A genomic search approach to identify carbonyl reductases in Gluconobacter oxydans for enantioselective reduction of ketones

Article abstract of DOI:10.1080/09168451.2014.925775

The versatile carbonyl reductases from Gluconobacter oxydans in the enantioselective reduction of ketones to the corresponding alcohols were exploited by genome search approach. All purified enzymes showed activities toward the tested ketoesters with different activities. In the reduction of 4-phenyl-2-butanone with in situ NAD(P)H regeneration system, (S)-alcohol was obtained with an e.e. of up to 100% catalyzed by Gox0644. Under the same experimental condition, all enzymes catalyzed ethyl 4-chloroacetoacetate to give chiral products with an excellent e.e. of up to 99%, except Gox0644. Gox2036 had a strict requirement for NADH as the cofactor and showed excellent enantiospecificity in the synthesis of ethyl (R)-4-chloro-3-hydroxybutanoate. For the reduction of ethyl 2-oxo-4-phenylbutyrate, excellent e.e. (>99%) and high conversion (93.1%) were obtained by Gox0525, whereas the other enzymes showed relatively lower e.e. and conversions. Among them, Gox2036 and Gox0525 showed potentials in the synthesis of chiral alcohols as useful biocatalysts.

Full text of DOI:10.1080/09168451.2014.925775

Bioscience, Biotechnology, and Biochemistry, 2014
Vol. 78, No. 8, 1350–1356
A genomic search approach to identify carbonyl reductases in Gluconobacter
oxydans for enantioselective reduction of ketones
Rong Chen1,2, Xu Liu1, Jinping Lin1,* and Dongzhi Wei1,*
1State Key Laboratory of Bioreactor Engineering, New World Institute of Biotechnology, East China University of
2
Science and Technology, Shanghai, China; Center for Biomedicine and Health, Division of Basical Medicine,
Hangzhou Normal University; Hangzhou, China
Received December 24, 2013; accepted March 10, 2014
http://dx.doi.org/10.1080/09168451.2014.925775
The versatile carbonyl reductases from Gluconob-
acter oxydans in the enantioselective reduction of
ketones to the corresponding alcohols were exploited
by genome search approach. All purified enzymes
showed activities toward the tested ketoesters with
different activities. In the reduction of 4-phenyl-2-
butanone with in situ NAD(P)H regeneration system,
(S)-alcohol was obtained with an e.e. of up to 100%
catalyzed by Gox0644. Under the same experimental
condition, all enzymes catalyzed ethyl 4-chloroace-
toacetate to give chiral products with an excellent e.e.
of up to 99%, except Gox0644. Gox2036 had a strict
requirement for NADH as the cofactor and showed
excellent enantiospecificity in the synthesis of ethyl
(R)-4-chloro-3-hydroxybutanoate. For the reduction
of ethyl 2-oxo-4-phenylbutyrate, excellent e.e. (>99%)
and high conversion (93.1%) were obtained by
Gox0525, whereas the other enzymes showed rela-
tively lower e.e. and conversions. Among them,
Gox2036 and Gox0525 showed potentials in the
synthesis of chiral alcohols as useful biocatalysts.
and pharmacologically important materials: (R)-carni-
tine, (R)-4-amino-3-hydroxy-butayric acid, and (R)-4-
hydroxy-2-pyrrolidone.4) Another chiral ketoester,
enantiopure ethyl 2-hydroxy-4-phenylbutyrate is com-
mon precursor in the manufacture of a variety of angio-
tensin-converting
enzyme
inhibitors
including
benazepril, cilazapril, and enalapril.5,6) Preparing these
chiral alcohols by asymmetric reduction of carbonyl
substrates via biocatalyst is an efficient and practical
way.7) These biocatalysts, such as carbonyl reductases,
have drawn considerable attention due to not only the
advantage of their high enantioselectivity and regiose-
lectivity, but also economic viability and environment
friendliness.8,9) Carbonyl reductases widely used in
asymmetric reduction can achieve up to 100% theoreti-
cal yield.8,10) Up to now, many carbonyl reductases
have been obtained either by culture-based strain
enrichment or by a metagenomic screening with differ-
ent substrate specificities and stereoselectivities.11)
However, the former excludes a large fraction of micro-
bial organisms in environments and the latter is time
consuming and laborious. In comparison with tradi-
tional approaches, genome mining has become more
effective and promising which involves searching new
enzymes from databases with those of known
enzymes.12) Some new carbonyl reductases have been
obtained in this way from many species.13)
Key words: carbonyl reductase; enantioselective
reduction; chiral alcohols
Chiral alcohols with additional functional groups
play an increasingly important role as building blocks
for the synthesis of enantiomeric pure pharmaceuticals,
agrochemicals, and natural products. Optical active sec-
ondary alcohols, especially chiral aryl alcohols and ke-
toesters, are widely used as intermediates. For example,
(S)-4-phenyl-2-butanol is used as a precursor for anti-
