RCE

CHa

Kb

c

a

ARRA

KCGGp

1

wtlmteb[psp

GPm

0h

Enzyme and Microbial Technology 51 (2012) 171– 176

Contents lists available at SciVerse ScienceDirect

Enzyme and Microbial Technology

jou rn al h om epage: www.elsev ier .com/ locate /emt

obust production of gamma-amino butyric acid using recombinantorynebacterium glutamicum expressing glutamate decarboxylase fromscherichia coli

hihiro Takahashia, Junki Shirakawaa, Takeyuki Tsuchidatea, Naoko Okaib, Kazuki Hatadaa,ideki Nakayamab, Toshihiro Tatenoc, Chiaki Oginoa, Akihiko Kondoa,∗

Department of Chemical Science and Engineering, Graduate School of Engineering, Graduate School of Science and Technology, Kobe University, 1-1 Rokkodaicho, Nada-ku,obe 657-8501, JapanOrganization of Advanced Science and Technology, Graduate School of Science and Technology, Kobe University, 1-1 Rokkodaicho, Nada-ku, Kobe 657-8501, JapanDepartment of Molecular Science and Material Engineering, Graduate School of Science and Technology, Kobe University, 1-1 Rokkodaicho, Nada-ku, Kobe 657-8501, Japan

r t i c l e i n f o

rticle history:eceived 20 April 2012eceived in revised form 27 May 2012ccepted 28 May 2012

eywords:orynebacterium glutamicum

a b s t r a c t

Gamma-amino butyric acid (GABA) is a component of pharmaceuticals, functional foods, and thebiodegradable plastic polyamide 4. Here, we report a simple and robust system to produce GABA fromglucose using the recombinant Corynebacterium glutamicum strain GAD, which expresses GadB, a gluta-mate decarboxylase encoded by the gadB gene of Escherichia coli W3110. As confirmed by HPLC analysis,GABA fermentation by C. glutamicum GAD cultured at 30 ◦C in GABA Production 1 (GP1) medium con-taining 50 g/L glucose without the addition of glutamate yielded 8.07 ± 1.53 g/L extracellular GABA after

′

amma-amino butyric acidlutamate decarboxylaseyridoxal 5′-phosphate

96 h. Addition of 0.1 mM pyridoxal 5 -phosphate (PLP) was found to enhance the production of GABA,whereas Tween 40 was unnecessary for GABA fermentation. Using the optimized GABA Production 2(GP2) medium, which contained 50 g/L glucose and 0.1 mM PLP, fermentation was performed in a flaskat 30 ◦C with 10% (v/v) seed culture of C. glutamicum GAD. GABA was produced in the culture supernatantwith a yield of 12.37 ± 0.88 g/L after 72 h with a space–time yield of 0.172 g/L/h, which is the highest yieldobtained to date for GABA from fermentation with glucose as a main carbon source.

. Introduction

Gamma-amino butyric acid (GABA) is a non-protein amino acididely found in microorganisms, animals, and plants. GABA func-

ions as a neurotransmitter signal in humans, has blood pressureowering activity [1], and has been used as a component of phar-

aceuticals and functional foods [2]. Recently, it was reportedhat GABA also represents a new building block of bio-plastics. Forxample, polyamide 4 (PA4) is a linear polymer of GABA which cane chemically synthesized from 2-pyrrolidone, a lactam of GABA3]. PA4 has excellent physical properties based on its high melting

oint of 260 ◦C and biodegradability in the soil [4] and in activatedludge [5]. In contrast, polyamide 6, which is commonly used inlastics and nylon materials, is not biodegradable in soil [6]. It is

Abbreviations: BHI, brain heart infusion; GABA, gamma-amino butyric acid;AD, glutamate decarboxylase; LB, Luria–Bertani; MSG, monosodium glutamate;A4, polyamide 4; PLP, pyridoxal 5′-phosphate; Tween 40, polyoxyethylene sorbitanonopalmitate.∗ Corresponding author. Tel.: +81 78 803 6196; fax: +81 78 803 6196.

E-mail address: akondo@kobe-u.ac.jp (A. Kondo).

141-0229/$ – see front matter © 2012 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.enzmictec.2012.05.010

© 2012 Elsevier Inc. All rights reserved.

anticipated that new bio-plastic materials can be synthesized atlow cost from abundantly available biomass resources when GABAis produced by recombinant microorganisms.

GABA was originally produced in traditional fermented foodssuch as Korean kimchi, Chun-gu-chan, yoghurt and cheese by lac-tic acid bacteria, including Lactobacillus brevis, Lactobacillus lactis,and Streptococcus salivarius [7–9]. These bacteria possess intracel-lular glutamate decarboxylase (GAD, EC 4.1.1.15) activities andcan be used for GABA fermentation. GAD catalyzes the alpha-decarboxylation reaction of l-glutamate to GABA [10]. Bacterial gadgenes have been identified in E. coli [11], L. brevis [12], L. paraca-sei [13], and several other Lactobacillus and Enterobacteria species.Using lactic acid bacteria, GABA is produced by adding glutamateto the fermentation medium as a precursor. As this method is notcost-effective for producing chemicals, a new approach for GABAfermentation from biomass is required for the sustainable indus-trial production of monomers for bio-plastics.

