Mutations of the Corynebacterium glutamicum NCgl1221 Gene, Encoding a Mechanosensitive Channel Homolog, Induce l-Glutamic Acid Production

Bacterial strains, growth conditions, and plasmids. Corynebacterium glutamicum ATCC 13869 (formerly known as Brevibacterium lactofermentum) was used as the wild-type strain. The l-glutamate-producing mutants 2A-1, 2A-1R, BL1 [NCgl1221(V419::IS1207)], BL2 [NCgl1221(A111V)], and BL3 [NCgl1221(W15CSLW)] were constructed in this study. Escherichia coli JM109 (36) was used for constructing and propagating plasmids. E. coli and C. glutamicum cells were cultured at 37°C on LB medium (14) or at 31.5°C on CM2B (16), respectively. If necessary, the C. glutamicum odhA mutants were cultured at 31.5°C on CM-Dex (per liter, 5 g glucose, 10 g yeast extract, 10 g tryptone, 1 g KH₂PO₄, 0.4 g MgSO₄·7H₂O, 0.01 g FeSO₄·7H₂O, 0.01 g MnSO₄·5H₂O, 3 g urea, 34.5 ml soybean protein hydrolysate [total nitrogen, 1.2 g], adjusted to pH 7.0 with NaOH). Plasmid pBS4S (17) was used as a nonreplicating integration vector. pBS4S was constructed by ligating the PCR-amplified sacB gene of Bacillus subtilis to pHSG299 (TAKARA BIO Inc., Shiga, Japan) cleaved with AvaII, and the SmaI recognition site in the kanamycin-resistant gene was destroyed by nucleotide substitution using crossover PCR without a change of amino acid residue. Plasmid pVK9 (17), derived from pHM1519 (16), was used to construct a genomic library.

Construction of mutants.To disrupt the odhA gene, an internal fragment of odhA was generated by PCR with the primer pair 5′-GGG GAT CCA TCG GTA TGC CAC ACC GTG GTC GCC-3′ and 5′-GGG GAT CCA CGA CTG CTT CTG CAT CTT CGT TGG-3′ (underlining indicates BamHI restriction site), using ATCC 13869 chromosomal DNA as a template. The amplified fragment was cleaved with BamHI and ligated to BamHI-cleaved pBS4S, resulting in pBS4odhAint. This plasmid, which lacks a functional replication origin in C. glutamicum, was introduced into ATCC 13869 by electroporation. Integration of the plasmid into the chromosomal odhA gene was verified by PCR.

To construct BL1 [NCgl1221(V419::IS1207)], a NCgl1221 gene fragment was amplified by PCR with the primer pair 5′-GGG AGC TCG ACT TTC TGG CTC CTT TAC T-3′ and 5′-GGG AGC TCG CCG ATG ACT AAT AAT GAG A-3′ (underlining indicates SacI restriction site), using the C. glutamicum 2A-1R chromosome as a template. The fragment was cleaved with SacI and ligated to SacI-cleaved pBS4S, yielding pBSo15832A-1.

To construct BL2 [NCgl1221(A111V)] and BL3 [NCgl1221(W15CSLW)], NCgl1221 gene fragments were amplified by PCR with the primer pair 5′-GGG AGC TCG ACT TTC TGG CTC CTT TAC T-3′ and 5′-GGG AGC TCC ACG GCA TGC CGA CCA CCG T-3′ (underlining indicates SacI restriction site), using the C. glutamicum odhA mutant L30-2 or A1 chromosome as a template. The fragment was digested with SacI and ligated to SacI-cleaved pBS4S, yielding pBS4S1583L30 and pBS4S1583A1, respectively.