hypertensive agents (labetalol) and spasmolytics or
anti-epileptics(emepronium)).1,2) Ethyl (S)-4-chloro-3-
hydroxybutanoate ((S)-CHBE) is used for production
of the HMG-CoA reductase inhibitors(statins),3) while
(R)-CHBE is useful for the synthesis of biologically
In the current study, we characterized novel enzymes
based on the genome of Gluconobacter oxydans, which
is commonly applied in many industries for its abilities
of stereoselective and regioselective oxidoreduc-
tion.14,15) The G. oxydans 621H genome revealed the
presence of 75 open reading frames (ORFs) that
encoded putative dehydrogenases/oxidoreductases based
on sequence analysis.16) Among these, a number of
ORFs had been expressed and functionally determined,
some of which showed important biological function
and great biotechnological interest. Gox1615 was
*Corresponding authors. Email: jplin@ecust.edu.cn (J. Lin); dzhwei@ecust.edu.cn (D. Wei)
Abbreviations: BsGDH, Bacillus subtilis-derived glucose dehydrogenase; CHBE, ethyl 4-chloro-3-hydroxybutanoate; COBE, ethyl 4-chloroacetoac-
etate; CsKR, Ketoreductase from the Cyanobacterium Synechococcus sp. Strain PCC 7942; e.e., enantiomeric excess; fabG, 3-ketoacyl-acyl carrier
protein reductases; HPBE, ethyl 2-hydroxy-4-phenylbutyrate; HPLC, high-performance liquid chromatography; IPTG, isopropyl β-D-thiogalactopy-
ranoside; NAD(P)H, Nicotinamide adenine dinucleotide (phosphate) reduced tetrasodium salt; OPBE, ethyl 2-oxo-4-phenylbutyrate; ORFs, open
reading frames.
© 2014 Japan Society for Bioscience, Biotechnology, and Agrochemistry

Genome mining of microorganic carbonyl reductases
1351
classified as an NADP-dependent glycerol dehydroge-
nase that can be applied in the kinetic resolution of
glyceraldehyde and in the production of enantiopure
L-glyceraldehyde on a preparative scale.17) Gox2181, as
a polyol dehydrogenase, had potential applications in
asymmetric oxidation of sugars and polyols for the syn-
thesis of certain rare ketoses.18) Considering the huge
utilization potentiality, we noticed that 21 putative car-
bonyl reductases maybe catalyze the reversible oxidation
of alcohol to aldehydes or ketones, therefore we cloned
some of these genes and obtained their corresponding
enzymes for determining their chemo-, regio-, and stere-
oselectivities. The results of this study can promote the
application of G. oxydans in many industries.
Gene cloning and expression of carbonyl reduc-
tases. All DNA manipulations and bacterial transfor-
mations were carried out according to standard
protocols.21) The putative carbonyl reductases gene
gox2036 (YP_192428.1), gox0525 (YP_190959),
gox0644 (YP_191077.1), gox1615 (YP_192012.1),
gox1598 (YP_191995.1), gox1462 (YP_191867.1), and
gox0290 (YP_190729.1) were amplified by polymerase
chain reaction (PCR). The primers, vectors, and bacte-
rial strains used in this study were listed in Table 1.
The amplified products were digested with relative dou-
ble restriction enzymes and then ligated to vectors. The
constructed plasmids were introduced into E. coli,
BL21 (DE3) or Rosetta (DE3). The positive clones
were screened by PCR amplification and further
verified by sequencing. The E. coli host carrying the
recombinant plasmid was induced with isopropyl
β-D-thiogalactopyranoside (IPTG) ranging from 0.1 to
1.0 mM and induced at 25, 30, and 37 °C to obtain sol-
uble recombinant enzymes. The cell-free extract was
prepared after lysis of the cell in a 20 mM phosphate
buffer by ultrasonication at 4 °C. Cell debris was
removed by centrifugation. The supernatant was used
to determine the enzyme expression by sodium dodecyl
sulfate-polyacrylamide gel electrophoresis, and the
enzyme concentration was determined by the Bradford
method. The soluble enzymes were purified by Ni-NTA
agarose according to the operation manual (Qiagen,
Hilden, Germany), and enzyme activity was assayed.
Material and methods
Chemicals and micro-organism cultivation.
Ketones and chiral alcohol standard samples were
purchased from Sigma–Aldrich and TCI companies.
All other chemicals used were commercially available
and of analytical grade.
G. oxydans DSM2343 was cultured in yeast–sorbitol
medium containing 8% (w/v) sorbitol and 2% (w/v)
yeast extract at 30 °C. Cells were harvested by centri-
fugation at 8000 g for 10 min and used for the total
genomic DNA preparation according to the cetyltrim-
ethyl ammonium bromide extraction protocol.19) Esche-
richia coli DH5α, Rosetta (DE3), and BL21 (DE3)
were routinely cultured in Luria–Bertani medium.
Ampicillin, kanamycin, and chloramphenicol were
added to the medium at 100, 50, and 25 μg mL−1,
respectively, when necessary. Recombinant Bacillus
subtilis-derived glucose dehydrogenase (BsGDH) was
prepared as previously described.20)
Activity assay and determination of kinetic parame-
ters.
