A tandem enzyme reaction to produce optically active halohydrins, epoxides and diols

Article abstract of DOI:10.1016/S0957-4166(99)00308-0

The recombinant halohydrin dehalogenase from Agrobacterium radiobacter AD1 was used to obtain enantiomerically pure halohydrins and epoxides by kinetic resolution. By adding an excess of the recombinant epoxide hydrolase from the same organism the reversible conversion was drawn to completion. Halohydrins such as (S)-2,3-dichloro-1-propanol (E>100) and (S)-2-chloro-1-phenylethanol (E=73) were obtained with an enantiomeric excess of higher than 99%. This is a novel biocatalytic route for obtaining enantiomerically pure aromatic halohydrins and epoxides.

Full text of DOI:10.1016/S0957-4166(99)00308-0

Pergamon
Tetrahedron: Asymmetry 10 (1999) 2863–2870
A tandem enzyme reaction to produce optically active
halohydrins, epoxides and diols
a,b
a
a
Jeffrey H. Lutje Spelberg, Johan E. T. van Hylckama Vlieg, Tjibbe Bosma,
b
a,∗
Richard M. Kellogg and Dick B. Janssen
a
Department of Biochemistry, Groningen Biomolecular Sciences & Biotechnology Institute, University of Groningen,
Nijenborgh 4, 9747 AG, Groningen, The Netherlands
Department of Organic & Molecular Inorganic Chemistry, University of Groningen, Nijenborgh 4, 9747 AG, Groningen,
The Netherlands
b
Received 31 May 1999; accepted 26 July 1999
Abstract
The recombinant halohydrin dehalogenase from Agrobacterium radiobacter AD1 was used to obtain enantio-
merically pure halohydrins and epoxides by kinetic resolution. By adding an excess of the recombinant epoxide
hydrolase from the same organism the reversible conversion was drawn to completion. Halohydrins such as (S)-2,3-
dichloro-1-propanol (E>100) and (S)-2-chloro-1-phenylethanol (E=73) were obtained with an enantiomeric excess
of higher than 99%. This is a novel biocatalytic route for obtaining enantiomerically pure aromatic halohydrins and
epoxides. © 1999 Elsevier Science Ltd. All rights reserved.
1. Introduction
Microorganisms provide us with a variety of enzymes that have evolved to help the organism deal with
the presence of synthetic compounds in their environment. In single organisms multiple enzymes can be
involved in the degradation of xenobiotics. Examples can be found in bacteria that metabolize epoxides or
1
haloaliphatics. If such enzymes are enantioselective, they can be useful for the preparation of optically
active intermediates. In the fine chemicals industry, enantiomerically pure epoxides are used as building
blocks for various pharmaceutical products. Several biocatalytic methods to produce optically active
2
epoxides using enzymes involving epoxide degradation or formation have been reported. Examples are
direct epoxidation of alkenes by mono-oxygenases or kinetic resolution of racemic epoxides. Kinetic
resolutions have been carried out by means of lipases, using another nearby functionality such as an
3
ester, or by using epoxide hydrolases for the direct resolution of the epoxide. Outstanding examples
∗
Corresponding author. Tel: 31-50-3634209; fax: 31-50-3634165; e-mail: d.b.janssen@chem.rug.nl
0957-4166/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(99)00308-0

