31-50-363-4209; fax: 31-50-363-4165;
e-mail: d.b.janssen@chem.rug.nl
0957-4166/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(02)00222-7

1084
J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 13 (2002) 1083–1089
indicator such as phenol red in a weakly buffered
medium allows continuous monitoring of the dechlori-
nation reaction, but the sensitivity is low9 and further-
more, a weak buffer does not allow the addition of
compounds that influence the pH of the medium such
as strong nucleophiles.
phenone. Racemic para-nitrostyrene oxide 2 was pre-
pared by ring closure of the halohydrin 1. The extinc-
tion coefficient of the halohydrin 1 was found to be
3050 M−1 cm−1 at 310 nm. The rate of ring-closure
without the presence of enzyme was 2.3×10−5 s−1 at
22°C (Tris buffer pH 7.5). Under similar conditions, no
degradation of para-nitrostyrene oxide 2 was observed.
Thus, a spectrophotometric assay would be very valu-
able. However, chromogenic substrates for measuring
halohydrin dehalogenase activity have not yet been
described in the literature. The goal of the work
described herein is the development of such an assay
for continuous monitoring of reactions catalysed by
halohydrin dehalogenases. The substrates should be
suitable for studying ring-closure and ring-opening
reactions. Evaluation of the shape of the progress curve
of a reaction with a racemic substrate can give informa-
tion about the enantioselectivity of the conversion since
a sequential conversion of the two enantiomers will
result in a biphasic curve. The assay must also be
applicable for studying the reverse reaction, the ring
opening of an epoxide by a nucleophile. A halohydrin
dehalogenase from Corynebacterium sp. N-1074 cataly-
ses the ring opening of epoxides by cyanide, resulting in
the formation of b-hydroxynitriles.10 More recently, we
have shown that azide is accepted as a nucleophile by
the halohydrin dehalogenase from A. radiobacter
AD1.11 Apart from these examples, ring-opening reac-
tions with alternative nucleophiles catalysed by a halo-
hydrin dehalogenase have not been described.
Halohydrin dehalogenases from A. radiobacter AD1,
Arthrobacter sp. AD2 and Mycobacterium sp. GP1
catalysed the ring closure of racemic para-nitro-2-
bromo-1-phenylethanol 1, resulting in the formation of
para-nitrostyrene oxide 2 (Scheme 2). The difference in
extinction coefficient of the halohydrin and epoxide
enabled the on-line monitoring of the reaction by fol-
lowing the increase in absorbance at 310 nm (Fig. 1).
2.2. Enantioselectivity of the ring-closure reaction
The progress curve of the ring-closure reaction of
racemic para-nitro-2-bromo-1-phenylethanol (at 250
mM concentration) catalysed by the halohydrin dehalo-
genase from A. radiobacter AD1 showed a biphasic
shape (Fig. 2). An explanation for this can be a differ-
ent rate of conversion of the two enantiomers. The
biphasic shape was not observed in reactions catalysed
by the halohydrin dehalogenases obtained from
Arthrobacter sp. AD2 and Mycobacterium sp. GP1.
To determine the enantioselectivity of the ring-closure
reactions, kinetic resolution experiments were per-
formed with racemic halohydrin 1 at an initial substrate
concentration of 3 mM. The conversion was monitored
by taking samples from the reaction mixture at differ-
Herein, we describe a convenient on-line spectrophoto-
metric assay for studying halohydrin dehalogenase-
catalysed ring-closure and ring-opening reactions. The
method is based on the difference in the absorbance
spectrum of the epoxide and the corresponding ring
opened product.
2. Results
2.1. Development and validation of the assay
Scheme 2.
