Enhanced Glucose Consumption and Organic Acid Production by Engineered Corynebacterium glutamicum Based on Analysis of a pfkB1 Deletion Mutant

ABSTRACT

In the analysis of a carbohydrate metabolite pathway, we found interesting phenotypes in a mutant strain of Corynebacterium glutamicum deficient in pfkB1, which encodes fructose-1-phosphate kinase. After being aerobically cultivated with fructose as a carbon source, this mutant consumed glucose and produced organic acid, predominantly l-lactate, at a level more than 2-fold higher than that of the wild-type grown with glucose under conditions of oxygen deprivation. This considerably higher fermentation capacity was unique for the combination of pfkB1 deletion and cultivation with fructose. In the metabolome and transcriptome analyses of this strain, marked intracellular accumulation of fructose-1-phosphate and significant upregulation of several genes related to the phosphoenolpyruvate:carbohydrate phosphotransferase system, glycolysis, and organic acid synthesis were identified. We then examined strains overexpressing several of the identified genes and demonstrated enhanced glucose consumption and organic acid production by these engineered strains, whose values were found to be comparable to those of the model pfkB1 deletion mutant grown with fructose. l-Lactate production by the ppc deletion mutant of the engineered strain was 2,390 mM (i.e., 215 g/liter) after 48 h under oxygen deprivation, which was a 2.7-fold increase over that of the wild-type strain with a deletion of ppc.

IMPORTANCE Enhancement of glycolytic flux is important for improving microbiological production of chemicals, but overexpression of glycolytic enzymes has often resulted in little positive effect. That is presumably because the central carbon metabolism is under the complex and strict regulation not only transcriptionally but also posttranscriptionally, for example, by the ATP/ADP ratio. In contrast, we studied a mutant strain of Corynebacterium glutamicum that showed markedly enhanced glucose consumption and organic acid production and, based on the findings, identified several genes whose overexpression was effective in enhancing glycolytic flux under conditions of oxygen deprivation. These results will further understanding of the regulatory mechanisms of glycolytic flux and can be widely applied to the improvement of the microbial production of useful chemicals.

INTRODUCTION

Corynebacterium glutamicum has been used as a powerful workhorse in the industrial production of amino acids, such as l-glutamate and l-lysine (1, 2). This organism behaves rather similarly to an aerobe because little growth is identified anaerobically without an alternative electron acceptor, nitrate (3, 4). In contrast, under oxygen-deprived conditions, C. glutamicum metabolizes glucose to l-lactate, succinate, and acetate even though its cell growth is arrested (5–7). Under these conditions, the glucose consumption rate becomes higher than that attained during aerobic cultivation (the Pasteur effect [8]), and improved product yields are additionally expected due to repression of carbon loss by cell growth and respiration. On the basis of these characteristics, we have demonstrated overproduction of organic acids (9, 10), amino acids (11–13), and a biofuel (14, 15) by engineered C. glutamicum R strains under oxygen deprivation.

The strategies for metabolic engineering to increase productivity have been commonly focused on improvement of the downstream synthesis pathway of the target chemicals as well as repression of the competitive pathways that form by-products (16–18). In addition, enhancement of upstream central carbon metabolism, i.e., uptake and metabolism of carbon sources, should be important so as to supply the precursor metabolites and coenzymes more efficiently. Nonetheless, it still remains unclear whether overexpression of glycolytic enzymes is effective for enhancement of glucose consumption and productivity; many attempts have been unsuccessful while some studies have led to positive results, and little is known about what has led to these discrepant results (19). In any case, a more detailed understanding of the regulatory mechanism of central carbon metabolism is essential for further improvement of productivity.

In C. glutamicum, the genes related to the central carbon metabolism are under complex and strict transcriptional control by multiple regulators in response to extracellular and/or intracellular conditions, such as carbon sources (20). One of the pivotal regulators is SugR, which globally represses transcription of genes related to the phosphoenolpyruvate:carbohydrate phosphotransferase system (PTS) (21–24), glycolysis (24, 25), and l-lactate dehydrogenase (LDH) (22, 24, 26). SugR-driven repression is relieved when the bacteria are cultivated with sugars such as glucose or fructose because sugar metabolite intermediates, sugar phosphates, are potent effectors to inhibit SugR function (21, 22, 25, 26). In addition, RamA, RamB (27), GlxR (28, 29), GntR1, and GntR2 (30, 31) are known as transcriptional regulators of multiple genes involved in carbon metabolism.

On the basis of the studies of transcriptional regulation of the central carbon metabolism in C. glutamicum to date, we are particularly interested in the regulatory mechanism of PTS and fructose metabolism genes. PTS catalyzes the coupled uptake and phosphorylation of sugars by using phosphoenolpyruvate (PEP), not ATP, as a high-energy phosphoryl group donor (32). The components of PTS consist of two common cytoplasmic proteins, enzyme I (EI) and HPr, and sugar-specific membrane-bound proteins, enzyme IIs (EIIs). There are three EIIs, encoded by ptsG, ptsF, and ptsS, which are specific to glucose, fructose, and sucrose, respectively, that have been identified in C. glutamicum, (33, 34). In addition, unlike ATCC 13032, one of the most representative strains of C. glutamicum, C. glutamicum R has two other EIIs encoded by bglF and bglF2, which are specific to β-glucoside (35, 36). In many bacteria, the genes encoding EI and HPr are generally located near those of the glucose-specific EII (EII^Glc) and are subject to joint transcriptional regulation such that glucose is the most effective inducer of EI and HPr expression (32). In contrast, as shown in Fig. 1, ptsI and ptsH, which encode EI and HPr, are flanked by the fruR-pfkB1-ptsF operon, which encodes a DeoR-type regulator, fructose-1-phosphate kinase (F1PK), and fructose-specific EII (EII^Fru), respectively, in C. glutamicum. In addition, the global DeoR-type regulator gene sugR is located in the vicinity of the fructose operon. Consequent to the cluster arrangement, the PTS's common components are transcriptionally regulated together with the fructose metabolism enzymes. FruR represses the transcription of not only the fruR-pfkB1-ptsF operon but also ptsI and ptsH in the presence of fructose. Both ptsI and ptsH are upregulated to a greater extent, concomitantly with the fruR-pfkB1-ptsF operon, by fructose than by glucose (37). In addition, expression levels of several glycolytic genes are greater with fructose than with glucose (38). Therefore, we expected that analysis of the fructose metabolism in C. glutamicum would lead to useful knowledge about how to increase glucose consumption and productivity. In this study, we found interesting phenotypes, i.e., markedly enhanced glucose consumption and organic acid production of a mutant strain deficient in a fructose metabolism gene, as expected. Further analysis of this strain revealed metabolic and transcriptional changes in central carbon metabolism that would likely be involved in the phenotypes. On the basis of these findings, we examined the overexpression of identified glycolytic genes and observed increased glucose consumption and organic acid production.

