Efficient one-step preparation of γ-aminobutyric acid from glucose without an exogenous cofactor by the designed Corynebacterium glutamicum

Article abstract of DOI:10.1039/c4gc00607k

Lactobacillus plantarum CCTCC M209102 efficiently produces γ-aminobutyric acid (GABA) from l-glutamate, in which glutamate decarboxylase and pyridoxal kinase are involved in the transformation. Pyridoxal kinase catalyzes ATP-dependent phosphorylation of pyridoxal to produce pyridoxal-5′-phosphate, which is the cofactor required for glutamate decarboxylase to biotransform GABA from l-glutamate. Corynebacterium glutamicum G01 is a good producer of l-glutamate from glucose. However, it cannot yield GABA from l-glutamate due to the absence of glutamate decarboxylase and pyridoxal kinase. In this work, to realize the efficient one-step preparation of GABA from glucose without exogenous pyridoxal-5′-phosphate, the metabolic module from L-glutamate to GABA based on glutamate decarboxylase and pyridoxal kinase in L. plantarum was grafted into C. glutamicum. To further improve the GABA production, the pathways to by-product pools of L-arginine, L-proline and L-lysine were blocked using the insertional mutation technique. The engineered C. glutamicum APLGGP carrying argB::tacgad, proB::tacgad and dapA::tacplk could efficiently convert glucose into GABA in one-step without an exogenous co-factor. In fed-batch cultures, the recombinant C. glutamicum APLGGP produced 70.6 g L^-1 GABA at 30 °C and 70 h through a two-stage pH control strategy. To our knowledge, this is the highest reported GABA production using glucose as a substrate, and this designed C. glutamicum should be an excellent candidate for producing GABA on an industrial scale. This work is expected to pave the way to redesign the bioreactor for efficient one-step biosynthesis of GABA from glucose without an exogenous co-factor. the Partner Organisations 2014.

Full text of DOI:10.1039/c4gc00607k

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Efficient one-step preparation of γ-aminobutyric
acid from glucose without an exogenous cofactor
by the designed Corynebacterium glutamicum†
Cite this: Green Chem., 2014, 16,
4190
Rongzhen Zhang,a,b Taowei Yang,a Zhiming Rao,*‡a,b Hongmei Sun,a Meijuan Xu,a
Xian Zhang,a Zhenghong Xuc and Shangtian Yangd
Lactobacillus plantarum CCTCC M209102 efficiently produces γ-aminobutyric acid (GABA) from L-gluta-
mate, in which glutamate decarboxylase and pyridoxal kinase are involved in the transformation. Pyridoxal
kinase catalyzes ATP-dependent phosphorylation of pyridoxal to produce pyridoxal-5’-phosphate, which
is the cofactor required for glutamate decarboxylase to biotransform GABA from L-glutamate. Corynebac-
terium glutamicum G01 is a good producer of L-glutamate from glucose. However, it cannot yield GABA
from L-glutamate due to the absence of glutamate decarboxylase and pyridoxal kinase. In this work, to
realize the efficient one-step preparation of GABA from glucose without exogenous pyridoxal-5’-phos-
phate, the metabolic module from L-glutamate to GABA based on glutamate decarboxylase and pyridoxal
kinase in L. plantarum was grafted into C. glutamicum. To further improve the GABA production, the path-
ways to by-product pools of L-arginine, L-proline and L-lysine were blocked using the insertional mutation
technique. The engineered C. glutamicum APLGGP carrying argB::tacgad, proB::tacgad and dapA::tacplk
could efficiently convert glucose into GABA in one-step without an exogenous co-factor. In fed-batch
cultures, the recombinant C. glutamicum APLGGP produced 70.6 g L−1 GABA at 30 °C and 70 h through a
two-stage pH control strategy. To our knowledge, this is the highest reported GABA production using
glucose as a substrate, and this designed C. glutamicum should be an excellent candidate for producing
GABA on an industrial scale. This work is expected to pave the way to redesign the bioreactor for efficient
one-step biosynthesis of GABA from glucose without an exogenous co-factor.
Received 8th April 2014,
Accepted 19th June 2014
DOI: 10.1039/c4gc00607k
www.rsc.org/greenchem
recent special attention has been focused on GABA preparation by
employing ‘‘green” and efficient bioprocesses since they have a
Introduction
Gamma-aminobutyric acid (GABA), a non-protein amino acid, simple reaction procedure, high catalytic efficiency, mild reac-
functions as a major inhibitory neurotransmitter in animals, tion conditions and environmental compatibility.2
and it has many physiological properties related to anti-anxiety
In eukaryotes and prokaryotes, including bacteria, fungi
and hypotension as tranquilizers, diuretics, and analgesics.1 and yeasts, glutamate decarboxylase (GAD, EC 4.1.1.15) cata-
GABA can be used as a bioactive component in the food, feed, lyzes the conversion of L-glutamate to GABA.3 However, GAD is
and pharmaceutical fields. Although there have been many a pyridoxal-5′-phosphate-dependent enzyme that catalyzes ir-
attempts for the GABA synthesis chemically or biologically, reversible α-decarboxylation of L-glutamate to GABA with pyri-
doxal-5′-phosphate as a cofactor.4 Very recently, Rowley and
Whisstock’s group have done excellent work on GAD structure
aThe Key Laboratory of Industrial Biotechnology, Ministry of Education, School of
Biotechnology, Jiangnan University, Wuxi, 214122, P. R. China.
