Effect of Overexpression of Actinobacillus succinogenes Phosphoenolpyruvate Carboxykinase on Succinate Production in Escherichia coli

Abstract

Succinate fermentation was investigated in Escherichia coli strains overexpressing Actinobacillus succinogenes phosphoenolpyruvate carboxykinase (PEPCK). In E. coli K-12, PEPCK overexpression had no effect on succinate fermentation. In contrast, in the phosphoenolpyruvate carboxylase mutant E. coli strain K-12 ppc::kan, PEPCK overexpression increased succinate production 6.5-fold.

Succinate is a four-carbon dicarboxylic acid, which has many applications in agriculture, food, and pharmaceutical industries. It is also a potential intermediary commodity chemical feedstock derived from biomass. Succinate production from glucose involves reductive CO₂ fixation. Several enzymes involved in CO₂ fixation have been overexpressed in Escherichia coli to enhance succinate production. Succinate production was increased by overexpressing pyruvate carboxylase, phosphoenolpyruvate (PEP) carboxylase (PPC), and malic enzyme, whereas overexpression of E. coli PEP carboxykinase (PEPCK) had no effect (3, 4, 8, 12, 18). In this paper, we call PPC the enzyme that catalyzes the formation of oxaloacetate plus P_i from PEP plus CO₂ (i.e., EC 4.1.1.38), and we call PEPCK the enzyme that catalyzes the formation of oxaloacetate plus ATP from PEP, ADP, and CO₂ (i.e., EC 4.1.1.49) (5). Actinobacillus succinogenes is a natural succinate producer (6, 7). In contrast to E. coli, PEPCK is the main CO₂-fixing enzyme in the A. succinogenes succinate production pathway (17). We have cloned and sequenced the A. succinogenes pckA gene (P. Kim, M. Laivenieks, J. McKinlay, C. Vieille, and J. G. Zeikus, submitted for publication). In this report, we describe the effect of A. succinogenes PEPCK overexpression on succinate production in wild-type and ppc mutant E. coli strains.

The characteristics of the E. coli strains used in this study are described in Table 1. The molecular biology techniques used in this study were performed as described previously (15). To construct a knockout K-12 ppc mutant, a ppc-5::kan gene was introduced into E. coli K-12 by P1 transduction with a P1 lysate of strain JCL1242 (Table 1) as described previously (13). E. coli DH5α was used as the host for the subcloning experiments. To clone the A. succinogenes pckA gene into an expression system, the A. succinogenes pckA gene (Kim et al., submitted) was amplified by using A. succinogenes chromosomal DNA as the template and oligonucleotides 5′-GCGAGAGTACTGACTTAAACAAACTCG (where the underlined sequence creates a ScaI site) and 5′-ACGCGTCGACCTCAGCCTTATTTTTCAG (where the underlined sequence creates a SalI site) as the forward and reverse primers, respectively. The 1.6-kb PCR product was cloned into pCRII (Invitrogen, Carlsbad, Calif.) and sequenced. The pckA gene was then subcloned into the EheI and SalI sites of expression vector pProEx-1 (Invitrogen), yielding plasmid pAsPCK. In this construct, PEPCK is expressed with an N-terminal His tag followed by a TEV protease cleavage site. Plasmid pAsPCK was used to overexpress A. succinogenes PEPCK in E. coli.

TABLE 1.

Strains used in this study

Strain	Characteristics	Source
DH5α	F− φ80 dlacZ ΔM15 Δ(lacZYA-argF) U169 endA1 recA1 hsdR17 (rK−mK+) deoR thi-1 phoA supE44 λ−gyrA96 relA1	Invitrogen Carlsbad, Calif.
K-12	Wild-type E. coli strain	Laboratory collection
K-12 ppc	E. coli K-12 ppc-5::kan	This study
JCL1242	ppc-5::kan DE(argF-lac)169 λ−	CGSC 7728
342-167	thr-1 leuB6(Am) fhuA2 lacY1 glnV44(AS) gal-6 hisG1(Fs) rfbD1 galP63 rpsL9 malT1(LamR) xylA7 mtlA2 ppc-1 argC32 thi-1 λ−	CGSC 3594
DL10	Δ(aroP-aceF)73 galK2(Oc) λ− IN(rnnD-rnnE)1 rpsL200 pck-11, ppc-1 maeA1 maeB3 maeR2	CGSC 7061
NJ27	[araD139] Δ(argF-lac)169 λ−manZ26 Δ(his-gnd)296 fruA25 relA1 rpsL150 (strR) ΔargH1 Δ(xylE-malB)262 deoC1 ppc-3	CGSC 6812