2864
J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 10 (1999) 2863–2870
of the synthesis of enantiomerically pure epoxides by kinetic resolution with metal containing salen
complexes have recently been published by Jacobsen et al.4,5
Halohydrins can be considered as direct precursors of epoxides. Ring closure of an enantio-
merically pure halohydrin generally yields an enantiomerically pure epoxide. Aromatic halohydrins
such as 2-chloro-1-phenylethanol can be obtained enantiomerically pure by microbial reduction of α-
haloacetophenones or by kinetic resolution using lipases.6
,7
8
Enantiomerically pure halohydrins can also be produced using halohydrin dehalogenases. These
enzymes catalyze the ring closure of a vicinal halohydrin to an epoxide and the reverse reaction.
Several organisms capable of dehalogenating halohydrins have recently been isolated. Two halohydrin
9
dehalogenases from Corynebacterium sp. were characterised and the genes were cloned and sequenced.
10
Recently, a halohydrin dehalogenase from Arthrobacter erithii H10a has been characterised. This
enzyme could be used to synthesize optically active epihalohydrins from 1,3-dichloro-2-propanol by
enantioselective dehalogenation and halogenation.11
We have isolated several organisms that can degrade chlorinated aliphatic compounds, including
bacteria that could grow on halohydrins and epichlorohydrin. A halohydrin dehalogenase from the 3-
chloro-1,2-propanediol-utilizing bacterium Arthrobacter strain AD2 was purified and characterised.12
Another organism, Agrobacterium radiobacter AD1, degraded the toxic epichlorohydrin to glycerol by
an epoxide hydrolase and a halohydrin dehalogenase (Fig. 1).13
Figure 1. Degradation route of epichlorohydrin to glycerol by Agrobacterium radiobacter AD1. Epoxide hydrolase (steps a and
c) and halohydrin dehalogenase (step b)
Recently, we have cloned and overexpressed a gene that encodes the epoxide hydrolase of A.
14
radiobacter AD1. With this enzyme enantiomerically pure aromatic epoxides could be obtained
15
by kinetic resolution although the E-values were only moderate. Styrene oxide, for example, was
hydrolysed with an E-value of 16. Non-aromatic epoxides such as epichlorohydrin were hydrolysed with
low enantioselectivity (E<2). We have also cloned and overexpressed the second enzyme involved in the
degradation pathway, the halohydrin dehalogenase.16 Research with halohydrin dehalogenases has so far
mainly been focused on C3 epoxides and halohydrins, such as 2,3-dichloro-1-propanol and 1,3-dichloro-
1
-propanol. One paper has been published describing the conversion of aromatic halohydrins by resting
17
cells of Flavobacterium rigense, but the enantioselectivity was very low.
The goal of our research was to establish the enantioselectivity of the halohydrin dehalogenase from
A. radiobacter AD1 towards a variety of substrates, including halohydrins with an aromatic moiety
(Fig. 2). Since the application of epoxides and halohydrins as precursors for drugs requires them to
be available in enantiomerically pure form it was the objective to obtain the investigated halohydrins
with an enantiomeric excess (ee) of higher than 99%.
Figure 2. Halohydrins used as substrates by the halohydrin dehalogenase from Agrobacterium radiobacter AD1

J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 10 (1999) 2863–2870
2865
2. Results and discussion
The halohydrin dehalogenase from A. radiobacter enantioselectively dehalogenated 2,3-dichloro-1-
propanol 1 and 2,3-dibromo-1-propanol 2. The conversion of 1 proceeded with a very high enantioselec-
tivity (E>100) for the (R)-enantiomer, leaving the (S)-enantiomer unreacted. Initially the (R)-enantiomer
−1
−1
was converted with a specific activity of 0.9 µmol min mg , but the reaction gradually slowed down
and after 20 h an ee of 96% was reached. The yield of the remaining (S)-enantiomer was higher then
49.5% (maximum theoretical yield is 50.0%). Instead of enantiomerically pure epichlorohydrin, racemic
epichlorohydrin and prochiral 1,3-dichloro-2-propanol were formed. This can be explained by the fact
that the enzyme also catalysed the reverse reaction, the cleavage of the epoxide ring by an halide (Fig. 3).
Figure 3. Conversion of 1 by halohydrin dehalogenase and epoxide hydrolase. Epoxide hydrolase (steps 2 and 3) and halohydrin
dehalogenase (steps 1, 3, 4a and 4b)
The favoured attack of the chloride in the halogenation reaction was at the β-position of epichloro-
hydrin to yield 1,3-dichloro-2-propanol (step 4a) instead of α-attack to yield 1 (step 1). This prochiral
1
,3-dichloro-2-propanol was then converted to both enantiomers of epichlorohydrin (steps 4a and 4b).
−1
−1
The specific activity for 1,3-dichloro-2-propanol was 3.9 µmol min mg , which is more than fourfold
higher than the activity for 1. Because of these side reactions the conversion of 1 was inhibited during
the course of the reaction.
To overcome the unwanted side reactions (steps 4a and 4b), an excess of recombinant epoxide
hydrolase from A. radiobacter AD1 was added. In this way the formed (S)-epichlorohydrin was converted
immediately to (S)-3-chloro-1,2-propanediol (step 2). This was then further converted via glycidol
to glycerol (step 3). In this case, formation of 1,3-dichloro-2-propanol and epichlorohydrin were not
observed. The enantiomeric excess of (S)-3-chloro-1,2-propanediol was dependent on the amount of the
epoxide hydrolase that was added due to competition with chemical hydrolysis and the reverse reaction.
The maximum measured ee was 91%. In the presence of epoxide hydrolase, the conversion of (R)-1 was
accelerated and a maximum ee of higher than 99% for (S)-1 was reached in less than 5 h (Fig. 4). The
analytical yield of 1 was higher than 49.5%.
Another method to circumvent the reversibility of a halohydrin dehalogenase was recently described
by Assis et al.11 They studied the conversion of 1,3-dichloro-2-propanol to (R)-epichlorohydrin, which
could be obtained with an ee of 89.5%. By adding an excess of bromide, only the contaminating (S)-

