Production of (R)-3-quinuclidinol by a whole-cell biocatalyst with high efficiency

Article abstract of DOI:10.1080/10242422.2017.1400019

Optically pure (R)-3-quinuclidinol [(R)-3-Qui] is widely used as a chiral building block for producing various antimuscarinic agents. An asymmetric bioreduction approach using 3-quinuclidinone reductases is an effective way to produce (R)-3-Qui. In this study, a biocatalyst for producing (R)-3-Qui was developed by using Escherichia coli that coexpressed Kaistia granuli (KgQR) and mutant glucose dehydrogenase (GDH). KgQR catalyses the synthesis of (R)-3-Qui through the efficient reduction of 3-quinuclidinone. The specific activity of recombinant KgQR was 254 U/mg, and the Michaelis–Menten constant (K_m) for 3-quinuclidinone was 0.51 mM. The thermal stability of KgQR was relatively high compared with ArQR. Approximately 73% of the residual activity remained after incubation in 0.2?M potassium phosphate buffer (KPB) (pH 7.0) for 8 h at 30 °C. In addition, 80% residual activity remained for the double-mutant GDH (Q252L and E170K) after incubation in a buffer (pH 7.0) for 8 h at 30 and 40 °C. 3-Quinuclidinone (242 g/L) can be reduced to (R)-3-Qui in 3 h by coexpressing KgQR and mutant GDH in E. coli. The conversion rate reached 80.6 g/L/h, which is the highest reported to date. The results demonstrates that this whole-cell biocatalyst will have a great potential in industrial manufacturing.

Full text of DOI:10.1080/10242422.2017.1400019

Biocatalysis and Biotransformation
ISSN: 1024-2422 (Print) 1029-2446 (Online) Journal homepage: http://www.tandfonline.com/loi/ibab20
Production of (R)-3-quinuclidinol by a whole-cell
biocatalyst with high efficiency
Zhenhua Jia, Hong Ma, Yali Huang, Yuanyuan Huang, Pengju Ren, Shuishan
Song, Meirong Hu & Yong Tao
To cite this article: Zhenhua Jia, Hong Ma, Yali Huang, Yuanyuan Huang, Pengju Ren, Shuishan
Song, Meirong Hu & Yong Tao (2017): Production of (R)-3-quinuclidinol by a whole-cell biocatalyst
with high efficiency, Biocatalysis and Biotransformation, DOI: 10.1080/10242422.2017.1400019
To link to this article: http://dx.doi.org/10.1080/10242422.2017.1400019
View supplementary material
Published online: 10 Nov 2017.
Submit your article to this journal
Article views: 4
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=ibab20
Download by: [Gothenburg University Library]
Date: 20 November 2017, At: 00:31

BIOCATALYSIS AND BIOTRANSFORMATION, 2017
https://doi.org/10.1080/10242422.2017.1400019
RESEARCH ARTICLE
Production of (R)-3-quinuclidinol by a whole-cell biocatalyst with
high efficiency
a
a
a
a
a
a
ꢀ
ꢀ
Zhenhua Jia , Hong Ma , Yali Huang , Yuanyuan Huang , Pengju Ren , Shuishan Song ,
Meirong Hub and Yong Taob
b
aBiology Institute, Hebei Academy of sciences, Shijiazhuang, P. R. China; CAS Key Laboratory of Microbial Physiological and
Metabolic Engineering, Institute of Microbiology, Chinese Academy of Sciences, Beijing, P.R. China
ABSTRACT
ARTICLE HISTORY
Received 21 April 2017
Revised 7 October 2017
Accepted 26 October 2017
Optically pure (R)-3-quinuclidinol [(R)-3-Qui] is widely used as a chiral building block for produc-
ing various antimuscarinic agents. An asymmetric bioreduction approach using 3-quinuclidinone
reductases is an effective way to produce (R)-3-Qui. In this study, a biocatalyst for producing (R)-
3-Qui was developed by using Escherichia coli that coexpressed Kaistia granuli (KgQR) and
mutant glucose dehydrogenase (GDH). KgQR catalyses the synthesis of (R)-3-Qui through the effi-
cient reduction of 3-quinuclidinone. The specific activity of recombinant KgQR was 254 U/mg,
and the Michaelis–Menten constant (Km) for 3-quinuclidinone was 0.51 mM. The thermal stability
of KgQR was relatively high compared with ArQR. Approximately 73% of the residual activity
remained after incubation in 0.2 M potassium phosphate buffer (KPB) (pH 7.0) for 8 h at 30 ꢁC. In
addition, 80% residual activity remained for the double-mutant GDH (Q252L and E170K) after
incubation in a buffer (pH 7.0) for 8 h at 30 and 40 ꢁC. 3-Quinuclidinone (242 g/L) can be reduced
to (R)-3-Qui in 3 h by coexpressing KgQR and mutant GDH in E. coli. The conversion rate reached
80.6 g/L/h, which is the highest reported to date. The results demonstrates that this whole-cell
biocatalyst will have a great potential in industrial manufacturing.