FIG 1.

Gene clusters of common and fructose-specific PTSs in C. glutamicum.

RESULTS AND DISCUSSION

Effect of fruR, pfkB1, and ptsF deletions and carbon sources on aerobic growth and oxygen-deprived glucose consumption.

We disrupted the fructose metabolism genes fruR, pfkB1, and ptsF, thus creating the respective deletion mutants of C. glutamicum, the ΔfruR, ΔpfkB, and ΔptsF strains (39) (Table 1). The fructose metabolism pathway in C. glutamicum is shown in Fig. 2. Fructose is mainly imported and phosphorylated by EII^Fru, encoded by ptsF, and consequently fructose-1-phosphate (F1P) is formed. In addition, EII^Glc, encoded by ptsG, also weakly catalyzes fructose uptake, resulting in fructose-6-phosphate (F6P) in this case (34, 38). F1P is channeled to subsequent glycolysis after conversion to fructose-1,6-bisphosphate by F1PK, encoded by pfkB1, as well as F6P by fructose-6-phosphate kinase (PFK), encoded by pfkA. The fruR gene encodes a transcriptional repressor of the fruR-pfkB1-ptsF operon as well as of ptsI and ptsH (37). Consequently, fructose metabolism was expected to be interrupted in the ΔpfkB and ΔptsF strains while it could be accelerated in the ΔfruR strain.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s)	Reference or source
Strains
C. glutamicum R	Wild type (JCM 18229)	55
ΔfruR strain	Markerless fruR deletion mutant	This study
ΔpfkB strain	Markerless pfkB1 deletion mutant	This study
ΔptsF strain	Markerless ptsF deletion mutant	39
R_PTS	Wild type with induction of ptsI, ptsH, and ptsG	This study
R_GPT	Wild type with induction of gapA, pgk, and tpi	This study
R_LDH	Wild type with induction of ldhA	This study
R_PG	Wild type with induction of ptsI, ptsH, ptsG, gapA, pgk, and tpi	This study
R_PL	Wild type with induction of ptsI, ptsH, ptsG, and ldhA	This study
R_GL	Wild type with induction of gapA, pgk, tpi, and ldhA	This study
R_PGL	Wild type with induction of ptsI, ptsH, ptsG, gapA, pgk, tpi, and ldhA	This study
R_PGLi	R_PGL with induction of pgi	This study
R_PGLf	R_PGL with induction of pfkA	This study
R_PGLif	R_PGL with induction of pgi and pfkA	This study
Δppc strain	Markerless ppc deletion mutant	This study
R_PGLif(ΔP)	R_PGLif with markerless deletion of ppc	This study
Plasmids
pCRA725	Kmr; Ptac-sacR-sacB in pHSG298	14
pCRD904	Apr; Ptac-SD-rrnB terminator in pCold Ia	49

^aSD, sequence containing a ribosome binding site (49).

FIG 2.

Glucose and fructose uptake by PTS and the downstream glycolytic and organic acid synthesis pathways in C. glutamicum. The dashed line indicates weak activity toward fructose. G6P, glucose-6-phosphate; F1P, fructose-1-phosphate; F6P, fructose-6-phosphate; FBP, fructose-1,6-bisphosphate; DHAP, dihydroxyacetone phosphate; GAP, glyceraldehyde-3-phosphate; BPG, 1,3-bisphosphoglycerate; 3PG, 3-phosphoglycerate; 2PG, 2-phosphoglycerate; PEP, phosphoenolpyruvate; OXA, oxaloacetate.

First, we examined the effect of each gene deletion on the aerobic growth of C. glutamicum on nutrient-rich A medium (see Materials and Methods) with glucose or fructose. As shown in Fig. 3A and C, the four strains showed almost the same profiles of growth and carbohydrate consumption when cultivated with glucose. They exponentially grew from 4 to 8 h and shifted to the stationary phase, where cell growth almost ceased due to an insufficient oxygen supply and/or low pH resulting from the production of l-lactate. For example, when the wild type was cultivated with glucose, about 30 mM l-lactate accumulated, and the pH of the medium decreased from 7.2 to below 5.0. Even when cultivated with fructose, the wild-type and ΔfruR strains grew and consumed the carbohydrate as well as they did with glucose, except for an earlier shift from exponential growth to the stationary phase at 6 h (Fig. 3B and D). In contrast, the growth and fructose consumption of the ΔptsF mutant were significantly decreased especially in the exponential growth phase, and consequently it was difficult to distinguish the exponential phase from the stationary phase. This resulted from the insufficient fructose uptake ability of the ΔptsF strain, which depends only on the weak-side activity of EII^Glc (Fig. 2) (34). Significant retardation of growth and of fructose consumption was also observed in the ΔpfkB mutant because F1P, which is formed after import by EII^Fru, could not be metabolized further (Fig. 2). Nonetheless, the retardation of both was accelerated afterward; the ΔpfkB strain consumed more fructose than the ΔptsF strain did after 12 h and reached an amount comparable to that consumed by the wild type at 24 h. Since there was probably no other pathway present to import and metabolize fructose in the ΔpfkB strain except that through insufficient EII^Glc, ptsG (in addition, perhaps ptsI and ptsH) may be upregulated in this strain. In contrast to other conditions, the ΔpfkB and ΔptsF mutants cultivated with fructose hardly produced l-lactate at 12 h, and consequently both strains subsequently grew until 24 h. Even after 24 h, little l-lactate was produced by the ΔptsF strain with fructose, and cell growth continued thereafter but stopped after 48 h as well (Fig. S1 in the supplemental material), at which time the l-lactate concentration was over 20 mM and pH was near 5.0. These smaller production levels of l-lactate and the prolonged growth of the ΔptsF and ΔpfkB strains would result from the insufficient fructose uptake. That is, the amounts of fructose taken up were so small that almost all of it was preferentially used for cell growth, and little remained for l-lactate production until cell growth was stopped by a depleted oxygen supply (high cell density).

