Synthesis of a chiral alcohol using a rationally designed Saccharomyces cerevisiae reductase and a NADH cofactor regeneration system

Article abstract of DOI:10.1016/j.molcatb.2012.01.016

Ethyl (S)-4-chloro-3-hydroxy butanoate (ECHB) is a building block for the synthesis of hypercholesterolemia drugs that function as 3-hydroxy-3- methylglutaryl coenzyme A reductase inhibitors. Yeast reductase YDL124W has been shown to convert ethyl 4-chloro-3-oxo butanoate (ECOB) into (S)-ECHB in an enantioselective manner. In that reduction, YDL124W absolutely prefers NADPH as a cofactor to NADH. NADPH is, however, much more expensive and unstable than NADH. In this study, four amino acid residues that directly interact with the 2′-phosphate group of adenosine ribose of NADPH were identified based on the homology model and docking model of YDL124W with ECOB and NADPH. Among various single mutants altered in these target residues, R264H evidenced noticeably increased reductase activity with NADH. The double mutant, R264H/R27Y, had more profoundly enhanced reductase activity with NADH. Kinetic analysis results demonstrated that these mutants had reduced K_m values toward NADH and increased k_cat/K_m values in comparison with the wild-type. Computer modeling showed that no hydrogen bond could be formed between the wild-type and 2′-hydroxyl group of adenosine ribose of the NADH cofactor, whereas three hydrogen bonds and one π-π stacking interaction could be formed between R264H or R264H/R27Y mutants and the NADH cofactor. These two mutants could efficiently produce (S)-ECHB using NADH as a cofactor in conjunction with formate dehydrogenase as a coupling enzyme.

Full text of DOI:10.1016/j.molcatb.2012.01.016

Journal of Molecular Catalysis B: Enzymatic 84 (2012) 15–21
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Synthesis of a chiral alcohol using a rationally designed Saccharomyces cerevisiae
reductase and a NADH cofactor regeneration system
Jihye Junga, Seongsoon Parkb, Hyung Kwoun Kima,∗
a Division of Biotechnology, The Catholic University of Korea, Bucheon 420-743, Republic of Korea
b Department of Chemistry, Center for NanoBio Applied Technology, Institute of Basic Sciences, Sungshin Women’s University, Seoul 136-742, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 25 January 2012
Ethyl (S)-4-chloro-3-hydroxy butanoate (ECHB) is a building block for the synthesis of hypercholes-
terolemia drugs that function as 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors. Yeast
reductase YDL124W has been shown to convert ethyl 4-chloro-3-oxo butanoate (ECOB) into (S)-ECHB
in an enantioselective manner. In that reduction, YDL124W absolutely prefers NADPH as a cofactor to
NADH. NADPH is, however, much more expensive and unstable than NADH. In this study, four amino
acid residues that directly interact with the 2ꢀ-phosphate group of adenosine ribose of NADPH were
identified based on the homology model and docking model of YDL124W with ECOB and NADPH. Among
various single mutants altered in these target residues, R264H evidenced noticeably increased reductase
activity with NADH. The double mutant, R264H/R27Y, had more profoundly enhanced reductase activity
with NADH. Kinetic analysis results demonstrated that these mutants had reduced Km values toward
NADH and increased kcat/Km values in comparison with the wild-type. Computer modeling showed that
no hydrogen bond could be formed between the wild-type and 2ꢀ-hydroxyl group of adenosine ribose of
the NADH cofactor, whereas three hydrogen bonds and one ␲–␲ stacking interaction could be formed
between R264H or R264H/R27Y mutants and the NADH cofactor. These two mutants could efficiently
produce (S)-ECHB using NADH as a cofactor in conjunction with formate dehydrogenase as a coupling
enzyme.
Keywords:
Reductase
Protein engineering
Cofactor preference
NADH
Chiral compound
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
cofactors, rather than NADH [10]. Several research groups have
attempted to alter this cofactor preference by rationally altering
The asymmetric reduction of ketones to optically pure alcohols
has become the focus of a great deal of attention, primarily in regard
to their possible use as intermediates for the synthesis of chiral
building block. Some microbial reductases are capable of reducing
the carbonyl groups of various compounds in a chemo-, regio-, and
stereo-selective manner [1,2].
Most reductases require NAD(P)H cofactor for the reduction of
ketone substrates. Although many reductases efficiently mediate a
variety of enantioselective bioconversion, the cofactor is too expen-
sive for the enzymatic bioconversion to be performed economically
[1,3]. To alleviate this cost factor, a cofactor recycling system is
usually introduced to the reductase bioconversion: namely, the
enzyme coupling reaction of glucose dehydrogenase (GDH) [4–6]
or formate dehydrogenase (FDH) [7–9].
the protein structure [11–14].
