- Email: [email protected]
Engineering synergetic CO2-fixing pathways for malate production
Metabolic Engineering 47 (2018) 496–504
Contents lists available at ScienceDirect
Metabolic Engineering journal homepage: www.elsevier.com/locate/meteng
Engineering synergetic CO2-fixing pathways for malate production a,c,1
b,1
a,c
a,c
a,c
T a,c
Guipeng Hu , Jie Zhou , Xiulai Chen , Yuanyuan Qian , Cong Gao , Liang Guo , ⁎ Peng Xud, Wei Chena, Jian Chena,c, Yin Lib, Liming Liua,c, a
State Key Laboratory of Food Science and Technology, Jiangnan University, 1800 Lihu Road, Wuxi 214122, China CAS Key Laboratory of Microbial Physiological and Metabolic Engineering, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China Key Laboratory of Industrial Biotechnology, Ministry of Education, Jiangnan University, Wuxi, China d Chemical Biochemical and Environmental Engineering, University of Maryland Baltimore County, Baltimore, MD, USA b c
A R T I C LE I N FO
A B S T R A C T
Keywords: CO2 fixation Malate production ATP balance Pathway engineering
Increasing the microbial CO2-fixing efficiency often requires supplying sufficient ATP and redirecting carbon flux for the production of metabolites. However, addressing these two issues concurrently remains a challenge. Here, we present a combinational strategy based on a synergetic CO2-fixing pathway that combines an ATP-generating carboxylation reaction in the central metabolic pathway with the ATP-consuming RuBisCO shunt in the carbon fixation pathway. This strategy provides enough ATP to improve the efficiency of CO2 fixation and simultaneously rewires the CO2-fixing pathway to the central metabolic pathway for the biosynthesis of chemicals. We demonstrate the application of this strategy by increasing the CO2-fixing rate and malate production in the autotroph Synechococcus elongatus by 110% and to 260 μM respectively, as well as increasing these two factors in the heterotrophic CO2-fixing Escherichia coli by 870% and to 387 mM respectively.
1. Introduction
fixation in vitro was built and optimized using enzyme engineering and metabolic proofreading, in which an efficient ATP recycling system was devised to meet the need for ATP (Schwander et al., 2016). These pathways can realize the high efficiency of an ATP supply, but it is difficult to rewire them into central metabolic pathways in vivo (Gong and Li, 2016). To circumvent this difficulty, the power of natural CO2fixing pathways has been harnessed. First, several intermediate metabolites have been engineered as the branching points for value-added chemical production in autotrophic microbes, such as 2,3-butanediol from pyruvate (Kanno et al., 2017; McEwen et al., 2016; Oliver et al., 2013), lipid from acetyl-CoA (Ajjawi et al., 2017), and ethylene from oxoglutarate (Xiong et al., 2015). Second, since the genetic tools for autotrophic microbes remain limited, some natural CO2-fixing pathways, such as the Calvin-Benson-Bassham (CBB) cycle and 3-hydroxypropionate bicycle, have been transplanted into heterotrophs to pave the way for the creation of completely synthetic autotrophs (Antonovsky et al., 2016; Kerfeld, 2016; Mattozzi et al., 2013). Although these CO2-fixing pathways are more easily rewired into central metabolic pathways than synthetic pathways, they still exhibit ATP deficiencies. In addition, researchers have paid attentions to the WoodLjungdahl (WL) pathway, which is the most ATP-efficient among the six natural CO2-fixing pathways. However, the genetic tools available for the microbes that harbor the WL pathway remain limited to date, and
Elevated atmospheric CO2 concentrations have been caused by human activities globally. CO2 emissions for 2007 were 28.8 Gt, which is expected to grow to 40.3 Gt by 2030 and to 50 Gt by 2050 (Venkata Mohan et al., 2016). To decrease atmospheric CO2 concentrations, there is an urgent need to reduce CO2 emissions and develop CO2 sequestration strategies. Conventional methods for CO2 sequestration, including CO2 capture (Leung et al., 2014), CO2 separation (Chou, 2013), and CO2 storage (Boot-Handford et al., 2014), have obvious shortfalls, such as energy-intensive processes and high operational costs (Hicks et al., 2017). To overcome these disadvantages, biological CO2-fixing techniques have garnered much attention, and such biorefinery approaches have been used to produce value-added chemicals (Liao et al., 2016). However, a remaining key issue is how to enhance biological CO2-fixing efficiency by replenishing ATP deficiencies in CO2-fixing pathways and engineering efficient biosynthetic pathways for valueadded chemical production. To explore more efficient CO2-fixing pathways able to reduce the demand for ATP, several synthetic carbon fixation pathways, which combine existing metabolic building blocks from various organisms, were computationally identified based on pathway kinetics and thermodynamics (Bar-Even et al., 2010). Then, a synthetic pathway for CO2
⁎
1
Corresponding author at: State Key Laboratory of Food Science and Technology, Jiangnan University, 1800 Lihu Road, Wuxi 214122, China. E-mail address: [email protected] (L. Liu). These authors contributed equally to this work.
https://doi.org/10.1016/j.ymben.2018.05.007 Received 1 March 2018; Received in revised form 10 May 2018; Accepted 10 May 2018 1096-7176/ © 2018 Published by Elsevier Inc. on behalf of International Metabolic Engineering Society.
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
Fig. 1. Overview of ATP balance in CO2-fixing pathways to improve malate production. ATP balance was achieved by introducing the PCK pathway. (a) Stoichiometric calculations of ATP balance for CO2 fixation with malate production in S. elongatus. (b) Stoichiometric calculations of ATP balance for CO2 fixation with malate production in E. coli. Abbreviations: hv, photon energy; G-6-P, glucose 6-phosphate; Ru5P, ribulose-5-phosphate; RuBP, ribulose-1,5-bisphosphate; 3PGA, 3-phosphoglyceric acid; PEP, phosphoenolpyruvate; OAA, oxaloacetate; MAL, malate; PRK, phosphoribulokinase; RuBisCO, ribulose-1,5-bisphosphate carboxylase/oxygenase; PCK, phosphoenolpyruvate carboxykinase; MDH, malate dehydrogenase; CBB, Calvin-Benson-Bassham cycle.
2. Materials and methods
this CO2-fixing pathway is strictly anaerobic (Schuchmann and Muller, 2014). These findings indicate the importance of developing novel strategies to improve CO2 biorefineries by considering the balance between efficient ATP supply and major metabolic pathway rewiring. In this study, we provide a practical solution for this balance. Based on stoichiometric calculations, we first identified a synergetic CO2fixing pathway that combines the ATP-generating phosphoenolpyruvate carboxykinase (PCK) pathway with the ATP-consuming RuBisCO shunt (or the CBB cycle) (Fig. 1). Then, the enzymes for building the PCK pathway were screened using a cell-free system. We subsequently demonstrated the application of the synergetic CO2-fixing pathway in Synechococcus elongatus, which resulted in a higher CO2-fixing rate (RCf) and malate titer. Finally, the synergetic CO2-fixing pathways were reconstructed in Escherichia coli, and the RCf and yield of malate form glucose were both significantly improved.
2.1. Strains and plasmids All strains and plasmids used in this study are listed in Supplementary Table 1 and Supplementary Table 2. S. elongatus UTEX 2973 and E. coli MG1655 were used as autotrophic and heterotrophic hosts, respectively, for investigating CO2 fixation and malate production. E. coli DH5a was used to construct vectors for engineering S. elongatus and E. coli. The plasmid pEM (a modified pETM6 (Xu et al., 2012) plasmid formed by replacing the promoter and multiple cloning enzyme cleavage sites with those from pQE-80l-kan) was used to provide a platform for the combinatorial optimization of CA and RuBisCO in E. coli. During the construction of S. elongatus mutants, cultures were grown in BG-11 medium with 50 mM NaHCO3 at 38 °C, and illumination intensity was controlled at 200 μmol photons/m2/s. During the construction of E. coli mutants, cultures were grown aerobically at 497
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
2.5. Carbon fixation rate determination
30 °C, 37 °C, or 42 °C in Luria–Bertani (LB) broth. Ampicillin, kanamycin, erythromycin, or chloramphenicol were added appropriately.
A GC-2014 gas chromatograph (Shimadzu, Kyoto, Japan) equipped with a TCD detector was used to calculate the net CO2 uptake rate (n = 3). The RTX-QBOND capillary column (30 m; 0.32 mm inside diameter; 10 mm film thickness; RESTEK Corporation, Bellefonte, PA, USA) was used to quantify CO2 in samples. The detailed process for determining net CO2 uptake rate by S. elongatus strains has been described by a previous study (Zhou et al., 2016). To calculate the net CO2 uptake rate by E. coli strains, cells in mid-log phase were harvested, washed with M9 medium and resuspended in 80 mL fresh M9 medium with 5% glucose and 20 mM NaHCO3 at a final OD600 = 3. Then the 80 mL cell suspension was transferred to a serum bottle with a volume of 100-mL and cultivated for 2 h. The total inorganic carbon of cell suspension was released by injecting concentrated HCl, and the concentration of CO2 in the headspace gas was quantified using gas chromatograph.
2.2. Genetic manipulations For the construction of S. elongatus mutants, double crossover homologous recombination based on tri-parental conjugation was used to integrate exogenous genes into the chromosome of S. elongatus (Yu et al., 2015). The conjugal strain was constructed by transforming the conjugal plasmid pRL443, and cargo strains were constructed by transforming pBR322 carrying the gene of interest into E. coli HB101 competent cells harboring the helper plasmid pRL623. Then, the conjugal strain (OD600 = 1.5–2.0), cargo strain (OD600 = 1.5–2.0), and S. elongatus strain (OD730 = 0.5–0.8) were centrifuged, washed, resuspended with BG-11 medium, mixed at a ratio of 1:1:2, and incubated for 4 h. These mixed cells were cultured in BG-11 + 5% LB (v/v) agar plates for 12 h and then transferred to BG-11 agar plates containing selective antibiotics. All primers used for genetic manipulations in S. elongatus are listed in Supplementary Table 3. For the construction of E. coli mutants, Red recombinase technology was used to knockout target genes (Dong et al., 2017). To fine-tune the expression level of CA and RuBisCO (which are come from Synechocystis PCC 6803), all exogenous genes were integrated into one plasmid (Supplementary Fig. 9). First, each gene was inserted into the pEM using the restriction sites SacI and PstI, thus forming component vectors. Then, the restriction sites AvrII, SpeI, and SalI were used for assembling these components to form pEM-CF5x. To replace the RBS of pEM-eGFP, inverse primers containing different RBS sequences were designed, and the restriction sites EcoRI and BamHI were used for replacing the original RBS of pEM-eGFP. All primers used for genetic manipulations in E. coli are listed in Supplementary Table 4 and Supplementary Table 5.
2.6. ATP assay To prepare samples for ATP assay, cells of S. elongatus (Kanno et al., 2017) and E. coli (Gong et al., 2015) in the exponential growth phase were harvested and washed. An ATP Assay Kit S0027 (Beyotime Biotechnology, Shanghai, China) was used to determine ATP content (n = 3). The concentration of intracellular ATP was calculated according to an ATP standard curve. 2.7. Production of malate For determination of malate production by S. elongatus, strains were initially grown in BG-11 medium containing 50 mM NaHCO3 at 38 °C, and illumination intensity was controlled at 200 μmol photons/m2/s (n = 3 per time point). For determination of malate production by E. coli, strains were initially grown aerobically in LB broth (n = 3 per time point). The cultures were induced for 8 h using 0.1 mM IPTG at mid-log phase. Then, cells were harvested, concentrated, and resuspended in fresh M9 medium with 5% glucose and 20 mM NaHCO3 at a final OD600 = 9. These cultures were transferred to a 100-mL serum bottle with an 80mL working volume. Anaerobic conditions were achieved by purging the headspace with N2 gas and sealing the bottle with a rubber plug (Bogorad et al., 2013). The initial pH was controlled at 7.0, and 1.6 mL of 1.0 M NaHCO3 was added every 12 h. For fed-batch fermentation in a 1.0-L bioreactor, cells were cultured at an initial OD600 = 27. The working volume was controlled at 0.8 L, and the pH of the medium was controlled at 7.0.