The aim of our study was to express the gad gene in a glutamate-

producing microorganism to create an efficient GABA produc-tion process. Corynebacterium glutamicum is a non-pathogenic,non-sporulating, non-motile, Gram-positive soil bacterium belong-ing to the order Actinomycetales, which includes species of

1 icrobi

Cidamainota

tGnadmpm1

2

2

E53mctCw2lf

2

mpwr

TB

72 C. Takahashi et al. / Enzyme and M

orynebacteria, Nocardia, Rhodococci, and other related microorgan-sms [14]. C. glutamicum is an important industrial microorganismue to its high productivity of glutamate and amino acids, whichre widely used in medicine, animal feed, and as food supple-ents [15,16]. Genetically engineered strains of C. glutamicum are

lso superior for producing various kinds of organic compoundsncluding bio-ethanol [17], cadaverine as a component of bio-basedylons [18], and succinic acid as a polymer building block underxygen-deprivation conditions [19,20]. Recently, it was reportedhat recombinant C. glutamicum expressing GAD from L. brevis wasble to produce extracellular GABA [21].

In this study, we developed an efficient system for the produc-ion of GABA using a recombinant C. glutamicum strain expressingAD from E. coli. The results showed that addition of glutamate wasot necessary for GABA production by the C. glutamicum GAD strainnd fermentation conditions were optimized to enhance GABA pro-uction from glucose as the main carbon source. As a result ofedium optimization, we found that the addition of pyridoxal 5′-

hosphate (PLP), which is a cofactor of GAD, in the fermentationedium greatly increased the productivity of GABA to more than

2 g/L after 72 h of fermentation.

. Materials and methods

.1. Bacterial strains and media

The bacterial strains and plasmids used in this study are listed in Table 1.scherichia coli strains were grown in Luria–Bertani (LB) medium (10 g/L tryptone,

g/L yeast extract, and 5 g/L sodium chloride) containing 50 �g/mL kanamycin at7 ◦C. C. glutamicum ATCC 13032 and all recombinant strains were grown in BYedium (10 g/L peptone, 10 g/L meat extract, 5 g/L yeast extract, and 5 g/L sodium

hloride) containing 25 �g/mL kanamycin at 30 ◦C [22]. For the selection of C. glu-amicum transformants, Brain Heart Infusion (BHI) medium (Becton, Dickinson ando., Franklin Lakes, NJ, USA) supplemented with 25 �g/mL kanamycin and 1.5% agaras used. The transformants were first precultivated in 5 mL BHI medium containing

5 �g/mL kanamycin in a test tube at 30 ◦C for 24 h. The culture was then inocu-ated into 20 mL GP medium containing 25 �g/mL kanamycin in a 200-mL flask forermentation.

.2. Construction of plasmids

All genetic manipulations were performed using E. coli SCS110 to avoid DNAethylation, and polymerase chain reactions (PCRs) were conducted using KOD-

lus2 DNA polymerase (Toyobo, Osaka, Japan). C. glutamicum–E. coli shuttle vectorsith the High-Constitutive Expression (HCE) promoter, pCH, were constructed as

eported in our previous study [23].

able 1acterial strains and plasmids used in this study.

Strains or plasmids Relevant characteristics Reference orsource

Escherichia coliSCS110 rpsL (Strr) thr leu endA thi-l

lacY galK galT ara tonA tsxdam dcm supE44�(lac-proAB) [F’traD36 proABlaclqZ�M15]

STRATAGENE

W3110 Wild-type NBRC

Corynebacterium glutamicumATCC 13032 Wild-type ATCCGAD C. glutamicum ATCC13032

harboring pCH-gadBThis study

GAD2 C. glutamicum ATCC13032harboring pCH-gadA

This study

W C. glutamicum ATCC13032harboring pCH

This study

PlasmidspCH E. coli–C. glutamicum

shuttle vector, Kmr[Tateno]

pCH-gadB pCH containing gadB fromE. coli W3110, Kmr

This study

pCH-gadA pCH containing gadA fromE. coli W3110, Kmr

This study

al Technology 51 (2012) 171– 176

Genomic DNA from E. coli W3110 grown in LB medium was purified using theWizard Genomic DNA Purification Kit (Promega, Madison, WI, USA). The GAD gene,gadB (GeneBank accession no. BAA15163.1), from E. coli was amplified by PCR fromthe genomic DNA of E. coli W3110 using the following primer pairs: W3110-gadB F(5′-GGC GAG CTC ATG TTT AAA GCT GTT CTG TTG GGC A-3′ , SacI restriction site isunderlined) and W3110-gadB R (5′-CCG CTC GAG TTA CTT GTC ATC GTC ATC CTT GTAGTC AGG TCG GAA CTA CTC GAT TCA CG-3′ , XhoI restriction site is underlined, and theFLAG-tagged sequence is italicized). The amplified DNA fragment was purified froma 1.0% agarose gel using the Wizard SV Gel and PCR Clean-Up System (Promega) aftergel electrophoresis. The purified 1.37-kbp gadB fragment was digested with SacI andXhoI (New England Biolabs, MA, USA) and cloned into pCH to yield pCH-gadB. Thesequence of the constructed plasmid was confirmed by DNA sequencing analysisusing an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems). A second GADgene, gadA, was cloned and used for expression in C. glutamicum. gadA (GeneBankaccession no. M84024.1) was amplified from E. coli W3110 genomic DNA using thefollowing primer pairs: W3110-gadA F (5′-GGC GAG CTC ATG GAC CAG AAG CTGTTA ACG GAT TT-3′ , SacI sites are underlined) and W3110-gadA R (5′-CCG CTC GAGTCA CTT GTC ATC GTC ATC CTT GTA GTC GGT GTG TTT AAA GCT GTT CTG CTG-3′ ,XhoI sites are underlined, and FLAG-tagged sequence is italicized). The amplified1.40-kbp fragment was cloned into pCH to yield pCH-gadA.

