Enantioselective synthesis of pure (R,R)-2,3-butanediol in Escherichia coli with stereospecific secondary alcohol dehydrogenases

Article abstract of DOI:10.1039/b913501d

We characterized the activity and stereospecificity of four secondary alcohol dehydrogenases (sADHs) towards acetoin reduction and constructed synthetic pathways in E. coli to produce enantiomerically pure (R,R)-2,3-butanediol (2,3-BDO) from glucose with

Full text of DOI:10.1039/b913501d

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
Enantioselective synthesis of pure (R,R)-2,3-butanediol in Escherichia coli
with stereospecific secondary alcohol dehydrogenases†
Yajun Yan, Chia-Chi Lee and James C. Liao*
Received 7th July 2009, Accepted 23rd July 2009
First published as an Advance Article on the web 3rd August 2009
DOI: 10.1039/b913501d
We characterized the activity and stereospecificity of four
secondary alcohol dehydrogenases (sADHs) towards acetoin
reduction and constructed synthetic pathways in E. coli
to produce enantiomerically pure (R,R)-2,3-butanediol
(2,3-BDO) from glucose with a titer of 6.1 g/L (enantio
purity >99%), and yield of 0.31 g product/g glucose (62% of
theoretical maximum).
BDO was also produced using engineered E. coli by feeding
diacetyl as precusor7 (Table 1). However, no enantiomerically
pure (R,R)-2,3-BDO production was achieved with microbial
fermentation. In this study, we characterized the activity and
stereospecificity of four sADHs towards (R)-acetoin reduction
and found that three of them produce specifically (R,R)-2,3-
BDO, while one produces meso-2,3-BDO. Using these enzymes,
we further constructed a synthetic pathway in E. coli to produce
enantiomerically pure (R,R)- or meso-2,3-BDO from glucose. To
our knowledge, production of pure (R,R)-form from glucose has
not been achieved previously (Table 1).
The 2,3-BDO biosynthesis pathway (Fig. 1) involves the
decarboxylation of acetolactate by ALDC to form acetoin,
which is stereospecific and only leads to the formation of the
(R)-enantiomer.8 When the ketone group of acetoin was reduced
to a hydroxyl group by sADH, either (S)-configuration or (R)-con-
figuration is formed depending on the enzyme stereospecificity.
This step creates the second stereo-center and yields either meso-
2,3-BDO or (R,R)-2,3-BDO. Production of enantiomerically pure
Synthesis of high-value pharmaceutical or agricultural com-
pounds often requires enantiomerically pure molecules, such as
secondary alcohols, as the important building blocks. However,
chiral synthesis or separation remains a costly step in chemical
synthesis. As an alternative, using enzymes or whole living cells
provides an effective approach to synthesize compounds with
high enantiomeric purity.1–4 In this study, we demonstrated chiral
synthesis of (R,R)-2,3-butanediol (2,3-BDO) from glucose with an
engineered pathway in Escherichia coli.
2,3-BDO contains two stereo centers and has three stereo
isomers including (R,R)-, meso- and (S,S)-forms. The optically
active isomers can be used as antifreeze agents. 2,3-BDO can
also be converted to 1,3-butandiene with established chemical
processes in the rubber industry. Moreover, dehydration of
2,3-BDO generates 2-butanone, which is an effective liquid fuel
additive.5 In general, 2,3-BDO can be naturally produced by
several microorganisms as a mixture of (R,R)- and meso-2,3 BDO.
Bacillus and Klebsiella are two native producers that have potential
for industrial applications in 2,3-BDO production. However, the
ratio of produced 2,3-BDO stereo isomers can vary dramatically
depending on the microorganisms and fermentation conditions
(Table 1).