KEYWORDS
Whole-cell biocatalyst;
(R)-3-quinuclidinol; 3-quinu-
clidinone reductase; Kaistia
granuli; glucose
dehydrogenase;
3-quinuclidinone
1. Introduction
catalysts are unresolved difficulties in the pharmaceut-
ical industry (Tsutsumi et al. 2009). Kinetic resolution
of a racemic mixture of n-butyl-3-Qui with protease
has also been used to produce (R)-3-Qui (Tsutsumi
et al. 2009). After 24 h of reaction, (R)-3-Qui is accessed
with an overall yield of 42% and an optical purity of
96% ee (Ikunaka 2004). However, the theoretical yield
of the product is only 50%, when using the racemic
substrate, which limits the industrial application. In
recent years, more studies have been focusing on
asymmetric bioreduction as a method to produce (R)-
3-Qui because asymmetric reductions can generate
highly optically pure products at a relatively low cost,
and the process is environmentally friendly.
An asymmetric bioreduction approach was used to
produce (R)-3-Qui using 3-quinuclidinone reductases.
For catalysing the synthesis of (R)-3-Qui, 3-quinuclidi-
none reductase uses the substrate 3-quinuclidinone
and requires cofactor NAD(P)H. To accelerate the bio-
reduction process, it is necessary to regenerate
NAD(P)H from NAD(P)þ. There have been many reports
on the regeneration of NAD(P)H via coupling reactions
Chiral compounds are essential building blocks of
drugs with a market share of nearly 50% (Hoelsch
et al. 2013; Karsten et al. 2015; Weyler and Heinzle
2015; Willrodt et al. 2017). Optically, pure Qui is an
important chiral pharmaceutical intermediate. Different
isomers of Qui [(R)-3-Qui or (S)-3-Qui] are widely used
in different fields (Zhang et al. 2016). Optically pure
(R)-3-Qui is widely used as a chiral building block for
producing various antimuscarinic agents such as talsa-
clidine, solifenacin, aclidinium bromide, and revatro-
pate (Leusch et al. 2000; Ishihara et al. 2004; Prat et al.
2009). For the production of (R)-3-Qui, chemical syn-
thesis, enzymatic resolution, and asymmetric bioreduc-
tion have been developed. Chemical synthesis is one
of the most efficient routes to produce (R)-3-quinculi-
dinol. A previous study demonstrated that the chiral
metal catalyst RuBr2 [(S,S)-xylskewphos] has been used
to synthesize (R)-3-Qui with 88–90% enantiomeric
excess (ee); however, the trace metal contamination
left in the product and the high cost of chiral metal
CONTACT Meirong Hu
hmr1211@126.com
CAS Key Laboratory of Microbial Physiological and Metabolic Engineering, Institute of Microbiology,
Chinese Academy of Sciences, Beijing, P.R. China
ꢀ
These authors contributed equally to this work.
Supplemental data for this article can be accessed here.
ß 2017 Informa UK Limited, trading as Taylor & Francis Group

2
Z. JIA ET AL.
using alcohol dehydrogenase (ADH) (Machielsen et al.
2009; Soni et al. 2015; van Rossum et al. 2016; Solanki
et al. 2017), glucose/glucose dehydrogenase (GDH)
(Nagao et al. 1992; Ramadas et al. 2013; Takeshita
et al. 2014) and formate/formate dehydrogenase
(Hoelsch et al. 2013). Many researchers have cloned
GDH for the bioreduction reaction to regenerate
NAD(P)H (Nagao et al. 1992; Kataoka et al. 1998;
Schewe et al. 2008; Pham et al. 2013). However, the
stability of these GDHs has been poor, and many stud-
ies have focused on increasing their thermal stability
(Makino et al. 1989; Baik et al. 2003; Pire et al. 2001,
2004; Ding et al. 2013; Liang et al. 2013). The crystal
structure of GDH (GlcDH) from Bacillus megaterium
IWG3 has been determined (Yamamoto et al. 2001). It
is a tetrameric protein composed of four identical sub-
units. The tetramer could rapidly dissociate into
inactive monomers under high temperatures and alka-
line conditions. Therefore, the poor stability of GDH
from Bacillus sp. is the bottleneck in its practical indus-
trial applications. Previous studies have indicated that
introducing Q252L and E170K mutation sites can
increase the stability of GDH (Baik et al. 2005).
whole-cell biocatalyst was constructed to produce (R)-
3-Qui by coexpressing KgQR and double-mutant GDH.
2. Material and methods
2.1. Strains, reagents, and enzymes
Escherichia coli DH5a was used for DNA manipulation.
Escherichia coli K12/BW25113 (Keio collection, National
BioResource Project NIG, Tokyo, Japan) was used for
protein expression. Bacillus megaterium ATCC 14581
was obtained from ATCC. Vector pBAD–hisB
(Invitrogen, Carlsbad, CA) was used for gene expres-
sion. 3-Quinuclidinone hydrochloride was obtained
from Sigma-Aldrich Chemical Co. (Milwaukee, WI). (R)-
3-Qui and (S) 3-Qui were purchased from Santa Cruz
Biotechnology, Inc. (Santa Cruz, CA). Other commercial
reagents were purchased in the highest purity avail-
able. All restriction enzymes, Pfu DNA polymerase, T4
DNA ligase, and Gibson AssemblyTM Master Mix were
obtained from New England Biolabs (Ipswich, UK).
2.2. Construction of expression plasmids
A number of enzymes have been reported as asym-
metric reductases to produce (R)-3-Qui, including RrQR
from Rhodotorula rubra, QNR and BacC from
Microbacterium luteolum JCM 9174, and ArQR from
Agrobacterium radiobacter ECU2556 (Uzura et al. 2009;
Isotani et al. 2012; Isotani et al. 2013; Zhang et al.