2866
J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 10 (1999) 2863–2870
Figure 4. Kinetic resolution of 1 with halohydrin dehalogenase (ꢀ,ꢁ), with addition of excess epoxide hydrolase (ꢂ,◦). (R)-1:
ꢂ,ꢀ and (S)-1: ◦,ꢁ
epichlorohydrin was halogenated and the (R)-epichlorohydrin was obtained with an ee of higher than
95%, but only in a 10.7% yield.
Surprisingly halohydrins containing an aromatic group such as halohydrins 3 and 5 were also con-
verted, with E-ratios of 73 and 23, respectively (Table 1). This high activity and enantioselectivity of
a halohydrin dehalogenase towards aromatic halohydrins has not been described before. In the case of
the kinetic resolution of halohydrin 3, (R)-styrene oxide was formed with an ee of 90%. Because of the
reversibility of the reaction the conversion was incomplete and the ee of (S)-3 reached a maximum of
only 92%. Formation of halohydrin 4 was not observed during the reaction. With addition of an excess
of epoxide hydrolase, resulting in removal of the inhibiting epoxide, the conversion was completed and
an ee of the (S)-enantiomer of higher than 99% was reached.
The diols produced by hydrolysis of the aromatic epoxides had a lower ee than the corresponding
epoxides. This was mainly due to the regioselectivity (β-attack versus α-attack) of the epoxide hydrolase.
When, in a separate experiment, enantiomerically pure (R)-styrene oxide or (R)-m-chlorostyrene oxide
were hydrolysed by epoxide hydrolase, the (R)-diols were formed with an ee of, respectively, 85% and
90%.
The kinetic resolution of halohydrin 4 was strongly influenced by the high chemical conversion of
the substrate in the buffer solution (42% degradation per h in buffer pH 7.5, 30°C) leading besides the
epoxide to major side products. This was probably due to dimerisation of the halohydrin.18 When the
kinetic resolution of 4 was performed without the addition of epoxide hydrolase, (R)-styrene oxide was
formed with an ee of 89%, indicating an E-value of at least 50. The chemical instability limits a practical
application of the kinetic resolution of this halohydrin. No significant chemical conversion was observed
with halohydrins 1, 2, 3 and 5.
The importance of the addition of epoxide hydrolase was substrate dependent as shown in Table 2.
In case of the conversion of halohydrin 1, without epoxide hydrolase the kinetic resolution was strongly
inhibited due to the side reactions shown in Fig. 3. The formed optically active epichlorohydrin rapidly
racemised. In the case of the aromatic halohydrins 3 and 5, no side reactions of the epoxide to other

