Production of Glyoxylate from Glucose in Engineered Escherichia coli

: Glyoxylates are essential intermediates in several metabolic pathways and have a broad range of industrial applications. In this study, we propose a novel method for producing glyoxylate from glucose using engineered Escherichia coli BW25113. To direct the production of glyoxylate from glucose, malate synthase A ( aceB ), malate synthase G ( glcB ), glyoxylate carboligase ( gcl ), and glyoxylate/hydroxypyruvate reductase A ( ycdW) genes were disrupted, and the glyoxylate shunt was reinforced in the disruptants by the overexpression of citrate synthase ( gltA ) and isocitrate lyase ( aceA ). In ﬂask cultivation using M9 medium supplemented with 1% glucose, the disruptant E. coli BW25113 ∆ aceB ∆ glcB ∆ gcl ∆ ycdW produced 0.93 ± 0.17 g/L of glyoxylate. Further overexpression of gltA and aceA in the disruptant resulted in an improvement in glyoxylate production to 1.15 ± 0.02 g/L. By expressing a heterologous gene, pyc , in the engineered E. coli , the accumulation of intracellular oxaloacetate remarkably improved, leading to glyoxylate production of up to 2.42 ± 0.00 g/L with speciﬁc productivity at 4.22 ± 0.09 g/g-cell. To date, this is the highest reported titer and speciﬁc productivity of glyoxylate in E. coli .


Introduction
Escherichia coli has garnered research attention owing to its fast growth rate and straightforward capacity for scale-up [1,2] because of its well-documented genome [3]. Gene modification in E. coli paves the way for the production of novel products. Recent studies have reported the production of multiple important substances by E. coli, including acetate, ethanol, itaconate, lactate, and propanediol [4][5][6][7][8]. These studies demonstrate the advantages of using an E. coli model to produce various biochemicals.
As a carboxylate, glyoxylate is of considerable interest owing to its functional properties [9,10]. Typical glyoxylate molecules harboring both aldehyde and carboxyl groups exhibit flexible chemical interactions. This stimulates pathways that produce various biochemical products including anticancer agents such as bestatin [11], anti-inflammatory substances [12], and antihypertensive agents such as atenolol [13]. One of the most popular applications of glyoxylates is the production of vanillin, a well-known aroma derivative used in the perfume and food industries [14]. Glyoxylate is chemically synthesized from malate in a three-step process [15]. Although maleate synthesis is complex, this approach produces glyoxylate with 97% purity [15]. Another approach involves the production of glyoxylate from oxalate via a reduction reaction or glyoxal via anodic oxidation [16].
The glyoxylate shunt serves as an alternative to the tricarboxylic acid (TCA) cycle and benefits bacteria through various biochemical processes [17]. First, this shunt enables bacteria to utilize acetate, fatty acids, and ethanol to recover TCA cycle intermediates, which are required for gluconeogenesis and other biosynthetic processes [18]. Second, the metabolites derived from the glyoxylate shunt are mainly organic acids such as glycolate are required for gluconeogenesis and other biosynthetic processes [18]. Second, the metabolites derived from the glyoxylate shunt are mainly organic acids such as glycolate and tartronate semialdehyde, which have high potential for application as industrial chemicals [19]. However, a considerable drawback of the glyoxylate shunt in bacteria is the production of multiple byproducts, such as lactate and malate [7,20]. Glyoxylate is also synthesized by glyoxylate oxidases in plant leaves and animal livers, with hydrogen peroxide as a by-product [21][22][23][24]. Although chemical production would facilitate the large-scale production of glyoxylate, this is associated with a large amount of physical and economic effort owing to the requirement of malate, oxalate, and glyoxal, as discussed above [15,16].
Recently, Li et al. produced 0.74 g/L of glyoxylate from xylose via a glyoxylate shunt using engineered E. coli K-12 [20]. Overexpression of dte, fucK, fucA, aldA, and glcDEF assimilated xylose to glyoxylate and disruption of aceB, glcB, gcl, and ycdW accumulated glyoxylate at higher concentrations. However, the conversion of xylose to glyoxylate resulted in a limited titer (0.74 g/L). Despite successful attempts to produce glycolate from glucose via the glyoxylate shunt in E. coli, resulting in high glycolate accumulation [25], related studies on the conversion of glucose to glyoxylate are currently lacking. The application of a glyoxylate shunt with positive results suggests the application of this pathway from glucose to glyoxylate.
Recent studies have attempted to improve the operation of the TCA cycle by introducing a heterologous pyruvate carboxylase gene (pyc) into E. coli. The expression of pyc has positive effects on the production of various chemicals in the TCA cycle, such as succinate [26], malate [27], and fumarate [28]. In a study by Chen et al. (2020), the expression of pyc in E. coli significantly improved the operation of the glyoxylate shunt, leading to an increase in fumarate concentration [28]. This suggests the expression of pyc would activate the glyoxylate shunt in E. coli. However, no study has demonstrated an improvement in glyoxylate production by expression of pyc in E. coli.