FIG 3.

Aerobic growth (A and B) and sugar consumption (C and D) of the wild-type (wt), ΔfruR, ΔpfkB, and ΔptsF strains on A medium with glucose (A and C) or fructose (B and D). Data are presented as means ± standard deviations of three independent experiments.

Under oxygen deprivation, C. glutamicum cannot grow but produces significant amounts of organic acid, predominantly l-lactate, from glucose (5–7), and this property is expected to be an appropriate indicator of glycolytic flux. Accordingly, we then analyzed the oxygen-deprived glucose consumption and organic acid production rates of the deletion mutants after growth with glucose or fructose. Figure 4 shows the glucose consumption of each strain on minimal BT-U medium (see Materials and Methods) under oxygen-deprived conditions. The organic acid production rates of the reaction are summarized in Table S1, together with the respective glucose consumption rates, showing that organic acid production was enhanced concomitantly with the increased glucose consumption. In the case of glucose-grown cells, the wild-type strain consumed 25.9 mM/h glucose in 12 h (Fig. 4A). The ΔfruR and ΔpfkB strains showed about 10% higher glucose consumption levels (29.5 and 28.6 mM/h, respectively) than the wild type, and the level in the ΔptsF strain was indistinguishable from that of the wild type (26.0 mM/h). On the other hand, when fructose was used as a carbon source for cell growth, the glucose consumption rate of each strain was significantly different (Fig. 4B). The fructose-grown wild type consumed glucose at a 21% greater rate (31.3 mM/h) than the glucose-grown wild-type strain. The ΔfruR strain consumed glucose 20% faster (37.6 mM/h) than this wild-type strain, and the ΔpfkB mutant consumed glucose more than twice as fast (63.5 mM/h). In contrast, the ΔptsF mutant consumed glucose much more slowly than the wild type (18.1 mM/h). These results suggest that deletion of fructose metabolite genes significantly affected not only fructose metabolism but also oxygen-deprived glucose metabolism. The increased rates of the ΔfruR strain suggest that the rate-limiting enzymes were upregulated by disruption of fruR even when the cells were cultivated with glucose in addition to fructose, in contrast to a previous report where PTS-related genes were upregulated only when cultivated with fructose in an fruR deletion mutant (37). Deletion of ptsF had an effect only when the cells were grown with fructose, during which the central carbon metabolism might be downregulated coordinately with the weak capacity for fructose uptake. Among these results, the enhanced glucose consumption and organic acid production of the ΔpfkB strain grown with fructose, with levels more than 2-fold greater than the level of the wild type, are outstanding. Although this strain showed the increased rates even when grown with glucose, the amount of increase was only 10% of that attained in the wild type. Thus, both pfkB1 deletion and cultivation with fructose are needed to induce the enhanced glucose consumption and organic acid production. The notably enhanced glucose consumption and organic acid production of the ΔpfkB strain grown with fructose are very interesting in terms of the regulation mechanism of glycolytic flux and highly useful for improvement of productivity. Accordingly, we further characterized the ΔpfkB mutant to elucidate how the remarkable fermentative capacity was achieved.

FIG 4.

Glucose consumption of each strain under oxygen deprivation on BT-U medium after aerobic growth with glucose (A) or fructose (B). The cell concentration was 20 g (cell dry weight)/liter. Data are presented as means ± standard deviations of three independent experiments. wt, wild type.

Metabolome analysis of the ΔpfkB mutant during aerobic growth with fructose and oxygen-deprived glucose consumption.

In order to unveil what changes in the central carbon metabolism are associated with the enhanced fermentative capacity, we analyzed intracellular metabolites of the ΔpfkB mutant during aerobic cultivation with glucose or fructose in the late exponential phase (at 6 h with glucose and at 12 h with fructose) (Fig. 3). As a result, almost 200-fold greater accumulation of F1P was detected when the ΔpfkB strain was cultivated with fructose than with glucose; F1P concentrations were 152 mM and 0.79 mM in the cells cultivated with fructose and glucose, respectively (Table 2). These results are reasonable because the ΔpfkB strain cannot metabolize F1P, which is formed from fructose after import by EII^Fru (Fig. 2).

TABLE 2.

Intracellular metabolite concentrations in the ΔpfkB strain during aerobic cultivation and subsequently oxygen-deprived organic acid production

Metabolite or coenzyme ratio	Concn or ratio for the cells under indicated conditionsb
	Aerobic cultivation		Oxygen deprivation
	Glucose	Fructose	Glucose	Fructose
Metabolites (mM)a
F1P	0.79 ± 0.05	152.34 ± 12.52	2.78 ± 0.46	2.16 ± 0.16
G6P	5.96 ± 0.16	0.94 ± 0.07	1.99 ± 0.20	5.31 ± 0.62
F6P	3.59 ± 0.11	0.87 ± 0.11	1.12 ± 0.16	1.48 ± 0.14
FBP	14.30 ± 2.25	1.11 ± 0.04	OD	OD
DHAP	3.56 ± 0.19	3.86 ± 0.13	8.09 ± 1.26	9.64 ± 1.51
GAP	1.16 ± 0.21	ND	0.82 ± 0.09	1.08 ± 0.06
BPG	0.18 ± 0.03	0.16 ± 0.02	0.06 ± 0.01	0.05 ± 0.01
3PG + 2PG	3.89 ± 0.14	3.21 ± 0.12	1.99 ± 0.20	1.63 ± 0.26
PEP	0.36 ± 0.02	0.86 ± 0.10	0.30 ± 0.05	0.21 ± 0.02
Pyruvate	35.28 ± 4.48	117.33 ± 24.87	ND	1.76 ± 0.16
Ac-CoA	0.12 ± 0.00	0.18 ± 0.01	0.04 ± 0.01	0.08 ± 0.01
Cit + Icit	3.14 ± 0.14	2.80 ± 0.20	0.46 ± 0.05	0.49 ± 0.08
α-KG	9.70 ± 0.51	22.63 ± 1.30	1.06 ± 1.83	2.64 ± 0.18
Suc-CoA	0.57 ± 0.04	1.32 ± 0.11	0.21 ± 0.06	0.32 ± 0.04
Succinate	OD	OD	OD	OD
Fumarate	2.12 ± 0.42	2.37 ± 0.48	0.32 ± 0.05	0.57 ± 0.05
Malate	8.85 ± 0.76	8.92 ± 0.77	1.42 ± 0.39	1.67 ± 0.08
Oxaloacetate	1.47 ± 0.15	2.18 ± 0.12	0.12 ± 0.02	0.04 ± 0.01
Coenzyme ratios
ATP/ADP	3.69 ± 0.35	1.97 ± 0.13	1.93 ± 0.13	2.89 ± 0.13
NADH/NAD+	0.05 ± 0.01	0.10 ± 0.02	1.03 ± 0.12	0.89 ± 0.10