Single colonies of E. coli cells (parental strains and transformants) were transferred into 5 ml of Luria-Bertani broth and incubated at 37°C for 12 h with shaking. Five hundred microliters of these precultures was used to inoculate 50 ml of glucose-based medium (per liter, 9 g [50 mM] of glucose, 5 g of yeast extract, 10 g of NaHCO₃, 8.5 g of NaH₂PO₄ · H₂O, and 15.5 g of K₂HPO₄ [pH 7.0]) in 125-ml butyl rubber-stoppered serum vials under a CO₂ atmosphere. Ampicillin was supplemented at 50 μg/ml for the strains harboring a plasmid; kanamycin was added at 20 μg/ml for the ppc::kan K-12 derivatives. IPTG (isopropyl-β-d-thiogalactopyranoside [0.6 mM]) was added to induce A. succinogenes PEPCK expression. Fermentations were performed without shaking at 37°C. Samples were withdrawn every 2 h during the fermentation to determine the biomass, glucose, and metabolites. Biomass was estimated by measuring optical density at 600 nm (OD₆₀₀). Glucose and metabolites were analyzed on a Breeze high-performance liquid chromatograph (Waters, Milford, Mass.) equipped with an Aminex-87H ion-exchange column (Bio-Rad, Hercules, Calif.) and a refractive index (RI) detector. A UV detector was also used to confirm the organic acid RI data. The mobile phase was 4 mM H₂SO₄ (isocratic flow, 0.6 ml · min⁻¹), and the column was maintained at 40°C. To measure PEPCK activity, bacteria were harvested in the exponential phase and disrupted in a French press as described previously (17). Protein content in the extracts was determined with the Bio-Rad protein assay kit with bovine serum albumin as the standard. PEPCK activity was calculated by measuring ATP formation at 37°C with the Sigma Diagnostics ATP kit (Sigma, St. Louis, Mo.). The reaction mixture contained 100 mM Tris-HCl (pH 6.6), 35 mM NaHCO₃, 16 mM MgCl₂, 0.3 mM NADH, 2 U of phosphoglycerate phosphokinase/glyceraldehyde phosphate dehydrogenase, 1 mM dithiothreitol, 10 mM ADP, 1.8 mM 3-phosphoglycerate, 5 mM PEP, and the cell extract. The extinction coefficient for NADH was 6.22 cm⁻¹ mM⁻¹ at 340 nm.

The growth properties and fermentation balances of the E. coli K-12 and K-12 ppc strains were compared in the presence and absence of A. succinogenes PEPCK overexpression. Fermentation balances were determined at the 18-h time point [K-12, K-12(pAsPCK), and K-12 ppc(pAsPCK)] and at the 22-h time point (K-12 ppc) (Table 2), when all of the glucose was just depleted and before any fermentation product could be recycled. The E. coli K-12 ppc strain grew slower and more poorly than strain K12 (as indicated by a 40% reduction in biomass). This result is not surprising, since E. coli ppc strains do not grow on glucose-minimal medium (11). In our fermentation, the K-12 ppc strain derives its tricarboxylic acid cycle intermediates from the yeast extract present in the medium. Succinate production by the K-12 ppc strain decreased by more than 70%, while its lactate production increased. The two strains harboring plasmid pAsPCK showed at least 60-fold-higher PEPCK activity than the strains containing no plasmid, indicating that A. succinogenes PEPCK (with an N-terminal His tag and TEV protease site) is expressed as a functional enzyme in E. coli. With the same amount of biomass produced by K-12 ppc(pAsPCK) as that by K-12, A. succinogenes PEPCK overexpression restored full growth to K-12 ppc. In E. coli K-12, A. succinogenes PEPCK overexpression did not significantly affect succinate production, nor did it affect the overall fermentation balance. In contrast, PEPCK overexpression in K-12 ppc increased succinate production more than sixfold. The amount of succinate produced by K-12 ppc(pAsPCK) reached twice the amount produced by K-12 (Table 2). Concomitantly, lactate production decreased by almost 30%, while levels of acetate, formate, and ethanol production remained unchanged. In all cases, the theoretical redox balance based on the flux estimations was between 0.68 and 0.89 (data not shown). Because the carbon recoveries were all about 100%, these redox balance values significantly below 1 suggest that some of the electrons are directed toward the production of hydrogen (not determined) or that they are consumed in uncounted reactions involving components of the yeast extract.

