Engineering of a Glycerol Utilization Pathway for Amino Acid Production by Corynebacterium glutamicum

Abstract

The amino acid-producing organism Corynebacterium glutamicum cannot utilize glycerol, a stoichiometric by-product of biodiesel production. By heterologous expression of Escherichia coli glycerol utilization genes, C. glutamicum was engineered to grow on glycerol. While expression of the E. coli genes for glycerol kinase (glpK) and glycerol 3-phosphate dehydrogenase (glpD) was sufficient for growth on glycerol as the sole carbon and energy source, additional expression of the aquaglyceroporin gene glpF from E. coli increased growth rate and biomass formation. Glutamate production from glycerol was enabled by plasmid-borne expression of E. coli glpF, glpK, and glpD in C. glutamicum wild type. In addition, a lysine-producing C. glutamicum strain expressing E. coli glpF, glpK, and glpD was able to produce lysine from glycerol as the sole carbon substrate as well as from glycerol-glucose mixtures.

Although biofuels such as biodiesel and bioethanol are thought to represent a secure, renewable, and environmentally safe alternative to fossil fuels, their economic viability is a major concern. As biodiesel production by transesterification of plant seed oil with methanol yields glycerol as the main by-product (10% by weight), crude glycerol may be treated as a waste product with a disposal cost attributed to it. Therefore, glycerol-rich streams generated in large amounts by the biodiesel industry present an opportunity to establish biorefineries. Existing bulk product fermentations such as fermentative amino acid production may be suitable for adaptation to such biorefinery approaches (37).

Corynebacterium glutamicum is a nonpathogenic, gram-positive soil bacterium (13) that is used for the biotechnological production of >1,500,000 tons of l-glutamate per year, >750,000 tons of l-lysine per year, and several other amino acids (48, 56). Current amino acid production processes with C. glutamicum focus on media containing glucose, fructose, or sucrose, with glucose being the preferred carbon source of this bacterium (3). C. glutamicum can grow aerobically on a variety of sugars, such as glucose, fructose, sucrose, ribose, or maltose, on the alcohols myo-inositol and ethanol, and on organic acids, such as acetate, propionate, pyruvate, l-lactate, citrate, and l-glutamate as the sole carbon and energy sources (12, 15, 19, 28, 31, 38). Metabolic engineering has been used to broaden the carbon substrate spectrum of C. glutamicum, and this bacterium has been engineered to grow on starch (47, 51), on the hemicellulosic pentoses xylose and arabinose (25, 26), on cellobiose (30), and on the whey sugars lactose and galactose (4, 7). Although glycerol has been observed in supernatants of cultivations of several C. glutamicum strains (28), to the best of our knowledge, amino acid production by C. glutamicum from glycerol has not yet been reported.

Glycerol utilization has been well studied in Escherichia coli (34, 36). Glycerol transport is facilitated by aquaglyceroporin (GlpF), and under aerobic conditions glycerol is then phosphorylated by glycerokinase (GlpK) to yield glycerol 3-phosphate, which is oxidized by glycerol 3-phosphate dehydrogenase (GlpD) to yield the glycolytic intermediate dihydroxyacetone 3-phosphate (34, 36). In nearly all Mycobacterium species, which belong to the suborder Corynebacterineae, glycerol is regarded as the preferred carbon source, as among the tested carbon sources glycerol was most effective in stimulating oxygen consumption (34). As exceptions, Mycobacterium bovis, M. microti, and M. africanum cannot grow with glycerol unless pyruvate is added because of a single nucleotide polymorphism in pykA, the gene encoding pyruvate kinase (27). This mutation was reverted during the selection of M. bovis BCG, the glycerol metabolism of which in Roisin's minimal medium chemostat cultures has recently been characterized (5). M. smegmatis, but not M. tuberculosis, also possesses a gene encoding a putative aquaglyceroporin for facilitated diffusion of glycerol (52). Glycerol utilization by Corynebacterium species is not well studied. While Corynebacterium diphtheriae possesses a putative operon comprising glpK (DIP2235), glpF (DIP2236), and glpD (DIP2237), the genome of C. glutamicum lacks this operon (24). Although two genes of the genome of C. glutamicum ATCC 13032 (cg3198 and cg1853), which are separated by about 1.3 Mb, have been annotated as glpK (glycerol kinase gene) and glpD (glycerol 3-phosphate dehydrogenase gene), respectively, a glpF homolog is missing (24), and C. glutamicum ATCC 13032 is not able to grow with glycerol as the sole carbon source. Here, we describe the engineering of C. glutamicum for growth on glycerol as the sole or combined carbon and energy source and its application to amino acid production.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