^aF1P, fructose-1-phosphate; G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; FBP, fructose-1,6-bisphosphate; DHAP, dihydroxyacetone phosphate; GAP, glyceraldehyde-3-phosphate; BPG, 1,3-bisphosphoglycerate; 3PG, 3-phosphoglycerate; 2PG, 2-phosphoglycerate; PEP, phosphoenolpyruvate; Ac-CoA, acetyl coenzyme A; Cit, citrate; Icit, isocitrate; α-KG, α-ketoglutarate; Suc CoA, succinyl-coenzyme A.

^bAerobic cultivation, cells under aerobic conditions on A medium with glucose or fructose in the late exponential phase; oxygen deprivation, cells under oxygen deprivation conditions on BT-U medium with glucose after aerobic growth with glucose or fructose. The data are from five analytical replicates from one experiment. OD, overdose; ND, not detected.

We also conducted metabolome analysis in the glucose- and fructose-grown ΔpfkB strain during organic acid production from glucose under oxygen deprivation (at 2 h) (Fig. 4). The resultant data on intracellular metabolite concentrations are shown in Table 2. The extraordinary F1P accumulation, observed during aerobic cultivation with fructose for the cell preparation (Table 2), was abrogated in the reactions. In this connection, previous studies have demonstrated that the intracellular NADH/NAD⁺ ratio is one of the predominant factors that influence the reaction rate under oxygen-deprived conditions (6, 12, 13, 40). Nonetheless, despite the considerable differences in the respective glucose consumption rates (Fig. 4), there were no significant differences in the ratios between the cells grown with glucose and fructose; the ratios with glucose and fructose were 1.03 and 0.89 (P > 0.05), respectively. Thus, the enhanced rates of the fructose-grown ΔpfkB strain resulted from mechanisms that are independent of the NADH/NAD⁺ ratio. On the other hand, the amounts of almost all metabolites in the fructose-grown cells are higher than those in the glucose-grown cells coordinately with the increased glucose consumption and organic acid production rates, whereas those of 1,3-bisphosphoglycerate (BPG), 3-phosphoglycerate, 2-phosphoglycerate, PEP, and oxaloacetate were exceptionally smaller. These profiles suggest unique mechanisms, i.e., upregulation of the enzymes catalyzing the reactions from BPG to pyruvate, such as phosphoglycerate kinase (PGK), phosphoglycerate mutase, enolase, pyruvate kinase (PYK), and PTS (Fig. 2). Among the listed enzymes, significant upregulation of genes encoding PGK and PTS was also identified in transcriptome analysis, as described below (Fig. 5). In addition, a notably increased pyruvate pool (Table 2) also implies that upregulated PTS would play a dominant role in the changed metabolome profiles because PTS catalyzes a reaction from PEP to pyruvate so as to obtain a high-energy phosphate group needed to import and phosphorylate glucose (32). That is, overexpressed PTS can lead to a decrease in the PEP pool, and this change in turn promotes the upstream enzyme reactions and a decrease in the respective substrates in the glycolytic pathway toward BPG (Fig. 2).

FIG 5.

Transcriptional analysis of genes in the central carbon metabolism by DNA microarray analysis during aerobic growth. (A) The wild-type strain cultivated with glucose, analyzed at 4 h and 12 h. (B) ΔpfkB strain cultivated with fructose, analyzed at 4 h and 12 h.

Transcriptome analysis of the ΔpfkB strain during aerobic growth with fructose.

To identify transcriptional changes that contributed to the significantly increased glucose consumption and organic acid production in the ΔpfkB strain grown with fructose, we performed transcriptome analysis by DNA microarray and compared the data with those from the wild type grown with glucose as a control. RNA was isolated from cells during aerobic cultivation with glucose or fructose in the early exponential phase and late exponential/stationary phases (at 4 and 12 h) (Fig. 3). Figure 5 shows the relative mRNA levels of genes in central carbon metabolism compared with those of the wild type with glucose at 4 h. In the ΔpfkB strain with fructose at 4 h, we identified several genes, such as ptsI, ptsH, ptsF, ptsS, tpi, gapA, pgk, ldhA, ppc, sucC, and sucD, which were upregulated more than 2-fold over the levels in the control (Fig. 5B). In addition, although the transcription levels of almost all genes at 12 h were much decreased compared with those at 4 h in the wild type with glucose (Fig. 5A), the relatively higher transcriptional levels were still maintained in many genes even at 12 h in the ΔpfkB strain grown with fructose (Fig. 5B). In this regard, in the cases of the wild type grown with fructose and the ΔpfkB strain grown with glucose, the central carbon metabolism genes were also upregulated over the level of the wild type with glucose, but the transcription levels were lower than those in the ΔpfkB strain with fructose, especially at 12 h (Fig. S2), suggesting that either the pfkB1 deletion or the growth with fructose alone was insufficient to lead to the significant upregulation observed in the ΔpfkB strain grown with fructose. In contrast to the wild type with glucose or fructose and the ΔpfkB strain with glucose, where the exponential growth phase and stationary phase were clearly distinguishable, the ΔpfkB strain with fructose grew slowly in the exponential phase, but subsequently its cell growth was continued even at 12 h and thereafter (Fig. 3B); this growth profile might contribute to its long-lasting gene transcription at higher levels. The significant upregulation of ptsI, ptsH, ptsF, ptsS, gapA, and ldhA in the ΔpfkB strain with fructose at 4 h and the relatively high transcription at 12 h of the pgk and ptsG genes were also confirmed by quantitative reverse transcription-PCR (qRT-PCR) analysis (Fig. S3). Furthermore, increased expression of EI and EII^Glc was confirmed by Western blot analysis (Fig. S4). These upregulated enzymes are expected to cause the enhanced glucose consumption and organic acid production.

