Purification and characterization of a yeast carbonyl reductase for synthesis of optically active (R)-styrene oxide derivatives

Article abstract of DOI:10.1271/bbb.69.79

Optically active styrene oxide derivatives are versatile chiral building blocks. Stereoselective reduction of phenacyl halide to chiral 2-halo-1-phenylethanol is the key reaction of the most economical synthetic route. Rhodotorula glutinis var. dairenensis IFO415 was discovered on screening as a potent microorganism reducing a phenacyl halide to the (R)-form of the corresponding alcohol. An NADPH-dependent carbonyl reductase was purified to homogeneity through four steps from this strain. The relative molecular mass of the enzyme was estimated to be 40,000 on gel filtration and 30,000 on SDS-polyacrylamide gel electrophoresis. This enzyme reduced a broad range of carbonyl compounds in addition to phenacyl halides. Some properties of the enzyme and preparation of a chiral styrene oxide using the crude enzyme are reported herein.

Full text of DOI:10.1271/bbb.69.79

Biosci. Biotechnol. Biochem., 69 (1), 79–86, 2005
Purification and Characterization of a Yeast Carbonyl Reductase
for Synthesis of Optically Active (R)-Styrene Oxide Derivatives
1
1
1
1;y
2
Noriyuki KIZAKI, Ikuo SAWA, Miho YANO, Yoshihiko YASOHARA, and Junzo HASEGAWA
1
Fine Chemicals Research Laboratories, Kaneka Corporation, 1-8 Miyamae, Takasago, Hyogo 676-8688, Japan
Life Science RD Center, Kaneka Corporation, 1-8 Miyamae, Takasago, Hyogo 676-8688, Japan
2
Received August 11, 2004; Accepted October 15, 2004
Optically active styrene oxide derivatives are versa-
tile chiral building blocks. Stereoselective reduction of
phenacyl halide to chiral 2-halo-1-phenylethanol is the
key reaction of the most economical synthetic route.
Rhodotorula glutinis var. dairenensis IFO415 was dis-
covered on screening as a potent microorganism reduc-
ing a phenacyl halide to the (R)-form of the correspond-
ing alcohol. An NADPH-dependent carbonyl reductase
was purified to homogeneity through four steps from
this strain. The relative molecular mass of the enzyme
was estimated to be 40,000 on gel filtration and 30,000
on SDS-polyacrylamide gel electrophoresis. This en-
zyme reduced a broad range of carbonyl compounds in
addition to phenacyl halides. Some properties of the
enzyme and preparation of a chiral styrene oxide using
the crude enzyme are reported herein.
There are several methodologies for synthesis of
optically active styrene oxide. Microbial enantioselec-
tive hydrolysis of racemic oxide to chiral oxide and the
corresponding diol has been reported, but the stereo-
2
)
selectivity was not sufficiently high. One of the most
practical methods is the use of optically active 2-halo-1-
phenylethanol as an intermediate; it is smoothly con-
verted to an oxide without loss of optical purity (Fig. 1).
3
)
Enzymatic stereoselective hydrolysis and transesterifi-
cation including the practical preparation of racemic
4
)
substrate for synthesis of optically active 2-halo-1-
phenylethanol (halohydrin) have been reported, but the
theoretical yield of the product by optical resolution is
50%, making it necessary to reuse the undesired isomer
in order to be economical.
Asymmetric reductions of phenacyl halide by Geo-
,6)
trichum5
baker’s yeast
7,8)
and phenylacetaldehyde
9
)
Key words: enantioselective reduction; carbonyl reduc-
tase; styrene oxide; Rhodotorula glutinis
reductase from Corynebacterium sp. have been re-
ported. Enzymatic reduction has been one of the most
powerful and practical methodologies for the synthesis
1
0)
The importance of optically active compounds has
increasingly been recognized in the field of pharma-
ceuticals. Optically active styrene oxide derivatives are
useful as versatile chiral building blocks. For example,
of optically active alcohols. In this paper, we describe
the screening and isolation of a novel carbonyl reductase
catalyzing the conversion of phenacyl halide to optically
active halohydrin from a Rhodotorula yeast.
(
for antidiabetic and antiobesity drugs.
R)-3-chlorostyrene oxide is an important intermediate
)
1
O
OH
R3
O
X
X
Enzyme
NaOH
R1
R3
R1
R3
R1
R2
R2
R2
Phenacyl halide
2-Halo-acetophenone)
Halohydrin
(2-Halo-1-phenylethanol)
Styrene oxide
(
Fig. 1. Synthesis of Optically Active Styrene Oxide Derivatives.
y
To whom correspondence should be addressed. Fax: +81-794-45-2668; E-mail: Yoshihiko.Yasohara@kaneka.co.jp
Abbreviations: COBE, ethyl 4-chloro-3-oxobutanoate; CPH, 2-chloro-1-(3-chlorophenyl)ethanol; CPN, 3-chlorophenacyl chloride; e.e., enan-
tiomeric excess; KPL, ketopantoyl lactone

80
N. KIZAKI et al.
Materials and Methods
2,000 ml of toluene twice. The organic layer was
evaporated under a vacuum. The oily residue was
Chemicals. Substituted phenacyl bromides were pre-
pared from substituted acetophenones using bromine.
