Purification and characterization of a (r)-1-phenyl-1, 3-propanediol- producing enzyme from trichosporon fermentans aj-5152 and enzymatic (r)-1-phenyl-1, 3-propanediol production

Article abstract of DOI:10.1271/bbb.90159

An (R)-1-phenyl-1, 3-propanediol-producing enzyme was purified from Trichosporon fermentans AJ-5152. It was NADPH-dependent and converted 3-hydroxy-1- phenylpropane-1-one (HPPO) to (R)-1-phenyl-1, 3-propanediol [(R)-PPD] with anti-Prelog's specificity. It

Full text of DOI:10.1271/bbb.90159

Biosci. Biotechnol. Biochem., 73 (7), 1640–1646, 2009
Purification and Characterization of a (R)-1-Phenyl-1,3-propanediol-producing
Enzyme from Trichosporon fermentans AJ-5152
and Enzymatic (R)-1-Phenyl-1,3-propanediol Production
y
Ikuo KIRA and Norimasa ONISHI
AminoScience Laboratories, Ajinomoto Co., Inc., 1-1 Suzuki-cho, Kawasaki-ku, Kawasaki 210-8681, Japan
Received February 27, 2009; Accepted April 13, 2009; Online Publication, July 7, 2009
[doi:10.1271/bbb.90159]
An (R)-1-phenyl-1,3-propanediol-producing enzyme
was purified from Trichosporon fermentans AJ-5152. It
was NADPH-dependent and converted 3-hydroxy-1-
phenylpropane-1-one (HPPO) to (R)-1-phenyl-1,3-pro-
panediol [(R)-PPD] with anti-Prelog’s specificity. It
showed maximum activity at pH 7.0 and 40 ꢀC. Its Km
and Vmax values toward HPPO were 20.1 mM and
3.4 ꢀmol minÀ1 mg proteinÀ1 respectively. The relative
molecular weight of the enzyme was estimated to be
68,000 on gel filtration and 32,000 on SDS-polyacryl-
amide gel electrophoresis. An (R)-PPD-producing reac-
tion using the (R)-PPD-producing enzyme and an
NADPH recycling system was carried out by successive
feeding of HPPO. A total (R)-PPD yield of 8.9 g/l was
produced in 16 h. The molar yield was 76%, and the
optical purity of the (R)-PPD produced was over
99% e.e.
The (S)-PPD-producing enzyme from W. saturnus var.
mrakii AJ-5620 also follows Prelog’s rule, as do other
stereospecific oxidoreductases.5–8) In contrast, some
enzymes that show anti-Prelog stereospecificity during
the formation of (R)-alcohols have been reported,9–11) but
(R)-specific oxidoreductases are still limited and are not
sufficient for the production of (R)-alcohols. Therefore,
the discovery of oxidoreductases with unusual stereo-
specificity would be valuable in the synthesis of both
enantiomers of useful chiral alcohols.
Here, the purification and characterization of an (R)-
PPD-producing enzyme from T. fermentans AJ-5152
and (R)-PPD production by an (R)-PPD-producing
enzyme coupled with glucose dehydrogenase NADPH-
regeneration is reported.
Materials and Methods
Key words: (R)-1-phenyl-1,3-propanediol; 3-hydroxy-
1-phenylpropane-1-one; Trichosporon fer-
mentans; stereospecific oxidoreductase;
chiral alcohol
Materials. Both (S)-and (R)-PPD were purchased from the Chisso
Corporation (Tokyo, Japan). HPPO was prepared as described
previously.2) The derivatives of HPPO were prepared as described
below.
Preparation of 3-hydroxy-1-(40-methylphenyl) propane-1-one, 3-
hydroxy-1-(40-chlorophenyl) propane-1-one, and 3-hydroxy-2-methyl-
1-phenylpropane-1-one.
