Introduction

1,3-Propanediol (1,3-PD) is a valuable and important chemical that is widely used in polyester, cosmetic, and pharmaceutical industries (Durgapal et al. 2014; Szymanowska-Powalowska and Kubiak 2015). Because of the large amount of low-cost raw glycerol produced during biodiesel production, the biological production of 1,3-PD from glycerol has attracted significant interest recently (Khan et al. 2013; Szymanowska-Powalowska and Bialas 2014). Klebsiella pneumoniae, a facultative anaerobe, has the ability to produce high yields of 1,3-PD from glycerol (Celinska 2012; Yen et al. 2014).

The metabolic pathway responsible for the microbial production of 1,3-PD from glycerol has been well studied in K. pneumoniae (Kumar et al. 2012; Seo et al. 2009; Skraly et al. 1998; Zhuge et al. 2010). Generally, glycerol is converted through reductive and oxidative pathways (Fig. 1), the reductive branch leading to 1,3-PD production, and the oxidative branch providing the reducing power for 1,3-PD synthesis; during the fermentation process, the formation of by-products derived from oxidative branch, including acetate, lactate, and 2,3-butanediol (2,3-BD), leads to the decreases of carbon flux and reducing power towards 1,3-PD synthesis (Huang et al. 2012; Petrov and Petrova 2009; Tang et al. 2009). To improve the production of 1,3-PD in K. pneumoniae, various efforts have been made to delete the genes responsible for the formation of these by-products. It was found that lactate formation was blocked by deleting the ldhA gene (Xu et al. 2009; Zheng et al. 2008), and that 2,3-BD formation was also completely abolished by deleting the genes responsible for its synthesis (Cui et al. 2014). Because blocking 2,3-BD formation results in an extreme growth defect and a decrease in 1,3-PD production (Cui et al. 2014; Zhu et al. 2015), many high 1,3-PD producers are based on lactate-deficient (ldhA-deleted) K. pneumoniae strains, although the deletion of ldhA is actually more beneficial for 2,3-BD production than 1,3-PD production (Durgapal et al. 2014; Kumar et al. 2013; Xu et al. 2009).

Fig. 1
figure 1

Metabolic pathways of glycerol metabolism in K. pneumoniae. GLY glycerol, PYR pyruvate, PEP phosphoenolpyruvate, 3-HPA 3-hydroxypropinaldehyde, 1,3-PD 1,3-propanediol, 2,3-BD 2,3-butanediol, LAC lactate, ACE acetate, SUC succinate, α-KG α-ketoglutarate, TCA tricarboxylic acids cycle, Acetyl-P acetylphosphate, pta encoding phosphotransacetylase, ackA encoding acetate kinase, poxB encoding pyruvate oxidase, ldhA encoding lactate dehydrogenase, acs encoding acetyl-CoA synthetase

Full size image

Compared with lactate and 2,3-BD, few reports have focused on reducing the other main by-product, acetate, which is commonly considered to be more toxic than lactate and 2,3-BD (Celinska 2010). Recently, Lee et al. (2014) reported that deleting the pta gene decreased acetate formation in a both 2,3-butanediol- and lactate-deficient K. pneumoniae strain. In Escherichia coli, acetate is synthesized by two major routes: the first occurs via the actions of phosphotransacetylase (Pta, encoded by the pta gene) and acetate kinase (AckA, encoded by the ackA gene), which convert acetyl-CoA to acetate and one molecule of adenosine triphosphate (ATP); the other route is via pyruvate oxidase (PoxB, encoded by the poxB gene), which oxidizes pyruvate to generate acetate and one molecule of CO2 (Ashok et al. 2011). Generally, the Pta-AckA pathway is the primary one, while deletion of poxB seems to affect central carbon metabolism, such as the tricarboxylic acid (TCA) cycle. Deleting both pathways results in decreased cell growth and acetate formation in E. coli (De Mey et al. 2007; Li et al. 2012). On the other hand, acetate can be assimilated by E. coli, mainly through acetyl-CoA synthetase (ACS, encoded by the acs gene) to generate acetyl-CoA. The recycling of acetate by ACS preserves the pool of free reduced CoA (CoA-SH) and mitigates acetate overflow in aerobic E. coli cultures (Dittrich et al. 2005; Peebo et al. 2014). Because of its close genetic relatedness to E. coli, K. pneumoniae is expected to have the same regulatory mechanism for acetate synthesis. However, to date, no detailed study has focused on the acetate synthesis pathway, especially, its effect on 1,3-PD production and physiology of K. pneumoniae.