The NCgl1221 gene disruptant pBS4ΔNCgl1221 was constructed by two successive rounds of PCR. In the first round, one-half of the fragment was amplified using the chromosomal DNA of ATCC 13869 as a template and the primer pair 5′-GGG AGC TCA CCT TTG TGG AGG AAT AGA G-3′ (underlining indicates a SacI restriction site) and 5′-GGG ACA CGT CTG TAA TCA GCG TCC TAG AGC CAA GAT TAG CGC TGA A-3′, while the other half was amplified with the primer pair 5′-TTC AGC GCT AAT CTT GGC TCT AGG ACG CTG ATT ACA GAC GTG TCC C-3′ and 5′-GGG AGC TCA GTA CTC TTC CTT GGA CAT C-3′ (underlining indicates a SacI restriction site). In the second round, the resulting products were mixed in a 1:1 molar ratio and subjected to PCR using the following primer pair: 5′-GGG AGC TCA CCT TTG TGG AGG AAT AGA G-3′ and 5′-GGG AGC TCA GTA CTC TTC CTT GGA CAT C-3′ (underlining indicates SacI restriction sites). The amplified product was digested with SacI and ligated with SacI-digested pBS4S, giving pBS4ΔNCgl1221.

Each plasmid was introduced into C. glutamicum ATCC 13869, and clones with the plasmid integrated into the NCgl1221 gene were selected for further use.

Enzyme preparation and assay.All steps were performed at 4°C. C. glutamicum cells were harvested, washed twice in 0.2% KCl, and suspended in 100 mM N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid-NaOH buffer, pH 7.5, containing 30% glycerol. Cells were disrupted by sonication, and the cellular debris was removed by centrifugation (15,000 × g for 15 min). The crude enzyme extracts were gel filtered in the same buffer on PD10 columns (GE Healthcare Bio-Science Corp., Piscataway, NJ) to remove intracellular metabolites. ODHC activity was measured at 30°C in a photometric assay by following the initial rate in absorbance of 3-acetylpyridine adenine dinucleotide at 365 nm. The standard reaction mixture contained 100 mM N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid-NaOH buffer, pH 7.7, 0.2 mM CoA, 0.3 mM thiamine pyrophosphate, 1 mM sodium 2-oxoglutarate, 3 mM l-cysteine, and 5 mM MgCl_l2. A control experiment was performed without sodium 2-oxoglutarate.

l-Glutamate fermentation.For flask scale fermentation, we used four conditions. For l-glutamate fermentation without induction, cells were cultured in 20 ml medium GH1 [per liter, 30 g glucose, 15 g (NH₄)₂SO₄, 1 g KH₂PO₄, 0.4 g MgSO₄·7H₂O, 0.01 g FeSO₄·7H₂O, 0.01 g MnSO₄·4-5H₂O, 200 μg vitamin B₁ HCl, 300 μg biotin, 13.8 ml soybean protein hydrolysate (total nitrogen, 0.48 g), 50 g CaCO₃ (separately sterilized) adjusted to pH 8.0 with KOH] in 500-ml Sakaguchi flasks with shaking. For fatty acid ester surfactant or penicillin treatments, cells were cultured in 20 ml GH2 medium [GH1 medium but with 80 g glucose, 30 g (NH₄)₂SO₄, and 60 μg biotin] in 500-ml Sakaguchi flasks with shaking until the glucose was completely consumed. Aliquots (2 ml) of these cultures were inoculated into 20 ml GH2 medium. When the optical density at 620 nm (OD₆₂₀) reached about 20, Tween 40 or penicillin G was added to a final concentration of 5 mg/ml or 0.4 U/ml, respectively. For l-glutamate fermentation by biotin depletion, 2-ml aliquots of cell culture were inoculated into 20 ml GH3 medium (GH2 medium without biotin).

For l-glutamate fermentation in a jar fermentor, cells were grown in 300 ml GH4 medium (per liter, 60 g glucose, 1.54 g H₃PO₄, 1.45 g KOH, 0.9 g MgSO₄·7H₂O, 0.01 g FeSO₄·7H₂O, 670 μg vitamin B₁ HCl, 60 μg biotin, 0.28 g dl-methionine, 44 ml soybean protein hydrolysate [total nitrogen, 1.54 g]). For the main fermentation, 30-ml aliquots of cell culture were inoculated into 270 ml GH5 medium (per liter, 111 g glucose, 3.84 g KH₂PO₄, 1.11 g MgSO₄·7H₂O, 0.011 g FeSO₄·7H₂O, 256 μg vitamin B₁ HCl, 11 ml soybean protein hydrolysate [total nitrogen, 0.39 g]). Aeration was controlled by agitation to keep the dissolved oxygen level above 5%. All cells were cultured at 31.5°C. The pH was maintained at 7.2 with ammonia gas.