Recent investigations have clearly established that
individual enantiomers contained in racemic mixtures of
pharmaceutical drugs can have different pharmacoki-
netic and bioavailability profiles. The manufacturing of
the active forms of drugs is consequently becoming the
norm in the industry.1) Chiral alcohols are very
important precursors for a large number of pharmaceut-
icals, and (S)-1-phenyl-1,3-propanediol [(S)-PPD] is an
important intermediate in the synthesis of therapeutic
agents such as the serotonin-uptake inhibitor (S)-fluox-
etine. By screening the microbial production of (S)-PPD
from 3-hydroxy-1-phenylpropane-1-one (HPPO), we
found that Williopsis saturnus var. mrakii AJ-5620
produced (S)-PPD and Trichosporon fermentans AJ-
5152 produced (R)-PPD with high enantioselectivity
(Fig. 1). Microbial production of (S)-PPD from HPPO
by intact cells of W. saturnus var. mrakii AJ-5620 and
characterization of the (S)-PPD-producing enzyme have
been reported.2,3)
To
a
solution of 4-(40-methylphenyl)-1,3-dioxane (1.00 g,
5.61 mmol) in 6 ml of CH2Cl2 and 4 ml of water, a solution of 4-(40-
chlorophenyl)-1,3-dioxane (329 mg, 1.66 mmol) in 1.5 ml of CH2Cl2
and 1 ml of water or a solution of 5-methyl-4-phenyl-1,3-dioxane
(509 mg, 2.86 mmol) in 3 ml of CH2Cl2 and 2 ml of water, 6.21, 1.94 or
3.11 mmol Br2 were added and this was stirred for 2.5 h at 4 ꢀC. After
the addition of 10 ml of saturated Na2SO3 solution, each mixture was
extracted with CH2Cl2 (10 ml ꢁ 2). The combined organic layer was
washed with saturated NaCl solution, dried over anhydrous Na2SO4,
and concentrated at reduced pressure. Each residue was purified by
silica-gel column chromatography to obtain the required compound.
i) 3-Hydroxy-1-(40-methylphenyl) propane-1-one (0.46 g, 2.80
mmol, 49.9%): 1H-NMR (CDCl3) ꢀ 2.42 (3H, s, ArCH3), 2.70 (1H,
t, J ¼ 6:6 Hz, CH2–OH), 3.21 (2H, t, J ¼ 5:3 Hz, CO–CH2–CH2), 4.03
(2H, dt, J ¼ 5:3 Hz, 6.6 Hz, CH2–CH2–OH), 7.2–7.3 (2H, m, ArCH2),
7.8–7.9 (2H, m, ArCH2).
ii) 3-Hydroxy-1-(40-chlorophenyl) propane-1-one (225 mg, 1.22
mmol, 73.4%): 1H-NMR (CDCl3) ꢀ 2.55 (1H, t, J ¼ 6:6 Hz, CH2–OH),
3.20 (2H, t, J ¼ 5:3 Hz, CO–CH2–CH2), 4.03 (2H, dt, J ¼ 5:3 Hz,
6.6 Hz, CH2–CH2–OH), 7.4–7.5 (2H, m, ArCH2), 7.8–7.9 (2H, m,
ArCH2).
Also, a number of stereospecific oxidoreductases have
been purified and reported. Of these oxidoreductases,
most follow Prelog’s rule, that hydride is transferred
from the cofactor to the re face of the carbonyl to give
an (S)-alcohol.4)
iii) 3-Hydroxy-2-methyl-1-phenylpropane-1-one (239 mg, 1.46
mmol, 51.0%): 1H-NMR (CDCl3) ꢀ 1.25 (3H, d, J ¼ 7:3 Hz, CH–
CH3), 2.32 (1H, t, –OH), 3.68 (1H ddq, J ¼ 4:2 Hz, 7.3 Hz, 7.3 Hz,
CO–CH–CH3), 3.8–3.9 (2H, m, CH–CH2–OH), 7.4–7.5 (2H, m,
ArH2), 7.5–7.6 (1H, m, ArH), 7.9–8.0 (2H, m, ArH2).
y
To whom correspondence should be addressed. Fax: +81-59-346-0140; E-mail: ikuo kira@ajinomoto.com

Purification of an (R)-1-Phenyl-1,3-propanediol-producing Enzyme
1641
(molecular weight): bovine serum albumin (67,000), lipase from
Candida cylindracea (53,000), ovalbumin (43,000), and chymotrypsi-
nogen (25,000).
OH
T. fermentans AJ-5152
OH
OH
O
Enzyme purification. All purification procedures were performed at
0–8 ꢀC.
(R)-PPD
OH
Preparation of cell-free extracts. The washed cells (800 g wet
weight) from 8 liters of culture broth were suspended in 1.6 liters of
50 mM potassium phosphate buffer (pH 7.0). To the cell suspension, an
equal volume of glass beads (0.45 mm diameter) was added, and the
cells were disrupted using a BeadBeater (Biospec, Bartlesville, OK) for
5 min. The disrupted cell suspension was decanted and centrifuged at
12;000 g for 20 min. To 2 liters of the obtained supernatant, 348 mg of
phenylmethanesulfonyl fluoride (PMSF) were added, and this was
stirred for 30 min at 4 ꢀC.
OH
HPPO
W. saturnus var. mrakii
AJ-5620
(S)-PPD
Fig. 1. Stereospecific Reduction of HPPO to (S)- and (R)-PPD.
Ammonium sulfate fractionation. The cell-free extracts were
fractionated with ammonium sulfate. Proteins precipitating between
60 and 80% saturation were collected by centrifugation (12;000 ꢁ g,
20 min), dissolved in 5 ml of 50 mM potassium phosphate buffer
(pH 7.5), and dialyzed against the same buffer.