E-mail: raozhm@jiangnan.edu.cn; Tel: +86-510-85916881
bNational Laboratory for Food Science, Jiangnan University, Wuxi 214122,
P. R. China
determination and kinetic studies for regulation of GABA
production.5
Using lactic acid bacteria harboring GAD to produce GABA-
enriched food has attracted great attention since they are safe
and recognized as probiotics.6 During the GABA production by
lactic acid bacteria, glutamate is added as a precursor to the
fermentation medium, and some expensive nitrogen sources
are supplemented for microorganism cultivation.7,8 The fer-
mentation method is not cost-effective, so, a new approach is
required for the sustainable industrial application.9 Kook et al.
reported the maximum GABA production of 27.3 g L−1 by
cSchool of Medicine and Pharmaceuticals, Jiangnan University, Wuxi, 214122,
P. R. China
dDepartment of Chemical Engineering, Ohio State University, 140 West 19th Avenue,
Columbus, OH 43210, USA
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4gc00607k
‡Present address: School of Biotechnology, Jiangnan University, 1800 Lihu
Avenue, Wuxi, Jiangsu, 214122, P. R. China.
4190 | Green Chem., 2014, 16, 4190–4197
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
expressing Lactobacillus plantarum GAD gene in Lactobacillus
The enzyme GAD is active to produce GABA with pyridoxal
Sakei.10 Corynebacterium glutamicum is an important industrial 5′-phosphate as a cofactor. Pyridoxal kinase (PLK) was reported
producer of various amino acids with great potential. Takaha- to catalyze the ATP-dependent phosphorylation reaction of
shi et al. successfully produced GABA from glucose in one step pyridoxal to produce pyridoxal 5′-phosphate.4 Although the fer-
using the recombinant C. glutamicum expressing Escherichia mentation medium, such as corn steep liquor and yeast
coli GadB.11 As is known, the enzyme GAD is the key enzyme extract, contains pyridoxal, the precursor of pyridoxal 5′-phos-
for the reaction from L-glutamate to GABA, the introduction of phate, it is not enough to support the quick biotransformation
one extra copy will promote the GABA production improve- of GABA. So the introduction of the enzyme PLK involved in
ment. C. glutamicum is an important industrial producer of the formation of pyridoxal 5′-phosphate will be beneficial to
various amino acids with great potential for the production of the GABA biosynthesis.
other metabolites. Shi et al. improved GABA production using
In our previous work, we reported that a modified strain
the recombinant C. glutamicum coexpressing two GAD genes through multi-mutagenesis, L. plantarum GB 01-21, showed
gadB1 and gadB2 from Lactobacillus brevis Lb85.13 Through the high GAD activity, and the engineered E. coli BL21 harboring
fermentation condition optimization, they further increased L. plantarum GAD gene (gad, accession no: JN248358) yielded
GABA production to a maximum level of 27.13 g L−1 after 204.5 g L−1 GABA from L-glutamate at 24 h. The mole conver-
120 h flask cultivation and 26.32 g L−1 after 60 h fed-batch fer- sion rate of the substrate reached 97.9%.14 By analysing the
mentation.13 However, the expensive pyridoxal 5′-phosphate is metabolic pathways, L-arginine and L-proline are the main by-
required for GABA production. The reported methods and pro- products obtained during the biotransformation of GABA from
ductivity of GABA are summarized in Table 1.
L-glutamate in L. plantarum GB 01-21 (Fig. 1). We also isolated
Table 1 The reported methods and productivity of GABA
Strains
Engineered methods
Substrates
Fermentation durations (h)
GABA (g L−1
)
Ref.
Lactobacillus brevis NCL912
L. brevis TCCC13007
Corynebacterium glutamicum
Lactobacillus Sakei
C. glutamicum
C. glutamicum ATCC 13032
Wild-type
Wild-type
Expressing GAD gene
Expressing GAD gene
Expressing GAD gene
Expressing GAD gene
Expressing GAD gene
Sodium L-glutamate
Monosodium glutamate
L-glutamate
L-glutamate
Glucose
Glucose
L-glutamate
48
66
60
48
72
72
24
35.66
38
26.32
27.3
12.37
2.15
204.5
7
6
13
10
11
12
14
Escherichia coli
Fig. 1 The strategy for designed C. glutamicum to produce GABA from glucose in one-step. Lactobacillus plantarum CCTCC M209102 is a good
producer of γ-aminobutyric acid (GABA) from L-glutamate. The glutamate decarboxylase and pyridoxal kinase are involved in the above biotrans-
formation. Glutamate decarboxylase biotransforms L-glutamate to GABA with pyridoxal-5’-phosphate as a cofactor. Pyridoxal catalyzes ATP-depen-
dent phosphorylation of pyridoxal to generate pyridoxal-5’-phosphate, which is supplied for the reaction by glutamate decarboxylase.
