Engineering of Acetate Recycling and Citrate Synthase to Improve Aerobic Succinate Production in Corynebacterium glutamicum

Corynebacterium glutamicum lacking the succinate dehydrogenase complex can produce succinate aerobically with acetate representing the major byproduct. Efforts to increase succinate production involved deletion of acetate formation pathways and overexpression of anaplerotic pathways, but acetate formation could not be completely eliminated. To address this issue, we constructed a pathway for recycling wasted carbon in succinate-producing C. glutamicum. The acetyl-CoA synthetase from Bacillus subtilis was heterologously introduced into C. glutamicum for the first time. The engineered strain ZX1 (pEacsA) did not secrete acetate and produced succinate with a yield of 0.50 mol (mol glucose)−1. Moreover, in order to drive more carbon towards succinate biosynthesis, the native citrate synthase encoded by gltA was overexpressed, leading to strain ZX1 (pEacsAgltA), which showed a 22% increase in succinate yield and a 62% decrease in pyruvate yield compared to strain ZX1 (pEacsA). In fed-batch cultivations, strain ZX1 (pEacsAgltA) produced 241 mM succinate with an average volumetric productivity of 3.55 mM h−1 and an average yield of 0.63 mol (mol glucose) −1, making it a promising platform for the aerobic production of succinate at large scale.


Introduction
Succinic acid is considered as one of the most important platform chemicals because of its extensive applications in many industrial fields. It can be converted to a wide range of products including specialty chemicals, food ingredients, green solvents, pharmaceuticals and biopolymers [1,2]. Currently, succinic acid production is accomplished by chemical synthesis using petroleum-derived maleic anhydride. However, this process requires high temperature, high pressure and heavy metal catalysts, which makes the conversion costly and ecologically questionable [1,3]. Thus, microbial production of succinate from cheap and renewable carbon sources attracted great interest as more sustainable replacement for conventional petrochemical routes in recent years. In 2004, succinic acid was identified by the U.S. Department of Energy as one of the top 12 building block chemicals that could be produced from renewable resources [4].
Succinate can be produced anaerobically by naturally isolated bacteria and recombinant organisms, such as rumen bacteria Actinobacillus succinogenes [5], Anaerobiospirillum succiniciproducens [6] and Mannheimia succiniciproducens [7], as well as recombinant Escherichia coli [8]. The rumen bacteria have shown great potential for industrial succinate production, but some of them are potentially pathogenic or strictly anaerobic and difficult to culture [1]. Corynebacterium glutamicum, a rapidly growing Gram-positive soil bacterium that can utilize a broad spectrum of carbon sources, represents another promising platform for succinate production. It does not grow under anaerobic conditions, but has the ability to utilize sugars and excrete lactate, succinate and small amounts of acetate in the presence of additional bicarbonate [9]. Based on this capability, metabolically engineered C. glutamicum strains (ATCC 13032 and R) were developed and the culture conditions were optimized for anaerobic production of succinate [9][10][11].
Even though anaerobic processes for succinate production have resulted in high yields and titers, some problems are still to be solved, e.g. the optimal redox balance was limited [12,13]. The aerobic cell culture conditions were reported to have advantages over anaerobic conditions due to higher biomass generation, faster carbon throughput and product formation [13,14]. Thus, aerobic platforms were designed and constructed. Several aerobic succinate-producing E. coli strains were designed and could achieve high succinate yields using various sugars [12,14,15]. The yeasts Saccharomyces cerevisiae and Yarrowia lipolytica which showed high tolerance toward acidity were also genetically modified for succinate production, but with low volumetric productivities [3,16]. Recently, Litsanov et al [13] reported that C. glutamicum ATCC 13032 lacking the succinate dehydrogenase complex showed greater potential for aerobic succinate production than other producers with equivalent genetic backgrounds.
Disruption of all known acetate formation pathways as well as overexpression of anaplerotic pathways resulted in a strain that produced 90 mM succinate with a yield of 0.45 mol (mol glucose) 21 under optimized production conditions [13]. But this strain produced succinate with a low volumetric productivity of 0.77 mM h 21 and accumulated acetate with a yield of 0.08 mol (mol glucose) 21 [13]. The acetate accumulation resulted in a waste of carbon source, therefore negatively affecting succinate production.
The aims of this study are to eliminate acetate accumulation as well as to further increase succinate production by C. glutamicum ATCC 13032 under aerobic conditions. First, the acetate utilization pathway from Bacillus subtilis was introduced into C. glutamicum ATCC 13032. Moreover, the native citrate synthase was overexpressed to redirect more carbon flux towards the tricarboxylic acid (TCA) cycle. Also, the effect of overexpression of the succinate exporter sucE gene, normally induced under anaerobic conditions, on succinate production was examined. Finally, a succinate titer of 241 mM with a volumetric productivity of 3.55 mM h 21 and a yield of 0.63 mol (mol glucose) 21 were achieved by fed-batch cultivations.