The enzyme activity toward the reduction of
ketones was determined by spectrophotometrically
measuring the oxidation of NAD(P)H at 340 nm
(ε = 6.22 mM−1 cm−1) in the presence of an excess
Table 1. Primers, test vectors and strains used.
Test
Gene
Primer sequence
vectors
Strains
gox2036
F: 5′-gaattcATGTCCCTTTCTGGAAAAA-3′
R: 5′- aagcttTCAGCGGAAAACGAGAC-3′
pET28a
BL21(DE3)
gox0525
gox0644
F: 5′-gaattcATGACCCACAGAGTAGCGC-3′
R: 5′-aagcttTCAGGCGACGAGACCGCC-3′
pET28a
BL21(DE3)
BL21(DE3)
F: 5′-catatgAGCGCCTCGTCACAGGTTC-3′
R: 5′-aagcttATATCAGAATTTCGCCGTAT-3′
pSE380 or
pET32a/
28a
gox1615
gox1598
F: 5′-ggatccAGCGCCGCATCCGACACCA-3′
R: 5′-ctcgagATATCAGTCCCGTGCCGGGG-3′
pET32a
BL21(DE3)
BL21(DE3)/
F: 5′-ggatccCTGGGGGCTTCCCCCCCATT-3′
R: 5′-ctcgagTCAGGGAGCCGGGTTGGAG-3′
pET32a/
28a
Rosetta
(DE3)
gox1462
gox0290
F: 5′-ggatccCAGTATCGTCAGCTTGGTC-3′
pET32a/
28a
BL21(DE3)/
R: 5′-ctcgagTCATTTTTTCGTTTCAAGGGCC-3′
Rosetta
(DE3)
F: 5′-gtcgacATGAAAACAGTCACACTCAGGA-3′
R: 5′-gcggccgcGATCATGGCGAGAGATACCT-3′
pET32a
Rosetta
(DE3)

1352
R. Chen et al.
amount of ketones. The change in absorbance of NAD
(P)H was monitored at 340 nm using an ultraviolet-visi-
ble spectrophotometer with a temperature-controlled
cuvette holder (Shimadzu Co., Kyoto, Japan). One unit
(U) activity was defined as the amount of enzymes
required to catalyze the oxidation of 1 μmol NAD(P)H
per minute at 30 °C. The data were expressed as U/mg
of protein. The standard reaction mixture contained
100 mM phosphate buffer (pH 7.0), 0.1 mM NAD(P)H,
5 mM substrate, and enzyme solvent in a total volume
of 1 mL. The reaction was initiated by adding 20 μL
solvent containing 10 μg–40 μg of enzymes. Blanks
without the enzyme were carried out for each substrate,
and data were collected in triplicate. Protein concentra-
tions were determined with the bicinchoninic acid assay
using bovine serum albumin as a standard.
NAD(P)H or NADH was assayed under standard
reaction conditions for the study of coenzyme depen-
dence. A range of substrates from 1 to 20 mM concen-
tration was assayed under the standard reaction
conditions for the study of kinetics. Apparent values of
Michaelis constant (Km) and kcat were calculated by fit-
ting the data into Michaelis–Menten equation using the
SigmaPlot (Systat Software Inc., San Jose, CA, USA).
All reactions followed Michaelis–Menten-type kinetics.
by sequence analysis and possibly reduce carbonyl
substrates.16) Three of them, Gox0458, Gox1458, and
Gox1538, were previously expressed in E. coli and had
no activities of aldo/keto reduction and alcohol/aldehyde
oxidation (Schweiger PB, unpublished results). In our
previous study, the gox1139, gox1034, gox0601, and
gox2371 genes were heterologously overexpressed in
E. coli, and their corresponding proteins had no activi-
ties toward ketones. The Gox2378 and Gox0950, pre-
dicted to locate on the membrane, were insoluble
expressed in E. coli, as same as Gox0716. The insoluble
proteins had no detected activity with tested ketones
after detergent treatment. Gox1899 is an aldehyde reduc-
tase that can efficiently catalyze the reduction of alde-
hydes while Gox2181 can oxidize polyol as polyol
dehydrogenase. Both of them showed no activity toward
ketones.18,24) Gox0646 had a very narrow substrate
spectrum, exclusively exhibited activities toward dike-
tones rather than ketoester, hydroxy ketone, and aryl
ketone.25) Gox0644 and Gox1615 displayed broad sub-
strate specificities and asymmetric reduction of aliphatic
ketones to produce chiral α-hydroxycarbonyls.26,27)
Furthermore, analysis using a nonredundant Basic
Local Alignment Search Tool (BLAST) in GenBank
and protein databases revealed that Gox2036 shared
50% and 54% of amino acid identities with diacetyl
reductase from Klebsiella terrigena and Corynebacte-
rium glutamicum, respectively.28) Gox0525 exhibited a
similar domain of the 3-ketoacyl-acyl carrier protein
reductase gene (fabG), which was responsible for the
reduction of prochiral ketones. The FabG homolog
from E. coli can reduce ethyl 4-chloroacetoacetate;
meanwhile, the recombinant FabG from Synechococcus
sp. can asymmetrically reduce various prochiral ketones
with good to excellent enantioselectivities.13,29)
BLAST-P analysis revealed that the amino acid
sequences of Gox1462, Gox1598, and Gox0290 had
putative conserved domains of the aldo–keto reductase
superfamily. The analyzed carbonyl reductases shared a
key catalytic tetrad for substrate binding, composed of
four amino acids: N, S, Y, and K residues in Gox0525
and Gox2036, and D, Y, K, and H residues in
Gox0644, Gox1615, Gox1598, Gox 1462, and
Gox0290 (Fig. 1), respectively. These strict conserva-
tive residues presumably formed the framework for a
proton relay system according to previous mutagenetic
and structural studies on reaction mechanism.30,31)
Rather, the specificity and stereoselectivity were proba-
bly defined by the geometry of the active site, resulting
from X-ray crystallographic studies.32) Therefore, the
enzyme set, consisting of Gox2036, Gox0525,
Gox1615, Gox0644, Gox1462, Gox1598, and
Gox0290, was assembled so as to investigate their
substrate selectivity and stereoselectivity toward the
reduction of ketones.