Corynebacterium glutamicum G01 is an excellent producer of L-glutamate from glucose. However, it cannot yield GABA from L-glutamate due to
the absence of glutamate decarboxylase and pyridoxal kinase. To realize the efficient one-step production of GABA from glucose, the C. glutamicum
reactor is redesigned. The metabolic module from L-glutamate to GABA in L. plantarum was grafted into C. glutamicum, making C. glutamicum
prepare GABA from glucose in one-step. FA, fragment of argB::tacgad; FB, fragment of proB::tacgad; FC, fragment of plk::tacdapA.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4190–4197 | 4191

View Article Online
Paper
Green Chemistry
a GRAS (generally regarded as safe) strain, C. glutamicum G01, nique, the coding genes of NAGK, GAK and DHDPS (argB, proB
which could accumulate about 106 g L−1 L-glutamate from and dapA) were expected to be knocked out.
glucose.13 L-Lysine is the main by-product obtained during the
The gad gene with the tac promoter from the plasmid
biosynthesis of L-glutamate from glucose in C. glutamicum G01 pDXW-gad was inserted into pMD-RargB and pMD-RproB to
(Fig. 1). create the recombinant pMD-RargB::tacgad and pMD-RproB::
In this work, to realize the efficient production of GABA tacgad, respectively. The gene plk with the tac promoter from
from glucose in one-step without exogenous pyridoxal-5′-phos- pDXW-plk was cloned into the pMD-RdapA to create the
phate, we grafted the metabolic modular from L-glutamate to plasmid pMD-dapA::tacplk. Then the RargB::tacgad, RproB::
GABA based on GAD and PLK in L. plantarum GB 01-21 to C. tacgad and dapA::tacplk fragments were inserted into the
glutamicum G01. To further improve the GABA production, we suicide plasmid pK18mobsacB, respectively, to construct the
not only knocked out the genes involved in the by-products recombinant plasmids pK18-RargB::tacgad, pK18-RproB::
pathways of L-arginine, L-proline and L-lysine through inser- tacgad and pK18-dapA::tacplk. After verification by restriction
tional inactivation, but also introduced two copies of GAD enzyme analysis and DNA sequencing, the recombinant plas-
genes into the engineered C. glutamicum. Finally, GABA pro- mids pK18-RargB::tacgad, pK18-RproB::tacgad and pK18-dapA::
duction reached 70.6 g L−1 after 70 h of fermentation under a tacplk were introduced into C. glutamicum G01 one by one
two-stage pH control strategy.
using the electroporation method.17 The recombinant C. gluta-
micum AG carrying pK18-RargB::tacgad, C. glutamicum APGG
carrying pK18-RargB::tacgad and pK18-RproB::tacgad, C. gluta-
micum APLGGP carrying pK18-RargB::tacgad, pK18-RproB::
tacgad and pK18-dapA::tacplk were constructed. In C. glutami-
cum AG, the gad gene was introduced and the gene argB was
deleted. In C. glutamicum APGG, the gad gene with two copies
Results and discussion
Construction of the engineered C. glutamicum
Lactobacillus plantarum GB 01-21 catalyzes the biotransform- was introduced and the argB and proB genes were knocked
ation of γ-aminobutyric acid (GABA) with an excellent perform- out. In C. glutamicum APLGGP, the targets genes gad and plk
ance. The glutamate decarboxylase (GAD) and pyridoxal kinase involved in the biotransformation of GABA from L-glutamate
(PLK) are the key enzymes involved in the conversion of L-glu- were introduced, and the genes argB, proB and dapA involved
tamate to GABA. The enzyme GAD catalyzes L-glutamate to in the by-product pathways of L-arg, L-pro and L-lys were
GABA with pyridoxal-5′-phosphate as a cofactor. Pyridoxal knocked out. More important, glutamate decarboxylase gene
kinase catalyzes ATP-dependent phosphorylation of pyridoxal was integrated into C. glutamicum APLGGP genome with two
to produce pyridoxal-5′-phosphate, which is required by GAD.4 copies.
Although Corynebacterium glutamicum G01 is an excellent pro-
ducer of L-glutamate from glucose, it cannot yield GABA due to
the absence of GAD and PLK. In order to produce GABA from
Identification of correct insertional mutation by enzyme
assays
glucose in one-step by C. glutamicum, the coding genes of GAD C. glutamicum is an important industrial producer of various
and PLK from L. plantarum were cloned into C. glutamicum. amino acids with great potential for the production of other
The GAD gene (gad) from the L. plantarum was inserted into metabolites. In order to identify whether the insertion
the plasmid pDXW-10 to generate pDXW-gad. In the same way, mutation was correctly integrated into the C. glutamicum
the PLK gene (plk) from L. plantarum was cloned into the genome, the specific activities of enzymes were determined
plasmid pDXW-10 to create pDXW-plk. The recombinant plas- using cell-free extracts of the recombinant strain C. glutamicum
mids were verified by DNA sequencing.