The constructed plasmid pCH-gadB, pCH-gadA or pCH as a control, was individ-ually introduced into C. glutamicum ATCC 13032. The transformation was conductedby electroporation with a 2.5-kV, 200-�, 25-�F electric pulse in a 0.2-cm cuvetteusing a Gene Pulser Xcell (Bio-Rad, Richmond, CA, USA), followed by a heat shock of46 ◦C for 6 min. The cells were then incubated in 1 mL BHI medium at 30 ◦C for 1.5 h.Transformants were selected on BHI agar plates containing 25 �g/mL kanamycin,and the presence of the gadB was confirmed by cell-directed PCR using KOD FX(Toyobo). The resulting strains, C. glutamicum ATCC13032 (pCH-gadB), C. glutam-icum ATCC13032 (pCH-gadA), and C. glutamicum ATCC13032 (pCH), were named C.glutamicum GAD, C. glutamicum GAD2, and C. glutamicum W, respectively.

2.3. Western blotting analysis

C. glutamicum GAD, C. glutamicum GAD2, and C. glutamicum W were preculti-vated in 5 mL BHI medium containing 25 �g/mL kanamycin in a test tube at 30 ◦Cfor 24 h. A small volume (0.2 mL) of the preculture solution was transferred to 20 mLBY medium containing 25 �g/mL kanamycin in a 200 mL shaker flask. After 24 h offermentation, the cells from 1 mL culture were centrifuged at 8000 × g for 5 min,washed once in 50 mM Tris–HCl (pH 6.8) buffer, suspended in 1 mL buffer, and then0.7 g of 0.1-mm glass beads YGB01 (Yasui Kikai, Japan) was added to the tube. Thecells were disrupted using a Shake Master Neo (Bio Medical Science) by shaking thetube three times at 1500 rpm for 1 min with 1-min intervals. After centrifugationat 9000 × g for 5 min, the supernatants were subjected to SDS-PAGE analysis. Theseparated proteins were electroblotted onto a PVDF membrane (Millipore, Boston,MA, USA) and then reacted sequentially with a mouse monoclonal anti-FLAG M2(Sigma, St. Louis, MO, USA) and goat anti-mouse IgG alkaline phosphate conjugate(Promega). The membrane was stained with 4-nitro-blue tetrazolium chloride (NBT;Promega) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP; Promega) according tothe manufacturer’s instructions.

2.4. Culture conditions for GABA fermentation from glucose

C. glutamicum GAD and C. glutamicum W were precultivated in 5 mL BHI mediumcontaining 25 �g/mL kanamycin in a test tube at 30 ◦C for 22 h. A preculture solu-tion (0.2 mL) was transferred to 20 mL GABA Production 1 (GP1) medium containing25 �g/mL of kanamycin in a 200 mL shaker flask. The composition of the GP1medium was 50 g glucose, 50 g (NH4)2SO4, 1 g K2HPO4, 3 g urea, 0.4 g MgSO4·7H2O,50 g soy peptone, 0.01 g FeSO4·7H2O, 0.01 g MnSO4·5H2O, 200 �g thiamine, 0.5 mgbiotin, 0.265 g PLP, and 5 g Tween 40 per liter. Stock solutions of thiamine, biotin,and PLP were filtrated using a 0.22 �m filter membrane and added to the mediumprior to the addition of cells. The pH of the GP1 medium in the flask was not adjusted.The fermentation was carried out at 30 ◦C with an agitation speed of 120 rpm in aBR-13FR BioShaker (Taitec, Japan). Throughout the 120 h cultivation, 1 mL of the cul-ture was collected every 24 h, centrifuged at 8000 × g for 5 min at 4 ◦C, and filtratedthrough a 0.45 �m DISMIC Mixed Cellulose Ester (Advantec, Tokyo, Japan). GABA,glutamate, and glucose concentrations of culture supernatants were analyzed asdescribed below. The optical density at 600 nm (OD600) was monitored at the sametime.

2.5. Determination of the effects of PLP, Tween 40, and cell culture volume onGABA fermentation

For the determination of the effects of Tween 40, 0 (control) and 5 g/L Tween 40(polyoxyethylene sorbitan monopalmitate; Nacalai Tesque, Kyoto, Japan) was addedto the GP1 medium. After 22 h of preculture of C. glutamicum GAD, 0.2 mL of the

preculture solution was transferred to 20 mL of GP1 medium containing 25 �g/mLkanamycin in a 200 mL shaker flask. For evaluation of the effects of Tween 40, theconcentration of PLP in the medium was fixed at 1 mM.