C. glutamicum strains and plasmids used in this work are listed in Table 1. For analyzing lysine or glutamate production, a brain heart infusion medium (Difco) preculture was inoculated from a fresh LB plate and cultivated overnight. After cells were washed in CGXII medium (14) without a carbon source, the main culture with CGXII medium was inoculated to an optical density at 600 nm (OD₆₀₀) of 1. For growth experiments and glutamate production, C. glutamicum wild type (WT) was used, while for lysine production, C. glutamicum DM1730 was used (18). To trigger glutamate production, 500 μg/ml ethambutol was added to the minimal medium (45). After 72 h in the case of lysine production or after 27 h in the case of glutamate production, samples were withdrawn from the cultures for the determination of the amino acid and sugar concentrations in the medium. C. glutamicum cultivations were performed in baffled 500-ml Erlenmeyer flasks with 60 ml medium at 30°C and 120 rpm. E. coli MG1655 and DH5α were cultivated in LB medium or on LB agar plates at 37°C. E. coli DH5α was used as the host for cloning. When appropriate, kanamycin was used at a concentration of 25 to 50 μg/ml and isopropyl β-d-1-thiogalactopyranoside (IPTG) at concentrations up to 1 mM.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Source and/or referencea
E. coli strains
DH5α	F−thi-1 endA1 hsdR17(r− m−) supE44 ΔlacU169 (φ80lacZΔM15) recA1 gyrA96 relA1	Bethesda Research Laboratories; 21
MG1655	K-12 WT, CGSC6300	CGSC
C. glutamicum strains
WT	WT strain ATCC 13032, auxotrophic for biotin	ATCC
DM1730	lysCP458ShomV59ApycT311I Δpck	18
Plasmids
pVWEx1	Kanr P lacIq	42
pVWEx1-glpF	Kanr; pVWEx1 carrying glpF from E. coli MG1655	This study
pVWEx1-glpK	Kanr; pVWEx1 carrying glpK from E. coli MG1655	This study
pVWEx1-glpD	Kanr; pVWEx1 carrying glpD from E. coli MG1655	This study
pVWEx1-glpFK	Kanr; pVWEx1 carrying glpF and glpK from E. coli MG1655	This study
pVWEx1-glpFD	Kanr; pVWEx1 carrying glpF and glpD from E. coli MG1655	This study
pVWEx1-glpFKD	Kanr; pVWEx1 carrying glpF, glpK and glpD from E. coli MG1655	This study

^aCGSC, E. coli Genetic Stock Center; ATCC, American Type Culture Collection.

Cloning of expression vectors.

Chromosomal DNA from E. coli MG1655 was isolated as described previously (16). Plasmid DNA was isolated with a QIAprep spin miniprep kit (Qiagen, Hilden, Germany). Transformation of E. coli was performed using the rubidium chloride method (21), while C. glutamicum was transformed by electroporation as described previously (14). The genes glpF, glpK, and glpD were amplified using genomic DNA from E. coli MG1655 and the following primer pairs (indicated restriction sites are in bold, and artificial ribosome binding sites are in italics): glpF-for (GCGTCGACAGGGAGATATAGATGAGTCAAAC; SalI) and glpF-rev (TCTAGATTACAGCGAAGCTTTTTGTTCTGAAGG; XbaI), glpK-for (GGGACGTCGACAAGGAGATATAGATGACTGAAAAAAAATATATC; SalI) and glpK-rev (TCTAGATTATTCGTCGTGTTCTTCCCACGCC; XbaI), and glpD-for (TCTAGAAAGGAGATATAGATGGAAACCAAAGATCTG; XbaI) and glpD-rev (GTTAATTCTAGATTACGACGCCAGCGATAA; XbaI). The PCR products were cloned into pVWEx1 (42) using the restriction sites introduced by the primers to yield pVWEx1-glpF, pVWEx1-glpK, and pVWEx1-glpD. To obtain pVWEx1-glpFK, both genes of the E. coli glpFKX operon were amplified using the primers glpF-for and glpK-rev and cloned into pVWEx1. The 1.5-kb XbaI fragment of pVWEx1-glpD was cloned into the XbaI sites of pVWEx1-glpF, pVWEx1-glpK, and pVWEx1-glpFK to yield pVWEx1-glpFD, pVWEx1-glpKD, and pVWEx1-glpFKD, respectively.