Katzberg et al. engineered yeast reductase Gre2p via site-
directed mutagenesis. The N9E mutant evidenced an enhanced
cofactor preference (NADH/NADPH) of 0.9, whereas the wild-type
had a preference of 0.007 [12]. Zhang et al. noted that mutant
S67D/H68D of Candida parapsilosis carbonyl reductase had an
altered specificity from NADPH to NADH [15]. Additionally, Banta
et al. reported that three amino acids in the cofactor-binding
pocket of Corynebacterium 2,5-diketo-d-gluconic acid reductase
were important for determining cofactor preference, and they con-
structed a mutant enzyme via combinatorial mutagenesis [11,16].
Shimizu et al. reported Candida magnolia reductase could synthe-
size optically pure ethyl (S)-4-chloro-3-hydroxybutanoate [17] and
developed Escherichia coli transformant cells coexpressing the car-
bonyl reductase and glucose dehydrogenase genes [18].
In the previous study, we reported Saccharomyces cerevisiae
YDL124W enzyme capable of generating ethyl (S) 4-chloro-3-
hydroxy butanoate ((S)-ECHB), which was a chiral building block for
the synthesis of Atorvastatin, a hyperlipidemia drug [19]. A homol-
ogy model of YDL124W was constructed on the basis of the X-ray
crystal structure of human liver ꢀ4-3-ketosteroid 5␤-reductase
NADPH cofactor is much more expensive and unstable than
NADH. Unfortunately, many reductases, including aldo-keto reduc-
tase (AKR) family members, employ NADPH exclusively as their
∗
Corresponding author. Tel.: +82 2 2164 4890; fax: +82 2 2164 4865.
E-mail address: hkkim@catholic.ac.kr (H.K. Kim).
1381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2012.01.016

16
J. Jung et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 15–21
(PDB ID 3BUV) [20]. In addition, its docking model with ethyl 4-
chloro-3-oxo butanoate (ECOB) substrate and NADPH cofactor was
also constructed, and its possible enantioselective reaction mech-
anism was elucidated [19].
exactly the same as those used in the previous single mutant prepa-
ration.
2.4. Expression and purification of mutant reductases
In this research, we attempted to determine the amino acid
residues that interact with the 2ꢀ-phosphate group of adenosine
ribose of NADPH based on the docking structure. We developed
mutant reductases capable of producing (S)-ECHB with the NADH
cofactor. Their kinetic parameters were determined and computer
modeling was conducted. Furthermore, two mutants were used to
generate (S)-ECHB by using NADH and formate dehydrogenase as
a cofactor and a coupling enzyme, respectively.
The recombinant plasmids were transformed into E. coli BL21
(DE3) cells. E. coli cells containing each plasmid were cultured at
20 ◦C in Luria–Bertani medium (1% tryptone, 0.5% yeast extract
and 0.5% NaCl) containing 100 ␮g/mL of ampicillin. When the
cells reached an optical density (OD600 nm) of 0.5, expression was
induced via the addition of 1 mM isopropyl thio-␤-d-galactoside
(IPTG). Cultivation was continued for an additional 24 h, after which
the cells were harvested via 10 min of centrifugation at 10,000 × g
at 4 ◦C and re-suspended in a 1/20 volume of 50 mM potassium
phosphate buffer (pH 7.0).
2. Experimentals
The recombinant E. coli cells were disrupted by sonication and
each mutant enzyme in the cell extracts was purified as follows.
First, each cell-free extract was loaded onto Ni-NTA spin columns
(QIAGEN GmbH, Hilden, Germany) equilibrated with a 50 mM
potassium phosphate buffer (pH 7.0) containing 10 mM imida-
zole and 300 mM NaCl, then washed with the 50 mM imidazole
buffer. The mutant reductases were then eluted from the column
via the application of a 300 mM imidazole buffer. The active frac-
tions were then collected and desalted using PD-MiniTrap G10
desalting columns (GE Healthcare Bio-Sciences AB, UK). The puri-
fied enzymes were analyzed by SDS-PAGE (Supplementary data
2) and their protein amounts were assayed through the Bradford
method.
The secondary structures of wild-type and mutant, R264H and
R264H/R27Y, were assessed by circular dichroism spectroscopy.
The ellipticity was measured in the range 190–260 nm with a scan-
ning speed of 100 nm/min in 1 s response time at 20 ◦C on a JASCO
J-815 spectrometer. Spectra were obtained by the average of three
successive spectra with purified enzyme at 50 mM potassium phos-
phate buffer (pH 7.0).