For determination of the effects of the coenzyme PLP on GABA fermentation,0 (control), 0.1 mM, and 1 mM PLP was added to the GP1 medium. After 22 h of

icrobi

p2st

tst1

Gmip3c

2c

waSPpswc0GamT

3

3

Eipahadiia

Fc2ogm

C. Takahashi et al. / Enzyme and M

reculture, 0.2 mL of the preculture solution of C. glutamicum GAD was transferred to0 mL PLP-supplemented GP1 medium containing 25 �g/mL kanamycin in a 200 mLhaker flask. For evaluation of the effects of PLP, the concentration of Tween 40 inhe medium was fixed at 5 g/L.

For determination of the optimal cell culture volume to be added to the fermen-ation culture, 1% (v/v), 5% (v/v), or 10% (v/v) of the C. glutamicum GAD precultureolution grown in BHI medium was transferred to 20 mL GP1 medium in a flask. Inhis assay, the concentrations of PLP and Tween 40 were fixed in the medium at

mM and 5 g/L, respectively.For optimization of the fermentation conditions, 0.1 mM PLP was added to the

P1 medium, and the resultant medium for GABA fermentation was named GP2edium. The cell culture volume of precultivated C. glutamicum GAD or C. glutam-

cum W added to 20 mL GP2 medium in a culture flask was set at 2 mL (10% (v/v)). TheH of the medium in each flask was unadjusted. Fermentations were performed at0 ◦C with an agitation speed of 120 rpm. All GP medium used for the fermentationontained 25 �g/mL kanamycin.

.6. Analysis of cell growth, GABA and L-glutamate production, and glucoseonsumption

The growth of C. glutamicum strains was monitored by measuring OD600

ith a UVmini-1240 UV-vis spectrophotometer (Shimadzu, Kyoto, Japan). GABAnd L-glutamate concentrations in the supernatant were analyzed using ahim-pack Amino-Li column (0.5 �m, 100 mm × 6.0 mm I.D.; Shimadzu) using arominence Amino Acid Analyzer System (Shimadzu) after derivation with ortho-hthalaldehyde (OPA). The mobile phase was a gradient of a lithium citrate–borateolution ranging in pH from 2.68 to 10.00 at a flow rate of 0.6 mL/min, and the columnas maintained at 39 ◦C. Amino acid mixtures Type AN-II and Type B (Wako Chemi-

als, Japan) were used as standards. In addition, 0.1 mM GABA (Nacalai Tesque) and.1 mM L-glutamate in sodium citrate buffer (pH 2.2) were also used as standards.lucose was analyzed using a Prominence HPLC System (Shimadzu) equipped with

SPR-Pb column (0.5 �m, 250 mm × 4.0 mm I.D.; Shimadzu). Water was used as theobile phase at a flow rate of 0.6 mL/min and the column was maintained at 80 ◦C.

he peak elution profile was monitored using a refractive index detector.

. Results

.1. Construction of a GAD-expressing C. glutamicum strain

The gadA and gadB genes amplified from the genomic DNA of. coli W3110 were separately inserted into pCH for construct-ng pCH-gadB and pCH-gadA, respectively. Each GAD-expressionlasmid was introduced into C. glutamicum ATCC 13032, which is

glutamate producing strain. The resultant recombinant strainsarboring pCH-gadB or pCH-gadA were named C. glutamicum GADnd C. glutamicum GAD2, respectively. As a control, pCH was intro-

uced into the same host to generate strain C. glutamicum W. The

ntracellular expression levels of GAD in the engineered C. glutam-cum strains were monitored by western blotting analysis usingn antibody raised against the FLAG-tagged sequence that was

ig. 1. Time course of extracellular GABA production by C. glutamicum GAD and C. glutamontaining 25 �g/mL kanamycin for 22 h, and 0.2 mL (1% [v/v]) of each preculture solutio0 mL of GP1 medium containing 1.0 mM PLP and 0.5% (w/v) Tween 40 using a 200-mL sf 120 rpm. Every 24 h, 1 mL of the culture supernatant was collected and assayed for GAlucose consumption (open symbols) of C. glutamicum GAD (circles) and C. glutamicum Wonitored throughout the fermentation. Data are expressed as mean and standard error

al Technology 51 (2012) 171– 176 173

incorporated into the cloned genes. The 53 kDa protein signal ofGadB was detected in the cytoplasmic fraction of C. glutamicumGAD cultured for 24 h in BY medium, whereas no protein signal wasobserved in the cytoplasmic fraction of C. glutamicum W as the neg-ative control (data not shown). Thus, this indicated the successfulintracellular expression of GadB in C. glutamicum.

The protein signal corresponding to GadA of E. coli was alsoobserved in the cytoplasmic fraction of C. glutamicum GAD2 (datanot shown). As the protein expression level of GadB from C. glu-tamicum GAD was the highest among the strains tested, we choseC. glutamicum GAD for the fermentation of GABA.

