Characterization of 3-phosphoglycerate kinase from Corynebacterium glutamicum and its impact on amino acid production

Background Corynebacterium glutamicum cg1790/pgk encodes an enzyme active as a 3-phosphoglycerate kinase (PGK) (EC 2.7.2.3) catalyzing phosphoryl transfer from 1,3-biphosphoglycerate (bPG) to ADP to yield 3-phosphoglycerate (3-PG) and ATP in substrate chain phosphorylation. Results C. glutamicum 3-phosphoglycerate kinase was purified to homogeneity from the soluble fraction of recombinant E. coli. PGKHis was found to be active as a homodimer with molecular weight of 104 kDa. The enzyme preferred conditions of pH 7.0 to 7.4 and required Mg2+ for its activity. PGKHis is thermo labile and it has shown maximal activity at 50–65°C. The maximal activity of PGKHis was estimated to be 220 and 150 U mg-1 with KM values of 0.26 and 0.11 mM for 3-phosphoglycerate and ATP, respectively. A 3-phosphoglycerate kinase negative C. glutamicum strain ∆pgk was constructed and shown to lack the ability to grow under glycolytic or gluconeogenic conditions unless PGK was expressed from a plasmid to restore growth. When pgk was overexpressed in L-arginine and L-ornithine production strains the production increased by 8% and by 17.5%, respectively. Conclusion Unlike many bacterial PGKs, C. glutamicum PGK is active as a homodimer. PGK is essential for growth of C. glutamicum with carbon sources requiring glycolysis and gluconeogenesis. Competitive inhibition by ADP reveals the critical role of PGK in gluconeogenesis by energy charge. Pgk overexpression improved the productivity in L-arginine and L-ornithine production strains.


Background
Central carbon metabolism uses a complex series of enzymatic steps to convert sugars into metabolic precursors [1]. When terminal electron acceptors are not available, glycolysis supplies all of the ATP molecules required for cellular activity through substrate chain phosphorylation and the glycolytic intermediates are direct precursors of many cellular building blocks. In glycolysis, substrate chain phosphorylation has a net yield of two moles of ATP per mole of glucose that are generated in the reactions catalyzed by 3-phosphoglycerate kinase (PGK) and pyruvate kinase. PGK catalyzes the phosphoryl transfer from 1,3-biphosphoglycerate (bPG) to ADP yielding 3phosphoglycerate (3-PG) and ATP [2].
C. glutamicum is a Gram-positive soil bacterium belonging to the order Corynebateriales within the class of Actinobacteria [3]. Since the discovery of this organism, it has been used for the industrial production of Lamino acids, and strains have been developed for the production of D-amino acids, organic acids, diamines or biofuels from different carbon sources [4][5][6][7][8]. C. glutamicum was metabolically engineered for the production of amino acids from alternative substrates, such as starch, the hemicellulose components xylose and arabinose, cellobiose, the whey components galactose and lactose, and glycerol [9][10][11]. The phosphotransferase system substrates such as glucose, fructose and sucrose are used for production of amino acids on an industrial scale in the form of starch hydrolysate or molasses [12].
C. glutamicum has been extensively studied and the central carbon metabolism genes in C. glutamicum are under the control of a transcriptional regulatory network composed of several global regulators and various transcriptional regulators which have been characterized [13,14]. The genes of glycolytic enzymes are not clustered except for the gapA-pgktpi-ppc cluster encoding glyceraldehyde-3-phosphate dehydrogenase, phosphoglycerate kinase, triose phosphate isomerase and phosphoenolpyruvate carboxykinase, respectively [14]. The transcription of the gapA-pgk-tpi-ppc cluster involves mono-, di-and tricistronic mRNAs [15]. Expression of the gapA-pgk-tpi operon is coordinately regulated by SugR, RamA, and GlxR [16][17][18].
However, despite of the importance as an energy generating metabolic reaction in glycolysis in C. glutamicum and other bacteria, limited information is available on PGK. In this work, we aimed to identify and enzymatically characterize the PGK from C. glutamicum and to analyze its physiological role.