Extraction of intracellular metabolites.Extraction of intracellular metabolites was performed by a modification of the method previously described (29). Cells were separated from the medium by rapid vacuum filtration (Durapore HV, 0.45 μm; Millipore, Billerica, MA), followed by two washes with 10 ml MilliQ water and incubation of the filter and attached cells for 10 min in 2 ml methanol containing 5 μM morpholineethanesulfonic acid and 5 μM methionine sulfone. The extract (1.6 ml) was mixed with 1.6 ml CHCl₃ and 640 μl MilliQ water, and the aqueous phase was applied to a spin column to remove protein contamination (Ultrafree-MC; Millipore). The filtered extracts were analyzed using capillary electrophoresis mass spectrometry by Human Metabolome Technologies Inc. (Yamagata, Japan). The cell volume used to calculate intracellular metabolite concentrations was 1.6 μl mg⁻¹ dry cell weight (DCW) (4). The DCW was calculated from the OD₆₂₀ by the following experimentally obtained formula: DCW = 22.707 × (OD₆₂₀)/100 (g liter⁻¹).

Disruption of odhA does not result in l-glutamate production.A significant decrease in ODHC activity is observed under l-glutamate-producing conditions (8), suggesting that its attenuation is a prerequisite for l-glutamate production by C. glutamicum. However, it has been observed that some odhA disruptants often produce little, if any, l-glutamate. It has been speculated that l-glutamate productivity is lost due to the genetic instability of odhA disruptants (1), but our results suggest the converse, that additional mutations are responsible for the l-glutamate productivity of some odhA disruptants.

To examine this hypothesis, we constructed odhA disruptants by homologous recombination between the chromosomal odhA gene and a truncated odhA gene. The plasmid pBS4odhAint, which carries an internal fragment of the odhA gene of C. glutamicum, was introduced into wild-type C. glutamicum ATCC 13869 cells and kanamycin-resistant transformants selected, which should harbor the plasmid integrated into the host chromosome by homologous recombination at the odhA locus. The resultant odhA-disrupted strains had extremely unstable phenotypes, forming colonies of various sizes on CM-Dex agar plates.

We cultured several odhA-disrupted clones in liquid GH1 medium and analyzed the metabolites in the culture broth. Most clones produced 2-oxoglutarate, lactate, acetate, and pyruvate but not l-glutamate; however, clone 2A-1 produced high levels of l-glutamate in the presence of excess biotin. In order to determine whether 2A-1 had gained an additional mutation(s) that rendered it an l-glutamate producer, its wild-type odhA gene was restored by homologous recombination using pBS4odhAint. The odhA mutants grew very poorly on medium without sugar because their energy generation depended primarily on glycolysis. Hence, we grew cells of strain 2A-1 on CM2B plates without sugar to cure the plasmid and selected a clone that formed larger colonies. Curing of the plasmid and restoration of the wild-type odhA gene were confirmed by PCR and sequencing as well as measuring of ODHC activity: the odhA⁺ revertant, named 2A-1R, and the wild-type strain had specific ODHC activities of 1.8 ± 0.5 and 3.7 ± 1.0 nmol min⁻¹ (mg protein)⁻¹, respectively, while 2A-1 had no detectable ODHC activity.

The l-glutamate productivity of this odhA⁺ clone was examined. As expected, 2A-1R retained the ability to produce l-glutamate in the presence of excess biotin (Fig. 1). The rates of l-glutamate production were almost the same in 2A-1 (odhA) and 2A-1R (odhA⁺), but the maximum yield of 2A-1R was slightly lower than that of 2A-1. The l-glutamate produced by 2A-1R was assimilated after the glucose had been completely consumed. Since the ability to assimilate l-glutamate depends on ODHC activity, this indirectly confirmed restoration of ODHC activity in 2A-1R. These findings show that odhA disruption does not confer the ability to produce l-glutamate efficiently. Rather, 2A-1 carries an unknown mutation(s) that independently confers the ability to produce l-glutamate.

• Open in new tab
• Download powerpoint

FIG. 1.