Evaluation of the oxygen-deprived glucose consumption rate of recombinant strains overexpressing the genes upregulated in the ΔpfkB strain cultivated with fructose.

On the basis of the above findings, we first selected ptsI, ptsH, ptsG, tpi, gapA, pgk, and ldhA as candidate genes whose upregulation should contribute to an increase in glucose consumption and organic acid production in the ΔpfkB mutant cultivated with fructose. The candidate genes were classified into the following three groups. The first group consists of ptsI, ptsH, and ptsG, which are associated with PTS although ptsG was not significantly upregulated at 4 h in the ΔpfkB mutant with fructose but was at 12 h (Fig. 5). The second group consists of gapA, pgk, and tpi (this set of three genes is designated GPT), which encode the glycolytic enzymes glyceraldehyde-3-phosphate dehydrogenase (GAPDH), PGK, and triosephosphate isomerase, respectively (Fig. 2), and whose transcription is synchronously regulated as a single gapA-pgk-tpi-ppc operon (41, 42). The third group consists of the gene ldhA, which encodes LDH that catalyzes the synthesis of l-lactate, a predominant product of C. glutamicum under oxygen deprivation. In accordance with the classification, we constructed recombinant strains overexpressing genes in the first (ptsI, ptsH, and ptsG), second (gapA, pgk, and tpi), and third (ldhA) groups and designated them R_PTS, R_GPT, and R_LDH, respectively (Table 1).

We evaluated the resultant strains in terms of the glucose consumption rate under conditions of oxygen deprivation after growth with glucose. Figure 6 shows the glucose consumption rate of each recombinant strain, and the summary containing organic acid production rates is shown in Table S1. Glucose consumption rates were increased by about 10%, from 25.9 to 29.8, 28.5, and 28.8 mM/h, by overexpressing the PTS, GPT, and LDH genes, respectively. We then constructed recombinant strains R_PG, R_PL, R_GL, and R_PGL (where P is PTS, G is GPT, and L is LDH), which concomitantly overexpress two or three sets of the genes of PTS, GPT, and LDH (Table 1), to confirm their synergistic effects. As a result, glucose consumption rates of the recombinant strains were increased synergistically, and, consequently, the rate of the strain overexpressing all genes of the three groups (triple-overexpressing), R_PGL, reached 38.5 mM/h, which is 48% higher than that of the wild type (Fig. 6).

FIG 6.

The glucose consumption rates of the wild-type, ΔpfkB, and recombinant strains under oxygen deprivation on BT-U medium. The cells used for the reaction were prepared by aerobic cultivation on A medium with glucose, with the exception of the ΔpfkB strain, which was grown with fructose. Reaction duration was 12 h or before depletion of the initial amount of glucose (400 mM). The cell concentration was 20 g (cell dry weight)/liter. Data are presented as means ± standard deviations of three independent experiments.

The glucose consumption rate was indeed increased by overexpressing the PTS, GPT, LDH genes, but even the rate of the triple-overexpressing strain R_PGL was still lower than that of the ΔpfkB strain grown with fructose (Fig. 6). We accordingly tested the other seven glycolytic genes to find additional gene candidates that enhance glucose consumption. As a result, overexpression of pgi and pfkA, which encode glucose-6-phosphate isomerase and PFK, respectively (Fig. 2), enhanced glucose consumption individually and synergistically. R_PGLi, R_PGLf, and R_PGLif, which are R_PGL-derived strains additionally overexpressing pgi, pfkA, and both pgi and pfkA, respectively (Table 1), consumed 42.7, 54.0, and 65.2 mM/h glucose, respectively; these values are equivalent to 1.6-, 2.1-, and 2.5-fold increases, respectively, relative to the wild-type level (Fig. 6). Note that the resultant rate of 65.2 mM/h in R_PGLif is comparable to 63.5 mM/h attained in the ΔpfkB strain grown with fructose. These results demonstrate that overexpression of the PTS, glycolytic, and organic acid synthesis genes is effective at increasing glucose consumption and organic acid production in the oxygen-deprived reaction.

l-Lactate production by the recombinant strain overexpressing the PTS, glycolysis, and LDH genes under oxygen deprivation.

In this study, several candidate genes related to PTS, glycolysis, and organic acid production whose overexpression enhanced glucose consumption and organic acid production were identified. We finally evaluated the effect of the candidate genes on oxygen-deprived l-lactate production because l-lactate is one of the most representative products of microbial fermentation. In this respect, C. glutamicum produces predominantly l-lactate under oxygen deprivation, but significant amounts of succinate are concomitantly formed (Table S1), which is not preferable for improvement of l-lactate productivity and yield. On the other hand, succinate is known to be synthesized mainly via phosphoenolpyruvate carboxylase which is encoded by ppc and catalyzes carboxylation from PEP to oxaloacetate (Fig. 2), and deletion of the ppc gene has successfully suppressed succinate formation (6, 14). Accordingly, we constructed R_PGLif(ΔP), in which ppc was disrupted from R_PGLif, as an engineered strain for l-lactate production (Table 1). A deletion mutant of ppc (Δppc strain) (Table 1) was evaluated together as a control.