Strains, Plasmids and Media
All strains and plasmids used in this study, their sources and relevant characteristics were given in Table 1. During plasmids construction, E. coli DH5a was used and cultured in lysogeny broth complex medium (LB). Plasmid DNA transferring into C. glutamicum ATCC 13032 was carried out by electroporation, and the recombinant strains were selected on brain heart infusionsorbitol (BHIS) agar plates containing 25 mg mL 21 kanamycin.

Genetic Methods
All oligonucleotides used in this study were given in Table 2. For construction of plasmid pEacsA, the acsA gene from B. subtilis was amplified from chromosomal DNA by PCR with primer pair acsA1/acsA2. The resulting fragment was digested with SacI and XbaI and ligated into SacI/XbaI-restricted vector pEC-XK99E. For construction of plasmid pEacsAgltA, the gltA gene from C. glutamicum ATCC 13032 was amplified from chromosomal DNA by PCR with primer pair gltA1/gltA2. The resulting fragment was digested with XbaI and SbfI and ligated into XbaI/SbfI-restricted pEacsA. For construction of plasmid pEacsAgltAsucE, the sucE gene from C. glutamicum ATCC 13032 was amplified from chromosomal DNA by PCR with primer sucE1/sucE2. The resulting fragment was digested with SbfI and ligated into SbfI-restricted pEacsAgltA. The inframe deletion of sdhCAB operon in C. glutamicum ATCC 13032 was achieved via a two-step homologous recombination procedure using the suicide vector pDsacB. The flanking regions of sdhCAB operon were amplified from chromosomal DNA using primer pairs sdh1/sdh2 and sdh3/sdh4. The flanking fragments were gel purified, mixed in equal amounts, and subjected to crossover PCR. The resulting fusion products containing the upstream and downstream regions were ligated into EcoRI/PstI-restricted pDsacB and transformed into C. glutamicum ATCC 13032 by electroporation. The procedure of gene deletion was carried out as described previously [17].

Cultivation Conditions
For the precultivation of C. glutamicum ATCC 13032 strains, single clones were grown in 5 mL of modified CGIII medium containing 10 g L 21 tryptone, 10 g L 21 yeast extract and 42 g L 21 3-morpholinopropanesulfonic acid (MOPS) (pH 7.4) at 30uC and 220 rpm. After overnight incubation, cells of the preculture were inoculated into 500 mL shake flasks containing 50 mL modified CGIII medium supplemented with 55 mM glucose to an initial OD 600 of 0.06. After inoculation, cultivations for succinate production were performed at 250 rpm and 30uC. For induction, up to 1 mM isopropyl b-D-1-thiogalactopyranoside (IPTG) was added to the culture medium.
The fed-batch cultures were performed in a 1.3-L glass bioreactor (BioFlo 110, New Brunswick Scientific, USA). 50 mL modified CGIII medium supplemented with 55 mM glucose was inoculated with 1 mL frozen stock culture of strain ZX1 (pEacsAgltA) and the culture was incubated on a rotary shaker at 30uC and 250 rpm. Subsequently, the cells were harvested and inoculated into 600 mL modified CGIII medium (pH 7.0) supplemented with about 222 mM glucose to an initial OD 600 of about 1. The aeration rate was set to 1.0 L min 21 and the dissolved oxygen was maintained above 30% saturation level by regulating the stirrer speed from 500 to 1000 rpm. The pH was maintained at 7.0 by automatic addition of the mixture of 6 M potassium hydroxide combined with 2 M potassium carbonate, and the temperature was maintained at 30uC. During the fedbatch process, adequate amounts of 80% (w/v) glucose were injected.