Enantioselective reduction of ketones. The enanti-
oselectivity of the enzymes was determined by
examining the reduction of aryl ketones, ethyl 4-chloro-
acetoacetate (COBE), and ethyl 2-oxo-4-phenylbutyrate
(OPBE) using an NAD(P)H regeneration system con-
sisting of BsGDH and glucose. The general procedure
was as follows: D-glucose (0.5%), recombinant BsGDH
(10 U), NAD(P)+ (0.1 mM), the recombinant cell
(30 g L−1, wet weight), and ketone solvated in ethanol
[1 g L−1, 10%(v/v)] were mixed in a potassium phos-
phate buffer (10 mL, 100 mM, pH 7.0). The mixture
was shaken at 30 °C for 12 h.22) Upon termination of
the reaction, each sample was extracted twice with
equivalent ethyl acetate. The organic layer was
removed, dried, diluted in the mobile phase, and then
subjected to chiral high-performance liquid chromatog-
raphy (HPLC) to determine the conversion and enantio-
meric excess (e.e.). Chiral HPLC analysis was
performed on an Agilent 1100 series HPLC system with
a UV detector.23) Chiral CHBE was analyzed on a chira-
cel OB-H column (Daicel, Japan) at λ210 nm using hex-
ane/2-propanol (90/10, v/v) as eluent at a flow rate of
0.8 mL min−1 and a temperature of 25 °C. Chiral HPBE
and 4-phenyl-2-butanol were analyzed on a chiracel
OD-H column at λ210 nm and λ254 nm using hexane/2-
propanol (98/2, v/v) as eluent at a flow rate of 1.0 mL
min−1 and a temperature of 30 °C. Authentic (relevant)
standards were used for peak identification, and quantifi-
cation was based on the peak area that was suitably
calibrated with standards of known concentration.
Expression and protein purification
These ORFs were amplified from the genomic DNA
of G. oxydans DSM2343, and the amplified DNA frag-
ments were cloned and successfully expressed in
E. coli. The pSE380, pET28a, and pET32a vectors
were used to carry genes to express in E. coli for
obtaining soluble recombinant enzymes. Gox1598 and
Results and discussion
Selection of carbonyl reductases and analysis of the
sequences
According to the genome of G. oxydans 621H, 21
ORFs were predicted to be putative carbonyl reductases

Genome mining of microorganic carbonyl reductases
1353
Fig. 1. Sequence alignment of amino-acid sequences of putative carbonyl reductases.
Notes: The key catalytic tetrad of N, S, Y and K residues found in Gox0525 and Gox2036, and D, Y, K and H residues found in Gox0644,
Gox1615, Gox1598, Gox1462 and Gox0290 are marked with *. CsKR, Ketoreductase from the Cyanobacterium Synechococcus sp. Strain PCC
7942 (ABB56716.1); ScCR, NADH-dependent reductase from the Streptomyces coelicolor (NP_631416.1).
Gox1462 obtained a soluble expression using Rosetta
(DE3) rather than BL21 (DE3) as the expressive host.
The pET32a vector was more suitable for protein
expression in soluble fraction than pET28a when used
to carry the genes of gox1598 and gox1462. Gox0644
was expressed in soluble recombinant protein in BL21
(DE3) by inserting gox0644 into pSE380, whereas the
recombinant pET28a or pET32a carrying Gox0644
formed an inclusion body in BL21 (DE3) during IPTG
induction. The optimal expression conditions were
screened by varying inducer (IPTG) concentrations
(0.1–1 mM) and culture temperatures (25, 30, and
37 °C). Gox2036 and Gox 0525 were induced by 0.5
mM IPTG at 30 °C for 6 h; others were induced by
0.25 mM IPTG at 25 °C overnight. Finally, all enzymes
were successfully expressed in soluble form in E. coli.