2013). Among these, ArQR is the most efficient
enzyme. Its specific activity using 3-quinuclidinone as
a substrate was 200 U/mg, and a high yield of
(R)-3-Qui (up to 916 g/L/day) was obtained using ArQR.
However, the thermal stability of ArQR is poor.
Therefore, there is a need for 3-quinuclidinone reduc-
tases with both high thermal stability and high sub-
strate-binding affinity. Hou et al. (2012, 2014) reported
that 3-quinuclidinone reductase from A. tumefaciens
(AtQR) have high substrate-binding affinity and enan-
tioselectivity. However, the synthesis of (R)-3-Qui using
AtQR has not been studied and fully evaluated. In the
present study, 3-quinuclidinone reductase from Kaistia
granuli (KgQR), which possesses 95% sequence similar-
ity with AtQR (GenBank:3AK4_A), was expressed in
Escherichia coli. According to above evidence, we
hypothesized that KgQR produces (R)-3-Qui from 3-qui-
nuclidinone with high efficiency and enantioselectivity.
In the present study, GDH from B. megaterium was
cloned and used to regenerate NAD(P)H. Double-
mutant Q252L and E170K GDH was constructed to
produce (R)-3-Qui more efficiently. KgQR was
expressed, purified, and characterized. In addition, a
Standard methods were used for genomic DNA and
plasmid DNA extraction, plasmid construction,
and transformation (Sambrook and Russell 2001).
Five plasmids, pBAD–hisB–KgQR, pBAD–hisB–ArQR,
pBAD–hisB–BmGDH,
pBAD–hisB–Q252LE170K,
and
pBAD–hisB–QG, were constructed. pBAD–hisB–KgQR
was constructed as follows: the gene encoding KgQR
was synthesized on the basis of the protein sequence
of 3-ketoacyl-ACP reductase (GenBank: WP_029076252)
in GenScript USA Inc. (Township, NJ). The nucleotide
sequence of KgQR has been deposited in the GenBank
databases (GenBank: AB733448). The gene fragment
encoding KgQR with restriction enzymes XhoI and PstI
was inserted into pBAD–hisB at the same enzyme sites
to generate pBAD–hisB–KgQR. pBAD/hisB–ArQR was
constructed as follows: the gene encoding ArQR
(GenBank: YP_002542435.1), which was synthesized in
GenScript USA Inc., was inserted into pBAD–hisB at the
sites of restriction enzymes XhoI and PstI to form
pBAD–hisB–ArQR. pBAD–hisB–BmGDH was constructed
as follows: the gene encoding BmGDH (GenBank:
CP009920.1) was amplified by PCR using primers P1
and P2 with genomic DNA of B. megaterium
ATCC 14581 as the template and inserted
into pBAD–hisB at the sites of XhoI and PstI to
generate pBAD–hisB–BmGDH. To construct pBAD–hisB–
Q252LE170K, PCR was performed using the primers
shown in Supplementary Table S1 (P3 and P4, and P5
and P6 in succession) with pBAD–hisB–BmGDH as

BIOCATALYSIS AND BIOTRANSFORMATION
3
a template. The PCR products were digested with DpnI
and transformed in E. coli DH5a to generate mutants.
To construct pBAD–hisB–QG, which was used to reduce
3-quinuclidinone to (R)-3-Qui, the gene fragment
encoding Q252LE170K was amplified by PCR using the
primers P7 and P8 with pBAD–hisB–Q252LE170K encod-
ing Q252LE170K was amplified by PCR using P7 and P8
with Pbad–hisBq as a template and then assembled
with pBAD–hisB–KgQR digested with PstI using Gibson
AssemblyTM Master Mix to generate pBAD–hisB–QG.
After DNA sequencing, the appropriate plasmids were
used to transform E. coli K12/BW25113 to express
recombinant enzymes or as a whole-cell biocatalyst to
produce (R)-3-Qui.
NADH, 0.2 M KPB (pH 7.0), and 10 lL enzyme solu-
tion in a total volume of 1.0 mL. One unit of enzyme
activity was defined as the amount of enzyme that
catalyses the oxidation of 1 lmol of NADH/min under
these conditions. The activity of BmGDH was calcu-
lated by measuring the increase in absorbance of
NADH at 340 nm. The reaction mixture contained
2 mM NADþ, 0.1 M glucose, 0.2 M KPB (pH 7.0), and
an appropriate amount of enzyme in a total volume
of 1 mL. One unit of enzyme activity was defined as
the amount of enzyme that catalyses the generation
of 1 lmol NADH/min under these conditions. Km val-
ues were obtained by assaying 3-quinuclidinone
reductase activity in a range of substrate concentra-
tions from 0.1 to 5.0 mM for 3-quinuclidinone and
from 0.01 to 0.25 mM for NADH. The resulting curves
were fitted to the Michaelis–Menten equation by
non-linear least squares regression using Sigma Plot
11.0 (Systat Software, Chicago, IL) (Km). The activity
for each concentration was assayed in triplicate,
resulting in a mean value for each substrate concen-
tration. Data are the mean value of three independ-
ent measurements.
