• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. Mar 2009; 75(6): 1635–1641.
Published online Jan 9, 2009. doi:  10.1128/AEM.02027-08
PMCID: PMC2655454

Reengineering of a Corynebacterium glutamicum l-Arginine and l-Citrulline Producer[down-pointing small open triangle]


Toward the creation of a robust and efficient producer of l-arginine and l-citrulline (arginine/citrulline), we have performed reengineering of a Corynebacterium glutamicum strain by using genetic information of three classical producers. Sequence analysis of their arg operons identified three point mutations (argR123, argG92up, and argG45) in one producer and one point mutation (argB26 or argB31) in each of the other two producers. Reconstitution of the former three mutations or of each argB mutation on a wild-type genome led to no production. Combined introduction of argB26 or argB31 with argR123 into a wild type gave rise to arginine/citrulline production. When argR123 was replaced by an argR-deleted mutation (ΔargR), the production was further increased. The best mutation set, ΔargR and argB26, was used to screen for the highest productivity in the backgrounds of different wild-type strains of C. glutamicum. This yielded a robust producer, RB, but the production was still one-third of that of the best classical producer. Transcriptome analysis revealed that the arg operon of the classical producer was much more highly upregulated than that of strain RB. Introduction of leuC456, a mutation derived from a classical l-lysine producer and provoking global induction of the amino acid biosynthesis genes, including the arg operon, into strain RB led to increased production but incurred retarded fermentation. On the other hand, replacement of the chromosomal argB by heterologous Escherichia coli argB, natively insensitive to arginine, caused a threefold-increased production without retardation, revealing that the limitation in strain RB was the activity of the argB product. To overcome this, in addition to argB26, the argB31 mutation was introduced into strain RB, which caused higher deregulation of the enzyme and resulted in dramatically increased production, like the strain with E. coli argB. This reconstructed strain displayed an enhanced performance, thus allowing significantly higher productivity of arginine/citrulline even at the suboptimal 38°C.

We have shown reverse engineering of a high-production strain of Corynebacterium glutamicum by using l-lysine fermentation as a model (10, 11, 21). The characteristic that the methodology aims at is the robustness of the resulting strain. The classical approach, based on random mutation and selection, sacrifices the native robustness of an organism in exchange for enhancing the production abilities to the limits. The high production abilities and delicate constitutions of classical industrial producers are the merits and demerits of the classical approach. Such an inevitable consequence of the classical approach could be understood also from the fact that more than 1,000 mutations have accumulated in the genome of an industrial l-lysine producer of C. glutamicum (11). We examined those mutations and identified mutations relevant to l-lysine production. Subsequent assembly of the useful mutations in a robust wild-type strain was shown to substantially improve producer performance (18, 21). In addition, the mechanisms of l-lysine hyperproduction were unraveled through this strain reconstruction (7, 15, 20, 21). As demonstrated in these studies, our work involves reengineering a more efficient producer using the knowledge regarding the mutations that have accumulated over years of industrial strain development. The methodology starts by tracing backwards to an existing classical producer and thus can be called reverse engineering.

With the accumulated knowledge on mutations relevant to production, it is possible to combine positive mutations derived from different lines of classical producers in a single wild-type background. Such an advanced approach has recently led to an impressive result in l-arginine and l-citrulline production by C. glutamicum. The procedure and impact of this reengineering methodology are described here.

l-Arginine, a semiessential amino acid, has lately attracted considerable attention because the amino acid has been shown to be a precursor to nitric oxide (NO), a key component of endothelium-derived relaxing factor (1). Because of l-arginine's NO-stimulating effect, the amino acid helps, for example, to relax and dilate blood vessels, and thus can be utilized in numerous clinical areas (1). On the other hand, l-citrulline, a precursor of l-arginine biosynthesis (Fig. (Fig.1),1), is also an important amino acid for our health since it is a source of endogenous l-arginine in the body (5). Since the issue of which amino acid is preferably accumulated is not the subject of this paper, we report here the results for the sum of the two amino acids as arginine/citrulline.

FIG. 1.
l-Arginine biosynthesis pathways and the relevant genes in C. glutamicum and E. coli. The broken arrows indicate the reactions specific to E. coli, and the corresponding E. coli genes are shown in parentheses. The separate reactions specified by the ...


Bacterial strains and plasmids.