Analytical Techniques
Extracellular organic acids were measured by HPLC (HP 1100 LC, Agilent Technologies, USA) equipped with a cationexchange column (HPX-87H, BioRad, USA), an UV absorbance detector (Agilent Technologies, G1315D) and a refractive index (RI) detector (Agilent Technologies, HP1047A). A mobile phase of 2.5 mM H 2 SO 4 solution at a 0.4 ml min 21 flow rate was used and the column operated at 65uC. Succinate and acetate were measured by the RI detector while pyruvate was measured by the UV detector at 210 nm. Glucose concentration was monitored using a SBA sensor machine (Institute of Microbiology, Shandong, China). Growth was determined by measuring the optical density at 600 nm (OD 600 ) and one unit of absorbance at 600 nm corresponded to 0.25 g cell dry weight (CDW) per liter. For determination of intracellular acetyl-CoA, 100 mL mid-exponential phase cell culture was taken into precooled centrifuge tubes and centrifuged at 90006g and 4uC for 10 min. The cell pellets were washed twice with 50 mL 100 mM sodium phosphate buffer (pH 3.0, 4uC). Acetyl-CoA was analyzed by HPLC (HP 1100 LC, Agilent Technologies, USA) equipped with a Hypersil ODS (C18) column (12564 mm, 5 mm, Phenomenex). The sample preparation and the analysis procedure were carried out as previously described [18].

Enzyme Assays
C. glutamicum ATCC 13032 strains were cultivated in modified CGIII complex medium supplemented with 55 mM glucose. At exponential phase, the cells were harvested by centrifugation for 10 min at 90006g and 4uC, washed twice with 15 mL 50 mM Tris-HCl (pH 7.5), centrifuged again and resuspended in the same buffer. The cells were then subjected to sonication for 15 min in an ice bath (130 W, 20 kHz, pulse: 5 s on; 5 s off). Cell debris was removed by centrifugation for 30 min at 100006g and 4uC and the supernatants were used as cell extracts. The acetyl-CoA synthetase activity was determined as previously described [19]. The citrate synthase activity was determined as described elsewhere [20]. Total protein concentrations were determined by the Bradford method using bovine serum albumin as standard. The specific enzyme activities were expressed as units per mg protein.

Real-time Quantitative PCR (RT-qPCR)
C. glutamicum ATCC 13032 strains were grown to the exponential phase. Then the cells were harvested by centrifugation for 2 min at 130006g and 4uC and washed once with distilled water. Total RNA was extracted with RNAprep pure Kit DP430 (Tiangen, Beijing, China). Firstly, 500 ng of total RNA was transcribed into cDNA using Quant Reverse Transcriptase with random primers (Tiangen, Beijing, China). Samples were then analyzed using a Light CyclerH 480 II (Roche, Basel, Switzerland) with Real Master Mix (SYBR Green). The 16S rRNA gene was used as reference for normalization and the primers used for gene quantification were list in Table 2.