2.3. Cloning, expression, and purification
Recombinant BW25113 cells harboring expression vec-
tors were cultured in 200 mL of 2YT broth (1.6% tryp-
tone, 1.0% yeast extract, 0.5% NaCl) containing
100 lg/mL ampicillin. Following this, 0.2% arabinose
was added until the OD600 reached 0.6, and the solu-
tion was incubated at 25 ꢁC for further 13 h. The cells
were then harvested by centrifugation (7000g for
10 min at 4 ꢁC). The expressed KgQR, ArQR, BmGDH,
and Q252LE170K with N-terminal 6ꢂ His-tag were
purified using an immobilized nickel-ion chromatog-
2.5. Thermal stability of KgQR, ArQR, BmGDH, and
the Q252LE170K mutant
€
The thermal stability of KgQR, ArQR, BmGDH, and the
Q252LE170K mutant was determined by incubating
the purified recombinant enzymes at different temper-
atures (30 ꢁC for KgQR and ArQR; 30 and 40 ꢁC for
BmGDH and the Q252LE170K mutant GDH, respect-
ively) in 0.2 M KPB (pH 7.0) for 8 h. Samples were col-
lected at different time intervals, and the residual
activity was assayed under standard conditions. Data
are the mean value of three independent measure-
ments with standard deviations.
raphy using an AKTA FPLC system GE Healthcare
(Piscataway, NJ) equipped with a HisTrap HP column,
according to the manufacturer’s instructions. The
bound recombinant KgQR, ArQR, BmGDH, and
Q252LE170K were eluted with an elution buffer
[50 mM NaH2PO4, 300 mM NaCl, and 250 mM imidazole
(pH 7.0)], and the elution fraction containing the
recombinant enzymes was dialysed with 20 mM
sodium phosphate buffer (pH 7.4) to remove excess
imidazole. The purified recombinant enzymes were
confirmed using sodium dodecyl sulphate–polyacryla-
mide gel electrophoresis (SDS–PAGE) electrophoresis
containing a 10% (w/v) acrylamide gel. Protein concen-
trations were determined using the method of
Bradford (1976) with bovine serum albumin as a stand-
ard. Purified recombinant enzymes were used to per-
form enzyme characterization assays.
2.6. Product analysis by GC
The enantiomeric purity of (R)-3-Qui was analysed
using
a GC system (HP 6890; Hewlett Packard,
Palo Alto, CA) equipped with a chiral capillary column
(CP-cyclodextrin-N-236-N19, 0.25 mm ꢂ 25 m; Varian
Medical Systems, Palo Alto, CA) with a flame ionization
detector. The GC conditions were as follows: the col-
umn temperature program was increased to 120 ꢁC,
maintained for 10 min, and then ramped from 120 to
150 ꢁC at 0.5 ꢁC/min. The injection and detection tem-
peratures were 220 ꢁC. The retention times were
14.28 min for 3-quinuclidinone, 32.17 min for (S)-3-Qui,
and 33.249 min for (R)-3-Qui.
2.4. Enzyme activity and kinetic parameters assay
A
spectrophotometric assay of 3-quinuclidinone
reductase activity was performed by measuring the
decrease in absorbance of NADH at 340 nm
(e ¼ 6.22 mMcmꢃ1). The reaction mixture consisted
of 200 lM 3-quinuclidinone hydrochloride, 0.1 mM

4
Z. JIA ET AL.
GDHR (Supplementary Table S1). Three expression vec-
tors, pBAD–KgQR, pBAD–ArQR, and pBAD–BmGDH,
which encoded KgQR, ArQR, and BmGDH, respectively,
were constructed. pBAD–hisB–Q252LE170K was con-
structed on the basis of pBAD–BmGDH (see “Material
and Methods” section for details). ORFs for these
genes were fused with an N-terminal 6 ꢂ His-tag.
Recombinant BW25113 cells harboring expression vec-
tors were cultured at 37 ꢁC until the OD600 was 0.6.
Then, the expression of recombinant cells were
induced by 0.2% arabinose. KgQR, ArQR, BmGDH, and
mutant GDH were purified by Ni–NTA agarose column
chromatography. Homogeneous purification was con-
firmed by SDS–PAGE (Figure 1(B,C)). SDS–PAGE
revealed that the molecular masses of recombinant
ArQR, KgQR, and BmGDH were 28, 25, and 37 kDa,
respectively. SDS–PAGE of GDH mutant (not shown)
was the same as that of BmGDH. The four enzymes
expressed in E. coli were soluble.
2.7. Conversion of 3-quinuclidinone to (R)-3-Qui
using a whole-cell biocatalyst
To construct the highly efficient whole-cell biocatalyst,
the plasmid pBAD–hisB–QG containing KgQR and
the Q252LE170K mutant was used to transform E. coli
K12/BW25113 to generate E. coli K12/BW25113
(pBAD–hisB–QG). The 3-quinuclidinone conversion
reaction was performed using the E. coli whole-cell
biocatalyst. The reaction mixture consisted of 65 mL of
200 mM pH 7.0 KPB, 2 M 3-quinuclidinone hydrochlor-
ide (242 g/L), 2 M glucose (360 g/L), 0.1 mM NADþ, and
0.65 g of lyophilized cells of E. coli K12/BW25113
(pBAD–hisB–QG). The reaction mixture was stirred at
30 ꢁC with pH adjustment to 7.0 by the automatic add-
ition of a 2 M NaOH solution. The reaction mixture
was sampled at appropriate intervals, and the produc-
tion of (R)-3-Qui was analysed. Samples were alkalified
with NaOH and extracted twice with an equal volume
of ethyl acetate. The collected organic phases were
combined and analysed by GC.