In this study, the effects of blocking the two acetate synthesis pathways, PoxB and Pta-AckA, on cell growth and 1,3-PD production in K. pneumoniae were investigated under different culture conditions. It was shown that deletion of poxB severely inhibited cell growth. Interestingly, the growth defect of the poxB mutant was alleviated by further deleting the Pta-AckA acetate synthesis pathway. With comprehensive analysis, the roles of PoxB and Pta-AckA in K. pneumoniae were represented. And, as a result, a strategy for efficient production of 1,3-PD by directly blocking acetate synthesis pathways was proposed.

Materials and methods

Strains, plasmids, primers, media, and culture conditions

Strains, plasmids, and primers are listed in Table 1. The lactate-deficient K. pneumoniae strain KG2 (Zhu et al. 2015), which was derived from K. pneumoniae strain KG (CCTCC M2014574) by deleting the ldhA gene, is a high 1,3-PD producer, and it was used as the parent strain in this study.

Table 1 Strains, plasmids, and primers used in this study
Full size table

Batch culture in 250-mL shake flask containing 50 mL of medium with 5 % inoculum for the assessment of mutants was performed at 37 °C for 24 h with shaking at 200 rpm under aerobic and anaerobic conditions. Pre-culture was cultivated in 10-mL test tube with 5 mL Luria-Bertani broth (LB) (37 °C, 18 h). The shake flask medium (pH 7.0) contained 60 g/L glycerol (or glucose), 1 g/L KCl, 2.2 g/L KH2PO4·2H2O, 4 g/L (NH4)2SO4, 0.4 g/L MgSO4·7H2O, and 2 g/L yeast extract. To test the effect of CO2 on bacterial growth, 1 g/L NaHCO3 was added to the medium. For aerobic cultures, flasks were covered with eight layers of gauze to permit air penetration; for anaerobic cultures, the air in the flasks was replaced with nitrogen gas before cultivation, and then, the flasks were plugged with a gas-impermeable rubber stopper.

Flask-scale batch fermentation for 1,3-PD production was performed micro-aerobically at 37 °C for 24 h with shaking at 50 rpm; the medium and other procedure were the same as those for aerobic flask culture mentioned above. Fed-batch fermentation for 1,3-PD production was performed in a 5-L stirred reactor (Shanghai Bailun Bio-technology Co., Ltd., Shanghai, China) with a working volume of 2 L. The reactor was inoculated with 5 % (v/v) of an overnight culture, and then, the bacteria were incubated under micro-aerobic conditions at 37 °C for 30 h. The aeration rate was 0.1 vvm, and the agitation speed was 140 rev/min. The initial glycerol concentration was 40 g/L, and it was maintained between 15 and 25 g/L by continuously feeding with 800 g/L of glycerol into the reactor throughout the fermentation process. The pH was maintained at 6.8 by automatically adding 50 % (w/v) KOH. The fermentation medium was the same as that reported by Cui et al. (2014).

Construction of poxB and poxB-pta-ackA mutant strains using lambda Red recombination

The poxB and pta-ackA genes were deleted by lambda Red recombination (Yamamoto et al. 2009). Gene disruption cassettes with a kanamycin-resistance marker and a flippase recognition site were amplified by two-step polymerase chain reactions from the chromosomal DNA of K. pneumoniae strain KG2 and the pKD4 vector, respectively, using the primers listed in Table 1. To construct the poxB deletion mutant, pKD46-Tc was first transformed into the KG2 strain, followed by transformation with the poxB disruption cassette. The recombinant strain was screened by growth on LB plates supplemented with kanamycin at 37 °C for 24 h. Finally, pCP20-Tc was transformed into the recombinant strain to remove the kanamycin-resistance gene. Cells were cultured overnight at 42 °C, diluted, and plated onto solid LB to obtain single colonies, which were further screened for loss of kanamycin and tetracycline hydrochloride resistance; the resulting poxB mutant strain was named KG4 (Table 1). Using the same procedure, the pta-ackA genes were deleted from strain KG4; the poxB-pta-ackA mutant strain was named KG5 (Table 1).