3.2. GABA fermentation by C. glutamicum GAD

Using C. glutamicum GAD, we attempted to perform GABAfermentation directly from glucose. First, we set the compositionof the GABA Production 1 (GP1) medium, which is an originalmedium for GABA fermentation developed in this study, based onthe GH1 medium used for glutamate production by C. glutamicum[24]. The GP1 medium contained 50 g/L glucose and soy peptoneas carbon and nitrogen sources, respectively, and was also supple-mented with 5 g/L Tween 40 to sustain glutamate production inthe batch culture. In addition, as a cofactor of GAD, 1.0 mM PLP wasadded to the GP1 medium. After adding 1% (v/v) of a preculturesolution of C. glutamicum GAD or C. glutamicum W to 20 mL GP1medium, the fermentation was performed at 30 ◦C for 120 h withshaking at 120 rpm, and the supernatant was periodically assayedfor GABA and glucose (Fig. 1). As the glucose was consumed, GABAformation was observed in the culture supernatant of C. glutam-icum GAD during the stationary phase (Fig. 1a and b). The GABAconcentration in the culture supernatant reached 8.07 ± 0.53 g/L in120 h, with a space–time yield of 0.0672 g/L/h (Fig. 1a). In contrast,C. glutamicum W did not produce any detectable GABA in the cul-ture supernatant (Fig. 1a). Under these fermentation conditions,C. glutamicum GAD consumed 45.17 ± 0.30 g/L of glucose within120 h (Fig. 1b). The productivity of GABA production from glucoseby C. glutamicum GAD reached 0.311 mol/mol.

3.3. Effects of cell concentration and Tween 40 on GABAfermentation by C. glutamicum GAD

For the stable production of GABA by C. glutamicum GAD, theeffects of the medium composition and culture conditions wereinvestigated. First, we examined the effects of the concentration of

icum W. C. glutamicum GAD and C. glutamicum W were precultured in BHI mediumn was transferred to GP1 medium. The two strains were cultured independently inhake flask. Fermentation was performed at 30 ◦C for 120 h with an agitation speedBA, glucose and OD600. (a) Extracellular GABA concentrations (closed symbols) and

(triangles) were monitored. (b) The OD600 (open symbols) of the two strains werefrom three independent experiments.

174 C. Takahashi et al. / Enzyme and Microbial Technology 51 (2012) 171– 176

Fig. 2. Effect of Tween 40 in the culture medium on GABA production by C. glutamicum GAD. C. glutamicum GAD was precultured in BHI medium containing 25 �g/mLkanamycin for 22 h, and 0.2 mL of seed culture was transferred to GP1 medium with or without Tween 40. For GABA fermentation, C. glutamicum GAD was cultured in 20 mLof GP1 medium containing 0.5% (w/v) Tween 40 (squares) in a shake flask at 30 ◦C for 120 h. GP1 medium without Tween 40 (circles) was used as a control. (a) Extracellularg h. (b)G as m

ctamfst

aecfptgatt3tstoToG

FcGP(t

lutamate (open symbols) and GABA (closed symbols) levels were monitored for 120AD in each flask were monitored throughout the fermentation. Data are expressed

ells added to the fermentation culture broth. The preculture solu-ion of C. glutamicum GAD cells grown in BHI medium was addedt concentrations of 1% (v/v), 5% (v/v), or 10% (v/v) to 20 mL GP1edium containing 1 mM PLP and 5 g/L Tween 40. After 96 h of

ermentation, the batch culture initiated with 10% (v/v) precultureolution exhibited a GABA yield of 6.5 g/L, which was higher thanhe other examined culture conditions (data not shown).

During glutamate fermentation by C. glutamicum GAD, theddition of Tween 40 is reported to promote the production ofxtracellular glutamate by assisting its permeability through theell membrane [25]. As we hypothesized that the intracellularormation and secretion of glutamate might also promote theroduction of GABA, we examined the effects of different concen-rations of Tween 40 in the medium on GABA fermentation by C.lutamicum GAD. For the assay, the concentration of PLP was fixedt 1.0 mM, while 0 g/L (control) or 5.0 g/L Tween 40 was added tohe GP1 medium (Fig. 2). After 120 h of fermentation, the produc-ivity of GABA was 5.25 g/L in the media without Tween40 and.18 g/L in the media with 5.0 g/L Tween40. The peak concentra-ion of GABA formation in the medium without added Tween 40hifted at 96 h, with a yield of 8.22 ± 2.05 g/L (Fig. 2a). In addition,he growth rate of C. glutamicum GAD in the GP1 medium with-

ut Tween 40 was higher than that in the medium containing 5 g/Lween 40 (Fig. 2b). From this result, we concluded that the additionf Tween 40 did not affect the production of GABA by C. glutamicumAD.

ig. 3. Effect of pyridoxal 5′-phosphate (PLP) addition to the culture broth of C. glutamicontaining 25 �g/mL of kanamycin for 22 h, and 0.2 mL of seed culture was transferred toAD was cultured in 20 mL modified GP1 medium containing 0.1 mM (closed circles) or 1LP (open circles) was used as a control. (a) Extracellular GABA (solid line) and glutamate (solid line) and OD600 (dotted line) of C. glutamicum GAD in each flask were monitored three independent experiments.

Glucose consumption (closed symbols) and OD600 (open symbols) of C. glutamicumean and standard error from three independent experiments.