Racemic para-nitro-2-bromo-1-phenylethanol 1 was
considered to be a suitable chromogenic substrate based
on previous work that demonstrated that halohydrin
dehalogenase from A. radiobacter AD1 catalysed the
enantioselective ring closure of aromatic halohydrins
such as 2-chloro-1-phenylethanol (enantioselectivity
factor, E=73).3 A spectrophotometric continuous assay
for measuring epoxide hydrolase activity has been
described.12 This assay is based on the difference in
extinction coefficients of para-nitrostyrene oxide (m310
4289 M−1 cm−1) and para-nitrophenylethanediol (m310
3304 M−1 cm−1), the product of epoxide hydrolysis. The
substrate is broadly applicable since it can be used to
investigate various aspects of epoxide hydrolase-
catalysed conversions such as kinetic properties,13
enzyme
stability
in
various
media,14
and
enantioselectivity.15
Figure 1. Absorption spectrum changes during the conversion
of (R)-para-nitro-2-bromo-1-phenylethanol 1 to (R)-para-
nitrostyrene oxide 2 catalysed by the halohydrin dehalogenase
from A. radiobacter AD1. Concentration of (R)-1, 285 mM.
Racemic para-nitro-2-bromo-1-phenylethanol 1 was
synthesised by reduction of v-bromo-para-nitroaceto-

J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 13 (2002) 1083–1089
1085
AD2 and Mycobacterium sp. GP1 converted the halo-
hydrin 1 with a low enantioselectivity (E <3). The
enantiopreference of these two enzymes is opposite to
that of A. radiobacter AD1 as the (S)-enantiomer was
converted at a slightly higher rate than the (R)-
enantiomer.
The exact E-value of the conversion could not be
determined from the spectrophotometric progress
curve, but the bend in the curve gives an approximate
indication of the enantioselectivity of the conversion
(Fig. 2). The monophasic progress curves for the con-
version of racemic 1 with the enzymes obtained from
Arthrobacter sp. AD2 and Mycobacterium sp. GP1
indicated a low enantioselectivity, which was confirmed
by monitoring the enantiomeric excess values of the
substrate and product. The structurally similar 2-
chloro-1-phenylethanol was converted by both enzymes
with comparably low enantioselectivity (E <10).4
Because our main interest is to use halohydrin dehalo-
genases to synthesise enantiomerically pure building
blocks, further experiments focused on the enzyme
from A. radiobacter AD1.
Figure 2. Progress curves of the conversion of racemic para-
nitro-2-bromo-1-phenylethanol 1, catalysed by the halohydrin
dehalogenase from A. radiobacter AD1 (a), Arthrobacter sp.
AD2 (b) and Mycobacterium sp. GP1 (c). Concentration 1,
255 mM.
ent times and determining the enantiomeric excess of
both the halohydrin 1 and the formed epoxide 2 by
chiral HPLC. The enantioselectivity factor (E-value)
was estimated from the enantiomeric excess values of
the substrate and product.16 An E-value of 92 ( 4) was
obtained for the kinetic resolution catalysed by the
halohydrin dehalogenase from A. radiobacter AD1
(Fig. 3).
The kinetic basis of the high enantioselectivity can be
established by studying the steady-state kinetic parame-
ters of the separate enantiomers of para-nitro-2-bromo-
1-phenylethanol 1. The E-value is defined by Eq. (1),
where kcat and Km represent the Michaelis–Menten
parameters for both enantiomers. The enantiomers of
the halohydrin 1 and epoxide 2 were obtained in enan-
tiomerically pure form (e.e. >99%) by separating them
using a HPLC equipped with a chiral column (Chiral-
pak AS, Daicel). Enantiomerically pure (R)- and (S)-
para-nitro-2-bromo-1-phenylethanol 1 (250 mM) were
converted by the halohydrin dehalogenase from A.
radiobacter AD1 with an initial activity of 35 and 2.6
mmol min−1 mg−1, respectively. The Km for the (R)-
enantiomer was too low to be determined by initial rate
measurements (Km <50 mM) allowing only the calcula-
tion of a kcat value of 16.3 s−1.