TABLE 2.

Effect of A. succinogenes PEPCK overexpression on the fermentation balance of E. coli K-12 and K-12 ppc

Strain	Concn of biomass and metabolites (mM)a						PEPCK activity (nmol/min · mg of protein−1)	Succinate ratioc	C recoveryd
	Biomassb	Succinate	Lactate	Acetate	Formate	Ethanol
K-12	33.8 ± 0.5	11.4 ± 0.4	37.4 ± 1.2	31.1 ± 0.6	44.4 ± 1.8	6.5 ± 0.2	3.1 ± 1.1	0.132	1.00 ± 0.04
K-12(pAsPCK)	34.7 ± 0.4	10.6 ± 0.5	35.9 ± 2.2	29.9 ± 1.6	43.9 ± 2.0	6.5 ± 0.2	194.4 ± 14.8	0.128	0.97 ± 0.03
K-12 ppc	20.0 ± 1.2	3.1 ± 0.1	48.8 ± 1.6	27.4 ± 0.4	45.2 ± 1.4	6.7 ± 0.7	1.3 ± 0.3	0.074	0.96 ± 0.07
K-12 ppc(pAsPCK)	35.2 ± 0.6	20.2 ± 0.9	35.0 ± 0.9	28.0 ± 0.8	45.9 ± 2.1	6.0 ± 0.1	200.9 ± 18.2	0.226	1.04 ± 0.02

^aBiomass and metabolite concentrations were determined at the 18-h time point [K-12, K-12(pAsPCK), and K-12 ppc(pAsPCK)] and at the 22-h time point (K12 ppc), when glucose was just completely depleted.

^bBiomass carbon was calculated by using the ratio 13.001 mM biomass/OD₆₆₀ based on the formula for CH₂O_0.5N_0.2 (16).

^cCalculated as succinate/(lactate + acetate + ethanol + succinate).

^dThe balance was corrected for CO₂ fixation to succinate (17).

Since no PPC activity is detectable in A. succinogenes, PEPCK is the only enzyme that mediates PEP carboxylation (17). In this organism, the ratio of pyruvate kinase and PEPCK activities probably determines the balance between the C3 and C4 metabolic branches, as reported in other rumen bacteria (1). In E. coli, however, PPC is the enzyme that carboxylates PEP during growth on glucose. PEPCK seems better suited than PPC for succinate production, though, because it generates ATP-conserving energy during glycolysis, whereas PPC dissipates that energy. Still, PPC overexpression in E. coli JCL1208 (a Δlac strain containing a chromosomally inserted lacI^q gene (12) increased succinate production, whereas overexpression of E. coli PEPCK had no effect (12). As expected, the same result was obtained when we overexpressed A. succinogenes PEPCK in E. coli K-12. A possible explanation for these results is the difference in kinetics between PPC and PEPCK, as extensively discussed in reference 12. E. coli PPC has a K_m toward bicarbonate of 0.15 mM (14), whereas the K_ms toward bicarbonate of E. coli and Anaerobiospirillum succiniciproducens PEPCKs are 13 mM (9) and 30 mM (10), respectively. With 74% sequence identity and 84% similarity between A. succinogenes and E. coli PEPCKs (Kim et al., submitted), it is likely that A. succinogenes PEPCK has a K_m toward bicarbonate in the same range as E. coli PEPCK does. This hypothesis could be verified in the future by kinetic studies of A. succinogenes PEPCK. Millard et al. (12) expressed the alternative hypothesis that “unknown regulatory controls at the enzyme level may prevent E. coli PEPCK from functioning in the reverse direction in vivo.” In the K-12 ppc strain, though, A. succinogenes PEPCK can replace PPC as the PEP-carboxylating enzyme in a glucose-based medium. Since A. succinogenes and E. coli PEPCKs are highly similar, it is likely that E. coli PEPCK would be able to replace PPC as the PEP-carboxylating enzyme in a ppc mutant E. coli strain. Our results suggest that in strains JCL1208(pCK601) (12) and K-12(pAsPCK) (this study), PEPCK does not function as the PEP-carboxylating enzyme due to PPC's and PEPCK's respective kinetic properties in relation to intracellular metabolite concentrations. This hypothesis concurs with the recent findings of Yang et al. (19). Contrary to the common belief that E. coli PEPCK is only expressed during gluconeogenesis, Yang et al. showed that PEPCK shows significant activity in glucose-grown cultures, in which it decarboxylates oxaloacetate back to PEP. A detailed metabolic flux analysis suggested that the opposing fluxes generated by PPC and PEPCK activities are controlled at the kinetic level by PEP and oxaloacetate concentrations (19).