C. glutamicum arginine producers used for comparative sequence analysis were strains A-27, I-30, and D-77, which were derived through multiple rounds of mutagenesis from wild-type ATCC 13870, ATCC 13032, and KY10025, respectively (2, 17). Strain ATCC 13870 was previously classified as Corynebacterium acetoacidophilum, but based on recent molecular taxonomic studies it is currently reclassified in the original species, C. glutamicum (13). The wild-type strains used for comparisons of potentials for arginine/citrulline production were C. glutamicum ATCC 13870, ATCC 13032, and other representative strains listed in our previous report (19). Escherichia coli DH5α (26) was used as a donor of the genomic DNA for amplifying the E. coli argB gene and also as a host for cloning of the PCR products. Plasmid pESB30, which is nonreplicative in C. glutamicum, is a vector for gene replacement in C. glutamicum (16). Plasmids pCargR123, pCargRG45, pCargB26, pCargB31, pCleuC456, and pCargB2631, which contain the mutated DNAs in vector pESB30, were used to replace the wild-type chromosomal DNAs by the mutated DNAs. Plasmid pCargRd, which contains the internally deleted argR gene in vector pESB30, was used to replace the wild-type chromosomal gene by the deleted gene. Plasmid pC-EargB, which contains the open reading frame (ORF) of E. coli argB in the vector pESB30, was used to replace the chromosomal argB ORF with the heterologous E. coli argB ORF.


Complete medium BY (27) and minimal medium (27) were used for cultivation of C. glutamicum. Solid plates were made by the addition of Bacto agar (Difco) to 1.6%. When required, kanamycin was added at the final concentration of 20 μg/ml for BY plates. RG2 medium used for production in a 300-ml flask consisted of (per liter) 60 g of glucose, 5 g of corn steep liquor, 30 g of (NH4)2SO4, 8 g of KCl, 2 g of urea, 0.5 g of KH2PO4, 0.5 g of K2HPO4, 1 g of MgSO4·7H2O, 1 g of NaCl, 20 mg of FeSO4·7H2O, 10 mg of MnSO4·5H2O, 20 mg of nicotinic acid, 20 mg of β-alanine, 10 mg of thiamine-HCl, 0.2 mg of d-biotin, and 30 g of CaCO3 (pH 7.7). RSG1 medium used for second-seed cultures in jar fermentations consisted of (per liter) 60 g of glucose, 22 g of corn steep liquor, 5 g of (NH4)2SO4, 5 g of urea, 2 g of KH2PO4, 1 g of MgSO4·7H2O, 10 mg of FeSO4·7H2O, 10 mg of MnSO4·5H2O, 30 mg of CaCl2·2H2O, 30 mg of CuSO4·5H2O, 1 mg of ZnSO4·7H2O, 1 mg of NiCl2·6H2O, 1 mg of CoCl2·6H2O, 1 mg of (NH4)6Mo7O24·4H2O, 10 mg of β-alanine, 10 mg of nicotinic acid, 10 mg of thiamine-HCl, and 0.3 mg of d-biotin (pH 7.2). RPG1 medium used for 5-liter jar fermentors consisted of (per liter) 60 g of glucose, 3 g of corn steep liquor, 30 g of (NH4)2SO4, 2.8 g of KH2PO4, 1 g of MgSO4·7H2O, 1 g of NaCl, 20 mg of FeSO4·7H2O, 10 mg of MnSO4·5H2O, 40 mg of CaCl2·2H2O, 2 mg of CuSO4·5H2O, 2 mg of ZnSO4·7H2O, 2 mg of NiCl2·6H2O, 2 mg of CoCl2·6H2O, 20 mg of β-alanine, 20 mg of nicotinic acid, 10 mg of thiamine-HCl, and 0.2 mg of d-biotin (pH 6.8). For growth of E. coli, Luria-Bertani broth or agar (26) was used.

Cultivations for arginine/citrulline production.

For flask fermentations, a 2.0-ml sample of the seed culture grown to early stationary phase at 30°C in BYG medium (containing 1.0% glucose in medium BY) was inoculated into 20 ml of RG2 medium in a 300-ml flask and cultivated aerobically at 30°C or 38°C until the sugar was consumed.

For jar fermentations, cells grown on a BY plate at 30°C for 1 day were inoculated into 200 ml of RSG1 medium in a 2-liter flask. After growth to early stationary phase at 30°C on a rotary shaker, the seed broth was transferred into a 5-liter jar fermentor containing 1,000 ml of RPG1 medium. After the sugar initially added was consumed, a solution containing (per liter) 500 g of glucose, 30 g of (NH4)2SO4, 1 g of NH4H2PO4, 1.2 g of MgSO4·7H2O, 8 g of KCl, 150 mg of CaCl2·2H2O, 20 mg of β-alanine, 20 mg of nicotinic acid, 20 mg of thiamine-HCl, and 0.2 mg of d-biotin was continuously fed until the total amount of glucose in the medium reached 572 g. The feeding rate of the solution was controlled to maintain the glucose concentration in the medium at a low concentration (below 0.5%). The culture was basically performed with an agitation speed of 800 rpm, with aeration at 2 liter/min, and at 30°C or 38°C. The pH was maintained at 6.8 with NH4OH.

Recombinant DNA techniques.