2.3. Enzyme assays Cells were harvested by centrifugation (4 °C, 5 min, 5000× g) during mid-log phase and then washed and resuspended in cold 100 mM Tris-HCl buffer (pH 7.5 or 8.0). After cell disruption using FastPrep-24 (MP Biomedicals, Solon, OH, USA), preparations were clarified by centrifugation (13,000× g for 10 min). Protein concentration was measured by the BCA method (Pierce, Rockford, IL, USA). We performed the experiment in triplicate (n = 3). RuBisCO activity was measured as described by Bonacci et al. (Bonacci et al., 2012). First, the reaction buffer containing 100 mM Tris (pH 8.0), 20 mM MgCl2, 10 mM KCl, 25 mM NaHCO3, 3.5 mM ATP, 5 mM creatine phosphate, 2 mM DTT, 5 U/mL 3-phosphoglycerate kinase, 5 U/mL creatine phosphokinase, 5 U/mL glyceraldehyde 3-phosphate dehydrogenase, and appropriate amounts of enzyme solution was activated for 10 min at 30 °C. Then, the reaction was initiated by adding 0.5 mM RuBP and 0.25 mM NADH and was immediately monitored at 340 nm (30 °C) in a plate reader (Wallac/Spectromax) for 5 min. PRK activity was measured as described by Masahiro Kannoet al.(Kanno et al., 2017). PCK activity was determined as described by Van der Werf et al. (VanderWerf et al., 1997). MDH activity was measured at A340 to monitor NADH oxidation (Chen et al., 2013).
2.8. High-performance liquid chromatography (HPLC) detection An HPLC system (Dionex UltiMate 3000 Series; Thermo Scientific, Waltham, MA, USA) was used for determining concentrations of glucose and organic acid. This involved an Aminex HPX-87H column (7.8× 300 mm; Bio-Rad Laboratories, Inc., Hercules, CA, USA) at 35 °C with 0.05 mM sulfuric acid as the mobile phase. The injection volume was 10 μL, and the flow rate was 0.6 mL/min.
2.4. Cell free synthesis of malate from PEP 2.9. Ultra-performance liquid chromatography–mass spectrometry (UPLCMS) detection
Cell free reactions were performed at a volume of 25 μL in 1.5 mL tubes and incubated at 37 °C. The standard reaction contained the following components: 50 mM Tris-HCl (pH 7.5), 10 mM PEP, 10 mM NADH, 10 mM ADP, 10 mM NaH13CO3 and 2.5 mM MgCl2. PCKs and MDHs were individually overexpressed and purified from E. coli BL21 (DE3). Purified PCK and MDH were both added with the final protein concentration of 0.2 μM. The reactions were incubated for 30 min at 37 °C.
Samples were centrifuged for 5 min at 12,000 ×g and diluted 1:10 in H2O for injection into the UPLC-MS system. The relative abundance of malate was detected using a Dionex UltiMate 3000 Q-TOF LC-MS system equipped with an electrospray ionization source set to negative ionization mode. UPLC-MS analysis was performed as described previously (Schwander et al., 2016). 498
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
48 photons + 8 CO2 + 6H2O + 0.42 ADP + 0.42 Pi = 2 malate + 0.42 ATP + 6 O2 (8)
2.10. UPLC–tandem mass spectrometry (UPLC-MS/MS) detection To prepare samples for UPLC-MS/MS detection, cells of S. elongatus (Kanno et al., 2017) and E. coli (Gong et al., 2015) in the exponential growth phase were harvested and washed. The detailed MRM parameters for 3-PGA and 1,5-RuBP analysis were optimized as described previously (Gong et al., 2015).
5·5 glucose + 15 CO2 + 6 NADH + 6H+ = 12 malate +3H2O + 6 NAD+ (9)
3.2. Screening enzymes for the PCK pathway To build an efficient PCK pathway for supplying ATP for CO2 fixation with malate biosynthesis, efficient PCK and MDH enzymes were needed. Based on recent studies, the pck genes from E. coli (Zhang et al., 2011), Actinobacillus succinogenes (VanderWerf et al., 1997), and Agrobacterium (Fu et al., 2010) and the mdh genes from E. coli (Zhang et al., 2011), A. succinogenes (VanderWerf et al., 1997), and Rhizopus oryzae (Zhang et al., 2012) were selected, overexpressed, and purified. Then, the specific enzyme activity and kinetic parameters of these PCKs and MDHs were measured. AsPCK had the best specific enzyme activity and kcat/Km, reaching 13 ± 0.8 U mg protein−1 and 1.7 × 105 M−1 s−1, respectively (Supplementary Table 6). Similarly, AsMDH also possessed the best corresponding parameters at 120 ± 9.3 U mg protein−1 and 1.4 × 106 M−1 s−1, respectively (Supplementary Table 6). Thus, AsPCK and AsMDH were selected to construct the PCK pathway for improving malate production and CO2 fixation. To test the efficiency of AsPCK and AsMDH for converting PEP to malate, these two enzymes were added to cell-free cocktails containing 10 mM substrate (PEP), catalytic amounts of cofactors (10 mM NADH and 10 mM ADP), and ions (10 mM NaH13CO3 and 2.5 mM MgCl2) to mimic the cytoplasmic environment. As a result, 6.7 ± 0.4 mM malate was detected in the supernatant after 30 min. These results indicated that AsPCK and AsMDH could be used for efficiently transforming PEP to malate and further building an efficient PCK pathway in living cells.
3. Results 3.1. Calculating biological CO2 fixation based on stoichiometry In autotrophic microbes such as S. elongatus, CO2 is fixed to sugar phosphate through the CBB cycle in dark reactions, requiring 3 ATP/2 NADPH. However, only 2.57 ATP/2 NADPH are generated by light reactions (Zhou et al., 2016). Thus, the stoichiometry of glucose-6phosphate (G-6-P) production from CO2 can be calculated by equation ① (Supplementary Note 1). Assuming that all the carbons can be channeled to malate, the optimal stoichiometry for converting G-6-P to malate is normally calculated by equation ②. Thus, combining equations ① + ②, the stoichiometry of malate from CO2 can be calculated by equation ③, meaning that an extra 1.58 ATP are required for this conversion. In heterotrophic microbes such as E. coli, the optimal stoichiometry of malate production from glucose is normally described by equation ④. It has been demonstrated that the CBB cycle enzymes [phosphoribulokinase (PRK) and ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO)] can be used for building a CO2-fixing bypass (termed the RuBisCO shunt) in E. coli (Antonovsky et al., 2016; Gong et al., 2015). When the RuBisCO shunt is constructed in E. coli, more CO2 can be transformed to malate according to equation ⑤ (Supplementary Note 2), but extra ATP are also needed for this conversion (Supplementary Fig. 7). In summary, ATP deficiency is a common issue for CO2 fixation in both autotrophic and heterotrophic CO2-fixing microbes.
3.3. Engineering the PCK pathway in S. elongatus We first demonstrated this strategy in the autotrophic microbe S. elongatus UTEX 2973. To reconstruct the PCK pathway in this microbe, the Aspck and Asmdh genes were integrated into the chromosome of S. elongatus with strong promoters Pcpc6803 and Pcpc560, respectively, obtaining strain SH006 (Fig. 2a,b). As a result, the RCf of strain SH006 was 73% ± 5% higher than that of wild-type S. elongatus (Fig. 2d). Simultaneously, malate production was increased from below the detection limit (~ 10 μM) to 200 ± 19 μM in strain SH006 after 9 days of cultivation (Fig. 2e). To test whether the CO2 fixation of the CBB cycle was improved by integrating AsPCK and AsMDH, the enzyme activities of PRK and RuBisCO were analyzed. Increases in the activities of PRK (1.4 ± 0.1-fold) and RuBisCO (1.3 ± 0.1-fold) were confirmed in mutant SH006 compared to activity levels in wild-type S. elongatus (Fig. 2f). Moreover, we observed that the concentrations of intracellular RuBP and 3-PGA were increased by 63% ± 4% and 51% ± 3%, respectively (Fig. 2g). These results demonstrated that engineering the PCK pathway not only improved malate production, but also enhanced the flux of the CBB cycle. Furthermore, the concentration of intracellular ATP (CATP) in strain SH006 was 1.7 ± 0.12 μmol gDCW−1, which was 85% ± 6% higher than that of wild-type S. elongatus (Fig. 2d), indicating that an enhanced ATP supply was beneficial for improving CO2 fixation in S. elongatus. To further promote CO2 fixation in S. elongatus through fine-tuning the ATP supply, AsPCK and AsMDH were expressed at different levels. First, gene expression was divided into two levels using the strong promoters Pcpc6803 or Pcpc560 and the weak promoter Pcpc300. Next, a series of AsPCK and AsMDH expression cassettes with various strengths were designed and assembled in S. elongatus to obtain the best ATP distribution for enhancing CO2 fixation (Fig. 2c). Finally, we found that controlling the strength of AsPCK at a low level and AsMDH at a high level resulted in a 17% ± 0.5% decrease in CATP compared with that of
48 photons + 6 CO2 + 5H2O + 2·58 ATP = 1 G-6-P + 2·58 ADP + 1·58 Pi + 6 O2 (1) G-6-P + 2 CO2 + H2O + ADP = 2 malate + ATP
(2)
48 photons + 8 CO2 + 6H2O + 1·58 ATP = 2 malate + 1·58 ADP + 1·58 Pi + 6 O2 (3) glucose + 2 CO2 = 2 malate
(4) +
2·5 glucose + 9 CO2 + 6 ATP + 6 NADH + 6H + 6 NAD+
= 6 malate +3H2O (5)
In order to promote CO2 fixation, an effective strategy is to rewire metabolic pathways to rescue this ATP deficiency and simultaneously transfer the fixed carbon to the production of a high-value chemical. Here, we selected malate as the target chemical for our calculations. An extra ATP can be generated by overexpressing PCK and malate dehydrogenase (MDH) for malate production, resulting in the theoretical maximum yield of malate from glucose/G-6-P (2 mol/mol) (Chen et al., 2017). Consequently, the optimal stoichiometry for the conversion of G6-P/glucose to malate can be improved to equation ⑥/⑦. Thus, when the ATP deficiency is replenished, the stoichiometry of malate production in S. elongatus is transformed into equation ⑧ = ① + ⑥ (Fig. 1a), and in CO2-fixing E. coli is transformed into equation ⑨ = ⑤ + 3 × ⑦ (Fig. 1b). In conclusion, engineering the PCK pathway for malate production can enhance CO2 fixation. In turn, this enhanced CO2 fixation can improve malate production, thus resulting in a mutually reinforcing cycle. G-6-P + 2 CO2 + H2O + 3 ADP + 2 Pi = 2 malate + 3 ATP
(6)
glucose + 2 CO2 = 2 malate + 2 ATP
(7) 499
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
Fig. 2. Engineering the PCK pathway improves CO2 fixation and malate production in S. elongatus. (a, b) Schematic representation of recombination to integrate Asmdh and Aspck. (c) A series of PCK and MDH expression cassettes were designed at different expression levels. (d) CO2-fixing rate and concentration of intracellular ATP in wild-type and engineered S. elongatus strains. (e) Time course of malate production by wild-type and engineered S. elongatus strains. (f) Crude enzyme activities of PRK and RuBisCO in wild-type S. elongatus and strain SH006. (g) Concentrations of intracellular RuBP and 3-PGA in wild-type and S. elongatus engineered strains. *P < 0.05 as determined by t-test; N = 3 biological replicates; the s.d. is shown as black error bars. Abbreviations are defined in the Fig. 1 legend.
