Metabolic evolution of Corynebacterium glutamicum for increased production of L-ornithine

Background L-ornithine is effective in the treatment of liver diseases and helps strengthen the heart. The commercial applications mean that efficient biotechnological production of L-ornithine has become increasingly necessary. Adaptive evolution strategies have been proven a feasible and efficient technique to achieve improved cellular properties without requiring metabolic or regulatory details of the strain. The evolved strains can be further optimised by metabolic engineering. Thus, metabolic evolution strategy was used for engineering Corynebacterium glutamicum to enhance L-ornithine production. Results A C. glutamicum strain was engineered by using a combination of gene deletions and adaptive evolution with 70 passages of growth-based selection. The metabolically evolved C. glutamicum strain, named ΔAPE6937R42, produced 24.1 g/L of L-ornithine in a 5-L bioreactor. The mechanism used by C. glutamicum ΔAPE6937R42 to produce L-ornithine was investigated by analysing transcriptional levels of select genes and NADPH contents. The upregulation of the transcription levels of genes involved in the upstream pathway of glutamate biosynthesis and the elevated NADPH concentration caused by the upregulation of the transcriptional level of the ppnK gene promoted L-ornithine production in C. glutamicum ΔAPE6937R42. Conclusions The availability of NADPH plays an important role in L-ornithine production in C. glutamicum. Our results demonstrated that the combination of growth-coupled evolution with analysis of transcript abundances provides a strategy to engineer microbial strains for improving production of target compounds.

Background L-ornithine, a non-essential amino acid and an important constituent of the urea cycle, is the precursor of other amino acids, such as citrulline and arginine. It is effective for the treatment and prophylaxis of liver diseases [1], and has also been applied to wound healing [2]. Recently, it was demonstrated that L-ornithine supplementation increased serum levels of growth hormone and insulin-like growth factor-1 after heavy-resistance exercise in strength-trained athletes [3]. Many studies have reported that high yields of L-ornithine can be produced from a citrulline-or arginine-requiring mutant of a coryneform bacterium obtained by classical mutagenesis [4][5][6][7]. Although this mutant can produce a high yield of L-ornithine, its culture is unstable owing to reversion of the auxotrophic phenotype, which causes the production of L-ornithine to drop markedly. Several recent reports have described progress in metabolic engineering of microorganisms for L-ornithine production. Lee and Cho reported that an engineered Escherichia coli produced 13.2 mg L-ornithine per gram of dry cell weight (DCW), and that addition of glutamate to the culture favoured L-ornithine production in the engineered E. coli [8]. Hwang et al. reported that cooverexpression of argCJBD in a triple-gene knockout strain C. glutamicum ATCC 13032 (ΔargFΔargRΔproB) resulted in a cellular L-ornithine content of 16.49 mg/g DCW and a concentration of L-ornithine in the culture medium of 179.14 mg/L [9]. Proline can be converted into L-ornithine by ornithine cyclodeaminase, which is a key enzyme responsible for enhancing L-ornithine production by C. glutamicum in proline-supplemented media [10]. Huang and Cho reported that over-expression of the Ncgl1469 open reading frame, exhibiting N-acetylglutamate synthase activity, increased L-ornithine production in C. glutamicum by 39% [11]. Recently, the same workers deleted the gluconate kinase gene gntK of C. glutamicum ATCC 13032 (ΔargFΔargR) to obtain C. glutamicum SJC8399, which produced 13.16 g/L of L-ornithine [12]. The L-ornithine producing strain C. glutamicum ATCC 13032 (ΔargFΔargR) named ORN1 was constructed and shown to produce L-ornithine from arabinose when araBAD from E. coli was expressed [13]. Recently, this group also constructed an engineered C. glutamicum ORN1 (pEKEx3-xylA Xc -xylB Cg ) to effectively produce Lornithine from xylose [14]. In our previous paper [15], we constructed a strain of C. glutamicum in which three genes had been deleted. This strain, named ATCC13032 (ΔargFΔproBΔkgd), produced L-ornithine of 18.17 g/L in the optimal medium [16].
Adaptive evolution strategies have been proven a feasible and efficient technique to achieve improved cellular properties without requiring metabolic or regulatory details of the strain [17][18][19][20]. The defining feature of adaptive evolution involves applying a selection pressure that favours the growth of mutants with the traits of interest. Growth-coupled adaptive evolution can significantly increase yields [21]. When combined with metabolic engineering, adaptive evolution is known as metabolic evolution engineering. The evolved strains can be further optimised by metabolic engineering. Metabolic evolution has been successfully employed for the improved production of succinate [20,22], L-alanine [23], and dihydroxyacetone [24]. However, to our knowledge, metabolic evolution engineering has never been reported to boost the production of L-ornithine.
In this study, we first deleted the speE gene of C. glutamicum ATCC 13032 (ΔargFΔproB) to obtain C. glutamicum ATCC 13032 (ΔargFΔproBΔspeE), which was then modified using growth-coupled adaptive evolution to improve L-ornithine production. After comparing the transcriptional levels of select genes of the evolved strain with those of the parent strain, additional genetic modifications were introduced to the evolved strain to further improve L-ornithine production. The mechanism of L-ornithine production by the metabolic evolved strain was also investigated by analysing transcriptional levels of genes of interest, as well as NADPH concentrations.