The percentage of recombinant target enzymes in crude
cell extract was more than 30% and was quantified by
Bandscan analysis software. The recombinant enzymes
were purified to homogeneity using a Ni–NTA affinity
chromatography for the evaluation of enzymatic
properties (Fig. 2).
Substrate and coenzyme specificity
It is reported that carbonyl reductases had abilities to
asymmetrically reduce carbonyl compounds to produce
optically active alcohols, differing in stereospecificities
and substrate specificities.10) Therefore, the enzymes
were evaluated for their abilities to reduce various
ketones and ketoesters with diverse side groups.
As shown in Table 2, all enzymes showed activities
toward OPBE, ethyl pyruvate, and COBE, all of which
belonged to ketoesters containing α-carbonyl or
Fig. 2. SDS-PAGE analysis of purified enzymes. The purified pro-
teins were resolved by SDS-PAGE on a 12% polyacrylamide gel and
stained with Coomassie Brilliant Blue G-250.
Notes: Lane 1, molecular mass standard; Lane 2, purified Gox2036;
Lane 3, purified Gox0525; Lane 4, purified Gox0644; Lane 5, puri-
fied Gox1615; Lane 6, purified Gox1598; Lane 7, purified Gox1462;
Lane 8, purified Gox0290.
Table 2. Substrate spectrum of carbonyl reductases from G. oxydans.
Gox2036
activity
Gox0525
activity
Gox0644
activity
Gox1615
activity
(U/mg)
Gox1598
activity
Gox1462
activity
Gox0290
activity
Substrate
Acetophenone
1.29
3.53
8.89
8.01
121.2
39.8
2.62
6.7
2.1
0.6
2.0
10.1
32.4
21.5
3.9
2.39
2.48
3.72
70.4
63.4
51.9
NMA
54.6
4.5
NMAa
NMA
NMA
NMA
6.91
15.4
NMA
1.12
NMA
NMA
NMA
NMA
0.34
0.21
1.06
2.89
2.16
NMA
NMA
NMA
NMA
NMA
NMA
4.1
3.0
28.8
5.69
13.7
NMA
NMA
NMA
NMA
4.56
3.60
2.08
3.14
4.06
1-phenyl-2-propanone
4-phenyl-2-butanone
1-phenyl-1,2-propanedione
2,3-butanedione
2,3-pentanedione
Ethyl benzoylacetate
Ethyl pyruvate
OPBE
COBE
Ethyl phenylacetate
3.7
3.44
3.07
NMA
7.15
2.82
NMA
2.31
2.92
NMA
5.71
NMA
0.21
0.20
1.30
2.08
aNMA, no measurable activity.

1354
Table 3. Kinetic parameters.
R. Chen et al.
4-phenyl-2-butanone
COBE
OPBE
kcat
Km
(mM)
kcat
kcat/Km
(M−1 −1
Km
(mM)
kcat
kcat/Km
(M−1 −1
Km
(mM)
kcat/Km
(M−1 −1
s )
Enzyme
(s−1
)
s
)
(s−1
)
s
)
(s−1
)
Gox2036
Gox0525
Gox0644
Gox1615
Gox1598
Gox1462
0.34
1.71
13.6
NMA
NMA
NMA
0.04
0.51
1.02
NMA
NMA
NMA
120
300
70
NMA
NMA
NMA
0.77
1.27
79.01
6.76
1.58
2.30
1.08
5.20
48.68
4.11
1.11
0.83
1.40×103
4.09×103
620
1.47
2.97
0.81
3.73
0.47
2.21
0.42
51.04
1.57
25.69
0.96
290
1.719×104
1.94×103
6.89×103
2.04×103
1.04×103
a
610
2.37×103
360
2.30
aNMA, no measurable activity.
(a)
β-carbonyl groups. Gox2036, Gox0525, and Gox0644
reduced aryl ketones, whereas others showed no activi-
ties. The reduction of aliphatic diketone can be detected
using all enzymes, except Gox1462 (Table 2).
Gox1462 exhibited no detectable activities on the tested
aryl and aliphatic compounds. Gox2036, Gox0525, and
Gox0644 displayed broad substrate acceptances, includ-
ing aryl, aliphatic, α- and β-carbonyl ketones.
Enzyme
con. (%)
e.e. (%)
type
Gox2036
Gox0525
Gox0644
66.8
73.8
8.2
56.0
97.9
100
S
R
S
Kinetic parameters (Km, kcat, and kcat/Km) were
determined for the three important ketones by nonlinear
regression using a Sigma Plot program for each
enzyme (Table 3). Gox0525 showed useful specificity
for 4-phenyl-2-butanone, COBE, and OPBE, with kcat/
Km values of 300, 4.09 × 103, and 1.719 × 104 M−1 s−1,
respectively. Gox2036, Gox0525, and Gox0644 exhib-
ited low-catalytic efficiency for the reduction of 4-phe-
nyl-2-butanone. The apparent Km value of Gox2036
for COBE was 0.77 mM, which is lowest in the test
enzymes. The Km value is similar to that of BYueD
(0.70 mM) from Bacillus sp.ECU001327)33) and much
lower than those of ARII (1.49 mM) from Sporobol-
omyces salmonicolor, S1 (4.6 mM) from Candida mag-
nolia, and PsCRI (4.9 mM) and PsCRII (3.3 mM) from
Pichia stipitis.34) These results demonstrated that the
Gox2036 could reach its maximal reactive velocity at a
lower substrate concentration, which is important
because high concentrations of substrate can destabilize
the enzyme. Although Gox1598 had the lowest Km of
0.47 mM, its kcat was also low at 0.96 s−1. That illus-
trated Gox1598 had the strongest affinity for OPBE.