The present study demonstrated the production of glyoxylate from glucose in E. coli BW25113 engineered by disrupting aceB, glcB, gcl, and ycdW, along with the expression of gltA, aceA, and pyc, to achieve a higher production titer of glyoxylate ( Figure 1). As a result, for the first time, glyoxylate was produced from glucose with the highest reported titer (2.42 ± 0.00 g/L) and productivity (4.22 ± 0.09 g/g-cell). These results could also help improve the bioproduction of value-added chemicals derived from glyoxylates.  reductase A (ycdW) were disrupted, and citrate synthase (gltA) and isocitrate lyase (aceA) were overexpressed. Pyruvate carboxylase (pyc) from Corynebacterium glutamicum NBRC 12168 was also expressed. Red and blue arrows indicate reinforced and disrupted pathways, respectively.

Gene Disruption
The bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli BW25113 cells were obtained from the NITE Biological Resource Center (NBRC, Chiba, Japan). Gene disruption was performed using the tetA-sacB disruption method [29]. Specifically, DNA fragments containing the tetA-sacB cassette were amplified using the following primer sets: aceB-F-tetA and aceB-R-sacB for aceB, glcB-F-tetA and glcB-R-sacB for glcB, gcl-F-tetA and gcl-R-sacB for gcl, and ycdW-F-tetA and ycdW-R-sacB for ycdW (Table S1). These primers were designed with homologous 50-bp upstream and downstream regions of the target disrupted genes, using the tetA-sacB cassette encoded on the genomic DNA of E. coli T-SACK [29] as the template. Polymerase chain reaction (PCR) was performed under the following conditions: denaturation at 98 • C for 30 s, 35 cycles of denaturation at 98 • C for 10 s, annealing at 60 • C for 30 s, and elongation at 72 • C for 2 min. All PCR master mixes were formulated using the Q5 High-Fidelity PCR Kit (New England Biolabs, Inc., MA, USA). The amplified fragments were introduced into E. coli BW25113 harboring pKD46 [30] by electroporation (2500 V, 200 Ω, 25 µF), resulting in a series of E. coli strains carrying the tetA-sacB cassette. The upstream and downstream regions of the target genes were amplified via PCR using the following primer sets: aceB-up-F, aceBup-R, aceB-dn-F, and aceB-dn-R for aceB; glcB-up-F, glcB-up-R, glcB-dn-F, and glcB-dn-R for glcB; gcl-up-F, gcl-up-R, gcl-dn-F, and gcl-dn-R for gcl; and ycdW-up-F, ycdW-up-R, ycdW-dn-F, and ycdW-dn-R for ycdW. PCR was performed under the following conditions: denaturation at 98 • C for 30 s, 35 cycles of denaturation at 98 • C for 10 s, annealing at 60 • C for 30 s, and elongation at 72 • C for 2 min. The amplified upstream and downstream regions were connected by SOEing PCR [31] and introduced into the resulting mutants, followed by screening on Luria-Bertani (LB) agar medium [32] containing 6% sucrose to eliminate the tetA-sacB cassette from their genomes. Finally, pKD46 was removed by overnight incubation at 45 • C [30], resulting in the construction of E. coli EG_1, EG_2, EG_3, and EG_4.

Plasmid Construction
For the construction of the pSR13 plasmid, gltA and aceA were amplified from the E. coli BW25113 genome using gltA-F, gltA-R, aceA-F, and aceA-R primers and were ligated with the fragment of the pGV3 expression vector [6] without oprI', which was amplified using pGV3-F and pGV3-R primers. The pLB6 plasmid was prepared by ligation with PCR amplicons of pyc from Corynebacterium glutamicum NBRC 12168 using the pyc-F and pyc-R primers and its genomic DNA as the template, and the pTV118N expression vector (Takara Bio Inc., Shiga, Japan) using the pTV118N-pyc-F and pTV118N-pyc-R primers and the vector as the template. The ligation was performed using the Gibson assembly technique [33]. Thereafter, E. coli ECOS DH5α (Nippon Gene, Tokyo, Japan) was transformed with the mixture using the heat shock method [34]. Cells were grown in LB medium for 1 h and plated on LB agar containing 100 µg/mL spectinomycin and 50 µg/mL carbenicillin for pSR13 and pLB6, respectively. The resulting plasmids were extracted from DH5α cells using the MagExtractor Plasmid kit (Toyobo Co., Ltd., Osaka, Japan), and E. coli BW25113 and its disruptants were transformed with the constructed plasmids by electroporation, resulting in the construction of E. coli EG_5, EG_6, and EG_7.

