Metabolic Engineering 23 (2014) 22–33

Contents lists available at ScienceDirect

Metabolic Engineering journal homepage: www.elsevier.com/locate/ymben

Systematic metabolic engineering of Escherichia coli for high-yield production of fuel bio-chemical 2,3-butanediol Youqiang Xu a, Haipei Chu a, Chao Gao a, Fei Tao b, Zikang Zhou b, Kun Li a, Lixiang Li b, Cuiqing Ma a,n, Ping Xu b,nn a

State Key Laboratory of Microbial Technology, Shandong University, Jinan 250100, People's Republic of China State Key Laboratory of Microbial Metabolism & School of Life Sciences & Biotechnology, Shanghai Jiao Tong University, Shanghai 200240, People's Republic of China

b

art ic l e i nf o

a b s t r a c t

Article history: Received 5 September 2013 Received in revised form 15 January 2014 Accepted 3 February 2014 Available online 11 February 2014

The production of biofuels by recombinant Escherichia coli is restricted by the toxicity of the products. 2,3-Butanediol (2,3-BD), a platform and fuel bio-chemical with low toxicity to microbes, could be a promising alternative for biofuel production. However, the yield and productivity of 2,3-BD produced by recombinant E. coli strains are not sufficient for industrial scale fermentation. In this work, the production of 2,3-BD by recombinant E. coli strains was optimized by applying a systematic approach. 2,3-BD biosynthesis gene clusters were cloned from several native 2,3-BD producers, including Bacillus subtilis, Bacillus licheniformis, Klebsiella pneumoniae, Serratia marcescens, and Enterobacter cloacae, inserted into the expression vector pET28a, and compared for 2,3-BD synthesis. The recombinant strain E. coli BL21/pETPT7-EcABC, carrying the 2,3-BD pathway gene cluster from Enterobacter cloacae, showed the best ability to synthesize 2,3-BD. Thereafter, expression of the most efficient gene cluster was optimized by using different promoters, including PT7, Ptac, Pc, and Pabc. E. coli BL21/pET-RABC with Pabc as promoter was superior in 2,3-BD synthesis. On the basis of the results of biomass and extracellular metabolite profiling analyses, fermentation conditions, including pH, agitation speed, and aeration rate, were optimized for the efficient production of 2,3-BD. After fed-batch fermentation under the optimized conditions, 73.8 g/L of 2,3-BD was produced by using E. coli BL21/pET-RABC within 62 h. The values of both yield and productivity of 2,3-BD obtained with the optimized biological system are the highest ever achieved with an engineered E. coli strain. In addition to the 2,3-BD production, the systematic approach might also be used in the production of other important chemicals through recombinant E. coli strains. & 2014 International Metabolic Engineering Society. Published by Elsevier Inc. All rights reserved.

Keywords: Escherichia coli 2,3-Butanediol Metabolic engineering Promoter Fermentation Biofuel

1. Introduction With fluctuations in petroleum prices and political considerations, alternative and renewable energy sources have become attractive (Ragauskas et al., 2006). In recent years, many microorganisms, including microalgae, fungi, and bacteria, have been used for the production of fuel bio-chemicals (Domínguez, de María, 2011; Gross, 2012). Although algae could produce biofuel feedstocks directly from CO2 and sunlight, significant challenges such as large-scale cultivation, harvesting, and product separation restrict its practical utilization (Chen et al., 2011). Saccharomyces cerevisiae is now commonly used in biofuel ethanol production, but it is unable to consume xylose, a very important sugar derived

n

Corresponding author. Fax: þ86 531 88369463. Corresponding author. Fax: þ 86 21 34206723. E-mail addresses: [email protected] (C. Ma), [email protected] (P. Xu).

nn

from lignocellulosic feedstock (Ha et al., 2011). In contrast, Escherichia coli, which is able to use both pentose and hexose sugars from lignocellulose, has been widely used as a model system for biofuel production (Bokinsky et al., 2011). Recently, recombinant E. coli strains have been used to produce various fuel bio-chemicals, e.g., acetone (May et al., 2013), α-pinene (Yang et al., 2013a), 1-butanol (Atsumi et al., 2008; Bokinsky et al., 2011), ethanol (Yomano et al., 1998), isobutanol (Baez et al., 2011), isoprenol (Zheng et al., 2013), and fatty alcohols (Zheng et al., 2012), through exogenous biosynthetic pathways. However, because of the high toxicity of the biofuels, such systems require constant isolation of the produced end-products to achieve high productivity and high yield (Baez et al., 2011). Similar to well-studied biofuels, 2,3-BD is a bulk fuel bio-chemical that can be produced via biotechnological routes (Ji et al., 2011). It has a high heating value of 27,200 J/g and can be used as liquid fuel or fuel additive (Xiao et al., 2012). However, compared with acetone, α-pinene, 1-butanol, isobutanol, isoprenol, and fatty alcohols, 2,3-BD exhibits lower toxicity to

http://dx.doi.org/10.1016/j.ymben.2014.02.004 1096-7176 & 2014 International Metabolic Engineering Society. Published by Elsevier Inc. All rights reserved.