2.1. Chemicals
Ethyl-4-chloro-3-oxo butanoate (ECOB), formate dehydro-
genase (FDH, Candida boidinii), NAD(P)H, and NAD(P)+ were
purchased from Sigma–Aldrich Co. (St. Louis, MO). (R)- and (S)-
ethyl-4-chloro-3-hydroxy butanoate ((R)- and (S)-ECHB) were
received from Cronbio Com (Korea). All other chemicals were of
analytical grade.
2.2. Reductase and formate dehydrogenase activity assay
Reductase activity was evaluated at 30 ◦C by measuring the
reduction in absorbance at 340 nm for 5–10 min using a spec-
trophotometer. The reaction mixture (1 mL) consisted of 1 mM
ECOB (100 mM stock in DMSO), 0.2 mM NAD(P)H, 50 mM potas-
sium phosphate (pH 6.5) buffer, and 5–10 ␮L of cell-free extract
or 15 ␮g (for NADPH) or 30 ␮g (for NADH) of the purified reduc-
tase. One unit of reductase was defined as the quantity of enzyme
required to catalyze the oxidation of 1 ␮mol NAD(P)H in 1 min at
30 ◦C.
The oxidation reaction mixture (1 mL) of formate dehydroge-
nase consisted of 10 mM sodium formate, 0.2 mM NAD+, 50 mM
potassium phosphate (pH 6.5) buffer, and 10 ␮L of commercial
formate dehydrogenase. The reaction rate was monitored using a
spectrophotometer based on the increase in absorbance at 340 nm
for 5–10 min at 30 ◦C. One unit of formate dehydrogenase was
defined as the quantity required to reduce 1 ␮mol of NAD+ in 1 min
at 30 ◦C.
2.5. Kinetic studies of the mutant enzymes
Using the purified wild-type and six mutants (S260D, R264H,
R264Y, R264H/R27Y, R264H/S259E, and R264H/S260E), kinetic
studies were conducted. In order to elucidate the Km and kcat val-
ues of wild-type and mutants toward ECOB, enzyme assays were
conducted in the reaction mixtures containing 1–15 mM ECOB,
0.5 mM NAD(P)H, and 0.19–14 ␮M purified enzyme. The Km and
kcat values of each enzyme toward ECOB were calculated from the
Lineweaver–Burk plots. Then, in order to determine the Km values
of the wild-type and mutants toward NAD(P)H cofactor, enzyme
assays were conducted in the reaction mixtures containing excess
ECOB (10-fold concentration of Km ECOB) and 0.01–0.5 mM NAD(P)H
cofactor, and 0.19–14 ␮M purified enzyme. The Km values of each
enzyme were calculated from the Lineweaver–Burk plots.
2.3. Site-directed mutagenesis
Arg27, Ser259, Lys261, Arg264 and Ser260 residues were
mutated into other selected amino acids using a QuikChange light-
ning site-directed mutagenesis kit (Stratagene). The PCR primers
employed for the preparation of the mutants are provided in
Supplementary data 1.
For the preparation of single mutants, the S. cerevisiae YDL124W
gene (pETR124) was used as template DNA for the PCR reac-
tions [19]. The reaction conditions were as follows: 30 cycles of
denaturation at 95 ◦C for 20 s, annealing at 60 ◦C for 10 s, and
polymerization at 72 ◦C for 3.5 min. After the PCR reaction, the
reaction mixtures were treated with DpnI at 37 ◦C for 5 min and
subsequently transformed into E. coli XL10-Glod ultracompetent
cells from QuikChange lightning site-directed mutagenesis kit. The
recombinant plasmids were purified from the recombinant E. coli
cells and their nucleotide sequences were analyzed to confirm the
mutations.
2.6. Molecular modeling
Molecular modeling was carried out using the SYBYL model-
ing package 8.1 (Tripos, Inc.). The previously reported homology
model of the enzyme containing the substrate and NADPH was
employed [19]. The AMBER99 force field [21] was used to describe
the atom types of the enzyme containing the substrate and cofactor.
Some missing parameters were adapted from the general AMBER
force field (GAFF) [22]. Nonstandard partial charges were calculated
using the Pullman method for the enzyme and the Gasteiger-Huckel
method for the substrate and cofactor. A nonbonded cutoff dis-
For the preparation of double mutant enzymes, the R264H
reductase gene was employed as template DNA for the PCR reac-
tions. The PCR reaction conditions and the cloning methods were
˚
tance of 8 A and a distance-dependent dielectric function were
employed in all calculations. Energy minimizations were conducted

J. Jung et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 15–21
17
with the Powell minimizer and continued until the rms derivatives
A
−1
.
fell below 0.005 kcal mol−1
A
O
O
HO
O
˚
H
YDL124W
GDH
Cl
Cl
O
O
2.7. Enzymatic coupling reaction and reaction mixture analysis
ECOB
(S)-ECHB
NADPH
NADP+
glucose
The enantioselectivity of the enzymatic reduction of ECOB was
evaluated. The general procedure was as follows: 10 mM ECOB,
1 mM NADH, 4.8 mg of purified reductase, 20 mM sodium formate,
and 1.5 units of commercial formate dehydrogenase (toward NAD+)
were mixed in a total volume of 10 mL of potassium phosphate
buffer (50 mM, pH 6.5), and the mixture was incubated at 25 ◦C.