However, only a few numbers of substrate molecules
turned over per enzyme molecule per second.
(b)
Enzyme
con. (%)
e.e. (%)
type
Gox2036
Gox0525
Gox0644
Gox1615
Gox1598
Gox1462
Gox0290
71.4
73.4
84.9
35.8
25.9
27.1
15.9
>99
>99
21.6
>99
>99
>99
>99
R
S
S
R
R
R
R
(c)
In terms of coenzyme specificity, Gox2036 was
strictly an NADH-dependent enzyme, whereas the oth-
ers were NADPH-dependent enzymes. For industrial
utility, NADH cofactor specificity would be preferable
over NADPH, due to the greater stability and one-tenth
to one-fifth price of the former.35)
Enzyme
con. (%)
e.e. (%)
type
Gox2036
Gox0525
Gox0644
Gox1615
Gox1598
Gox1462
Gox0290
62.3
93.1
49.1
31.5
17.8
15.2
10.9
52.7
>99
R
S
S
R
S
R
S
91.4
61.6
75.0
76.3
40.7
Enantioselective reductions of ketones
BsGDH was successfully expressed in E. coli to
build the cofactor-regenerating system because of its
high activity toward both NAD+ and NADP+.36) In this
study, BsGDH/glucose as an external NAD(P)H regen-
eration system was used to analyze the enantioselectivi-
ty of the enzyme.
Fig. 3. Asymmetric reduction of ketones and ketoesters catalyzed
by carbonyl reductases from G. oxydans in the presence of BsGDH/
glucose with an additionally coupled NAD(P)H regeneration system.
Notes: (a) 4-phenyl-2-butanone; (b) COBE; and (c) OPBE.
The results of three representative reactions using 4-
phenyl-2-butanone (aryl ketone), COBE (β-keto ester),
and OPBE (α-ketoester and aryl group) as the
substrates were summarized in Fig. 3. Gox2036
demonstrated moderate stereoselectivity (S-type, 56%
e.e.) toward 4-phenyl-2-butanone, but excellent
enantioselectivity (approximately 100%) was obtained

Genome mining of microorganic carbonyl reductases
1355
Program, No. 2012CB721003], the Natural Science Founda-
tion of China [No. 21002029/B020706 and 21276084/
B060804] and National Major Science and Technology
Projects of China [No.2012ZX09304009], Natural Science
Foundation of Zhejiang Province LY13C050004, Science and
Technology Development Foundation of Hangzhou
[20101131N05], Technology Research and Development
Program for Institute of Hangzhou [20130432B05].
for the bioreduction of acetophenone and 1-phenyl-2-
propanone.37) Gox0644 produced an (S)-type alcohol
with 100% e.e. and the lowest conversion (8.2%). By
contrast, Gox0525 produced an R-type alcohol with
97.9% e.e. and 73.8% conversion (Fig. 3(a)). All the
tested recombinant enzymes catalyzed COBE with an
excellent e.e. of up to 99%, except Gox0644. R-CHBE
was obtained by Gox2036, Gox1615, Gox1598,
Gox1462, and Gox0290 while Gox0525 and Gox0644
afforded the configuration as S. Gox2036, Gox0525,
and Gox0644 had higher conversions than the other
tested enzymes with regard to CHBE production
(Fig. 3(b)). Excellent e.e. values (>99%) were obtained
for OPBE reduction by Gox0525, whereas other
enzymes showed relatively low e.e. A 93.1% conver-
sion of (S)-HPBE production was obtained by
Gox0525. Gox0644 had high enantioselectivities for
the reduction of 4-phenyl-2-butanone and OPBE (100
and 91.4%); however, it showed relatively low conver-
sions (8.2 and 49.1%)(Fig. 3(c)).
References
[1] Chen YZ, Lie F, Li Z. Enantioselective benzylic hydroxylation
with Pseudomonas monteilii TA-5: a simple method for the syn-
theses of (R)-benzylic alcohols containing reactive functional
groups. Adv. Synth. Catal. 2009;351:2107–2112.
[2] Liese A, Seelbach K, Wandry C. Industrial biotransformations.
New York (NY): Wiley-VCH; 2000. p. 423.
[3] Ye Q, Ouyang P, Ying H. A review—biosynthesis of optically
pure ethyl (S)-4-chloro-3-hydroxybutanoate ester: recent
advances and future perspectives. Appl. Microbiol. Biotechnol.
2011;89:513–522.