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

microbial systems (Oliver et al., 2013; Zeng and Sabra, 2011). Thus, it could be a promising alternative for biofuel production through recombinant E. coli strains. As a Voges-Proskauer test-negative strain, E. coli lacks a functional 2,3-BD synthesis pathway (Blattner et al., 1997). Thus, an exogenous 2,3-BD synthesis pathway was introduced into E. coli to produce 2,3-BD (Ui et al., 1997). Many low-price substrates such as seaweed hydrolysate and cellodextrin were used to produce 2,3-BD by recombinant E. coli strains (Mazumdar et al., 2013; Shin et al., 2012). However, the output and productivity of 2,3-BD produced by recombinant E. coli strains are not sufficient for potential industrial scale production (Zeng and Sabra, 2011). Until now, the highest amount of 2,3-BD produced by an E. coli strain was only 17.7 g/L (Ui et al., 1997). Many 2,3-BD producers such as Klebsiella pneumoniae, Klebsiella oxytoca, Enterobacter cloacae, and Serratia marcescens have been used to produce 2,3-BD with high yields and productivities through optimization of fermentation conditions or genetic engineering (Celińska and Grajek, 2009; Ji et al., 2011; Kim et al., 2013a). However, the strains used above belong to class 2 microorganisms, which make them unsuitable for industrial scale fermentation (Ji et al., 2011; Kim et al., 2013b). Studies have used engineered Saccharomyces cerevisiae, a non-pathogenic strain, for 2,3-BD production, which resulted in a high yield of 96.2 g/L after 244 h of fermentation (Kim et al., 2013b). E. coli BL21(DE3) has been reported to be a non-pathogenic strain that does not carry the pathogenic mechanisms causing enteric infections (Chart et al., 2000). Thus, it is a preferred candidate for the safe synthesis of chemicals (Chart et al., 2000; Zhang et al., 2012). In this work, we used E. coli BL21(DE3) as host strain and aimed to establish a method for efficient 2,3-BD production. In bacterial metabolism, three key enzymes are involved in the 2,3-BD biosynthesis from pyruvate, i.e., α-acetolactate synthase (ALS, EC 4.1.3.18), α-acetolactate decarboxylase (ALDC, EC 4.1.1.5), and 2,3-butanediol dehydrogenase (BDH, EC 1.1.1.76; also called acetoin reductase, EC 1.1.1.4) (Celińska and Grajek, 2009; Ji et al., 2011; Zhang et al., 2013). First, ALS catalyzes the condension of two molecules of pyruvate to yield α-acetolactate (Celińska and

23

Grajek, 2009; Ji et al., 2011). Then, α-acetolactate is converted to (3R)-acetoin ((3R)-AC) by ALDC (Celińska and Grajek, 2009; Ji et al., 2011). Finally, (3R)-AC is reduced to meso-2,3-BD and (2R,3R)-2,3BD by meso-2,3-BDH and (2R,3R)-2,3-BDH, respectively (Celińska and Grajek, 2009; Ji et al., 2011; Zhang et al., 2013). The 2,3-BD gene cluster of Bacillus subtilis 168 has been cloned and well characterized (Renna et al., 1993). Although the 2,3-BD gene cluster of B. subtilis 168 was studied for the construction of recombinant E. coli strains (Li et al., 2010; Shin et al., 2012), the highest yield of 2,3-BD mentioned above was obtained from utilizing the gene cluster of K. pneumoniae (Ui et al., 1997). Therefore, it is important to utilize or construct an efficient 2,3BD biosynthesis pathway that includes the related gene cluster to improve the 2,3-BD production. Promoters play an important role in controlling metabolic pathways because gene expression in prokaryotes is mainly regulated by transcription and translation processes (Keasling, 2012; Xu et al., 2013). Various promoters with different strengths and regulatory features have been developed (Peretti and Bailey, 1987). To successfully construct an entire metabolic pathway in a heterologous host, the expression of the foreign genes needs to be properly regulated; this can reduce the metabolic burden (Alper et al., 2005; Dueber et al., 2009). Therefore, optimization of the promoter strength is an important step for appropriate gene expression (Alper et al., 2005; Dueber et al., 2009), which is also essential for the construction of efficient recombinant E. coli strains for 2,3-BD production. In this work, we used a systematic approach to construct and optimize 2,3-BD production by engineering E. coli BL21(DE3) (Fig. 1). First, 2,3-BD biosynthesis gene clusters of several native 2,3-BD producers were cloned and compared for 2,3-BD synthesis. Second, the expression of the most efficient gene cluster was optimized by using different promoters. Third, extracellular metabolite profiling of the recombinant E. coli strains was conducted to optimize the key fermentation conditions. Finally, the production of 2,3-BD with high yield (73.8 g/L) and productivity (1.19 g/[L h]) was achieved through fed-batch fermentation using the efficient recombinant E. coli strain constructed in this work.