A standard protocol (26) was used for the construction, purification, and analysis of plasmid DNA and transformation of E. coli. Chromosomal DNA of C. glutamicum was extracted from protoplasts by the method of Saito and Miura (24). The protoplasts were prepared by the method of Katsumata et al. (12). Transformation of C. glutamicum by electroporation was carried out by the method of van der Rest et al. (29), using a Gene Pulser and a Pulse Controller (Bio-Rad). PCR was performed with a DNA thermal cycler GeneAmp 9600 (Perkin-Elmer), using TaKaRa La Taq DNA polymerase (Takara Bio, Otsu, Japan).

Nucleotide sequence analysis of the arg operon.

The entire arg operon argCJBDFRGH (Cgl1394 to -1401) was PCR amplified using primers arg op-F and arg op-R with the genomic DNA of each of three arginine producers, A-27, I-30, and D-77, and individual natural ancestors as a template. Primers used in this study are listed in Table Table1.1. Amplified 9.0-kb PCR fragments were purified using a GENECLEAN III kit (Qbiogene, CA). With each PCR fragment as a template, the inner segments of each fragment were amplified by PCR at intervals of approximately 500 bp using primers designed based on the genome sequence of C. glutamicum (BA000036), which is publicly available at http://gib.genes.nig.ac.jp/single/index.php?spid=Cglu_ATCC13032 (10). The nucleotide sequences of the PCR products were then analyzed with an ABI Prism 377 DNA sequencer from Applied Biosystems, with the ABI Prism Big Dye Terminator cycle sequencing kit (Perkin-Elmer). The subsequent electrophoresis analysis was carried out by using Pageset SQC-5ALN 377 (Toyobo, Japan).

Oligonucleotide primers

Introduction of specific mutations into the genome.

The mutated argB gene regions were PCR amplified using primers argB-F and argB-R with genomic DNAs of arginine producers I-30 and D-77 as templates. On the other hand, the mutated argR gene region was PCR amplified using primers argR-F and argR-R with genomic DNA of arginine producer A-27 as a template. Similarly, the mutated argRG gene region was PCR amplified using primers argR-F and argG-R with genomic DNA of arginine producer A-27 as a template. Each PCR fragment was cloned into pESB30 by the TA cloning method (26). The recombinant plasmids constructed by cloning the PCR fragments containing the mutated argB gene regions of strains I-30 and D-77 into pESB30 were designated pCargB26 and pCargB31, respectively. The recombinant plasmids constructed by cloning the PCR fragments containing the mutated argR and argRG gene regions of strain A-27 into pESB30 were designated pCargR123 and pCargRG45, respectively. The specific mutations on the recombinant plasmids were introduced into C. glutamicum strains via two recombination events as described previously (21). For introduction of the leuC456 mutation into strain RB, plasmid pCleuC456 (7) was used.

Generation of a strain carrying both argB mutations.

The argB26 and arg31 mutations were introduced into the genomic argB gene using plasmid pCargB2631, which was constructed as follows. The 5′ region and 3′ region of the argB gene were amplified by PCR with pCargB31 as a template using the primer pairs argB-F with argB26-R and argB26-F with argB-R, respectively. As the two primers argB26-R and argB26-F contained regions complementary to each other, fusion PCR was performed using the purified 5′-region and 3′-region fragments as templates and the primers argB-F and argB-R. The resulting 1.0-kb fragment contained the intended argB gene region, on which both argB mutations coexisted. The fragment was cloned into pESB30 by the TA cloning method to yield pCargB2631. This plasmid was used to replace the corresponding chromosomal gene with the double-mutated gene.

Chromosomal deletion of argR.

Plasmid pCargRd containing the internally deleted argR gene was constructed as follows and was used to replace the wild-type chromosomal gene with the deleted gene. The 5′-upstream region of the argR gene was amplified by PCR using primers argR-up-F and argR-up-R with ATCC 13032 genomic DNA as a template. Similarly, the 3′-downstream region of the gene was amplified using primers argR-down-F and argR-down-R. Fusion PCR was then performed using the purified 5′-upstream and 3′-downstream fragments as templates and the primers argR-up-F and argR-down-R. The resulting 1.0-kb fragment contained the deleted argR gene, which was shortened from 513 bp to 132 bp by in-frame deletion of the inner sequence. The fragment was digested by BamHI and then ligated to BamHI-digested pESB30 to yield pCargRd. Defined chromosomal deletion of the argR gene was accomplished using pCargRd via two recombination events as described previously (21).

Replacement of chromosomal argB with heterologous E. coli argB.