Fermentation Production
For pre-culture, all strains were grown overnight in LB medium. When required, 100 µg/mL of spectinomycin and 50 µg/mL of carbenicillin were added to the medium. For cultivation in a flask, 1 mL of E. coli preculture was transferred to flasks with 100 mL of M9 medium containing 1% glucose, 10 g/L glucose, 6 g/L Na 2 HPO 4 , 3 g/L KH 2 PO 4 , 0.5 g/L NaCl, 1 g/L NH 4 Cl, 2 mM MgSO 4 , 0.2 mM CaCl 2 , and 20 mg/L thiamine [7]. When required, 1 g/L of yeast extract was added to the medium. To analyze glyoxylate toxicity, 0-1.5% (w/v) sodium glyoxylate was prepared in the M9 medium. Escherichia coli BW25113 was pre-cultured in LB medium and inoculated into glyoxylate-supplemented M9 medium. Optical density at 600 nm (OD600) was checked at 24-h intervals ( Figure S1). For fermentation, flasks were capped with aerobic caps and shaken at 120 rpm and 30 • C under aerobic conditions. Supplementation with 5 mM of isopropyl-β-D-thiogalactopyranoside (Apollo Scientific, Stockport, UK) was performed for the induction of gene expression when OD600 reached 0.4. Cultures were sampled at appropriate intervals to determine the OD600 and to analyze the metabolite and residual glucose concentrations.

Analytical Methods
Culture supernatants were analyzed using a high-performance liquid chromatography (HPLC) system equipped with an ion exclusion column (Aminex HPX-87H; Bio-Rad, Hercules, CA, USA) and a refractive index detector (RID-20A; Shimadzu Corporation, Kyoto, Japan) for acetate, lactate, and formate quantification. Five millimolars of H 2 SO 4 was used as the mobile phase with a 0.4 mL/min flow rate at 65 • C. Glyoxylate in the culture supernatant was derivatized with phenylhydrazine reagent [35] and analyzed by HPLC. Briefly, 200 µL of glyoxylate sample (0.01-2 g/L) was mixed with 200 µL of 0.4 M phosphate buffer (pH 8.0) prepared by mixing K 2 HPO 4 and NaH 2 PO 4 in 70% (v/v) ethanol. Then, 20 µL of 5% (v/v) phenylhydrazine in 0.4 M phosphate buffer was added to the samples. Subsequently, the reaction occurred in mixing at 37 • C for 15 min. Then, the derivatized samples were injected into an HPLC system. The glyoxylate concentration was determined using an HPLC system equipped with a Cosmosil 5C18-AR-II (Nacalai Tesque, Inc., Kyoto, Japan) and a prominent diode array detector (SPD-M20A) at 324 nm using a 5% ethanol solution as the mobile phase in isocratic mode at a flow rate of 0.4 mL/min. The HPLC profiles of glyoxylate are shown in Figure S2. Glucose was quantified using a Glucose CII Test Kit (Fujifilm Wako Co., Osaka, Japan).
The correlation between cell weight (g/L) and OD600 was determined by the following method: E. coli cells were grown in a 5 mL test tube containing LB medium overnight. Then, 1 mL of cells was sampled, centrifuged to collect the cells, and washed three times with physiological saline water. Cell densities were prepared at different OD600 val- ues by dilution with physiological saline water. Then, 1 mL of each sample at different OD600 values was completely dried overnight at 70 • C, and the cell weight was measured. Subsequently, the cell weight was measured at OD600 using the correlation equation (1 OD600 = 0.33 g-cell/L).
Intracellular oxaloacetate was extracted using the methanol freeze-thaw method [36]. Bacterial cells (1 mL) were collected from the cultures and centrifuged at 15,000 rpm for 5 min. The cells were washed once with 1 mL saline solution. Subsequently, 250 µL of 50% methanol precooled at −40 • C was added to the cells. Samples were mixed, incubated in liquid nitrogen for 5 min, and thawed on ice for 5 min. The experiment was conducted in triplicate. The samples were then analyzed at 210 nm using a Hitachi LaChrom Elite HPLC System (Hitachi High-Tech Corp., Tokyo, Japan) equipped with a Cosmosil 5C18-AR-II column and UV detector (L-2400; Hitachi High-Tech Corp.). HPLC was performed in gradient mode with acetonitrile and water supplemented with 0.01% trifluoroacetic acid as the mobile phase. From 0 to 20 min, the concentration of acetonitrile containing trifluoroacetic acid increased from 0% to 100%. The flowrate was set at 1 mL/min and the temperature was stabilized at 25 • C. The standard samples were prepared with 0 to 2.5 g/L fresh standard oxaloacetate solutions (Tokyo Chemical Industry, Tokyo, Japan). Oxaloacetate was detected at a retention time of 3.2 min ( Figure S3). All fermentation samples were diluted with water to ensure that the oxaloacetate concentration was within the quantifiable range (0-2.5 g/L).