Figure 7 shows l-lactate production from glucose under oxygen deprivation. The Δppc strain consumed 495 mM glucose and produced 872 mM l-lactate during the 48-h reaction. In contrast, glucose consumption and l-lactate production of R_PGLif(ΔP) were 1,250 mM and 2,390 mM, respectively, which are 2.6- and 2.7-fold higher than the levels of the Δppc strain. In addition, the l-lactate yield of R_PGLif(ΔP) was improved to 1.91 from 1.76 mol/mol glucose attained in the Δppc strain. The resultant l-lactate production of 2,390 mM, i.e., 215 g/liter, after 48 h is one of the highest values reported so far (e.g., 198 g/liter after 36 h by Lactobacillus casei G-03 [43] and 192 g/liter after 48 h by Lactobacillus paracasei subsp. paracasei CHB2121 [44]). These results demonstrate that overexpression of the candidate genes identified in this study is highly effective for enhancing productivity even in a prolonged reaction.

FIG 7.

Fed-batch production of l-lactate from glucose by the Δppc and R_PGLif(ΔP) strains on BT-U medium under oxygen deprivation. The cells used for the reaction were prepared by aerobic growth on A medium with glucose. The cell concentration was 20 g (cell dry weight)/liter. Data are presented as means ± standard deviations of three independent experiments.

Putative mechanisms that led to the enhanced glucose consumption and organic acid production in the ΔpfkB strain grown with fructose.

In this study, we analyzed the pfkB1 deletion mutant of C. glutamicum grown with fructose and identified several genes related to PTS, glycolysis, and organic acid synthesis whose upregulation and overexpression were effective for enhancing glucose consumption and productivity under oxygen deprivation. These results are very interesting and valuable for increasing the productivity of other useful chemicals, but in order to use these findings in practice, greater understanding of the detailed mechanisms that led to the improved results is important. Glycolytic flux is predominantly controlled by the ATP demand in aerobic cultivation through PFK and PYK activities that are allosterically regulated by such molecules as ATP and ADP (19, 45, 46). On the other hand, as shown by the Pasteur effect, the glucose consumption rate is enhanced under O₂-deficient conditions in which the ATP supply is insufficient due to the lack of respiration that efficiently produces ATP. Similarly, Yokota and coworkers succeeded in enhancing glucose consumption and l-glutamate (47) and l-valine (48) production by an H⁺-ATPase-defective mutant of C. glutamicum even in aerobic cultivation. Thus, larger amounts of glycolytic enzymes than those required to maintain glycolysis flux under aerobic conditions should already be present. In other words, glucose consumption and productivity would not be increased even by overexpression of glycolytic enzymes in this case.

In contrast, under anaerobic conditions where the regulation of glycolytic flux by ATP demand is relieved, the flux would in turn be limited by an insufficient NAD⁺ supply (the higher NADH/NAD⁺ ratio) due to lack of respiration that efficiently reoxidizes NADH. This limitation is attributed to simple inhibition by surplus NADH of the GAPDH reaction (38), in contrast to the complicated allosteric regulation depending on ATP demand in aerobic cultivation. Consequently, under anaerobic conditions, overexpression of GAPDH and perhaps of other glycolytic enzymes should be effective to enhance glycolytic flux. Indeed, we have demonstrated that glucose consumption and productivity were enhanced by overexpressing gapA (11), pyk, pfkA, pgi (49), and tpi (15) under oxygen deprivation, whereas gapA overexpression had no effect on the glucose consumption rate under aerobic conditions (11). In this study, overexpression of pgi, pfkA, tpi, gapA, and pgk increased oxygen-deprived glucose consumption and organic acid production as well (Fig. 6). In addition, our study revealed that overexpression of the general and glucose-specific PTS genes, i.e., ptsI, ptsH, and ptsG, was also effective. Interestingly, in addition to GAPDH itself, all of the enzymes that improved glucose consumption and productivity identified in this study are located upstream and just downstream of GAPDH (Fig. 2) (50). Therefore, these overexpressed enzymes presumably enhanced glycolytic flux predominantly by accelerating the GAPDH reaction; i.e., in addition to overexpression of GAPDH itself, an increase in substrate supply and a decrease in the product would synergistically promote the GAPDH reaction. The metabolome profiles of exceptionally decreased pools of BPG to PEP in the glycolysis pathway (Fig. 2) in the fructose-grown ΔpfkB mutant (Table 2) support this assumption; i.e., the decreased BPG pool, which is a direct product of the GAPDH reaction, would result in acceleration of the enzyme reaction, and consequently glucose consumption and organic acid production would be enhanced.

Possible involvement of SugR affected by accumulated F1P in the phenotype of the ΔpfkB mutant grown with fructose.

As mentioned above, we identified several genes whose upregulation/overexpression enhanced glucose consumption and organic acid production under oxygen deprivation, based on the analysis of a ΔpfkB mutant grown with fructose. Although the detailed mechanisms that caused the upregulation are still unclear, the resultant data in this study suggest involvement of SugR, one of the important transcriptional regulators of sugar metabolism (20). The extraordinarily accumulated F1P in the ΔpfkB strain cultivated with fructose (Table 2) is the most powerful effector to inhibit the repressor activity of SugR (22, 25, 26). In addition, almost all of the genes whose upregulation was identified in the ΔpfkB strain grown with fructose (Fig. 5) and whose overexpression was effective at improving glucose consumption and organic acid production (Fig. 6 and Table S1) are consistent with the genes that have been reported to be subjected to repression by SugR, such as ptsI (21–24), ptsH (22–24), fruR-pfkB1-ptsF (21–24), ptsG (21–24), ptsS (21, 22, 24), pfkA, fba, eno, pyk, pyc (24), gapA-pgk-tpi-ppc (25), and ldhA (22, 24, 26). The putative mechanisms of enhanced glucose consumption and organic acid production of the ΔpfkB strain grown with fructose are as follows. First, fructose is imported by EII^Fru and converted to F1P, but F1P is not metabolized further due to the deficiency in pfkB1, resulting in intracellular F1P accumulation (Fig. 2). The accumulated F1P then inhibits binding of SugR to each promoter region of the genes in the central carbon metabolism; i.e., the transcriptional repression by SugR is relieved, and the genes are upregulated. Consequently, the overexpressed enzymes will lead to enhanced glucose consumption and organic acid production in the ΔpfkB strain grown with fructose. Very recently, a similar model of sugar metabolism regulated through F1P and SugR has been proposed based on the study of ptsF and pfkB deletion mutants of C. glutamicum grown on sucrose (51).