J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 10 (1999) 2863–2870
2867
Table 1
The kinetic resolution of halohydrins with recombinant halohydrin dehalogenase and epoxide hydro-
lase
products occurred. With these compounds, the importance of epoxide hydrolase was dependent on the
ratio of the concentrations of epoxide and halohydrin at equilibrium. Without the addition of epoxide
hydrolase, (S)-3 was obtained with a lower ee than (S)-5 because with the latter the equilibrium was
positioned further towards the epoxide.
Table 2
The kinetic resolution of halohydrins with recombinant halohydrin dehalogenase, with and without
epoxide hydrolase
It is important to state that the excess epoxide hydrolase was only added to draw the kinetic resolution
of the halohydrins to completion by removing the inhibiting epoxide. The epoxide hydrolase itself has
no influence on the enantioselectivity of the conversion.
These results are an example of a tandem enzyme reaction. The combined use of these enzymes was
developed by evolution in nature for the degradation of xenobiotic compounds. We used this combination
as a biocatalyst to produce optically active halohydrins, epoxides and diols. Since the genes for both
enzymes were cloned and the enzyme was brought to overexpression, application on an industrially
interesting scale may become feasible.

2868
J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 10 (1999) 2863–2870
3. Experimental
3.1. General
The enantiomeric excesses and the yields of the halohydrins, epoxides and derivatised diols were
determined with a Hewlett–Packard 5890 gas chromatograph equipped with a FID-detector, using a
Chiraldex G-TA capillary column (col I, 50 m, Astec), Chiraldex B-TA capillary column (col II, 30
m, Astec) or β-dex 120 column (col III, 30 m, Supelco) all of 0.25 mm inside diameter. NMR spectra
were recorded in CDCl . Halohydrin 1 was purchased from Merck, halohydrin 2 from Fluka, halohydrin
3
3
from Lancaster and m-chlorostyrene oxide was a gift from DSM. All other chemicals were purchased
from Aldrich.
3.2. Synthesis of halohydrins
Halohydrins 4 and 5 were synthesised by treating the corresponding epoxides with 36% aqueous HCl
19
according to described methods. To a solution of styrene oxide (1.0 g, 8.3 mmol) in CHCl (80 ml),
3
36% aqueous HCl was added (25 ml) and stirred for 1 h. After separating, the organic phase was washed
with saturated aqueous NaHCO and water, dried with MgSO , and the solvent removed by a rotary
3
4
1
evaporator giving halohydrin 4 in 90% yield as a colourless oil. H NMR, δ: 2.2–2.4 (s, 1H), 3.89 (dd,
1
3
2
H), 4.98 (t, 1H), 7.3 (m, 5H); C NMR, δ: 62.25, 65.34, 125.01, 126.30, 126.37, 135.48.
To a solution of m-chlorostyrene oxide (0.75 g, 4.8 mmol) in acetonitrile (5 ml) were added LiClO4
(
0.80 g, 7.5 mmol) and NH Cl (0.40 g, 7.5 mmol) and the resulting mixture was stirred for 4 h at 80°C.
4
The reaction mixture was diluted with water and extracted with ether. The ether was washed with water,
dried with MgSO , and removed by a rotary evaporator yielding a colourless oil. Flash chromatography
4
1
on silica 60 H using petroleum ether:ether (7:3) as eluent yielded pure halohydrin 5 in 25% yield. H
NMR, δ: 2.6 (s, 1H), 3.55 (dd, 1H), 3.65 (dd, 1H), 4.85 (dd, 1H), 7.2–7.3 (m, 3H), 7.36 (s, 1H); 13
NMR, δ: 48.12, 70.86, 121.71, 123.79, 126.08, 127.42, 132.13, 139.38.
C
3.3. Production of halohydrin dehalogenase and epoxide hydrolase
A gene library of A. radiobacter AD1 was constructed in the cosmid vector pLAFR3. After in
vitro packaging, the library was transduced to Escherichia coli HB101. Transconjugants were screened
for dehalogenase activity with 1,3-dichloro-2-propanol. The halohydrin dehalogenase gene, designated
hheC, was sequenced and subsequently amplified by PCR and cloned behind the T7 promotor of the
+
20
expression vector pGEF , yielding pGEFhheC. The halohydrin dehalogenase gene was overexpressed
up to 30% of soluble protein by introduction of pGEFhheC in E. coli BL21(DE3).
For the described kinetic resolutions purified enzyme was used. Plasmid DNA was transformed by
electroporation to competent E. coli BL21(DE3) cells, which were then plated out on LB medium
containing tetracycline and incubated overnight at 30°C. A preculture was started by inoculating 100 ml
of LB medium containing tetracycline with the transformants from a plate to a initial OD600 of 0.1. The
culture was incubated at 30°C until an OD600 of 1–2 was reached, diluted in 1 l of LB medium containing
tetracycline and incubated overnight at 20°C. The cells were subsequently centrifuged, washed and
resuspended. A crude extract was prepared by ultrasonic disruption and centrifugation of the cells.
This was followed by a purification step with a Resource Q column. The purified recombinant epoxide
hydrolase from A. radiobacter AD1 was prepared as described before.