Fermentation kinetics of the odhA mutant and revertant in the presence of excess biotin. The wild-type, C. glutamicum ATCC 13869, and the l-glutamic acid-producing odhA mutant 2A-1 and its odhA⁺ derivative 2A-1R were grown with excess biotin at 31.5°C. (a) Growth; (b) residual glucose; (c) l-glutamate accumulation. Symbols: filled circles, wild-type; filled triangles, odhA mutant 2A-1; open squares, 2A-1R.

Identification of a mutation in strain 2A-1R inducing l-glutamate production.Growth of 2A-1R was slow on minimal medium, probably because the substrate (glucose) is used not only for growth but also for l-glutamate production. Based on this observation, we shotgun cloned a gene capable of restoring the growth of 2A-1R on minimal medium. A Sau3AI library of wild-type C. glutamicum chromosomal DNA, constructed in the E. coli-C. glutamicum shuttle vector pVK9, was introduced into 2A-1R by electroporation, and transformants that grew faster on minimal medium plates were selected. From these we isolated a plasmid, designated pL5k, that abolished l-glutamate production and restored rapid growth on minimal medium. Sequencing of a 3-kb Sau3AI fragment from the plasmid showed that it contained only one intact gene, NCgl1221. The 533-amino-acid product of this gene has previously been reported to catalyze betaine efflux (21). Its deduced amino acid sequence has significant homology to mechanosensitive ion channels, such as the E. coli yggB gene product. Mechanosensitive channels are gated by alterations of membrane tension, thus preventing cell disruption by hypo-osmotic shock (13, 30). The NCgl1221 protein has an N-terminal region similar to that of YggB and an additional C-terminal region unlike any known sequence. This type of protein is also present in other corynebacteria, such as C. glutamicum ATCC 13032, Corynebacterium efficiens YS314, Corynebacterium diphtheriae NCTC 13129, and Corynebacterium jeikeium K411 (GenBank accession no. NC_004369, NC_004369, NC_002935, and CR931997, respectively).

The coding region of the endogenous NCgl1221 gene of 2A-1R was found to contain an insertion sequence with 99.9% identity to IS1207 at nucleotide 1258, resulting in the production of a C-terminally truncated NCgl1221 protein of 423 amino acids. To determine whether this mutation was responsible for l-glutamate production, we cloned the mutant NCgl1221 gene [named NCgl1221(V419::IS1207)] and introduced it into wild-type C. glutamicum ATCC 13869. The cloning process requires integration into the chromosome by homologous recombination, since the plasmid lacks a replication origin that functions in C. glutamicum. We isolated a derivative, designated BL1, in which the chromosomal NCgl1221 gene had been replaced by the NCgl1221(V419::IS1207) gene. This strain displayed remarkably high l-glutamate productivity without induction treatment (Table 1), showing that this mutation alone can lead to l-glutamate production.

In the course of constructing BL1, we obtained an integrant carrying both wild-type NCgl1221 and NCgl1221(V419::IS1207). The l-glutamate productivity of this integrant was one-fourth that of BL1 under biotin-sufficient conditions (data not shown). It is therefore possible that the NCgl1221 gene product forms oligomers, like E. coli YggB, which is predicted to form hexamers (30), and that therefore NCgl1221(V419::IS1207) is semidominant. This might explain the ability of the wild-type NCgl1221 gene cloned in the multicopy plasmid to restore the growth of 2A-1R on minimal medium and reduce its l-glutamate productivity.

We also sequenced the NCgl1221 genes of other odhA disruptants of C. glutamicum that produce large amounts of l-glutamate. As expected, most had mutations in the NCgl1221 gene. Substitutions identified were Ala100→Thr [named NCgl1221(A100T)], Ala111→Val [NCgl1221(A111V)], Ala111→Thr [NCgl1221(A111T)], Pro424→Leu [NCgl1221(P424L)], and Trp15→Cys-Ser-Leu-Trp [insertion of three amino acids; NCgl1221(W15CSLW)]. odhA disruptants that produced very little or no l-glutamate carried the wild-type NCgl1221 allele. The PHDhtm program (http://www.predictprotein.org/newwebsite/submit.html ) predicts that NCgl1221 has three transmembrane segments in its N-terminal domain and one transmembrane segment in its C-terminal domain. The NCgl1221(W15CSLW), NCgl1221(A100T), NCgl1221(A111T), and NCgl1221(A111V) substitutions are located in the N-terminal domain, while the NCgl1221(V419::IS1207) and NCgl1221(P424L) substitutions are in the C-terminal domain (Fig. 2a).