Results
Construction of the triple gene deletion C. glutamicum for adaptive evolution In our previous paper [15], proteomic analysis demonstrated that the spermidine synthase encoded by the speE gene was more abundant in C. glutamicum engineered to overproduce L-ornithine than in wild-type C. Glutamicum. The upregulation might result in the degradation of L-ornithine. Thus, we deleted the speE gene of C. glutamicum (ΔargFΔproB) to obtain C. glutamicum ΔAPE. As expected, the deletion of the speE gene enhanced L-ornithine production. The C. glutamicum ΔAPE strain produced 11.3 ± 0.3 g/L of L-ornithine, which is higher than that of C. glutamicum (ΔargFΔproB) (10.2 ± 0.2 g/L; Table 1). Thus, C. glutamicum ΔAPE was used as the parent strain for adaptive evolution.

Adaptive evolution
To improve L-ornithine production, C. glutamicum ΔAPE was subjected to adaptive evolution driven by growth-based selection. This process comprised two stages ( Figure 1). First, glucose and L-ornithine were added into the fermentation medium to overcome their inhibitions for growth and overproduction of metabolite. Screening a mutant resistant to substrate and end product is one common strategy for strain improvement by classical mutagenesis. At the later stage, only glucose was added at gradually increasing levels, and no L-ornithine was needed. After 70 days of adaptive evolution, one of the clones, referred to as C. glutamicum ΔAPE 6937, was chosen for further study. The C. glutamicum ΔAPE 6937 strain produced 13.6 ± 0.5 g/L of L-ornithine. The yield was about 20% higher than that of the parent strain C. glutamicum ΔAPE (11.3 ± 0.3 g/L; Table 1). However, the L-ornithine content of the evolved strain C. glutamicum ΔAPE 6937 was similar to that of the parent strain C. glutamicum ΔAPE. This suggests that the increased Lornithine level of the evolved strain C. glutamicum ΔAPE 6937 is caused by the increased cell density.
Characterisation of the evolved strain using qRT-PCR and sequence analysis In order to understand the mechanism of the increased L-ornithine level in the evolved strain, we analysed the transcriptional levels of genes that encode enzymes involved in L-ornithine biosynthesis in the evolved strain C. glutamicum ΔAPE6937. These genes comprise pgi (encoding glucose-6-phosphate isomerase), pfkA (encoding 6-phosphofructokinase), gap (encoding glyceraldehyde-3-phosphate dehydrogenase), pyk (encoding pyruvate kinase), pyc (encoding pyruvate carboxylase), gltA (encoding citrate synthase), gdh (encoding glutamate dehydrogenase), argB (encoding acetylglutamate kinase) and argJ (encoding the bifunctional ornithine acetyltransferase/N-acetylglutamate synthase). And then we compared the data with that obtained in the parent strain C. glutamicum ΔAPE. The results are presented in Figure 2A. All of these genes in the evolved strain  C. glutamicum ΔAPE6937 are upregulated, with the smallest degree of upregulation seen for the argB gene.
In an attempt to identify mutations that confer the high-yield phenotype, we sequenced the above genes of the evolved strain ΔAPE6937 and compared the results with sequences of the same genes from C. glutamicum ATCC 13032. Analysis of the sequences of the pgi, pfkA, gapA, pyk, pyc, gltA, gdh, argB, and argJ genes in the evolved strain failed to identify any mutations relative to the parent strain.