Plasmid pC-EargB containing the ORF of E. coli argB sandwiched between the 3′ region of C. glutamicum argJ and the 5′ region of C. glutamicum argD was constructed as follows and was used to replace the endogenous C. glutamicum argB ORF with the E. coli argB ORF. The 3′ region of C. glutamicum argJ was amplified by PCR using two primers, argJ-F and argJ-R, with ATCC 13032 genomic DNA as a template. On the other hand, the ORF of E. coli argB was also amplified with two primers, EargB-F and EargB-R, with E. coli DH5α genomic DNA as a template. In the design of the primer EargB-F, four low-usage codons in the N-terminal coding region of E. coli argB were replaced with synonymous high-usage codons in C. glutamicum as shown in Table Table11 (AAT→AAC, TTA→CTC, ATT→ATC, and AAA→AAG; underlining indicates changed nucleotides). Fusion PCR was performed using the purified C. glutamicum argJ and E. coli argB gene fragments as templates and the primers argJ-F and EargB-R. The resulting 1.5-kb fragment contained the E. coli argB ORF, which was preceded by the 3′ region of C. glutamicum argJ. The fragment was digested with BglII and BamHI and then ligated to BamHI-digested pESB30 to yield pESBargB. Similarly, the 5′ region of C. glutamicum argD was amplified by PCR using two primers, argD-F and argD-R, with ATCC 13032 genomic DNA as a template. The resulting 0.7-kb fragment was digested with BamHI and BfrI and then ligated to pESBargB digested by the same restriction enzymes to yield pC-EargB. Replacement of chromosomal argB with heterologous E. coli argB was conducted using pC-EargB via two recombination events as described previously (21).

Transcriptome analysis.

Total RNAs were prepared from mid-exponential-phase cultures of 5-liter jar fermentors as described previously (8). Transcriptome analysis was performed using GeneChip (Affymetrix). Labeling of RNA transcripts, hybridization, and scanning were performed according to the manufacturer's instructions. Gene expression data were analyzed using Microarray Suite 5.0 software (Affymetrix). Changes in expression levels that had a change call of decrease or increase together with a P value of <0.001 and a signal ratio of more than 1.5-fold were considered significant. The reproducibility of the measurements was confirmed by duplicate, independent cultures and experiments.

Enzyme assay.

Crude cell extracts were prepared by sonic disruption of cells from mid-exponential-phase cultures of 5-liter jar fermentors as described previously (22). Protein quantity was determined by the method of Bradford (3). The activities of N-acetyl-l-glutamate kinase (ArgB) in crude cell extracts were measured colorimetrically at 30 °C, essentially as described by Fernández-Murga et al. (6).


Cell growth was monitored by measuring the optical density at 660 nm (OD660) of the culture broth with a U-1080 Auto Sipper photometer (Hitachi, Japan). The glucose concentration was determined using a Determinar GL-E apparatus (Kyowa Medex Co., Ltd., Japan). l-Arginine and l-citrulline titers were determined by using a high-performance liquid chromatography system (Shimazu, Japan) after derivatization with o-phthalaldehyde.


Identification of the mutation(s) causing arginine/citrulline production.

Our first task was to identify the basal mutation(s), namely, the mutation(s) that conferred the ability to produce arginine/citrulline in wild-type C. glutamicum. For this purpose, three arginine producers, strain A-27, strain I-30, and strain D-77, were used as gene resources. These are classical mutants derived independently from different lines of C. glutamicum wild-type strains. From their genomic information, we attempted to identify the basal mutation(s). Generally speaking, such a basal mutation(s) is assumed to exist on the argB gene encoding the key regulatory enzyme for arginine biosynthesis (Fig. (Fig.1),1), but to our surprise, there was no such mutation in strain A-27, the best arginine producer of the three. Instead, a point mutation was found in each of the argR gene, the upstream noncoding region of the argG gene, and the argG gene (Fig. (Fig.2A):2A): a C-to-T exchange at position 368 in argR, leading to an amino acid replacement of Ala-123 by Val (designated mutation argR123); a G-to-A exchange at 92 bp upstream of argG (designated mutation argG92up); a G-to-A exchange at position 136 in argG, leading to an amino acid replacement of Asp-45 by Asn (designated mutation argG45). On the other hand, the argB genes of strains I-30 and D-77 were found to have a different point mutation, as expected (Fig. (Fig.2A):2A): a C-to-T exchange at position 77 in argB, leading to an amino acid replacement of Ala-26 by Val (designated mutation argB26); an A-to-G exchange at position 91 in argB, leading to an amino acid replacement of Met-31 by Val (designated mutation argB31).

FIG. 2.
Reconstitutions of defined mutations on a wild-type genome and their effects on arginine/citrulline production. (A) The five specific mutations identified on the arg operons of three classical producers, A-27, I-30, and D-77, are indicated under the corresponding ...