Glucose dehydrogenase (GDH; Bacillus sp.) was purchased from
Amano Enzyme. (Nagoya, Japan). All other chemicals used were
commercial products.
Microorganism and cultivation conditions. Trichosporon fermen-
tans AJ-5152, which was kept at Ajinomoto, was used as the enzyme
source. The medium (pH 7.0) comprised 2% glucose, 0.5% (NH4)2SO4,
DEAE-Toyopearl column chromatography. The dialyzed enzyme
solution was applied to a DEAE-toyopearl 650M column (22 mm ꢁ
160 mm; Tosoh, Tokyo) equilibrated with 50 mM potassium phosphate
buffer (pH 7.5). The enzyme was eluted with 200 ml of the buffer, and
the active fractions were collected.
.
0.1% KH2PO4, 0.3% K2HPO4, 0.05% MgSO4 7H2O, 0.001%
.
.
FeSO4 7H2O, 0.001% MnSO4 4H2O, 1% yeast extract, and 1%
peptone. The strain was cultured in a 500-ml Sakaguchi flask containing
50 ml of the medium at 30 ꢀC for 24 h with reciprocal shaking.
Butyl-Toyopearl column chromatography. Ammonium sulfate was
added (30% saturation) to the active fraction. The fraction was applied
to a Butyl-Toyopearl column (22 mm ꢁ 160 mm, Tosoh) equilibrated
with 50 mM potassium phosphate buffer (pH 6.5) containing ammo-
nium sulfate (30% saturation). The column was washed with 150 ml of
buffer, and the enzyme was eluted with a 450-ml linear gradient of
ammonium sulfate (30 to 0% saturation) in the buffer. The active
fractions were collected and dialyzed against 25 mM potassium
phosphate buffer (pH 5.5) containing 10 mM MgSO4.
Enzymatic preparation and analysis of PPD. The enzymatic
reduction of HPPO was performed as follows: A reaction mixture
(1 ml) comprised of 100 mM potassium phosphate buffer (pH 7.0),
20 mM HPPO, 10 mM NADPH, and 0.1 units of the enzyme was
incubated at 30 ꢀC. After incubation for 60 min, 900 ml of ethyl alcohol
was added to 100 ml of the reaction mixture, and the supernatant,
obtained by centrifugation (12;000 g, 10 min), was analyzed. The
optical purity of PPD was measured as described previously.2)
Enzyme activity assay. HPPO-reducing activity was spectrophoto-
metrically determined using HPPO and NADPH as substrate and
cofactor respectively. The standard assay mixture comprised 100 mM
potassium phosphate buffer (pH 7.0), 10 mM HPPO, 0.32 mM NADPH,
and an appropriate amount of enzyme solution. The decrease in
absorbance at 340 nm was monitored at 30 ꢀC. One unit of the enzyme
was defined as the amount required to catalyze the oxidation of 1 mmol
of NADPH per min.
Red-Toyopearl column chromatography. The resulting solution was
applied to a Red-Toyopearl column (14 mm ꢁ 70 mm, Tosoh) equili-
brated with 25 mM potassium phosphate buffer (pH 5.5) containing
10 mM MgSO4. The column was washed with 50 ml of buffer, and the
enzyme was eluted with 90 ml of buffer containing 0.1 M NaCl. The
active fractions were collected and dialyzed against 25 mM potassium
phosphate buffer (pH 5.5). The dialyzed solution was used as the
purified enzyme for characterization.
GDH activity was assayed spectrophotometrically at 340 nm and
30 ꢀC. The standard assay mixture comprised 100 mM potassium
phosphate buffer (pH 7.0), 100 mM glucose, 18 mM NADPþ, and an
appropriate amount of enzyme solution. One unit of the enzyme was
defined as the amount required to catalyze the reduction of 1 mmol of
NADPþ per min.
Results
Purification of the enzyme
Since the enzyme activity in the cell-free extract
decreased during preservation, we examined some
compounds that stabilize enzyme activity. Among the
compounds tested, PMSF kept the enzyme activity from
decreasing during preservation. Hence PMSF was added
to the cell-free extract.
Protein determination. Protein concentrations were measured with a
protein assay kit (Bio-Rad Laboratories, Hercules, CA) using bovine
serum albumin as the standard.12)
SDS–PAGE analysis. SDS–PAGE was performed in a 10–20%
gradient acrylamide slab gel (Daiichi Pure Chemicals, Tokyo, Japan),
as described by Laemmli.13) The following proteins were used as
marker proteins (molecular weight): phosphorylase b (Mr ¼ 94; 000),
bovine serum albumin (67,000), ovalbumin (43,000), carbonic anhy-
drase (30,000), trypsin inhibitor (20,000), and ꢁ-lactalbumin (14,400)
(Daiichi Pure Chemicals). The gel was stained with Coomassie
Brilliant Blue R-250 for protein.