Genetic modification of the evolved strain
The results shown in Figure 2A suggest that expression of the argB gene may be the bottleneck for L-ornithine production by the evolved strain. Thus, we first overexpressed either of the C. glutamicum or E. coli argB genes in the evolved strain. Over-expression of either of the two argB genes indeed enhanced L-ornithine concentration and content, with over-expression of C. glutamicum argB increasing levels of L-ornithine concentration to a greater degree than that achieved by the over-expression of E. coli argB ( Table 1).
The expression of the arg operon for control of the L-ornithine biosynthesis pathway is regulated by the arginine repressor ArgR. In addition, the DNA-binding affinity of ArgR to the upstream of argB gene was suggested to play an important role in L-ornithine biosynthesis in C. glutamicum [25]. Deletion of the argR gene is another strategy for enhancing the level of expression of the arg operon. Thus, we deleted the argR gene of the evolved strain C. glutamicum ΔAPE6937 to generate C. glutamicum ΔAPE6937R42. As expected, C. glutamicum ΔAPE6937R42 produced 17.3 ± 0.4 g/L of L-ornithine ( Table 1). The concentration of L-ornithine was 27% higher than that of C. glutamicum ΔAPE6937.
To investigate the mechanism of L-ornithine production in C. glutamicum ΔAPE6937R42, we analysed the transcriptional levels of the genes that encode enzymes involved in L-ornithine biosynthesis in C. glutamicum ΔAPE6937R42 by using qRT-PCR, and compared the data with that obtained in C. glutamicum ΔAPE ( Figure 2B). Deletion of the argR gene promoted the upregulations of the transcript levels of the pgi, pfkA, argB, and argJ genes. The respective transcriptional levels of the pgi, pfkA, argB, and argJ genes in C. glutamicum ΔAPE6937R42 are about 5.6-, 5.0-, 9.4-, and 16.6-fold (p < 0.05) higher than those in the parent strain C. glutamicum ΔAPE. In contrast, the respective levels of the same transcripts in C. glutamicum ΔAPE6937 are only about 3.7-, 3.5-, 2.3-, and 4.0-fold higher (p < 0.05) than those in the parent strain C. glutamicum ΔAPE (Figure 2A).
Three reactions in the L-ornithine biosynthesis pathway involve NADPH. These are the reactions catalysed by NADP-dependent isocitrate dehydrogenase, NADPdependent glutamate dehydrogenase, and NADP-dependent N-acetyl-gamma-glutamyl-phosphate reductase. To analyse the effect of NADPH availability on L-ornithine accumulation, we deleted the argR gene of the parent strain C. glutamicum ΔAPE to obtain C. glutamicum ΔAPER, and then analysed the NADPH contents of the two strains. As expected, C. glutamicum ΔAPE6937R42 produced more NADPH and L-ornithine than C. glutamicum ΔAPER (Table 2). To better understand the effect of NADPH, we compared the transcript levels of the genes involved in NADPH synthesis in the C. glutamicum strains ΔAPE6937R42 and ΔAPER. The results are presented in Figure 3. In C. glutamicum ΔAPE6937R42, the genes involved in NADPH synthesis (zwf, gnd, and icd) were upregulated by 3.8-, 2.7-and 2.5-fold, respectively (p < 0.05). The ppnk gene was also upregulated by 1.8-fold (p < 0.05). Therefore, we examined whether the increased NADPH levels are caused by the upregulations of the transcriptional levels of these genes in C. glutamicum ΔAPE6937R42. We overexpressed these genes in C. glutamicum ΔAPER. The results are presented in Table 3. Over-expression of the zwf, gnd and ppnK genes indeed increased the concentration of NADPH in C. glutamicum ΔAPER. However, only over-expression of the ppnK gene promoted Lornithine production. It suggests that only the increased NADPH level caused by the elevated transcriptional level of the ppnK gene promoted L-ornithine production.