(278 mM) was fed to strains MG1655 and FH010. We observed that the ratio of 13C-labeled 3-PGA to unlabeled 3-PGA (13C-3-PGA/3-PGA) was increased from 2% ± 0.2% (strain MG1655) to 8% ± 0.6% (strain FH010) (Fig. 3b). These results indicated that the RuBisCO shunt in E. coli was able to capture 13CO2 and convert it to 3-PGA (Fig. 4a). Next, 11 genes in CO2-fixing E. coli FH010 were knocked out, resulting in strain FH019 (ΔptsG ΔfrdBC ΔfumABC ΔmaeAB ΔaspC ΔldhA ΔadhE) for malate production (Supplementary Fig. 8 and Supplementary Note 3). AsPCK and AsMDH were introduced into strain FH019, obtaining strain FH089. As a result, the RCf in strain FH089 was 490% ± 32% higher than that of strain FH019 (Fig. 3d). This enhancement led to a 620% ± 41% increase in the yield of malate from glucose (Ym/g), with the titer of malate increasing from 7.9 ± 0.5 mM to 106 ± 8.9 mM (Fig. 3c and Supplementary Table 8). In contrast, strain FH029 (FH009 expressing only AsPCK and AsMDH) only produced 30 ± 1.9 mM malate; the RCf and Ym/g of strain FH029 were also 55% ± 3% and 53% ± 3% lower than those of strain FH089, respectively (Supplementary Table 7 and Fig. 3c,d). Even when CA was further expressed in strain FH029 (resulting in strain FH039), the titer of malate was increased to only 42 ± 4.1 mM, with RCf and Ym/g values lagging behind those of strain FH089 by 34% ± 2% and 42% ± 3%, respectively (Supplementary Table 7 and Fig. 3c,d). These results
strain SH006 but a 19% ± 0.9% increase in RCf, reaching 310 ± 22 mg gDCW−1 h−1 (Fig. 2d). Furthermore, the highest malate titer achieved was 260 ± 21 μM (Fig. 2e). These results demonstrated that CO2 fixation and malate production in S. elongatus could be further enhanced by optimizing the ATP supply to an appropriate level.
3.4. Engineering synergetic CO2-fixing pathway for malate production in E. coli To determine whether the PCK pathway can improve CO2 fixation by RuBisCO in heterotrophic microbes, we first constructed a CO2fixing E. coli harboring the RuBisCO shunt. The enzymes PRK and RuBisCO from Synechocystis sp. PCC 6803 were introduced into E. coli MG1655 to build the CO2-fixing bypass. Carbonic anhydrase (CA) from Synechocystis sp. PCC 6803 was expressed simultaneously to provide sufficient CO2 for RuBisCO, obtaining the CO2-fixing E. coli strain FH010. To test the function of this CO2-fixing shunt, we analyzed the concentration of intracellular 1,5-RuBP, which is the specific metabolite of the RuBisCO shunt in E. coli. As a result of building the RuBisCO shunt, this parameter was increased from below the detection limit (~0.01 μmol gDCW−1) to 2.1 ± 0.1 μmol gDCW−1 (Supplementary Fig. 6). Furthermore, NaH13CO3 (20 mM) along with unlabeled glucose 500
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
Fig. 3. Tuning the balance of CO2-fixing pathways improves CO2 fixation and malate production in E. coli. (a) Characteristics of strains used in this section; (b) Concentration of intracellular 3-PGA and 13C-3-PGA and the ratio of 13C-3-PGA/3-PGA in wild-type and engineered E. coli strains. (c) Time course of malate production and molar yield of malate from glucose by strains FH019, FH089, and FH389. (d) Concentration of intracellular ATP and CO2-fixing rate by strains FH019, FH089, and FH389. (e) A series of CA and RuBisCo expression cassettes were designed at different expression levels. (f) The yields of malate from glucose by strains with different CA and RuBisCO expression cassettes. *P < 0.05 as determined by the t-test; N = 3 biological replicates; the s.d. is shown as black error bars.
strengths: high (H), moderate (M), and low (L) (Supplementary Fig. 10 and Fig. 3e). Expressing RuBisCO at a low level and CA at a high level resulted in the highest RCf and Ym/g of 49 ± 3.1 mg gDCW−1 h−1 (Figs. 3d) and 1.37 ± 0.12 mol/mol (Fig. 3f), respectively. To further reveal the effect of engineering synergetic CO2-fixing pathways on carbon metabolism, we analyzed the carbon balance of strains FH039, FH089, and FH389. The partial apparent carbon efficiencies (CPro/CGlu) of strains FH039, FH089, and FH389 were 60%, 78%, and 99%, respectively (Supplementary Table 7). These results indicated that introducing the RuBisCO shunt was beneficial for redirecting carbon flux to organic acids (especially for malate in our engineered strains), and this effect was further improved by tuning the expression of RuBisCO and CA. Finally, strain FH389 produced 387 ± 43 mM malate in a bioreactor, with the Ym/g reaching 1.46 ± 0.12 mol/mol (Supplementary Fig. 12), which is the highest from engineered E. coli (Supplementary Table 8).
demonstrated that the RuBisCO shunt contributed to malate production. To further demonstrate that the flux of the RuBisCO shunt was improved by the introduction of AsPCK and AsMDH, 20 mM NaH13CO3 was used to supply CO2 for strains FH019 and FH089. As expected, the ratio of 13C-3-PGA/3-PGA in strain FH089 was 1.9 ± 0.1-fold higher than that in strain FH019 (Fig. 3b). Meanwhile, the relative abundance of malate-m2 (malate with two 13C atoms) in FH089 was increased from 9% ± 0.6–17% ± 0.9% (Fig. 4d,e). These results indicated that reconstructing the PCK pathway was effective at improving the ATP supply for the RuBisCO shunt, enhancing overall CO2 fixation and malate production in E. coli. Although the CATP in strain FH089 was increased to reinforce the overall CO2 fixation (Fig. 3d), the distribution of ATP could be further optimized through fine-tuning the expression levels of CO2 capture enzymes. In order to achieve a more appropriate distribution of ATP and redirect more carbon to malate, RuBisCO and CA were selected and expressed under three different ribosome-binding sequence (RBS) 501
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
Fig. 4. Effect of CO2 fixation on malate production in E. coli. (a) Schematics of carbon atom transits via the RuBisCO shunt and the PCK pathway. 13CO2 fixed into 3-PGA by the RuBisCO shunt can be redirected into the PCK pathway for biosynthesis of malate in the engineered E. coli. Abbreviations are defined in Fig. 1 legend. The labeled (13C) and unlabeled carbon atoms are represented by closed and open circles, respectively. (b–f) Relative abundance of mass isotopomers for malate in engineered strains. Malate-m0 represents malate without 13C atoms, malate-m1 represents malate with one 13C atom, and malate-m2 represents malate with two 13C atoms. Minimal malate-m1 was detected in the control strain (FH009 NaHCO3), mainly owing to the small fraction of 13C isotopes present in all natural 12Ccontaining compounds. The malate-m0 in the engineered strain presumably originated from other metabolic pathways or was synthesized from the unlabeled PEP combined with unlabeled CO2 released from decarboxylic reactions. N = 3 biological replicates; the s.d. is shown as black error bars.
4. Discussion
Recycling CO2 into central metabolic pathways is a promising approach for expanding the utilization of CO2 fixation to improve the yield of target metabolites. Recently, the power of autotrophs has been exploited in heterotrophic microbes to meet this goal (Claassens et al., 2016). For instance, the RuBisCO shunt was introduced into E. coli MZ for organic acid production (Yang et al., 2016), but the yield of products were not able to meet the requirement for industrial production. In this study, we rewired the RuBisCO shunt with the PCK pathway for malate production in E. coli. Thus, CO2 fixation capability was effectively enhanced, and an impressive Ym/g (1.46 mol/mol) was achieved, representing a 3% increase over the highest reported Ym/g (1.42 mol/ mol) in malate-producing E. coli XZ658 (using the PCK pathway by not the RuBisCO shunt) (Zhang et al., 2011). Furthermore, the strategy of engineering a synergy between an ATP-producing biosynthetic pathway and ATP-consuming CO2-fixing pathway can be utilized for the production of other chemicals in E. coli. An efficient ATP supply is required for both biological CO2 fixation and chemical biosynthesis. In this study, we rewired and tuned the balance of the ATP-producing PCK pathway and ATP-consuming RuBisCO shunt to redirect intracellular ATP for the biosynthesis of chemicals and CO2 fixation. According to recent studies, several principles can be proposed to obtain efficient microbial cell factories: (1) get a clear understanding of the entire carbon and energy metabolisms, for example, by dividing ATP cost into growth-associated maintenance and non-growth-associated loss using flux balance analysis (Wu et al., 2015); (2) take advantage of native pathways for product synthesis, for example, by recruiting ATP-generating pathways for chemical production (Zhang et al., 2009); (3) avoid extensive pathway reconstruction, for example, by combining microbial production with chemical conversions to reduce biosynthetic steps (Wheeldon et al., 2017); (4) develop novel non-model microbial workhorses with desired traits in terms of energy metabolism, for example, by exploiting genetic tools for cyanobacteria, which acquire extra energy from light (Xiong et al.,
Microbial production of chemicals from CO2 faces two challenges: (i) an economic and efficient ATP supply for CO2-fixing pathways and (ii) the rewiring of CO2-fixing pathways with biosynthetic pathways. In this study, the PCK pathway for malate production with ATP generation was engineered into autotrophic S. elongatus to enhance RCf and heterotrophic CO2-fixing E. coli to increase Ym/g. Using this approach, CO2 fixation capability and malate biosynthesis were simultaneously boosted to new levels. This study provides a novel method for replenishing ATP to enhance CO2 fixation and engineering CO2-fixing pathway to improve the production of value-added chemicals. CO2-fixing efficiency is important for chemical production in autotrophic microbes. Recent studies have engineered autotrophic microbes for synthesizing chemicals and biofuels from CO2 (Nybo et al., 2015; Oliver and Atsumi, 2014; Oliver et al., 2016; Woo, 2017). However, production performance remains far below industrial feasibility, partly due to the low efficiency of the ATP supply for the biosynthesis of compounds in autotrophic cell factories (Claassens et al., 2016). In addition, the production of C4-dicarboxylic acids such as succinate (Huang et al., 2016; Li et al., 2016) has lagged far behind the “workhorses” of biotechnology like E.coli, due to the inherent weak flux of the relevant pathways in autotrophic microbes (Angermayr et al., 2015). Here, we rewired an ATP-producing pathway into a cyanobacterium to enhance the metabolic flux for malate biosynthesis, resulting in a cyanofactory with higher efficiencies of CO2 fixation and malate production. It is noteworthy that the enhancement of CO2 fixation was not only caused by the specific biosynthetic pathway (PCK carboxylation), but also by the CBB cycle. Therefore, this strategy can be used to reinforce the efficiency of autotrophic cell factories and engineer autotrophic microbes for the production of other value-added chemicals in the future, such as aspartate from the PEP-OAA pathway and lactate from the PEP-pyruvate pathway (Xu et al., 2017). 502
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
experiments. G.H., J.Z., X.C., Y.Q., C.G., L.G., P.X., W.C., J.C. and Y.L. analyzed the results.
2017). Synthetic pathways can be effectively constructed in microbes through current metabolic engineering and synthetic biology techniques. By rewiring synthetic pathways into microbial metabolisms, diverse natural and non-natural products have been synthesized efficiently (Clomburg et al., 2017). Obviously, synthetic pathways can enable us to perform important tasks that are not possible in nature. However, energy molecules and carbon building blocks are both required for constructing and operating synthetic pathways (Wu et al., 2016). Thus, the problem of metabolic burden may be a limiting factor for harnessing the power of synthetic pathways. Generally, cyanobacteria are more sensitive to metabolic burdens than E. coli because they have to synthesize large amounts of protein for photosynthesis. In this study, we introduced synthetic pathways by inserting exogenous genes into the chromosome of S. elongatus to reduce the metabolic burden. In other words, we should exploit the advantages and avoid the disadvantages of synthetic pathways, thus achieving more efficient microbial cell factories. More efficient CO2-fixing pathways can be achieved through synthetic biology (SynBio). However, several drawbacks of this process need to be considered: (1) synthetic CO2-fixing pathways may be hardly connected to the central carbon metabolism of organisms (Gong and Li, 2016); (2) even if synthetic pathways can be rewired into the central carbon metabolism, large metabolic burdens may be induced since energy molecules and carbon building blocks are both required for constructing and operating synthetic pathways (Wu et al., 2016); and (3) microorganisms have evolved robust metabolic and regulatory networks to survive and grow in specific environments, and therefore more effort should be made to use these synthetic CO2-fixing pathways for further biological conversions. While nature has performed biological engineering experiments for billions of years, it is possible to improve upon natural systems. For example, the CO2-fixing rate of the synthetic CETCH cycle is higher than those of all natural CO2-fixing pathways that have so far been discovered (Schwander et al., 2016). Similarly, the design of a synthetic light-harvesting system for non-photosynthetic bacteria enabled a higher efficiency of solar-to-chemical synthesis than that observed among naturally photosynthetic bacteria (Sakimoto et al., 2016). Furthermore, several strategies can be used to improve the outcome of SynBio projects, including: (1) combining -omic data with protein design de novo to build more efficient CO2-fixing pathways that can be easily rewired into living cells (Siegel et al., 2015); (2) harnessing the power of adaptive evolution, 13C-MFA, and genome-scale models to decrease metabolic burdens caused by synthetic pathways (Antonovsky et al., 2016; G and M, 2015); and (3) providing sugars (e.g., glucose) for engineered autotrophs to achieve mixotrophic fermentation, achieving higher productivity for synthetic pathways (Matson and Atsumi, 2018). By overcoming the disadvantages of SynBio, it may be possible to further improve such systems in the future.
Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ymben.2018.05.007. References Ajjawi, I., Verruto, J., Aqui, M., Soriaga, L.B., Coppersmith, J., Kwok, K., et al., 2017. Lipid production in Nannochloropsis gaditana is doubled by decreasing expression of a single transcriptional regulator. Nat. Biotechnol. 35, 647–652. Angermayr, S.A., Rovira, A.G., Hellingwerf, K.J., 2015. Metabolic engineering of cyanobacteria for the synthesis of commodity products. Trends Biotechnol. 33, 352–361. Antonovsky, N., Gleizer, S., Noor, E., Zohar, Y., Herz, E., Barenholz, U., et al., 2016. Sugar synthesis from CO2 in Escherichia coli. Cell 166, 115–125. Bar-Even, A., Noor, E., Lewis, N.E., Milo, R., 2010. Design and analysis of synthetic carbon fixation pathways. Proc. Natl. Acad. Sci. USA 107, 8889–8894. Bogorad, I.W., Lin, T.S., Liao, J.C., 2013. Synthetic non-oxidative glycolysis enables complete carbon conservation. Nature 502, 693–697. Bonacci, W., Teng, P.K., Afonso, B., Niederholtmeyer, H., Grob, P., Silver, P.A., et al., 2012. Modularity of a carbon-fixing protein organelle. Proc. Natl. Acad. Sci. USA 109, 478–483. Boot-Handford, M.E., Abanades, J.C., Anthony, E.J., Blunt, M.J., Brandani, S., Mac Dowell, N., et al., 2014. Carbon capture and storage update. Energy Environ. Sci. 7, 130–189. Chen, X., Gao, C., Guo, L., Hu, G., Luo, Q., Liu, J., et al., 2017. DCEO biotechnology: tools to design, construct, evaluate, and optimize the metabolic pathway for biosynthesis of chemicals. Chem. Rev. 118, 4–72. Chen, X., Xu, G., Xu, N., Zou, W., Zhu, P., Liu, L., et al., 2013. Metabolic engineering of Torulopsis glabrata for malate production. Metab. Eng. 19, 10–16. Chou, C.-t., 2013. Carbon dioxide separation and capture for Global warming mitigation. J. Adv. Eng. Technol. 1, 2348–2931. Claassens, N.J., Sousa, D.Z., Dos Santos, V.A., de Vos, W.M., van der Oost, J., 2016. Harnessing the power of microbial autotrophy. Nat. Rev. Microbiol. 14, 692–706. Clomburg, J.M., Crumbley, A.M., Gonzalez, R., 2017. Industrial biomanufacturing: the future of chemical production. Science 355, aag0804. Dong, X., Chen, X., Qian, Y., Wang, Y., Wang, L., Qiao, W., et al., 2017. Metabolic engineering of Escherichia coli W3110 to produce L-malate. Biotechnol. Bioeng. 114, 656–664. Fu, W., Lin, J., Cen, P., 2010. Expression of a hemA gene from Agrobacterium radiobacter in a rare codon optimizing Escherichia coli for improving 5-aminolevulinate production. Appl. Biochem Biotechnol. 160, 456–466. Gong, F., Li, Y., 2016. Fixing carbon, unnaturally. Science 354, 830–831. Gong, F., Liu, G., Zhai, X., Zhou, J., Cai, Z., Li, Y., 2015. Quantitative analysis of an engineered CO2-fixing Escherichia coli reveals great potential of heterotrophic CO2 fixation. Biotechnol. Biofuels 8, 86. Hicks, N., Vik, U., Taylor, P., Ladoukakis, E., Park, J., Kolisis, F., et al., 2017. Using prokaryotes for carbon capture storage. Trends Biotechnol. 35, 22–32. Huang, C.H., Shen, C.R., Li, H., Sung, L.Y., Wu, M.Y., Hu, Y.C., 2016. CRISPR interference (CRISPRi) for gene regulation and succinate production in cyanobacterium S. elongatus PCC 7942. Microb. Cell Fact. 15, 196. Kanno, M., Carroll, A.L., Atsumi, S., 2017. Global metabolic rewiring for improved CO2 fixation and chemical production in cyanobacteria. Nat. Commun. 8, 14724. Kerfeld, C.A., 2016. Rewiring Escherichia coli for carbon-dioxide fixation. Nat. Biotechnol. 34, 1035–1036. Leung, D.Y.C., Caramanna, G., Maroto-Valer, M.M., 2014. An overview of current status of carbon dioxide capture and storage technologies. Renew. Sust. Energ. Rev. 39, 426–443. Li, H., Shen, C.R., Huang, C.H., Sung, L.Y., Wu, M.Y., Hu, Y.C., 2016. CRISPR-Cas9 for the genome engineering of cyanobacteria and succinate production. Metab. Eng. 38, 293–302. Liao, J.C., Mi, L., Pontrelli, S., Luo, S., 2016. Fuelling the future: microbial engineering for the production of sustainable biofuels. Nat. Rev. Microbiol. 14, 288–304. Matson, M.M., Atsumi, S., 2018. Photomixotrophic chemical production in cyanobacteria. Curr. Opin. Biotechnol. 50, 65–71. Mattozzi, M., Ziesack, M., Voges, M.J., Silver, P.A., Way, J.C., 2013. Expression of the sub-pathways of the Chloroflexus aurantiacus 3-hydroxypropionate carbon fixation bicycle in E. coli: toward horizontal transfer of autotrophic growth. Metab. Eng. 16, 130–139. McEwen, J.T., Kanno, M., Atsumi, S., 2016. 2,3 Butanediol production in an obligate photoautotrophic cyanobacterium in dark conditions via diverse sugar consumption. Metab. Eng. 36, 28–36. Nybo, S.E., Khan, N.E., Woolston, B.M., Curtis, W.R., 2015. Metabolic engineering in chemolithoautotrophic hosts for the production of fuels and chemicals. Metab. Eng. 30, 105–120. Oliver, J.W., Atsumi, S., 2014. Metabolic design for cyanobacterial chemical synthesis. Photosynth Res. 120, 249–261. Oliver, J.W., Machado, I.M., Yoneda, H., Atsumi, S., 2013. Cyanobacterial conversion of carbon dioxide to 2,3-butanediol. Proc. Natl. Acad. Sci. USA 110, 1249–1254. Oliver, N.J., Rabinovitch-Deere, C.A., Carroll, A.L., Nozzi, N.E., Case, A.E., Atsumi, S., 2016. Cyanobacterial metabolic engineering for biofuel and chemical production.
Acknowledgments This work is supported by the National Natural Science Foundation of China (21676118, 21706095), the Provincial Natural Science Foundation of Jiangsu Province (BK20160163), the Fundamental Research Funds for the Central Universities (JUSRP51611A), the Key Research Program of the Chinese Academy of Sciences (ZDRW-ZS2016-3) and the National First-class Discipline Program of Light Industry Technology and Engineering (LITE2018-08). The authors would like to acknowledge Qingsheng Qi for E.coli MG1655, Min Jiang for A.succinogenes and Ye Ni for plasmid pQE-80l-kan. Author contributions G.H., J.Z., X.C. and L.L. conceived the project and wrote the manuscript. G.H., J.Z., X.C. and Y.Q. designed and performed all the 503
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
engineering: yin-yang. Microb. Cell Fact. 14, 39. Xiong, W., Cano, M., Wang, B., Douchi, D., Yu, J., 2017. The plasticity of cyanobacterial carbon metabolism. Curr. Opin. Chem. Biol. 41, 12–19. Xiong, W., Morgan, J.A., Ungerer, J., Wang, B., Maness, P.-C., Yu, J., 2015. The plasticity of cyanobacterial metabolism supports direct CO2 conversion to ethylene. Nat. Plants 1, 15053. Xu, N., Ye, C., Chen, X., Liu, J., Liu, L., 2017. Genome-scale metabolic modelling common cofactors metabolism in microorganisms. J. Biotechnol. 251, 1–13. Xu, P., Vansiri, A., Bhan, N., Koffas, M.A., 2012. ePathBrick: a synthetic biology platform for engineering metabolic pathways in E. coli. ACS Synth. Biol. 1, 256–266. Yang, C.H., Liu, E.J., Chen, Y.L., Ou-Yang, F.Y., Li, S.Y., 2016. The comprehensive profile of fermentation products during in situ CO2 recycling by Rubisco-based engineered Escherichia coli. Microb. Cell Fact. 15, 133. Yu, J., Liberton, M., Cliften, P.F., Head, R.D., Jacobs, J.M., Smith, R.D., et al., 2015. Synechococcus elongatus UTEX 2973, a fast growing cyanobacterial chassis for biosynthesis using light and CO2. Sci. Rep. 5, 8132. Zhang, B., Skory, C.D., Yang, S.T., 2012. Metabolic engineering of Rhizopus oryzae: effects of overexpressing pyc and pepc genes on fumaric acid biosynthesis from glucose. Metab. Eng. 14, 512–520. Zhang, X., Jantama, K., Moore, J.C., Jarboe, L.R., Shanmugam, K.T., Ingram, L.O., 2009. Metabolic evolution of energy-conserving pathways for succinate production in Escherichia coli. Proc. Natl. Acad. Sci. USA 106, 20180–20185. Zhang, X., Wang, X., Shanmugam, K.T., Ingram, L.O., 2011. L-malate production by metabolically engineered Escherichia coli. Appl. Environ. Microbiol. 77, 427–434. Zhou, J., Zhang, F., Meng, H., Zhang, Y., Li, Y., 2016. Introducing extra NADPH consumption ability significantly increases the photosynthetic efficiency and biomass production of cyanobacteria. Metab. Eng. 38, 217–227.
Curr. Opin. Chem. Biol. 35, 43–50. G, S., M, C.D., 2015. 13C metabolic flux analysis at a genome-scale. Metab. Eng. 32, 12–22. Sakimoto, K.K., Wong, A.B., Yang, P., 2016. Self-photosensitization of nonphotosynthetic bacteria for solar-to-chemical production. Science 351, 74–77. Schuchmann, K., Muller, V., 2014. Autotrophy at the thermodynamic limit of life: a model for energy conservation in acetogenic bacteria. Nat. Rev. Microbiol. 12, 809–821. Schwander, T., von Borzyskowski, L.S., Burgener, S., Cortina, N.S., Erb, T.J., 2016. A synthetic pathway for the fixation of carbon dioxide in vitro. Science 354, 900–904. Siegel, J.B., Smith, A.L., Poust, S., Wargacki, A.J., Bar-Even, A., Louw, C., et al., 2015. Computational protein design enables a novel one-carbon assimilation pathway. Proc. Natl. Acad. Sci. USA 112, 3704–3709. VanderWerf, M.J., Guettler, M.V., Jain, M.K., Zeikus, J.G., 1997. Environmental and physiological factors affecting the succinate product ratio during carbohydrate fermentation by Actinobacillus sp. 130Z. Arch. Microbiol. 167, 332–342. Venkata Mohan, S., Modestra, J.A., Amulya, K., Butti, S.K., Velvizhi, G., 2016. A circular bioeconomy with biobased products from CO2 sequestration. Trends Biotechnol. 34, 506–519. Wheeldon, I., Christopher, P., Blanch, H., 2017. Integration of heterogeneous and biochemical catalysis for production of fuels and chemicals from biomass. Curr. Opin. Biotechnol. 45, 127–135. Woo, H.M., 2017. Solar-to-chemical and solar-to-fuel production from CO2 by metabolically engineered microorganisms. Curr. Opin. Biotechnol. 45, 1–7. Wu, G., Yan, Q., Jones, J.A., Tang, Y.J., Fong, S.S., Koffas, M.A.G., 2016. Metabolic burden: cornerstones in synthetic biology and metabolic engineering applications. Trends Biotechnol. 34, 652–664. Wu, S.G., He, L., Wang, Q., Tang, Y.J., 2015. An ancient Chinese wisdom for metabolic
504
Contents lists available at ScienceDirect
Metabolic Engineering journal homepage: www.elsevier.com/locate/meteng
Engineering synergetic CO2-fixing pathways for malate production a,c,1
b,1
a,c
a,c
a,c
T a,c
Guipeng Hu , Jie Zhou , Xiulai Chen , Yuanyuan Qian , Cong Gao , Liang Guo , ⁎ Peng Xud, Wei Chena, Jian Chena,c, Yin Lib, Liming Liua,c, a
State Key Laboratory of Food Science and Technology, Jiangnan University, 1800 Lihu Road, Wuxi 214122, China CAS Key Laboratory of Microbial Physiological and Metabolic Engineering, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China Key Laboratory of Industrial Biotechnology, Ministry of Education, Jiangnan University, Wuxi, China d Chemical Biochemical and Environmental Engineering, University of Maryland Baltimore County, Baltimore, MD, USA b c
A R T I C LE I N FO
A B S T R A C T
Keywords: CO2 fixation Malate production ATP balance Pathway engineering
Increasing the microbial CO2-fixing efficiency often requires supplying sufficient ATP and redirecting carbon flux for the production of metabolites. However, addressing these two issues concurrently remains a challenge. Here, we present a combinational strategy based on a synergetic CO2-fixing pathway that combines an ATP-generating carboxylation reaction in the central metabolic pathway with the ATP-consuming RuBisCO shunt in the carbon fixation pathway. This strategy provides enough ATP to improve the efficiency of CO2 fixation and simultaneously rewires the CO2-fixing pathway to the central metabolic pathway for the biosynthesis of chemicals. We demonstrate the application of this strategy by increasing the CO2-fixing rate and malate production in the autotroph Synechococcus elongatus by 110% and to 260 μM respectively, as well as increasing these two factors in the heterotrophic CO2-fixing Escherichia coli by 870% and to 387 mM respectively.