Measurement of enzyme activities.

For determination of enzyme activities, exponentially growing cells were harvested by centrifugation (4,500 × g, 5 min, 4°C), and crude extracts were prepared as described previously (50). Glycerokinase was assayed in 168 mM sodium carbonate (pH 9.5), 20 mM MgCl₂, 20 mM ATP, 1.5 mM NAD, 0.3 M hydrazine, and 0.1 mg/ml NAD-dependent glycerol 3-phosphate dehydrogenase (from rabbit muscle; EC 1.1.1.8; Roche) as described previously (35). The reaction was started by the addition of 10 mM glycerol at 30°C, and the increase in absorbance of NADH (ɛ₃₄₀ = 6.3 mM⁻¹ cm⁻¹) was determined. Because E. coli glycerol 3-phosphate dehydrogenase (GlpD) is FAD dependent (EC 1.1.99.5), glycerol 3-phosphate dehydrogenase activity was measured as described previously (35) in 84 mM Tris-HCl (pH 7.4), 30 μg/ml MTT [3(4,5-dimethylthiazolyl-1-2)2,5-diphenyl tetrazolium bromide], 60 μg/ml phenazine methosulfate, 10 μM FAD, and the reaction was started with 25 mM sn-glycerol 3-phosphate at 30°C. The rate of reduction of the tetrazolium dye MTT to formazan was monitored by the increase in absorbance (ɛ₅₇₀ = 17 mM⁻¹ cm⁻¹).

Determination of biomass and of glycerol, glucose, and amino acid concentrations.

Growth was followed by measuring the OD₆₀₀ with an Ultrospec 500-pro spectrophotometer (Amersham Biosciences). The biomass concentration was calculated from OD₆₀₀ values using an experimentally determined correlation factor of 0.25 g cells (dry weight [DW]) liter⁻¹ for an OD₆₀₀ of 1 (55). The carbon content of cells grown in CGXII glucose medium was previously determined to be 40% of the cell DW (33). During cultivation, samples were collected to determine glycerol, glucose, glutamate, and lysine concentrations. After centrifugation of the sample (13,000 × g, 5 min), d-glucose, glycerol, dihydroxyacetone, and l-lactate were quantified by high-pressure liquid chromatography on an instrument equipped with a refractive index detector (RID G1362A, 1200 series; Agilent Technologies) and a diode array detector (G1315B, 1200 series; Agilent Technologies), and separation was carried out with an organic acid resin column (300 by 8 mm, 10-μm particle size, 25-Å pore diameter; CS-Chromatographie, Langerwehe, Germany) with 5 mM H₂SO₄ at a flow rate of 1 ml/min. Alternatively, d-glucose was quantified enzymatically with a d-glucose kit as described by the manufacturer (R-Biopharm, Darmstadt, Germany), and glycerol was quantified enzymatically by conversion to glycerol 3-phosphate using glycerol kinase and by coupling with the pyruvate kinase and lactate dehydrogenase reactions to regenerate ATP. The concentration of glycerol was equivalent to the NADH concentration formed in the assay, which contained 100 mM glycylglycine (pH 7.4), 10 mM MgSO₄, 1.25 mM ATP, 0.3 mM NADH, 1.25 mM phosphoenolpyruvate, 2.25 U/ml pyruvate kinase-lactate dehydrogenase (glycerol-free solution; Roche) and 0.4 U/ml glycerol kinase (Roche). Sample concentrations were determined by comparison with external standards. Quantitative determination of l-lysine and l-glutamate in supernatants was carried out by reversed-phase high-pressure liquid chromatography as described previously (18).

Determination of intracellular glycerol 3-phosphate concentrations.