ECHB product was extracted by ethyl acetate and dried via evap-
oration. After acetylation step and washing steps, the organic phase
was subsequently analyzed using a GC system equipped with a
Chirasil-dex column (Varian, Ltd) [19]. The column temperature
was increased from 70 ◦C to 180 ◦C at a rate of 5 ◦C/min, and main-
tained for 2 min at 180 ◦C. By comparing the retention times and
peak areas of standard (R)- and (S)-alcohol peaks, the total quan-
tities of products in the reaction mixtures were calculated. The
retention times of (R)- and (S)-alcohols were 16.8 min and 17.0 min,
respectively.
glucono-
-lactone
B
O
O
HO
O
H
YDL124W mutant
Cl
Cl
O
O
ECOB
(S)-ECHB
NADH
NAD+
FDH
Formic acid
CO2
3. Results and discussion
Scheme 1. Schematic presentation of two coupling reactions. (A) YDL124W reduces
ECOB using the NADPH cofactor regenerated by glucose dehydrogenase (GDH). (B)
YDL124W mutant reduces ECOB using the NADH cofactor regenerated by formate
dehydrogenase (FDH).
3.1. Construction of mutant reductases based on rational design
In a previous study, yeast YDL124W reductase was reported
to convert ECOB substrate exclusively into the (S)-ECHB product
[19]. In that conversion, YDL124W utilized NADPH as a cofactor,
but not NADH. In that enzyme reaction, GDH was employed as
a coupling enzyme for NADPH recycling (Scheme 1A). NADPH is,
however, far more expensive and unstable than NADH. For this
reason, the development of YDL124W mutant enzyme able to use
NADH would render the production of (S)-ECHB more economi-
cal. In this new process, FDH can be employed for NADH recycling
(Scheme 1B). This FDH coupling reaction provides an extra advan-
tage; GDH produces gluconic acid and it may be cumbersome to
adjust the reaction pH frequently in the middle of the reaction pro-
cess, whereas FDH produces CO2, and it is not essential that the
reaction pH be adjusted during the reaction [1,3].
structural analysis of the cofactor-binding site using the dock-
ing model suggested that the 2ꢀ-phosphate group of adenosine
ribose of NADPH was surrounded by five amino acid residues
including Arg27, Ser259, Ser260, Lys261, and Arg264. The two
guanidino nitrogen atoms of Arg27 and Arg264 were located
˚
at a distance of 2.7–2.8 A from the phosphate oxygen atoms,
and seemed to form two salt bridges (Fig. 1). The hydroxyl
groups of Ser259 and the ␧-amino group of Lys261 were also
located close and formed hydrogen bonds with the phosphate
group; thus they made a hydrophilic pocket for the binding of
the phosphate group of the adenosine ribose of NADPH. The
S259–S260–K261–x–x–R264 sequence was similar to the previously
reported consensus sequence (K–S–x–x–x–x–R) [16]. Therefore, we
Homology model of YDL124W and its docking model with
ECOB and NADPH were recently constructed [19]. In this research,
Fig. 1. Homology model of YDL124W and NADPH binding site. (A) The 3-dimensional homology model was redrawn on the basis of the previously reported structure [19].
(B) Line diagrams of a docking structure of YDL124W with NADPH. The phosphate group at 2ꢀ-O of the adenosine group is surrounded by Arg27, Ser259, Lys261, and Arg264.

18
J. Jung et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 15–21
A
A
2.5
2.0
1.5
1.0
0.5
0.0
2.5
2.0
1.5
1.0
0.5
0.0
NADPH
NADPH
B
0.16
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
NADH
B
0.18
0.15
0.12
0.09
0.06
0.03
0.00
NADH
Fig. 2. Reductase activity of various single mutants. (A) Reductase activities were
measured using NADPH cofactor. (B) Reductase activities were measured using
NADH cofactor.
attempted to change this region, along with an additional residue
(Arg27).
Fig. 3. Reductase activity of various double mutants. (A) Reductase activities were
measured using NADPH cofactor. (B) Reductase activities were measured using
NADH cofactor.