Analytical methods

The optical density (OD) of bacterial cultures was related to the dry cell weight by an experimentally determined calibration curve. The OD at 620 nm of the bacterial cultures was measured after appropriate dilution. 1,3-PD and 2,3-BD concentrations in the fermentation samples were quantified using gas chromatography. The analysis conditions included N2 as the carrier gas, a detector temperature of 270 °C, and a column temperature of 120 °C. Other metabolites present in the fermentation broths were quantified using a high-performance liquid chromatography system equipped with a 2487 Dual-Wavelength Absorbance Detector (Waters Corporation, Milford, MA, USA) and a Plastisil ODS column (AQ-C18, 5 μm, 250 × 4.6 mm; Welch Material, Inc., Ellicott City, MD, USA) at a flow rate of 0.8 mL min−1 and a column temperature of 65 °C. The mobile phase was 0.005 M H2SO4. Samples were filtered through 0.45-μm filters before analysis.

Results

Inhibitory effect of the poxB deletion on cell growth

Because lactate-deficient K. pneumoniae strains are commonly used for industrial applications, the lactate-deficient strain KG2 was used as the parent strain in our study. The growth and acetate formation of the poxB mutant strain KG4 and the parent strain KG2 after 24 h of flask-scale cultivation under different conditions are shown in Fig. 2. An inhibitory effect of the poxB deletion on bacterial growth was found. Compared with the KG2 strain, the growth of the KG4 strain significantly decreased, regardless of the carbon source used and whether the cultures were grown under anaerobic or aerobic conditions. The KG4 strain exhibited larger growth defects under anaerobic conditions (64 and 55 % of the cell concentration of the KG2 strain in glycerol- and glucose-based media, respectively), while its cell concentrations were 70 and 76 % of those of the KG2 strain in glycerol and glucose-based media, respectively, under aerobic culture conditions (Fig. 2a).

Fig. 2
figure 2

Cell growth (a) and acetate formation (b) of KG2 and KG4 after 24 h of flask cultivation under anaerobic and aerobic conditions. Black bars KG2 in glycerol-based medium, cross-hatched bars KG4 in glycerol-based medium, white bars KG2 in glucose-based medium, hatched bars KG4 in glucose-based medium. Error bars represent the standard deviations from three independent experiments

Full size image

Under anaerobic culture conditions, although the total acetate formation in strain KG4 decreased, the acetate formation per cell (g/g) was 0.17 and 0.42 in glycerol- and glucose-based media, respectively, which were almost the same as those in the KG2 strain, indicating that deleting poxB did not affect acetate formation. In contrast, compared with strain KG2, acetate formation in the KG4 strain was dramatically increased under aerobic conditions; meanwhile, more acetate (9.84 g/L) accumulated in a glycerol-based medium at the end of the cultivation (Fig. 2b).

In addition to acetate formation, another role of poxB is to generate CO2. To assess the effect of CO2 on the growth defect of the KG4 strain, 1 g/L of NaHCO3, an efficient source of CO2 for bacterial growth (Kozliak et al. 1995; Repaske and Clayton 1978), was added to the medium. As shown in Fig. 3, the addition of NaHCO3 had no effect on the growth of the KG2 strain under all culture conditions, and it did not affect the growth of the KG4 strain under aerobic culture conditions. However, the addition of NaHCO3 stimulated the anaerobic growth of strain KG4 in both glycerol- and glucose-based media, indicating that the growth defect of the KG4 strain was probable due to the shortage of CO2 under anaerobic conditions.

Fig. 3
figure 3

The effect of NaHCO3 on cell growth of KG2 and KG4 at the end of flask cultivation under anaerobic and aerobic conditions in glycerol-based medium (a) or glucose-based medium (b). Black bars KG2 without NaHCO3, white bars KG4 without NaHCO3, cross-hatched bars KG2 with NaHCO3, hatched bars KG4 with NaHCO3. Error bars represent the standard deviations from three independent experiments