CgXII media containing 55 mM glucose and 110 mM glycerol were inoculated to an OD₆₀₀ of 1. After growth to an OD₆₀₀ of 4, gene expression was induced by addition of IPTG to final concentrations of 0.5 and 0.1 mM, respectively. Four hours after induction, cell samples were harvested and treated as described in reference 12. Glycerol 3-phosphate concentrations were determined spectrophotometrically in 200 mM triethanolamine (pH 7) containing 5 mM NAD and 0.1 mg/ml glycerol 3-phosphate dehydrogenase (from rabbit muscle; EC 1.1.1.8; Roche). The increase in levels of NADH was measured at 340 nm. Sample concentrations of glycerol 3-phosphate were calculated from external glycerol 3-phosphate standards (0.025 to 0.5 mM), treated similarly to the cell samples. Intracellular glycerol 3-phosphate concentrations were calculated based on the cell volume of C. glutamicum of 1.8 μl (mg [DW])⁻¹ (22).

RESULTS

Engineering of C. glutamicum for glycerol utilization.

C. glutamicum WT was unable to grow on minimal medium containing glycerol as the sole source of carbon and energy (Fig. 1A). Even after prolonged incubation, no spontaneous mutants that had acquired the ability to grow on glycerol were obtained. As C. glutamicum is able to coutilize many carbon sources with glucose, growth of C. glutamicum was also tested on mixtures containing both glucose and glycerol. However, biomass formation on a mixture of glucose and glycerol was as high as that on glucose alone, indicating that glycerol could not be utilized as a cosubstrate with glucose for growth of wild-type C. glutamicum (Fig. 1B).

FIG. 1.

Growth of C. glutamicum WT(pVWEx1) (open symbols) and WT(pVWEx1-glpFKD) (closed symbols) on mineral medium containing 110 mM glycerol (A) or 110 mM glycerol and 55 mM glucose (B) as carbon sources. Growth was monitored as OD₆₀₀ (circles); the concentrations of glycerol (squares) and glucose (triangles) are indicated. The data are averages and standard deviations of three replicates.

In order to test whether heterologous expression of the E. coli genes glpF, glpK, and glpD (encoding the glycerol facilitator, glycerol kinase, and glycerol 3-phosphate dehydrogenase, respectively) is sufficient for growth of C. glutamicum on glycerol, plasmids for IPTG-inducible expression of each individual gene and of different combinations of these genes arranged as operons were constructed. The ribosomal binding sites were introduced with the primers used for PCR amplification, with the exception of the native DNA sequence upstream of glpK present in pVWEx1-glpFK. C. glutamicum WT was transformed with the empty vector control pVWEx1 or with the plasmid pVWEx1-glpF, pVWEx1-glpK, pVWEx1-glpD, pVWEx1-glpFK, pVWEx1-glpFD, pVWEx1-glpKD, or pVWEx1-glpFKD. The resulting strains were cultured on complex medium with IPTG, and crude extracts were assayed for glycerol kinase and glycerol 3-phosphate dehydrogenase activities (Table 2). Crude extracts of the negative control strain C. glutamicum WT(pVWEx1) contained very little glycerol kinase activity (0.01 U/mg), while crude extracts of all strains expressing glpK from E. coli showed high specific activities of glycerol kinase (1.84 to 4.06 U/mg). Glycerol 3-phosphate dehydrogenase activity could be detected in crude extracts of all strains expressing glpD from E. coli (0.20 to 0.77 U/mg) but was absent from the other strains (<0.01 U/mg) (Table 2).

TABLE 2.

Specific activities of glycerol kinase and glycerol 3-phosphate dehydrogenase in recombinant C. glutamicum strains and growth on glycerol as the sole carbon source

C. glutamicum strain	Sp act (μmol min−1 mg−1)a of:		Growth on glycerol
	Glycerol kinase	Glycerol 3-phosphate dehydrogenase
WT(pVWEx1)	0.01	<0.01	−
WT(pVWEx1-glpF)	0.02	<0.01	−
WT(pVWEx1-glpK)	1.84	<0.01	−
WT(pVWEx1-glpD)	0.03	0.77	−
WT(pVWEx1-glpFK)	3.74	<0.01	−
WT(pVWEx1-glpFD)	0.01	0.20	−
WT(pVWEx1-glpKD)	4.06	0.53	+
WT(pVWEx1-glpFKD)	2.21	0.41	+

^aData are the means of at least two determinations for at least two independent cultures with errors of <20%. The cultures contained 1 mM IPTG.