3.3. Characterization of recombinant KgQR, ArQR,
BmGDH, and the Q252LE170K mutant
3. Results
The Km values of purified recombinant KgQR and
ArQR for 3-quinuclidinone in the reductive reaction
were 0.52 and 0.91 mM, respectively, from the
Hanes–Woolf plot. The Km value of purified KgQR for
3-quinuclidinone was lower than that of ArQR. It indi-
cated that the reaction velocity will be high even
with a low concentration of 3-quinuclidinone, as for
KgQR. The Km values of KgQR and ArQR for NADH in
the reductive reaction were 0.03 and 0.028 mM
(Table 1), respectively. The specific activities of KgQR
and ArQR were 254 and 200 U/mg, respectively. The
specific activities of BmGDH and the Q252LE170K
mutant were 356.9 and 350.0 U/mg (Table 1), respect-
ively. The effect of pH on recombinant KgQR, ArQR,
BmGDH, and the Q252LE170K mutant was measured
using various pH buffers; our results indicated that
the maximum activity of these enzymes was at
pH 7.0.
3.1. Sequence analysis of 3-quinuclidinone-
reducing enzymes
3-Quinuclidinone reductase from KgQR (GenBank:
WP_029076252) was isolated on the basis of the
assumption that this enzyme would possess 95%
sequence similarity with AtQR. Sequence analysis indi-
cated that KgQR belonged to the short-chain dehydro-
genases/reductases (SDR) family. The enzyme consisted
of 260 amino acid residues with a calculated molecular
mass of 27437.55 Da. The sequence of 3-quinuclidinone
reductase displayed moderate identities with the
NADH-dependent 3-quinuclidinone reductase ArQR
(ꢄ67%) (GenBank: YP_002542435.1) (Figure 1(A)). It also
showed a low identity with NADPH-dependent 3-quinu-
clidinone reductase from R. rubra (25.61%) (GenBank:
BAH28833.1) (Figure 1(A)). The results indicated that
KgQR be probably NADH-dependent 3-quinuclidinone
reductase with high specific activity.
3.4. Thermal stability of KgQR, ArQR, BmGDH, and
the Q252LE170K mutant
3.2. Gene cloning, enzyme expression, and
purification
The thermal stability of KgQR, ArQR, BmGDH, and the
Q252LE170K mutant was determined using purified
recombinant enzymes (Figure 2(A,B)). The results indi-
cated that the thermal stability of KgQR was relatively
high and a residual activity of ꢄ73% remained after
incubation in 0.2 M KPB (pH 7.0) for 8 h at 30 ꢁC. In
contrast, the residual activity of ArQR was only about
25% (Figure 2(A)). The thermal stability of BmGDH was
The gene encoding KgQR and ArQR was synthesized
on the basis of the codon usage of E. coli. It has been
submitted to GenBank (The GenBank numbers of
KgQR and ArQR were KT228249 and KT228250,
respectively). BmGDH (GenBank: CP009920.1) was
amplified using the genome of B. megaterium ATCC
14581 as a template with the primers GDHF and

BIOCATALYSIS AND BIOTRANSFORMATION
5
Figure 1. Comparison of amino acid sequences and SDS–PAGE analysis of purified 3-quinuclidinone reductases from different
organisms K. granuli and those of SDR superfamily proteins. (A) KgQR (GenBank: WP_018183582.1) from Kaistia granuli, AT3AK4
(GenBank: 3AK4_A) from Agrobacterium tumefaciens, ArQR (GenBank: YP_002542435.1) from A. radiobacter ECU2556, and RrQR
(GenBank: BAH28833.1) from R. rubra. Multiple sequence alignment was performed using the DNAMAN program. (B) SDS–PAGE
analysis of recombinant KgQR, and ArQR. M: Mark; 1: E. coli K12/pBAD–KgQR, E. coli K12/pBAD–ArQR; 2: the purified recombinant
enzymes KgQR, ArQR. (C) SDS–PAGE analysis of recombinant BmGDH. M: Mark; 1: E. coli K12/pBAD–BmGDH; 2: the purified recom-
binant enzyme BmGDH.
Table 1. Characterization of recombinant KgQR, ArQR, BmGDH, and the Q252LE170K mutant.
3-Quinuclidinone
Km (mM)
Specific activity
(U/mg)
Stereoselectivity for
(R)-3-quinuclidinol (%)
Property
NADH Km (mM)
Optimum pH
KgQR
ArQR
BmGDH
Q252LE170K mutant
0.030
0.028
0.52
0.91
251.41
200.23
356.95
350.00
>99.9
>99.9
7–7.5
7
7
7
very low, and 10% residual activity remained after
incubation in buffer (pH 7.0) for 1.5 h at 30 ꢁC. The
Q252LE170K mutant showed good thermal stability,
and 80% residual activity after incubation in buffer
(pH 7.0) for 8 h at both 30 and 40 ꢁC (Figure 2(B)).