The observed E-value is lower than the intrinsic E-
value, since the non-enzyme-catalysed ring-closure reac-
tion (kc=2.3×10−5 s−1) affected the enantiomeric excess
of the formed product. Correction of the enantiomeric
excess of the product epoxide for the chemical side
reaction yielded an E-value of 124 ( 10). The (R)-enan-
tiomer was converted preferentially, followed by a
much slower conversion of the (S)-enantiomer, causing
the biphasic shape of the progress curve observed when
the increase in absorbance was monitored at 310 nm.
The halohydrin dehalogenases from Arthrobacter sp.
For the (S)-enantiomer, on the other hand, no satura-
tion kinetics were observed (Km >250 mM), allowing
only the determination of a kcat/Km value of 4740 s−1
M−1. The above described kinetic parameters and the
E-value of 124 allowed calculation of a Km value of 28
mM for the (R)-enantiomer using Eq. (1).
2.3. Evaluation of nucleophiles for the ring-opening
reaction
The enzyme-catalysed ring opening of para-nitrostyrene
oxide was investigated by allowing different nucleo-
philic compounds to react with para-nitrostyrene oxide.
In Fig. 4, the progress curves are shown of the reaction
between (R)-para-nitrostyrene oxide 2 and Br− at vari-
ous concentrations. An equilibrium constant of 480
mM was calculated from the ring opening of (R)-2 at a
Br− concentration of 100 mM.
Figure 3. Kinetic resolution of racemic para-nitro-2-bromo-1-
phenylethanol 1, catalysed by the halohydrin dehalogenase
from A. radiobacter AD1. Concentration 1, 3 mM. Symbols:
ꢀ, (S)-enantiomer; ꢁ, (R)-enantiomer; ꢂ, non-enzyme-
catalysed reaction.
The halide ions I−, Cl−, Br− and F−, the ions OCN−,
SCN−, N3 , CN−, CH3COO− and NO3 , and the non-
−
−

1086
J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 13 (2002) 1083–1089
The enzyme-catalysed nucleophilic ring opening of
para-nitrostyrene oxide resulted in products with lower
extinction coefficients (310 nm), which allowed the
monitoring of the conversion by following the decrease
in absorption. The initial rate of enzymatic ring open-
ing of epoxide (R)-2 with Br− and Cl− was more than
100-fold higher than that of (S)-2 (Table 1). A decrease
in absorption was also observed when the (R)-enan-
tiomer was subjected to the anions N3 , CN− and NO2 ,
whereas no absorbance change was observed with the
(S)-enantiomer. No decrease in absorbance with either
−
−
enantiomer was observed with F−, OCN−, SCN−, NO3 ,
−
CH3COO−, ethanol and isopropylamine. To validate
the absence of a reaction between the epoxide and these
latter compounds, the concentration of remaining epox-
ide was monitored in time using (chiral) HPLC. In all
cases where no change in absorbance (DA <0.01 min−1
mg−1) occurred, no decrease in epoxide concentration
due to an enzymatic conversion was observed.
Figure 4. Progress curves of the ring opening of (R)-para-
nitrostyrene oxide 2 by Br−, catalysed by halohydrin dehalo-
genase from A. radiobacter AD1. Incubations contained: 0.25
mM Br− (a), 25 mM Br− (b), 50 mM Br− (c), 100 mM Br− (d);
225 mM (R)-2.
In order to investigate if nucleophiles that do not react
with the epoxide can bind in the active site, the inhibit-
ing effect of these compounds on the initial rate of the
conversion of 250 mM halohydrin (R)-2 was determined
(Table 2). Nucleophiles that reacted with the epoxide
were in general good inhibitors of the ring-closure
reaction. Inhibition by halide ions decreased in the
order of I>Br>Cl>F, which is also the order of decreas-
ing nucleophilicity. The non-ionic nucleophiles ethanol
and isopropylamine did not inhibit the ring-closure
reaction. OCN− was a good inhibitor and therefore
binds in the active site, although no enzyme-catalysed
reaction with the epoxide took place. On the other
hand, CN−, which reacted selectively with one enan-
tiomer of the epoxide, did not inhibit the ring-closure
reaction when added at a concentration of 10 mM.