3.4. Effect of PLP on GABA fermentation by C. glutamicum GAD

As the E. coli GAD requires PLP as a cofactor for its activity [26],we anticipated that the addition of an appropriate concentration ofPLP to the medium might affect the GABA yield by C. glutamicumGAD. The PLP concentration in the medium was further adjustedin the range of 0.1–1.0 mM for performing stable GABA fermenta-tion in shaker flask cultures with C. glutamicum GAD (Fig. 3). As acontrol, GP1 medium without PLP was used with the same strain.The extracellular concentration of GABA in the culture of C. glutam-icum GAD was higher in the PLP-containing GP1 medium than inthe medium without PLP. The productivity of GABA after 120 h offermentation was 3.26 g/L in the medium containing 1.0 mM PLP,6.06 g/L in the medium containing 0.1 mM PLP, and 2.15 g/L in themedium without PLP (Fig. 3a). Thus, the addition of 0.1 mM PLPto the GP1 medium was the most effective for prolonging GABAproduction by C. glutamicum GAD.

3.5. GABA production by C. glutamicum GAD under optimizedconditions

Using the selected culture conditions described above (0.1 mM

PLP and 0 g/L Tween 40), we performed GABA fermentation fromglucose using C. glutamicum GAD and C. glutamicum W as a negativecontrol. The optimized medium for GABA fermentation also con-tained 50 g/L glucose and was named GABA production medium

um GAD on GABA production. C. glutamicum GAD was precultured in BHI medium modified GP1 medium with or without PLP. For GABA fermentation, C. glutamicum.0 mM PLP (closed squares) in a shake flask at 30 ◦C for 120 h. GP1 medium withoutdotted line) levels in each flask were monitored for 120 h. (b) Glucose consumptionhroughout the fermentation. Data are expressed as mean and standard error from

C. Takahashi et al. / Enzyme and Microbi

Fig. 4. GABA production by C. glutamicum GAD under optimized conditions. C. glu-tamicum GAD and C. glutamicum W were precultured in BHI medium containing25 �g/mL of kanamycin for 22 h. Two milliliters (10% [v/v]) of seed culture wastransferred to 20 mL of optimized GP2 medium in a shake flask, and fermentationwas performed at 30 ◦C for 120 h with shaking at 120 rpm. The culture supernatantsoaa

2(if

tci7imtwG0

4

sEtmttgsf1yts(

wcTtofGpn

f C. glutamicum GAD (circles) and C. glutamicum W (diamonds) were collected andssayed for GABA (closed symbols) and glucose concentrations (open symbols). Datare expressed as mean and standard error from three independent experiments.

(GP2). To examine GABA production using this medium, 2 mL10% [v/v]) preculture solution of C. glutamicum GAD or C. glutam-cum W was added to 20 mL GP2 medium in a 200 mL flask, andermentation was performed at 30 ◦C with shaking at 120 rpm.

As the glucose in the GP2 medium consistently decreased fromhe beginning of the fermentation, the concentration of extra-ellular GABA produced by C. glutamicum GAD simultaneouslyncreased, reaching a maximum level of 12.37 ± 0.88 g/L after2 h (Fig. 4). The GABA production yield by C. glutamicum GAD

ncreased 1.53-fold in GP2 medium compared with that in GP1edium (Fig. 1). Under the optimized conditions for the fermenta-

ion of GABA by C. glutamicum GAD, the consumption of glucoseas 26.31 ± 2.83 g/L in 72 h, and the maximum productivity ofABA reached 52.3% (mol/mol). The space–time yield of GABA was.172 g/L/h after 72 h.

. Discussion

In this study, we established a simple and robust productionystem for GABA using a C. glutamicum strain expressing GAD from. coli. Under the optimized culture conditions, which includedhe addition of PLP to the medium, C. glutamicum GAD displayed

arked GABA-producing ability. For biotechnological applications,he selection of an appropriate host microorganism is a key fac-or that governs the efficient production of target materials. As C.lutamicum overproduces the GABA precursor L-glutamate fromugar, we focused on engineering an L-glutamate-producing strainor GABA synthesis. Introducing GadB from E. coli into strain ATCC3032 led to the initial production of GABA from glucose with aield of 8.07 ± 1.53 g/L after 96 h (Fig. 1). Thus, the direct produc-ion of GABA from glucose was successful using this recombinanttrain without the addition of glutamate or monosodium glutamateMSG) as precursors of GABA.

In an attempt to optimize the fermentation conditions of GABA,e investigated the effects of Tween 40, PLP, and initial cell con-

entration on the yields of GABA produced by C. glutamicum GAD.he addition of Tween 40 was necessary for the production of glu-amate by the wild type strain of C. glutamicum under conditionsf biotin supplementation [24]. However, the removal of Tween 40

rom the GP1 medium led to a faster growth rate of C. glutamicumAD, and a steady GABA synthesis (Fig. 2). Therefore, for the GABAroduction by C. glutamicum GAD, no addition of any detergent waseeded. The clear diminishment of extracellular glutamate levels

al Technology 51 (2012) 171– 176 175

in the culture medium of C. glutamicum GAD suggested the efficientintracellular conversion of glutamate to GABA (Fig. 2a). The additionof 0.1 mM PLP, which is a necessary coenzyme of E. coli GAD, to theGP1 medium was most effective at prolonging GABA fermentation(Fig. 3). GADs of E. coli (GadA and GadB) are homohexamers com-prised of proteins with a molecular weight of 53 kDa [10]. Structuraland functional studies of E. coli GadB have revealed that PLP bindsto a lysine residue near the active site of the enzyme [26]. A portionof cellular GAD is present as apoGAD, which is not bound to PLP andremains inactive until forming a GAD-PLP complex (holoGAD) [27].Thus, a sufficient concentration of PLP (0.1 mM) in the GP1 mediumwas necessary to complex with intracellular GadB expressed in C.glutamicum GAD to form holoGAD, leading to efficient GABA pro-duction.