3.5. Production of (R)-3-Qui using a whole-cell
biocatalyst containing KgQR and the Q252LE170K
mutant in E. coli cells
Highly efficient production of (R)-3-Qui was obtained
using coexpressed 3-quinuclidinone reductase and

6
Z. JIA ET AL.
Figure 3. Time course of production of (R)-3-Qui by E. coli
biocatalyst containing pBAD–hisB–QG. The concentration of
(R)-3-Qui is represented by a solid circle and that of 3-quinucli-
dinone by a solid square. Data are the mean value of three
independent measurements with standard deviation.
from A. radiobacter ECU2556 (Zhang et al. 2013). In
the present study, a novel and efficient whole-cell bio-
catalyst was constructed to produce (R)-3-Qui. Using E.
coli K12/BW25113 (pBAD–hisB–QG) coexpressing the 3-
quinuclidinone reductase and double-mutant GDH,
27thinsp;M (242 g/L) 3-quinuclidinone hydrochloride
was stereospecifically converted to (R)-3-Qui with 99%
conversion in 3 h (Figure 3). The conversion rate
reached 80.6 g/L/h, which is higher than that reported
by Zhang et al. (2013). In that study, 2 M (242 g/L)
3-quinuclidinone hydrochloride was completely con-
verted to (R)-3-Qui with 99% conversion and >99%
ee within 4.5 h, and the conversion rate was only
53.7 g/L/h.
The efficient biocatalyst system in the present study
was based on a 3-quinuclidinone reductase with good
catalytic features and GDH with high specific activity
and thermal stability. KgQR can be widely used for the
production of (R)-3-Qui because of its catalytic feature.
The Km value (0.51 mM) of KgQR for 3-quinuclidinone
in the reductive reaction was much lower than that of
RrQR (145 mM) (Uzura et al. 2009) and even lower
than that of ArQR (0.91 mM) (Table 1). The specific
activity of KgQR (254 U/mg) was higher than that of
ArQR (200 U/mg). Furthermore, the thermal stability
of KgQR was better than that of ArQR. The activity of
KgQR was relatively high, and up to 75% residual
activity remained after incubation in 0.2 M KPB buffer
(pH 7.0) for 8 h at 30 ꢁC. These results also imply that
the conformation of KgQR is more favourable to the
catalytic reaction than that of ArQR. KgQR is an NADH-
dependent 3-quinuclidinone reductase. For the large-
scale production of (R)-3-Qui, NADþ, which is less
expensive than NADPþ, can be added to the reaction
buffer. Therefore, NADH-dependent quinuclidinone
reductase is more suitable for the production of
(R)-3-Qui. The Q252LE170K mutant also had good
Figure 2. Thermostability of 3-quinuclidinone reductases. (A)
Thermostability of ArQR and KgQR. (B) Thermostability of WT
(BmGDH) and Q252LE170K at 30 and 40 ꢁC, respectively.
GDH with good catalytic properties. Our results indi-
cated that the specific activity of KgQR was higher
than that of ArQR and the thermal stability of KgQR
was considerably better than that of ArQR. The ther-
mal stability of the Q252LE170K mutant was consider-
ably better than that of WT. Therefore, KgQR and the
Q252LE170K mutant were used to construct a whole-
cell biocatalyst by transforming pBAD–hisB–QG into
E. coli K12/BW25113. According to previous studies
(Tomoko et al. 2000; Hou et al. 2014), the conversion
rate can be increased by adding NADþ. Therefore,
0.1 mM NADþ was added to the reaction mixture, and
242 g/L of the substrate was loaded. After 3 h of the
reaction, the substrate was completely converted to
(R)-3-Qui (Figure 3). The conversion rate in the present
study was 80.6 g/L/h. GC analysis of the product indi-
cated that the optical purity of the (R)-3-Qui produced
by KgQR was 99% ee, which was same as that pro-
duced by ArQR (Table 1).
4. Discussion
(R)-3-Qui is a useful compound applicable to the syn-
thesis of various pharmaceuticals (Rzeszotarski et al.
1988; Ishihara et al. 2004). Using carbonyl reductases
catalysing the asymmetric reduction of 3-quinuclidi-
none to (R)-3-Qui have been attempted, such as reduc-
tases of Datura stramonium (Portseffen et al. 1994) and
Hyoscyamus niger (Hashimoto et al. 1992), RrQR of
R. rubra (Uzura et al. 2009), QNR and BacC from
M. luteolum JCM 9174 (Isotani et al. 2012), and ArQR

BIOCATALYSIS AND BIOTRANSFORMATION
7
Hashimoto T, Nakajima K, Ongena G, Yamada Y. 1992. Two
tropinone reductases with distinct stereospecificities from
cultured roots of Hyoscyamus niger. Plant Physiol 100:
836–845.
Hoelsch K, Suhrer I, Heusel M, Weuster-Botz D. 2013.
Engineering of formate dehydrogenase: synergistic effect
of mutations affecting cofactor specificity and chemical
stability. Appl Microbiol Biotechnol 97:2473–2481.
Hou F, Miyakawa T, Kataoka M, Takeshita D, Kumashiro S,
Uzura A, Urano N, Nagata K, Shimizu S, Tanokura M. 2014.
Structural basis for high substrate-binding affinity and
enantioselectivity of 3-quinuclidinone reductase AtQR.
Biochem Biophys Res Commun 446:911–915.
thermal stability. It was incubated in a buffer (pH 7.0)
for 8 h at 30 and 40 ꢁC, with 80% residual activity
remaining. Therefore, in our study, a highly efficient
whole-cell biocatalyst system was obtained using
KgQR with good catalytic properties and
a
Q252LE170K mutant with high stability. A high yield of
(R)-3-Qui, greater than 200 g/L, was obtained in 3 h.
The conversion rate based on substrate was 80.6 g/L/h,
demonstrating that the process has a great potential
for the industrial production.
In the present study, we developed a highly effi-
cient whole-cell biocatalyst to produce (R)-3-Qui by
coexpressing KgQR and the Q252LE170K mutant (GDH)
in E. coli. Our result indicated that KgQR has good
catalytic features such as high specific activity, sub-
strate affinity, stereoselectivity, and thermal stability.