All data were analyzed in duplicate and are presented as averages with standard errors.

Effects of Gene Disruption on Glyoxylate Production
Glyoxylate can be toxic because of its highly reactive properties and accumulation inside cells as an aldehyde [37]. In this study, low (0.47 ± 0.06 g-cell/L) and no cell growth was noted in the presence of 1% and 1.5% glyoxylate, respectively ( Figure S1). Possible solutions to overcome the toxicity of glyoxylate to E. coli growth may be the utilization of extractive fermentation or continuous fermentation to avoid the accumulation of glyoxylate in the culture. Furthermore, appropriate channel proteins would be improved to efflux extracellular glyoxylate. In this study, glyoxylate production of up to 5 g/L did not negatively affect E. coli growth. As an intermediate metabolite in E. coli, the secreted concentration of glyoxylate in wild-type E. coli is theoretically low. The wild-type BW25113 produced 0.01 ± 0.00 g/L glyoxylate after 48 h of cultivation ( Figure 2A).
To improve glyoxylate production, the competing genes were disrupted. Sequential disruption of aceB, glcB, gcl, and ycdW led to an increase in the final glyoxylate content. Escherichia coli EG_1, EG_2, and EG_3 produced 0.52 ± 0.07, 0.62 ± 0.12, and 0.71 ± 0.05 g/L of glyoxylate after 48 h of cultivation, respectively ( Figure 2B-D). The highest production of glyoxylate was achieved in E. coli EG_4 with 0.93 ± 0.17 g/L ( Figure 2E). Previously, a deletion of aceB, glcB, and gcl in E. coli K-12 was reported to stimulate glyoxylate production from xylose [20]. Of note, in the previous study, the production of glyoxylate in a flask was lower (0.74 g/L) than that in the present study. These results suggested that glucose is a better carbon source than xylose for glyoxylate production by E. coli. Further disruption of ycdW and the use of glucose as a carbon source positively affected glyoxylate production.
A significant decrease in glyoxylate content was observed after 48 h of cultivation, although tertiary and quaternary gene disruptions were attempted. This may be owing to the central role of glyoxylate in E. coli metabolism, which converts glyoxylate into other metabolites. A potential side pathway is the conversion of glyoxylate to 4-hydroxyoxoglutarate, which is catalyzed by an enzyme encoded by the 2-keto-4-hydroxyglutarate aldolase gene (eda) [38]. This gene limits the accumulation of glyoxylate and drives its conversion to 4-hydroxyoxoglutarate. A possible solution to overcome the assimilation of glyoxylate conferred by eda is to limit the accumulation of pyruvate, as pyruvate is required for condensation with glyoxylate to form 2-keto-4-hydroxyglutarate [38]. We found that the  As shown in Figure 2, EG_1 (0.51 ± 0.02 g-cell/L), EG_2 (0.54 ± 0.03 g-cell/L), EG_3 (0.60 ± 0.02 g-cell/L), and EG_4 (0.49 ± 0.03 g-cell/L) showed a lower growth at 72 h than BW25113 (1.05 ± 0.01 g-cell/L). This may be because of the effect of gene disruption on metabolism. Disruption of aceB results in a significant decrease in cell growth. In fact, the disruption of aceB and glcB can interfere with the glyoxylate shunt assimilating acetate in E. coli [39]. This is compatible with the increase in the residue acetate concentration observed in EG_2 (4.82 ± 0.14 g/L) compared to that in the wild-type BW25113 (3.31 ± 0.03 g/L).
Among the disruptants, the deletion of ycdW created a significant impact on E. coli growth, with a final cell biomass of 0.50 ± 0.04 g-cell/L. Although the effects of ycdW on cell growth are unclear, an attempt at gene disruption of ycdW results in 10% lower cell biomass [40]. Since the eda gene plays an important role in the bacterial Entner-Doudoroff pathway, further disruption of eda was not performed in the present study to avoid possible effects on cell growth because glyoxylate is produced during cell growth.

Effects of Gene Overexpression on Glyoxylate Production
To further improve glyoxylate production, the glyoxylate shunt was enhanced by overexpressing gltA and aceA using plasmid pSR13. Overall, glyoxylate production increased with an increase in the number of disrupted genes. The maximum titers of glyoxylate in E. coli EG_5 and EG_6 were 0.12 ± 0.00 and 1.15 ± 0.02 g/L, respectively (Figure 3). Among these, E. coli EG_6 showed better glyoxylate production than the other E. coli EG_5. Escherichia coli EG_6 showed decreased glyoxylate content at the end of the cultivation period. This trend may have originated from the potential conversion of glyoxylate into other compounds, as discussed above. Escherichia coli EG_6 showed decreased glyoxylate content at the end of the cultivation period. This trend may have originated from the potential conversion of glyoxylate into other compounds, as discussed above.