Full size image

Recovery of cell growth in the poxB mutant by deleting pta and ackA

In contrast to the adverse effects of the poxB deletion, deleting the other main acetate pathway, Pta-AckA, in strain KG2 had almost no effect on cell growth and the formation of metabolites, such as acetate and 1,3-PD, in all of the test conditions (data not shown). However, according to the aforementioned experiments, deleting poxB resulted in an extreme accumulation of acetate under aerobic culture conditions, which could be the reason for the aerobic growth defect of strain KG4. Thus, the pta and ackA genes were deleted from strain KG4. The growth and acetate formation of the resulting mutant, KG5, in a glycerol-based medium after 24 h of flask cultivation are shown in Table 2. Compared with strain KG4, acetate formation in strain KG5 dramatically decreased to 1.85 g/L (19 % of that of the KG4 strain) under aerobic conditions (Table 2). As a result of the decreased accumulation of acetate, the concentration of KG5 cells significantly increased at the end of aerobic cultivation, almost reaching the level of the parent strain KG2 (Table 2). More interestingly, it was shown that deleting the pta-ackA genes alleviated the growth defect of the poxB mutant, even under aerobic condition. At the end of anaerobic cultivation, the cell concentration of KG5 was the same as that of KG2 (Table 2). Such recovery of cell growth in the KG5 strain under both aerobic and anaerobic conditions was confirmed in a glucose-based medium (data not shown).

Table 2 Cell growth and production of KG2, KG4, and KG5 in glycerol-based medium under aerobic and anaerobic conditions at 24 h in shake flask
Full size table

Although the growth of strain KG5 was almost the same as that of the KG2 strain in aerobic and anaerobic cultures, the varieties of metabolites formed, especially those involved in the TCA cycle, differed (Table 2). The levels of α-ketoglutarate and succinate in strain KG5 were only 62 and 41 %, respectively, of those in strain KG2 at the end of aerobic cultivation. In contrast, the levels of α-ketoglutarate and succinate substantially increased in strain KG5 at the end of anaerobic cultivation, as they were approximately 211 and 53 %, respectively, greater than those in strain KG2. Interestingly, the highest level of 1,3-PD production (20.8 g/L) at the end of anaerobic cultivation was found in strain KG5, which was 15 % greater than that in strain KG2 (Table 2).

Pyruvate accumulation is usually reported to be the reason for the growth defects in Pta-AckA-deficient E. coli strains (Chang et al. 1999; Liao et al. 1996). Based on our experiments, deleting poxB resulted in a high accumulation of pyruvate in strain KG4 under anaerobic culture condition, while strain KG5 did not accumulate high levels of pyruvate at the end of anaerobic cultivation (Table 2). Because of the importance of the pyruvate node in metabolism, variations in the level of pyruvate in strain KG5 during flask cultures were further analyzed. As shown in Fig. 4a, pyruvate accumulation in strain KG5 was found during the early stage of flask cultivation. The peak values of pyruvate occurred after approximately 6 h and reached 0.72 g/L (aerobic conditions) and 0.31 g/L (anaerobic conditions). The growth of strain KG5 during flask cultivation is also shown in Fig. 5. Compared with the KG2 strain, the growth of strain KG5 was retarded during the early stage of flask cultivation. The growth retardation was greater under aerobic conditions. At 10 h of cultivation, the cell concentration of strain KG5 was 66 % of that of strain KG2, while under anaerobic conditions, it reached 77 % of that of the KG2 strain (Fig. 4b).

Fig. 4
figure 4

Pyruvate formation (a) and cell growth (b) of KG2 (open symbol) and KG5 (solid symbol) during the shaker flask cultivation. Aerobic (square) and anaerobic (triangle). Date points are averages of three identical experiments

Full size image
Fig. 5
figure 5

Cell growth and metabolites formation in KG2 (open symbol) and KG5 (solid symbol) during the fed-batch fermentation process under micro-aerobic condition. a 1,3-PD (triangle), glycerol (circle), and cell growth (square); b acetate (square) and pyruvate (triangle); c α-ketoglutarate (square), 2,3-BD (circle), and succinate (triangle). Date points are averages of three identical experiments

Full size image

1,3-PD production by the poxB-pta-ackA mutant KG5 during fed-batch fermentation

Compared with anaerobic fermentation, micro-aerobic fermentation for 1,3-PD production by K. pneumoniae is commonly used recently due to its high process productivity and convenient industrial application (Chen et al. 2003; Durgapal et al. 2014). As expected, the strain of KG5 also showed the high ability for 1,3-PD production under micro-aerobic condition in a flask batch process (Table 3). To confirm the ability of strain KG5 to produce large amounts of 1,3-PD, fed-batch fermentations under micro-aerobic conditions were performed in a 5-L bioreactor. The detailed fermentation results are shown in Fig. 5 and Table 3. During the fed-batch fermentation, the cell concentration increased in the early stage (0–16 h) and decreased in the later stage (16–30 h). Compared with the KG2 strain, the growth of strain KG5 was lower in the early stage, although a small decrease in the cell concentration was found in the later stage. In contrast to growth, a high level of 1,3-PD production was observed in strain KG5, especially in the later stage of fermentation (Fig. 5a). At the end of the fed-batch fermentation, 1,3-PD production by strain KG5 reached 76.8 g/L, which was 15 % greater than that of strain KG2.