ionic nucleophiles isopropylamine and ethanol were
also tested. The conversion rates are expressed in DA
min−1 mg−1 since, apart from the reaction with Br−,
nucleophilic ring opening leads to products with
unknown extinction coefficients. The enantiopreference
of the ring opening with a certain nucleophile was
determined by allowing the (R)- and (S)-enantiomers of
the epoxide to react separately. The effect of the non-
enzyme-catalysed ring opening was minimised by mea-
suring the initial rate within the first 2 min. During this
period, no substantial decrease or increase of
absorbance was observed in a control reaction in which
the enzyme was omitted. An exception is I−, which gave
an apparent chemical reaction that caused an increase
in absorbance.
Table 1. Initial rates of ring opening of (R)- and (S)-para-nitrostyrene oxide 2 by various nucleophiles, catalysed by the
halohydrin dehalogenase from A. radiobacter AD1
Nucleophile
Concentration of nucleophile (mM)
(R)-2 (DA min−1 mg−1
)
(S)-2 (DA min−1 mg−1
)
Br−
Cl−
N3
25
25
2.5
2.5
2.5
2.33
2.83
0.17
0.25
0.45
0.02
0.01
B0.01
B0.01
0.01
−
−
NO2
CN−
Table 2. Inhibition of the ring closure of (R)-para-nitro-2-bromo-1-phenylethanol 1 by various nucleophiles, catalysed by the
halohydrin dehalogenase from A. radiobacter AD1
Nucleophile
Inhibition constant I50 (mM)a
Nucleophile
Inhibition constant I50 (mM)a
I−
2.3
4.3
13
29
2.7
OCN−
SCN−
4.5
30
44
33
\50
\50
Br−
Cl−
F−
−
NO3
CH3COO−
−
N3
CH3CH2OH
(CH3)2CHNH2
−
NO2
20
\50
CN−
a I50 represents the nucleophile concentration at which the initial rate of ring closure of 250 mM (R)-1 is 50% of the initial rate in the absence of
the nucleophile.

J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 13 (2002) 1083–1089
1087
Table 3. Enantioselective ring opening of racemic para-nitrostyrene oxide 2a with various nucleophiles, catalysed by halohy-
drin dehalogenase from A. radiobacter AD1
Nucleophile
Nucleophile conc. (mM)
Init. activityb (mmol min−1 mg−1
)
E-value
Conversion (%) e.e. epoxidec (%)d
Cl−
100
10
10
1.57
0.09
0.55
0.18
\40
45
105
26
51
52
51
36 (S)
92 (S)
97 (S)
CN−
−
NO2
−
N3
1.3
\200
\99 (S)
a Concentration racemic 2, 3 mM, except with N3−, 2 mM.
b Initial activity towards (R)-para-nitrostyrene oxide 2 in a kinetic resolution experiment.
c Highest observed e.e.
d Value between parentheses is the absolute configuration of the unreacted enantiomer.
On the basis of the screening, we conclude that Br−,
lectivity towards the (R)-enantiomer (E >200) and the
remaining (S)-para-nitrostyrene oxide 2 was obtained
with an e.e. of more than 99%.
Cl−, N3 , NO2 and CN− are accepted and F−, OCN−,
−
−
SCN−, NO3 , CH3COO−, ethanol and isopropylamine
−
are not accepted as nucleophiles in the ring opening of
epoxide 2 catalysed by the halohydrin dehalogenase
from A. radiobacter AD1. The substantial differences in
initial activities towards both enantiomers indicate that
kinetic resolutions will likely occur with high
enantioselectivity.