Fermentation of the strain generated by metabolic evolution
For a more detailed view on L-ornithine production, C. glutamicum ΔAPE6937R42 was cultured in a 5-L bioreactor (Figure 4), and was found to grow in a diauxic manner. The maximum L-ornithine concentration (24.1 ± 1.5 g/L) and yield (0.3 g/g) was obtained at 33 h, at which time all glucose in the medium had been consumed.

Discussion
In this study, we first deleted the speE gene to enhance L-ornithine production ( Table 1). Spermidine synthase encoded by speE catalyzes the formation of spermidine from putrescine. Although the genes involved in the biosynthesis of putrescine remain unknown, cells of C. glutamicum contain putrescine and polyaminies [26]. The deletion of the speE gene blocked putrescine to be converted into spermidine and might alleviate the degradation of L-ornithine. However, the real reason of promoting L-ornithine production by the deletion of the speE gene should be further investigated. Sequence data of the nine genes of C. glutamicum ΔAPE6937 we characterised at the transcript level failed to uncover any mutations in the evolved strain. It is possible that mutations in other genes may have conferred the observed phenotypes. These might include other genes that encode enzymes responsible for L-ornithine biosynthesis or transcription factors that regulate the expression of the sequenced genes. Kutyna et al. also reported that there were no apparent mutations in the Saccharomyces cerevisiae B2-c3 they evolved to generate elevated yields of glycerol [27]. Further work will be required to identify the genetic determinants of these traits.
In C. glutamicum ΔAPE6937 and ΔAPE6937R42, elevated transcriptional levels of the genes involved in the upstream pathway of glutamate biosynthesis (pgi, pfkA, gap, pyk, pyc, gltA and gdh, Figure 2) indicated increased availability of endogenous glutamate compared with the parental strain. Pyruvate kinase is a major bottleneck for glutamate production, and over-expression of the pyc gene improved glutamate production in C. glutamicum ATCC 13032 [28]. Shirai et al. reported that the fluxes of the reactions catalysed by Pgi, PfkA, Gap, Pyk, Pyc, GltA, and Gdh were increased with glutamate production in C. glutamicum [29]. In our previous paper [15], these enzymes were more abundant in C. glutamicum engineered to overproduce L-ornithine than in wild-type C. glutamicum. It is thought that increasing the availability of endogenous glutamate might increase Lornithine production. Lee and Cho reported that the availability of glutamate was a rate-limiting factor in Lornithine accumulation, and that addition of glutamate in the media increased L-ornithine production in engineered E. coli [8]. However, the same group reported that the intracellular concentration and supply of glutamate is not a rate-limiting step for L-ornithine production in an L-ornithine-producer C. glutamicum SJ8074 owing to the presence of rate limiting steps in Lornithine biosynthesis downstream of glutamate synthesis in C. glutamicum SJ8074 [9]. They also reported that the increased availability of glutamate might increase Lornithine accumulation in certain genetic backgrounds (e.g., that of the L-ornithine-producer C. glutamicum SJ8039) [9]. The transcriptional levels of the argB and argJ genes were much higher than those of the genes that encode enzymes that act upstream of glutamate in C. glutamicum ΔAPE6937R42 ( Figure 2B). This indicates that there is no rate-limiting step in L-ornithine synthesis downstream of the synthesis of glutamate in C. glutamicum ΔAPE6937R42. Thus, the elevated transcriptional levels of the genes involved in the upstream pathway of glutamate biosynthesis may be one reason of the increased L-ornithine production in C. glutamicum ΔAPE6937R42.   