By using these five specific mutations, we examined their relevance for arginine/citrulline production in a wild-type ATCC 13032 background. First of all, three strain A-27-derived mutations (argR123, argG92up, and argG45) were reconstituted simultaneously on the wild-type genome, but unexpectedly arginine/citrulline production was not observed (Fig. (Fig.2B).2B). Separate introduction of the argB31 or argB26 mutation into the wild type also resulted in no production (Fig. (Fig.2B).2B). However, when the argB31 or argB26 mutation was combined with the argR123 mutation on the wild-type genome, we observed arginine/citrulline production for the first time (Fig. (Fig.2B).2B). Based on this finding, the missense mutation, argR123, was assumed to impair the repressor function of the ArgR protein. We also examined the effect of in-frame deletion of the argR inner sequence (designated mutation ΔargR) on production. As a result, increased production was observed in combination with either argB31 or argB26 (Fig. (Fig.2B).2B). The level was higher in the combination of argB26 and ΔargR than in the combination of argB31 and ΔargR (Fig. (Fig.2B).2B). Taken all together, we specified the basal mutations as argB26 and ΔargR.

The consequences of these mutations on production were more prominent when the culture temperature was shifted from the traditional 30°C to a suboptimal 38°C (Fig. (Fig.2B).2B). This indicated that arginine/citrulline fermentation by this organism could be potentially realized even at higher temperatures than are traditionally practiced.

Screening for a wild-type background with best performance.

In the reengineering approach, it is important to start from different wild-type strains to obtain the best performance, since a strain engineered is supposed to inherit the properties of its original ancestor. For this purpose, the basal mutation set, argB26 and ΔargR, was introduced into six C. glutamicum wild-type strains to generate isogenic mutants, which were then screened for the abilities to produce arginine/citrulline at the flask level (data not shown). Among those, we chose two typical producers, the derivatives of wild-type strains ATCC 13032 and ATCC 13870, and compared their performance in more detail using 5-liter jar fermentors. The evaluation was conducted under suboptimal 38°C conditions, because fermentation at elevated temperatures (35 to 40°C) is industrially advantageous, as it leads to a reduction of cooling costs compared with the traditional 30°C conditions. Figure Figure33 shows the profiles of arginine/citrulline production, which revealed that strain ATCC 13032 had a significantly higher potential for arginine/citrulline production at elevated temperatures than the other strain. The selected wild-type ATCC 13032 carrying the basal mutation set, argB26 and ΔargR, on its genome was designated strain RB and used for subsequent analyses.

FIG. 3.
Comparison of the abilities to produce arginine/citrulline between two isogenic mutants, strain ATCC 13032 carrying the argB26 and ΔargR mutations (•) and strain ATCC 13870 carrying the same two mutations ([filled square]). The molar ratios ...

Transcriptome analysis.

Strain RB, which carries the argB26 and ΔargR mutations, can grow and consume glucose almost as fast as the wild-type strain, and so the fermentation period can be shortened to below half of that traditionally required. In addition, fermentation at high temperatures around 38°C is possible. Nevertheless, the final titer of arginine/citrulline production was about one-third of that of the classical producer, A-27 (Fig. (Fig.2B).2B). What is the limitation in strain RB, compared with the classical producer? To answer this, we attempted here a transcriptome analysis of each producer. When total RNAs from mid-exponential-phase cultures of 5-liter jar fermentors were used to study differential transcription profiles between each producer and its parental wild type, an interesting finding that could explain the difference in production levels emerged: the difference in the expression levels of the arg operon. In strain RB, carrying the ΔargR mutation, the expression of the arg genes was indeed derepressed by around 10-fold, but to our surprise, an additional upregulation was observed in the classical producer A-27 (Fig. (Fig.4).4). Although strain A-27 carries the argR123 mutation, such high upregulations could not be explained by only the mutation. So we hypothesized that a sort of global response, probably the stringent-like response, which we have observed in the classical lysine producer B-6 (9), might occur also in strain A-27, leading to further induction of the arg operon. In fact, the expression profiles observed for the central metabolic pathways and amino acid biosynthetic pathways in strain A-27 had features similar to those of the lysine producer (9).

FIG. 4.
Ratio of mRNA levels of the arg genes in strain RB (white bars) and classical producer A-27 (gray bars) to those in their parental wild-type strains, ATCC 13032 and ATCC 13870, respectively. Total RNAs from mid-exponential-phase cultures of 5-liter jar ...

Induction of a global response.

If the hypothesis mentioned above were true, introduction of a specific mutation provoking the stringent-like response, the global induction of the amino acid biosynthesis genes, into strain RB would be expected to result in increased production of arginine/citrulline. Since we have already defined such a mutation, namely, leuC456, in the classical lysine producer B-6 (7), the mutation was introduced into strain RB. As expected, the presence of the mutation enabled strain RB to accumulate a level of arginine/citrulline comparable to that of the classical producer A-27 in flask cultures (data not shown). However, with 5-liter jar fermentor cultivation at 38°C, the engineered strain showed retarded fermentation due to decreased growth and sugar consumption (data not shown), thus leaving a problem to be solved for realization of high-speed fermentation.