Phylogenetic analysis of PGK from C. glutamicum
Databank searches with the amino acid sequences for PGK of C. glutamicum revealed similarities to biochemically characterized PGKs of different multimerization states such as monomeric PGK from E. coli, dimeric PGKs from Methanothermus fervidus [19] and Pyrococcus woesi, and tetrameric PGKs from Sulfolobus solfataricus and Trypanosoma brucei, the latter shown to be a dimer of dimers [20][21][22][23]. The N-terminal sequences of all PGKs showed high amino acid sequence similarity and a proline residue mostly conserved in all proteins in the hinge region. Based on its amino acid sequence, PGK from C. glutamicum is closely related to the monomeric enzyme from E. coli (Data not shown).
Purification of PGK of C. glutamicum from recombinant E. coli expressing cg1790/pgk For purification, PGK from C. glutamicum was overproduced as an N-terminally His-tagged protein in recombinant E. coli. Metal chelate chromatography allowed to purify the enzyme to homogeneity as determined by SDS-PAGE ( Figure 1) revealing a molecular mass of 47 kDa for the monomer. To determine the multimeric state of C. glutamicum PGK His , gel filtration chromatography was performed revealing a single peak with a molecular mass of about 104 kDa and only this fraction showed activity as PGK (data not shown). Thus, unlike the monomeric PGK from E. coli, PGK from C. glutamicum appears to be active as a homodimer as also reported for several archaeal PGKs [22].

Characterization of PGK from C. glutamicum
The pH and metal ion concentrations were varied to determine the optimal buffer conditions for activity of PGK His . Within a pH range of 4.5 to 9.8, PGK activity was optimal at pH 7.0-7.4 in 100 mM TEA-Cl buffer with only 10% of PGK activity remaining at pH 4.5 and 33% at pH 8.9. PGK required bivalent cations with Mg 2+ being the most effective. In the presence of 1 mM Ni 2+ , Co 2+ , Mn 2+ , Cd 2+ , Ca 2+ or Zn 2+ , PGK His activity was reduced to between 80% and 40% in comparison to the presence of Mg 2+ only. To determine its thermal stability, PGK His was incubated at temperatures ranging from 25 to 65°C prior to activity measurements at 30°C. A temperature optimum between 50-65°C was observed. Protein precipitation was observed above 65°C due to the instability of rabbit muscle GAPDH used as indicator enzyme in the assay.

Kinetic parameters of PGK from C. glutamicum
The kinetic parameters of PGK His for the substrates 3phosphoglycerate and ATP were determined at 30°C, the optimal growth temperature of the bacterium. The activity of PGK His with 3-phosphoglycerate and ATP in the gluconeogenetic direction followed Michaelis-Menten kinetics (data not shown). The K M values of C. glutamicum PGK for 3-phosphoglycerate and ATP, respectively, were determined to be 0.26 mM and 0.11 mM, respectively. The Vmax values were 220 U mg −1 and 150 U mg −1 , respectively (Table 1). Catalytic efficiencies were about 733 s −1 mM −1 and 592 s −1 mM −1 , respectively (Table 1).
Physiological studies of pgk deletion mutants pgk was deleted to study the role of PGK in C. glutamicum. Since pgk is part of the gapA-pgk-tpi-ppc operon [15] the specific activities of GapA, Tpi and Ppc were determined in crude extracts of the deletion mutant. While GapA activity in Δpgk was comparable to wild type, Δpgk showed decreased Tpi activity (0.1 U/mg as compared to 0.4 U/mg in wild type) and Ppc activity (0.03 U/mg as compared to 0.05 U/mg in wild type) which might be explained by polar effects of pgk deletion on the downstream genes tpi and ppc. Thus, plasmid pEKEx3-pgk was constructed for complementation analysis. C. glutamicum Δpgk showed no growth in CgXII minimal medium with glucose or pyruvate as sole source of carbon and energy, but grew with blends of glucose plus pyruvate ( Table 2). Complementation of C. glutamicum Δpgk led to comparable growth with either glucose or pyruvate as sole carbon source (Table 2) revealing the requirement of PGK for growth with glycolytic as well as with gluconeogenic carbon sources.
Effect of pgk overexpression on amino acid production pgk was overexpressed in the L-lysine producing strain DM1933 [7], the L-arginine production strain ARG1 [6], and the L-ornithine production strain ORN1 [6] to assay the affects of pgk overexpression on amino acid production rates. In each strain, PGK specific activities increased upon pgk overexpression to similar levels as when pgk was overexpressed in the wild type (Table 3). Amino acid production in CgXII minimal medium with 4% (w/v) glucose was performed and amino acid accumulation was followed. While pgk overexpression did not increase the L-lysine production rate, pgk overexpression accelerated L-arginine and L-ornithine production by 8% and 17.5%, respectively ( Figure 3). Glucose consumption was comparable between control and pgk overexpression strains.  Values for K M (mM), V max (U/mg), and catalytic efficiency (K cat /K M = s −1 mM −1 ) were determined for two independent protein purifications.