• Open in new tab
• Download powerpoint

FIG. 2.

Secondary structure of C. glutamicum NCgl1221 and E. coli YggB and their mutation sites. (a) Membrane topology of C. glutamicum NCgl1221 as predicted by the PHDhtm program. Arrows indicate positions of NCgl1221 gene mutations. Circled numbers indicate residues located on the borders of transmembrane segments. (b) Proposed membrane topology of E. coli YggB based on the PhoA fusion assay (24). Arrows indicate positions of GOF mutations. Circled numbers indicate residues used for the PhoA fusion assay.

E. coli yggB mutations have been described that confer constitutive K⁺ leakage phenotypes in the absence of hypo-osmotic shock, so-called gain-of-function (GOF) mutations (15, 23). Two GOF mutations of E. coli YggB have substitutions, A102P and L109S, respectively, in the third transmembrane helix (Fig. 2b). The positions of NCgl1221(A100T) and probably of NCgl1221(A111V) and NCgl1221(A111T) in the third transmembrane helix resemble these substitutions, suggesting that alterations in this region cause a structural transformation, producing the GOF phenotype: constitutive excretion of l-glutamic acid. The NCgl1221(V419::IS1207) and NCgl1221(P424L) substitutions in the C-terminal region also cause a GOF-like phenotype. In particular, NCgl1221(V419::IS1207) results in truncation of the C-terminal extracellular domain, suggesting that changes in this region can also produce alterations of the structure of the NCgl1221 protein, leading to l-glutamic acid excretion.

We constructed NCgl1221(A111V) and NCgl1221(W15CSLW)derivatives of wild-type C. glutamicum by replacing the NCgl1221 gene on the chromosome. These NCgl1221 gene mutants, designated BL2 and BL3, respectively, also produced remarkably high levels of l-glutamate without induction (Table 1), confirming that NCgl1221 gene mutations on their own cause l-glutamate overproduction.

NCgl1221 gene mutations render cells resistant to the l-glutamate analog, 4-fluoroglutamate.Overexpression of the E. coli arginine exporter YggA renders cells resistant to canavanine, an arginine analog that inhibits cell growth (19). We therefore examined the effect of NCgl1221 gene mutations on sensitivity to an l-glutamate analog. As shown in Fig. 3a, the NCgl1221(A111V) and NCgl1221(W15CSLW) strains grew faster than the wild type on minimal medium containing 4-fluoroglutamic acid. The colony-forming efficiency of the wild-type strain was only 3.2% in the presence of 1.25 mM 4-fluoroglutamic acid, while those of the NCgl1221(A111V) and NCgl1221(W15CSLW) strains were 46% and 60%, respectively (Fig. 3b). These data imply that the mutant NCgl1221 proteins reduce 4-fluoroglutamic acid entry. However, amplification of the wild-type NCgl1221 gene did not affect 4-fluoroglutamic acid sensitivity (data not shown). We speculate that the glutamate export system of wild-type C. glutamicum is inactive under normal growth conditions and that it is activated by conformational changes of NCgl1221 caused by certain NCgl1221 mutations or by treatments that induce l-glutamate production.

• Open in new tab
• Download powerpoint

FIG. 3.

Fluoroglutamic acid resistance of NCgl1221 gene mutants. (a) Serial dilutions (1/10 each) of cultures of wild-type C. glutamicum ATCC 13869 or NCgl1221 gene mutants were spotted onto minimal medium with or without 7.5 mM 4-fluoroglutamic acid, and the plates were incubated at 31.5°C. The upper numbers give cell titers estimated from the OD₆₂₀ of cultures. (b) Cultures of wild-type C. glutamicum ATCC 13869 and its NCgl1221 gene mutants were inoculated into fresh minimal medium with or without 4-fluoroglutamic acid, and the cultures were shaken at 31.5°C. When the OD₆₂₀ of each strain without 4-fluoroglutamic acid reached 1.0, cell numbers were examined by plating appropriate dilutions on CM2B plates. Colonies were counted after 2 days of incubation at 31.5°C. Symbols: filled circles, wild-type; open triangles, NCgl1221 gene deletion mutant; open squares, NCgl1221(A111V); X cross, NCgl1221(W15CSLW). Error bars indicate standard deviations (n = 3).