For food applications, GABA has been produced using lacticacid bacteria by the addition of glutamate or MSG to the fer-mentation medium. For example, Streptococcus salivarius subsp.thermophilus Y2, a cheese starter strain, produced 7.98 g/L of GABAafter 84 h of fermentation with continuous feeding of 15 g/L MSG,corresponding to a rate of 0.095 g/L/h [9]. Lactobacillus paraca-sei NFRI 7415, which was isolated from fermented fish, produced31.11 g/L (302 mM) GABA in 168 h, corresponding to a produc-tion rate of 0.185 g/L/h. Although the production rate by strainNFRI 7415 was relatively high, 500 mM (73.5 g/L) glutamate wasadded to the culture medium [28]. These microbial-based produc-tion systems involving supplementation with amino acids are notcost-effective for applications such as synthesizing chemicals. Incontrast, in our GABA production system using C. glutamicum GAD,high concentrations of GABA were produced from glucose in GP1medium without the addition of glutamate during the fermenta-tion. Using this recombinant strain of C. glutamicum, it is expectedthat fewer fermentation by-products will be produced and collec-tion of the target product will be simpler than the methods usingnon-genetically modified lactic acid bacteria. It was also reportedthat, using a recombinant C. glutamicum strain expressing gadRCB2from Lactobacillus, 2.16 ± 0.15 g/L GABA was produced in 72 h when160 g/L of glucose was added in the initial culture medium [21].Under our optimized conditions using GP2 medium, GABA wasproduced from glucose with a yield of 12.37 ± 0.88 g/L after 72 h,corresponding to a space–time yield of 0.172 g/L/h (Fig. 4), whichwas the highest yield of GABA from a fermentation process usingglucose as the main carbon source reported to date.

Thus, future work will involve the development of a process toproduce GABA from abundantly available plant biomass such asstarch or cellulose. We have previously developed a cell-surfacedisplay system of amylase in C. glutamicum and have performedlysine production from starch [23]. In addition, secretion systemsof endoglucanases in C. glutamicum can be applied to the utiliza-tion of glucan during GABA fermentation [29]. In the future, theproduction of GABA using our GABA fermentation system wouldallow the synthesis of 100% biomass-derived nylon PA4. Notably,as C. glutamicum is a microorganism with generally recognized assafe (GRAS) status, the system for GABA fermentation developedin the present study can be applied to the production of GABA as acomponent of functional foods and pharmaceuticals.

Acknowledgments

This work was partially supported by Special CoordinationFunds for Promoting Science and Technology, Creation of Inno-vation Centers for Advanced Interdisciplinary Research Areas

(Innovative Bioproduction Kobe) from the Ministry of Education,Culture, Sports, Science and Technology of Japan. We would liketo thank Dr. Pulla Kaothien-Nakayama for critical reading of thepaper.

1 icrobi

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

76 C. Takahashi et al. / Enzyme and M

eferences

[1] Hayakawa K, Kimura M, Kasaha K, Matsumoto K, Sansawa H, Yamori Y. Effectof a gamma-aminobutyric acid-enriched dairy product on the blood pressureof spontaneously hypertensive and normotensive Wistar-Kyoto rats. BritishJournal of Nutrition 2004;92:411–7.

[2] Cao YS, Li HX. Lactic acid bacterial cell factories for gamma-aminobutyric acid.Amino Acids 2010;39:1107–16.

[3] Kawasaki N, Nakayama A, Yamano N, Takeda S, Kawata Y, Yamamoto N, et al.Synthesis, thermal and mechanical properties and biodegradation of branchedpolyamide 4. Polymer 2005;46:9987–93.

[4] Hashimoto K, Hamano T, Okada M. Degradation of several polyamides in soils.Journal of Applied Polymer Science 1994;54:1579–83.

[5] Yamano N, Nakayama A, Kawasaki N, Yamamoto N, Aiba S. Mechanism andcharacterization of polyamide 4 degradation by Pseudomonas sp. Journal ofPolymers and the Environment 2008;16:141–6.

[6] Hashimoto K, Sudo M, Ohta K, Sugiyama T, Yamada H, Aoki H. Biodegrada-tion of Nylon4 and its blend with Nylon6. Journal of Applied Polymer Science2002;86:2307–11.

[7] Park KB, Oh SH. Cloning, sequencing and expression of a novel glutamate decar-boxylase gene from a newly isolated lactic acid bacterium, Lactobacillus brevisOPK-3. Bioresource Technology 2007;98:312–9.

[8] Park KB, Oh SH. Enhancement of gamma-aminobutyric acid productionin Chungkukjang by applying a Bacillus subtilis strain expressing gluta-mate decarboxylase from Lactobacillus brevis. Biotechnology Letters 2006;28:1459–63.