AG, APGG and APLGGP. As shown in Table 2, in the recombi-
Since L-arginine and L-proline are the main by-products nant strain C. glutamicum AG, NAGK activity was observed,
during the biotransformation of GABA from L-glutamate by C. while in the parent strain C. glutamicum G01, GAD activity was
glutamicum G01, the pathway to by-product pools need to be 2.13 U mg−1, but no any NAGK activity was detected. The
blocked to increase the carbon flux from glucose to L-gluta- result suggested that the enzyme GAD was expressed, while
mate. Lazcano’s group has reported that N-acetyl-L-glutamate NAGK involved in the L-arg biosynthesis was successfully
kinase (NAGK) is involved in the arginine biosynthesis knocked out in the recombinant C. glutamicum AG strain. In
pathway, and the three enzymes, glutamate kinase (GAK), glu- the engineered C. glutamicum APGG, the activities of NAGK
tamate-γ-semialdehyde dehydrogenase and pyrroline-5-carboxy- and GAK were not observed, while the GAD activity was almost
late reductase catalyzes the proline synthesis in bacteria.15 improved 2-fold (Table 2) due to the introduction of an extra
L-Lysine is the other by-product during the production of L-glu- copy of gad and the deletion of NAGK and GAK genes involved
tamate from glucose in C. glutamicum G01. Dihydropicolinate in the synthesis of L-arg and L-pro. In the recombinant C. gluta-
synthase (DHDPS) is a key enzyme for lysine biosynthesis.15 As micum APLGGP, GAD and PLK were active, but NAGK, GAK,
described in the Experimental section, the recombination plas- and DHDPS involved in the by-product pathways of L-arg, L-pro
mids pMD-RargB, pMD-RproB and pMD-RdapA containing and L-lys were inactive. More important, C. glutamicum
flanks of argB (RargB), proB (RproB) and dapA (RdapA) were APLGGP showed 4.2-fold PLK activity higher than the other
constructed.16 So, by the homologous recombination tech- strains: wild-type, C. glutamicum AG and C. glutamicum APGG,
4192 | Green Chem., 2014, 16, 4190–4197
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Table 2 Crude enzyme activities in the recombinant strains
Specific activities (U mg−1 protein)
Strains GAD
NAGK
GAK
DHDPS
PLK
G01
AG
APGG
—
0.30 0.02 0.22 0.02 0.21 0.01 0.11 0.01
2.13 0.13
4.20 0.25
—
—
—
0.23 0.02 0.20 0.01 0.10 0.01
—
—
0.21 0.01 0.10 0.01
— 0.43 0.03
APLGGP 4.18 0.23
In the recombinant C. glutamicum AG, the NAGK gene was knocked
out and GAD gene was expressed. In the recombinant C. glutamicum
APGG, the NAGK and GAK genes were knocked out, and two copies of
GAD genes were expressed. In the recombinant C. glutamicum
APLGGP, the genes of NAGK, GAK and DHDPS were knocked out, and
PLK and two copies of GAD were expressed. GAD, glutamate
decarboxylase; NAGK, N-Acetyl-L-glutamate kinase; GAK, glutamate
kinase; DHDPS, dihydropicolinate synthase; PLK, pyridoxal kinase; —,
non-detected.
Fig. 2 Growth curves of C. glutamicum G01, AG, APGG and APLGGP.
The engineered strains were cultivated at 37 °C in a 2.5 L flask bottle
with an initial working volume of 1.0 L. Error bars represent standard
deviations (n = 3).
making biotransformation of GABA possible without pyri-
doxal-5′-phosphate. These above results indicate that the key
enzymes, GAD and PLK catalyzing L-glutamate to GABA were
expressed, while the NAGK, GAK, and DHDPS involved in
the by-products L-arg, L-pro and L-lys were successfully
knocked out.
secrete L-arg into the fermentation liquid, while the parental
strain C. glutamicum G01 did (Table 3). These results further
confirmed that L. plantarum GAD was active, but NAGK was
inactive in the recombinant C. glutamicum AG. These above
results demonstrated that GAD expression is a key factor for
enhancing the GABA production, which is similar to the other
reports.12 More interesting, in the C. glutamicum AG harboring
one copy of GAD coding gene, 40.7 g L−1 L-glutamate was
detected in the broth, suggesting that much L-glutamate was
not transformed to GABA.
Since the GAD enzyme is a key factor for biotransformation
of GABA from glucose, one extra copy of GAD gene was intro-
duced into the recombinant C. glutamicum APGG. In this
recombinant strain, the by-product pathways of L-arg and L-pro
were also blocked to further improve GABA yield. As shown in
Table 3, the production of GABA reached 69.9 g L−1 in the
broth of C. glutamicum APGG after 80 h fed-batch cultivation.
Compared to C. glutamicum AG, the GABA production was
further improved from 45.8 g L−1 to 69.9 g L−1, and L-gluta-
mate was almost consumed by C. glutamicum APGG. Mean-
while, L-arg and L-pro were not detected. The enhanced GABA
yield is due to the two reasons. First, the introduction of an
The insertional mutation had little effect on cell growth
Since L-arg, L-pro and L-lys have separate and distinct roles in
the microbial metabolism, the deletion of genes involved in
their synthesis may have an affect on cell growth of the recom-
binant C. glutamicum. Meanwhile, the effects of introduction
of GAD and PLK into C. glutamicum on the cell-growth were
investigated. The fermentation characteristics of the recombi-
nants had been compared with the data from flask experi-
ments under the same conditions. Cell growth was determined
by measuring the turbidity of the culture at OD560 using a UV-
visible spectroscopy system. Based on the three-stage division
of cell growth curves (Fig. 2), the engineered C. glutamicum
strains AG, APGG, and APLGGP grew at a similar rate as com-
pared to the parent strain. The results suggested that the
knock out of genes involved in L-arg, L-pro and L-lys biosyn-
thesis and introduction of the key genes of gad and plk related
to the biotransformation of GABA had nearly no effect on cell
growth properties. The insertion mutation in C. glutamicum
did not inhibit cell growth. The insertion mutation system
based on the five enzymes, GAD, PLK, NAGK, GAK, and
DHDPS would be preferable for further studies of bioconver-
sion of GABA from glucose without an exogenous cofactor.