In this study, a variety of YDL124W mutants were constructed
to make ‘NADH-utilizing reductase’. At first, a total of 16 differ-
ent single mutant enzymes were constructed; Arg27 and Arg264
were mutated into Ala, Asp, Glu, His, and Tyr, respectively. Addi-
tionally, Ser259, Ser260, and Lys261 were mutated into Ala and
Asp, respectively. After the purification steps, the purities of mutant
reductases were ascertained via SDS-PAGE (Supplementary data
2). When the purified mutant enzymes were tested for their
reductase activities using ECOB and NADH as the substrate and
cofactor, respectively, the majority of the mutants evidenced
very little activity relative to the wild-type (Fig. 2). However,
among them, two mutant enzymes (R264H and R264Y) evi-
denced profoundly increased reductase activities with the NADH
cofactor. The purified R264H and R264Y exhibited specific activ-
ities of 0.136 and 0.079 U/mg, respectively, whereas the purified
wild-type reductase exhibited a specific activity of 0.007 U/mg.
Their reductase activities were increased by 19-fold and 11-fold,
respectively.
Secondly, various double mutant enzymes were designed using
R264H as a template enzyme. Arg27, Ser259, Ser260, and Lys261
were mutated into Glu, His, and Tyr, respectively. When the puri-
fied double mutant enzymes were assessed for reductase activity
using the same method as the single mutant enzymes, they evi-
denced substantially different activities (Fig. 3). Among them, the
R264H/R27Y mutant showed the most profoundly enhanced activi-
ties in comparison with R264H. The purified R264H/R27Y exhibited
a specific activity of 0.171 U/mg, which was 24-fold higher than that
of the wild-type.
3.2. Kinetic parameters of mutant reductases
In order to elucidate the enzyme activities in the presence
of NADPH or NADH cofactor, kinetic studies were conducted
against six mutants, as well as wild-type reductase (Table 1). The
six mutants tested included S260D, R264H, R264Y, R264H/R27Y,
R264H/S259E, and R264H/S260E. As anticipated, the wild-type
enzyme not only had a very low Km value (0.00436 mM) for NADPH
but also a very high Km value (43.6 mM) for NADH. Moreover, the
kcat/Km ECOB (104 min−1 mM−1) in the presence of NADPH cofactor
was quite a bit higher than that (0.453 min−1 mM−1) seen in the
presence of the NADH cofactor.
Most mutants evidenced rather elevated Km values
(0.03–0.9 mM) toward NADPH compared to the wild-type
(0.00436 mM). Conversely, many of the mutants had substan-
tially reduced Km values (0.05–4 mM) toward NADH relative to the
wild-type (43.6 mM). In the case of R264H, the Km value for NADH
was 1.15 mM, which suggested that the NADH cofactor was bound
tightly within the active pocket of the mutant enzyme. R264H/R27Y
evidenced the lowest Km value of 0.058 mM. These two mutants
exhibited substantially elevated kcat/Km ECOB values with the NADH

J. Jung et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 15–21
19
Table 1
Kinetic parameters of wild-type and mutant reductase enzymes.
NADPH
NADH
Km cofactor (mM)
Km ECOB (mM)
kcat (min−1
)
kcat/Km ECOB
Km cofactor (mM)
Km ECOB (mM)
kcat (min−1
)
kcat/Km ECOB
(min−1 mM−1
)
(min−1 mM−1
)
WT
S260D
R264H
R264Y
R264H/R27Y
R264H/S259E
R264H/S260E
0.00436
0.0350
0.0660
0.0390
0.0740
0.991
1.08
1.21
7.69
3.77
8.10
3.83
26.3
112
171
507
288
759
104
141
43.6
4.18
1.15
0.354
0.0580
ND
10.2
3.39
1.73
4.49
1.15
31.6
2.35
4.62
14.8
63.9
31.9
34.1
1.21
35.9
0.453
4.37
36.9
7.10
29.7
65.9
76.4
93.7
0.496
47.1
1.90
1240
0.0383
15.3
0.143
0.374
ND, not detected.
cofactor (36.9 and 29.7 min−1 mM−1, respectively) compared to
that of the wild-type with NADH cofactor (0.453 min−1 mM−1);
these values were 82- and 66-fold higher than those of the
wild-type (Table 1). Circular dichroism results indicated that the
secondary structures of both R264H and R264H/R27Y were almost
the same as that of the wild-type (Supplementary data 3).
Although the kcat/Km ECOB values of R264H and R264H/R27Y
mutants with NADH were slightly lower than that
(104 min−1 mM−1) of wild-type with NADPH, they seemed to
have sufficiently high activities for the production of (S)-ECHB
using NADH cofactor.
phosphate group. The distances between the oxygen atoms of the
phosphate group and the terminal nitrogen of Arg or Lys were
˚
measured at 2.5–2.8 A. The oxygen atoms (O␥) of the Ser and the
˚
phosphate are separated by a distance of 2.5–3.0 A. The distances
represent typical hydrogen bonding in a protein. These electronic
interaction and hydrogen bonds are presumably a key interaction
for the acceptance of NADPH as a cofactor.