2.4. Kinetic resolution of para-nitrostyrene oxide by
ring opening with selected nucleophiles
Kinetic resolutions of 3 mM racemic epoxide 2
catalysed by halohydrin dehalogenase from A.
radiobacter AD1 were performed in the presence of the
nucleophiles Cl−, N3 , NO2 or CN− (Table 3). The Br−
nucleophile was not included since the high equilibrium
constant makes the preparation of halohydrin 1 starting
from epoxide 2 and Br− impractical (since very high
concentrations of NaBr e.g. >10 M) are needed to
achieve complete conversion (>99%) of 3 mM epoxide
2.
−
−
All kinetic resolutions occurred with high enantioselec-
tivity towards the (R)-enantiomer leaving the (S)-enan-
tiomer of the epoxide behind. With 100 mM Cl− the e.e.
of the epoxide reached a maximum of 36% (Fig. 5). The
low conversion with this nucleophile, due to an unfa-
vorable equilibrium position, did not allow an exact
determination of the E-value (E >40). The ring opening
−
by NO2 occurred with high enantioselectivity (E=105)
Figure 5. Enantioselective ring opening of racemic para-
nitrostyrene oxide by NO2 (A) and Cl− (B), catalysed by
−
resulting in an almost enantiomerically pure unreacted
(S)-para-nitrostyrene oxide. To our knowledge, the
enzyme-catalysed nucleophilic ring opening of an epox-
halohydrin dehalogenase from A. radiobacter AD1. Incuba-
tions contained: (A), 10 mM NO2−, 3 mM 2 and 5.3 mM
enzyme; (B), 100 mM Cl−, 3 mM 2 and 0.74 mM enzyme.
Symbols: ꢀ, (S)-enantiomer; ꢁ, (R)-enantiomer.
−
ide by NO2 has not been described before. Ring open-
ing by CN− occurred with a low conversion rate (0.09
mmol min−1 mg−1) and moderate enantioselectivity (E=
45), resulting in (S)-para-nitrostyrene oxide with an e.e.
of 92% after 24 h incubation. Since CN− is a very poor
leaving group, it is highly unlikely that ring closure of
the formed b-hydroxynitrile occurs, making the reac-
tion irreversible. The low enantiomeric purity of the
remaining (S)-enantiomer of the epoxide is more likely
to be due to the low reaction rate or inactivation.
3. Conclusion and discussion
This paper describes the use of chromogenic substrates
to investigate diverse halohydrin dehalogenase-
catalysed reactions. The ring closure of para-nitro-2-
bromo-1-phenylethanol 1 and the ring opening of
para-nitrostyrene oxide 2 by Br− can be monitored
on-line since these reactions result in an increase and
decrease in absorbance, respectively. The bend in the
progress curve gives an indication of the enantioselec-
A complete conversion of the (R)-enantiomer of the
epoxide could be achieved with a small excess of N3 .
Racemic epoxide 2 was converted with high enantiose-
−

1088
J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 13 (2002) 1083–1089
tivity of the conversion of halohydrin. An obvious
limitation of the assay is that its predictive value with
regard to the activity and enantioselectivity of other
substrates will be limited to compounds that are struc-
turally related to the chromogenic substrates. On the
other hand, this method seems to be suitable since three
distinctly different types of halohydrin dehalogenases
converted the halohydrin 1.
closure and ring opening reaction gave (R)-g-chloro-b-
hydroxybutyronitrile in 95% e.e. and 65% yield.24
This paper described the use of chromogenic substrates
for halohydrin dehalogenases which can be used to
study various aspects of halohydrin-catalysed conver-
sions, such as steady state kinetics, enantioselectivity of
the ring-opening and ring-closure reaction and nucleo-
phile selectivity. The assay can be adopted to a 96-
wells format which provides an easy method for screen-
ing large number of mutant colonies, such as those
produced in directed evolution experiments.