Substituted phenacyl chlorides were prepared from
substituted acetophenones using sulfuryl chloride. Phen-
acyl chloride and phenacyl bromide were purchased
from Tokyo Kasei, Japan. Racemic halohydrins were
distilled to obtain (R)-2-chloro-1-(3-chlorophenyl)etha-
ꢁ
nol (CPH) (110–120 C/1 mmHg, 24.5 g, 90% yield,
over 99% e.e; NMR ꢁH (CDCl3): 2.71 (1H, s, OH), 3.62
(1H, dd, J ¼ 11:3 and 8.5 Hz, CH2Cl), 3.75 (1H, dd,
J ¼ 11:2 and 3.6 Hz, CH2Cl), 4.90 (1H, dd, J ¼ 8:3 and
3.4 Hz, CH), 7.25–7.30 (3H, m, Ph), 7.42 (1H, s, Ph);
2
0
prepared by NaBH reduction. Glucose dehydrogenase
4
½ꢂꢂ
ꢃ33:62 (c 1.0, CH OH). The mixture of 20 g of
D
3
was purchased from Amano Enzyme, Nagoya, Japan.
All other chemicals used in this study were of analytical
grade and commercially available.
(R)-CPH dissolved in 100 ml of toluene, and 6.5 g of
ꢁ
NaOH dissolved in water was stirred at 40 C for 3 h.
Then the reaction mixture was separated into two
phases: the organic phase was evaporated under a
vacuum, and the oily residue was distilled to obtain (R)-
Microorganisms and cultivation. Microorganisms
were obtained from our laboratory collection, the
collection of the National Institute of Technology and
Evaluation of Japan (IFO), and the Centraalbureau
voor Schimmelcultures of the Netherlands (CBS). The
medium was composed of 0.7% KH PO , 1.3%
ꢁ
3-chlorostyrene oxide (54–56 C/1 mmHg, 15.3 g, 95%
yield, over 99% e.e.); NMR ꢁH (CDCl3): 2.75 (1H, dd,
J ¼ 2:4 and 5.4 Hz, CH O), 3.15 (1H, dd, J ¼ 4:2 and
2
5.4 Hz, CH O), 3.82 (1H, dd, J ¼ 2:4 and 4.2 Hz, CH),
2
20
7.10 (1H, m, Ph), 7.20–7.26 (3H, m, Ph); ½ꢂꢂ
þ15:98
2
4
D
.
.
(
7
0
NH4)2HPO4, 0.08% MgSO4 7H2O, 0.007% ZnSO4
(c 1.0, CH3OH). The reported value for the (R)-isomer is
1)
25
.
.
H2O, 0.009% FeSO4 7H2O, 0.0005% CuSO4 5H2O,
½ꢂꢂD þ10:8.
.
.001% MnSO4 4H2O, 0.01% NaCl, 0.3% yeast extract,
and 4% glucose, pH 7.0. In the screening experiments,
each strain was inoculated into 5 ml of medium in a test
Enzyme assays and protein determination. Carbonyl
reductase activity for CPN was assayed spectrophoto-
metrically. The standard assay mixture, which was
composed of 100 mM potassium phosphate buffer
tube (24 mmꢀ ꢀ 200 mm), followed by incubation at
ꢁ
3
0 C with reciprocal shaking, usually for 24 h.
(
pH 6.5), 1 mM of CPN, 0.25 mM of NADPH, 0.3%
ꢁ
Screening method. Each reaction mixture, composed
(v/v) of dimethylsulfoxide, and the enzyme at 30 C in a
total volume of 3 ml, was monitored for decrease in
absorbance at 340 nm. One unit of the enzyme was
defined as the amount catalyzing the oxidation of 1 mmol
of coenzyme per min. Specific activity was expressed as
units per milligram of protein. Protein was measured
using the protein-dye binding method using bovine
of cells from 5 ml of culture broth, 5 mg of one of three
substituted monochlorophenacyl bromides or phenacyl
bromide, and 80 mg glucose in 1 ml of 100 mM potas-
sium phosphate buffer (pH 7.0), was shaken for 24 to
ꢁ
1
44 h at 30 C. Ethyl acetate (5 ml) was added to the
reaction mixture, followed by centrifugation. The
organic layer was then analyzed to determine the yield
and optical purity of the product.
1
1)
serum albumin as a standard.
Purification of carbonyl reductase. The purification
ꢁ
Enzymatic reduction of 3-chlorophenacyl chloride.
Rhodotorula glutinis var. dairenensis IFO415 was
cultured in 18-liter of the above medium at 30 C using
a 30-liter fermenter. Cells obtained from 500 ml of the
cultured broth by centrifugation were suspended in
procedure was performed at 0–4 C. Rhodotorula gluti-
nis var. dairenensis IFO415 was cultured in 18-liter of
ꢁ
ꢁ
the above medium at 30 C using a 30-liter fermenter.