1. Introduction
fixation in vitro was built and optimized using enzyme engineering and metabolic proofreading, in which an efficient ATP recycling system was devised to meet the need for ATP (Schwander et al., 2016). These pathways can realize the high efficiency of an ATP supply, but it is difficult to rewire them into central metabolic pathways in vivo (Gong and Li, 2016). To circumvent this difficulty, the power of natural CO2fixing pathways has been harnessed. First, several intermediate metabolites have been engineered as the branching points for value-added chemical production in autotrophic microbes, such as 2,3-butanediol from pyruvate (Kanno et al., 2017; McEwen et al., 2016; Oliver et al., 2013), lipid from acetyl-CoA (Ajjawi et al., 2017), and ethylene from oxoglutarate (Xiong et al., 2015). Second, since the genetic tools for autotrophic microbes remain limited, some natural CO2-fixing pathways, such as the Calvin-Benson-Bassham (CBB) cycle and 3-hydroxypropionate bicycle, have been transplanted into heterotrophs to pave the way for the creation of completely synthetic autotrophs (Antonovsky et al., 2016; Kerfeld, 2016; Mattozzi et al., 2013). Although these CO2-fixing pathways are more easily rewired into central metabolic pathways than synthetic pathways, they still exhibit ATP deficiencies. In addition, researchers have paid attentions to the WoodLjungdahl (WL) pathway, which is the most ATP-efficient among the six natural CO2-fixing pathways. However, the genetic tools available for the microbes that harbor the WL pathway remain limited to date, and
Elevated atmospheric CO2 concentrations have been caused by human activities globally. CO2 emissions for 2007 were 28.8 Gt, which is expected to grow to 40.3 Gt by 2030 and to 50 Gt by 2050 (Venkata Mohan et al., 2016). To decrease atmospheric CO2 concentrations, there is an urgent need to reduce CO2 emissions and develop CO2 sequestration strategies. Conventional methods for CO2 sequestration, including CO2 capture (Leung et al., 2014), CO2 separation (Chou, 2013), and CO2 storage (Boot-Handford et al., 2014), have obvious shortfalls, such as energy-intensive processes and high operational costs (Hicks et al., 2017). To overcome these disadvantages, biological CO2-fixing techniques have garnered much attention, and such biorefinery approaches have been used to produce value-added chemicals (Liao et al., 2016). However, a remaining key issue is how to enhance biological CO2-fixing efficiency by replenishing ATP deficiencies in CO2-fixing pathways and engineering efficient biosynthetic pathways for valueadded chemical production. To explore more efficient CO2-fixing pathways able to reduce the demand for ATP, several synthetic carbon fixation pathways, which combine existing metabolic building blocks from various organisms, were computationally identified based on pathway kinetics and thermodynamics (Bar-Even et al., 2010). Then, a synthetic pathway for CO2
⁎
1
Corresponding author at: State Key Laboratory of Food Science and Technology, Jiangnan University, 1800 Lihu Road, Wuxi 214122, China. E-mail address: [email protected] (L. Liu). These authors contributed equally to this work.
https://doi.org/10.1016/j.ymben.2018.05.007 Received 1 March 2018; Received in revised form 10 May 2018; Accepted 10 May 2018 1096-7176/ © 2018 Published by Elsevier Inc. on behalf of International Metabolic Engineering Society.
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
Fig. 1. Overview of ATP balance in CO2-fixing pathways to improve malate production. ATP balance was achieved by introducing the PCK pathway. (a) Stoichiometric calculations of ATP balance for CO2 fixation with malate production in S. elongatus. (b) Stoichiometric calculations of ATP balance for CO2 fixation with malate production in E. coli. Abbreviations: hv, photon energy; G-6-P, glucose 6-phosphate; Ru5P, ribulose-5-phosphate; RuBP, ribulose-1,5-bisphosphate; 3PGA, 3-phosphoglyceric acid; PEP, phosphoenolpyruvate; OAA, oxaloacetate; MAL, malate; PRK, phosphoribulokinase; RuBisCO, ribulose-1,5-bisphosphate carboxylase/oxygenase; PCK, phosphoenolpyruvate carboxykinase; MDH, malate dehydrogenase; CBB, Calvin-Benson-Bassham cycle.
2. Materials and methods
this CO2-fixing pathway is strictly anaerobic (Schuchmann and Muller, 2014). These findings indicate the importance of developing novel strategies to improve CO2 biorefineries by considering the balance between efficient ATP supply and major metabolic pathway rewiring. In this study, we provide a practical solution for this balance. Based on stoichiometric calculations, we first identified a synergetic CO2fixing pathway that combines the ATP-generating phosphoenolpyruvate carboxykinase (PCK) pathway with the ATP-consuming RuBisCO shunt (or the CBB cycle) (Fig. 1). Then, the enzymes for building the PCK pathway were screened using a cell-free system. We subsequently demonstrated the application of the synergetic CO2-fixing pathway in Synechococcus elongatus, which resulted in a higher CO2-fixing rate (RCf) and malate titer. Finally, the synergetic CO2-fixing pathways were reconstructed in Escherichia coli, and the RCf and yield of malate form glucose were both significantly improved.
2.1. Strains and plasmids All strains and plasmids used in this study are listed in Supplementary Table 1 and Supplementary Table 2. S. elongatus UTEX 2973 and E. coli MG1655 were used as autotrophic and heterotrophic hosts, respectively, for investigating CO2 fixation and malate production. E. coli DH5a was used to construct vectors for engineering S. elongatus and E. coli. The plasmid pEM (a modified pETM6 (Xu et al., 2012) plasmid formed by replacing the promoter and multiple cloning enzyme cleavage sites with those from pQE-80l-kan) was used to provide a platform for the combinatorial optimization of CA and RuBisCO in E. coli. During the construction of S. elongatus mutants, cultures were grown in BG-11 medium with 50 mM NaHCO3 at 38 °C, and illumination intensity was controlled at 200 μmol photons/m2/s. During the construction of E. coli mutants, cultures were grown aerobically at 497
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
2.5. Carbon fixation rate determination
30 °C, 37 °C, or 42 °C in Luria–Bertani (LB) broth. Ampicillin, kanamycin, erythromycin, or chloramphenicol were added appropriately.
A GC-2014 gas chromatograph (Shimadzu, Kyoto, Japan) equipped with a TCD detector was used to calculate the net CO2 uptake rate (n = 3). The RTX-QBOND capillary column (30 m; 0.32 mm inside diameter; 10 mm film thickness; RESTEK Corporation, Bellefonte, PA, USA) was used to quantify CO2 in samples. The detailed process for determining net CO2 uptake rate by S. elongatus strains has been described by a previous study (Zhou et al., 2016). To calculate the net CO2 uptake rate by E. coli strains, cells in mid-log phase were harvested, washed with M9 medium and resuspended in 80 mL fresh M9 medium with 5% glucose and 20 mM NaHCO3 at a final OD600 = 3. Then the 80 mL cell suspension was transferred to a serum bottle with a volume of 100-mL and cultivated for 2 h. The total inorganic carbon of cell suspension was released by injecting concentrated HCl, and the concentration of CO2 in the headspace gas was quantified using gas chromatograph.
2.2. Genetic manipulations For the construction of S. elongatus mutants, double crossover homologous recombination based on tri-parental conjugation was used to integrate exogenous genes into the chromosome of S. elongatus (Yu et al., 2015). The conjugal strain was constructed by transforming the conjugal plasmid pRL443, and cargo strains were constructed by transforming pBR322 carrying the gene of interest into E. coli HB101 competent cells harboring the helper plasmid pRL623. Then, the conjugal strain (OD600 = 1.5–2.0), cargo strain (OD600 = 1.5–2.0), and S. elongatus strain (OD730 = 0.5–0.8) were centrifuged, washed, resuspended with BG-11 medium, mixed at a ratio of 1:1:2, and incubated for 4 h. These mixed cells were cultured in BG-11 + 5% LB (v/v) agar plates for 12 h and then transferred to BG-11 agar plates containing selective antibiotics. All primers used for genetic manipulations in S. elongatus are listed in Supplementary Table 3. For the construction of E. coli mutants, Red recombinase technology was used to knockout target genes (Dong et al., 2017). To fine-tune the expression level of CA and RuBisCO (which are come from Synechocystis PCC 6803), all exogenous genes were integrated into one plasmid (Supplementary Fig. 9). First, each gene was inserted into the pEM using the restriction sites SacI and PstI, thus forming component vectors. Then, the restriction sites AvrII, SpeI, and SalI were used for assembling these components to form pEM-CF5x. To replace the RBS of pEM-eGFP, inverse primers containing different RBS sequences were designed, and the restriction sites EcoRI and BamHI were used for replacing the original RBS of pEM-eGFP. All primers used for genetic manipulations in E. coli are listed in Supplementary Table 4 and Supplementary Table 5.
2.6. ATP assay To prepare samples for ATP assay, cells of S. elongatus (Kanno et al., 2017) and E. coli (Gong et al., 2015) in the exponential growth phase were harvested and washed. An ATP Assay Kit S0027 (Beyotime Biotechnology, Shanghai, China) was used to determine ATP content (n = 3). The concentration of intracellular ATP was calculated according to an ATP standard curve. 2.7. Production of malate For determination of malate production by S. elongatus, strains were initially grown in BG-11 medium containing 50 mM NaHCO3 at 38 °C, and illumination intensity was controlled at 200 μmol photons/m2/s (n = 3 per time point). For determination of malate production by E. coli, strains were initially grown aerobically in LB broth (n = 3 per time point). The cultures were induced for 8 h using 0.1 mM IPTG at mid-log phase. Then, cells were harvested, concentrated, and resuspended in fresh M9 medium with 5% glucose and 20 mM NaHCO3 at a final OD600 = 9. These cultures were transferred to a 100-mL serum bottle with an 80mL working volume. Anaerobic conditions were achieved by purging the headspace with N2 gas and sealing the bottle with a rubber plug (Bogorad et al., 2013). The initial pH was controlled at 7.0, and 1.6 mL of 1.0 M NaHCO3 was added every 12 h. For fed-batch fermentation in a 1.0-L bioreactor, cells were cultured at an initial OD600 = 27. The working volume was controlled at 0.8 L, and the pH of the medium was controlled at 7.0.