The purification of the enzyme is summarized in
Table 1. The enzyme was purified 1,100-fold to homo-
geneity with an overall recovery of 16%. Red-Toyopearl
chromatography, an affinity chromatography method
deigned for NADPH-dependent enzymes, was effective
at purifying the enzyme. The purified enzyme gave
a single band on SDS–PAGE (Fig. 2). As shown in
Table 1, the optical purity of the (R)-PPD prepared from
the cell-free extracts from T. fermentans was 94% e.e.
On the other hand, the optical purity of the (R)-PPD pre-
pared by the enzyme solution fractionated by (NH4)2SO4
was more than 99% e.e. This suggests that both (S)- and
(R)-PPD producing enzymes exist in cell free-extracts
Determination of molecular weight. The molecular weight of
the purified enzyme was estimated by gel filtration on a Superose
12 HR 10/30 column (10 mm ꢁ 300 mm; GE Healthcare UK,
Buckinghamshire, UK) at a flow rate of 0.4 ml/min with an elution
buffer consisting of 50 mM Tris–HCl buffer (pH 7.2) containing 0.3 M
NaCl. A calibration curve was produced using the following proteins

1642
I. KIRA and N. ONISHI
Table 1. Purification of the (R)-PPD-Producing Enzyme from T. fermentans AJ-5152
Total activity
(units)
Total protein
(mg)
Specific activity
(units/mg)
Yield
(%)
Purification
fold
Stereospecificity
e.e. of (R)-PPD produced
Step
Cell-free extract
1.13
0.85
0.75
0.43
0.18
2717
612
300
35.0
0.4
0.0004
0.0014
0.0025
0.0126
0.458
100
75
1
3.5
94
>99
>99
>99
>99
Ammonium sulphate (30–60%)
DEAE-Toyopearl
Butyl-Toyopearl
66
39
6.0
30.3
1100
Red-Toyopearl
16
In addition, the pH dependence of the oxidation of
(R)-PPD was examined. The optimum pH was different
from that of the reduction of HPPO, and the enzyme
showed maximum activity between pH 9 and 10, similar
to the carbonyl reductase from Candida parapsilosis5)
and the alcohol dehydrogenase from Leifsonia sp.10)
A B
Effects of pH and temperature on the stability of the
enzyme
94,000
The enzyme was incubated for 60 min at 4 ꢀC in
several buffers at various pH values, and the remaining
activity was measured. As shown in Fig. 4A, the
enzyme was stable from pH 5.5 to pH 8.5. It was
inactivated rapidly below pH 5.0.
67,000
43,000
The enzyme was incubated for 60 min at various
temperatures in potassium phosphate buffer (pH 7.0),
and the remaining activity was measured. As shown in
Fig. 4B, when the enzyme was incubated at 40 ꢀC in
potassium phosphate buffer (pH 7.0) for 60 min, more
than 80% of the activity remained. About 50% of the
activity remained at 50 ꢀC, and most of the activity was
lost at 60 ꢀC.
30,000
20,100
14,400
The effects of various compounds on enzyme activity
As shown in Table 2, enzyme activity was measured
after incubation of the enzyme at 4 ꢀC for 15 min with
1 mM of various compounds. Sulfhydryl reagents, such as
p-chloromercuribenzoic acid, and chelating reagents,
such as ꢁ,ꢁ0-dipyridyl and o-phenanthroline, inhibited
the enzyme activity. p-chloromercuribenzoic acid de-
creased the enzyme activity to 67%, and ꢁ,ꢁ0-dipyridyl
and o-phenanthroline decreased it to 69% and 76%
respectively. On the other hand, the enzyme was not
inhibited by EDTA. Judging from the inhibition by
chelating reagents, this (R)-PPD-producing enzyme
might require a metal ion for activity. Among the
metallic compounds tested, HgCl2 inhibited the enzyme,
but FeSO4, MnSO4, ZnSO4, CuSO4, CoSO4, and NiSO4
did not.
Fig. 2. SDS–PAGE of the (R)-PPD-Producing Enzyme from Tricho-
sporon fermentans AJ-5152.
A, molecular weight standard. B, purified enzyme. The conditions
for SDS–PAGE are given in ‘‘Materials and Methods.’’
from T. fermentans, as is the case for aldehyde reduc-
tases from Sporobolomyces salmonicolor AKU4429.14)
Sequencing of the N-terminal amino acids of the (R)-
PPD-producing enzyme was attempted by the automated
Edman degradation method, but was not successful,
suggesting that the N terminus of the (R)-PPD-produc-
ing enzyme is blocked.
Molecular weight and subunit size
The molecular weight of the purified enzyme was
estimated to be 68,000 by gel filtration on Superose 12
HR, and the subunit molecular weight was estimated to
be about 32,000 by SDS–PAGE (Fig. 2). These results
suggest that the enzyme is composed of two identical
subunits.