Cells obtained by centrifugation of 5.6-liter of the
cultured broth were suspended with 1,000 ml of a
100 mM potassium phosphate buffer (pH 8.2) and dis-
150 ml of 100 mM potassium phosphate buffer (pH 7.0)
and disrupted with 0.25 mm-diameter glass beads (Dyno
ꢁ
rupted with 0.25 mm-diameter glass beads at 4 C. After
Mill KDL-A, W.A. Bachofen, Basel, Switzerland) at
ꢁ
centrifugation, the resulting supernatant was used as the
cell-free extract. The cell-free extract was fractionated
by the addition of solid ammonium sulfate to give a 45%
saturated solution while maintaining the pH at 7.5, and
the precipitate was discarded after centrifugation. The
resultant supernatant solution was then brought to 60%
saturation with ammonium sulfate and centrifuged
again. The supernatant was discarded while the precip-
itate was dissolved in a 10 mM potassium phosphate
buffer (pH 7.5) and dialyzed against the same buffer.
The dialyzed solution was applied to the DEAE-
Toyopearl 650M column (Tosoh, Tokyo, Japan) and
equilibrated with 10 mM potassium phosphate buffer
(pH 7.5). The enzyme was eluted with a linear NaCl
4
C. After centrifugation, the resulting supernatant was
used as the cell-free extract. In the case of the aqueous
reaction, a reaction mixture of 18 ml of the cell-free
extract, 2.7 g of 3-chlorophenacyl chloride (CPN), 2.7 g
þ
of glucose, 6 mg of NADP , and 1,200 units of glucose
ꢁ
dehydrogenase was stirred at 30 C. The pH of the
mixture was kept at 6.5 with 5 N sodium hydroxide. In
the case of the water/organic solvent two-phase reac-
tion, 18 ml of n-butyl acetate was added to the above
reaction mixture.
Synthesis of (R)-3-chlorostyrene oxide. The enzymat-
ic reaction mixture (2,000 ml) was extracted with

Carbonyl Reductase for Synthesis of Styrene Oxide
81
gradient solution (0–0.3 M). Active fractions were
collected and dialyzed against 10 mM potassium phos-
phate buffer (pH 7.5). Solid ammonium sulfate was
dissolved in the dialyzed solution (the final concentra-
tion was 1 M) and applied to the Phenyl-Toyopearl 650M
column (Tosoh, Tokyo, Japan) equilibrated with 10 mM
potassium phosphate buffer (pH 7.5) containing 1 M of
ammonium sulfate. The enzyme was eluted with an
ammonium sulfate linear gradient solution (1–0 M). The
active fractions were collected and dialyzed against
determined using an HPLC equipped with a Chiralcel
1
OB (4:6 mmꢀ ꢀ 250 mm) column (Daicel). H-NHR
spectra were recorded in CDCl , with a FT-NMR JM-
3
400 spectrometer (JEOL, Tokyo, Japan). Chemical shifts
were expressed in parts/million (ppm), with trimethyl-
silane as the internal standard. Optical rotation was
measured with a SEPA-200 digital polarimeter (Horiba,
Kyoto, Japan).
Results
1
0 mM potassium phosphate buffer (pH 7.5). The dia-
lyzed solution was applied to a Blue-Sepharose CL-6B
column (Amersham Biosciences, Piscataway, NJ,
U.S.A.). The enzyme was eluted with a linear NaCl
gradient (0–1 M). The active fractions were collected and
dialyzed against 10 mM potassium phosphate buffer
Screening of chlorophenacyl halide-reducing micro-
organisms
Phenacyl halide-reducing activity was widely dis-
tributed in various microorganisms, although the activity
was very weak. Approximately 430 yeast strains were
examined for phenacyl halide-reducing activity. For 3-
chlorophenyl bromide, 63 strains had reducing activity
and 47 strains reduced to the corresponding (R)-form
alcohol. For phenacyl bromide, 21 strains produced the
(R)-form alcohol. The strains producing the alcohol with
a high optical purity are listed in Table 1. The Candida
and Rhodotorula genera had especially high stereo-
selectivities for the desired (R)-isomer. The stereo-
selectivity of the reduction depended on the structure of
chlorophenacyl bromide. Candida blankii CBS2774
reduced 3-chlorophenacyl bromide to the (R)-form
alcohol and phenacyl bromide to the (S)-form alcohol.
Rhodotorula glutinis var. dairenensis IFO415 was
selected as the best enzyme source because it showed
high stereoselectivity for all tested substrates.
(pH 7.5). The active fraction was pooled and used as
the purified enzyme for characterization.
Determination of molecular mass of enzyme. The
molecular mass of the native carbonyl reductase was
estimated by column chromatography using a Superdex
2
00 HR 10/30 column and standard molecular marker
with 50 mM potassium phosphate buffer (pH 7.0) con-
taining 150 mM sodium chloride. The molecular mass of
the subunit was estimated by SDS-polyacrylamide gel
electrophoresis (10%) with an SDS-PAGE marker as the
standard.