J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 10 (1999) 2863–2870
2869
3.4. Kinetic resolution with halohydrin dehalogenase and epoxide hydrolase
A screw-capped bottle containing 30 ml Tris buffer (100 mM, pH 7.5) was incubated in a shaking
waterbath at 30°C. The halohydrin was added to a final concentration of 5 mM. The reaction was started
by the addition of the appropriate amount of purified halohydrin dehalogenase. In experiments where
epoxide hydrolase was used, this enzyme was added immediately afterwards. The reaction was monitored
by periodically taking samples from the reaction mixture. The samples were extracted with diethyl ether
containing mesitylene or 1-chlorohexane as an internal standard. Prior to analysis by chiral GLC, the
samples were dried over a short column containing MgSO . The diols were analysed after derivatisation
4
to their acetonides. The diethyl ether was vaporized under a stream of nitrogen and the sample was
redissolved in 0.5 ml 2,2-dimethoxypropane. The solution was then shaken for 1 h with 200 mg amberlite
+
IR-120 (H ). After addition of 50 mg NaHCO the organic phase was analysed by chiral GLC.
3
3.5. Calculation of enantioselectivity (E) and activity
The enantioselectivity of the kinetic resolutions with the halohydrin dehalogenase was determined
in the presence of an excess of epoxide hydrolase. The enantioselectivity of the kinetic resolutions of
halohydrins 3 and 5 was calculated using the enantiomeric ratio21 which is defined by Eq. 1, where
Vmax and K represent the Michaelis–Menten parameters of both enantiomers. To estimate the kinetic
m
parameters, the equations that describe competitive Michaelis–Menten kinetics were fitted by numerical
integration to the data, as described before. With the obtained kinetic parameters or in cases where no
unique solutions were obtained, with the lumped parameter Vmax/K , the E-value was calculated using
m
Eq. 1.
(1)
The exact enantioselectivity of the conversions of halohydrins 1 and 2 could not be determined because
no significant decrease of the remaining enantiomer could be measured, indicating an E-value of at least
higher than 100. With halohydrin 4, the conversion rate of slow reacting enantiomer was approximately
equal to the chemical conversion rate, indicating an E-value of at least 50.
The activities as presented in Table 1 were calculated from the substrate depletion at a concentration
of 5 mM. In the case of halohydrin 4, the initial activity was corrected for chemical conversion.
3.6. Analysis of chiral compounds
The data from chiral GLC were as follows. Halohydrin 1 on col II: temp. prog. 60°C for 2 min, increase
with 3°C per min to 90°C, 20 min on 90°C, t =20.2 min and 20.8 min; halohydrin 2 on col I: temp. 105°C,
r
t =53.4 min and 55.0 min; halohydrin 3 on col III: temp. 140°C, t =26.8 min and 28.2 min; halohydrin 4
r
r
on col III: temp. 140°C, t =25.4 min and 26.0 min; halohydrin 5 on col III: temp. 170°C, t =21.8 min and
r
r
2
2.4 min. Styrene oxide on col I: temp. 110°C, t =15.3 min and 18.0 min; m-chlorostyrene oxide on col I:
r
temp. 110°C, t =20.2 min and 31.1 min; epichlorohydrin on col II: temp. prog. 60°C for 2 min, increase
r
with 3°C per min to 90°C, 20 min on 90°C, t =7.8 min and 8.2 min; derivatised 3-chloro-1,2-propanediol
r
on col I: temp. 70°C, t =42.1 min and 43.4 min; derivatised phenylethanediol on col III: temp. 130°C,
r