Verification of the target to be engineered.

Plasmid-mediated amplification of the entire arg operon could be another strategy for increased expression of the operon. However, this engineering also led to retarded fermentation (data not shown). On the other hand, assays using crude extract of strain RB showed that the mutated ArgB enzyme was inhibited by relatively low concentrations of arginine, although it was less sensitive to arginine than the wild-type enzyme (Fig. (Fig.5).5). This finding reminded us of that if the metabolic flux toward arginine/citrulline were restricted at the regulatory step, the expressional shortage of the arg operon in strain RB might be compensated by a further qualitative change of the ArgB enzyme, namely, higher desensitization of the enzyme. To verify this, we engineered the genome of strain RB to generate a strain carrying the hybrid arg operon where the ORF of the endogenous argB gene was replaced with that of E. coli argB. This engineering is based on the information that E. coli has different control mechanisms for arginine biosynthesis than C. glutamicum (4, 14, 23, 25, 28): the feedback control in E. coli occurs at the ArgA enzyme, and the ArgB enzyme is natively insensitive to end product inhibition (Fig. (Fig.1).1). The result was beyond our expectations. In 5-liter jar fermentor cultivation at 38°C, arginine/citrulline production increased by about threefold during almost the same culture period as that for strain RB (data not shown). From this result, it was verified that the target to be engineered in strain RB was argB.

FIG. 5.
Arginine sensitivities of various ArgB enzymes. Symbols: [filled square], ArgB activity of wild-type strain ATCC 13032; •, ArgB activity of strain ATCC 13032 carrying the argB26 and ΔargR mutations; [filled triangle], ArgB activity of strain ATCC 13032 ...

Higher desensitization of the ArgB enzyme.

How could we achieve higher desensitization of the ArgB enzyme of strain RB? This became our next objective. To this end, we combined the argB31 mutation with the argB26 mutation on the ΔargR background, because the two changed amino acid residues were located very close to each other on the amino acid sequences of the ArgB enzyme and thus were assumed to occur on the same allosteric site of the enzyme. Enzyme assays revealed that the resulting ArgB enzyme was expectedly less sensitive to end product inhibition than the original enzyme carrying either of the two argB mutations (Fig. (Fig.5).5). The engineered strain which carried the argB26, argB31, and ΔargR mutations was designated strain RBid (Fig. (Fig.66).

FIG. 6.
Schematic diagram of the creation of new strain RBid. Useful mutations identified in classical producers are indicated together with unnecessary mutations (×).

Arginine/citrulline production by strain RBid.

We examined whether strain RBid displayed considerably increased production, similar to the strain carrying the heterologous E. coli argB on its genome. The result obtained from flask cultures met our expectations, and arginine/citrulline production by strain RBid reached a level comparable to that of the classical producer, A-27 (Fig. (Fig.2B).2B). With 5-liter jar fermentor cultivation, the engineered strain maintained its high production rate even at 38°C, thus leading to completion of fermentation in half the time required with the traditional strain (Fig. (Fig.77).

FIG. 7.
Fermentation kinetics of the newly developed strain RBid at 38°C in 5-liter jar fermentor cultivation. For comparison, the profiles of classical producer A-27, which was cultured under its optimal 30°C conditions, are shown as controls. ...


The producer's performance that we demanded in the face of this reengineering program is robustness which allows high fermentation yields in the shortest possible time, even under stressful conditions. The newly engineered strain, RBid, satisfies the requirements, and this is clearly reflected in the differences in the fermentation profiles between the new strain and the classical strain (Fig. (Fig.7).7). The engineered strain actually inherited wild-type levels of tolerance to various stresses, including elevated temperatures, high osmolarity, and low oxygen tension, whereas the classical producer showed considerably less sensitivities to such stresses (data not shown). The cellular robustness which is intrinsic to the wild type is occasionally regarded as a negative characteristic for metabolic engineering because of its counteractive effects on metabolic changes. However, in large-scale industrial fermentation, the ambient conditions vary considerably depending on the location of cells within the fermentor, and thus, the native robustness of an organism would be one of several important characteristics which should be retained in a producer strain.

As shown in Fig. Fig.6,6, strain RBid was generated ultimately by assembling the argB26, ΔargR, and argB31 mutations derived from three different lines of classical producers, I-30, A-27, and D-77, respectively, on the wild-type genome. This showed that combining relevant mutations from independently developed classical producers could have a prominent effect on amino acid production. Furthermore, the combination of such mutations and a specific host is also an important consideration in this reengineering approach. The host strain where the three mutations were incorporated is ATCC 13032, which was selected from among various C. glutamicum wild-type strains as a strain with high potential for industrial arginine/citrulline production. If another wild-type strain had been used as a host, such a clear result as presented here may not have been obtained.