Discussion
Phosphoglycerate kinase from C. glutamicum was purified to homogeneity and shown to be active as homodimer. This is unusual and was hitherto only found in the thermophilic archaea Pyrococcus and Methanothermus [19,22]. Most prokaryotes and eukaryotes possess monomeric PGKs [24,25]. Tetramers have been reported in rare cases as in Sulfolobus solfataricus [26] and Trypanosoma [21]. C. glutamicum PGK His showed higher temperature stability than monomeric PGKs, e.g., 40°C for Mus musculus [27] and 24°C for S. cervisiae [28]. Possibly, dimerization of PGK from C. glutamicum contributes to the stability of the protein by favoring hydrophobic interactions via the subunit contacts and by reducing the surface area exposed to the solvent in a similar manner as reported for dimeric PGKs [29]. Other characteristics are shared by many PGKs. An optimum pH 7.0 to 7.4 is found for phosphoglycerate kinases from Escherichia coli and Homo sapiens, while Saccharomyces cerivisiae shows optimum activity at pH 7.5 [24,30,31]. The K M for ATP of PGK His of 0.11 mM is similar to those reported for PGK from Mus musculus and Homo sapiens [32,33] and within the range reported for prokaryotic PGKs, e.g. 0.21 mM for PGK from Pseudomonas sp. and 0.31 mM for yeast PGK [33,34]. The K M value of 0.26 mM for 3-PG of PGK His is comparable to that of PGK from Mus musculus [33] and slightly lower than that of PGK from Pseudomonas sp. which is 0.48 mM [35]. The enzyme is very specific for Mg +2 ions for catalysis as reported for several other phosphoglycerate kinases. Other divalent cations such as Ni 2+ , Co 2+ , Mn 2+ , Cd 2+ , Ca 2+ or Zn 2+ could not replace Mg 2+ and were inhibitory at higher concentrations. Zn 2+ is a strong inhibitor with K i of 0.45 mM. The monovalent cations Na + and K + activated the enzyme whereas sulfate ions showed no influence as reported for multimeric PGKs from archaea [22]. PGK from C. glutamicum was shown to be subject to allosteric regulation. Of the cellular metabolites tested  only ADP affected the activity of PGK His . ADP inhibited PGK from C. glutamicum as competitive inhibitor with a low K i value (0.1 mM). Since ADP concentrations in C. glutamicum typically range from 0.5 to 1 mM in glucose batch cultures [36], this indicates that gluconeogenesis via PGK is tightly regulated by the energy charge. Regulation by the energy charge has also been described for eukaryotic PGKs, such as mouse PGK which has a very low K i of ≈ 0.08 mM [32] and yeast PGK which has a K i of ≈ 0.22 mM [37]. PGK from C. glutamicum was shown here to be required for growth with different glycolytic as well as with different gluconeogenic carbon sources. Besides PGK, other enzymes of glycolysis are expected to be essential for both glycolysis and gluconeogenesis in C. glutamicum, however, experimental evidence has only been obtained for fructose-1,6-bisphosphate aldolase [38] and phosphoglycerate mutase [39]. Deletion of the genes encoding phosphofructokinase and glyceraldehyde-3-phosphate dehydrogenase (GapA) [40] prevented growth with glycolytic substrates, while growth with gluconeogenic substrates was still possible. On the other hand, growth of C. glutamicum