(A) (B) Among the evaluated strains, E. coli EG_6 showed lower cell growths at the end of the cultivation, corresponding to 0.43 ± 0.01 g-cell/L than E. coli EG_5 (0.87 ± 0.01 g-cell/L). This was consistent with the lower growth rates of EG_4 than that of E. coli BW25113, as E. coli EG_6 shared the same disrupted genes as E. coli EG_4. The overexpression of gltA and aceA showed no negative effect on the growth rate of E. coli.
The overexpression of gltA and aceA further benefits E. coli BW25113 in the improvement of the titer and specific productivity of glyoxylate (Table 2). Among these strains, E. coli EG_6 showed the highest specific productivity at 2.88 ± 0.02 (g/g-cell). Deng et al. (2015) also reported that the production of glycolate from glyoxylate via a glyoxylate shunt could be reinforced by the overexpression of aceA [42].
We investigated the antagonistic effects of gene overexpression on acetate production. All of the strains equipped with the glyoxylate shunt reinforced by gene overexpression produced less acetate; EG_5 and EG_6 produced acetate at 2.62 ± 0.01 and 3.61 ± 0.17 g/L, respectively, at 72 h. These titers were lower than those of BW25113, EG_1, EG_2, and EG_3 (Figures 2 and 3). Renilla et al. (2012) showed that the disruption of genes involved in the glyoxylate shunt affects E. coli growth, and that the reinforced glyoxylate shunt can reduce acetate [43].
As shown in Figure 3, the residual glyoxylate at the end of fermentation of the overexpressed strains EG_5 and EG_6 was higher than that of BW25113 and EG_4, respectively, indicating that the overexpression of gltA and aceA not only enhanced glyoxylate production but also limited glyoxylate assimilation. In particular, although the majority of the produced glyoxylate was consumed in the fermentation of E. coli EG_4 (Figure 2), more than 70% glyoxylate concentration remained at the end of E. coli EG_6 fermentation with the residue glyoxylate noted at 0.74 ± 0.06 g/L compared to that of E. coli EG_4 with 0.28 ± 0.09 g/L (Figure 2). These results demonstrate that the overexpression of gltA and aceA is an efficient method to overcome glyoxylate consumption.
Similarly, the glyoxylate productivities of all strains were significantly improved compared to non-overexpressed strains with specific productivities of glyoxylate in EG_5 and EG_6 at 0.14 ± 0.03 and 2.88 ± 0.02 g/g-cell, respectively (Table 2). Among the evaluated strains, E. coli EG_6 showed lower cell growths at the end of the cultivation, corresponding to 0.43 ± 0.01 g-cell/L than E. coli EG_5 (0.87 ± 0.01 g-cell/L). This was consistent with the lower growth rates of EG_4 than that of E. coli BW25113, as E. coli EG_6 shared the same disrupted genes as E. coli EG_4. The overexpression of gltA and aceA showed no negative effect on the growth rate of E. coli.
The overexpression of gltA and aceA further benefits E. coli BW25113 in the improvement of the titer and specific productivity of glyoxylate (Table 2). Among these strains, E. coli EG_6 showed the highest specific productivity at 2.88 ± 0.02 (g/g-cell). Deng et al. (2015) also reported that the production of glycolate from glyoxylate via a glyoxylate shunt could be reinforced by the overexpression of aceA [42].
We investigated the antagonistic effects of gene overexpression on acetate production. All of the strains equipped with the glyoxylate shunt reinforced by gene overexpression produced less acetate; EG_5 and EG_6 produced acetate at 2.62 ± 0.01 and 3.61 ± 0.17 g/L, respectively, at 72 h. These titers were lower than those of BW25113, EG_1, EG_2, and EG_3 (Figures 2 and 3). Renilla et al. (2012) showed that the disruption of genes involved in the glyoxylate shunt affects E. coli growth, and that the reinforced glyoxylate shunt can reduce acetate [43].
As shown in Figure 3, the residual glyoxylate at the end of fermentation of the overexpressed strains EG_5 and EG_6 was higher than that of BW25113 and EG_4, respectively, indicating that the overexpression of gltA and aceA not only enhanced glyoxylate production but also limited glyoxylate assimilation. In particular, although the majority of the produced glyoxylate was consumed in the fermentation of E. coli EG_4 (Figure 2), more than 70% glyoxylate concentration remained at the end of E. coli EG_6 fermentation with the residue glyoxylate noted at 0.74 ± 0.06 g/L compared to that of E. coli EG_4 with 0.28 ± 0.09 g/L (Figure 2). These results demonstrate that the overexpression of gltA and aceA is an efficient method to overcome glyoxylate consumption.