Fig. 1. Diagram summarizing the systematic approach for the metabolic engineering of Escherichia coli to produce fuel bio-chemical 2,3-butanediol (2,3-BD).

24

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

2. Materials and methods

2.3. Cloning of 2,3-BD pathway gene clusters from various bacteria

2.1. Enzymes and chemicals

Primers used in this study are listed in Table 2. Isolation of vectors, restriction enzyme digestion, agarose gel electrophoresis, and other DNA manipulations were carried out using standard protocols (Sambrook and Russell, 2001). The genes alsS and alsD of B. subtilis 168 and B. licheniformis 10-1-A were amplified through PCR with the primer pairs PBs.f(NheI)/PBs.r(overlap) and PBl.f (NheI)/PBl.r(overlap), respectively (Fig. 2C and D). The gene bdhA of B. subtilis 168 and B. licheniformis 10-1-A was amplified through PCR with the primer pairs PBs.f(overlap)/PBs.r(XhoI) and PBl.f(overlap)/PBl.r(XhoI), respectively (Fig. 2C and D). The genes alsS, alsD, and bdhA of B. subtilis 168 and B. licheniformis 10-1-A were then ligated through gene splicing by overlap extension (Horton et al., 1993), respectively, and inserted into the NheI and XhoI sites of pET28a to produce the vectors designated pETPT7-BsABC and pETPT7-BlABC, respectively (Fig. S1). The genes slaA and slaB of S. marcescens ATCC 14041 were amplified through PCR with the primer pair PSm.f(NheI)/PSm.r(overlap) (Fig. 2E). The gene slaC of S. marcescens ATCC 14041 was amplified through PCR with the primer pair PSm.f(overlap)/PSm.r(HindIII) (Fig. 2E). The genes slaA, slaB, and slaC of S. marcescens ATCC 14041 were ligated through gene splicing by overlap extension (Horton et al., 1993) and inserted into the NheI and HindIII sites of pET28a to produce the vector designated pETPT7-SmABC (Fig. S1). The 2,3-BD gene clusters of K. pneumoniae CICC 10281 and Enterobacter cloacae subsp. dissolvens SDM were amplified through PCR with the primer pairs PKp.f(NheI)/PKp.r(EcoRI) and PEc.f(NheI)/PEc.r(HindIII) (Fig. 2F and G) and inserted into the NheI/EcoRI sites or NheI/ HindIII sites of pET28a to construct the vectors designated pETPT7-

Meso-2,3-BD (98.0%), (2R,3R)-2,3-BD (98.0%), and (2S,3S)-2,3BD (99.0%) were purchased from ACROS (Geel, Belgium). AC (99.0%) was purchased from Apple Flavor & Fragrance Group Co. (Shanghai, China). The restriction enzymes were obtained from Fermentas Bio Inc. (Beijing, China). FastPfu DNA polymerase was purchased from TransGen Biotech (Beijing, China). T4 DNA ligase was obtained from New England Biolabs (Beijing, China). Isopropyl-β-D-thiogalactoside (IPTG) was obtained from Klontech (Jinan, China). Kanamycin was purchased from Amresco (Solon, OH, USA). All other chemicals were analytical-grade reagents and commercially available.

2.2. Bacterial strains and vectors Strains used in this study are listed in Table 1. E. coli DH5α (Novagen) was used for the propagation of vector DNA. E. coli BL21 (DE3) (Invitrogen) was used as host strain for 2,3-BD production. The genomes of B. subtilis 168 (Kunst et al., 1997), Bacillus licheniformis 10-1-A (CGMCC 5461), K. pneumoniae CICC 10281, S. marcescens ATCC 14041, and Enterobacter cloacae subsp. dissolvens SDM (CGMCC 4230) (Xu et al., 2012) were used as templates for the amplification of 2,3-BD pathway gene clusters by polymerase chain reaction (PCR). The vector pET28a (Novagen) was used for gene expression. Table 1 Bacterial strains and vectors used in this study. Name

Characteristic

Strain Escherichia coli DH5α

Reference

F  , φ80d lacZΔM15, Δ(lacZYA-argF)U169, recA1, endA1, hsdR17(rk  , mk þ ), phoA, supE44λ  , thi  1, gyrA96, relA1 E. coli BL21(DE3) Used as host strain Bacillus subtilis 168 Wild type Bacillus licheniformis 10-1-A Wild type Serratia marcescens ATCC 14041 Wild type Klebsiella pneumoniae CICC 10281 Wild type Enterobacter cloacae subsp. dissolvens SDM Wild type E. coli BL21/pETPT7-BsABC E. coli BL21 (DE3) harboring pETPT7-BsABC E. coli BL21/pETPT7-BlABC E. coli BL21 (DE3) harboring pETPT7-BlABC E. coli BL21/pETPT7-SmABC E. coli BL21(DE3) harboring pETPT7-SmABC E. coli BL21/pETPT7-KpABC E. coli BL21 (DE3) harboring pETPT7-KpABC E. coli BL21/pETPT7-EcABC E. coli BL21 (DE3) harboring pETPT7-EcABC E. coli BL21/pETPtac-ABC E. coli BL21 (DE3) harboring pETPtac-ABC E. coli BL21/pETPc-ABC E. coli BL21 (DE3) harboring pETPc-ABC E. coli BL21/pET-RABC E. coli BL21 (DE3) harboring pET-RABC