on gluconeogenic carbon sources has been shown to be dependent on PEPCk [41] and FBPase [42], but the lack of these enzymes did not preclude growth with glycolytic substrates. Deletion of the phosphoglucoisomerase gene pgi perturbed growth on glucose indirectly since the glucose PTS permease gene ptsG was hardly expressed in the deletion mutant. The lack of pyruvate kinase cannot be compensated during growth with some glycolytic or gluconeogenic carbon sources [43]. Pyruvate kinase is essential for growth with glycolytic non-PTS carbon substrates such as maltose but not for PTS substrates such as glucose, fructose and sucrose. Similarly, pyruvate kinase is essential for growth on gluconeogenic carbon sources that do not enter the central carbon metabolism via pyruvate, such as acetate or citrate, but is not required for growth with pyruvate or lactate [43].
Overproduction of PGK did not notably accelerate growth of C. glutamicum with glucose (data not shown). However, ornithine and arginine production rates were increased upon pgk overexpression in the respective amino acid producing strains (Figure 3), while lysine production was not accelerated. It is tempting to speculate that arginine and ornithine biosynthesis are positively affected while lysine biosynthesis is not because the latter has a lower ATP requirement (1 ATP per Crude extracts were obtained by sonication of cells cultured in LB medium supplemnted with 1 mM IPTG and 25 μg/mL kanamycin. Figure 3 L-lysine, L-arginine and L-ornithine production by C. glutamicum DM1933, ARG1 and ORN1 strains either with empty vector (white column represents pVWEx1) or overproducing phosphoglycerate kinase (black column represents pVWEx1-pgk). Cells were grown in CgXII minimal medium with 4% glucose as a carbon source. Cell pellets of cultures grown in glucose CgXII minimal medium after consumption of the carbon source. Data represent mean values and standard deviations of six replicates from two independent cultivations with three flasks per strain each. Significant differences (p < 0.01 in a student's t-test) between empty vector control (pVWEx1) and strains carrying pVWEx1-pgk are highlighted by an asterisk.
lysine) than the former (2 ATP per ornithine or arginine which in addition requires carbamoylphosphate). Most metabolic engineering approaches of glycolysis in C. glutamicum focused on increasing the product yield e.g. by redirecting carbon flux to reduce glycolysis and increase pentose phosphate pathway flux and NADPH provision [44,45,46]. Overexpression of the gene encoding gluconeogenic enzyme FBPase during growth and production with sucrose increased gluconeogenic flux from fructose-1, 6-bisphosphate to glucose-6-phosphate and into the pentose phosphate pathway [47]. Production of D-lactic acid [48], succinic acid [49] and alanine [50] by C. glutamicum under oxygen-deprivation conditions is characterized by a high glycolytic flux which could be enhanced by overexpression of gapA. Also overexpression of the genes coding for phosphofructokinase, triosephosphate isomerase, and fructose-1,6-bisphosphate aldolase accelerated D-lactate production under oxygen-deprivation conditions [48].