Similarly, the glyoxylate productivities of all strains were significantly improved compared to non-overexpressed strains with specific productivities of glyoxylate in EG_5 and EG_6 at 0.14 ± 0.03 and 2.88 ± 0.02 g/g-cell, respectively (Table 2).

Effects of pyc Expression on Glyoxylate Production in E. coli
Recent studies have demonstrated an improvement in chemical production via the glyoxylate shunt by improving the flow of pyruvate into the TCA cycle. Here, we reinforced glyoxylate production by expressing the heterologous gene pyc from C. glutamicum, which helps to convert pyruvate to oxaloacetate. Figure 4A shows the fermentation profile of E. coli EG_7 with the expression of pyc using pLB6 plasmid.

Effects of pyc Expression on Glyoxylate Production in E. coli
Recent studies have demonstrated an improvement in chemical production via the glyoxylate shunt by improving the flow of pyruvate into the TCA cycle. Here, we reinforced glyoxylate production by expressing the heterologous gene pyc from C. glutamicum, which helps to convert pyruvate to oxaloacetate. Figure 4A shows the fermentation profile of E. coli EG_7 with the expression of pyc using pLB6 plasmid. Although the expression of pyc slowed E. coli growth, this strategy helped increase the glyoxylate concentration to 1.31 ± 0.11 g/L. The positive effects of pyc expression on succinate [26], malate [27], and fumarate production [28] have been demonstrated in recent studies. These studies concluded that the expression of pyc could activate the glyoxylate shunt of E. coli and improve the production of these organic acids. As shown in Figure 4A, EG_7, equipped with pyc, grew more slowly than E. coli EG_6 and stopped using glucose at the end of fermentation.
To improve the growth of E. coli EG_7, we conducted fermentation in a medium supplemented with 1 g/L yeast extract ( Figure 4B). As a result, E. coli EG_7 rapidly grew with yeast extract supplementation and the final cell biomass reached 0.61 ± 0.01 g-cell/L compared to a cell biomass of 0.30 ± 0.04 g-cell/L cells in the non-supplemented medium. In addition, glyoxylate production reached the highest at 2.42 ± 0.00 g/L with specific productivity at 4.22 ± 0.09 g/g-cell (Table 2) at 48 h. In this study, glucose was completely consumed after 144 h. Yeast extract can support E. coli growth as a source of amino acids.
In theory, glyoxylate production from glucose via the glyoxylate shunt follows the equation: 1 glucose + 6 H2O = 2 glyoxylate + 2 CO2 + 20 H + 1 O2. Therefore, 1 g of glucose can generate 0.82 g of glyoxylate under the optimal condition. In Figure 4B and Table 2, the highest glyoxylate concentration was 2.42 ± 0.00 g/L at 48 h with 1.72 ± 0.04 g/L acetate, 0.57 ± 0.01 g/L cells, and 0.05 ± 0.01 g/L formate from 7.85 ± 0.07 g/L consumed glucose. Malate and succinate were not detected in the cultures (data not shown). The total quantifiable products accounted for 60.6% of the consumed glucose mass, reaching 37.6% theoretical yield in mass and 0.30 mol-glyoxylate/mol-glucose. It was suggested that the assimilation of glyoxylate and the generation of non-detected metabolites, such as CO2, caused this decrease in the yield. The production from glucose in this study was higher than the yield from xylose, with a 30.0% yield compared to the theory and 0.30 mol-glyoxylate/mol-xylose in the previous study [20]. On the other hand, the related study reported that malate was accumulated when boosting the glyoxylate shunt [44]. In contrast, Although the expression of pyc slowed E. coli growth, this strategy helped increase the glyoxylate concentration to 1.31 ± 0.11 g/L. The positive effects of pyc expression on succinate [26], malate [27], and fumarate production [28] have been demonstrated in recent studies. These studies concluded that the expression of pyc could activate the glyoxylate shunt of E. coli and improve the production of these organic acids. As shown in Figure 4A, EG_7, equipped with pyc, grew more slowly than E. coli EG_6 and stopped using glucose at the end of fermentation.
To improve the growth of E. coli EG_7, we conducted fermentation in a medium supplemented with 1 g/L yeast extract ( Figure 4B). As a result, E. coli EG_7 rapidly grew with yeast extract supplementation and the final cell biomass reached 0.61 ± 0.01 g-cell/L compared to a cell biomass of 0.30 ± 0.04 g-cell/L cells in the non-supplemented medium. In addition, glyoxylate production reached the highest at 2.42 ± 0.00 g/L with specific productivity at 4.22 ± 0.09 g/g-cell (Table 2) at 48 h. In this study, glucose was completely consumed after 144 h. Yeast extract can support E. coli growth as a source of amino acids.