N-terminal amino acid sequence analysis. The N-
terminal amino acid sequence was analyzed on a model
ABI492 pulsed liquid protein sequencer equipped with
an on-line phenylthiohydantoin analyzer (Applied Bio-
systems, Foster City, CA, U.S.A.).
Enzymatic reduction of CPN and synthesis of (R)-3-
styrene oxide
For the synthesis of (R)-3-chlorostyrene oxide, opti-
cally active 2-chloro-1-(3-chlorophenyl)ethanol (CPH)
was synthesized by cell-free extract from R. glutinis
IFO415. 3-Chlorophenacyl chloride (CPN) was a better
substrate because it was more stable than 3-chlorophe-
nacyl bromide. For reduction of CPN by this cell-free
extract, NADPH was required as a cofactor. Then
NADPH was regenerated by the addition of glucose and
commercial glucose dehydrogenase. The cell-free ex-
tract accumulated approximately 130 g/l (680 mM) of
(R)-CPH. The turnover number of NADPH for CPH was
about 1,800. (R)-CPH was extracted from the reaction
mixture and purified by distillation in a high yield. This
alcohol was easily converted to (R)-3-chlorostyrene
oxide with sodium hydroxide. In the case of the butyl
acetate-water two-phase reaction, 90 g/l (471 mM) of
(R)-CPH was accumulated.
Reduction of various carbonyl compounds. Each
reaction mixture comprising 2.5 mg of substrate, 14 mg
of NADPH, and the enzyme in 0.5 ml of 100 mM
potassium phosphate buffer (pH 6.5) was stirred at
ꢁ
3
0 C for 20 h. Ethyl acetate (5 ml) was added, and the
reaction mixture was centrifuged. The organic layer was
then analyzed for optical purity of the product.
Analysis. The amounts and the optical purity of the
halohydrins were determined with an HPLC equipped
with a Chiralcel OJ (4:6 mmꢀ ꢀ 250 mm) column
(Daicel Chemicals, Osaka, Japan). The HPLC conditions
included n-hexane/isopropanol (39/1 (v/v)) as a mobile
phase, a flow rate of 1 ml/min, ambient column temper-
ature, and detection at 210 nm. The optical purity of
various halohydrins produced by the enzyme was
determined using an HPLC equipped with the column
(Daicel) described in Table 5. The optical purity of
Purification of the carbonyl reductase
pantolactone, produced from ketopantoyl lactone by the
enzyme, was determined using an HPLC equipped with
a Chiralpak AS (4:6 mmꢀ ꢀ 250 mm) column (Daicel).
And the optical purity of ethyl 4-chloro-3-hydroxybuta-
noate produced from 4-chloro-3-hydroxybutanoate was
The CPN-reducing enzyme was purified from R. glu-
tinis var. dairenensis IFO415 (Table 2). The enzyme
was purified 487-fold to homogeneity with an overall
recovery of 4%. The molecular weight of the enzyme
was found to be 40,000 by gel filtration. The relative

Table 1. Microbial Reduction of Chlorophenacyl Bromides and Phenacyl Bromidea)
Productb)
o-Cl
R1, R2 ¼ H, R3 ¼ Cl, X ¼ Br)
m-Cl
(R1, R3 ¼ H, R2 ¼ Cl, X ¼ Br)
p-Cl
Micoorganism
(
(R2, R3 ¼ H, R1 ¼ Cl, X ¼ Br)
(R1, R2, R3 ¼ H, X ¼ Br)
Yield
% e.e.
Yield
% e.e.
Yield
% e.e.
Yield
% e.e.
Ashbya gossypii IFO560
Brettanomyces custersianus IFO1585
Candida blankii CBS2774
n.t.c)
n.t.
23
23
5
—
—
92
92
99
96
96
91
86
98
99
92
—
98
99
99
98
—
99
92
88
55
10
40
54
31
80
44
n.t.
n.t.
59
63
n.t.
21
79
56
75
51
8
72
99
90
99
99
81
92
—
—
92
98
—
99
99
67
99
99
99
88
—
81
47
11
51
26
8
95
99
73
72
87
68
96
—
—
97
99
—
—
88
93
99
89
—
76
—
84
59
n.t.
49
45
—
(S) 98
—
—
11
99
—
—
—
—
—
—
3
Candida intermedia IFO761
Candida krusei IFO11
Candida magnoliae IFO705
n.t.
n.t.
99
98
32
13
32
15
39
75
n.t.
62
87
15
31
n.t.
7
89
25
n.t.
n.t.
24
9
Candida pini IFO1327
Candida saitoana IFO768
99
n.t.
n.t.
n.t.
n.t.
n.t.
n.t.
60
Candida tropicalis IFO1403
Cryptococcus albidus IFO378
Cryptococcus terreus IFO727
Geotrichum hirtum ATCC56047
Pichia membranefaciens IFO461
Rhodosporidium toruloides IFO871
Rhodotorula glutinis IFO1099
Rhodotorula glutinis var. dairenensis IFO415
Rhodotorula glutinis IFO190
Rhodotorula minuta IFO387
Trigonopsis variabilis IFO671
Trichosporon loubieri var.loubieri CBS7065
Yamadazyma farinosa IFO574
n.t.
n.t.