A range of nucleophiles was tested for their ability to
react with epoxide 2 in the presence of halohydrin
dehalogenase from A. radiobacter AD1. Ring opening
of racemic epoxide 2 with Cl−, N3 , NO2 and CN−
resulted in enantioselective kinetic resolutions, in which
the (R)-enantiomer was in all cases the preferred enan-
tiomer. A complex product mixture was obtained dur-
−
−
4. Experimental
4.1. General
ing the ring opening reactions with NO2 and CN−, as
−
judged by chiral HPLC.
The enantiomeric excess (e.e.) and yields of various
compounds were determined by chiral HPLC (Chiral-
pak AS, Daicel). Spectrophotometric assays were con-
The enzyme-catalysed ring opening of an epoxide by
−
NO2 is the most remarkable since, to our knowledge,
−
enzymes that accept NO2 as a nucleophile have not
ducted in
a
Perkin Elmer Lambda BIO 40
been described in the literature before. The ambient
−
spectrophotometer. NMR spectra were recorded in
CDCl3. Halohydrin dehalogenases from A. radiobacter
AD1, Arthrobacter sp. AD2 and Mycobacterium sp.
GP1 were overexpressed in E. coli and purified as
described before.4
nucleophilic nature of NO2 (possible attack by either
the oxygen or nitrogen atom) and the regioselectivity of
ring opening of the epoxide (a- or b-attack) make the
formation of several products possible. If a nitro ester
(R-O-NꢀO) is formed, various side products, such as a
diol, can be expected since this class of compounds is
unstable in aqueous media.17 A further study into the
regioselectivity and enantioselectivity of these ring-
opening reactions, which will require the synthesis of all
possible products, is now in progress.
4.2. Synthesis of substrates
4.2.1. Racemic para-nitro-2-bromo-1-phenylethanol, 1.
To a cooled solution of v-bromo-para-nitroacetophe-
none (5.0 g, 20 mmol) in methanol (50 mL), was added
sodium borohydride (1.0 g, 26 mmol) and stirred for 3
h. Water (50 mL) was added and the mixture was
extracted with diethyl ether. After separation, the
organic phase was washed with brine, dried with
MgSO4 and removed by a rotary evaporator yielding
an orange solid (4.1 g) that consisted of a 7:3 mixture
of 1 and para-nitrostyrene oxide 2. Pure 1 was obtained
−
The enantioselective ring opening of epoxides by N3
catalysed by halohydrin dehalogenase from A.
radiobacter AD1 was studied in more detail recently.11
Various (substituted) styrene oxides were converted to
the corresponding azido alcohols with high enantiose-
lectivity and regioselectivity. The regioselectivity
towards the b-carbon is opposite to the selectivity of
the non-enzyme-catalysed reaction. The highly enan-
by flash chromatography on silica 60
H using
−
tioselective ring opening of the epoxide by N3 gives
1
petroleum ether/diethyl ether (ratio 7:3). H NMR l:
2.69 (d, 1H, OH, J=3.7 Hz); 3.52 (m, 2H); 5.00 (m,
1H); 7.54 (d, 2Har, J=8.8 Hz); 8.20 (d, 2Har, J=8.8
Hz). 13C NMR l: 36.8 (C-1); 70.1 (C-2); 121.3; 124.4;
144.8; 145.3 (Car).