On the other hand, hydrogen bonds were not found in the dock-
ing structure of the wild-type enzyme with NADH (Fig. 4B). NADH
does not have a negatively charged phosphate group; thus no
electronic interaction is possible. The positively charged residues
repelled one another and moved away from the ribose ring of the
adenosine group. Therefore, the YDL124W enzyme would not effi-
ciently utilize NADH as a cofactor.
3.3. Molecular basis of utilizing NADH as a cofactor by the
mutants
A minimized model of the R264H mutant enzyme with NADPH
demonstrated that all hydrogen bonds from Arg27, Ser259, and
Lys261 to the phosphate group are available (Fig. 4C). Additionally,
the protonated nitrogen atom (N␦) [23] of the His serves a hydro-
gen bond donor similar to the Arg, although the His is incapable of
neutralizing the negative charge of the phosphate group. Addition-
ally, the imidazole ring of the His264 residue exhibited a stacking
interaction with the adenosine base of NADPH. This presumably
explains how the Arg264His mutant can still accept NADPH as a
cofactor.
In order to analyze the binding affinity of NADPH or NADH
to the wild-type or R264H mutant enzyme, we generated three
other docking structures from the docking model of the wild-type
enzyme containing the ECOB and NADPH via the substitution of
NADPH with NADH or replacement of the Arg264 residue with His.
After substitution, we minimized all four models of the docking
structures and analyzed the phosphate-binding region (Table 2).
An energy-minimized structure of the wild-type enzyme with
NADPH demonstrated that the negatively charged phosphate group
at the adenosine group is stabilized by several positively charged
amino acids, such as two Arg and one Lys (Fig. 4A). Additionally,
a Ser residue made hydrogen bonds with an oxygen atom of the
A docking model of the R264H mutant enzyme with NADH
also evidenced a similar hydrogen bonding interaction with the
residues, except for the Lys261 residue (Fig. 4D). The 2ꢀ-oxygen
Table 2
Hydrogen bonds in possible docking structures of the wild type and Arg264His mutant enzyme with NADPH or NADH.
Atom of residue
Atom of cofactor
Distancea
NADPH
O1 of the phosphate P2ꢀ
2.8 A
2.5 and 3.0 A
˚
N␻ (Arg27)
O␥ (Ser259)
N␨ (Lys261)
N␻ (Arg264)
O2 and O3 of the phosphate P2ꢀ
O1 and O2 of the phosphate P2ꢀ
O1 and O3 of the phosphate P2ꢀ
˚
Wild type
˚
˚
2.5 and 2.6 A
2.7 and 2.8 A
NADH
O2ꢀ of the adenosine ribose
O2ꢀ of the adenosine ribose
O2ꢀ of the adenosine ribose
O2ꢀ of the adenosine ribose
˚
4.0 A
N␻ (Arg27)
O␥ (Ser259)
N␨ (Lys261)
N␻ (Arg264)
˚
4.7 A
Wild type
˚
7.0 A
˚
4.8 A
NADPH
˚
2.7 and 2.8 A
N␻ (Arg27)
O␥ (Ser259)
N␨ (Lys261)
N␦ (His264)
O1 and O2 of the phosphate P2ꢀ
O2 of the phosphate P2ꢀ
2.4 A
2.7 and 2.7 A
˚
R264H mutant
O1 and O2 of the phosphate P2ꢀ
O3 of the phosphate P2ꢀand O2ꢀ of the adenosine ribose
˚
˚
2.9and 3.0 A
NADH
O2ꢀ of the adenosine ribose
O2ꢀ of the adenosine ribose
O2ꢀ of the adenosine ribose
O2ꢀ of the adenosine ribose
˚
3.1 A
N␻ (Arg27)
O␥ (Ser259)
N␨ (Lys261)
N␦ (His264)
˚
3.1 A
R264H mutant
˚
7.0 A
˚
2.9 A
a
˚
Distance between non-hydrogen atoms (N O or O O). Distances of 2.7–3.2 A are considered as a range for a hydrogen bond.