17
55
67
31
n.t.
42
n.t.
51
66
60
99
99
—
—
—
—
—
n.t.
n.t.
n.t.
n.t.
n.t.
76
n.t.
44
27
34
a)The assay conditions are given in ‘‘Materials and Methods’’.
b)The absolute configilation of the product is (R) form.
c)Not tested.

Carbonyl Reductase for Synthesis of Styrene Oxide
83
Table 2. Purification of Carbonyl Reductase from R. glutinis var. dairenensis IFO415
Total proteina)
mg)
Total activitya)
(units)
Specific activity
(units/mg)
Yield
(%)
Purification
(fold)
Step
(
Cell-free extract
Ammonium sulfate
DEAE-Toyopearl
Phenyl-Toyopearl
Blue-Sepharose
23,500
4,200
840
1,840
775
732
508
66
0.078
0.18
0.87
24
100
42
40
28
4
1
2.30
11.1
308
487
21.2
1.77
38
a)The assay conditions are given in ‘‘Materials and Methods’’.
molecular mass weight of the subunit was estimated to
be about 30,000 on SDS–polyacrylamide gel electro-
phoresis. The optimum temperature of the enzyme was
40–50 C. The optimum pH was 5.0–6.0 (potassium
phosphate).
dithiobis(2-nitrobenzoate), HgCl2, and p-chloromercuri-
benzoate. Quercetin, a nonspecific inhibitor of human
1
2)
brain carbonyl reductase, diphenylhydantoin, a potent
13)
ꢁ
inhibitor of aldehyde dehydrogenase, and 2,4-dinitro-
phenol, an inhibitor of NADPH dehydrogenase (qui-
1
4)
none),
compounds tested, Cu , Fe , and Mg had no effect
were also inhibitory. Among the metallic
2þ
2þ
2þ
N-terminal amino acid sequence analysis
2
þ
þ
Automated Edman degradation of the enzyme protein
with a pulsed liquid phase sequencer showed that the
N-terminal amino acid sequence was Pro–Ala–Ala–
Lys–The–Tyr–Phe–Ile–Ser–Gly–Ala–Ser–Arg–Gly–
Leu–Gly–Leu–Gly–Tyr–Thr–Arg–Glu–Leu–Leu. A
BLAST program search found this sequence to be
significantly similar to the partial amino acid sequences
of five putative dehydrogenases.
on enzyme activity, while Zn and Ag did. Chelating
reagents such as EDTA and 1,10-phenanthroline also
had no effect.
Substrate specificity
As shown in Table 4, a broad range of carbonyl
compounds was used to investigate substrate specificity.
The enzyme catalyzed some aromatic and aliphatic
aldehydes, aliphatic ketones and keto esters, and keto
acids. p-Nitrobenzaldehyde, ketopantoyl lactone (KPL)
and methyl pyruvate, and ethyl 2-chloro-3-oxobutanoate
were good substrates. 4-Acetylpyridine and m-nitro-
acetophenone were reduced while acetophenone and
m-chloroacetophenone were not reduced by the enzyme.
Effects of chemicals
The various compounds listed in Table 3 were added
to the reaction mixture, and then their relative activities
were measured with CPN as substrate. The enzyme was
0
inhibited by several sulfhydryl reagents, viz., 5,5 -
3-Oxoesters were reduced but 2-oxoesters did not serve
as substrates. Moreover, the enzyme reduced KPL and
ethyl 4-chloro-3-oxobutanoate (COBE). Asymmetric
reductions of these compounds are interesting reactions
from the standpoint of industrial applications. NADPH
was almost absolutely required as a cofactor; when
NADPH was replaced with NADH, the reaction rate was
reduced by 93%. Kinetic parameters were measured by
Lineweaver–Burk plots. For NADPH and NADH, Km
and Vmax were 0.056 mM and 70.4 mmol/min/mg, and
Table 3. Effects of Various Chemicals on Enzyme Activitya)
Concentration
mM)
Activity
(%)
Compound
Quercetin
(
1
0
0
0
.1
Diphenylhydantoin
0
0.5
1
32
0
5
,5 -dithiobis(2-nitrobenoate)
0
1
1
1
1
1
1
0
1
1
1
1
0
1
1
1
1
0
.1
89
81
24
70
96
98
0
Iodoacetate
0.54 mM and 15.0 mmol/min/mg, respectively. For CPN,
p-Chloromercuribenzoate
N-Ethylmaleimide
EDTA
Km and Vmax were 0.060 mM and 54.1 mmol/min/mg.