2870
J. H. Lutje Spelberg et al. / Tetrahedron: Asymmetry 10 (1999) 2863–2870
t =15.1 min and 16.3 min; derivatised m-chlorophenylethanediol on col III: temp. 150°C, t =18.4 min
r
r
and 19.7 min.
The absolute configurations of the halohydrins were determined by base-catalyzed ring closure of the
remaining enantiomers to the corresponding epoxides. The absolute configurations of epichlorohydrin,
styrene oxide and m-chlorostyrene oxide were determined by co-injection of the commercially available
optically pure epoxides. The absolute configurations of 3-chloro-1,2-propanediol and phenylethanediol
were determined by co-injection of the derivatised commercially available optically pure diols. The
absolute configuration of m-chlorophenylethanediol was determined by enzymatic hydrolysis of (R)-
m-chlorostyrene oxide and analysis of the formed and derivatised diol.
Acknowledgements
This research was financially supported by the Innovation Oriented Research Program (IOP) on
Catalysis (no. 94007a) of the Dutch Ministry of Economic Affairs.
References
1
2
3
4
5
6
7
8
9
. Fetzner, S.; Lingens, F. Microbiol. Rev. 1994, 58, 641.
. Faber, K. Biotransformations in Organic Chemistry, 3rd ed.; Springer-Verlag: New York, 1997.
. Archer, I. V. J. Tetrahedron 1997, 53, 15617.
. Furrow, M. E.; Schaus, S. E.; Jacobsen, E. N. J. Org. Chem. 1998, 63, 6776.
. Tokunaga, M.; Larrow, J. F.; Kakiuchi, F.; Jacobsen, E. N. Science 1997, 277, 936.
. de Carvalho, M.; Okamoto, M. T. Tetrahedron 1991, 47, 2073.
. Hiratake, J.; Inagaki, M.; Nishioka, T.; Oda, J. J. Org. Chem. 1988, 53, 6130.
. Kasai, N.; Suzuki, T.; Furukawa, Y. J. Mol. Catal. B 1998, 4, 237.
. Yu, F.; Nakamura, T.; Mizunashi, W.; Watanabe, I. Biosci. Biotechnol. Biochem. 1994, 58, 1451.
1
1
1
1
1
1
1
1
1
1
2
2
0. Assis, H. M. S.; Sallis, P. J.; Bull, A. T.; Hardman, D. J. Enzyme Microbial. Technol.1998, 22, 568.
1. Assis, H. M. S.; Bull, A. T.; Hardman, D. J. Enzyme Microbial. Technol. 1998, 22, 545.
2. van den Wijngaard, A. J.; Reuvekamp, P. T. W.; Janssen, D. B. J. Bacteriol. 1991, 173, 124.
3. van den Wijngaard, A. J.; Janssen, D. B.; Witholt, B. J. Gen. Microbiol. 1989, 135, 2199.
4. Rink, R.; Fennema, M.; Smids, M.; Dehmel, U.; Janssen, D. B. J. Biol. Chem. 1997, 272, 14650.
5. Lutje Spelberg, J. H.; Rink, R.; Kellogg, R. M.; Janssen, D. B. Tetrahedron: Asymmetry 1998, 9, 459.
6. van Hylckama Vlieg, J. E. T.; Lutje Spelberg, J. H.; Tang, L.; Janssen, D. B., manuscript in preparation.
7. Wandel, U.; Königsberger, K.; Griengl, H. Biocatalysis 1994, 10, 159.
8. Summerbell, R. K.; Kland-English, M. J. J. Am. Chem. Soc. 1955, 77, 5095.
9. Chini, M.; Crotti, P.; Gardelli, C.; Macchia, F. Tetrahedron 1992, 48, 3805.
0. Schanstra, J. P.; Rink, R.; Pries, F.; Janssen, D. B. Protein Express. Purif. 1993, 4, 479.
1. Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. Am. Chem. Soc. 1982, 104, 7294.