In theory, glyoxylate production from glucose via the glyoxylate shunt follows the equation: 1 glucose + 6 H 2 O = 2 glyoxylate + 2 CO 2 + 20 H + 1 O 2 . Therefore, 1 g of glucose can generate 0.82 g of glyoxylate under the optimal condition. In Figure 4B and Table 2, the highest glyoxylate concentration was 2.42 ± 0.00 g/L at 48 h with 1.72 ± 0.04 g/L acetate, 0.57 ± 0.01 g/L cells, and 0.05 ± 0.01 g/L formate from 7.85 ± 0.07 g/L consumed glucose. Malate and succinate were not detected in the cultures (data not shown). The total quantifiable products accounted for 60.6% of the consumed glucose mass, reaching 37.6% theoretical yield in mass and 0.30 mol-glyoxylate/mol-glucose. It was suggested that the assimilation of glyoxylate and the generation of non-detected metabolites, such as CO 2 , caused this decrease in the yield. The production from glucose in this study was higher than the yield from xylose, with a 30.0% yield compared to the theory and 0.30 mol-glyoxylate/mol-xylose in the previous study [20]. On the other hand, the related study reported that malate was accumulated when boosting the glyoxylate shunt [44]. In contrast, the production level of malate was negligible in this study. This may be due to the disruption of glcB and aceB involving the conversion of glyoxylate to malate.
In a previous study on pyc expression in E. coli for the production of succinate, the increase in succinate production was due to an increase in oxaloacetate concentration, which accelerates glyoxylate production [26]. To understand the effects of pyc expression on E. coli metabolism, we quantified intracellular oxaloacetate in E. coli EG_6 and EG_7 during fermentation ( Figure 5). As a result, the highest intracellular oxaloacetate was noted in E. coli EG_7 at 327.4 ± 46.5 mg/g-cell, which is higher than that of E. coli EG_6 (51.5 ± 10.1 mg/g-cell). During the first 24 h of induction by isopropyl-β-Dthiogalactopyranoside, the intracellular oxaloacetate content was undetectable in both E. coli EG_6 and EG_7. The accumulation of intracellular oxaloacetate was quickly improved after gene induction by isopropyl-β-D-thiogalactopyranoside, which indicated the effects of the introduced pyc gene on oxaloacetate production in E. coli. As shown in Figure 4B, the cell biomass and glyoxylate concentration steadily increased after 48 h, followed by a decrease in the concentration of intracellular oxaloacetate ( Figure 5), indicating that the assimilation of intracellular oxaloacetate improved glyoxylate production in E. coli. the production level of malate was negligible in this study. This may be due to the disruption of glcB and aceB involving the conversion of glyoxylate to malate. In a previous study on pyc expression in E. coli for the production of succinate, the increase in succinate production was due to an increase in oxaloacetate concentration, which accelerates glyoxylate production [26]. To understand the effects of pyc expression on E. coli metabolism, we quantified intracellular oxaloacetate in E. coli EG_6 and EG_7 during fermentation ( Figure 5). As a result, the highest intracellular oxaloacetate was noted in E. coli EG_7 at 327.4 ± 46.5 mg/g-cell, which is higher than that of E. coli EG_6 (51.5 ± 10.1 mg/g-cell). During the first 24 h of induction by isopropyl-β-D-thiogalactopyranoside, the intracellular oxaloacetate content was undetectable in both E. coli EG_6 and EG_7. The accumulation of intracellular oxaloacetate was quickly improved after gene induction by isopropyl-β-D-thiogalactopyranoside, which indicated the effects of the introduced pyc gene on oxaloacetate production in E. coli. As shown in Figure 4B, the cell biomass and glyoxylate concentration steadily increased after 48 h, followed by a decrease in the concentration of intracellular oxaloacetate ( Figure 5), indicating that the assimilation of intracellular oxaloacetate improved glyoxylate production in E. coli. Figure 5. Intracellular oxaloacetate produced by Escherichia coli EG_6 and EG_7. Escherichia coli EG_6 and EG_7 were cultivated in 100 mL of M9 medium supplemented with 10 g/L of glucose and 1 g/L of yeast extract. The experiment was performed in duplicate and the average is presented with error bars. Figure 4A also indicates a lower growth of E. coli EG_7, as the highest cell biomass reached 0.30 ± 0.04 g-cell/L, 30% lower than that of E. coli EG_6. One possible reason for poor cell growth is the accumulation of intracellular oxaloacetate, consequently causing the feedback inhibition of malate dehydrogenase, as reported in a recent study [45]. Oxaloacetate is considered a branch point that controls amino acid synthesis via lysine metabolism and the TCA cycle via the phosphoenolpyruvate-pyruvate-oxaloacetate node in bacteria [3]. Furthermore, the expression of pyc in E. coli was suggested to generate competitive effects on the production of acetyl-CoA from pyruvate. A possible solution is the overexpression of acetyl-CoA synthetase (acs) to recover acetyl-CoA from acetate since acetate concentrations were remarkable during the fermentation (Figure 4). Nevertheless, the expression of pyc can be a valuable technique for improving glyoxylate production.