[4] Yamamoto H, Matsuyama A, Kobayashi Y. Synthesis of ethyl
(R)-4-chloro-3-hydroxybutanoate with recombinant escherichia
coli cells expressing (S)-specific secondary alcohol dehydroge-
nase. Biosci. Biotechnol. Biochem. 2002;66:481.
[5] Wu XR, Wang YC, Ju JM, Chen C, Liu N, Chen YJ. Enantiose-
lective synthesis of ethyl (S)-2-hydroxy-4-phenylbutyrate by
Conclusions
In this study, we rapidly characterized biocatalysts
based on genome database from G. oxydans. Seven
ORFs were successfully isolated and overexpressed as
a functional recombinant protein in E. coli. All car-
bonyl reductases showed activities toward ethyl pyru-
vate, COBE, and OPBE. Three types of carbonyl
substrates were selected to be asymmetrically reduced
by these recombinant enzymes for their reductive chiral
products, which can be used to synthesize biologically
and pharmacologically important materials. Among the
tested enzymes, Gox2036 had a strict requirement for
NADH as the cofactor and showed excellent enantio-
specificity in synthesis of (R)-CHBE. Gox0525 reduced
tested ketones with high e.e. (from 97.9 to 99%) and
conversion (from 73.4 to 93.1%), even under the unop-
timized experimental conditions. Therefore, these
enzymes had potentials in the synthesis of chiral alco-
hols as useful biocatalysts. It is noted that the Gox0525
exhibited anti-Prelog selectivity in the sense of stereo-
chemistry outcomes.38）
recombinant
diketoreductase.
Tetrahedron:
Asymmetry.
2009;20:2504–2509.
[6] D’Arrigo P, Pedrocchi-Fantoni G, Servi S. Chemo-enzymatic
synthesis of ethyl (R)-2-hydroxy-4-phenylbutyrate. Tetrahedron:
Asymmetry. 2010;21:914–918.
[7] Nakamura K, Yamanaka R, Matsuda T, Harada T. Recent devel-
opments in asymmetric reduction of ketones with biocatalysts.
Tetrahedron: Asymmetry. 2003;14:2659–2681.
[8] Zheng GW, Xu JH. New opportunities for biocatalysis: driving
the synthesis of chiral chemicals. Curr. Opin. Biotechnol.
2011;22:784–792.
[9] Woodley JM. New opportunities for biocatalysis: making phar-
maceutical processes greener. Trends Biotechnol. 2008;26:
321–327.
[10] Huisman GW, Liang J, Krebber A. Practical chiral alcohol man-
ufacture using ketoreductases. Curr. Opin. Chem. Biol.
2010;14:122–129.
[11] Forrest GL, Gonzalez B. Carbonyl reductase. Chem-Biol. Inter-
act. 2000;129:21–40.
[12] Ferrer M, Martínez-Abarca F, Golyshin PN. Mining genomes
and ‘metagenomes’ for novel catalysts. Curr. Opin. Biotechnol.
2005;16:588–593.
[13] Yamamoto H, Matsuyama A, Kobayashi Y. Synthesis of ethyl
(S)-4-chloro-3-hydroxybutanoate using fabG-homologues. Appl.
Microbiol. Biotechnol. 2003;61:133–139.
Author contribution
D. Wei and J. Lin designed the study. R. Chen car-
ried out the bulk of the experiments. X. Liu contributed
to the expression and purification of the reductase. All
authors have read and approved the final manuscript.
[14] Rabenhorst J, Gatfield I, Hilmer JM. Natural aliphatic and
thiocarboxylic acids obtainable by fermentation and a microor
ganism therefore. European Patent 1078990. 2000 Feb. 28.
[15] Gao KL, Wei DZ. Asymmetric oxidation by Gluconobacter
oxydans. Appl. Microbiol. Biotech. 2006;70:135–139.
[16] Prust C, Hoffmeister M, Liesegang H, Wiezer A, Fricke WF,
Ehrenreich A, Gottschalk G, Deppenmeier U. Complete genome
sequence of the acetic acid bacterium Gluconobacter oxydans.
Nat. Biotechnol. 2005;23:195–200.
[17] Richter N, Neumann M, Liese A, Wohlgemuth R, Weckbecker
A, Eggert T, Hummel W. Characterization of a whole-cell cata-
lyst co-expressing glycerol dehydrogenase and glucose dehydro-
genase and its application in the synthesis of L-glyceraldehyde.
Biotechnol. Bioeng. 2010;106:541–552.
Supplemental material
The Supplemental material for this paper is available
at http://dx.doi.org/10.1080/09168451.2014.925775.
Funding
[18] Liu X, Yuan Z. Biochemical and structural analysis of Gox2181,
a new member of the SDR superfamily from Gluconobacter
oxydans. Biochem. Biophys. Res. Commun. 2011;415:410–415.
This work was financially supported by the National Key
Basic Research Development Program of China [“973”

1356
R. Chen et al.
[19] Ausubel FM. Preparation and analysis of genomic DNA from
bacteria. In: Ausubel FM, Brent R, Kingston RE, Moore DD,
Seidman JG, Struhl K, editors. Current protocols in molecular
biology. New York (NY): Wiley; 2002. p. 2–11.