Chao and Liao (2) tested the growth of E. coli strain JCL4212(pCK601) (a Δ[argF-lac]169 ppc mutant strain overexpressing E. coli PEPCK) on minimal medium containing glucose plus succinate. This strain grew poorly, and Chao and Liao concluded that PEPCK overexpression was inhibitory for growth. This observation probably led to the comment by Gokarn et al. (3) that PEPCK “was unable to complement the absence of PPC in E. coli.” Our results with K-12 ppc(pAsPCK) (i.e., normal growth rate and enhanced succinate production) contradict this statement and Chao and Liao's results. To confirm our results, we performed similar experiments with three other ppc mutant E. coli strains (i.e., strains 342-167, DL10, and NJ27) (Table 1). PEPCK activity levels increased between 68- and 189-fold in the three strains overexpressing A. succinogenes PEPCK (data not shown). As observed with strain K-12 ppc, PEPCK overexpression improved growth of the poorly growing strain 342-167 by 2.5-fold (Table 3). Finally, overexpression of A. succinogenes PEPCK in 342-167, DL10, and NJ27 increased succinate production 3.0-, 4.2-, and 2.6-fold, respectively. The results with these last three strains show that our results are not strain specific. In addition to other mutations, strain NJ27 contains the same Δ[argF-lac]169 deletion as that in strain JCL4212. Under our fermentation conditions, overexpression of PEPCK in NJ27 did not inhibit growth (Table 3), but it did increase succinate production significantly. What differentiates our experiment with NJ27(pAsPCK) from Chao and Liao's experiment with JCL4212(pCK601) is the presence in our fermentation medium of 0.5% yeast extract. In the K-12-MG1655 genomic sequence (at www.tigr.org), the argF-lac region mutation encompasses approximately 70 genes and putative genes. The poor growth of JCL4212(pCK601) on minimal medium-glucose-succinate (2) could have been due to the absence of any or several of these genes rather than to growth inhibition by PEPCK.

TABLE 3.

Effect of A. succinogenes PEPCK overexpression on succinate production by E. coli ppc strains 342-167, DL10, and NJ27

Strain	Growth (OD660)a	Succinate concn (mM)a
342-167	1.27 ± 0.06	8.2 ± 5.7
342-167(pAsPCK)	3.12 ± 0.04	26.4 ± 2.8
DL10	2.43 ± 0.11	4.7 ± 4.7
DL10(pAsPCK)	2.69 ± 0.24	12.1 ± 1.9
NJ27	1.73 ± 0.17	3.0 ± 0.5
NJ27(pAsPCK)	1.97 ± 0.16	12.5 ± 1.2

^aData were collected at the 24-h time point.

The main results of this work are that PEPCK can replace PPC as the PEP-carboxylating enzyme in E. coli and that PEPCK overexpression results in a significant increase in succinate production. K-12 ppc(pAsPCK) produces almost twice as much succinate as K-12 does. Upon PPC overexpression in E. coli JCL1208, Millard et al. (12) observed increases in succinate production in nonoptimized and optimized fermentation media of 1.36- and 3.5-fold, respectively. Because the fermentation medium we used is closer in composition to the nonoptimized medium used by Millard et al., one would expect PEPCK overexpression to increase succinate production more than it already does, under optimized conditions. In this respect, since PEPCK's K_m for HCO₃⁻ is 2 orders of magnitude higher than that of PPC, optimizing the HCO₃⁻ concentration might be a critical step toward increasing succinate production through PEPCK overexpression. Because the increase in succinate production observed in this study is coupled to the substitution of PPC by PEPCK (an enzyme that couples ATP production to PEP carboxylation), this engineering step represents an attractive option for the design of a succinate-producing E. coli strain.