Invitrogen Kunst et al. (1997) Li et al. (2012a) ATCC 14041 CICC 10281 Xu et al. (2012) This work This work This work This work This work This work This work This work

Vector pET28a pMMPc pMMB66EH pETPtac pETPc pETPT7-BsABC pETPT7-BlABC pETPT7-SmABC pETPT7-KpABC pETPT7-EcABC

Novagen Xu et al. (2013) Fürste et al. (1986) This work This work This work This work This work This work This work

pETPtac-ABC pETPc-ABC pET-RABC

pMB1 replicon, Kanr Used to offer Pc promoter Used to offer Ptac promoter The promoter PT7 of pET28a replaced by Ptac The promoter PT7 of pET28a replaced by Pc pET28a carrying 2,3-BD gene cluster originated from B. subtilis 168 pET28a carrying 2,3-BD gene cluster originated from B. licheniformis 10-1-A pET28a carrying 2,3-BD gene cluster originated from S. marcescens ATCC 14041 pET28a carrying 2,3-BD gene cluster originated from K. pneumoniae CICC 10281 pET28a carrying 2,3-BD gene cluster originated from Enterobacter cloacae subsp. dissolvens SDM pETPtac carrying 2,3-BD gene cluster originated from Enterobacter cloacae subsp. dissolvens SDM pETPc carrying 2,3-BD gene cluster originated from Enterobacter cloacae subsp. dissolvens SDM pET28a carrying 2,3-BD gene cluster with its operon originated from Enterobacter cloacae subsp. dissolvens SDM

Novagen

This work This work This work

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

KpABC and pETPT7-EcABC, respectively (Fig. S1). The vectors pETPT7-BsABC, pETPT7-BlABC, pETPT7-KpABC, pETPT7-SmABC, and

25

pETPT7-EcABC were transformed into E. coli BL21(DE3) to produce the respective engineered E. coli strains (Fig. 2A).

2.4. Gene expression optimization by using different promoters

Table 2 Primers used in this study. Primera

Sequence (50 –30 )b

PBs.f(NheI) PBs.r(overlap) PBs.f(overlap) PBs.r(XhoI) PBl.f(NheI) PBl.r(overlap) PBl.f(overlap) PBl.r(XhoI) PSm.f(NheI) PSm.r(overlap) PSm.f(overlap) PSm.r(HindIII) PKp.f(NheI) PKp.r(EcoRI) PEc.f(NheI) PEc.f(EcoRI) Poperon.Ec.f(BglII) PEc.r(HindIII) Pc.f(MluI) Pc.r(XbaI) Ptac.f(BglII) Ptac.r(EcoRI)

GCCAGCTAGCGTGTTGACAAAAGCAACAAA GCCATCTTGCTGCCTTCATTATTCAGGGCTTCCTTC GAAGGAAGCCCTGAATAATGAAGGCAGCAAGATGGC GGCACTCGAGTTAGTTAGGTCTAACAAGGA GGCAGCTAGCATGAATGAAAATGGAGGAGT TCCAGATACTTTACTCATTACTCGGGATTGCCTTC GAAGGCAATCCCGAGTAATGAGTAAAGTATCTGGA GGCACTCGAGTTAATTAAATACCATTCCG GAACGGCTAGCATGAACGAAAAACACGGGT ATTGTCAAAACGCATTAAATCATCTGGCTGAAG CAGCCAGATGATTTAATGCGTTTTGACAATAAAG GGGAAGCTTTTAGACGATATTCTGTTG GCCAGCTAGCATGAATCATTCTGCTGAATG GCCAGAATTCTTAGTTAAATACCATCCCGC GTTAGCTAGCATGATGCACTCATCTGCCTG GTTAGAATTCATGATGCACTCATCTGCCTG CGGTAGATCTCTACTCCTCGCTTATCATCG GCCTAAGCTTTTAGTTGAACACCATCCCA CATGACGCGTTGCCGATACAAGAACAA GTCATCTAGAACGGGTTCGCTACCTGC GGTTAGATCTCTGTGGTATGGCTGTGCA GCGCGAATTCTGTTTCCTGTGTGAAATT

a “.f” in the primer name means that this is the sense primer; “.r” in the primer name means that this is the antisense primer. b Restriction sites are underlined.