Table 3 GABA production by the recombinant strains
One-step preparation of GABA from glucose without an
exogenous cofactor was realized by C. glutamicum APLGGP
L-Glu
L-Arg
L-Pro
L-Lys
GABA
(g L−1
)
Strains
(g L−1
)
(g L−1
)
(g L−1
)
(g L−1
)
By fermentation with glucose, GABA production was compared
between the engineered C. glutamicum and parental strain.
The results are summarized in Table 3. The strain C. glutami-
cum AG accumulated 45.8 g L−1 GABA with a productivity of
0.48 g L−1 h−1 in the fermented broth, while the wild-type
C. glutamicum did not produce GABA after 80 h of fermenta-
tion (Table 3). However, the strain C. glutamicum AG did not
G01
AG
APGG
APLGGP
106.3 3.3
40.7 1.2
0.6 0.1
5.2 0.3
4.3 0.2
4.5 0.3
—
3.1 0.2
3.3 0.2
3.2 0.2
—
—
—
—
—
45.8 1.3
69.9 1.9
70.6 2.1
0.5 0.1
—
The recombinant C. glutamicum APLGGP catalyzes the bio-
transformation without addition of pyridoxal-5′-phosphate, while the
other strains catalyze the reaction with an appropriate amount of
pyridoxal-5′-phosphate. —, non-detected.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4190–4197 | 4193

View Article Online
Paper
Green Chemistry
extra copy of GAD induced the higher GAD activity for the
transformation of GABA from L-glutamate. Second, the inactive
NAGK and GAK by insertional mutation redistributed carbon
flux of glucose to the GABA pathway. Shi et al. also improved
GABA production (27.13 g L−1 at 120 h) by co-expressing gadB1
and gadB2 in C. glutamicum ATCC13032.13
As the GAD requires pyridoxal-5′-phosphate as a cofactor for
its activity,4 PLK involved in the generation of pyridoxal-5′-
phosphate and two copies of GAD were introduced into the
recombinant C. glutamicum APLGGP, in which the relative
genes involved in by-product pathways of L-arg, L-pro and L-lys
were knocked out. The fermentation results showed that C. glu-
tamicum APLGGP produced 70.6 g L−1 GABA from glucose
without addition of pyridoxal-5′-phosphate. Compared with
the fermentation results of C. glutamicum APGG with addition
of pyridoxal-5′-phosphate, C. glutamicum APLGGP slightly
improved GABA production without an exogenous cofactor.
Thus, the one-step biotransformation of GABA from glucose
was realized by the recombinant strain C. glutamicum APLGGP.
Enhanced production of GABA through a two-stage pH control
strategy
During the biotransformation of GABA from glucose, L-gluta-
mate is the precursor of GABA. It is very important to maintain
the neutral conditions to steadily biosynthesize GABA from
L-glutamate. Ammonia can not only take on this responsibility,
but also can be used as the nitrogen source to promote the
GABA production. Moreover, the L-glutamate dehydrogenase,
which is critical for L-glutamate biosynthesis, exhibits the
highest activity at pH 7.5, whereas at pH 6.5, half of its activity
is lost.18 So ammonia supplementation has a great influence
on the biosynthesis of GABA from glucose. In this work, since
L-glutamate is rapidly synthesized within 48 h of fermentation,
ammonia was supplemented during 0–48 h of fermentation to
retain the pH values at 7.3. As shown in Fig. 3A, it was found
that time profiles of GABA production could be divided into
two stages. At the first 48 h, pH values were constant at about
7.3, and L-glutamate was quickly accumulated, while GABA was
produced very slowly. When ammonia supplement was
stopped, the pH value dropped rapidly to about 5.4, GABA
began to secret to the fermentation broth.
Fig. 3 Fed-batch fermentation of GABA by the recombinant C. glutami-
cum APLGGP. (A) pH value was controlled at 7.3 until 48 h, and was then
not controlled; (B) pH value was controlled at 7.3 until 48 h, and then
controlled at 4.8. The strains were cultured in a 5 L bioreactor at 37 °C.
It was reported that the enzyme GAD shows activity under
relatively acidic conditions.4,13 Shi et al. observed that GAD
exhibited a stringent pH dependence for substrate transport
and functions at acidic pH, without any activity above pH
6.5.12 Seo et al. found that the optimal pH of GAD activity was
5.2.19 We further determined the effects on GABA conversion
in the range of pH 4.0–5.8. As shown in Fig. 4, GAD showed
the highest activity at about pH 4.8, and the conversion rate of
GABA was the highest at pH 4.6–5.0. Basing on the above pH
preference, a two-stage control strategy was proposed in favour
of GABA formation. In the first 48 h, the pH value was con-
trolled at about 7.3 for L-glutamate formation, and then pH
was adjusted to 4.8 to promote GABA accumulation. As
expected, through the strategy, the GABA production was up to
70.6 g L−1 after 70 h fed-batch cultivation, and the GABA for-
Fig. 4 Effects of pH on conversion rate of GABA.