2.3. Enzyme assays Cells were harvested by centrifugation (4 °C, 5 min, 5000× g) during mid-log phase and then washed and resuspended in cold 100 mM Tris-HCl buffer (pH 7.5 or 8.0). After cell disruption using FastPrep-24 (MP Biomedicals, Solon, OH, USA), preparations were clarified by centrifugation (13,000× g for 10 min). Protein concentration was measured by the BCA method (Pierce, Rockford, IL, USA). We performed the experiment in triplicate (n = 3). RuBisCO activity was measured as described by Bonacci et al. (Bonacci et al., 2012). First, the reaction buffer containing 100 mM Tris (pH 8.0), 20 mM MgCl2, 10 mM KCl, 25 mM NaHCO3, 3.5 mM ATP, 5 mM creatine phosphate, 2 mM DTT, 5 U/mL 3-phosphoglycerate kinase, 5 U/mL creatine phosphokinase, 5 U/mL glyceraldehyde 3-phosphate dehydrogenase, and appropriate amounts of enzyme solution was activated for 10 min at 30 °C. Then, the reaction was initiated by adding 0.5 mM RuBP and 0.25 mM NADH and was immediately monitored at 340 nm (30 °C) in a plate reader (Wallac/Spectromax) for 5 min. PRK activity was measured as described by Masahiro Kannoet al.(Kanno et al., 2017). PCK activity was determined as described by Van der Werf et al. (VanderWerf et al., 1997). MDH activity was measured at A340 to monitor NADH oxidation (Chen et al., 2013).
2.8. High-performance liquid chromatography (HPLC) detection An HPLC system (Dionex UltiMate 3000 Series; Thermo Scientific, Waltham, MA, USA) was used for determining concentrations of glucose and organic acid. This involved an Aminex HPX-87H column (7.8× 300 mm; Bio-Rad Laboratories, Inc., Hercules, CA, USA) at 35 °C with 0.05 mM sulfuric acid as the mobile phase. The injection volume was 10 μL, and the flow rate was 0.6 mL/min.
2.4. Cell free synthesis of malate from PEP 2.9. Ultra-performance liquid chromatography–mass spectrometry (UPLCMS) detection
Cell free reactions were performed at a volume of 25 μL in 1.5 mL tubes and incubated at 37 °C. The standard reaction contained the following components: 50 mM Tris-HCl (pH 7.5), 10 mM PEP, 10 mM NADH, 10 mM ADP, 10 mM NaH13CO3 and 2.5 mM MgCl2. PCKs and MDHs were individually overexpressed and purified from E. coli BL21 (DE3). Purified PCK and MDH were both added with the final protein concentration of 0.2 μM. The reactions were incubated for 30 min at 37 °C.
Samples were centrifuged for 5 min at 12,000 ×g and diluted 1:10 in H2O for injection into the UPLC-MS system. The relative abundance of malate was detected using a Dionex UltiMate 3000 Q-TOF LC-MS system equipped with an electrospray ionization source set to negative ionization mode. UPLC-MS analysis was performed as described previously (Schwander et al., 2016). 498
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
48 photons + 8 CO2 + 6H2O + 0.42 ADP + 0.42 Pi = 2 malate + 0.42 ATP + 6 O2 (8)
2.10. UPLC–tandem mass spectrometry (UPLC-MS/MS) detection To prepare samples for UPLC-MS/MS detection, cells of S. elongatus (Kanno et al., 2017) and E. coli (Gong et al., 2015) in the exponential growth phase were harvested and washed. The detailed MRM parameters for 3-PGA and 1,5-RuBP analysis were optimized as described previously (Gong et al., 2015).
5·5 glucose + 15 CO2 + 6 NADH + 6H+ = 12 malate +3H2O + 6 NAD+ (9)
3.2. Screening enzymes for the PCK pathway To build an efficient PCK pathway for supplying ATP for CO2 fixation with malate biosynthesis, efficient PCK and MDH enzymes were needed. Based on recent studies, the pck genes from E. coli (Zhang et al., 2011), Actinobacillus succinogenes (VanderWerf et al., 1997), and Agrobacterium (Fu et al., 2010) and the mdh genes from E. coli (Zhang et al., 2011), A. succinogenes (VanderWerf et al., 1997), and Rhizopus oryzae (Zhang et al., 2012) were selected, overexpressed, and purified. Then, the specific enzyme activity and kinetic parameters of these PCKs and MDHs were measured. AsPCK had the best specific enzyme activity and kcat/Km, reaching 13 ± 0.8 U mg protein−1 and 1.7 × 105 M−1 s−1, respectively (Supplementary Table 6). Similarly, AsMDH also possessed the best corresponding parameters at 120 ± 9.3 U mg protein−1 and 1.4 × 106 M−1 s−1, respectively (Supplementary Table 6). Thus, AsPCK and AsMDH were selected to construct the PCK pathway for improving malate production and CO2 fixation. To test the efficiency of AsPCK and AsMDH for converting PEP to malate, these two enzymes were added to cell-free cocktails containing 10 mM substrate (PEP), catalytic amounts of cofactors (10 mM NADH and 10 mM ADP), and ions (10 mM NaH13CO3 and 2.5 mM MgCl2) to mimic the cytoplasmic environment. As a result, 6.7 ± 0.4 mM malate was detected in the supernatant after 30 min. These results indicated that AsPCK and AsMDH could be used for efficiently transforming PEP to malate and further building an efficient PCK pathway in living cells.
3. Results 3.1. Calculating biological CO2 fixation based on stoichiometry In autotrophic microbes such as S. elongatus, CO2 is fixed to sugar phosphate through the CBB cycle in dark reactions, requiring 3 ATP/2 NADPH. However, only 2.57 ATP/2 NADPH are generated by light reactions (Zhou et al., 2016). Thus, the stoichiometry of glucose-6phosphate (G-6-P) production from CO2 can be calculated by equation ① (Supplementary Note 1). Assuming that all the carbons can be channeled to malate, the optimal stoichiometry for converting G-6-P to malate is normally calculated by equation ②. Thus, combining equations ① + ②, the stoichiometry of malate from CO2 can be calculated by equation ③, meaning that an extra 1.58 ATP are required for this conversion. In heterotrophic microbes such as E. coli, the optimal stoichiometry of malate production from glucose is normally described by equation ④. It has been demonstrated that the CBB cycle enzymes [phosphoribulokinase (PRK) and ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO)] can be used for building a CO2-fixing bypass (termed the RuBisCO shunt) in E. coli (Antonovsky et al., 2016; Gong et al., 2015). When the RuBisCO shunt is constructed in E. coli, more CO2 can be transformed to malate according to equation ⑤ (Supplementary Note 2), but extra ATP are also needed for this conversion (Supplementary Fig. 7). In summary, ATP deficiency is a common issue for CO2 fixation in both autotrophic and heterotrophic CO2-fixing microbes.
3.3. Engineering the PCK pathway in S. elongatus We first demonstrated this strategy in the autotrophic microbe S. elongatus UTEX 2973. To reconstruct the PCK pathway in this microbe, the Aspck and Asmdh genes were integrated into the chromosome of S. elongatus with strong promoters Pcpc6803 and Pcpc560, respectively, obtaining strain SH006 (Fig. 2a,b). As a result, the RCf of strain SH006 was 73% ± 5% higher than that of wild-type S. elongatus (Fig. 2d). Simultaneously, malate production was increased from below the detection limit (~ 10 μM) to 200 ± 19 μM in strain SH006 after 9 days of cultivation (Fig. 2e). To test whether the CO2 fixation of the CBB cycle was improved by integrating AsPCK and AsMDH, the enzyme activities of PRK and RuBisCO were analyzed. Increases in the activities of PRK (1.4 ± 0.1-fold) and RuBisCO (1.3 ± 0.1-fold) were confirmed in mutant SH006 compared to activity levels in wild-type S. elongatus (Fig. 2f). Moreover, we observed that the concentrations of intracellular RuBP and 3-PGA were increased by 63% ± 4% and 51% ± 3%, respectively (Fig. 2g). These results demonstrated that engineering the PCK pathway not only improved malate production, but also enhanced the flux of the CBB cycle. Furthermore, the concentration of intracellular ATP (CATP) in strain SH006 was 1.7 ± 0.12 μmol gDCW−1, which was 85% ± 6% higher than that of wild-type S. elongatus (Fig. 2d), indicating that an enhanced ATP supply was beneficial for improving CO2 fixation in S. elongatus. To further promote CO2 fixation in S. elongatus through fine-tuning the ATP supply, AsPCK and AsMDH were expressed at different levels. First, gene expression was divided into two levels using the strong promoters Pcpc6803 or Pcpc560 and the weak promoter Pcpc300. Next, a series of AsPCK and AsMDH expression cassettes with various strengths were designed and assembled in S. elongatus to obtain the best ATP distribution for enhancing CO2 fixation (Fig. 2c). Finally, we found that controlling the strength of AsPCK at a low level and AsMDH at a high level resulted in a 17% ± 0.5% decrease in CATP compared with that of
48 photons + 6 CO2 + 5H2O + 2·58 ATP = 1 G-6-P + 2·58 ADP + 1·58 Pi + 6 O2 (1) G-6-P + 2 CO2 + H2O + ADP = 2 malate + ATP
(2)
48 photons + 8 CO2 + 6H2O + 1·58 ATP = 2 malate + 1·58 ADP + 1·58 Pi + 6 O2 (3) glucose + 2 CO2 = 2 malate
(4) +
2·5 glucose + 9 CO2 + 6 ATP + 6 NADH + 6H + 6 NAD+
= 6 malate +3H2O (5)
In order to promote CO2 fixation, an effective strategy is to rewire metabolic pathways to rescue this ATP deficiency and simultaneously transfer the fixed carbon to the production of a high-value chemical. Here, we selected malate as the target chemical for our calculations. An extra ATP can be generated by overexpressing PCK and malate dehydrogenase (MDH) for malate production, resulting in the theoretical maximum yield of malate from glucose/G-6-P (2 mol/mol) (Chen et al., 2017). Consequently, the optimal stoichiometry for the conversion of G6-P/glucose to malate can be improved to equation ⑥/⑦. Thus, when the ATP deficiency is replenished, the stoichiometry of malate production in S. elongatus is transformed into equation ⑧ = ① + ⑥ (Fig. 1a), and in CO2-fixing E. coli is transformed into equation ⑨ = ⑤ + 3 × ⑦ (Fig. 1b). In conclusion, engineering the PCK pathway for malate production can enhance CO2 fixation. In turn, this enhanced CO2 fixation can improve malate production, thus resulting in a mutually reinforcing cycle. G-6-P + 2 CO2 + H2O + 3 ADP + 2 Pi = 2 malate + 3 ATP
(6)
glucose + 2 CO2 = 2 malate + 2 ATP
(7) 499
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
Fig. 2. Engineering the PCK pathway improves CO2 fixation and malate production in S. elongatus. (a, b) Schematic representation of recombination to integrate Asmdh and Aspck. (c) A series of PCK and MDH expression cassettes were designed at different expression levels. (d) CO2-fixing rate and concentration of intracellular ATP in wild-type and engineered S. elongatus strains. (e) Time course of malate production by wild-type and engineered S. elongatus strains. (f) Crude enzyme activities of PRK and RuBisCO in wild-type S. elongatus and strain SH006. (g) Concentrations of intracellular RuBP and 3-PGA in wild-type and S. elongatus engineered strains. *P < 0.05 as determined by t-test; N = 3 biological replicates; the s.d. is shown as black error bars. Abbreviations are defined in the Fig. 1 legend.
(278 mM) was fed to strains MG1655 and FH010. We observed that the ratio of 13C-labeled 3-PGA to unlabeled 3-PGA (13C-3-PGA/3-PGA) was increased from 2% ± 0.2% (strain MG1655) to 8% ± 0.6% (strain FH010) (Fig. 3b). These results indicated that the RuBisCO shunt in E. coli was able to capture 13CO2 and convert it to 3-PGA (Fig. 4a). Next, 11 genes in CO2-fixing E. coli FH010 were knocked out, resulting in strain FH019 (ΔptsG ΔfrdBC ΔfumABC ΔmaeAB ΔaspC ΔldhA ΔadhE) for malate production (Supplementary Fig. 8 and Supplementary Note 3). AsPCK and AsMDH were introduced into strain FH019, obtaining strain FH089. As a result, the RCf in strain FH089 was 490% ± 32% higher than that of strain FH019 (Fig. 3d). This enhancement led to a 620% ± 41% increase in the yield of malate from glucose (Ym/g), with the titer of malate increasing from 7.9 ± 0.5 mM to 106 ± 8.9 mM (Fig. 3c and Supplementary Table 8). In contrast, strain FH029 (FH009 expressing only AsPCK and AsMDH) only produced 30 ± 1.9 mM malate; the RCf and Ym/g of strain FH029 were also 55% ± 3% and 53% ± 3% lower than those of strain FH089, respectively (Supplementary Table 7 and Fig. 3c,d). Even when CA was further expressed in strain FH029 (resulting in strain FH039), the titer of malate was increased to only 42 ± 4.1 mM, with RCf and Ym/g values lagging behind those of strain FH089 by 34% ± 2% and 42% ± 3%, respectively (Supplementary Table 7 and Fig. 3c,d). These results
strain SH006 but a 19% ± 0.9% increase in RCf, reaching 310 ± 22 mg gDCW−1 h−1 (Fig. 2d). Furthermore, the highest malate titer achieved was 260 ± 21 μM (Fig. 2e). These results demonstrated that CO2 fixation and malate production in S. elongatus could be further enhanced by optimizing the ATP supply to an appropriate level.