access to optically active azido alcohols, which are
direct precursors to highly valuable and biologically
active amino alcohols.18,19
Enantioselective enzymatic conversions of epoxides can
occur in different ways. Hydrolysis of epoxides is
catalysed by epoxide hydrolases from various sources.20
Enantioselective ring opening by amines has been car-
ried out in the presence of liver microsomes and
lipases.21,22 The ring opening of an aliphatic epoxide by
4.2.2. Racemic para-nitrostyrene oxide, 2. Racemic
para-nitrostyrene oxide 2 was prepared by treating a
7:3 mixture of 1 and 2 (1.0 g) dissolved in diethyl ether
with aqueous KOH solution (15 mL, 1 M). The mixture
was heated under reflux for 15 min, cooled, diluted with
sulfuric acid (20 mL, 1 M) and extracted with diethyl
ether. After separating, the organic layer was dried with
MgSO4 and removed using a rotary evaporator to yield
pure 2 (0.76 g), which was used without further purifi-
−
N3 was catalysed by an immobilised enzyme prepara-
tion from a Rhodococcus sp.23 A halohydrin dehaloge-
nase from Corynebacterium sp. N-1074 catalysed the
ring opening of epichlorohydrin by cyanide, yielding
(R)-g-chloro-b-hydroxybutyronitrile in low enan-
tiomeric purity.24 The decrease in enantiomeric purity
of the product, caused by a competing non-enzyme-
catalysed conversion, could be overcome by in-situ
generation of the epoxide from 1,3-dichloro-2-propanol
catalysed by the same enzyme. The combined ring
1
cation. H NMR l: 2.72 (dd, 1H, J=2.6 Hz and 5.5
Hz); 3.17 (dd, 1H, J=4.0 Hz and 5.5 Hz); 3.91 (dd, 1H,
J=2.6 Hz and 4.0 Hz); 7.40 (d, 2Har, J=8.8 Hz); 8.16
(d, 2Har, J=8.8 Hz). 13C NMR l: 48.9 (C-1); 49.2
(C-2); 121.3; 123.7; 142.7; 145.3 (Car).

J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 13 (2002) 1083–1089
1089
4.2.3. Enantiomerically pure para-nitro-2-bromo-1-
References
phenylethanol, 1 and para-nitrostyrene oxide, 2. Enan-
tiomerically pure 1 and 2 were obtained by preparative
HPLC using an analytical Chiralpak AS column with
hexane/isopropanol (95:5) as eluent. Retention times of
individual enantiomers (e.e. >99%): (R)-1, 45.4 min;
(S)-1, 52.8 min; (R)-2, 17.2 min; (S)-2, 25.3 min.
1. Kasai, N.; Suzuki, T.; Furukawa, Y. J. Mol. Catal. B
1998, 4, 237.
2. Assis, H. M. S.; Bull, A. T.; Hardman, D. J. Enzyme
Microb. Technol. 1998, 22, 545.
3. Lutje Spelberg, J. H.; van Hylckama Vlieg, J. E. T.;
Bosma, T.; Kellogg, R. M.; Janssen, D. B. Tetrahedron:
Asymmetry 1999, 10, 2863.
4. van Hylckama Vlieg, J. E. T.; Tang, L.; Lutje Spelberg, J.
H.; Smilda, T.; Poelarends, G. J.; Bosma, T.; van
Merode, A.; Fraaije, M. W.; Janssen, D. B. J. Bacteriol.
2001, 183, 5058.
5. Yu, F.; Nakamura, W.; Mizunashi, W.; Watanabe, I.
Biosci. Biotechnol. Biochem. 1994, 58, 1451.
6. Verschueren, K. H.; Seljee, F.; Rozeboom, H. J.; Kalk,
K. H.; Dijkstra, B. W. Nature 1993, 363, 693.
7. Keuning, S.; Janssen, D. B.; Witholt, B. J. Bacteriol.
1985, 163, 635.
8. Holloway, P.; Trevors, J. T.; Lee, H. J. Microbiol. Meth-
ods 1998, 32, 31.
9. Marvanova´, S.; Nagata, Y.; Wimmerova´, M.; Sy´korova´,
J.; Hynkova´, K.; Damborsky´, J. J. Microbiol. Methods
2001, 44, 149.
10. Nakamura, T.; Nagasawa, T.; Yu, F.; Watanabe, I.;
Yamada, H. Biochem. Biophys. Res. Commun. 1991, 180,
124.
11. Lutje Spelberg, J. H.; van Hylckama Vlieg, J. E. T.;
Tang, L.; Janssen, D. B.; Kellogg, R. M. Org. Lett. 2001,
3, 41.
12. Westkaemper, R. B.; Hanzlik, R. P. Arch. Biochem.
Biophys. 1981, 208, 195.