There are several possible explanations for the mixed forma-
tion of 2,3-BDO isomers, including aeration conditions (redox
balance), non-stereospecific dehydrogenases, multiple pathways,
and multiple stereospecific dehydrogenases.5 Ui et al. introduced
gene fragments containing genes encoding acetolactate synthase
(ALS), acetolactate decarboxylase (ALDC) and a single meso-
dehydrogenase into E. coli, and obtained production of enan-
tiomerically pure meso-2,3-BDO6 (Table 1) which indicated that
mixed formation of 2,3-BDO is mainly due to the existence
of multiple pathways or dehydrogenases. Recently, (S,S)-2,3-
Department of Chemical and Biomolecular Engineering, University of
California, Los Angeles, 5531, Boelter Hall, 420 Westwood Plaza, Los
Angeles, CA 90095, USA. E-mail: Liaoj@ucla.edu; Fax: 310-206-4107;
Tel: 310-205-1656
† Electronic supplementary information (ESI) available: B. subtilis alsS
cDNA, B. subtilis alsD cDNA, B. subtilis bdhA cDNA, K. pneumoniae
MGH78578 budC cDNA, C. beijerinckii adh cDNA and T. brockii adh
cDNA sequences. See DOI: 10.1039/b913501d
Fig. 1 Schematic illustration of the 2,3-BDO biosynthesis pathway.
ALS: acetolactate synthase; ALDC: acetolactate decarboxylase; sADH:
secondary alcohol dehydrogenase.
3914 | Org. Biomol. Chem., 2009, 7, 3914–3917
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Table 1 Previous studies on enantioselective biosynthesis of stereo isomers of 2,3-BDO
Organism
sADH gene
Native
Precursor
glucose
Isomer
Titer(g/L)
Enantio purity
Reference
native producer
B. cereus
YUF-4
K. pneumoniae
IAM 1063
meso-
(R,R-)
meso-
(S,S)-
(R,R)-
meso-
(S,S)-
(R,R)-
0.3
0.2
21.2
1.6
6.0
17.7
2.2
60%
40%
74%
6%
20%
98%
98%
99%
13
10
Native
glucose
E. coli
E. coli
E. coli
K. pneumoniae IAM 1063 budC
B. saccharolyticum C-1012 bdh
B. subtilis bdhA
glucose
diacetyl
glucose
6
7
non-native producer
5.8
This
study
This
study
This
study
E. coli
E. coli
C. beijerinckii adh
T. brockii adh
glucose
glucose
(R,R)-
(R,R)-
5.1
6.1
99%
99%
Table 2 Kinetic parameters of secondary alcohol dehydrogenases on
2,3-BDO requires the use of a sADH with high stereospecifity. In
2,3-BDO native producers, the presence of multiple sADHs with
inconsistent stereospecificity contributes to formation of a mixture
of stereo isomers of 2,3-BDO. Therefore, reconstruction of the
2,3-BDO biosynthesis pathway in a non-native host with expres-
sion of only single sADH was expected to allow for the synthesis
of enantiomerically pure 2,3-BDO.
substrate acetoin
K
cat/Km
Organism
Gene Cofactor Km(mM)
K
cat(s-1)
(mM-1.s-1)
B. subtilis
C. beijerinkii
T. brockii
bdhA NADH
adh
adh
0.26 0.02 0.98 0.05 3.8
8.3 0.3 8.2 0.3 0.99
NADPH 0.23 0.02 0.91 0.06 4.0
0.85 0.04 58 68
NADPH
In order to obtain such a sADH, candidate sADH genes (the
sequences provided in the ESI†) were amplified from the genomic
DNA of 2,3-BDO native producer Bacillus subtilis and Klebsiella
pneumoniae MGH78578, since B. subtilis and K. pneumoniae
were reported to be able to produce (R,R)-form and meso-form
2,3-BDO, respectively.9,10 The genes bdhA from B. subtilis and
budC from K. pneumoniae were cloned. In addition, two adh genes
(Thermoanaerobacter brockii adh and Clostridium beijerinckii adh)
were completely synthesized with codon optimized for expression
in E. coli11 (Epoch Biolabs, Missouri City, TX). For protein
overexpression and purification, the amplified genes were inserted
into vector pETDuet-1 (Novagen, Madison, WI) with His-tag
at the N-terminal, generating expression plasmids pET-bdhA,
pET-budC, pET-CBADH, pET-TBADH. The proteins were ex-
pressed in the E. coli strain BL21(DE3) (Invitrogen, Carlsbad,
CA) harboring the constructed plasmids. The proteins were
purified with Ni-NTA spin columns (Qiagen, Valencia, CA). The
protein concentration was measured by Bradford assay (Bio-Rad,
Hercules, CA). Dehydrogenase activity was measured by monitor-
ing the absorbance decrease of NADH or NADPH at wavelength
340 nm. To determine the kinetic parameters, the assay reaction
was prepared with Tris-HCl buffer (50 mM, pH = 8.0) containing
100 mM of NADH or NADPH and various concentration of
acetoin ranging from 10 mM to 20 mM at room temperature.