3.4. Engineering synergetic CO2-fixing pathway for malate production in E. coli To determine whether the PCK pathway can improve CO2 fixation by RuBisCO in heterotrophic microbes, we first constructed a CO2fixing E. coli harboring the RuBisCO shunt. The enzymes PRK and RuBisCO from Synechocystis sp. PCC 6803 were introduced into E. coli MG1655 to build the CO2-fixing bypass. Carbonic anhydrase (CA) from Synechocystis sp. PCC 6803 was expressed simultaneously to provide sufficient CO2 for RuBisCO, obtaining the CO2-fixing E. coli strain FH010. To test the function of this CO2-fixing shunt, we analyzed the concentration of intracellular 1,5-RuBP, which is the specific metabolite of the RuBisCO shunt in E. coli. As a result of building the RuBisCO shunt, this parameter was increased from below the detection limit (~0.01 μmol gDCW−1) to 2.1 ± 0.1 μmol gDCW−1 (Supplementary Fig. 6). Furthermore, NaH13CO3 (20 mM) along with unlabeled glucose 500
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
Fig. 3. Tuning the balance of CO2-fixing pathways improves CO2 fixation and malate production in E. coli. (a) Characteristics of strains used in this section; (b) Concentration of intracellular 3-PGA and 13C-3-PGA and the ratio of 13C-3-PGA/3-PGA in wild-type and engineered E. coli strains. (c) Time course of malate production and molar yield of malate from glucose by strains FH019, FH089, and FH389. (d) Concentration of intracellular ATP and CO2-fixing rate by strains FH019, FH089, and FH389. (e) A series of CA and RuBisCo expression cassettes were designed at different expression levels. (f) The yields of malate from glucose by strains with different CA and RuBisCO expression cassettes. *P < 0.05 as determined by the t-test; N = 3 biological replicates; the s.d. is shown as black error bars.
strengths: high (H), moderate (M), and low (L) (Supplementary Fig. 10 and Fig. 3e). Expressing RuBisCO at a low level and CA at a high level resulted in the highest RCf and Ym/g of 49 ± 3.1 mg gDCW−1 h−1 (Figs. 3d) and 1.37 ± 0.12 mol/mol (Fig. 3f), respectively. To further reveal the effect of engineering synergetic CO2-fixing pathways on carbon metabolism, we analyzed the carbon balance of strains FH039, FH089, and FH389. The partial apparent carbon efficiencies (CPro/CGlu) of strains FH039, FH089, and FH389 were 60%, 78%, and 99%, respectively (Supplementary Table 7). These results indicated that introducing the RuBisCO shunt was beneficial for redirecting carbon flux to organic acids (especially for malate in our engineered strains), and this effect was further improved by tuning the expression of RuBisCO and CA. Finally, strain FH389 produced 387 ± 43 mM malate in a bioreactor, with the Ym/g reaching 1.46 ± 0.12 mol/mol (Supplementary Fig. 12), which is the highest from engineered E. coli (Supplementary Table 8).
demonstrated that the RuBisCO shunt contributed to malate production. To further demonstrate that the flux of the RuBisCO shunt was improved by the introduction of AsPCK and AsMDH, 20 mM NaH13CO3 was used to supply CO2 for strains FH019 and FH089. As expected, the ratio of 13C-3-PGA/3-PGA in strain FH089 was 1.9 ± 0.1-fold higher than that in strain FH019 (Fig. 3b). Meanwhile, the relative abundance of malate-m2 (malate with two 13C atoms) in FH089 was increased from 9% ± 0.6–17% ± 0.9% (Fig. 4d,e). These results indicated that reconstructing the PCK pathway was effective at improving the ATP supply for the RuBisCO shunt, enhancing overall CO2 fixation and malate production in E. coli. Although the CATP in strain FH089 was increased to reinforce the overall CO2 fixation (Fig. 3d), the distribution of ATP could be further optimized through fine-tuning the expression levels of CO2 capture enzymes. In order to achieve a more appropriate distribution of ATP and redirect more carbon to malate, RuBisCO and CA were selected and expressed under three different ribosome-binding sequence (RBS) 501
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
Fig. 4. Effect of CO2 fixation on malate production in E. coli. (a) Schematics of carbon atom transits via the RuBisCO shunt and the PCK pathway. 13CO2 fixed into 3-PGA by the RuBisCO shunt can be redirected into the PCK pathway for biosynthesis of malate in the engineered E. coli. Abbreviations are defined in Fig. 1 legend. The labeled (13C) and unlabeled carbon atoms are represented by closed and open circles, respectively. (b–f) Relative abundance of mass isotopomers for malate in engineered strains. Malate-m0 represents malate without 13C atoms, malate-m1 represents malate with one 13C atom, and malate-m2 represents malate with two 13C atoms. Minimal malate-m1 was detected in the control strain (FH009 NaHCO3), mainly owing to the small fraction of 13C isotopes present in all natural 12Ccontaining compounds. The malate-m0 in the engineered strain presumably originated from other metabolic pathways or was synthesized from the unlabeled PEP combined with unlabeled CO2 released from decarboxylic reactions. N = 3 biological replicates; the s.d. is shown as black error bars.
4. Discussion
Recycling CO2 into central metabolic pathways is a promising approach for expanding the utilization of CO2 fixation to improve the yield of target metabolites. Recently, the power of autotrophs has been exploited in heterotrophic microbes to meet this goal (Claassens et al., 2016). For instance, the RuBisCO shunt was introduced into E. coli MZ for organic acid production (Yang et al., 2016), but the yield of products were not able to meet the requirement for industrial production. In this study, we rewired the RuBisCO shunt with the PCK pathway for malate production in E. coli. Thus, CO2 fixation capability was effectively enhanced, and an impressive Ym/g (1.46 mol/mol) was achieved, representing a 3% increase over the highest reported Ym/g (1.42 mol/ mol) in malate-producing E. coli XZ658 (using the PCK pathway by not the RuBisCO shunt) (Zhang et al., 2011). Furthermore, the strategy of engineering a synergy between an ATP-producing biosynthetic pathway and ATP-consuming CO2-fixing pathway can be utilized for the production of other chemicals in E. coli. An efficient ATP supply is required for both biological CO2 fixation and chemical biosynthesis. In this study, we rewired and tuned the balance of the ATP-producing PCK pathway and ATP-consuming RuBisCO shunt to redirect intracellular ATP for the biosynthesis of chemicals and CO2 fixation. According to recent studies, several principles can be proposed to obtain efficient microbial cell factories: (1) get a clear understanding of the entire carbon and energy metabolisms, for example, by dividing ATP cost into growth-associated maintenance and non-growth-associated loss using flux balance analysis (Wu et al., 2015); (2) take advantage of native pathways for product synthesis, for example, by recruiting ATP-generating pathways for chemical production (Zhang et al., 2009); (3) avoid extensive pathway reconstruction, for example, by combining microbial production with chemical conversions to reduce biosynthetic steps (Wheeldon et al., 2017); (4) develop novel non-model microbial workhorses with desired traits in terms of energy metabolism, for example, by exploiting genetic tools for cyanobacteria, which acquire extra energy from light (Xiong et al.,
Microbial production of chemicals from CO2 faces two challenges: (i) an economic and efficient ATP supply for CO2-fixing pathways and (ii) the rewiring of CO2-fixing pathways with biosynthetic pathways. In this study, the PCK pathway for malate production with ATP generation was engineered into autotrophic S. elongatus to enhance RCf and heterotrophic CO2-fixing E. coli to increase Ym/g. Using this approach, CO2 fixation capability and malate biosynthesis were simultaneously boosted to new levels. This study provides a novel method for replenishing ATP to enhance CO2 fixation and engineering CO2-fixing pathway to improve the production of value-added chemicals. CO2-fixing efficiency is important for chemical production in autotrophic microbes. Recent studies have engineered autotrophic microbes for synthesizing chemicals and biofuels from CO2 (Nybo et al., 2015; Oliver and Atsumi, 2014; Oliver et al., 2016; Woo, 2017). However, production performance remains far below industrial feasibility, partly due to the low efficiency of the ATP supply for the biosynthesis of compounds in autotrophic cell factories (Claassens et al., 2016). In addition, the production of C4-dicarboxylic acids such as succinate (Huang et al., 2016; Li et al., 2016) has lagged far behind the “workhorses” of biotechnology like E.coli, due to the inherent weak flux of the relevant pathways in autotrophic microbes (Angermayr et al., 2015). Here, we rewired an ATP-producing pathway into a cyanobacterium to enhance the metabolic flux for malate biosynthesis, resulting in a cyanofactory with higher efficiencies of CO2 fixation and malate production. It is noteworthy that the enhancement of CO2 fixation was not only caused by the specific biosynthetic pathway (PCK carboxylation), but also by the CBB cycle. Therefore, this strategy can be used to reinforce the efficiency of autotrophic cell factories and engineer autotrophic microbes for the production of other value-added chemicals in the future, such as aspartate from the PEP-OAA pathway and lactate from the PEP-pyruvate pathway (Xu et al., 2017). 502
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
experiments. G.H., J.Z., X.C., Y.Q., C.G., L.G., P.X., W.C., J.C. and Y.L. analyzed the results.
2017). Synthetic pathways can be effectively constructed in microbes through current metabolic engineering and synthetic biology techniques. By rewiring synthetic pathways into microbial metabolisms, diverse natural and non-natural products have been synthesized efficiently (Clomburg et al., 2017). Obviously, synthetic pathways can enable us to perform important tasks that are not possible in nature. However, energy molecules and carbon building blocks are both required for constructing and operating synthetic pathways (Wu et al., 2016). Thus, the problem of metabolic burden may be a limiting factor for harnessing the power of synthetic pathways. Generally, cyanobacteria are more sensitive to metabolic burdens than E. coli because they have to synthesize large amounts of protein for photosynthesis. In this study, we introduced synthetic pathways by inserting exogenous genes into the chromosome of S. elongatus to reduce the metabolic burden. In other words, we should exploit the advantages and avoid the disadvantages of synthetic pathways, thus achieving more efficient microbial cell factories. More efficient CO2-fixing pathways can be achieved through synthetic biology (SynBio). However, several drawbacks of this process need to be considered: (1) synthetic CO2-fixing pathways may be hardly connected to the central carbon metabolism of organisms (Gong and Li, 2016); (2) even if synthetic pathways can be rewired into the central carbon metabolism, large metabolic burdens may be induced since energy molecules and carbon building blocks are both required for constructing and operating synthetic pathways (Wu et al., 2016); and (3) microorganisms have evolved robust metabolic and regulatory networks to survive and grow in specific environments, and therefore more effort should be made to use these synthetic CO2-fixing pathways for further biological conversions. While nature has performed biological engineering experiments for billions of years, it is possible to improve upon natural systems. For example, the CO2-fixing rate of the synthetic CETCH cycle is higher than those of all natural CO2-fixing pathways that have so far been discovered (Schwander et al., 2016). Similarly, the design of a synthetic light-harvesting system for non-photosynthetic bacteria enabled a higher efficiency of solar-to-chemical synthesis than that observed among naturally photosynthetic bacteria (Sakimoto et al., 2016). Furthermore, several strategies can be used to improve the outcome of SynBio projects, including: (1) combining -omic data with protein design de novo to build more efficient CO2-fixing pathways that can be easily rewired into living cells (Siegel et al., 2015); (2) harnessing the power of adaptive evolution, 13C-MFA, and genome-scale models to decrease metabolic burdens caused by synthetic pathways (Antonovsky et al., 2016; G and M, 2015); and (3) providing sugars (e.g., glucose) for engineered autotrophs to achieve mixotrophic fermentation, achieving higher productivity for synthetic pathways (Matson and Atsumi, 2018). By overcoming the disadvantages of SynBio, it may be possible to further improve such systems in the future.
Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ymben.2018.05.007. References Ajjawi, I., Verruto, J., Aqui, M., Soriaga, L.B., Coppersmith, J., Kwok, K., et al., 2017. Lipid production in Nannochloropsis gaditana is doubled by decreasing expression of a single transcriptional regulator. Nat. Biotechnol. 35, 647–652. Angermayr, S.A., Rovira, A.G., Hellingwerf, K.J., 2015. Metabolic engineering of cyanobacteria for the synthesis of commodity products. Trends Biotechnol. 33, 352–361. Antonovsky, N., Gleizer, S., Noor, E., Zohar, Y., Herz, E., Barenholz, U., et al., 2016. Sugar synthesis from CO2 in Escherichia coli. Cell 166, 115–125. Bar-Even, A., Noor, E., Lewis, N.E., Milo, R., 2010. Design and analysis of synthetic carbon fixation pathways. Proc. Natl. Acad. Sci. USA 107, 8889–8894. Bogorad, I.W., Lin, T.S., Liao, J.C., 2013. Synthetic non-oxidative glycolysis enables complete carbon conservation. Nature 502, 693–697. Bonacci, W., Teng, P.K., Afonso, B., Niederholtmeyer, H., Grob, P., Silver, P.A., et al., 2012. Modularity of a carbon-fixing protein organelle. Proc. Natl. Acad. Sci. USA 109, 478–483. Boot-Handford, M.E., Abanades, J.C., Anthony, E.J., Blunt, M.J., Brandani, S., Mac Dowell, N., et al., 2014. Carbon capture and storage update. Energy Environ. Sci. 7, 130–189. Chen, X., Gao, C., Guo, L., Hu, G., Luo, Q., Liu, J., et al., 2017. DCEO biotechnology: tools to design, construct, evaluate, and optimize the metabolic pathway for biosynthesis of chemicals. Chem. Rev. 118, 4–72. Chen, X., Xu, G., Xu, N., Zou, W., Zhu, P., Liu, L., et al., 2013. Metabolic engineering of Torulopsis glabrata for malate production. Metab. Eng. 19, 10–16. Chou, C.-t., 2013. Carbon dioxide separation and capture for Global warming mitigation. J. Adv. Eng. Technol. 1, 2348–2931. Claassens, N.J., Sousa, D.Z., Dos Santos, V.A., de Vos, W.M., van der Oost, J., 2016. Harnessing the power of microbial autotrophy. Nat. Rev. Microbiol. 14, 692–706. Clomburg, J.M., Crumbley, A.M., Gonzalez, R., 2017. Industrial biomanufacturing: the future of chemical production. Science 355, aag0804. Dong, X., Chen, X., Qian, Y., Wang, Y., Wang, L., Qiao, W., et al., 2017. Metabolic engineering of Escherichia coli W3110 to produce L-malate. Biotechnol. Bioeng. 114, 656–664. Fu, W., Lin, J., Cen, P., 2010. Expression of a hemA gene from Agrobacterium radiobacter in a rare codon optimizing Escherichia coli for improving 5-aminolevulinate production. Appl. Biochem Biotechnol. 160, 456–466. Gong, F., Li, Y., 2016. Fixing carbon, unnaturally. Science 354, 830–831. Gong, F., Liu, G., Zhai, X., Zhou, J., Cai, Z., Li, Y., 2015. Quantitative analysis of an engineered CO2-fixing Escherichia coli reveals great potential of heterotrophic CO2 fixation. Biotechnol. Biofuels 8, 86. Hicks, N., Vik, U., Taylor, P., Ladoukakis, E., Park, J., Kolisis, F., et al., 2017. Using prokaryotes for carbon capture storage. Trends Biotechnol. 35, 22–32. Huang, C.H., Shen, C.R., Li, H., Sung, L.Y., Wu, M.Y., Hu, Y.C., 2016. CRISPR interference (CRISPRi) for gene regulation and succinate production in cyanobacterium S. elongatus PCC 7942. Microb. Cell Fact. 15, 196. Kanno, M., Carroll, A.L., Atsumi, S., 2017. Global metabolic rewiring for improved CO2 fixation and chemical production in cyanobacteria. Nat. Commun. 8, 14724. Kerfeld, C.A., 2016. Rewiring Escherichia coli for carbon-dioxide fixation. Nat. Biotechnol. 34, 1035–1036. Leung, D.Y.C., Caramanna, G., Maroto-Valer, M.M., 2014. An overview of current status of carbon dioxide capture and storage technologies. Renew. Sust. Energ. Rev. 39, 426–443. Li, H., Shen, C.R., Huang, C.H., Sung, L.Y., Wu, M.Y., Hu, Y.C., 2016. CRISPR-Cas9 for the genome engineering of cyanobacteria and succinate production. Metab. Eng. 38, 293–302. Liao, J.C., Mi, L., Pontrelli, S., Luo, S., 2016. Fuelling the future: microbial engineering for the production of sustainable biofuels. Nat. Rev. Microbiol. 14, 288–304. Matson, M.M., Atsumi, S., 2018. Photomixotrophic chemical production in cyanobacteria. Curr. Opin. Biotechnol. 50, 65–71. Mattozzi, M., Ziesack, M., Voges, M.J., Silver, P.A., Way, J.C., 2013. Expression of the sub-pathways of the Chloroflexus aurantiacus 3-hydroxypropionate carbon fixation bicycle in E. coli: toward horizontal transfer of autotrophic growth. Metab. Eng. 16, 130–139. McEwen, J.T., Kanno, M., Atsumi, S., 2016. 2,3 Butanediol production in an obligate photoautotrophic cyanobacterium in dark conditions via diverse sugar consumption. Metab. Eng. 36, 28–36. Nybo, S.E., Khan, N.E., Woolston, B.M., Curtis, W.R., 2015. Metabolic engineering in chemolithoautotrophic hosts for the production of fuels and chemicals. Metab. Eng. 30, 105–120. Oliver, J.W., Atsumi, S., 2014. Metabolic design for cyanobacterial chemical synthesis. Photosynth Res. 120, 249–261. Oliver, J.W., Machado, I.M., Yoneda, H., Atsumi, S., 2013. Cyanobacterial conversion of carbon dioxide to 2,3-butanediol. Proc. Natl. Acad. Sci. USA 110, 1249–1254. Oliver, N.J., Rabinovitch-Deere, C.A., Carroll, A.L., Nozzi, N.E., Case, A.E., Atsumi, S., 2016. Cyanobacterial metabolic engineering for biofuel and chemical production.
Acknowledgments This work is supported by the National Natural Science Foundation of China (21676118, 21706095), the Provincial Natural Science Foundation of Jiangsu Province (BK20160163), the Fundamental Research Funds for the Central Universities (JUSRP51611A), the Key Research Program of the Chinese Academy of Sciences (ZDRW-ZS2016-3) and the National First-class Discipline Program of Light Industry Technology and Engineering (LITE2018-08). The authors would like to acknowledge Qingsheng Qi for E.coli MG1655, Min Jiang for A.succinogenes and Ye Ni for plasmid pQE-80l-kan. Author contributions G.H., J.Z., X.C. and L.L. conceived the project and wrote the manuscript. G.H., J.Z., X.C. and Y.Q. designed and performed all the 503
Metabolic Engineering 47 (2018) 496–504
G. Hu et al.
engineering: yin-yang. Microb. Cell Fact. 14, 39. Xiong, W., Cano, M., Wang, B., Douchi, D., Yu, J., 2017. The plasticity of cyanobacterial carbon metabolism. Curr. Opin. Chem. Biol. 41, 12–19. Xiong, W., Morgan, J.A., Ungerer, J., Wang, B., Maness, P.-C., Yu, J., 2015. The plasticity of cyanobacterial metabolism supports direct CO2 conversion to ethylene. Nat. Plants 1, 15053. Xu, N., Ye, C., Chen, X., Liu, J., Liu, L., 2017. Genome-scale metabolic modelling common cofactors metabolism in microorganisms. J. Biotechnol. 251, 1–13. Xu, P., Vansiri, A., Bhan, N., Koffas, M.A., 2012. ePathBrick: a synthetic biology platform for engineering metabolic pathways in E. coli. ACS Synth. Biol. 1, 256–266. Yang, C.H., Liu, E.J., Chen, Y.L., Ou-Yang, F.Y., Li, S.Y., 2016. The comprehensive profile of fermentation products during in situ CO2 recycling by Rubisco-based engineered Escherichia coli. Microb. Cell Fact. 15, 133. Yu, J., Liberton, M., Cliften, P.F., Head, R.D., Jacobs, J.M., Smith, R.D., et al., 2015. Synechococcus elongatus UTEX 2973, a fast growing cyanobacterial chassis for biosynthesis using light and CO2. Sci. Rep. 5, 8132. Zhang, B., Skory, C.D., Yang, S.T., 2012. Metabolic engineering of Rhizopus oryzae: effects of overexpressing pyc and pepc genes on fumaric acid biosynthesis from glucose. Metab. Eng. 14, 512–520. Zhang, X., Jantama, K., Moore, J.C., Jarboe, L.R., Shanmugam, K.T., Ingram, L.O., 2009. Metabolic evolution of energy-conserving pathways for succinate production in Escherichia coli. Proc. Natl. Acad. Sci. USA 106, 20180–20185. Zhang, X., Wang, X., Shanmugam, K.T., Ingram, L.O., 2011. L-malate production by metabolically engineered Escherichia coli. Appl. Environ. Microbiol. 77, 427–434. Zhou, J., Zhang, F., Meng, H., Zhang, Y., Li, Y., 2016. Introducing extra NADPH consumption ability significantly increases the photosynthetic efficiency and biomass production of cyanobacteria. Metab. Eng. 38, 217–227.
Curr. Opin. Chem. Biol. 35, 43–50. G, S., M, C.D., 2015. 13C metabolic flux analysis at a genome-scale. Metab. Eng. 32, 12–22. Sakimoto, K.K., Wong, A.B., Yang, P., 2016. Self-photosensitization of nonphotosynthetic bacteria for solar-to-chemical production. Science 351, 74–77. Schuchmann, K., Muller, V., 2014. Autotrophy at the thermodynamic limit of life: a model for energy conservation in acetogenic bacteria. Nat. Rev. Microbiol. 12, 809–821. Schwander, T., von Borzyskowski, L.S., Burgener, S., Cortina, N.S., Erb, T.J., 2016. A synthetic pathway for the fixation of carbon dioxide in vitro. Science 354, 900–904. Siegel, J.B., Smith, A.L., Poust, S., Wargacki, A.J., Bar-Even, A., Louw, C., et al., 2015. Computational protein design enables a novel one-carbon assimilation pathway. Proc. Natl. Acad. Sci. USA 112, 3704–3709. VanderWerf, M.J., Guettler, M.V., Jain, M.K., Zeikus, J.G., 1997. Environmental and physiological factors affecting the succinate product ratio during carbohydrate fermentation by Actinobacillus sp. 130Z. Arch. Microbiol. 167, 332–342. Venkata Mohan, S., Modestra, J.A., Amulya, K., Butti, S.K., Velvizhi, G., 2016. A circular bioeconomy with biobased products from CO2 sequestration. Trends Biotechnol. 34, 506–519. Wheeldon, I., Christopher, P., Blanch, H., 2017. Integration of heterogeneous and biochemical catalysis for production of fuels and chemicals from biomass. Curr. Opin. Biotechnol. 45, 127–135. Woo, H.M., 2017. Solar-to-chemical and solar-to-fuel production from CO2 by metabolically engineered microorganisms. Curr. Opin. Biotechnol. 45, 1–7. Wu, G., Yan, Q., Jones, J.A., Tang, Y.J., Fong, S.S., Koffas, M.A.G., 2016. Metabolic burden: cornerstones in synthetic biology and metabolic engineering applications. Trends Biotechnol. 34, 652–664. Wu, S.G., He, L., Wang, Q., Tang, Y.J., 2015. An ancient Chinese wisdom for metabolic
504