1
2
,10-Phenanthroline
,4-Dinitrophenol
Stereospecificity
The stereospecificities for some useful carbonyl
compounds were examined. The enzyme reduced var-
ious phenacyl halides with high enantioselectivity for
the (R)-isomer (Table 5). Ethyl 4-chloro-3-oxobutanoate
(COBE) was reduced to the useful (R)-form optically
pure active alcohol, which is versatile for the synthesis
.1
.1
.1
7
CaCl2
CuSO4
FeSO4
HgCl2
96
101
114
0
52
96
120
57
0
MgSO4
MnCl2
ZnSO4
AgNO3
15)
of pharmaceutical compounds. On the other hand,
ketopantoyl lactone (KPL) was reduced to (R)-panto-
lactone, the synthetic starting material of D-pantothenic
1
6)
39
acid, with only 52% e.e.
a)Enzyme activity was measured as described in ‘‘Materials and Methods’’,
except for the addition of the indicated compounds.

84
N. KIZAKI et al.
Table 4. Substrate Specificity of the Carbonyl Reductase from R. glutinis var. dairenensisa)
Relative activity
%)c)
Relative activity
(%)
b)
Substrate
Substrate
(
3
-Chlorophenacyl chloride (CPN)
100
Acetone
-Butanone
2-Heptanone
11
10
11
4
2
Propionaldehyde
n-Butylaldehyde
4
20
10
13
2
Diacetyl
n-Hexylaldehyde
Methy glyoxal
Glutaraldehyde
Camphorquinone
Cyclopentanone
p-Chloroacetophenone
4-Acetylpyridine
m-Nitroacetophenone
Propiophenone
33
9
5
o-Nitrobenzaldehyde
m-Nitrobenzaldehyde
p-Nitrobenzaldehyde
Benzaldehyde
31
64
162
25
24
45
11
55
44
35
14
o-chlorobenzaldehyde
m-chlorobenzaldehyde
p-chlorobenzaldehyde
Pyridine-4-aldehyde
Ketopantoyl lactone (KPL)
Methy pyruvate
238
208
12
Ethyl 4-chloro-3-oxobutanoate (COBE)
Octyl 4-chloro-3-oxobutanoate
Ethyl 2-chloro-3-oxobutanoate
72
111
2-Ketobutyric acid
Oxalactic acid
7
24
a)Enzyme activity was measured as described in ‘‘Materials and Methods’’.
b)The substrate concentration was 1 mM.
c)To calculate the relative activity, the activity for CPN was taken as 100%.
The following compounds did not serve as substrates: acetophenone, m-chloroacetophenone, benzylacetone, 2-octanone, ethyl 2-oxopentanoate, methyl 2-
oxodecanoate, methyl propionylacetate, nicotinaldehyde, glyoxal, and 2-oxo-n-valeric acid.
Table 5. Stereoselectivity of the Carbonyl Reductase from R. glutinis var. dairenensisa)
Product
Substrateb)
Analysisc)
Optical purity (% e.e.)
Configulation
Phenacyl chloride
Phenacyl bromide
(R1, R2, R3 ¼ H, X ¼ Cl)
>99
>99
>99
>99
>99
(R)
(R)
(R)
(R)
(R)
OB
OJ
(R1, R2, R3 ¼ H, X ¼ Br)
3-Chlorophenacyl chloride (CPN)
4-Chlorophenacyl chloride
4-Chlorophenacyl bromide
(R1, R3 ¼ H, R2 ¼ Cl, X ¼ Cl)
(R2, R3 ¼ H, R1 ¼ Cl, X ¼ Cl)
(R2, R3 ¼ H, R1 ¼ Cl, X ¼ Br)
OJ
OJ-H
AD
a)The reaction conditions are given in ‘‘Materials and Methods’’.
b)The structures are described in Fig. 1.
c)The HPLC columns used for the determination of optical purity are shown. The assay conditions are given in ’’Materials and Methods’’.
Discussion
R. glutinis var. dairenensis IFO415. The purified en-
zyme readily reduced several aldehydes, such as p-
nitrobenzaldehyde, p-chlorobezaldehyde, and pyridine
aldehyde. These aldehydes are typical substrates for
Optically active styrene oxide derivatives are useful
chiral building blocks for the synthesis of pharmaceut-
icals. In turn, a combination of biological asymmetric
reduction of phenacyl halide and chemical epoxidation
of 2-halo-1-phenylethanol is one of the most practical
and economical strategies for their synthesis.
Several microorganisms having the ability to reduce
phenacyl halide derivatives to optically active 2-halo-1-
phenylethanol derivatives have been reported. But no
phenacyl halide-reducing enzyme have been character-
ized except for phenylacetaldehyde reductase from
Corynebacterium sp.17) This is a tetrameric NADH-
dependent alcohol dehydrogenase (the subunit Mr is
1
2)
aldo-keto reductase family enzymes. Also, this en-
zyme reduced ethyl 4-chloro-3-oxobutanoate (COBE) to
(R)-formed alcohol with a 99% e.e., and ketopantoyl
lactone (KPL) to the (S)-form alcohol with 52% e.e.,
respectively. These optically active alcohols are inter-
1
5,16)
esting compounds from an industrial standpoint.
The N-terminal amino acid sequence of the enzyme
was similar to putative dehydrogenases. Also, the
sequence contained a conservative glycine-rich domain
(Gly–X–X–X–Gly–X–Gly) that is found in most of the
1
8)
short chain dehydrogenases.