Substrate specificity and catalytic properties
The substrate specificity of the enzyme is shown in
Table 3. Besides HPPO, the (R)-PPD-producing enzyme
reduced HPPO derivatives. The enzyme showed higher
activity toward 3-hydroxy-1-(40-chlorophenyl) propane-
1-one than toward HPPO. On the other hand, enzyme
activities toward 3-hydroxy-1-(40-methylphenyl) pro-
pane-1-one and 3-hydroxy-2-methyl-1-phenylpropane-
1-one were 35% and 7% of the activity toward HPPO
respectively. The Km and Vmax values toward HPPO,
calculated from Lineweaver-Burk plots, were 20.1 mM
and 3.4 mmol minꢂ1 mg proteinꢂ1 respectively. The en-
zyme was highly specific for NADPH as a coenzyme;
the Km value toward NADPH was 190 mM, and no
Effects of pH and temperature on enzyme activity
Enzyme activity was measured at various pH from 4.5
to 10.0 in several buffers. As shown in Fig. 3A, the
enzyme showed maximum activity at pH 7.0 in potas-
sium phosphate buffer. Enzyme activity was measured at
various temperatures from 30 ꢀC to 60 ꢀC. As shown in
Fig. 3B, the enzyme showed maximum activity at 40 ꢀC.

Purification of an (R)-1-Phenyl-1,3-propanediol-producing Enzyme
1643
A
B
100
80
60
40
20
0
100
80
60
40
20
0
30
40
50
60
4
5
6
7
8
9
10
pH
Temperature (°C)
Fig. 3. The Effects of pH (A) and Temperature (B) on Enzyme Activity.
A, Enzyme activity was measured at various pH values with 100 mM sodium acetate buffer (pH 4.5–6.0; ), potassium phosphate buffer
(pH 6.0–7.5; ), Tris–HCl buffer (pH 7.5–8.5; ), and glycine–NaOH buffer (pH 8.5–10.0; ). B, Enzyme activity was measured at various
temperatures in potassium phosphate buffer (pH 7.0).
A
B
100
80
60
40
20
0
100
80
60
40
20
0
4
5
6
7
8
9
10
0
10 20 30 40 50 60
pH
Temperature. (°C)
Fig. 4. The Effects of pH (A) and Temperature (B) on Enzyme Stability.
A, The enzyme was incubated for 60 min at 4 ꢀC in several buffers at various pH values with 100 mM sodium acetate buffer (pH 4.5–6.0; ),
potassium phosphate buffer (pH 6.0–7.5; ), Tris–HCl buffer (pH 7.5–8.5; ), and glycine–NaOH buffer (pH 8.5–10.0; ), and the remaining
activity was measured. B, The enzyme was incubated for 60 min at various temperatures in potassium phosphate buffer (pH 7.0), and the
remaining activity was measured.
Table 2. Effects of Various Compounds on the Activity of the (R)-
PPD-Producing Enzyme from T. fermentans AJ-5152
Table 3. The Substrate Specificity of the (R)-PPD-Producing
Enzyme from T. fermentans AJ-5152
Compound
Relative activityꢃ (%)
Substrate (10 mM)
Relative activity (%)
None
100
101
100
102
93
3-Hydroxy-1-phenylpropane-1-one (HPPO)
3-Hydroxy-1-(40-methylphenyl) propane-1-one
3-Hydroxy-1-(40-chlorophenyl) propane-1-one
3-Hydroxy-1-phenyl-2-methylpropane-1-one
Benzoylacetic acid ethyl ester
Ethyl 3-oxobutanoate
100
35
121
7
FeSO4
MnSO4
ZnSO4
CuSO4
CoSO4
NiSO4
HgCl2
88
75
32
1
99
100
47
Ethyl 4-chloro-3-oxobutanoate
Benzaldehyde
Pyridine-3-aldehyde
Iodoacetate
97
138
93
53
1
N-Ethylmaleimide
p-chloromercuribenzoic acid
Phenylmethane sulfonyl fluoride
EDTA
102
67
Pyridine-4-aldehyde
D-Glyceraldehyde
Hydroxyacetone
100
100
69
Dihydroxyacetone
Diacetyl
Glyoxal
3
ꢁ,ꢁ0-Dipyridyl
870
23
o-Phenanthroline
76
Enzyme activity was measured after the enzyme had been incubated at 4 ꢀC
for 15 min with various compounds in 100 mM potassium phosphate buffer
(pH 7.0). All compounds were tested at a final concentration of 1 mM.
To calculate relative activity, the activity produced with 10 mM HPPO was
taken to be 100%.
decrease was observed at 340 nm due to the reduction of
HPPO when NADPH was replaced with an equimolar
concentration of NADH.