20
J. Jung et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 15–21
Fig. 4. Proposed ECOB and NADPH or NADH binding structures. (A) Wild-type enzyme with NADPH. (B) Wild-type enzyme with NADH. (C) Arg264His mutant enzyme with
NADPH. (D) Arg264His mutant enzyme with NADH. ECOB (yellow), NADPH or NADH (yellow and atomic color), Arg27, Ser259, Lys261, Arg264, and His264 were represented
˚
as a stick model. The imidazole ring of the His264 evidenced a stacking interaction with the adenosine base of NADPH or NADH (distance: ∼3.5–4.3 A). (For interpretation of
the references to color in figure caption, the reader is referred to the web version of the article.)
atom of the ribose ring is stabilized by hydrogen bonds with Ser259
and Arg27. The positive charges of the Arg27 and Lys261 residues
repel each other, and only one of the residues could function as
a hydrogen bond donor. Additionally, the unprotonated nitrogen
atom (N␦) of the His264 residue appears likely to produce hydrogen
bonds to the 2ꢀ-oxygen atom of the terminal ribose ring of NADH
and the imidazole ring of the histidine has a stacking interaction
with the adenosine base. These interactions presumably explain
the Arg264His mutant enzyme can accept NADH with the lower
Km value relative to the wild-type enzyme.
3.4. Mutant reductase and FDH coupling reaction
Bioconversions of the wild-type and two mutants (R264H and
R264H/R27Y) were conducted using NADH and commercial FDH
Additionally, we modeled the double mutant enzyme
(R264H/R27Y) containing ECOB and NADH (Fig. 5). This mutant
enzyme possesses a lower Km value for NADH relative to the
wild-type and the R264H mutant enzyme. The 2ꢀ-oxygen atom
˚
of the ribose ring can make hydrogen bonds to Ser259 (3.2 A),
˚
˚
Lys261 (3.0 A), and His264 (3.0 A). The Tyr27 residue exhibits
no interaction with the 2ꢀ-oxygen atom. However, unlike the
R264H mutant enzyme, the Lys261 residue, which is located
away from the 2ꢀ-oxygen atom in a docking model of the R264H
mutant enzyme with NADH (Fig. 4B), functions as a hydrogen bond
donor. The substitution of Arg27 with tyrosine looses a hydrogen
bond between the Arg27 residue and the 2ꢀ-oxygen atom, but
gains hydrogen bonding between the 2ꢀ-oxygen atom and Lys261
because the repulsive interaction between Arg27 and Lys261
residues is diminished, and thus the Lys can be brought closer to
Fig. 5. A proposed ECOB and NADH binding structure of the R264H/R27Y mutant
enzyme. The color and representation were identical to those in Fig. 4. The 2ꢀ-oxygen
˚
˚
the oxygen atom. Additionally, the Tyr27 residue possesses a weak
atom of the ribose ring hydrogen bonds to Ser259 (3.2 A), Lys261 (3.0 A), and His264
ꢀ
ꢀ
˚
(3.0 A). The Tyr27 residue evidences a weak interaction with the 3 -oxygen atom
(3.4 A). (For interpretation of the references to color in figure caption, the reader is
˚
hydrogen bond (3.4 A) with the 3 -oxygen atom. These interactions
˚
may reduce the Km value for NADH.
referred to the web version of the article.)

J. Jung et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 15–21
21
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
A
B
C
R264H
WT
R264H/R27Y
0
30
60
90 120
0
30
60
90
120
0
30
60
90 120
Reaction time (min)
Reaction time (min)
Reaction time (min)
Fig. 6. Bioconversion using mutant reductases. Wild-type (A), R264H (B), and R264H/R27Y (C) were used in the bioconversion of ECOB into (S)-ECHB together with FDH and
NADH. Closed and open circles represent the conversion yield and e.e.p, respectively.
as a cofactor and coupling enzyme, respectively. In the case of
wild-type (Fig. 6A), the bioconversion yield and rate were 25% and
2.66 ␮mol h−1 mg−1, respectively. These low yield and slow con-
version rates were anticipated as the wild-type had a very low
kcat/Km value (0.453 min−1 mM−1) with NADH cofactor (Table 1).
Moreover, the wild-type showed an unexpected reduction in enan-
tiomeric excess of 91% for (S)-ECHB when NADH was employed as
a cofactor. This lower e.e. value seemed to be related with its high
Km value (10.2 mM) toward ECOB (Table 1).
However, in the case of R264H and R264H/R27Y, we noted
a conversion yield of 84% and 93%, respectively, within 120 min
of reaction (Fig. 6B and C). The conversion rates were 18.9 and
20.1 ␮mol h−1 mg−1, respectively, which were approximately 7-
fold higher than that of the wild-type. These results clearly
demonstrated that the two mutants could utilize NADH as a cofac-
tor, which was continuously recycled by FDH. The single mutant
and the double mutant produced (S)-ECHB enantioselectively with
e.e.p values of 98% and 99%, respectively.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcatb.2012.01.016.
References
[1] K. Goldberg, K. Schroer, S. Lu˝ tz, A. Liese, Appl. Microbiol. Biotechnol. 76 (2007)
237–248.
[2] J.C. Moore, D.J. Pollard, B. Kosjek, P.N. Devine, Acc. Chem. Res. 40 (2007)
1412–1419.