Growth experiments with these recombinant C. glutamicum strains were performed with either 55 mM glucose or 110 mM glycerol as the sole carbon source. While all strains grew at similar rates and to similar final optical densities on glucose minimal medium (data not shown), only C. glutamicum WT(pVWEx1-glpKD) and WT(pVWEx1-glpFKD) grew on glycerol as the sole carbon source (Table 2; Fig. 1A). Thus, heterologous expression of the E. coli genes for glycerol kinase and glycerol 3-phosphate dehydrogenase was sufficient to establish glycerol utilization by C. glutamicum.

Glycerol 3-phosphate may serve as a phosphorus source for E. coli, which takes up glycerol 3-phosphate via GlpT and generates inorganic phosphate from glycerol 3-phosphate intracellularly. Because the ugpAEBC operon, encoding an ABC transporter for uptake of glycerol 3-phosphate, a source of phosphorus for C. glutamicum, is induced during the phosphate starvation response of C. glutamicum (23, 29, 54), we tested whether glycerol 3-phosphate serves as a carbon source for C. glutamicum precultured with either a sufficient or limiting phosphate supply. However, neither phosphate-starved cells nor cells pregrown with a sufficient phosphate supply could grow with 100 mM glycerol 3-phosphate as the sole carbon source even when E. coli glpD was expressed (data not shown).

Growth performance of recombinant C. glutamicum WT(pVWEx1-glpFKD) on glycerol.

C. glutamicum WT(pVWEx1-glpFKD) grew faster (doubling time, 150 min versus 185 min) and formed more biomass (4.25 versus 2.75 g [DW] per liter) than WT(pVWEx1-glpKD) on 110 mM glycerol, indicating that expression of the glycerol facilitator gene glpF is advantageous for growth of C. glutamicum on glycerol. Therefore, growth rates and biomass yields on minimal medium with different concentrations of glycerol as the sole carbon source were determined with C. glutamicum WT(pVWEx1-glpFKD). On 25 to 200 mM glycerol, C. glutamicum WT(pVWEx1-glpFKD) grew at rates of 0.27 to 0.29 h⁻¹, while growth was slower at 400 mM glycerol (0.23 h⁻¹) and was severely perturbed at 1 M glycerol (0.07 h⁻¹), indicating substrate inhibition of growth at very high glycerol concentrations (Fig. 2). Similarly, the biomass yield decreased with increasing glycerol concentrations (Fig. 2). The biomass yield on 110 mM glycerol (about 0.5 g [DW] g⁻¹) was comparable to that on 55 mM glucose, indicating that carbon in glycerol can be used as efficiently as carbon in glucose. On 110 mM glycerol, the glycerol uptake rate was 105 nmol (mg [DW])⁻¹ min⁻¹ for WT(pVWEx1-glpFKD).

FIG. 2.

Biomass yields (squares) and growth rates (triangles) of C. glutamicum WT(pVWEx1-glpFKD) for growth on minimal medium containing different concentrations of glycerol as the sole carbon source.

Growth of recombinant C. glutamicum strains on glycerol-glucose mixtures.

To test whether glycerol can be utilized as a cosubstrate with glucose, growth experiments with different recombinant C. glutamicum strains were performed on minimal medium containing 55 mM glucose and 110 mM glycerol as carbon sources. C. glutamicum WT(pVWEx1-glpFKD) grew to a biomass concentration of 8.1 g (DW) liter⁻¹ on 55 mM glucose and 110 mM glycerol, whereas C. glutamicum WT(pVWEx1) grew to a biomass concentration of only 4.6 g (DW) liter⁻¹ (Fig. 1). A transient accumulation of up to 6.1 ± 0.6 mM dihydroxyacetone and up to 7.9 ± 0.8 mM l-lactate during the exponential growth phase of WT(pVWEx1-glpFKD) on a glucose-glycerol mixture was observed (data not shown). C. glutamicum WT(pVWEx1-glpFKD) showed no diauxic shift on the glucose-glycerol mixture and coutilized both carbon sources simultaneously (Fig. 1). Typically, growth rates of about 0.35 h⁻¹, glycerol uptake rates of about 44 nmol (mg [DW])⁻¹ min⁻¹, and glucose uptake rates of about 49 nmol (mg [DW])⁻¹ min⁻¹ were observed. Thus, the uptake rates of the coutilized substrates were lower than the uptake rates during growth on a single substrate. This phenomenon is typical for substrate coutilization by C. glutamicum, e.g., of glucose and acetate (55). The reduced glucose uptake rate in the presence of a second, coutilized substrate such as acetate was shown to be due at least in part to repression of ptsG, the gene for the glucose-specific phosphoenolpyruvate-dependent phosphotransferase system (PTS) component EII^Glc, by the transcriptional regulator SugR (17).