The enzyme had broad substrate specificity and
catalyzed the reduction of ꢂ-ketoesters such as benzoyl-
acetic acid ethyl ester, ethyl 3-oxobutanoate, and ethyl
4-chloro-3-oxobutanoate. It also reduced pyridine-3-
aldehyde, a typical substrate for microbial aldehyde
reductases.9)
The enzyme showed its highest activity toward
diacetyl, which is a good substrate for the (S)-PPD-
producing enzyme from W. saturnus var. mrakii3) and
the (R)-N-benzyl 3-pyrrolidinol-producing enzyme from
D. riboflavina.14)

1644
I. KIRA and N. ONISHI
The effects of HPPO and (R)-PPD on the activity and
stability of GDH
Since the specific activity of the (R)-PPD-producing
enzyme for (R)-PPD production was low in intact cells
of T. fermentans, we attempted (R)-PPD production
using a coenzyme recycling system.
As the (R)-PPD-producing enzyme was dependent on
NADPH, GDH was used as the enzyme for NADPH
recycling (Fig. 5). The effects of HPPO and (R)-PPD
on the activity and stability of GDH were examined.
As shown in Fig. 6A, GDH activity decreased with
increasing HPPO concentrations. On the other hand,
the enzyme activity was inhibited by 5% at 120 mM
(R)-PPD.
On incubation with 20 mM HPPO for 60 min at 4 ꢀC,
22% of the initial activity of GDH was lost, and 90% of
the initial activity of GDH was lost on incubation with
120 mM HPPO (Fig. 6B), whereas no inactivation of
GDH was observed on incubation with 120 mM (R)-PPD
for 60 min at 4 ꢀC.
The effects of coenzyme concentration
Reaction efficiency was examined at different con-
centrations (0.2 mM, 0.4 mM, and 1.0 mM) of NADPH.
The reaction mixture consisted of 20 mM HPPO, 30 mM
glucose, 0.2 U/ml (R)-PPD-producing enzyme, and
0.2 U/ml GDH. The reaction was performed in 4 ml of
100 mM potassium phosphate buffer (pH 7.0) at 25 ꢀC,
and the pH was adjusted to 7.0 with 2 N KOH.
The reaction time decreased with increases in
NADPH concentration. When 1.0 mM and 0.4 mM
NADPH were used, the reaction was completed in 5 h
and 6 h respectively. On the other hand, when only
0.2 mM NADPH was used, the reaction proceeded
efficiently and 18.3 mM (R)-PPD was produced after a
reaction time of 7 h. Hence we used 0.2 mM NADPH in
the reaction mixture for (R)-PPD production.
(R)-PPD production with an NADPH recycling
system
Production of (R)-PPD from HPPO was examined
using an NADPH recycling system. As shown in Fig. 6,
since a high concentration of HPPO inactivates GDH,
the reaction was performed at pH 7.0 and 25 ꢀC for 16 h,
and the HPPO concentration was kept below 20 mM by
successive feeding of HPPO. The time course of (R)-
PPD production is shown in Fig. 7. A total of 59 mM
(8.9 g/l) (R)-PPD was produced after a reaction time of
16 h. The molar yield was 76%, and the optical purity of
the (R)-PPD produced was over 99% e.e. The calculated
turnover of NADPþ, based on the amount of NADPþ
added and (R)-PPD produced, was about 300.
The effects of HPPO and (R)-PPD on the stability of
the (R)-PPD-producing enzyme
The effects of HPPO and (R)-PPD on the stability of
the (R)-PPD-producing enzyme were examined. The
(R)-PPD-producing enzyme was stable toward HPPO
and (R)-PPD, in contrast to GDH. On duplicate
incubation with 120 mM HPPO and (R)-PPD for
60 min at 4 ꢀC, 84% and 86% of the initial activity of
the (R)-PPD-producing enzyme remained.
Discussion
(R)-PPD-producing enzyme
OH
O
A number of NADPH dependent stereospecific
oxidoreductases have been used for the production of
chiral alcohols after being purified from various micro-
organisms. These oxidoreductases can be divided into
superfamilies such as the aldo-keto reductase family and
the alcohol dehydrogenase (ADH) superfamily: Group I
(zinc-dependent long chain ADH), Group II (zinc-
independent short chain ADH), and Group III (iron-
activated ADH), based on their molecular structure,
molecular size, and requirement for metal cofactors
for activity.15,16) Although the N-terminal amino acid
OH
OH
NADP+
NADPH
glucono-1,5-lactone
glucose
GDH
Fig. 5. (R)-PPD Production by an (R)-PPD-Producing Enzyme in
Conjunction with an NADPH Recycling System.