The constitutive Pc promoter (Xu et al., 2013) and IPTGinducible Ptac promoter (de Boer, et al., 1983; Fürste et al., 1986) were amplified through PCR with the primer pairs Pc.f(MluI)/Pc.r (XbaI) and Ptac.f(BglII)/Ptac.r(EcoRI), respectively, and inserted into the MluI/XbaI and BglII/EcoRI sites of pET28a, which replaced the PT7 promoter. The resultant vectors were renamed pETPc and pETPtac, respectively. Thereafter, the 2,3-BD pathway gene cluster of Enterobacter cloacae subsp. dissolvens SDM was amplified through PCR with the primer pair PEc.f(EcoRI)/PEc.r(HindIII) and inserted into the EcoRI and HindIII sites of pETPc and pETPtac to produce the vectors designated pETPc-ABC and pETPtac-ABC, respectively (Fig. S1). In B. subtilis 168, AlsR, a member of the LysR-type transcriptional regulator family, directly stimulates alsSD transcription by binding to characteristic sites preceding the alsS promoter (Renna et al., 1993). As shown in Fig. 3, in addition to the promoter region and transcription start site predicted by the method of Reese (Reese, 2001), the upstream region of budA in Enterobacter cloacae also contains a typical cisregulatory region similar to AlsR. Thus, in this work, Pabc was cloned together with lysR from Enterobacter cloacae. The 2,3-BD gene cluster containing its entire operon was amplified through PCR with the primer pair Poperon.Ec.f(BglII)/PEc.r(HindIII) and inserted into the BglII and HindIII sites of pET28a to produce the vector designated pET-RABC (Fig. S1). The vectors pETPc-ABC,

Fig. 2. Metabolic engineering of Escherichia coli BL21(DE3) for 2,3-butanediol (2,3-BD) production. (A) Engineered pathway for 2,3-BD production in E. coli BL21(DE3). Three key enzymes are involved in the 2,3-BD biosynthesis from pyruvate. als, the gene encoding α-acetolactate synthase (ALS); aldC, the gene encoding α-acetolactate decarboxylase (ALDC); bdh, the gene encoding 2,3-butanediol dehydrogenase (BDH); ori: replication origin; Kan: the gene encoding kanamycin resistance. (1) The BDH of Bacillus subtilis 168 could catalyze interconversion of (3S)–AC/meso-2,3-BD and (3R)–AC/(2R,3R)-2,3-BD (Celińska and Grajek, 2009; Zhang et al., 2013). (2) The BDH of Bacillus licheniformis 10-1-A, Klebsiella pneumoniae CICC 10281, Serratia marcescens ATCC 14041 and Enterobacter cloacae subsp. dissolvens SDM could catalyze interconversion of (3S)–AC/(2S,3S)-2,3-BD and (3R)–AC/meso-2,3-BD (Li et al., 2013; Li et al., 2012b; Zhang et al., 2013). (B) Location of 2,3-BD pathway gene clusters in bacterial genomes. , alsR/slaR/lysR, the gene encoding the transcriptional regulator; , alsS/slaB/budB, the gene encoding ALS; , alsD/slaA/budA, the gene encoding ALDC; , bdhA/slaC/budC, the gene encoding BDH. C1, Polymerase chain reaction (PCR) amplification of alsS and alsD from B. subtilis 168; C2, PCR amplification of bdhA from B. subtilis 168; C3, gene splicing of alsS, alsD, and bdhA from B. subtilis 168 by overlap extension and gel extraction purification; D1, PCR amplification of alsS and alsD from B. licheniformis 10-1-A; D2, PCR amplification of bdhA from B. licheniformis 10-1-A; D3, gene splicing of alsS, alsD, and bdhA from B. licheniformis 10-1-A by overlap extension and gel extraction purification; E1, PCR amplification of slaA and slaB from S. marcescens ATCC 14041; E2, PCR amplification of slaC from S. marcescens ATCC 14041; E3, gene splicing of slaA, slaB, and slaC from S. marcescens ATCC 14041 by overlap extension and gel extraction purification; F, PCR amplification of 2,3-BD gene cluster from K. pneumoniae CICC 10281; G, PCR amplification of 2,3-BD gene cluster from Enterobacter cloacae subsp. dissolvens SDM.

26

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

Fig. 3. The predicted Pabc promoter of the 2,3-butanediol (2,3-BD) pathway gene cluster from Enterobacter cloacae subsp. dissolvens SDM. The DNA sequence between the gene encoding the transcriptional regulator (lysR) and the gene encodingα-acetolactate decarboxylase (budA) was 105 base pairs as shown above (Xu et al., 2012). Promoter region and transcription start site were predicted by the method of Reese (Reese, 2001).

pETPtac-ABC, and pET-RABC were transformed into E. coli BL21 (DE3) to produce the respective engineered E. coli strains.