The Km and Kcat values were obtained by non-linear fitting with
the Michaelis–Menten equation.
K. pneumoniae budC NADH
3
In order to determine the stereospecificity of these sADHs,
enantiomerically pure (R)-acetoin was used for the enzyme assay.
(R)-Acetoin was prepared by culturing E. coli that overexpressed
B. subtilis alsS (encoding ALS) and alsD (encoding ALDC) in
M9 glucose medium containing 4% glucose. The assay condition
was as described above with (R)-acetoin as the substrate at
a concentration of 5 mM. After 5 min, the reaction mixture
was analyzed with GC-FID (Hewlett Packard) equipped with a
HP-chiral 20b column (30 m, 0.32-mm internal diameter, 0.25-mm
film thickness; Agilent Technologies). For the analysis, the GC
oven temperature was initially set at◦40 ◦C for 2 min, increased
with a gradient of 5 ◦C min-1 until 45 C, and held for 4 min. Then
it was increased with a gradient 15 ◦C min-1 until 230 ◦C and held
for 4 min. Helium was used as the carrier gas. The temperature of
◦
the injector and detector were set at 225 C. The stereo isomers
were identified by the standard (R,R)-2,3-BDO, meso-2,3-BDO
and (S,S)-2,3-BDO purchased from Sigma-Aldrich. As the results
show in Fig. 2(A–E), all four tested sADHs demonstrated strict
stereospecificity on (R)-acetoin reduction. Among them, the gene
products of B. subtilis bdhA, C. beijerinckii adh and T. brockii
adh exclusively catalyzed the formation of (R,R)-2,3-BDO, while
the gene product of K. pneumoniae MGH78578 budC exclusively
catalyzed the production of meso-2,3-BDO. To our knowledge,
the stereospecificity of these enzymes have not been characterized
previously for the synthesis of 2,3-BDO isomers, and the gene
product of bdhA is the first sADH characterized from B. subtilis
that is responsible for 2,3-BDO formation.
With the characterized sADHs, we further engineered E. coli
as a whole-cell biocatalyst to produce enantiomerically pure
2,3-BDO from glucose. Vector pZE12-Luc was utilized to carry
each of the synthetic operons consisting of the 2,3-BDO biosyn-
thesis pathway with different sADHs.11 The synthetic operons
included three structural genes: alsS and alsD from B. subtilis
The sADHs from B. subtilis bdhA and T. brockii adh demon-
strated similar activities (Kcat = 0.98 s-1 and 0.91 s-1, respectively)
towards acetoin with high affinities (Km = 0.26 mM and 0.23 mM,
respectively), although the latter enzyme was not previously
reported to reduce acetoin. On the other hand, the enzyme coded
by C. beijerinckii adh has a low affinity towards acetoin (Km
=
8.3 mM) although activity is relatively high (Kcat = 8.2 s-1). The
enzyme encoded by budC from K. pneumoniae shows the highest
activity (Kcat = 58s-1) despite having a slightly lower affinity (Km =
0.85 mM) (Table 2).
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3914–3917 | 3915
©

View Article Online
Fig. 2 Identification of 2,3-BDO isomers by GC-FID. A and F: 2,3-BDO isomer standards; B: Assay product of bdhA from B. subtilis; C: Assay product
of adh from C. beijerinckii; D: Assay product of adh from T. brockii; E: Assay product of budC from K. pneumoniae MGH78578; G: Fermentation
product with E. coli expressing B. subtilis bdhA in the pathway; H: Fermentation product with E. coli expressing C. beijerinckii adh in the pathway; I:
Fermentation product with E. coli expressing T. brockii adh in the pathway; J: Fermentation product with E. coli expressing K. pneumoniae MGH78578
budC in the pathway. Peak 1: (R,R)-2,3-BDO; peak 2: (S,S)-2,3-BDO; peak 3: meso-2,3-BDO; peak 4: an unknown component from culture medium.
encoding ALS and ALDC, respectively, and the four sADHs
characterized above, each in a separate plasmid. The operons were
regulated by a pLlacO1 promoter with separated RBS located
upstream of each structure gene to facilitate the translation.