A
B
100
80
60
40
20
0
100
80
60
40
20
0
0
20 40 60 80 100 120
0
20 40 60 80 100 120
HPPO or (R)-PPD (mM)
HPPO or (R)-PPD (mM)
Fig. 6. The Effects of HPPO and (R)-PPD on GDH.
A, The effects of HPPO and (R)-PPD on the Activity of GDH. Enzyme activity was measured as described in the text, except for the addition
of HPPO or (R)-PPD. HPPO: ; (R)-PPD: . B, The Effects of HPPO and (R)-PPD on the Stability of GDH. GDH was incubated with the
indicated amounts of HPPO or (R)-PPD in 100 mM potassium phosphate buffer (pH 7.0) at 4 ꢀC for 60 min, and then the remaining activity of
GDH was measured. HPPO: ; (R)-PPD:
.

Purification of an (R)-1-Phenyl-1,3-propanediol-producing Enzyme
1645
and the two enzymes are dissimilar in their optimum pH
and optimum temperature. Both enzymes show similar
substrate specificities; they can reduce diacetyl, pyri-
dine-3-aldehyde, pyridine-4-aldehyde, and ethyl acetoa-
cetate. However, the T. fermentans enzyme showed
almost identical activity toward pyridine-3-aldehyde,
pyridine-4-aldehyde, and HPPO, while the W. saturnus
enzyme showed 8 times and 6 times higher activity
toward pyridine-3-aldehyde and pyridine-4-aldehyde
than toward HPPO.
The T. fermentans enzyme showed its highest activity
toward diacetyl among the substrates tested, as did the
W. saturnus enzyme. Compared with diacetyl reductase
from Staphylococcus aureus,17) both the (R)-PPD-pro-
ducing enzyme and the diacetyl reductase catalyzed the
reverse reaction (oxidation reaction). On the other hand,
the subunit structure and the specificity toward the
coenzyme of the (R)-PPD-producing enzyme are differ-
ent from those of the diacetyl reductase from S. aureus.
The diacetyl reductase from S. aureus is a monomer and
is specific for NADH.
60
50
40
30
20
10
0
0
4
8
12
16
Reaction time (h)
Fig. 7. The Time Course of (R)-PPD Production.
The reaction mixture consisted of 180 mM glucose, 0.2 mM
NADPþ, 0.5 U/ml (R)-PPD-producing enzyme, and 0.5 U/ml
GDH. The reaction was performed in 50 ml of 100 mM potassium
phosphate buffer (pH 7.0) at 25 ꢀC. The pH was adjusted to 7.0 with
2 N KOH. The HPPO concentration was kept at 20 mM, and 25 U
GDH was added at 6 h, 12 h, and 18 h. HPPO: ; (R)-PPD:
.
The (R)-N-benzyl 3-pyrrolidinol-producing enzyme
from D. riboflavin also shows high activity toward
diacetyl, pyridine-3-aldehyde, and pyridine-4-aldehyde,
and is a dimer enzyme.14) However, the requirement for
a coenzyme and sensitivity to metal ions of the (R)-PPD-
producing enzyme are different from those of the
D. riboflavin enzyme. The D. riboflavin enzyme is
Table 4. Properties of the (R)-PPD-Producing Enzyme from T.
fermentans AJ-5152 and the (S)-PPD-Producing Enzyme from
Williopsis saturnus var. mrakii AJ-56203)
(R)-PPD
-Producing enzyme
(S)-PPD
-Producing enzyme
Molecular weight
(Gel filtration)
68,000
32,000
58,000
27,000
NADH-dependent and strongly inhibited by Fe2þ
.
Compared with the short-chain alcohol dehydrogen-
ase from Candida parapsilosis,11) which shows anti-
Prelog selectivity, the requirement of NADPH as a
coenzyme and the optimum temperature and temper-
ature stability of the (R)-PPD-producing enzyme are
similar. On the other hand, the subunit structure,
inhibition by metal ions, and oxidative activity toward
alcohol of (R)-PPD-producing enzyme are different
from those of the enzyme from C. parapsilosis. The
C. parapsilosis enzyme is a monomer and is inhibited
by Cu2þ, whereas the (R)-PPD-producing enzyme is a
dimer and is not inhibited by Cu2þ. In addition, the
C. parapsilosis enzyme showed no oxidative activity
toward 1-phenyl-1,2-ethanediol, whereas the (R)-PPD-
producing enzyme catalyzed the oxidation of (R)-PPD.
This is the first report regarding enzymatic synthesis of
(R)-PPD using an NADPH recycling system. Recently,
an enzymatic production process for (R)-4-chloro-3-
hydroxybutanoate using recombinant E. coli coexpress-
ing a 4-chloro-3-oxobutanoate-reducing enzyme and
glucose dehydrogenase was developed.18) Compared
with the accumulation of the product described above,
the current (R)-PPD accumulation was low. However, it
is thought that the accumulation of (R)-PPD can be
improved by optimization of the reaction conditions.