Aerobic Succinate Production by Strain ZX1
Efforts to increase succinate production by C. glutamicum ATCC 13032 under aerobic conditions have been made by disrupting the succinate dehydrogenase complex and pathways leading to the synthesis of acetate, as well as co-overexpressing pyruvate carboxylase and PEP carboxylase [13]. Similarly, starting with the wild-type ATCC13032, the genes of pqo, pta and cat were disrupted, and both of the native promoters of genes of ppc and pyc were replaced by the strong promoter of the superoxide dismutase (sod) (Figure 1). Also, the ldhA gene encoding L-lactate dehydrogenase was inactivated to avoid lactate accumulation in further strain improvement. Finally, the deletion of sdhCAB operon was introduced, yielding strain ZX1. To test its ability to produce succinate from glucose, shake-flask cultivations were carried out in modified CGIII medium with 55 mM glucose at 30uC and 250 rpm. In order to prevent acidification, 3-morpholinopropanesulfonic acid (MOPS) was added as a buffering agent. As described in Figure 2A and Table 3, strain ZX1 consumed glucose completely within 16 h and showed a growth rate of 0.35 h 21 . Also, it produced 24.60 mM succinate at the end of the cultivations with a succinate yield of 0.43 mol (mol glucose) 21 , which exhibited a 3.07-fold increase compared to the sdhCAB-dificient mutant. The specific succinate productivity was 0.94 mmol g CDW 21 h 21 . In addition, 7.33 mM pyruvate and 8.59 mM acetate were accumulated in the supernatant.
Accordingly, the ACS from B. subtilis was selected and expressed in strain ZX1, resulting in strain ZX1 (pEacsA). Because the expression of acetyl-CoA synthetase (ACS) in C. glutamicum ATCC 13032 has never been accomplished, crude extract enzyme assay was performed. The specific ACS activity of 8.0360.12 mU mgprotein 21 was verified in strain ZX1 (pEacsA) whereas no activity was detected with the empty plasmid (pEC-XK99E), indicating its functional expression in this host. In batch cultivations ( Figure 2B and Table 3), strain ZX1 (pEacsA) showed lower growth rate (0.27 h 21 ) compared to strain ZX1. In addition, there was no accumulation of acetate at the end of cultivations. In the case of succinate production, this strain produced 28.67 mM succinate at the end of cultivations and showed a slight increase in succinate yield (0.50 versus 0.43 mol (mol glucose) 21 ). The specific succinate productivity of strain ZX1 (pEacsA) was 1.20 mmol g CDW 21 h 21 , which was 28% higher than that of strain ZX1. Also, 14.75 mM pyruvate was formed with a yield of 0.26 mol (mol glucose) 21 . These results demonstrated that exogenous ACS expression in succinate-producing C. glutamicum ATCC 13032 was an efficient approach to cope with acetate accumulation.