To construct these plasmids, the genes alsS and alsD were first
amplified from B. subtilis genomic DNA. These two genes were
integrated together by splicing overlap extension PCR with the
introduction of RBS sequence (AGGAGATATACC) in front of
alsD. Various sADH genes were inserted downstream of alsD
between SalI and BamHI. The generated plasmids carrying four
synthetic operons were named pZE12-alsS-alsD-bdhA, pZE12-
alsS-alsD-CBADH, pZE12-alsS-alsD-TBADH, and pZE12-alsS-
alsD-BudC (the plasmid construct is provided in the ESI†). The
constructed plasmids were then transformed into E. coli strains
respectively. The transformants were tested for the production of
enantiomeric 2,3-BDO through fermentation.
To test the production of 2,3-BDO, 1 ml of seed culture was
prepared in LB medium containing 50mg/L of ampicillin and
grown overnight at 37 ◦C inside a shaker set at 250 rpm. After
overnight incubation, the culture was inoculated (1% vol/vol)
into 10 ml of M9 medium containing 4% glucose, 0.5% yeast
extract, 1 mM MgSO4, 0.1 mM CaCl2, 2 ¥ 10-4% thiamine, and
50 mg/L ampicillin with a pH value at around 7.0. After growing
at 37 ◦C for 3 hours, IPTG was added into the culture to a
final concentration of 0.2 mM to induce◦protein expression. Then
the fermentation was carried out at 30 C with rigorous shaking
(250 rpm). The samples were taken after 48 hours and analyzed
with GC-FID. The analytic method was as described above. The
stereo isomers were identified and quantified by the standard
(R,R)-2,3-BDO, meso-2,3-BDO and (S,S)-2,3-BDO purchased
from Sigma-Aldrich. Glucose consumption was measured using a
glucose analyzer (YSI Inc., OH).
The wild type E. coli JCL16 (BW25113/F’[traD36, proAB+,
12
lacIq ZDM15])
carrying the operon with budC was able to
produce 4.5 g/L meso-2,3-BDO with only a trace amount of
the (R,R)- form from 17.1 g/L of glucose consumed. Since the
theoretical maximum yield of 2,3-BDO from glucose is 0.5 (g
product/g glucose consumed), we accomplished 53% of the the-
oretical maximum. When budC from K. pneumoniae MGH78578
was replaced by bhdA from B. subtilis, only (R,R)-2,3-BDO was
detected as the exclusive product (Fig. 2(F–J)). The titer was
around 2.2 g/L with the yield of 0.16 g product/g glucose (32%
of theoretical maximum). When the operons containing either
T. brockii adh or C. beijerinckii adh were expressed in wild type
E. coli JCL16, (R,R)-2,3-BDO was produced exclusively with titers
of 1.5 g/L and 2.8 g/L, respectively, with the yields of 0.12 g
3916 | Org. Biomol. Chem., 2009, 7, 3914–3917
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Table 3 Enantiomeric 2,3-BDO production in wild type E. coli JCL16 with different sADHs
sADH gene
Glucose consumption (g/L)
(R,R)-2,3-BDO (g/L)
meso-2,3-BDO (g/L)
Enantio purity
Yield (g/g glucose)
B. subtilis bdhA
C.beijerinckii adh
T. brockii adh
13.8 0.6
12.8 0.4
15.2 0.7
17.1 1.1
2.2 0.2
1.5 0.1
2.8 0.3
<0.04
<0.01
<0.01
<0.01
4.5 0.4
>99%
>99%
>99%
>99%
0.16
0.12
0.19
0.27
K. pneumoniae budC
Table 4 Enantiomeric 2,3-BDO production in engineered E. coli JCL260 with different sADHs
sADH gene
Glucose consumption (g/L)
(R,R)-2,3-BDO (g/L)
meso-2,3-BDO (g/L)
Enantio purity
Yield (g/g glucose)
B. subtilis bdhA
C.beijerinckii adh
T.brockii adh
18.9 0.5
17.8 0.8
19.7 0.7
29.6 2.8
5.8 0.3
5.1 0.4
6.1 0.5
<0.06
<0.03
<0.03
<0.03
10 0.8
>99%
>99%
>99%
>99%
0.30
0.29
0.31
0.34
K. pneumoniae budC
product/g glucose (24% of theoretical maximum) and 0.19 g
product/g glucose (38% of theoretical maximum), respectively
(Table 3).
enantiomeric purity over 99%. Due to the strict stereospecificity
and promiscuous activity of the characterized sADHs, they
will have broad applications in chiral synthesis. Furthermore,
we also demonstrated the advantage of using E. coli as a
whole cell biocatalyst with expression of certain biosynthesis
pathway by avoiding the costly reducing cofactor recycling proce-
dures.