This enzyme might
4
0,299) that catalyzes various asymmetric reductions of
belong to the short-chain dehydrogenase/reductase
family judging by its molecular weight, its independence
of metals, its response to certain inhibitors, and the
conservative domain in its N-terminus.
carbonyl compounds. We purified an NADPH-depend-
ent carbonyl reductase that catalyzed asymmetric re-
duction of phenacyl halide to homogeneity from

Carbonyl Reductase for Synthesis of Styrene Oxide
85
Table 6. Comparison of Stereoselectivities of Various Yeast Carbonyl Reductasesa)
3
-Chlorophenacyl chlorideb) Ethyl 4-chloro-3-oxobutanoate Ketopantoyl lactone
Substrate
(
CPN)
(COBE)
(KPL)
Enzyme
Carbonyl reductase
Origin
R. glutinis
c)
c)
(R)
(R)
(S) 52% e.e
(R) 67% e.e.
d)
Menadione reductase
Aldehyde reductase
C. macedoniensis
S. salmonicolor
n.t.
n.r.
(S)
(R)
n.r.
n.r.
e)
n.r.
(R)
(R)
Conjugated polyketone reductase C1 C. parapsilosis
Conjugated polyketone reductase C2 C. parapsilosis
n.r.
n.r.
a)The configulation and optical purity of the product are listed.
b)The assay conditions are given in ‘‘Materials and Methods’’.
c)The analytical methods are given in ‘‘Materials and Methods’’.
d)Not tested.
e)Not reacted.
Several monomeric NADPH-dependent carbonyl re-
ductases have been purified from yeast and character-
ized. The stereoselectivities of several enzymes for
CPN, COBE, and KPL are compared in Table 6.
Aldehyde reductase from Sporobolomyces salmonicolor
reduces COBE to (R)-form alcohol but does not reduce
References
1
2
)
)
Ainsworth, A. T., and Smith, D. G., European Patent
Application 28105 (May 6, 1981).
Pedragosa-Moreau, S., Archelas, A., and Furstoss, R.,
Enantiocomplementary epoxide hydrolyses as a prepa-
rative access to both enantiomers of styrene oxide.
J. Org. Chem., 58, 5533–5536 (1993).
1
9)
CPN or KPL. Two conjugated polyketone reductases
from Candida parapsilosis reduce KPL to (R)-form
3
)
Kutsuki, H., Sawa, I., Hasegawa, J., and Watanabe, K.,
Asymmetric hydrolysis of (dl)-1-acyloxy-2-halo-1-phen-
ylethanes with lipases. Agric. Biol. Chem., 50, 2369–
2373 (1986).
2
0)
alcohol but do not reduce CPN or COBE. Moreover,
carbonyl reductase from R. glutinis and menadione
reductase from Candida macedoniensis reduce COBE
and KPL to the corresponding alcohols with opposite
4
)
)
Tanaka, K., and Yasuda, M., A practical synthesis of
(R)-chlorostyrene oxide starting from 3-chloroethylben-
zene. Tetrahedron: Asymmetry, 9, 3275–3282 (1998).
Hamada, H., Miura, T., Kumobayashi, H., Matsuda, T.,
Harada, T., and Nakamura, K., Asymmetric synthesis of
2
1,22)
stereoselectivities.
The wide variation among these
useful compounds in terms of their reactivities and
stereoselectivities for reductases suggests that the each
suitable reductase is necessary for the preparation of the
desired optically active compound.
5
(
R)-2-chloro-1-(m-chlorophenyl)ethanol using acetone
powder of Geotrichum candidum. Biotechnol. Lett., 23,
603–1606 (2001).
Preparations of optically active 2-halo-1-phenyletha-
nol derivatives using microbial cells have been reported.
For enzymatic reduction, the reduced form of the
coenzyme is needed; therefore, an effective coenzyme
regeneration system is essential. But the enzymes
required for regeneration of coenzymes in microbial
cells are often weak and complicated. Further, these
enzymes might be damaged by phenacyl halide, which is
a potent acylating reagent. Previous studies have there-
fore examined the effects of the addition of an organic
1
6
)
)
Barbieri, C., Bossi, L., D’Arrigo, P., Fantoni, G. P., and
Servi, S., Bioreduction of aromatic ketones: preparation
of chiral benzyl alcohols in both enantiomeric forms.
J. Mol. Cat. B: Enzymatic, 11, 415–421 (2001).
Ema, T., Moriya, H., Kofukuda, T., Ishida, T., Maehara,
K., Utaka, M., and Sakai, T., High enantioselectivity and
broad substrate specificity of a carbonyl reductase:
toward a versatile biocatalyst. J. Org. Chem., 66,
7
8
682–8684 (2001).
8
)
)
De Carvalho, M., Okamoto, M. T., Samenho Moran, P.
J., and Rodrigues, J. A. R., Baker’s yeast reduction of ꢂ-
haloacetophenones. Tetrahedron, 47, 2073–2080 (1991).