In order to establish an industrial (R)-PPD production
process, further studies are in progress to identify the
gene that encodes the (R)-PPD-producing enzyme.
Molecular weight
(SDS–PAGE)
Number of subunits
N-terminal sequence
Thermal stability
Optimum pH
2
2
Not detected
<50 ꢀC (60 min)
7.0
PGQVYLIPGAHRGI
<55 ꢀC (60 min)
6.5
Optimum temperature
Stereoselectivity for
HPPO reduction
Km toward HPPO
Km toward NADPH
Vmax toward HPPO
40 ꢀC
50 ꢀC
>99% e.e. for (R)
>99% e.e. for (S)
20.1 mM
18.3 mM
190 mM
176 mM
3.4 mmol minꢂ1 mgꢂ1 2.8 mmol minꢂ1 mgꢂ1
sequence of the (R)-PPD-producing enzyme has not
been elucidated, judging from the inhibition of enzyme
activity by chelating reagents and its subunit structure
and molecular weight, the (R)-PPD-producing enzyme
might belong to alcohol dehydrogenase superfamily
Group II.
In a previous study, we purified and characterized an
(S)-PPD-producing enzyme from Williopsis saturnus
var. mrakii AJ5620 and suggested that it is a short chain
alcohol dehydrogenase.3) The enzyme purified here
catalyzes the stereoselective reduction of HPPO, that is
the conversion of HPPO to (R)-PPD. The properties of
the (R)-PPD-producing enzyme from T. fermentans and
the (S)-PPD-producing enzyme from W. saturnus var.
mrakii are summarized in Table 4. Both enzymes are
dimers and require NADPH as a coenzyme. In addition,
the enzyme activity of both enzymes is inhibited by
p-chloromercuribenzoic acid, ꢁ,ꢁ0-dipyridyl, o-phenan-
throline, and HgCl2, and the Km and Vmax values of the
two enzymes are similar. On the other hand, the relative
molecular mass of the enzyme from T. fermentans is
larger than that of the W. saturnus var. mrakii enzyme,
Acknowledgments
We thank Dr. K. Kubota, Dr. T. Tanaka, Dr. T.
Suzuki, Dr. E. Nakanishi, and Dr. K. Izawa of our
laboratory for their encouragement and helpful sugges-
tions throughout this study.

1646
I. KIRA and N. ONISHI
6, 333–339 (1999).
References
10) Inoue K, Makino Y, and Itoh N, Appl. Environ. Microbiol., 71,
3633–3641 (2005).
1) Ariens E, TiPS, 14, 68–75 (1993).
2) Kira I, Suzuki T, Oonishi N, and Nakanishi E, J. Mol. Cat. B:
Enzymatic, 56, 283–287 (2008).
11) Nie Y, Xu Y, Mu XQ, Wang HY, Yang M, and Xiao R, Appl.
Environ. Microbiol., 73, 3759–3764 (2007).
12) Bradford MM, Anal. Biochem., 72, 248–254 (1976).
13) Laemmli UK, Nature, 227, 680–685 (1970).
14) Kizaki N, Yasohara Y, Nagashima N, and Hasegawa J, J. Mol.
Cat. B: Enzymatic, 51, 73–80 (2008).
3) Kira I and Oonishi N, J. Biosci. Bioeng., 107, 116–118 (2009).
4) Prelog V, Pure Appl. Chem., 9, 119–130 (1964).
5) Peters J, Minuth T, and Kula MR, Enzyme Microb. Technol., 15,
950–958 (1993).
15) Bohren KM, Bullock B, Wermuth B, and Gabbay KH, J. Biol.
Chem., 264, 9547–9551 (1989).
6) Zelinski T, Peters J, and Kula MR, J. Biotechnol., 33, 283–292
(1994).
16) Matthew FR and Fewson CA, Crit. Rev. Microbiol., 20, 13–56
(1994).
7) Itoh N, Mizuguchi N, and Mabuchi M, J. Mol. Cat. B:
Enzymatic, 6, 41–50 (1999).
17) Isabel V, Josefa G, Ana B, and Roberto M, Biochem. J., 251,
461–466 (1988).
8) Wada M, Kataoka M, Kubota H, Yasohara Y, Kizaki N,
Hasegawa J, and Shimizu S, Biosci. Biotechnol. Biochem., 62,
280–285 (1998).
18) Kataoka M, Yamamoto K, Kawabata H, Wada M, Kita K,
Yanase H, and Shimizu S, Appl. Microbiol. Biotechnol., 51,
486–490 (1999).
9) Wada M, Kawabata H, Kataoka M, Yasohara Y, Kizaki
N, Hasegawa J, and Shimizu S, J. Mol. Cat. B: Enzymatic,