[20] Xu Z, Jing K, Liu Y, Cen P. High-level expression of recombi-
nant glucose dehydrogenase and its application in NADPH
regeneration. J. Ind. Microbiol. Biotech. 2007;34:83–90.
Cyanobacterium Synechococcus sp. strain PCC 7942. Appl.
Environ. Microbiol. 2008;74:6697–6702.
[30] Kallberg Y, Oppermann U, Jörnvall H, Persson B. Short-chain
dehydrogenases/reductases (SDRs). Eur. J. Biochem. 2002;269:
4409–4417.
[31] Oppermann U, Filling C, Hult M, Shafqat N, Wu X, Lindh M,
Shafqat J, Nordling E, Kallberg Y, Persson B, Jörnvall H. Short-
chain dehydrogenases/reductases (SDR): the 2002 update.
Chem.-Biol. Interact. 2003;143–144:247–253.
[32] Kwiecień RA, Ayadi F, Nemmaoui Y, Silvestre V, Zhang BL,
Robins RJ. Probing stereoselectivity and pro-chirality of hydride
transfer during short-chain alcohol dehydrogenase activity: A
combined quantitative 2H NMR and computational approach.
Arch. Biochem. Biophys. 2009;482:42–51.
[21] Sambrook J, Russell DW. Molecular cloning:
a laboratory
manual. 3rd ed. Cold Spring Harbor (NY): Cold Spring Harbor
Laboratory Press; 2001.
[22] Jung J, Park S. Synthesis of a chiral alcohol using a rationally
designed Saccharomyces cerevisiae reductase and
a NADH
cofactor regeneration system. J. Mol. Catal. B: Enzym.
2012;84:15–21.
[23] Nie Y, Xiao R, Xu Y, Montelione GT. Novel anti-Prelog stereo-
specific carbonyl reductases from Candida parapsilosis for
asymmetric reduction of prochiral ketones. Org. Biomol. Chem.
2011;9:4070–4078.
[33] Ni Y, Li CX, Wang LJ, Zhang J, Xu JH. Highly stereoselective
reduction of prochiral ketones by a bacterial reductase coupled
with cofactor regeneration. Org. Biomol. Chem. 2011;9:
5463–5468.
[24] Schweiger P, Deppenmeier U. Analysis of aldehyde reductases
from Gluconobacter oxydans 621H. Appl. Microbiol. Biotech-
nol. 2010;85:1025–1031.
[25] Schweiger P, Gross H, Zeiser J, Deppenmeier U. Asymmetric
reduction of diketones by two Gluconobacter oxydans oxidore-
ductases. Appl. Microbiol. Biotechnol. 2013;97:3475–3484.
[26] Chen M, Lin J, Ma Y, Wei D. Characterization of a novel
NADPH-dependent oxidoreductase from Gluconobacter oxydans.
Mol. Biotechnol. 2010;46:176–181.
[34] Wang Q, Shen LH, Ye T, Cao D, Chen R, Pei X, Xie T,
Li Y, Gong W, Yin X. Overexpression and characterization of a
novel (S)-specific extended short-chain dehydrogenase/reductase
from Candida parapsilosis. Bioresour. Technol. 2012;
123:690–694.
[35] Machielsen R, Looger LL, Raedts J, Dijkhuizen S, Hummel W,
Hennemann H-G, Daussmann T, van der Oost J. Cofactor
engineering of Lactobacillus brevis alcohol dehydrogenase by
computational design. Eng. Life Sci. 2009;9:38–44.
[27] Schweiger P, Gross H, Deppenmeier U. Characterization of two
aldo–keto reductases from Gluconobacter oxydans 621H capable
of regio- and stereoselective α-ketocarbonyl reduction. Appl.
Microbiol. Biotechnol. 2010;87:1415–1426.
[28] Blomqvist K, Nikkola M, Lehtovaara P, Suihko ML, Airaksinen
U, Stråby KB, Knowles JK, Penttilä ME. Characterization of the
genes of the 2,3-butanediol operons from Klebsiella terrigena
and Enterobacter aerogenes. J. Bacteriol. 1993;175:1392–1404.
[29] Holsch K, Havel J, Haslbeck M, Weuster-Botz D. Identification,
cloning, and characterization of a novel ketoreductase from the
[36] Xiao Z, Lv C, Gao C, Qin J, Ma C, Liu Z, Liu P, Li L, Xu P.
A novel whole-cell biocatalyst with NAD+ regeneration for
production of chiral chemicals. PLOS one. 2010;5:e8860.
[37] Liu X, Chen R, Yang Z, Wang J, Lin J, Wei D. Characterization
of a putative stereoselective oxidoreductase from Gluconobacter
oxydans and Its application in producing ethyl (R)-4-chloro-3-
hydroxybutanoate ester. Mol. Biotechnol. 2014;56:285–295.
[38] Musa MM, Phillips RS. Recent advances in alcohol dehydroge-
nase-catalyzed asymmetric production of hydrophobic alcohols.
Catal. Sci. Technol. 2011;1:1311–1323.