C. glutamicum WT(pVWEx1-glpKD) also coutilized glycerol with glucose, exhibiting growth parameters similar to those obtained for C. glutamicum WT(pVWEx1-glpFKD). Thus, as for growth on glycerol as the sole carbon source, the combined overexpression of glpK and glpD was necessary and sufficient for biomass formation from glycerol present in glycerol-glucose mixtures (Fig. 1 and data not shown).

When the glycerol kinase gene glpK, but not the glycerol 3-phosphate dehydrogenase gene glpD, was overexpressed, growth on glycerol-glucose mixtures was perturbed (Fig. 3), while growth on glucose was not affected (data not shown). Growth of the empty vector control and of the strain expressing only the glycerol facilitator gene glpF was not perturbed on glucose-glycerol medium (Fig. 3). When different IPTG concentrations were added, it was observed that the extent of growth inhibition of C. glutamicum WT(pVWEx1-glpK) correlated with the specific glycerol kinase activities measured in crude extracts (Fig. 3), suggesting that the growth inhibition might be caused by intracellular accumulation of glycerol 3-phosphate due to increased glycerol kinase levels. When gene expression of exponentially growing cells was induced for 4 h with 0.1 or 0.5 mM IPTG, C. glutamicum WT(pVWEx1-glpFK) accumulated glycerol 3-phosphate to intracellular concentrations of 550 and 565 mM, respectively (Fig. 4). The strains WT(pVWEx1), WT(pVWEx1-glpF), and WT(pVWEx1-glpFKD) accumulated a maximum of 51 mM intracellular glycerol 3-phosphate. Thus, balanced overexpression of glpK and glpD is required to prevent intracellular accumulation of glycerol 3-phosphate and to ensure efficient coutilization of glycerol and glucose.

FIG. 3.

Growth rates and specific glycerol kinase activities of various recombinant C. glutamicum strains during growth on glycerol-glucose mixtures with different IPTG concentrations. Growth rates of C. glutamicum WT(pVWEx1) (squares), WT(pVWEx1-glpF) (triangles), and WT(pVWEx1-glpFK) (solid circles) and specific glycerol kinase activities of WT(pVWEx1-glpFK) (open circles) are plotted against the IPTG concentration used for inducible gene expression from plasmid pVWEx1. Crude extracts of WT(pVWEx1) and WT(pVWEx1-glpF) contained specific glycerol kinase activities of 0.01 to 0.02 U mg⁻¹.

FIG. 4.

Intracellular glycerol 3-phosphate concentrations of various recombinant C. glutamicum strains during growth on 110 mM glycerol and 55 mM glucose. Concentrations were measured 4 h after induction with 0.1 (white bars) or 0.5 mM (black bars) IPTG at an OD₆₀₀ of 4 and are given as arithmetic means and absolute errors of two independent cultivations.

Glutamate production by recombinant C. glutamicum on glycerol.

In order to test whether glycerol supports glutamate production by C. glutamicum WT(pVWEx1-glpFKD), this strain was cultured in minimal medium with 20 g liter⁻¹ glycerol as the sole carbon source, and glutamate production was triggered by the addition of 500 μg/ml ethambutol (45). After 33 h C. glutamicum WT(pVWEx1-glpFKD) utilized glycerol completely and accumulated up to 15 mM glutamate in the culture medium. Thus, expression of E. coli glpF, glpK, and glpD in C. glutamicum WT enabled glutamate production from glycerol as the sole carbon source with a product yield of 0.11 g glutamate per g glycerol (or 0.069 mol/mol), which is in the same range as the glutamate yield on glucose (0.20 g per g) for ethambutol-triggered glutamate production by C. glutamicum WT (45).

Lysine production by recombinant C. glutamicum on glycerol.