In this connection, the above hypothesis suggests that deletion of sugR should also give rise to the same result as that of the ΔpfkB mutant grown with fructose. We thus tested a sugR deletion mutant, the ΔsugR strain (25), but this strain unexpectedly did not show positive results. Aerobic growth and concomitant carbohydrate consumption of this strain were almost the same as those of the wild type (21, 25), but growth and final cell density were relatively low in the case with glucose, and exponential growth was somewhat delayed with fructose (Fig. S5). In contrast, oxygen-deprived glucose consumption and organic acid production were significantly decreased when either glucose or fructose was used as a carbon source for cell growth (Fig. S6 and Table S1). The effect of sugR deletion on glucose consumption and productivity has remained unclear also in previous studies. Engels et al. reported that the sugR deletion mutant of C. glutamicum formed 3-fold more l-lactate than the wild type under oxygen deprivation conditions although the productivities were very low (about 15 mM and 5 mM after 3 h, respectively) (24). Blombach et al. disrupted sugR to increase the l-valine productivity of pyruvate dehydrogenase complex-deficient (ΔPDHC) C. glutamicum. In the presence of acetate in addition to glucose, the ΔPDHC strain consumed mainly acetate for cell growth but less glucose for l-valine production. By disrupting sugR, the strain's glucose consumption was increased even in the presence of acetate, but l-valine production was not increased (52, 53). What caused these discrepant results remains unclear, and further research will be needed in order to elucidate the still unknown mechanisms of SugR. A sugR deletion affected transcription of many genes other than those involved in the glycolytic pathway (21, 22), which may have some effects on glucose consumption and organic acid production indirectly. On the other hand, accumulated F1P might also influence other regulators besides SugR. Given this assumption, the enhanced glucose consumption and organic acid production of the ΔpfkB strain grown with fructose presumably resulted from successful circumvention of repression by both SugR and other unknown negative factors.

Conclusion.

In this study, we identified the PTS and glycolytic enzymes as rate-limiting factors in glycolytic flux based on the analysis of the ΔpfkB strain grown with fructose and demonstrated enhanced glucose consumption and organic acid production by overexpressing the genes under oxygen deprivation. Notably, in contrast to the ΔpfkB mutant that requires expensive fructose as a carbon source for cell growth to induce the enhanced productivity, the engineered strains that can use inexpensive glucose are cost-effective in commercial production. Although it has already been reported that overexpression of the glycolytic genes increases glucose consumption and productivity (11, 15, 49), we further demonstrate that overexpression of the PTS genes is also effective in C. glutamicum. These findings will be helpful in understanding the regulatory mechanisms of glycolytic flux and can be applicable to the improvement of microbial production of various kinds of useful chemicals.

MATERIALS AND METHODS

Bacterial strains, media, and cultivation conditions.

Bacterial strains used in this study are listed in Table 1. Escherichia coli strains were grown at 37°C in LB medium (54). C. glutamicum R (JCM 18229) (55) and its recombinants were grown at 33°C in A medium containing 2 g/liter yeast extract, 7 g/liter Casamino Acids, 2 g/liter urea, 7 g/liter (NH₄)₂SO₄, 0.5 g/liter KH₂PO₄, 0.5 g/liter K₂HPO₄, 0.5 g/liter MgSO₄·7H₂O, 6 mg/liter FeSO₄·7H₂O, 4.2 mg/liter MnSO₄·H₂O, 0.2 mg/liter biotin, 0.2 mg/liter thiamine-HCl (6), and 4% glucose as a carbon source unless otherwise stated. If necessary, 50 μg/ml kanamycin or 50 μg/ml chloramphenicol for E. coli and 50 μg/ml kanamycin or 5 μg/ml chloramphenicol for C. glutamicum were added to the culture medium. Aerobic growth of C. glutamicum strains on glucose or fructose was evaluated on 100 ml of A medium with 200 mM carbohydrate in a 500-ml flask at 33°C on a rotary shaker at 180 rpm.

Construction of plasmids and strains.

General DNA manipulations were performed as previously described (54). Primers, plasmids, and strains used and constructed in this study are listed in Tables 1 and 3. Chromosomal gene deletion and induction were achieved via a markerless system using suicide vector pCRA725 carrying the sacB gene as previously described (14, 49). The induction of gapA (49), pgk, tpi, pgi, and pfkA (50) and the deletion of ppc (14) were performed according to previous reports. In order to truncate the middle region of the fruR and pfkB1 genes, each 5′ and 3′ region was amplified by PCR using primers 1 to 8 using chromosomal DNA of C. glutamicum as a template. The amplified 5′ and 3′ regions of each gene were cloned into pCRA725, and the resultant plasmids bearing each gene deficient in the middle region were used for gene disruption via homologous recombination. For the construction of recombinant strains overexpressing enzymes in central carbon metabolism, each gene under the control of the tac promoter was integrated into strain-specific islands (SSIs) of C. glutamicum R (56). The ptsI, ptsH, ptsG, and ldhA genes, as well as the regions SSI3-8, SSI9-2, SSI4-3, and SSI4-7, into which the respective genes were integrated, were amplified by PCR using primers 9 to 24. The amplified genes were inserted into pCRD904 (49) and integrated with the tac promoter and rrnB terminator, while the amplified SSI regions were cloned into pCRA725. Each gene together with the promoter and terminator was excised from the respective pCRD904 derivative and inserted into the middle of the respective SSI region in pCRA725. The resulting plasmids were used for chromosomal induction of each gene.

TABLE 3.