Table 3 Fermentation results of KG2 and KG5 under micro-aerobic condition in flask batch and bioreactor fed-batch processes
Full size table

Deleting the poxB, pta, and ackA genes resulted in a significant decrease in acetate formation during fed-batch fermentation. Acetate production by strain KG5 was 1.58 g/L, only 23 % of that of strain KG2 at the end of the fed-batch fermentation. In contrast with strain KG2, greater accumulations of α-ketoglutarate and succinate were observed in strain KG5 during fed-batch fermentation, and a high accumulation of pyruvate during the early stage of fermentation was also found. As shown in Table 3, strain KG5 completely consumed 140.1 g/L of glycerol within 30 h of fed-batch fermentation, which was less than that consumed by strain KG2 (141.3 g/L). Meanwhile, the formation of the dominant by-product, 2,3-BD, was almost unchanged; thus, the total yield of 1,3-PD from glycerol by strain KG5 in fed-batch fermentation reached 0.66 (mol/mol), a 16 % increase compared with strain KG2 (Table 3).

Discussion

Acetate, which is highly toxic, is a major by-product that is excreted during the formation of 1,3-PD from glycerol in K. pneumoniae (Celinska 2010; Zeng et al. 1994). Although a low level of acetate seems to have a positive effect on 1,3-PD fermentation (Grahame et al. 2013; Lu et al. 2013), acetate overflow could lead to inhibition of cell growth and a decrease in 1,3-PD production (Cui et al. 2014; He et al. 2013; Szymanowska-Powalowska and Kubiak 2015). In this study, the possibility of improving 1,3-PD production by directly blocking acetate synthesis pathways was investigated. Our results indicated that, in addition to decreasing acetate formation, deleting the poxB, pta, and ackA genes significantly improved 1,3-PD production in K. pneumoniae.

PoxB catalyzes the oxidative decarboxylation of pyruvate to acetate. In E. coli, pyruvate is mainly converted into acetyl-CoA by the dehydrogenase complex (PDHC) and/or pyruvate formate-lyase (PFL), so PoxB generally has been regarded as non-essential (Chang et al. 1994; Grabau and Cronan 1984). Although previous studies have elucidated some functions of PoxB, deleting poxB had almost no effect on aerobic growth and acetate formation in E. coli (Chang and Cronan 1983; Phue et al. 2010). In our experiments, an extreme growth defect occurred in the poxB mutant, indicating that, in contrast to E. coli, PoxB plays a more important role in K. pneumoniae. As the recycling of PoxB-generated acetate to acetyl-CoA via ACS requires additional energy, the involvement of PoxB in pyruvate oxidation seems wasteful. However, PoxB was actually confirmed to contribute to aerobic growth in E. coli (Abdel-Hamid et al. 2001). Thus, the sizeable acetate flux generated by PoxB could be an adaptation to ensure good growth in K. pneumoniae.