13. Rink, R.; Kingma, J.; Lutje Spelberg, J. H.; Janssen, D.
B. Biochemistry 2000, 39, 5600.
14. Baldascini, H.; Ganzeveld, K. J.; Janssen, D. B.;
Beenackers, A. A. C. M. Biotech. Bioeng. 2001, 73, 44.
15. Rink, R.; Lutje Spelberg, J. H.; Pieters, R. J.; Kingma, J.;
Nardini, M.; Kellogg, R. M.; Dijkstra, B. W.; Janssen, D.
B. J. Am. Chem. Soc. 1999, 121, 7417.
16. Straathof, A. J. J.; Jongejan, J. A. Enzyme Microb.
Technol. 1997, 21, 559.
17. Iglesias, E.; Garcia-Rio, L.; Leis, R. J.; Pena, M. E.;
Williams, D. L. H. J. Chem. Soc. Perkin Trans. 2 1992,
1673.
4.3. Enzyme assays with chromogenic substrates
In a typical experiment, a stock solution of 1 or 2 in
DMSO (2 mL) was added to a cell containing buffer
(1.00 mL) at 22°C to give a final concentration of
approximately 250 mM. In ring-opening and inhibition
experiments the pH of the medium was 7.2 (300 mM
Tris-SO4) and in ring-closure reactions 7.5 (100 mM
Tris-SO4). In ring-opening and inhibition experiments,
the nucleophile was added as its sodium salt (except for
ethanol and isopropylamine) from a concentrated
aqueous stock solution. The conversion was started by
addition of the enzyme and the change in absorbance at
310 nm was monitored. The extinction coefficients of 1,
3050 M−1 cm−1 and 2, 4289 M−1 cm−1 at 310 nm were
used to calculate concentrations of 1 and 2.
4.4. Kinetic resolution experiments
Substrate 1 or 2 was added to a screw-capped bottle
containing Tris buffer (20 mL) at 22°C to give a final
concentration of 3 mM. In ring-opening experiments
the pH of the medium was 7.2 (300 mM Tris-SO4) and
in ring-closure reactions 7.5 (100 mM Tris-SO4). In
ring-opening experiments, the nucleophile was added to
a concentration indicated in Table 3. The reaction was
started by addition of the enzyme. The reaction was
monitored by periodically taking samples from the
reaction mixture and extracting them with diethyl ether
containing an internal standard such as mesitylene or
acetophenone. The e.e. and conversion of 1 and 2 were
determined by chiral HPLC analysis as described
above. In ring-closure experiments the E-value of a
kinetic resolution was calculated from the enantiomeric
excess of the halohydrin 1 and the epoxide 2 and in
ring-opening experiments the E-value was calculated
from the conversion and the enantiomeric excess of
epoxide 2.16
18. Martinez, L. E.; Leighton, J. L.; Carsten, D. H.; Jacob-
sen, E. N. J. Am. Chem. Soc. 1995, 117, 5897.
19. Farrow, J. F.; Schaus, S. E.; Jacobsen, E. N. J. Am.
Chem. Soc. 1996, 118, 7420.
20. Archer, I. V. J. Tetrahedron 1997, 53, 15617.
21. Kamal, A.; Rao, A. B.; Rao, M. Tetrahedron Lett. 1992,
33, 4077.
Acknowledgements
22. Kamal, A.; Rao, M. Tetrahedron: Asymmetry 1994, 5,
1881.
23. Mischitz, M.; Faber, K. Tetrahedron Lett. 1994, 35, 81.
24. Nakamura, T.; Nagasawa, T.; Yu, F.; Watanabe, I.;
Hideaki, Y. Tetrahedron 1994, 50, 11821.
This research was financially supported by STW
(GCH.3877) and by Innovation Oriented Research Pro-
gram (IOP) on Catalysis (No. 94007a) of the Dutch
Ministry of Economic Affairs.