mation rate was up to 1.04 g L−1 h−1 (Fig. 3B), which is the
highest production report of GABA with glucose as a substrate
in one-step. Thus, an excellent GABA producer from glucose
was redesigned through combining the metabolic modules
4194 | Green Chem., 2014, 16, 4190–4197
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
(from L-glutamate to GABA) in L. plantarum and C. glutamicum culture was then inoculated (4% v/v) into the fermentation
(from glucose to L-glutamate), and blocking the by-product medium (160 g L−1 glucose, 30 g L−1 corn steep liquor, 10 g
pathways of L-arg, L-pro and L-lys. The engineered C. glutami- L−1 yeast extract, 5.5 g L−1 urea, 1.5 g L−1 K2HPO4·3H2O, 0.8 g
cum directly converted glucose to GABA without an exogenous L−1 MgSO4·7H2O, 0.02
g g
L−1 MnSO4·4H2O, 0.02 L−1
cofactor through a two-stage pH control strategy.
FeSO4·7H2O, 200 μg L−1 vitamin B1, 0.1 mmol L−1 pyridoxal-5′-
phosphate, pH 7.5). Only for the recombinants carrying a pyri-
doxal kinase gene, 0.1 mmol L−1 pyridoxal-5′-phosphate was
not added into the fermentation medium. For all the recombi-
nant strains, 1 μmol L−1 isopropyl β-D-1-thiogalactopyranoside
was added to the fermentation medium to induce GAD
expression. Fed-batch fermentation was carried out in a 5 L
bioreactor (BIOTECH-2002, Baoxing Biological Equipment Co.,
Shanghai, China) containing 2.0 L initial medium at 30 °C,
350 rpm and an airflow rate 0.66 vvm. Ammonia (NH4OH) was
added during the first 48 h of fermentation to maintain the
alkaline environment. And then, the addition of ammonia was
stopped to form an acidic fermentation environment, which
was suitable for the GABA production. Glucose at 80% (w/v)
was fed into the bioreactor to maintain the glucose concen-
tration between 20 and 30 g L−1 for 16–44 h.
Conclusions
An excellent producer of GABA was successfully redesigned by
grafting the metabolic modular of L-glutamate from glucose in
L. plantarum into C. glutamicum, making the engineered C. glu-
tamicum highly yield GABA from glucose in one-step. Further-
more, the by-product pathways of L-arg, L-pro and L-lys were
blocked to redistribute the metabolic flux to GABA, and a two-
stage pH control strategy was proposed based on the pH pre-
ference of the key enzymes involved in the GABA synthesis.
The production of GABA reached up to 70.6 g L−1 after 70 h.
More important, C. glutamicum is a GRAS microorganism and
its engineered strain can grow in no-selective medium. There-
fore, the newly designed C. glutamicum system is an excellent
GABA producer from glucose in food and pharmaceutical
industry. This work is hoped to pave the way to redesign the
bioreactor for efficient one-step preparation of GABA from
glucose.
Plasmid construction and gene deletion
The glutamate decarboxylase gene (gad) from L. plantarum GB
01-21 was amplified by PCR using primers gadP1 and gadP2.
The amplified gad gene was inserted into the Nhe I and Bgl II
sites of the pDXW-10 plasmid to create pDXW-gad. For dele-
tion of argB and proB in C. glutamicum G01 by homologous
recombination, the recombination flanks of argB (RargB) and
proB (RproB) were constructed and individually cloned into
pMD18-T vector (Takara Co., Dalian, China), resulting in the
vectors pMD-RargB and pMD-RproB. Recombination flanks of
argB and proB were individually amplified from the C. glutami-
cum G01 chromosome by PCR as follows: flank A (445 bp) of
argB and flank A (646 bp) of proB were generated with the
primer pairs argBP1/argBP2 and proBP1/proBP2, respectively;
flank B (529 bp) of argB and flank B (613 bp) of proB were
created with the primer pairs argBP3/argBP4 and proB3/
proBP4, respectively; the recombination flanks were then con-
structed by splicing the corresponding flanks A and B. The Bgl
II restriction site was introduced into the splicing sites of
flanks A and B. The gad gene, containing the tac promoter
from the pDXW-gad plasmid, was then PCR-amplified using
the primers tacgadP and gadP2. The amplified tac-gad gene
was inserted into the pMD-RargB and pMD-RproB plasmids to
create the pMD-RargB::tacgad and pMD-RproB::tacgad plas-
mids, respectively. The pyridoxal kinase gene (plk) from
L. plantarum GB 01-21 was amplified using primers plkP1 and
plkP2. The amplified plk gene was inserted into the Pst I and
Hind III sites of the pDXW-10 plasmid to create the pDXW-plk
Experimental
General information
All chemicals, i.e., glutamate, N-acetyl-L-glutamate, pyridoxal
etc. were purchased from the Sigma-Aldrich Chemical Co. Inc.
(Shanghai, China), and used without further purification. The
precise determination of the amino acids was carried out on
an automatic amino acid analyzer by high-performance liquid
chromatography.
Organisms and growth conditions
C. glutamicum G01 has been deposited in the China Center for
Type Culture Collection (CCTCC) under collection number
CCTCC M2013418. E. coli JM109 strain was grown at 37 °C in
Luria–Bertani (LB) medium with ampicillin (50 μg mL−1) and/
or kanamycin (50 μg mL−1). C. glutamicum was grown in LBG
medium (LB medium supplemented with 5 g L−1 glucose) at
30 °C with 30 μg mL−1 kanamycin, if necessary. The competent
cells of C. glutamicum were cultured in the LB medium with
30 g L−1 glycine and 1 g L−1 Tween 80.