Conclusions
C. glutamicum 3-phosphoglycerate kinase encoded by pgk was shown to be essential for growth with glycolytic as well as with gluconeogenetic carbon sources. ADP was shown to be a competitive inhibitor of PGK, which unlike most bacterial PGKs is active as a dimer. Since overexpression of pgk increased amino acid productivity for ornithine and arginine, but not for lysine, PGK may be a promising target to accelerate production processes requiring high glycolytic flux but needs careful testing.

Microorganisms and cultivation conditions
Microorganisms are listed in Table 4. ATCC 13032 was used as C. glutamicum wild type (WT) [51] along with the amino acid producing strains DM1933, ORN1, ARG1 [6], PUT21 [52]. C. glutamicum strains were precultured in lysogeny broth (LB) medium [53] with antibiotics added when appropriate. E. coli strains DH5α [54] and BL21 (DE3) [55] were used as host for cloning and heterologous expression, respectively. For growth and amino acid production experiments, exponentially growing cells of LB precultures (50 ml) were harvested by centrifugation (3200 × g, 10 min), and washed in CgXII medium [56] without carbon source. Cultures of 50 ml CgXII media containing 4% (w/v) glucose, 100 μg/ml spectinomycin or 25 μg/ml kanamycin, and 1 mM IPTG were inoculated to a final optical density (OD 600 ) of 1 and incubated in 500 ml baffled shake flasks at 30°C. The OD 600 was measured in dilutions resulting in an OD 600 between 0.05 and 0.25 using a Shimadzu UV-1202 spectrophotometer (Duisburg, Germany). For enzymatic activity determination in cell-free extracts, cells were grown in LB medium to mid-exponential phase Overexpression of pgk in C. glutamicum For overexpression of pgk, the gene was amplified via PCR from genomic DNA of C. glutamicum WT. The PCR was performed using the oligonucleotide primers listed in Table 5. To allow IPTG inducible expression of pgk in C. glutamicum the PCR-product was ligated into SmaI restricted vector pEKEx3 resulting in pEKEx3-pgk and XbaI restricted vector pVWEx1 resulting pVWEx1pgk. Sequencing confirmed the integrity of the construct.
Overproduction of PGK in E. coli, protein purification and molecular weight determination For heterologous expression of the 3-phosphoglyceratekinase gene pgk (Cg1790) in E. coli BL21 (DE3), pgk was amplified via PCR from genomic DNA of C. glutamicum WT using the following oligonucleotide primers listed in Table 5. The 1208 bp amplification product was cloned into vector pGEM-T (Promega, Mannheim, Germany) resulting in vector pGEM-T-pgk. After restriction with NdeI, the 1208 bp product from pGEM-T-pgk was ligated to NdeI restricted pET16b (Novagen, Madison, WI, USA). The vector, pET16b-pgk, allows IPTGinducible expression of an N-terminal tenfold Histagged pgk in E. coli BL21 (DE3). Cultivation of BL21 (DE3) (pET16b-pgk), Cells were grown in an incubator/ shaker at 37°C to an absorbance reading (A 600 ) of 0.6-0.8 at which point IPTG (0.5 mM) was added and the flasks were cooled to 22°C and further incubated for 4 hours. Cells were harvested by centrifugation (20 min at 3200 × g) and the cell pellet was washed with 20 mM Tris, 300 mM NaCl, 5 mM imidazole (TNI buffer), 5% (vol/vol) glycerol and stored at −80°C. Prior to lysis by French press, cells were resuspended in TNI buffer, and protease activity was inhibited by addition of 1 mM phenylmethylsulfonyl fluoride (PMSF) and 1 mM diisopropylfluorophosphate (DFP). The extract was cleared by centrifugation for 1 h at 25000 × g. Peak fractions of Ninitrilotriacetic acid (Ni-NTA) agarose affinity chromatography eluted with 20 mM Tris, 300 mM NaCl, 100, 200, or 400 mM imidazol, and 5% (vol/vol) glycerol were pooled, and the pooled fractions were desalted using Sephadex G25 gel filtration (Amersham Bioscience, Uppsala) and buffered in 100 mM triethanolamine hydrochloride (TEA-Cl), pH 7.4. The molecular weight of PGK His was determined using gel filtration and by cross-linking experiments. For gel filtration, a Bio-Prep SE-1000/17 column (BioRad, Richmond, VA, USA) was used and calibrated with Gel Filtration Standard (BioRad) containing thyroglobulin (670 kDa), bovine γ -globulin (158 kDa), chicken ovalbumin (44 kDa), equine myoglobin (17 kDa), and vitamin B 12 (1.35 kDa). The buffer used contained the optimal conditions for activity of PGK plus 100 mM TEA-Cl pH 7.4). To determine the PGK elution volume, two gel filtrations were performed, one in a mix with the indicated protein standards containing 2 mg/ml of PGK and a second with 4 mg/ml of PGK. Protein elution time was measured at 280 nm. Eluting PGK His was collected in 0.5 ml fractions and confirmed by enzyme activity measurement.

Assay conditions
The phosphoglycerate kinase activity was determined with purified enzyme and cell-free extracts of WT (pEKEx3) and WT(pEKEx3-pgk). Cell-free extracts were prepared according to Stansen et al. [57]. In vitro enzyme assays were carried out spectrophotometrically in a coupled assay in which, product of the first reaction