This study demonstrated that NADPH availability and L-ornithine production are strongly correlated. Production of L-ornithine requires 2 mol of NADPH per mole of L-ornithine produced by C. glutamicum. Increases in the levels of zwf, gnd, and icd transcripts can increase NADPH availability in C. glutamicum ΔAPE6937R42. However, increased levels of zwf and gnd transcripts can drive the carbon metabolic flux from the Embden-Meyerhof-Parnas pathway to the pentose phosphate (PP) pathway. Moreover, given that the PP pathway is coupled with CO 2 production, direct enhancement of the PP pathway may result in release of CO 2 , thereby decreasing the yield of L-ornithine. Thus, enhanced carbon flux through the PP pathway, caused by the upregulation of zwf and gnd at the transcriptional level, inhibits Lornithine production. Our results also demonstrated this point. Over-expression of the zwf and gnd genes could not increase L-ornithine production in C. glutamicum ΔAPER although it could enhance NADPH concentration (Table 3). These suggest that the increased level of L-ornithine production in the evolved strain C. glutamicum ΔAPE6937R42 was not caused by the increased levels of zwf and gnd transcripts.
ATP-dependent NAD kinase encoded by ppnK catalyzes the phosphorylation of NAD to NADP. Increased transcriptional levels of ppnK increased the size of the NADP pool, thus potentially increasing the abundance of NADPH. Our results demonstrated that the increased level of ppnK transcript indeed enhanced the abundance of NADPH in the evolved strain C. glutamicum ΔAPE6937R42 (Figure 3 and Table 2). Lindner et al. reported that over-expression of ppnk improved L-lysine production in C. glutamicum by 12% [30]. Overexpression of nadk, which encodes NAD kinase, increased the NADPH/NADP ratio, which in turn enhanced thymidine biosynthesis in E. coli [31]. Our result also demonstrated that over-expression of the ppnK gene enhanced Lornithine production in C. glutamicum ΔAPER. Thus, the elevated level of ppnK transcript that increase the availability of NADPH may be another reason of increased Lornithine production in C. glutamicum ΔAPE6937R42.
Other strategies have been developed to improve NADPH availability. Over-expression of E. coli pntAB genes, which encode a membrane-bound transhydrogenase enhanced NADPH availability, and thus increased L-lysine levels in C. glutamicum [32]. In this study, overexpression of E. coli pntAB genes enhanced NADPH availability (about 5.2-fold), and thus increased L-ornithine levels (23.7%) in C. glutamicumΔAPER (Table 3). Simultaneous chromosomal overexpression of transhydrogenase (pntAB) and NAD kinase (yfjB) genes had a effect on increasing NADPH supply and improving anaerobic isobutanol production [33]. Replacement of the endogenous NAD-dependent glyceraldehyde 3-phosphate dehydrogenase with the NADP-dependent glyceraldehyde 3-phosphate dehydrogenase from Streptococcus mutans also increased both NADPH availability and L-lysine production in C. glutamicum [34]. The inactivation of the gluconate kinase gene (gntK) led to a 51.8% increase in intracellular NADPH concentration and a 49.9% increase in L-ornithine production [12]. These strategies may be useful for further improving Lornithine production in C. glutamicum ΔAPE6937R42.
To the best of our knowledge, this is the first report of the use of metabolic evolution engineering to increase production of L-ornithine by C. glutamicum. The yield of our engineered stain (24.1 g/L) is unprecedented in any engineered C. glutamicum strain of which we are aware, including that described in our previous paper (18.17 g/L) [16] and that reported by Hwang and Cho (13.16 g/L) [12]. To date, the highest titre of L-ornithine reported in the literature was 74 g/L, as reported by Lee et al. [6]. Those researchers achieved this yield by using a 7-L fed-batch fermentation process with a glucosefeeding strategy and an L-arginine auxotrophic mutant of Brevibacterium ketoglutamicum ATCC 21092. The titre of this auxotrophic mutant, which was generated using classical mutagenesis, was only 2 g/L in batch culture [35]. This suggests that the yield of the new strains described in this study may be enhanced by growth using fed-batch fermentation technology.