2.5. 2,3-BD production using E. coli strains M9 minimal medium (Howard-Flanders and Theriot, 1966) containing 50 g/L glucose, 5 g/L yeast extract, and 50 μg/mL kanamycin was used to cultivate the E. coli strains. Kanamycin was used to maintain plasmid stability. Yeast extract was added to M9 minimal medium according to previous work (Oliver et al., 2013). Previous studies also showed that organic nutrient sources are beneficial for 2,3-BD production (Li et al., 2013; Yang et al., 2013b). We used 500-mL flasks filled with 100 mL of medium that was inoculated with 1% (v/v) of seed cultures to cultivate the strains. The strains were grown at 37 1C on a rotary shaker at 180 rpm. For gene expression using the strains harboring pETPT7BsABC, pETPT7-BlABC, pETPT7-KpABC, pETPT7-SmABC, pETPT7EcABC, or pETPtac-ABC, IPTG was added to a final concentration of 1 mM when the optical density of the cell suspension at 620 nm was about 0.6. Glucose consumption, products, and byproducts were analyzed after 24 h of cultivation.

2.6. Enzyme activity assays To measure enzyme activity, the cells of the strains were grown for 12 h, then centrifuged at 13,000g for 5 min, and washed twice with 100 mM potassium phosphate (pH 7.0). The cells were finally resuspended with 100 mM potassium phosphate (pH 7.0) and disrupted with an ultrasonic cell breaking apparatus (Xinzhi, Ningbo, China). Cell debris was removed through centrifugation (13,000g, 15 min). Enzyme activity was assayed in the resulting supernatant. ALS activity was measured by monitoring the conversion of pyruvate to α-acetolactate (Stormer, 1975). One unit of enzyme activity was defined as the amount of enzyme that produced 1 μmol of α-acetolactate per min. ALDC activity was assayed by detecting the production of AC from α-acetolactate (Phalip et al., 1994). α-Acetolactate was prepared immediately before use from ethyl 2-acetoxy-2-methylacetoacetate (Sigma-Aldrich), which could be transformed to α-acetolactate by adding two equivalents of NaOH (Phalip et al., 1994). One unit of ALDC activity was defined as the amount of protein that formed 1 μmol of AC per min. BDH activity was assayed by measuring the change in absorbance at 340 nm corresponding to the oxidation of NADH (ε340 ¼ 6.22 mM  1 cm  1) to NAD þ during the conversion of AC to 2,3-BD (Li et al., 2012b). One unit of enzyme activity was defined as the amount of enzyme that consumed 1 μmol of NADH/min.

2.7. Transcript analysis by quantitative reverse transcription-PCR Cells of the strains were grown for 12 h and collected for RNA isolation using the RNAprep Pure Cell/Bacteria Kit (Tiangen Biotech, Beijing, China). RNA was quantified using a NanoDrop ND-1000 UV– vis spectrophotometer (NanoDrop Technologies, Wilmington, DE). cDNA was prepared using the TransScript First-Strand cDNA Synthesis SuperMix Kit (TransGen Biotech, Beijing, China) and used for quantitative PCR analysis. The Beacon designer software was used to design the primers for each gene of the 2,3-BD gene clusters from different strains (Table S1). Quantitative PCR was performed with the MyiQ™2 Two Color Real-Time PCR detection system (Bio-Rad, Hercules, CA) using a RealMaster Mix (SYBR Green) (Tiangen Biotech, Beijing, China) according to the manufacturer's instructions. 16S rRNA was used as reference. The relative abundance of mRNAs was calculated with the ΔCT method. 2.8. Optimization of fermentation conditions M9 minimal medium supplemented with 5 g/L yeast extract and 50 μg/mL kanamycin was used to cultivate E. coli BL21/pETRABC in 500-mL flasks. The medium (100 mL) was inoculated with 1% (v/v) of the seed culture. The initial glucose concentration was 50 g/L. A quadruple biological fermenter (4 fermenters, Infors AG, Bottmingen, Switzerland) was used to optimize pH, agitation speed, and aeration rate. The fermentation medium was M9 minimal medium containing 50 g/L glucose and 5 g/L yeast extract. The medium was inoculated with E. coli BL21/pET-RABC seed culture at 10% (v/v) in a 1-L fermenter with an operating volume of 0.8 L. Kanamycin at a final concentration of 50 μg/mL was added to the medium. For pH optimization, the pH of the medium was maintained at 5.0, 6.0, 7.0, or 8.0 during the whole fermentation process by automatic addition of 5 M NaOH or 4 M H3PO4. Aeration rate and agitation speed were maintained at 1.0 vvm and 400 rpm, respectively. For optimization of the agitation speed, fermentations were carried out at 200, 300, 400, and 500 rpm; pH and aeration rate were kept constant at 7.0 and 1.0 vvm, respectively. Thereafter, 0.5, 1.0, 1.5, and 2.0 vvm were used for the optimization of the aeration rate at a pH of 7.0 and an agitation speed of 400 rpm, respectively. 2.9. Fed-batch fermentation For fed-batch fermentation, the experiment was carried out by using the same quadruple biological fermenter. The medium was inoculated with seed culture (10% [v/v]) in a 1-L fermenter with an operating volume of 0.8 L. The initial pH of the medium was 7.0; during the fermentation process, the pH value was maintained at 7.0 by automatic addition of 5 M NaOH or 4 M H3PO4. The initial