Culture conditions for L-glutamate and GABA fermentation
from glucose
C. glutamicum strains were subcultured weekly and stored at plasmid. The plk gene, containing the tac promoter from the
4 °C or frozen at −70 °C in 15% (v/v) glycerol. Stock and inocu- pDXW-plk plasmid, was then amplified using the primers
lum cultures were grown at 30 °C with shaking at 160 rpm in tacplkP and plkP2. The dihydropicolinate synthase gene (dapA)
seed medium (25 g L−1 glucose, 20 g L−1 corn steep liquor, from C. glutamicum G01 was amplified using primers dapAP1
10 g L−1 yeast extract, 6 g L−1 urea, 1.5 g L−1 K2HPO4·3H2O, and dapAP2. The amplified dapA gene was cloned into the
0.4 g L−1 MgSO4·7H2O, pH 7.5) for 12 h. The incubated seed pMD18-T vector to create pMD-dapA. The amplified tac-plk
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4190–4197 | 4195

View Article Online
Paper
Green Chemistry
gene was inserted into the pMD-dapA plasmid to create the the National Basic Research Program of China (973 Program)
plasmid pMD-dapA::tacplk. (2012CB725202), the National Natural Science Foundation of
The pK18-RargB::tacgad, pK18-RproB::tacgad and pK18- China (21276110), the Research Project of Chinese Ministry of
dapA::tacplk vectors were created by inserting a RargB::tacgad, Education (113033A), the Fundamental Research Funds for the
RproB::tacgad and dapA::tacplk fragments of pMD-RargB:: Central Universities (JUSRP51306A and JUSRP21121), the 111
tacgad digested by Xba I and Hind III, pMD- RproB::tacgad Project (111-2-06) and a Project Funded by the Priority Aca-
digested by Xba I and Hind III and dapA::tacplk digested by demic Program Development of Jiangsu Higher Education
Xba I and Sal I (agarose-purified, Gel DNA Purification Kit, Institution.
Takara) into same restriction enzymes linearized pK18mob-
sacB, respectively. The deletion of argB, proB and dapA and
introduction of gad and plk genes in C. glutamicum were pre-
Notes and references
formed by homologous recombination.20 The genetic modifi-
cations were verified by enzyme assays and gene sequencing.
The primers (restriction sites were underlined) used in this
work are listed in Table S1 in the ESI.† All the oligonucleotide
primers were synthesized at Takara Co. (Dalian, China).
1 (a) I. Mody, Y. De Koninck, T. Otis and I. Soltesz, Trends
Neurosci., 1994, 17, 517–525; (b) C. Wong, T. Bottiglieri and
O. Snead, Ann. Neurol., 2003, 54, S3–S12; (c) Y. Cao and
H. Li, Amino Acids, 2010, 29, 1107–1116.
2 (a) S. I. Choi, J. W. Lee, S. M. Park, M. Y. Lee, J. I. Ge,
M. S. Park and T. R. Heo, J. Microbiol. Biotechnol., 2006, 16,
562–568; (b) J. Huang, L. Mei, H. Wu and D. Lin, World
J. Microbiol. Biotechnol., 2007, 23, 865–871; (c) K. Inoue,
T. Shirai, H. Ochiai, M. Kasao, K. Hayakawa, M. Kimura
and H. Sansawa, Eur. J. Clin. Nutr., 2003, 57, 490–495;
(d) H. Ueno, J. Mol. Catal. B: Enzym., 2000, 10, 67–79.
3 (a) G. Capitani, D. De Biase, C. Aurizi, H. Gut, F. Bossa and
G. Gruetter, EMBO J., 2003, 22, 4027–4037; (b) Y. Su,
J. Wang, T. Lin and T. Pan, J. Ind. Microbiol. Biotechnol.,
2003, 30, 41–46; (c) K. Masuda, X.-f. Guo, N. Uryu,
T. Hagiwara and S. Watabe, Biosci. Biotechnol. Biochem.,
2008, 72, 3265–3272.
4 (a) G. Battaglioli, H. Liu and D. Martin, J. Neurochem.,
2003, 86, 879–887; (b) P. Cheung, C. Fong, K. Ng, W. Lam,
Y. Leung, C. Tsang, M. Yang and M. Wong, J. Biochem.,
2003, 134, 731–738.
5 G. Fenalti, R. H. P. Law, A. M. Buckle, C. Langendorf,
K. Tuck, C. J. Rosado, N. G. Faux, K. Mahmood,
C. S. Hampe, J. P. Banga, M. Wilce, J. Schmidberger,
J. Rossjohn, O. El-Kabbani, R. N. Pike, A. I. Smith,
I. R. Mackay, M. J. Rowley and J. C. Whisstock, Nat. Struct.
Mol. Biol., 2007, 14, 280–286.
Construction of the recombinant C. glutamicum strains
By electroporation, the plasmid pK18-RargB::tacgad was intro-
duced into the competent cells of C. glutamicum G01 to gen-
erate the strain C. glutamicum AG, in which the gad gene was
introduced and argB was knocked out. The recombinant pK18-
RargB::tacgad and pK18-RproB::tacgad were both introduced
into C. glutamicum G01 to construct C. glutamicum APGG, in
which two copies of gad genes were introduced, and argB and
proB were deleted. The recombinant pK18-RargB::tacgad,
pK18-RproB::tacgad and pK18-dapA::tacplk were simul-
taneously introduced into C. glutamicum G01 to create C. gluta-
micum APLGGP, in which plk gene and two copies of gad genes
were introduced, and argB, proB and dapA were knocked out.