Conclusion
We first deleted the speE gene of C. glutamicum ATCC 13032 (ΔargFΔproB), then evolved by a growth-based selection process for 70 passages to generate C. glutamicum ΔAPE6937, and finally deleted the argR gene of the evolved strain to obtain C. glutamicum ΔAPE6937R42. The C. glutamicum ΔAPE6937R42 strain produced 24.1 g/L of L-ornithine in a 5-L bioreactor, a level unprecedented in any engineered strain. It has been demonstrated that the increased L-ornithine production in C. glutamicum ΔAPE6937R42 is dependent on the increased availabilities of glutamate caused by the elevated levels of transcripts involved in the upstream pathway of glutamate biosynthesis and the increased availabilities of NADPH caused by the elevated level of ppnK transcript. The availability of NADPH plays an important role in L-ornithine production in C. glutamicum.

Strains, primers and plasmids
All strains and plasmids used in this study are listed in Table 4. The C. glutamicum strain ATCC 13032 (ΔargFΔproB) [15] was used as the starting strain for L-ornithine production. Primers used in this study are listed in Table 5. L-ornithine was purchased from Sigma.

L-ornithine production in shake flasks
For L-ornithine fermentations, a 1.0-mL sample of the seed culture that had been grown at 150 rpm and 30°C for 12 hours was inoculated into 10 mL of the fermentation medium in a 100-mL flask and incubated at 30°C and 150 rpm for 72 hours. Each litre of the seed medium contained 25 g of glucose, 10 g of yeast extract, 10 g of corn steep liquor, 15 g of (NH 4 ) 2 SO 4 , 2.5 g of MgSO 4 ⋅7H 2 O, 1 g of KH 2 PO 4 , 0.5 g of K 2 HPO 4 , 0.5 g of Na 2 HPO 4 , and 10 g of CaCO 3 . Each litre of the fermentation medium contained 100 g of glucose, 20 g of corn steep liquor, 50 g of (NH 4 ) 2 SO 4 , 2.5 g of MgSO 4 ⋅7H 2 O, 1 g of KH 2 PO 4 , 0.5 g of K 2 HPO 4 , 0.5 g of Na 2 HPO 4 , 20 mg of FeSO 4 ⋅7H 2 O, 20 mg of MnSO 4 ⋅4H 2 O, 2 g of molasses, 1 mL of Tween-80, and 10 g of CaCO 3 . The initial pH of both media described above was adjusted to 7.0.

Adaptive evolution
The adaptive evolution process is presented in Figure 1. Adaptive evolution was conducted in a test tube that contained 5 mL of fermentation medium supplemented with 50 g/L glucose and 15 g/L L-ornithine, and incubated at 30°C and 200 rpm. Cultures were serially passed into fresh medium (initial OD 600 of 0.2) daily. After repeating this transfer procedure 30 times, the culture was then sequentially transferred to fermentation medium containing 70 g/L glucose. The daily transfer procedure at the glucose concentration was repeated 20 times. Finally, the culture was then sequentially transferred to fermentation medium with 100 g/L glucose, and the daily transfer procedure was repeated 20 times. Cultures were frozen and stored at −80°C at every 10 passages throughout adaptive evolution.
After the 70-day adaptive evolution process, cultures stored at −80°C were spread onto LB medium plates (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl). Single colony was transferred to MM medium plates (5 g/L of glucose, 1 g/L of (NH 4 ) 2 SO 4 , 0.5 g/L of sodium citrate, 0.1 g/L of MgSO 4 ⋅7H 2 O, 4.5 g/L K 2 HPO 4 , and 10.5 g of KH 2 PO 4 (pH 7.0)) that was supplemented with 5 g/L L-ornithine and 5 g/L L-arginine. Only colonies that grew on the LB medium plates were further cultured in shake flasks to evaluate their levels of L-ornithine production.

Quantitative real-time PCR (qRT-PCR)
Total RNA from C. glutamicum cells grown for 54 h in shake flasks was isolated using an RNA extraction kit (Dongsheng Biotech, Guangzhou, China), following the manufacturer's instructions. The first-strand cDNA was synthesized using an All-in-One™ First-Strand cDNA Synthesis Kit (GeneCopoeia, Guangzhou, China). The qRT-PCR was performed with the All-in-One™ qPCR   Table 5. The quantification technique used to analyse data was the 2 -ΔΔCt method described by Livak and Schmittgen [36]. The data were normalized using expression of 16S rRNA.