Although the deletion of poxB inhibited growth in both aerobic and anaerobic culture conditions, the reasons for the growth defects differ. Under aerobic conditions, the extra carbon flux obtained by blocking PoxB should directly flow to acetyl-CoA via PDHC and/or PFL. Meanwhile, the TCA cycle and acetate assimilation were reported to be suppressed in a poxB mutant (Kumari et al. 2000; Li et al. 2006; Martinez-Gomez et al. 2012; Phue and Shiloach 2004); thus, a high concentration of acetate accumulated in the poxB mutant, probably through the Pta-AckA pathway. Compared with the glucose-based medium, greater acetate accumulation in the poxB mutant correlated with a more severe aerobic growth defect in the glycerol-based medium. Consequently, the accumulation of acetate in aerobic cultures could explain the growth defect of the poxB mutant. Carbon-dependent accumulation of acetate in the poxB mutant may be caused by regulatory differences in acetate assimilation (Wolfe 2005). The suppression of the TCA cycle in an aerobic culture of the poxB mutant was confirmed, as evidenced by the decrease in metabolites involved in the TCA cycle, such as α-ketoglutarate and succinate (Table 2). Additionally, under anaerobic conditions, CO2 is important for bacterial growth (Dharmadi et al. 2006; Merlin et al. 2003; Tran et al. 2014), and metabolic pathways, such as the PoxB pathway, should be involved in producing an endogenous supply of CO2 because of the lack of TCA. Thus, the serious growth defect of the poxB mutant implies that an appreciable amount of PoxB-generated CO2 is required for bacterial growth. The alleviation of the anaerobic growth defect of the poxB mutant by the addition of NaHCO3 (an external CO2 donor) to the medium strongly supports the above assumption (Fig. 3). It was reported that the requirement for CO2 for bacterial growth mainly depended on PDHC in anaerobically grown E. coli (Murarka et al. 2010). Here, it was shown that PoxB affected the generation of CO2 during the anaerobic growth of K. pneumoniae. Based on our data, compared with aerobic growth, deleting poxB resulted in a great accumulation of pyruvate and a more severe inhibition of anaerobic growth of K. pneumoniae (Table 2 and Fig. 2).

The Pta-AckA pathway is considered to be the primary acetate synthesis pathway in E. coli (De Mey et al. 2007; Merlin et al. 2003). In our study, the deletion of pta-ackA, unlike the deletion of poxB, had no effect on bacterial growth and metabolite formation in K. pneumoniae. However, the growth and metabolic characteristics of the poxB-pta-ackA mutant significantly differed from those of the poxB mutant. Under aerobic conditions, a large decrease in acetate formation occurred in the poxB-pta-ackA mutant, which strongly suggests that Pta-AckA also plays an important role in acetate formation, although the Pta-AckA pathway is dispensable in a K. pneumoniae strain that expresses PoxB. Accompanying the decrease in acetate formation, the aerobic growth of the poxB-pta-ackA mutant was better than that of the poxB mutant. The cell concentration of the poxB-pta-ackA mutant at the end of flask cultivation almost recovered to the same level of the parent strain. Thus, this further confirmed that the high accumulation of acetate was responsible for the aerobic growth defect of the poxB mutant. More interestingly, the alleviation of the growth defect of the poxB mutant by deleting the pta-ackA even occurred under aerobic conditions. This could mean that, in addition to the decrease in acetate formation in aerobic conditions, deleting pta-ackA could also overcome the shortage of CO2 in the poxB mutant under anaerobic conditions. According to our data, the increase in metabolites involved in the TCA cycle (α-ketoglutarate and succinate) clearly suggests that carbon flux could be forced to the TCA cycle in the poxB-pta-ackA mutant under anaerobic conditions, and an increased flux of the TCA cycle can supply sufficient CO2 for the anaerobic growth of K. pneumoniae.

Increasing the carbon flux to the TCA cycle could generate extra dihydronicotinamide adenine dinucleotide (NADH), which is helpful for 1,3-PD production. Further studies indicated that the cell growth rate of poxB-pta-ackA mutant was still little lower than that of parent strain, and accumulation of pyruvate was also found in the early stage of cultivation, which was observed in Pta-AckA deficient E. coli strains (Chang et al. 1999; Liao et al. 1996). Nevertheless, the poxB-pta-ackA mutant showed an excellent ability to produce 1,3-PD under both anaerobic and micro-aerobic conditions. Based on our data, deleting poxB-pta-ackA severely decreased acetate formation and increased the carbon flux to the TCA cycle, while there was no effect on 2,3-BD formation. As a result, the 1,3-PD concentration and yield reached 76.8 g/L and 0.66 (mol/mol), respectively, after 30 h of fed-batch fermentation, which were 15 and 16 %, respectively, greater than those of the parent strain.

In summary, a K. pneumoniae poxB mutant exhibited an extreme growth defect. The growth defect of the poxB mutant was alleviated by deleting pta-ackA, probably due to the decrease in acetate formation under aerobic conditions and the improvement in the CO2 supply under anaerobic conditions. According to our data, deleting poxB, pta, and ackA in K. pneumoniae severely decreased acetate formation and increased carbon flux to the TCA cycle, which was improved 1,3-PD production.