Reduction of Pyruvate Accumulation by Overexpressing Citrate Synthase
As described above, strain ZX1 (pEacsA) showed restricted growth and accumulated pyruvate as the major byproduct. In E. coli, inactivating acetate formation pathways led to reduced cell growth and/or increased pyruvate accumulation, which were attributed to the accumulation of intracellular metabolic intermediates, such as acetyl-CoA [24,25]. Thus, it could be that intracellular acetyl-CoA of strain ZX1 (pEacsA) would accumulate, especially in the presence of an acetate assimilation pathway as well as the absence of major acetate formation pathways ( Figure 1). It was also reported that activity of the pyruvate carboxylase, which contributed over 90% of the oxaloacetate replenishment when grown on glucose, was inhibited by acetyl-CoA (110 mM) [26]. Actually, intracellular metabolites analysis revealed that the level of acetyl-CoA of strain ZX1 (pEacsA) was 89% higher than that of strain ZX1 (Figure 3), which could explain the phenomena of higher accumulation of extracellular pyruvate and reduced growth rate with strain ZX1 (pEacsA) ( Figure 2B and Table 3). Recently, Lin et al [27] developed an efficient aerobic succinate-producing E. coli HL27659K and also found that this strain had high-level citrate synthase activity. In contrast, a lysine-overproducing mutant of C. glutamicum was reported to have low-level citrate synthase activity, which represented low flux towards TCA cycle [28].
Thus, in order to alleviate the accumulation of intracellular acetyl-CoA and pull more carbon towards succinate synthesis, the native citrate synthase encoded by gltA gene was overexpressed by introducing plasmid pEacsAgltA into strain ZX1, resulting in strain ZX1 (pEacsAgltA). The specific CS activity of this strain was 4.5160.09 U mg-protein 21 , which exhibited about six-fold increased activity compared to strain ZX1 (pEacsA) (0.6960.03 U mg-protein 21 ), indicating its functional overexpression. In subsequent cultivations, the cell growth rate and organic acids production, as well as intracellular acetyl-CoA level were The growth rate and specific succinate productivity were calculated within the exponential phase of the cultivations. The product concentration was determined at the end of cultivations when glucose was consumed. Values were given as the averages and standard deviations of three independent cultures. ND, not detected. doi:10.1371/journal.pone.0060659.t003 analyzed. As described in Figure 2C and Table 3, this strain showed a growth rate of 0.30 h 21 and accumulated 5.64 mM pyruvate with a yield of 0.10 mol (mol glucose) 21 , which was 62% lower compared to strain ZX1 (pEacsA). Also, the succinate yield of 0.61 mol (mol glucose) 21 and the specific succinate productivity of 1.26 mmol g CDW 21 h 21 were increased by 22% and 5% compared to strain ZX1 (pEacsA), respectively. In addition, the intracellular level of acetyl-CoA was decreased by 38% after introducing this modification ( Figure 3). These results indicated that enhanced expression of the citrate synthase was effective for the efficient production of succinate under aerobic conditions.
Effect of modifying export system on aerobic succinate production. In C. glutamicum ATCC 13032, several C4-dicarboxylic acid transporters have been identified. The dccT and dctA genes were responsible for C4-dicarboxylate uptake under aerobic conditions [29,30], while the sucE gene was responsible for succinate export under anaerobic conditions [31]. To examine whether succinate production could be further increased by sucE overexpression, plasmid pEacsAgltAsucE was introduced into strain ZX1, resulting in strain ZX1 (pEacsAgltAsucE). As expected, the sucE expression level of this strain was significantly increased by nearly seven-fold compared to strain ZX1 (pEac-sAgltA), reflecting its successful overexpression ( Figure 4). In batch cultivations, strain ZX1 (pEacsAgltAsucE) showed a growth rate of 0.29 h 21 , whereas the succinate yield and specific succinate productivity were almost unchanged compared to strain ZX1 (pEacsAgltA) ( Table 3), indicating that the expression level of this gene or the ability of other export systems [31] might be sufficient under this condition.
Fed-batch cultures for aerobic succinate production. In order to evaluate the suitability of strain ZX1 (pEacsAgltA) as a host for the aerobic succinate production with higher level, fedbatch cultures were carried out in modified CGIII medium, initially containing 222 mM glucose. The detailed results of the fed-batch cultivation were described in Figure 5. After the glucose concentration was below 55 mM, adequate amounts of 80% glucose (w/v) were added at 23 h. Strain ZX1 (pEacsAgltA) grew to a maximum OD 600 of 75 within 35 h. The succinate concentration continuously increased to 76 mM within 23 h, resulting in a volumetric productivity of 3.29 mM h 21 . The volumetric productivity during the second stage (after glucose feeding) was further increased to 3.69 mM h 21 . The overall succinate yield was 0.63 mol (mol glucose) 21 . The pyruvate yield obtained in bioreactor (0.02 mol (mol glucose) 21 ) was lower compared to that obtained in shake flask (0.10 mol (mol glucose) 21 ). It could be due to the different dissolved oxygen levels between bioreactor and shake flask. In shake flask, the cells faced   oxygen limitation when OD 600 was above 10 [32], while the oxygen was surplus (above 30% saturation level) in bioreactor. Recently, Wieschalka et al [17] described a similar observation for a pyruvate production process with C. glutamicum ATCC 13032. They found that surplus oxygen could reduce pyruvate production and the physiological state of the cells had a significant impact on the overall production behavior. Taken together, these results indicated that strain ZX1 (pEacsAgltA) was a useful platform for the production of succinate under aerobic conditions.

Conclusions
In summary, this work demonstrated that introduction of an acetate assimilation pathway from B. subtilis was an efficient approach to reduce acetate accumulation in succinate-producing C. glutamicum ATCC 13032. It was also shown that high level expression of citrate synthase could assure efficient production of succinate under aerobic conditions. To the best of our knowledge, the succinate titer, yield and productivity obtained in fed-batch cultivations with strain C. glutamicum ATCC 13032 ZX1 (pEacsAgltA) are the highest currently reported with Corynebacterium species under fully aerobic conditions. Moreover, The complete elimination of byproduct acetate in the production process and the high succinate productivity with strain C. glutamicum ATCC 13032 ZX1 (pEacsAgltA) offer advantages over other microorganisms, making it a promising platform for the aerobic production of succinate [3,12,14,15,16]. Even so, further strain improvement and optimization of culture conditions will be pursued to make better performance.