Enzyme assays
The activities of GAD, NAGK, GAK, PLK and DHDPS were
assayed according to previously reported methods.21,22 All
measurements were repeated at least three times.
Abbreviations
6 S. Siragusa, M. De Angelis, R. Di Cagno, C. Rizzello,
R. Coda and M. Gobbetti, Appl. Environ. Microbiol., 2007,
73, 7283–7290.
argB
dapA
N-Acetyl-L-glutamate kinase gene
Dihydropicolinate synthase gene
DHDPS Dihydropicolinate synthase
7 H. Li, T. Qiu, D. Gao and Y. Cao, Amino Acids, 2010a, 38,
1439–1445.
GABA
GAD
gad
GAK
NAGK
PLK
plk
γ-Aminobutyric acid
Glutamate decarboxylase
Glutamate decarboxylase gene
Glutamate kinase
N-Acetyl-L-glutamate kinase
Pyridoxal kinase
8 (a) K. Park and S. Oh, Bioresour. Technol., 2007, 98, 1675–
1679; (b) H. Li, T. Qiu, D. Gao and Y. Cao, Amino Acids,
2010a, 38, 1439–1445; (c) Y. Zhang, L. Song, Q. Gao, S. Yu,
L. Li and N. Gao, Appl. Microbiol. Biotechnol., 2012, 94,
1619–1627.
Pyridoxal kinase gene
Glutamate kinase gene.
9 S. Yang, F. Lü, Z. Lu, X. Bie, Y. Jiao, L. Sun and B. Yu,
Amino Acids, 2008, 34, 473–478.
proB
10 M.-C. Kook, J. Korean Appl. Biol. Chem., 2010, 53, 816–820.
11 C. Takahashi, J. Shirakawa, T. Tsuchidate, N. Okai,
K. Hatada, H. Nakayama, T. Tateno, C. Ogino and
A. Kondo, Enzyme Microb. Technol., 2012, 51, 171–176.
Acknowledgements
This work was supported by the Program for New Century
Excellent Talents in University (NCET-10-0459, NCET-13-0833), 12 F. Shi and Y. Li, Biotechnol. Lett., 2011, 33, 2469–2474.
4196 | Green Chem., 2014, 16, 4190–4197
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
13 F. Shi, J. Jiang, Y. F. Li, Y. X. Li and Y. Xie, J. Ind. Microbiol.
Biotechnol., 2013, 40, 1285–1296.
768; (b) E. Bormann, B. Eikmanns and H. Sahm, Mol.
Microbiol., 1992, 6, 317–326.
14 (a) T. Liu, T. Yang, S. Zhang, H. Xia and Z. Rao, J. Food Sci. 19 M. J. Seo, Y. D. Nam, S. Y. Lee, S. L. Park, S. H. Yi
Biotechnol., 2010, 29, 742–747; (b) L. Tian, M. Xu and
Z. Rao, Chin. J. Biotechnol., 2012, 1, 65–75.
and S. I. Lim, Biosci. Biotechnol. Biochem., 2013, 77,
853–856.
15 (a) V. Sakanyan, P. Petrosyan, M. Lecocq, A. Boyen, 20 (a) A. Schäfer, A. Tauch, W. Jäger, J. Kalinowski,
C. Legrain, M. Demarez, J. N. Hallet and N. Glansdorff,
Microbiology, 1996, 142, 99–108; (b) E. Adams and L. Frank,
Annu. Rev. Biochem., 1980, 49, 1005–1061; (c) A. M. Velasco,
G. Thierbach and A. Pühler, Gene, 1994, 145, 69–73;
(b) J. Nešvera and M. Pátek, Appl. Microbiol. Biotechnol.,
2011, 90, 1641–1654.
J. I. Leguina and A. Lazcano, J. Mol. Evol., 2002, 55, 445– 21 D. McCormick, M. Gregory and E. Snell, J. Biol. Chem.,
449. 1961, 236, 2076–2084.
16 K. L. Heckman and L. R. Pease, Nat. Protocols, 2007, 2, 22 (a) S. Kitaoka and Y. Nakano, J. Biochem., 1969, 66, 87–94;
924–932.
(b) M. Xu, Z. Rao, W. Dou, J. Yang, J. Jin and Z. Xu, Amino
Acids, 2012, 43, 255–266; (c) C. Smith, A. Deutch and
K. Rushlow, J. Bacteriol., 1984, 157, 545–551;
(d) C. Dereppe, G. Bold, O. Ghisalba, E. Ebert and
H.-P. Schär, Plant Physiol., 1992, 98, 813–821.
17 A. Schäfer, A. Tauch, W. Jäger, J. Kalinowski, G. Thierbachb
and A. Pühler, Gene, 1994, 145, 69–73.
18 (a) S. Delaunay, P. Gourdon, P. Lapujade, E. Mailly, E. Oriol
and J. Engasser, Enzyme Microb. Technol., 1999, 25, 762–
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 4190–4197 | 4197