[3] A. Weckbecker, H. Gröger, W. Hummel, Adv. Biochem. Eng. Biotechnol. 120
(2010) 195–242.
[4] Y.H. Choi, H.J. Choi, D. Kim, K.N. Uhm, H.K. Kim, Appl. Microbiol. Biotechnol. 87
(2010) 185–193.
[5] H.J. Park, J. Jung, H.J. Choi, K.N. Uhm, H.K. Kim, J. Microbiol. Biotechnol. 20 (2010)
1300–1306.
[6] Q. Ye, P. Ouyang, H. Ying, Appl. Microbiol. Biotechnol. 89 (2011) 513–522.
[7] M. Bozic, S. Pricelius, G.M. Guebitz, V. Kokol, Appl. Microbiol. Biotechnol. 85
(2010) 563–571.
[8] R. Kratzer, M. Pukl, S. Egger, M. Vogl, L. Brecker, B. Nidetzky, Biotechnol. Bioeng.
108 (2010) 797–803.
[9] K. Schroer, U. Mackfeld, I.A.W. Tan, C. Wandrey, F. Heuser, S. Bringer-Meyer, A.
Weckbecker, W. Hummel, T. Daussmann, R. Pfaller, A. Liese, S. Lu˝ tz, J. Biotech-
nol. 132 (2007) 438–444.
4. Conclusion
[10] V.Z. Pletnev, C.M. Weeks, W.L. Duax, Proteins 57 (2004) 294–301.
[11] S. Banta, B.A. Swanson, S. Wu, A. Jarnagin, S. Anderson, Protein Eng. 15 (2002)
131–140.
[12] M. Katzberg, N. Skorupa-Parachin, M.F. Gorwa-Grauslund, M. Bertau, Int. J. Mol.
Sci. 11 (2010) 1735–1758.
[13] G.A. Khoury, H. Fazelinia, J.W. Chin, R.J. Pantazes, P.C. Cirino, C.D. Maranas,
Protein Sci. 18 (2009) 2125–2138.
[14] S. Watanabe, A. Abu Saleh, S.P. Pack, N. Annaluru, T. Kodaki, K. Makino, Micro-
biology 153 (2007) 3044–3054.
We developed several YDL124W mutants capable of using
NADH as a cofactor by rational design in this study. Among them,
the R264H and R264H/R27Y mutants evidenced very low Km values
toward NADH and highly increased kcat/Km values toward ECOB.
Although the kcat/Km values were still less than that of the wild-type
toward NADPH, the two mutants could carry out the bioconver-
sion efficiently in the presence of NADH and FDH as the cofactor
and coupling enzyme, respectively. These results suggest that these
mutants can be employed for the economic production of (S)-ECHB.
[15] R. Zhang, Y. Xu, Y. Sun, W. Zhang, R. Xiao, Appl. Environ. Microbiol. 75 (2009)
2176–2183.
[16] S. Banta, S. Anderson, J. Mol. Evol. 55 (2002) 623–631.
[17] M. Wada, H. Kawabata, A. Yoshizumi, M. Kataoka, S. Nakamori, Y. Yasohara, N.
Kizaki, J. Hasegawa, S. Shimizu, J. Biosci. Bioeng. 87 (1999) 144–148.
[18] N. Kizaki, Y. Yasohara, J. Hasegawa, M. Wada, M. Kataoka, S. Shimizu, Appl.
Microbiol. Biotechnol. 55 (2001) 590–595.
Acknowledgments
[19] J. Jung, H.J. Park, K.N. Uhm, D. Kim, H.K. Kim, Biochim. Biophys. Acta 1804 (2010)
1841–1849.
[20] L. Di Costanzo, J.E. Drury, T.M. Penning, D.W. Christianson, J. Biol. Chem. 283
(2008) 16830–16839.
[21] J. Wang, P. Cieplak, P.A. Kollman, J. Comput. Chem. 21 (2000) 1049–1074.
[22] J. Wang, W. Wang, P.A. Kollman, D.A. Case, J. Mol. Graph. Model. 25 (2006)
247–260.
[23] J.L. Sudmeier, E.M. Bradshaw, K.E. Haddad, R.M. Day, C.J. Thalhauser, P.A. Bul-
lock, W.W. Bachovchin, J. Am. Chem. Soc. 125 (2003) 8430–8431.
This work was supported by the 21C Frontier Microbial
Genomics and Applications Center Program, Ministry of Educa-
tion, Science & Technology, Republic of Korea, and by Basic Science
Research Program through the National Research Foundation of
Korea (NRF) funded by the Ministry of Education, Science and Tech-
nology (2011-0009692) and the 2011 Research Fund of the Catholic
University of Korea. We appreciate the KRICT for utilizing the SYBYL
package.