The thermal stability of GDH was clearly increased by
introducing Q252L and E170K mutations. The max-
imum conversion rate of 80.6 g/L/h, which was
achieved by the whole-cell biocatalyst coexpressing
KgQR and the Q252L and E170K mutant GDH, demon-
strates that this whole-cell biocatalyst has a great
effect on industrial manufacturing.
Hou F, Miyakawa T, Takeshita D, Kataoka M, Uzura A, Nagata
K, Shimizu S, Tanokura M. 2012. Expression, purification,
crystallization and X-ray analysis of 3-quinuclidinone
reductase from Agrobacterium tumefaciens. Acta
Crystallogr F Struct Biol Cryst Commun 68:1237–1239.
Ikunaka M. 2004. Biocatalysis from the perspective of an
industrial practitioner: let a biocatalyst do a job that no
chemocatalyst can. Catalysis Today 96:93–102.
Ishihara T, Kakuta H, Moritani H, Ugawa T, Yanagisawa I.
2004. Synthesis and biological evaluation of quinuclidine
derivatives incorporating phenothiazine moieties as squa-
lene synthase inhibitors. Chem Pharm Bull 52:1204–1209.
Isotani K, Kurokawa J, Itoh N. 2012. Production of (R)-
3-quinuclidinol by E. coli biocatalysts possessing NADH-
dependent 3-quinuclidinone reductase (QNR or bacC)
from Microbacterium luteolum and Leifsonia alcohol
dehydrogenase (LSADH). Int J Mol Sci 13:13542–13553.
Isotani K, Kurokawa J, Suzuki F, Nomoto S, Negishi T,
Matsuda M, Itoh N. 2013. Gene cloning and characteriza-
tion of two NADH-dependent 3-quinuclidinone reductases
from Microbacterium luteolum JCM 9174. Appl Environ
Microbiol 79:1378–1384.
Disclosure statement
The authors declare that they have no potential conflict of
interests.
Karsten L, Katja B, Andreas S. 2015. Multistep synthesis of
Funding
(S)-3-hydroxyisobutyric
acid
from
glucose
using
This study was supported by the Science Foundation of
Hebei Science Academy [Grant Nos. 12342, 15004022 and
2016023290-01] and Hebei High Technology Plan [Grant No.
16212801D].
Pseudomonas taiwanensis VLB120 B83 T7 catalytic biofilms.
Adv Synth Catal 357:1919–1927.
Kataoka M, Sri Rohani LP, Wada M, Kita K, Yanase H, Urabe I,
Shimizu S. 1998. Escherichia coli transformant expressing
the glucose dehydrogenase gene from Bacillus megate-
rium as a cofactor regenerator in a chiral alcohol produc-
tion system. Biosci Biotechnol Biochem 62:167–169.
Leusch A, Troger W, Greischel A, Roth W. 2000.
Pharmacokinetics of the M1-agonist talsaclidine in mouse,
rat, rabbit and monkey, and extrapolation to man.
Xenobiotica 30:797–813.
Liang B, Lang Q, Tang X, Liu A. 2013. Simultaneously improv-
ing stability and specificity of cell surface displayed glu-
cose dehydrogenase mutants to construct whole-cell
biocatalyst for glucose biosensor application. Bioresour
Technol 147:492–498.
Machielsen R, Looger LL, Raedts J. 2009. Cofactor engineer-
ing of Lactobacillus brevis alcohol dehydrogenase by com-
putational design. Eng Life Sci 9:38–44.
Makino Y, Negoro S, Urabe I, Okada H. 1989. Stability-increas-
ing mutants of glucose dehydrogenase from Bacillus meg-
aterium IWG3. J Biol Chem 264:6381–6385.
References
Baik SH, Ide T, Yoshida H, Kagami O, Harayama S. 2003.
Significantly enhanced stability of glucose dehydrogenase
by directed evolution. Appl Microbiol Biotechnol
61:329–335.
Baik SH, Michel F, Aghajari N, Haser R, Harayama S. 2005.
Cooperative effect of two surface amino acid mutations
(Q252L and E170K) in glucose dehydrogenase from
Bacillus megaterium IWG3 on stabilization of its oligomeric
state. Appl Environ Microbiol 71:3285–3293.
Bradford MM. 1976. A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing
the principle of protein-dye binding. Anal Biochem
72:248–254.
Ding H, Gao F, Liu D, Li Z, Xu X, Wu M, Zhao Y. 2013.
Significant improvement of thermal stability of glucose 1-
dehydrogenase by introducing disulfide bonds at the
tetramer interface. Enzyme Microb Technol 53:365–372.
Nagao T, Mitamura T, Wang XH, Negoro S, Yomo T, Urabe I,
Okada H. 1992. Cloning, nucleotide sequences, and
enzymatic
properties
of glucose
dehydrogenase

8
Z. JIA ET AL.
isozymes from Bacillus megaterium IAM1030. J Bacteriol
174:5013–5020.
mutations or insertions distal to the 2’-phosphate group
of NADP(H). Protein Eng Des Sel 8:1–8.
Pham SQ, Gao P, Li Z. 2013. Engineering of recombinant E.
coli cells co-expressing P450pyrTM monooxygenase and
glucose dehydrogenase for highly regio- and stereoselect-
ive hydroxylation of alicycles with cofactor recycling.