Primers used in this study

Primer no.	Target	Sequence (5′–3′)a	Restriction enzyme
1b	fruR	GCGTCTAGAGCTGGGGAGAGGTCATCTGC	XbaI
2b		TGGCTTGCTTGATGGGAAGCTTAACCTCGCTGATGTTGGAG	HindIII
3c		ACATCAGCGAGGTTAAGCTTCCCATCAAGCAAGCCATGATC	HindIII
4c		GCGGCATGCATTACGCTATTCATGCTGATTCTTTCAATC	SphI
5b	pfkB1	CTCTGTCGACACACCAACTCCAACATCAGC	SalI
6b		GGGATAGCACCTTCGGAGACTTGATGCCTTTACCACCTGC
7c		GTCTCCGAAGGTGCTATCCCATTC
8c		CTCTGTCGACTGGTTGCTGTAGCAACCTGC	SalI
9	ptsI	CTCTCATATGGCTACTGTGGCTGATGTG	NdeI
10		CTCTCATATGTTAGACTGCTGCGTCGATCAC	NdeI
11	SSI3-8	CTCTCCTGCAGGCTCTGGATACTGACGGATG	Sse8387I
12		CTCTCCTGCAGGCTACGACAAGCTATCGGTG	Sse8387I
13	ptsH	CTCTCATATGGCTTCCAAGACTGTAACC	NdeI
14		CTCTCATATGTTTACTCAGCGTCGAGGTC	NdeI
15	SSI9-2	CTCTGTCGACAATCGCTTGACACAATCGGC	SalI
16		CTCTGTCGACTCGTTGGTTGTAATCGACCG	SalI
17	ptsG	CTCTCATATGGCGTCCAAACTGACGAC	NdeI
18		CTCTCATATGTTACTCGTTCTTGCCGTTGAC	NdeI
19	SSI4-3	CTCTGTCGACATTCCTGCGTCTGGTGGTCT	SalI
20		CTCTGTCGACCCGACCAATGATGTAACTGC	SalI
21	ldhA	CTCTCATATGAAAGAAACCGTCGGCAATAAG	NdeI
22		CTCTCATATGTCAGAAGAACTGCTTCTGAATTTC	NdeI
23	SSI4-7	CTCTGTCGACCTGTGGTGACTTTATTGTCTAGG	SalI
24		CTCTGTCGACGCCAGCTTCTGTAAGTAACTC	SalI

^aThe restriction sites used in the cloning procedures are underlined.

^bPrimer used for amplification of the 5′ region of the gene.

^cPrimer used for amplification of the 3′ region of the gene.

Evaluation of glucose consumption and organic acid production under oxygen deprivation.

The conditions of aerobic cultivation for the preparation of cells and the oxygen-deprived reactions using the prepared cells were described in a previous report (6). Briefly, C. glutamicum strains were aerobically cultivated at 33°C for 14 h on 500 ml of A medium with 4% glucose or fructose in a 2-liter flask at 180 rpm to an optical density at 610 nm (OD₆₁₀) of 4.0 to 8.0 from 0.1. Harvested cells were washed with BT-U medium, whose composition was identical to that of the A medium except for the absence of yeast extract, Casamino Acids, and urea. The washed cells were resuspended to a concentration of 20 g (cell dry weight)/liter in 50 ml of BT-U medium containing 400 mM glucose. These cell suspensions were incubated at 33°C with no aeration but with gentle agitation. The pH was maintained at 7.5 using a pH controller (DJ-1023P; Biott, Tokyo, Japan) by supplementation with 5.0 M NH₄OH. In the case of a longer experiment, glucose was replenished before its depletion. Due to the addition of the NH₄OH solution and glucose, the reaction volume increased during organic acid production, and glucose and organic acid concentrations were accordingly calculated to correct the data for the increased volume and consequent dilution.

Analytical procedures.

The concentrations of glucose and fructose in samples of aerobic cultivation were measured by an ultraperformance liquid chromatography system (UPLC) (Acquity H-class; Waters Corporation, Milford, MA) equipped with a BEH (ethylene bridged hybrid) amide column (Waters Corporation) and refractive index (RI) detector. The column temperature was maintained at 32°C, and the mobile phase, consisting of 0.1% NH₃, 40% acetone, and 40% acetonitrile, was run at 0.5 ml/min. An enzyme electrode glucose sensor (BF-4; Oji Scientific Instruments, Hyogo, Japan) was used for quantification of glucose in samples of oxygen-deprived reactions. Organic acid concentrations were determined by a high-performance liquid chromatography (HPLC) system (Prominence 20A; Shimadzu Corporation, Kyoto, Japan) equipped with TSKgel OApak-A and OApack-P columns (Tosoh Corporation, Tokyo, Japan) and a photodiode array (PDA) detector. The column temperature was maintained at 40°C, and 0.75 mM H₂SO₄ was run at 1.0 ml/min as a mobile phase.

Metabolome analysis was performed by a liquid chromatography-tandem mass spectrometry (LC-MS/MS) system consisting of an HPLC column (Prominence 20A) and a linear ion trap mass spectrometer (4000 Q TRAP; Applied Biosystems/MDS Sciex) (57) after extraction of intracellular metabolites from C. glutamicum cells by means of cold ethanol and chloroform according to previously described protocols (12). A factor of 1.8 ml/g (cell dry weight) was assumed as the cell volume for the calculation of intracellular concentrations. We conducted metabolome analysis of the cells under two conditions: aerobic growth and oxygen-deprived organic acid production. In the former case, the metabolites were extracted from the cells during aerobic cultivation on A medium with glucose or fructose in the late exponential phase (at 6 h with glucose and at 12 h with fructose). In the latter case, the cells used for the oxygen-deprived reaction were first prepared by aerobic cultivation on A medium with glucose or fructose; the metabolites were then extracted from the cells 2 h after they were exposed to oxygen deprivation on BT-U medium with glucose.

A DNA microarray assay for transcriptome analysis was conducted on the Agilent eArray platform (Agilent Technologies, Palo Alto, CA) as described previously (31, 58). RNA was extracted from the cells during aerobic cultivation on A medium with glucose or fructose in the early exponential phase (at 4 h) and in late exponential/stationary phases (at 12 h).

Statistical analysis was performed by using Welch's t test for appropriate comparison between the data measured. Significance was assumed at a P value of <0.05.

Accession number(s).

DNA microarray data were deposited in NCBI's Gene Expression Omnibus (GEO) under accession number GSE83383.