The three specific mutations introduced into the wild type are all relevant to the terminal pathway for arginine biosynthesis, and no mutation relevant to other metabolic pathways and regulations is included in the final strain. Nevertheless, the reengineered strain achieved a high level of production above that of the classical producer, which has a long breeding history of more than 10 years. This result seems to suggest that carbon will be directed toward a desired end product from central metabolism if appropriate deregulation of a relevant terminal pathway is attained. However, it is also true that a comparable level of production as with the classical producer was attained by the introduction of the leuC456 mutation into strain RB, although fermentation was retarded in that case. Furthermore, it is certain that the classical producer A-27 has no mutation in argB. These findings suggest alternative mechanisms of production: a specific mutation such as the leuC456 mutation might give rise to the stringent-like response, leading to the global induction of the amino acid biosynthetic genes. This induction, coupled with the derepression of the arg operon by the argR defect, likely contributes to production through tremendous overexpression of the operon, which we observed in the transcriptome of the classical producer.

C. glutamicum has a long history of classical breeding, which has resulted in a large variety of industrially useful mutants. Their beneficial mutations have mostly lain idle within individual mutant strains, but their exploitation is now progressing in our laboratories. In the near future, most of them will be regenerated for useful knowledge that is widely available for the amino acid industry. Thereby, the conventional style of selecting improved strains by phenotypes will undoubtedly be replaced by the new style of reengineering strains by assembling only desired genotypes. This work as well as our previous studies on lysine fermentation (11, 21) will be a paradigm for future strain development in fermentation industries.


We thank A. Ozaki for encouraging support of our work and also S. Hashimoto, S. Koizumi, T. Nakano, Y. Yonetani, M. Yagasaki, S. Nakagawa, A. Yasuhara, T. Abe, J. Ohnishi, H. Mizoguchi, and S. Takeno for useful discussions.


[down-pointing small open triangle]Published ahead of print on 9 January 2009.