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

glucose concentration was 50 g/L; glucose (stock solution: 800 g/L) was added to the medium to a final concentration of about 50 g/L when the residual glucose concentration was reduced to about 20 g/L. 2.10. Analytical methods The culture media were centrifuged at 13,000g for 10 min. The supernatant was used for detection. Residual glucose concentration was measured using a SBA-40C bio-analyzer (Shandong Academy of Sciences, Jinan, China). 2,3-BD and AC were analyzed by gas chromatography (GC; Varian 3800, Varian) with the method described by Ma et al. (Ma et al., 2009). The GC system was equipped with a 30-m SPB-5 capillary column (0.32-mm inner diameter, 0.25-μm film thickness; Supelco, Bellefonte, PA) and a flame ionization detector. Nitrogen was used as carrier gas. Injector and detector temperatures were both set at 280 1C. The column oven temperature was maintained at 40 1C for 3 min and then raised to 240 1C at a rate of 20 1C/min. The injection volume was 1 μL. Calibration curves were used to calculate the concentration of the products. Byproducts were assayed by highperformance liquid chromatography (Agilent 1100 series, Hewlett-Packard) with the method described by Li et al. (Li et al., 2012b). The high-performance liquid chromatography system was equipped with a refractive index detector and a Bio-Rad Aminex HPX-87 H column (300  7.8 mm). The analysis was performed using a mobile phase of 10 mM H2SO4 at 55 1C with a flow rate of 0.4 mL/min. The injection volume was 5 μL. The concentrations of the byproducts were determined with calibration curves. Meso2,3-BD purity was defined as (([M])/([M] þ[S] þ[R]))  100%, where [M], [S], and [R] represent the concentrations of meso-2,3BD, (2S,3S)-2,3-BD, and (2R,3R)-2,3-BD, respectively.

3. Results and discussion 3.1. Cloning of 2,3-BD pathway gene clusters 2,3-BD has three stereoisomers, i.e., meso-2,3-BD, (2R,3R)-2,3BD, and (2S,3S)-2,3-BD (Celińska and Grajek, 2009; Ji et al., 2011). Various microorganisms produce different stereoisomers of 2,3-BD (Ji et al., 2011). Among the typical 2,3-BD producers, B. subtilis 168 produces (2R,3R)-2,3-BD as the major product (Ji et al., 2011; Zhang et al., 2013). B. licheniformis 10-1-A produces (2R,3R)-2,3-BD and meso-2,3-BD with a ratio of nearly 1:1 (Li et al., 2013). K. pneumoniae and Enterobacter cloacae produce meso-2,3-BD and (2S,3S)-2,3-BD as the major products (Ji et al., 2011; Wang et al., 2012). S. marcescens produces nearly enantiomerically pure meso-2,3-BD (Celińska and Grajek, 2009). Thus, the 2,3-BD pathway gene clusters of the typical 2,3-BD producers B. subtilis 168

27

(Kunst et al., 1997), B. licheniformis 10-1-A (Li et al., 2012a), K. pneumoniae CICC 10281, S. marcescens ATCC 14041, and Enterobacter cloacae subsp. dissolvens SDM (Xu et al., 2012) were firstly amplified to compare their 2,3-BD synthesis capabilities. As shown in Fig. 2B, in K. pneumoniae and Enterobacter cloacae, the genes encoding ALDC, ALS, and BDH are sequentially clustered in one operon called budABC. However, in other bacteria such as B. subtilis, B. licheniformis, and S. marcescens, only the genes encoding ALDC and ALS are in the same gene cluster. Thus, the gene encoding BDH in B. subtilis, B. licheniformis, and S. marcescens was ligated with the genes encoding ALS and ALDC through overlap extension to form three non-native 2,3-BD gene clusters (Horton et al., 1993). As shown in Fig. 2C to G, the 2,3-BD pathway gene clusters of these typical 2,3-BD producers were successfully cloned from the respective strains. 3.2. Expression of different 2,3-BD pathway gene clusters in E. coli BL21(DE3) The cloned 2,3-BD pathway gene clusters from B. subtilis 168, B. licheniformis 10-1-A, K. pneumoniae CICC 10281, S. marcescens ATCC 14041, and Enterobacter cloacae subsp. dissolvens SDM were inserted into the expression vector pET28a (Novagen) to construct the expression vectors pETPT7-BsABC, pETPT7-BlABC, pETPT7KpABC, pETPT7-SmABC, and pETPT7-EcABC, respectively. As useful metabolic engineering components, gene clusters have been widely used for the construction of many recombinant strains to produce a large number of chemicals (Keasling, 2012; Keasling, 2010). In this work, each of the 2,3-BD pathway gene clusters was used as a whole component, and their efficiencies for the 2,3-BD production were compared. All of the cloned 2,3-BD pathway gene clusters were overexpressed under the control of the same IPTGinducible promoter, PT7. The constructed expression vectors were transformed into E. coli BL21(DE3) and the 2,3-BD synthesis abilities of the recombinant strains were assayed. As shown in Table 3, E. coli BL21(DE3) produced only 0.25 g/L AC within 24 h. Although B. subtilis 168, B. licheniformis 10-1-A, K. pneumoniae CICC 10281, and S. marcescens ATCC 14041 are good 2,3-BD producers, direct translation of their 2,3-BD pathway gene clusters in E. coli did not result in the production of high amounts of 2,3-BD as compared to the native strains. E. coli BL21/pETPT7-EcABC expressing the gene cluster from Enterobacter cloacae subsp. dissolvens SDM had the best ability to produce 2,3-BD. This strain produced 12.8 g/L 2,3-BD and 1.88 g/L AC within 24 h. The carbon balances of these strains were also analyzed, and E. coli BL21/pETPT7-EcABC showed the highest 2, 3-BD carbon ratio per glucose among the strains harboring different 2,3-BD gene clusters (Table S2). These results indicate that screening of multiple gene clusters is necessary to achieve optimal 2,3-BD production in recombinant E. coli.