Itoh, N., Matsuda, M., Mabuchi, M., Dairi, T., and
Wang, J., Chiral alcohol production by NADH-depend-
ent phenylacetaldehyde reductase coupled with in situ
regeneration of NADH. Eur. J. Biochem., 269, 2394–
solvent,2 surfactant, or absorbing resin to reduce
the damage, but none of these trials has resulted in a
fundamental solution to the problem. Construction of an
effective enzymatic reduction system using a trans-
formant overexpressing both reductase and glucose
3)
24)
6)
9
2
5)
dehydrogenase genes has been reported. The use of
this system for the preparation of optically active styrene
oxide derivatives as explained in the present study has
great promise for industrial applications.
2
402 (2002).
1
1
0) Kataoka, M., Kita, K., Wada, M., Yasohara, Y.,
Hasegawa, J., and Shimizu, S., Novel bioreduction
system for the production of chiral alcohols. Appl.
Microbiol. Biotechnol., 62, 437–445 (2003).
1) Badford, M. M., A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing
the principle of protein-dye binding. Anal. Biochem., 72,
Acknowledgment
The authors are grateful to Dr. K. Kita, Dr. M.
Kataoka, and Dr. S. Shimizu of Kyoto University for
gifts of aldehyde reductase of S. salmonicolor and
menadione reductases of C. parapsilosis.
248–254 (1976).
12) Wermuth, B., Purification and properties of NADPH-
dependent carbonyl reductase from human brain. J. Biol.

86
N. KIZAKI et al.
Chem., 256, 1206–1213 (1981).
3) Ris, M. M., and von Wartburg, J.-P., Heterogeneity of
20) Hidalgo, A.-R. G. D., Akond, M. A., Kita, K., Kataoka,
1
M., and Shimizu, S., Isolation and primary structural
analysis of two conjugated polyketone reductases from
Candida parapsilosis. Biosci. Biotechnol. Biochem., 65,
2785–2788 (2001).
NADPH-dependent aldehyde reductase from human and
rat brain. Eur. J. Biochem., 37, 69–77 (1973).
4) Kori, K. K., Yearby, C., Scott, W., and Donaldson, O.
K., Purification and properties of three separate mena-
dione reductases from hog liver. J. Biol. Chem., 244,
1
21) Kataoka, M., Doi, Y., Sim, T.-S., Shimizu, S., and
Yamada, H., A novel NADPH-dependent carbonyl
reductase of Candida macedoniensis: purification and
characterization. Arch. Biochem. Biophys., 294, 469–474
(1992).
22) Kataoka, M., Shimizu, S., and Yamada, H., Carbonyl
reductases: their functions and application to the pro-
duction of optically active compounds. Hakkokougaku
(in Japanese), 70, 479–493 (1992).
23) Liu, X., Zhu, T.-S., Sun, P.-D., and Xu, J.-H., Asym-
metric reduction of aromatic ketones by the baker’s yeast
in organic solvent systems. Synth. Commun., 31, 1521–
1526 (2001).
24) Goswami, A., Bezbaruah, L. R., Borthakur, N., Dey, D.,
and Hazarika, K. A., Microbial reduction of !-bromo-
acetophenones in the presence of surfactants. Tetrahe-
dron: Asymmetry, 11, 3701–3709 (2000).
621–629 (1969).
1
1
1
5) Zhou, B., Gopalan, A. S., van Middlesworth, F., Shieh,
W. R., and Sih, C. J., Stereochemical control of yeast
reduction. 1. Asymmetric synthesis of L-carnitine. J. Am.
Chem. Soc., 105, 5925–5926 (1983).
6) Ogawa, J., and Shimizu, S., Industrial microbial en-
zymes: their discovery by screening and use in large-
scale production of useful chemicals in Japan. Curr.
Opin. Biochem., 13, 367–375 (2002).
7) Itoh, N., Morikawa, R., Wang, J., Okada, K., and
Mizuguchi, N., Purification and characterization of
phenylacetaldehyde reductase from a styrene-assimilat-
ing Corynebacterium strain, ST-10. Appl. Environ.
Microbiol., 63, 3783–3788 (1997).
8) Hummel, W., New alcohol dehydrogenases for the
synthesis of chiral compounds. Adv. Biochem. Eng./
Biotechnol., 58, 145–184 (1997).
1
25) Kizaki, N., Yasohara, Y., Hasegawa, J., Wada, M.,
Kataoka, M., and Shimizu, S., Synthesis of optically
pure ethyl (S)-4-chloro-3-hydroxybutanoate by Esche-
richia coli transformant cells coexpressing carbonyl
reductase and glucose dehydrogenase. Appl. Microbiol.
Biotechnol., 55, 590–595 (2001).
1
9) Kataoka, M., Sakai, H., Morikawa, T., Katoh, M.,
Miyoshi, T., Shimizu, S., and Yamada, H., Character-
ization of aldehyde reductase of Sporobolomyces sal-
monicolor. Biochim. Biophys. Acta, 1122, 57–62 (1992).