As discussed above, the assimilation of glyoxylate was noted during the stationary phase of E. coli growth, which may be due to a decrease in TCA cycle operation and an increase in the accumulation of pyruvate [46], a precursor for eda function in reducing glyoxylate. Figure 5. Intracellular oxaloacetate produced by Escherichia coli EG_6 and EG_7. Escherichia coli EG_6 and EG_7 were cultivated in 100 mL of M9 medium supplemented with 10 g/L of glucose and 1 g/L of yeast extract. The experiment was performed in duplicate and the average is presented with error bars. Figure 4A also indicates a lower growth of E. coli EG_7, as the highest cell biomass reached 0.30 ± 0.04 g-cell/L, 30% lower than that of E. coli EG_6. One possible reason for poor cell growth is the accumulation of intracellular oxaloacetate, consequently causing the feedback inhibition of malate dehydrogenase, as reported in a recent study [45]. Oxaloacetate is considered a branch point that controls amino acid synthesis via lysine metabolism and the TCA cycle via the phosphoenolpyruvate-pyruvate-oxaloacetate node in bacteria [3]. Furthermore, the expression of pyc in E. coli was suggested to generate competitive effects on the production of acetyl-CoA from pyruvate. A possible solution is the overexpression of acetyl-CoA synthetase (acs) to recover acetyl-CoA from acetate since acetate concentrations were remarkable during the fermentation (Figure 4). Nevertheless, the expression of pyc can be a valuable technique for improving glyoxylate production.
As discussed above, the assimilation of glyoxylate was noted during the stationary phase of E. coli growth, which may be due to a decrease in TCA cycle operation and an increase in the accumulation of pyruvate [46], a precursor for eda function in reducing glyoxylate.
In this study, pH-controlled jar fermentation was conducted; however, the glyoxylate titer was not improved (data not shown). One possible solution to overcome the assimilation of glyoxylate is to apply fermentation optimization to extend the cell growth phase using DO-stat fermentation [47]. This approach limits the operation of the glycolytic pathway, which controls pyruvate accumulation [48]. In addition, DO-stat cultivation can effectively control the pH and the production of organic acids, such as acetate, lactate, and formate, which enable higher cell growth [47] and glyoxylate production.
Another disadvantage of the fermentation in the present study was the accumulation of acetate as the major by-product, with the final content generated by E. coli EG_7 at 4.42 ± 0.18 g/L. Acetate, the major by-product, accumulates at a high concentration in the cultures of engineered E. coli. According to recent studies, overflow metabolism originating from a high pyruvate content results in high acetate production and can affect E. coli growth [38]. A possible solution for acetate accumulation is the overexpression of the acetyl-CoA synthetase gene (acs), because Lin et al. (2006) showed that acs overexpression results in the reduction of acetate in E. coli [49]. Another approach was reported by Negrete et al. (2013) using the overexpression of sgrS, which works as a small RNAi to slow down the glucose uptake rate and could decrease acetate by 80% in E. coli [50]. These results provide promising insights into improving glyoxylate production from glucose in future studies.

Conclusions
This study demonstrated an improvement in glyoxylate production from glucose in E. coli BW25113. Glyoxylate concentrations of up to 5 g/L did not affect E. coli growth. The disruption of aceB, glcB, gcl, and ycdW, along with the overexpression of gltA and aceA, improved the glyoxylate production titer and specific productivity at 1.15 ± 0.02 g/L and 2.88 ± 0.02 g/g-cell, respectively, which were much higher than those of the wild-type (0.01 ± 0.00 g/L and 0.01 ± 0.00 g/g-cell, respectively). In addition, the expression of pyc significantly enhanced glyoxylate production to achieve the highest titer and specific productivity at 2.42 ± 0.00 g/L and 4.42 ± 0.09 g/g-cell, respectively, which is the highest reported to date.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/fermentation9060534/s1, Figure S1: Effects of glyoxylate concentrations on Escherichia coli BW25113 growth; Figure S2: HPLC chromatograms of glyoxylate after derivatization with phenylhydrazine; Figure S3: HPLC chromatograms of oxaloacetate; Table S1: Sequence information of primers used in this study.