Effect of NCgl1221 gene amplification on l-glutamate production.The amplification of an amino acid exporter gene is expected to increase excretion of the corresponding amino acid. For example, amplification of the lysE gene encoding the lysine exporter leads to excretion of l-lysine, although a large amount of l-lysine is not excreted by the wild-type strain without treatments to increase the intracellular l-lysine pool (35). It is reported that most amino acid exporters are regulated by intracellular metabolite concentrations. Thus, lysE expression is positively regulated by the transcriptional regulator LysG, which senses the intracellular l-lysine pool (2). brnFE, encoding the l-methionine/l-isoleucine exporter, is also positively regulated by intracellular l-methionine and l-isoleucine levels (33).

We examined the effects of amplifying the NCgl1221 gene on l-glutamate production. The resultant strain produced no l-glutamic acid in the presence of excess biotin; however, when induced by biotin limitation, it produced larger amounts of l-glutamic acid than the wild type (Table 2). Similar results were obtained using other induction conditions, such as Tween 40 or penicillin treatment. Under these conditions, the times required for l-glutamate production to reach plateaus were almost identical (data not shown), although the biomass yield of the NCgl1221 gene-amplified strain was reduced. This means that the l-glutamate secretion rate per cell was increased by NCgl1221 gene amplification. It is also suggested that l-glutamate export is activated by the NCgl1221 protein only under conditions that induce l-glutamate production.

Effects of NCgl1221 gene disruption on l-glutamate production.If the NCgl1221 protein is the major l-glutamate exporter of C. glutamicum, l-glutamate production should be diminished by disrupting the NCgl1221 gene. We deleted the chromosomal NCgl1221 locus and tested the resulting NCgl1221 gene disruptant for l-glutamate productivity in limiting biotin. l-Glutamate production and the glucose consumption rate were both greatly diminished (Fig. 4). After 8 h of cultivation, l-glutamate accumulation in the wild type and the NCgl1221 gene disruptant reached 66.6 mM and 12.2 mM, respectively. The intracellular l-glutamate levels of the wild type and the NCgl1221 gene disruptant decreased gradually from the nonproducing period to the producing period. Although the wild type and the NCgl1221 gene disruptant had similar levels of intracellular l-glutamate in the nonproducing period, intracellular l-glutamate in the disruptant was about 4- to 10-fold higher than that of the wild type during the producing period. We also examined other metabolites. l-Aspartate behaved similarly to l-glutamate. Extracellular l-aspartate was lower in the disruptant than in the wild type, while the reverse was true of intracellular l-aspartate. By contrast, the ratio of intracellular 2-oxoglutarate to extracellular 2-oxoglutarate was almost the same in the wild type and disruptant (data not shown). Similar results were obtained when the l-glutamate production was induced with Tween 40 or penicillin treatment: the amount of l-glutamate after 24 h of cultivation was 6.8-fold or 7.1-fold lower in the disruptant than that in the wild type, respectively. These results suggest that the NCgl1221 protein is the major exporter of the acidic amino acids l-glutamic acid and l-aspartic acid in C. glutamicum. An alternative possibility is that it is a positive regulator that activates an acidic amino acid export system in response to those treatments that induce l-glutamate excretion.

• Open in new tab
• Download powerpoint

FIG. 4.

Effect of NCgl1221 gene deletion on l-glutamate production and the intracellular l-glutamate pool. Wild-type C. glutamicum ATCC 13869 and its NCgl1221 gene deletion derivative were grown on biotin-depleted medium. (a) Growth; (b) residual glucose (symbols: filled circles, wild type; open circles, NCgl1221 gene deletion mutant); (c) extracellular and intracellular l-glutamate; (d) extracellular and intracellular l-aspartate (symbols: filled circles, extracellular pool of the wild type; open circles, extracellular pool of the NCgl1221 gene deletion mutant; filled triangle, intracellular pool of the wild type; open triangle, intracellular pool of NCgl1221 gene deletion mutant).