Gene knockout
Chromosomal DNA of C. glutamicum was isolated as described by Eikmanns et al. [37]. The preparation of competent cells and electroporation for C. glutamicum was performed as described by Van de Rest et al. [38]. The correct mutants of C. glutamicum were confirmed by PCR. The lethality of the sacB gene in corynebacteria depends on its expression levels. Therefore, a 1.85-kb DNA fragment containing the sacB gene cluster was amplified by PCR using the primers sacBF and sacBR (Table 5), and the plasmid pK18mobsacB [39] as template. This converted the native promoter of the sacB gene cluster to a tac-M promoter, which is a strong promoter in corynebacteria [40]. The entire backbone of pK18mobsacB, except for the sacB gene, was amplified using the primers pk18msF and pk18msR ( Table 5). The two fragments were digested with SpeI and EcoRV, and ligated together to form the inducible suicide vector pK-JL (5,570 bp).
Disruption of the gene was performed using the nonreplicable integration vector pK-JL, which allows for marker-free deletion of the target gene [39]. The flanking sequence of the argR gene (1,153 bp and 921 bp) was amplified from the genomic DNA of C. glutamicum using the primers argR-F5/argR-R5 and argR-F3/argR-R3, and ligated separately into pMD18-T. The two fragments were excised using BamHI/XbaI and XbaI/SalI, respectively, and then ligated into the BamHI/SalI sites of pK-JL to obtain the non-replicable integration vector pK-ΔargR, which contains an internal 506-bp deletion in the argR gene. The above nonreplicable integration vector pK-ΔargR was transferred into C. glutamicum to disrupt the site-specific gene using the protocol described by Schäfer et al. [39].

Plasmid construction
The argB gene, which encodes acetylglutamate kinase, was amplified from the genomic DNA of C. glutamicum and E. coli, using the primer pairs GargBF/GargBR and EargBF/EargBR, respectively. The fragments were ligated separately into the pMD18-T vector. Both fragments were excised using XbaI/SalI and then inserted into the XbaI/SalI sites of pEC-XK99E [41] to obtain the overexpression vectors pEC-argB CG and pEC-argB EC . The zwf, ppnK and gnd gene was amplified from the genomic DNA of C. glutamicum using the corresponding primer pairs (Table 5), and then inserted into the corresponding sites of pEC-XK99E, respectively. The pntAB genes were amplified from the genomic DNA of E. coli, using the primer pair pntAB-F/pntAB-R (Table 5) and then inserted into the KpnI/PstI sites of pEC-XK99E.

NADPH assay
After aerobic cultivation of C. glutamicum on a rotary shaker (150 rpm) at 30°C for 54 h, the cells were harvested by centrifugation and washed twice with water. Intracellular NADPH was extracted and quantified using the Enzychrom TM NADP + /NADPH Assay kit (BioAssay Systems, Hayward, CA) following the manufacturer's instructions.

Batch culture in bioreactor
Batch culture was carried out at 30°C in a 5-L jar fermentor (Biostat B5, B. Braun, Germany) that contained 3 L of fermentation medium. To prepare the inocula, 10 mL of LB medium was inoculated with a small aliquot of cell glycerol stock that had been stored at −80°C, and was cultured overnight at 30°C. One millilitre of the overnight culture was subsequently transferred into a 250-mL Erlenmeyer flask containing 50 mL of the seed medium, and incubated for 24 h at 30°C and 150 rpm in a shaking incubator. The seed cultures (300 mL) were inoculated into the fermenter for batch cultivation. The pH was maintained at 7.0 by adding NH 4 OH. Antifoam was added manually as needed. The aeration rate was 1.0 L/L/min and the agitation rate was 400 rpm. Samples were periodically taken for the measurements of OD 600 , residual glucose concentration, and L-ornithine concentration. Fermentation experiments were carried out in duplicate.