1. Appleton, J. 2002. Arginine: clinical potential of a semi-essential amino acid. Altern. Med. Rev. 7:512-522. [PubMed]
2. Azuma, T., T. Nakanishi, and M. Sugimoto. 1988. Isolation and characterization of a stable L-arginine producer from continuous culture broth of Corynebacterium acetoacidophilum. J. Ferment. Technol. 3:27-284.
3. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:248-254. [PubMed]
4. Caldara, M., G. Dupont, F. Leroy, A. Goldbeter, L. De Vuyst, and R. Cunin. 2008. Arginine biosynthesis in Escherichia coli. Experimental perturbation and mathematical modeling. J. Biol. Chem. 283:6347-6358. [PubMed]
5. Curis, E., P. Crenn, and L. Cynober. 2007. Citrulline and the gut. Curr. Opin. Clin. Nutr. Metab. Care 10:620-626. [PubMed]
6. Fernández-Murga, M. L., F. Gil-Ortiz, J. L. Llácer, and V. Rubio. 2004. Arginine biosynthesis in Thermotoga maritima: characterization of the arginine-sensitive N-acetyl-l-glutamate kinase. J. Bacteriol. 186:6142-6149. [PMC free article] [PubMed]
7. Hayashi, M., H. Mizoguchi, J. Ohnishi, S. Mitsuhashi, Y. Yonetani, S. Hashimoto, and M. Ikeda. 2006. A leuC mutation leading to increased l-lysine production and rel-independent global expression changes in Corynebacterium glutamicum. Appl. Microbiol. Biotechnol. 72:783-789. [PubMed]
8. Hayashi, M., H. Mizoguchi, N. Shiraishi, M. Obayashi, S. Nakagawa, J. Imai, S. Watanabe, T. Ota, and M. Ikeda. 2002. Transcriptome analysis of acetate metabolism in Corynebacterium glutamicum using a newly developed metabolic array. Biosci. Biotechnol. Biochem. 66:1337-1344. [PubMed]
9. Hayashi, M., J. Ohnishi, S. Mitsuhashi, Y. Yonetani, S. Hashimoto, and M. Ikeda. 2006. Transcriptome analysis reveals global expression changes in an industrial l-lysine producer of Corynebacterium glutamicum. Biosci. Biotechnol. Biochem. 70:546-550. [PubMed]
10. Ikeda, M., and S. Nakagawa. 2003. The Corynebacterium glutamicum genome: features and impacts on biotechnological process. Appl. Microbiol. Biotechnol. 62:99-109. [PubMed]
11. Ikeda, M., J. Ohnishi, M. Hayashi, and S. Mitsuhashi. 2006. A genome-based approach to create a minimally mutated Corynebacterium glutamicum strain for efficient l-lysine production. J. Ind. Microbiol. Biotechnol. 33:610-615. [PubMed]
12. Katsumata, R., A. Ozaki, T. Oka, and A. Furuya. 1984. Protoplast transformation of glutamate-producing bacteria with plasmid DNA. J. Bacteriol. 159:306-311. [PMC free article] [PubMed]
13. Kinoshita, S. 1999. Glutamic acid producing microorganisms, p. 1330-1336. In M. C. Flickinger, and S. W. Drew (ed.), Encyclopedia of bioprocess technology: fermentation, biocatalysis, and bioseparation. John Wiley & Sons, Inc., New York, NY.
14. Lu, C. D. 2006. Pathways and regulation of bacterial arginine metabolism and perspectives for obtaining arginine overproducing strains. Appl. Microbiol. Biotechnol. 70:261-272. [PubMed]
15. Mitsuhashi, S., M. Hayashi, J. Ohnishi, and M. Ikeda. 2006. Disruption of malate:quinone oxidoreductase increases l-lysine production by Corynebacterium glutamicum. Biosci. Biotechnol. Biochem. 70:2803-2806. [PubMed]
16. Mitsuhashi, S., J. Ohnishi, M. Hayashi, and M. Ikeda. 2004. A gene homologous to β-type carbonic anhydrase is essential for the growth of Corynebacterium glutamicum under atmospheric conditions. Appl. Microbiol. Biotechnol. 63:592-601. [PubMed]
17. Nakayama, K., and H. Yoshida. 1972. Fermentation of l-arginine. Agric. Biol. Chem. 36:1675-1684.
18. Ohnishi, J., M. Hayashi, S. Mitsuhashi, and M. Ikeda. 2003. Efficient 40°C fermentation of l-lysine by a new Corynebacterium glutamicum mutant developed by genome breeding. Appl. Microbiol. Biotechnol. 62:69-75. [PubMed]
19. Ohnishi, J., and M. Ikeda. 2006. Comparisons of potentials for l-lysine production among different Corynebacterium glutamicum strains. Biosci. Biotechnol. Biochem. 70:1017-1020. [PubMed]
20. Ohnishi, J., R. Katahira, S. Mitsuhashi, S. Kakita, and M. Ikeda. 2005. A novel gnd mutation leading to increased l-lysine production in Corynebacterium glutamicum. FEMS Microbiol. Lett. 242:265-274. [PubMed]
21. Ohnishi, J., S. Mitsuhashi, M. Hayashi, S. Ando, H. Yokoi, K. Ochiai, and M. Ikeda. 2002. A novel methodology employing Corynebacterium glutamicum genome information to generate a new l-lysine-producing mutant. Appl. Microbiol. Biotechnol. 58:217-223. [PubMed]
22. Ozaki, A., R. Katsumata, T. Oka, and A. Furuya. 1985. Cloning of the genes concerned in phenylalanine biosynthesis in Corynebacterium glutamicum and its application to breeding of a phenylalanine producing strain. Agric. Biol. Chem. 49:2925-2930.
23. Rajagopal, B. S., J. DePonte III, M. Tuchman, and M. H. Malamy. 1998. Use of inducible feedback-resistant N-acetylglutamate synthetase (argA) genes for enhanced arginine biosynthesis by genetically engineered Escherichia coli K-12 strains. Appl. Environ. Microbiol. 64:1805-1811. [PMC free article] [PubMed]
24. Saito, H., and K. Miura. 1963. Preparation of transforming deoxyribonucleic acid by phenol treatment. Biochim. Biophys. Acta 72:619-629. [PubMed]
25. Sakanyan, V., P. Petrosyan, M. Lecocq, A. Boyen, C. Legrain, M. Demarez, J. N. Hallet, and N. Glansdorff. 1996. Genes and enzymes of the acetyl cycle of arginine biosynthesis in Corynebacterium glutamicum: enzyme evolution in the early steps of the arginine pathway. Microbiology 142:99-108. [PubMed]
26. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratory manual, 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
27. Takeno, S., J. Ohnishi, T. Komatsu, T. Masaki, K. Sen, and M. Ikeda. 2007. Anaerobic growth and potential for amino acid production by nitrate respiration in Corynebacterium glutamicum. Appl. Microbiol. Biotechnol. 75:1173-1182. [PubMed]
28. Utagawa, T. 2004. Arginine metabolism: enzymology, nutrition, and clinical significance. J. Nutr. 134:2854S-2857S. [PubMed]
29. van der Rest, M. E., C. Lange, and D. Molenaar. 1999. A heat shock following electroporation of Corynebacterium glutamicum with xenogeneic plasmid DNA. Appl. Microbiol. Biotechnol. 52:541-545. [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...