[9] Yang SY, Lu FX, Lu ZX, Bie XM, Jiao Y, Sun LJ, et al. Production of gamma-aminobutyric acid by Streptococcus salivarius subsp thermophilus Y2 undersubmerged fermentation. Amino Acids 2008;34:473–8.

10] Fonda ML. L-glutamate decarboxylase from bacteria. Methods in Enzymology1985;113:11–6.

11] DeBiase D, Tramonti A, John RA, Bossa F. Isolation, overexpression, and bio-chemical characterization of the two isoforms of glutamic acid decarboxylasefrom Escherichia coli. Protein Expression and Purification 1996;8:430–8.

12] Hiraga K, Ueno YH, Oda KH. Glutamate decarboxylase from Lactobacillus brevis:activation by ammonium sulfate. Bioscience, Biotechnology, and Biochemistry2008;72:1299–306.

13] Komatsuzaki N, Nakamura T, Kimura T, Shima J. Characterization ofglutamate decarboxylase from a high gamma-aminobutyric acid (GABA)-producer, Lactobacillus paracasei. Bioscience, Biotechnology, and Biochemistry2008;72:278–85.

14] George M. Burgey’s manual of systematic bacteriology. 2nd ed. New York:Springer–Verlag; 2001.

15] Leuchtenberger W, Huthmacher K, Drauz K. Biotechnological production ofamino acids and derivatives: current status and prospects. Applied Microbi-ology and Biotechnology 2005;69:1–8.

[

al Technology 51 (2012) 171– 176

16] Hermann T. Industrial production of amino acids by Coryneform bacteria. Jour-nal of Biotechnology 2003;104:155–72.

17] Sakai S, Tsuchida Y, Okino S, Ichihashi O, Kawaguchi H, Watanabe T, et al. Effectof lignocellulose-derived inhibitors on growth of and ethanol production bygrowth-arrested Corynebacterium glutamicum R. Applied and EnvironmentalMicrobiology 2007;73:2349–53.

18] Tateno T, Okada Y, Tsuchidate T, Tanaka T, Fukuda H, Kondo A. Direct pro-duction of cadaverine from soluble starch using Corynebacterium glutamicumcoexpressing alpha-amylase and lysine decarboxylase. Applied Microbiologyand Biotechnology 2009;82:115–21.

19] Inui M, Murakami S, Okino S, Kawaguchi H, Vertes AA, Yukawa H. Metabolicanalysis of Corynebacterium glutamicum during lactate and succinate produc-tions under oxygen deprivation conditions. Journal of Molecular Microbiologyand Biotechnology 2004;7:182–96.

20] Okino S, Noburyu R, Suda M, Jojima T, Inui M, Yukawa H. An efficient suc-cinic acid production process in a metabolically engineered Corynebacteriumglutamicum strain. Applied Microbiology and Biotechnology 2008;81:459–64.

21] Shi F, Li Y. Synthesis of gamma-aminobutyric acid by expressing Lactobacil-lus brevis-derived glutamate decarboxylase in the Corynebacterium glutamicumstrain ATCC 13032. Biotechnology Letters 2011;33:2469–74.

22] Katsumata R, Ozaki A, Oka T, Furuya A. Protoplast transformation ofglutamate-producing bacteria with plasmid DNA. Journal of Bacteriology1984;159:306–11.

23] Tateno T, Fukuda H, Kondo A. Direct production of l-lysine from raw corn starchby Corynebacterium glutamicum secreting Streptococcus bovis alpha-amylaseusing cspB promoter and signal sequence. Applied Microbiology and Biotech-nology 2007;77:533–41.

24] Yao WJ, Zhong H, Liu M, Zheng P, Sun ZH, Zhang Y, et al. Double deletionof dtsR1 and pyc induce efficient L-glutamate overproduction in Corynebac-terium glutamicum. Journal of Industrial Microbiology and Biotechnology2009;36:911–21.

25] Duperray F, Jezequel D, Ghazi A, Letellier L, Shechter E. Excretion of glutamatefrom Corynebacterium glutamicum triggered by amine surfactants. Biochimicaet Biophysica Acta 1992;1103:250–8.

26] Capitani G, De Biase D, Aurizi C, Gut H, Bossa F, Grutter MG. Crystal structure andfunctional analysis of Escherichia coli glutamate decarboxylase. EMBO Journal2003;22:4027–37.

27] Bazhulina NP, Darii EL, Lobachev VM, Stel’mashchuk VY, Sukhareva BS. Struc-ture of glutamate decarboxylase from E. coli: spectral studies. Journal ofBiomolecular Structure and Dynamics 2002;19:999–1006.

28] Komatsuzaki N, Kawamoto S, Momose H, Kimura T, Shima J. Production of

gamma-aminobutyric acid (GABA) by Lactobacillus paracasei isolated from tra-ditional fermented foods. Food Microbiology 2005;22:497–504.

29] Tsuchidate T, Tateno T, Okai N, Tanaka T, Ogino C, Kondo A. Glutamateproduction from beta-glucan using endoglucanase-secreting Corynebacteriumglutamicum. Applied Microbiology and Biotechnology 2011;90:895–901.