Biotechnol Bioeng 110:363–373.
Pire C, Esclapez J, Ferrer J, Bonete MJ. 2001. Heterologous
overexpression of glucose dehydrogenase from the
halophilicarchaeonHaloferaxmediterranei, an enzyme of
the medium chain dehydrogenase/reductase family. FEMS
Microbiol Lett 200:221–227.
Soni JY, Premasagar V, Thakore S. 2015. An improved and
simple route for the synthesis of 3-quinuclidinone hydro-
chloride. Lett Org Chem 12:277–279.
Takeshita D, Kataoka M, Miyakawa T, Miyazono K, Kumashiro
S, Nagai T, Urano N, Uzura A, Nagata K, Shimizu S, et al.
2014. Structural basis of stereospecific reduction by quinu-
clidinone reductase. AMB Express 4:1–6.
Tomoko M, Tadao H, Nobuyoshi N, Kaoru N. 2000.
Mechanism for improving stereoselectivity for asymmetric
reduction using acetone powder of microorganism. Tetreh
41:4135–4138.
Tsutsumi K, Katayama T, Utsumi N. 2009. Practical asymmet-
ric hydrogenation of 3-quinuclidinone catalyzed by the
XylSkewphos/PICA-Ruthenium(II) complex. Org Process Res
Dev 13:625–628.
Pire C, Marhuenda-Egea FC, Esclapez J. 2004. Stability and
enzymatic studies of glucose dehydrogenase from the
Archaeon Haloferax mediterranei in reverse micelles.
Biocatal Biotransform 22:17–23.
€
Portseffen A, Drager B, Nahrstedt A. 1994. The reduction of
tropinone in Datura stramonium root cultures by two spe-
cific reductases. Phytochemistry 37:391–400.
Uzura A, Nomoto F, Sakoda A, Nishimoto Y, Kataoka M,
Shimizu S. 2009. Stereoselective synthesis of (R)-3-quinucli-
dinol through asymmetric reduction of 3-quinuclidinone
with 3-quinuclidinone reductase of Rhodotorula rubra.
Appl Microbiol Biotechnol 83:617–626.
Prat M, Fernandez D, Buil MA, Crespo MI, Casals G, Ferrer M,
Tort L, Castro J, Monleon JM, Gavalda A, et al. 2009.
Discovery of novel quaternary ammonium derivatives of
(3R)-quinuclidinol esters as potent and long-acting mus-
carinic antagonists with potential for minimal systemic
exposure after inhaled administration: identification of
(3R)-3-{[hydroxy(di-2-thienyl)acetyl]oxy}-1-(3-phenoxy-
propyl)-1-azoniabicyclo [2.2.2]octane bromide (aclidinium
bromide). J Med Chem 52:5076–5092.
Ramadas C, Narayana RM, Srinivasa RC. 2013. Industrially
viable preparation of (R)-3-quinuclidinol, a key building
block of muscarine M1, M3 agonists and M3 antagonists.
Org Prep Proc Int 45:507–509.
Rzeszotarski WJ, McPherson DW, Ferkany JW, Kinnier WJ,
Noronha-Blob L, Kirkien-Rzeszotarski A. 1988. Affinity and
selectivity of the optical isomers of 3-quinuclidinyl benzi-
late and related muscarinic antagonists. J Med Chem
31:1463–1466.
Sambrook J, Russell DW. 2001. Molecular cloning: a labora-
tory manual. Cold Spring Harbor: Laboratory Press.
Schewe H, Kaup BA, Schrader J. 2008. Improvement of P450
(BM-3) whole-cell biocatalysis by integrating heterologous
cofactor regeneration combining glucose facilitator and
dehydrogenase in E. coli. Appl Microbiol Biotechnol
78:55–65.
Van Rossum HM, Kozak BU, Pronk JT, van Maris AJ. 2016.
Engineering cytosolic acetyl-coenzyme
A
supply in
Saccharomyces cerevisiae: pathway stoichiometry, free-
energy conservation and redox-cofactor balancing. Metab
Eng 36:99–115.
Weyler C, Heinzle E. 2015. Multistep synthesis of UDP-glu-
cose using tailored, permeabilized cells of E. coli. Appl
Biochem Biotechnol 175:3729–3736.
Willrodt C, Halan B, Karthaus L, Rehdorf J, Julsing MK,
Buehler K, Schmid A. 2017. Continuous multistep synthesis
of perillic acid from limonene by catalytic biofilms under
segmented flow. Biotechnol Bioeng 114:281–290.
Yamamoto K, Kurisu G, Kusunoki M, Tabata S, Urabe I, Osaki
S. 2001. Crystal structure of glucose dehydrogenase from
Bacillus megaterium IWG3 at 1.7 A resolution. J Biochem
129:303–312.
Zhang Y, Feng J, Li J, Wang Y, Wang M. 2016. Asymmetric
enzymatic reduction of quinuclidinone to (S)-3-quinuclidi-
nol. Chin J Appl Environ Biol 22:0854–0859.
Zhang WX, Xu GC, Huang L, Pan J, Yu HL, Xu JH. 2013.
Highly efficient synthesis of (R)-3-quinuclidinol in a space-
time yield of 916 g L(-1) d(-1) using a new bacterial reduc-
tase ArQR. Org Lett 15:4917–4919.
Solanki K, Abdallah W, Banta S. 2017. Engineering the cofac-
tor specificity of an alcohol dehydrogenase via single