The plasmids pVWEx1 and pVWEx1-glpFKD were introduced into the genetically defined lysine-producing strain DM1730 (18). C. glutamicum DM1730(pVWEx1) and DM1730(pVWEx1-glpFKD) were cultured in minimal medium with 20 g liter⁻¹ glycerol or with a mixture of 20 g liter⁻¹ glycerol and 20 g liter⁻¹ glucose. While C. glutamicum DM1730(pVWEx1) could not grow on glycerol as the sole carbon source, C. glutamicum DM1730(pVWEx1-glpFKD) grew, utilized glycerol completely, and accumulated 26 mM lysine in the culture medium. The product yield of 0.19 g lysine per g glycerol (0.12 mol/mol) is comparable to the lysine yield on glucose previously determined with this strain (0.26 g lysine per g glucose) (32).

On a mixture of 20 g liter⁻¹ glycerol and 20 g liter⁻¹ glucose, C. glutamicum DM1730(pVWEx1-glpFKD) consumed both substrates completely and accumulated 37 mM lysine in the culture medium. On the other hand, C. glutamicum DM1730(pVWEx1) could utilize only glucose and accumulated only 22 mM lysine in the culture medium. On glycerol-glucose mixtures, a product yield of 0.136 g lysine per g carbon source was obtained with C. glutamicum DM1730(pVWEx1-glpFKD).

DISCUSSION

Expression of two E. coli genes, the glycerol kinase gene glpK and the glycerol 3-phosphate dehydrogenase gene glpD, was necessary and sufficient to enable growth of C. glutamicum on glycerol as the sole carbon and energy source. This indicates that although the gene products of cg3198 and cg1853 have been annotated as putative glycerol kinase and glycerol 3-phosphate dehydrogenase (24), either they are not functional equivalents of GlpK and GlpD or they are not synthesized under the chosen conditions. cg3198 was annotated to encode glycerol kinase, as its product shares 49% identical amino acids with E. coli GlpK. cg3198 lies downstream of cg3199 and cg3200, the latter of which putatively encodes 1-acyl-sn-glycerol 3-phosphate acyltransferase, an enzyme which is involved in glycerophospholipid biosynthesis and catalyzes esterification of glycerol 3-phosphate with activated fatty acids. Thus, if the low glycerol kinase activity observed in C. glutamicum crude extracts (0.01 to 0.02 μmol min⁻¹ mg⁻¹) (Table 2) is due to Cg3198, it might play a role in biosynthesis of lipids by scavenging glycerol generated from lipid turnover and/or breakdown. Nevertheless, the low glycerol kinase activity observed in C. glutamicum was clearly not sufficient for growth on glycerol, even when the glycerol 3-phosphate dehydrogenase gene glpD from E. coli was expressed (Table 2).

Enzyme assays revealed that crude extracts of C. glutamicum WT did not contain detectable glycerol 3-phosphate dehydrogenase activity (<0.01 μmol min⁻¹ mg⁻¹) (Table 2). Cg1853 shows sequence similarities over its entire length to proteins belonging to those encoded by the cluster of orthologous genes COG0578. This cluster comprises genes encoding proteins similar to E. coli GlpA, a subunit of the anaerobic glycerol 3-phosphate dehydrogenase GlpABC, and GlpD, the aerobic glycerol 3-phosphate dehydrogenase (36). The product of C. glutamicum cg1853 shares 27% and 31% identical amino acids with E. coli GlpA and GlpD, respectively. However, while the three subunits of anaerobic glycerol 3-phosphate dehydrogenase are encoded in one operon in E. coli, the genes adjacent to cg1853 are not glpB or glpC homologs, as cg1853 lies in opposite direction to the upstream serine deaminase gene sdaA (40) and downstream of the histidyl-tRNA synthetase gene cg1854. The inability of C. glutamicum to grow on glycerol even when the glycerol kinase gene from E. coli was expressed and the absence of glycerol 3-phosphate dehydrogenase activity indicate that cg1853 either does not encode a functional glycerol 3-phosphate dehydrogenase or is not expressed under the tested conditions.