In order to conserve the carbon source and reducing power
(either NADH or NADPH) for 2,3-BDO synthesis, the previously
created E. coli strain JCL260 with knockouts, adhE, ldhA, frdBC,
fnr, pta and pflB, was employed.12 With this host, the 2,3-BDO
production was dramatically increased without changing the
product chirality (Table 4). When budC-contained operon was
expressed in JCL260, meso-2,3-BDO was produced at 10 g/L
(122% increase) with a yield of 0.34 g product/g glucose (68% of
theoretical maximum). Similarly, (R,R)-2,3-BDO was produced
with expression of B. subtilis bdhA, C. beijerinckii adh or T.
brockii adh-contained operons at titers of 5.8 g/L, 5.1 g/L and
6.1 g/L, respectively, which were 160%, 240% and 120% increases
of titers in JCL16, with yields of 0.30 g product/g glucose
(60% of theoretical maximum), 0.29 g product/g glucose (58%
of theoretical maximum), and 0.31 g product/g glucose (62% of
theoretical maximum), respectively.
In conclusion, we described for the first time the enantioselective
synthesis of pure (R,R)-2,3-BDO isomer in E. coli from glucose
by employing stereospecific sADHs which were characterized
to produce the R-configuration specifically. By using engineered
whole cell biocatalysts expressing sADHs with strict stereospeci-
ficity instead of native producers, we were able to increase the
enantiomeric selectivity in (R,R)-2,3-BDO biosynthesis. Switching
the R-specific sADHs to the S-specific sADH in the pathway led
to the synthesis of enantiomerically pure meso-2,3-BDO. Both
(R,R)- and meso-2,3BDO were synthesized from glucose with
Notes and references
1 H. C. Tseng, C. H. Martin, D. R. Nielsen and K. L. Prather, Appl.
Environ. Microbiol., 2009, 75, 3137–3145.
2 B. Pscheidt and A. Glieder, Microb. Cell Fact, 2008, 7, 25.
3 K. Goldberg, K. Schroer, S. Lutz and A. Liese, Appl. Microbiol.
Biotechnol., 2007, 76, 249–255.
4 K. Goldberg, K. Schroer, S. Lutz and A. Liese, Appl. Microbiol.
Biotechnol., 2007, 76, 237–248.
5 M. J. Syu, Appl. Microbiol. Biotechnol., 2001, 55, 10–18.
6 S. Ui, Y. Okajima, A. Mimura, H. Kanai and T. Kudo, J. Ferm. Bioeng.,
1997, 84, 185–189.
7 S. Ui, Y. Takusagawa, T. Sato, T. Ohtsuki, A. Mimura, M. Ohkuma
and T. Kudo, Lett. Appl. Microbiol., 2004, 39, 533–537.
8 S. Najmudin, J. T. Andersen, S. A. Patkar, T. V. Borchert, D. H. Crout
and V. Fulop, Acta Crystallogr., D: Biol. Crystallogr., 2003, 59, 1073–
1075.
9 W. L. Nicholson, Appl. Environ. Microbiol., 2008, 74, 6832–6838.
10 S. Ui, H. Masuda and H. Muraki, J. Ferm. Technol., 1983, 61, 253–259.
11 T. Hanai, S. Atsumi and J. C. Liao, Appl. Environ. Microbiol., 2007, 73,
7814–7818.
12 S. Atsumi, A. F. Cann, M. R. Connor, C. R. Shen, K. M. Smith, M. P.
Brynildsen, K. J. Chou, T. Hanai and J. C. Liao, Metab. Eng., 2008, 10,
305–311.
13 S. Ui, T. Hosaka, K. Watanabe and A. Mimura, J. Ferm. Bioeng., 1998,
85, 79–83.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 3914–3917 | 3917
©