Expression of glycerol utilization genes is generally regulated in bacteria. In Listeria, efficient utilization of glycerol and expression of glycerol utilization genes depend on the alternative sigma factor σ^B (1). The Bacillus subtilis glycerol utilization genes are induced by glycerol 3-phosphate, which involves the glycerol 3-phosphate-activated antiterminator GlpP, and they are repressed by rapidly metabolizable sugars through CcpA (9, 11). In addition, Crh and the similar phosphoenolpyruvate:sugar phosphotransferase protein Hpr phosphorylate and activate glycerol kinase (9, 11). While, e.g., Lactobacillus can use glycerol only aerobically, producing lactate, acetate, and diacetyl (2), E. coli can utilize glycerol aerobically and anaerobically (34, 36, 39, 49). In E. coli, the glycerol 3-phosphate-dependent regulator GlpR represses transcription of the three glycerol utilization operons, which are also controlled by cyclic AMP-cyclic AMP receptor, ArcA-ArcB, and FNR (36). Glycerol kinase from E. coli is inhibited by the unphosphorylated form of the glucose-specific PTS component EIIA^Glc, the form prevailing in the presence of glucose, and this control is the basis of inducer inclusion when E. coli grows on glucose-glycerol mixtures (10, 36). The C. glutamicum glucose-specific PTS EIIBCA^Glc appears not to inhibit E. coli glycerol kinase, as indicated by the efficient coutilization of glycerol and glucose by the C. glutamicum strains expressing E. coli glpK and glpD described here.

The C. glutamicum strain WT(pVWEX1-glpK), which expressed only the glycerol kinase gene glpK, showed perturbed growth on glucose in the presence of glycerol (Fig. 3) and accumulated glycerol 3-phosphate to a very high intracellular concentration (Fig. 4). Strains WT(pVWEx1) and WT(pVWEx1-glpF), which did not express glpK and therefore could not efficiently phosphorylate glycerol to glycerol 3-phosphate, did not show a growth defect and did not accumulate high intracellular glycerol 3-phosphate concentrations (Fig. 3 and 4). Moreover, strain WT(pVWEx1-glpFKD) showed good growth on glucose-glycerol mixtures (Fig. 1B), as it was able to catabolize the intermediate glycerol 3-phosphate to dihydroxyacetone 3-phosphate and further in glycolysis. Accordingly, glycerol 3-phosphate did not accumulate to high intracellular concentrations in WT(pVWEx1-glpFKD) (Fig. 4). Mutants of B. subtilis accumulating glycerol 3-phosphate intracellularly show perturbed growth (34). For E. coli, it is known that intracellular glycerol 3-phosphate inhibits growth (8) and 1,3-propanediol production (57). The molecular cause is not known, but it is assumed that inhibition of phospholipid metabolism rather than ATP depletion is responsible for growth inhibition (8, 34, 44).

C. glutamicum has been engineered to utilize pentoses and hexoses derived from starch, whey, and lignocellulosic biomass (4, 7, 25, 26, 30, 47, 51) as carbon sources and phytic acid as a source of phosphorus (53). However, the use of the engineered strains has rarely been tested with substrates of technical quality. For raw glycerol qualities, it has been observed that production of 1,3-propanediol, butanol, and lipids from glycerol by Clostridium species or by Yarrowia lipolytica (6, 20, 41, 43, 46) is influenced by the glycerol purity, and the resistance of C. glutamicum to impurities present in raw glycerol remains to be studied. The C. glutamicum strains for amino acid production from glycerol described here might prove valuable for efficient use of glycerol, which arises in large quantities in the biodiesel process as a major by-product of plant seed oil transesterification with methanol. This study showed that lysine production from glycerol is possible in principle with recombinant C. glutamicum. The strains described here need to be improved further, e.g., to be able to cope with high glycerol concentrations, as we observed growth perturbation at glycerol concentrations exceeding 200 mM (Fig. 2). As glycerol feeds into glycolysis downstream of the pentose phosphate pathway, which is known to provide NADPH for lysine biosynthesis, metabolic engineering strategies to optimize regeneration of the cofactor NADPH need to be pursued. Currently, the product yields obtained with the model producer strain are too low to be commercially viable. However, the model producer strain showed lysine yields on glycerol that are in the same range as lysine yields on glucose obtained with this strain. The ability to grow on glycerol can easily be transferred to many more advanced lysine producer strains via the constructed vector pVWEx1-glpFKD. Amino acid production from glycerol has the potential to be integrated into a biorefinery, because glutamate production and lysine production by C. glutamicum are proven bulk production processes.