Table 3 Cell growth, glucose consumption, product and byproduct production analyses of Escherichia coli strains harboring vectors carrying 2,3-butanediol (2,3-BD) pathway gene clusters from different bacteria in 24-h flask culturesa. Strain

Cell density (OD620)

Glucose consumed (g/L)

2,3-BD (g/L)

Acetoin (g/L)

Succinate (g/L)

Lactate (g/L)

Acetate (g/L)

Ethanol (g/L)

Diol Diol productivity yield (g/g) (g/[L h])

Escherichia coli BL21(DE3) E. coli BL21/pETPT7-BsABC E. coli BL21/pETPT7-BlABC E. coli BL21/pETPT7-SmABC E. coli BL21/pETPT7-KpABC E. coli BL21/pETPT7-EcABC

4.27 70.05 4.48 70.17 4.61 70.07 5.04 70.09 5.84 70.04 5.98 70.09

7.337 1.53 9.677 1.53 24.007 1.00 24.337 1.15 20.007 1.73 32.337 0.58

ND 2.86 7 0.01 5.86 7 0.06 0.26 7 0.02 6.42 7 0.05 12.777 0.13

0.25 7 0.03 0.767 0.01 4.86 7 0.13 10.08 7 0.14 1.56 7 0.02 1.88 7 0.07

0.197 0.07 0.22 7 0.03 0.60 7 0.18 0.677 0.03 0.677 0.19 1.54 7 0.01

0.38 7 0.30 0.30 7 0.20 0.137 0.02 2.75 7 0.02 0.02 7 0.00 0.077 0.01

3.75 7 0.24 2.03 7 0.12 2.08 7 0.07 0.29 7 0.00 2.69 7 0.06 1.63 7 0.04

0.54 70.17 0.53 70.17 1.25 70.12 1.32 70.04 0.99 70.08 1.69 70.23

0.03 0.15 0.45 0.43 0.33 0.61

ND: not detected; OD620: optical density at 620 nm. a

Data are the means 7 standard deviations (SDs) from three parallel experiments.

0.01 0.37 0.45 0.42 0.40 0.45

28

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

In this study, biomass and extracellular metabolic byproduct concentrations obtained by using different recombinant E. coli strains were also determined. As shown in Table 3, the final cell density of the wild-type BL21(DE3) strain at 620 nm appeared to be 4.27 70.05, which was lower than the cell density measured for the other recombinant strains. As for the five recombinant strains, biomass increased in the order of E. coli BL21/pETPT7-BsABC, E. coli BL21/pETPT7-BlABC, E. coli BL21/pETPT7-SmABC, E. coli BL21/ pETPT7-KpABC, and E. coli BL21/pETPT7-EcABC. Acetate, lactate, succinate, and ethanol were the main byproducts in all cultures. As shown in Table 3, the production of acetate decreased among the recombinant strains compared with that of the host strain. In E. coli, acetate formation is due to excessive carbon influx from glucose and accumulation of acetate would inhibit the growth of biomass (Contiero et al., 2000). When 2,3-BD pathway gene clusters were expressed in E. coli, the recombinant strains would have the ability to channel excess pyruvate to the much less toxic compound, 2,3-BD, which would lead to suppressed acetate production and enhanced biomass growth (Celińska and Grajek, 2009; Zeng and Sabra, 2011). 3.3. Activities of 2,3-BD biosynthesis enzymes in recombinant E. coli strains The activities of ALDC, ALS, and BDH in the recombinant strains were also assayed (Table 4). Conversion of pyruvate to αacetolactate can be catalyzed by ALS or acetohydroxy acid synthase (Ji et al., 2011; Steinmetz et al., 2010). Because acetohydroxy acid synthase is indispensable for the production of branched amino acids in bacteria, the low ALS activity detected in E. coli BL21(DE3) might be due to the presence of acetohydroxy acid synthase in the strain (Steinmetz et al., 2010). On the other hand, consistent with the result of AC production, E. coli BL21(DE3) exhibited low ALDC activity. E. coli BL21(DE3) showed BDH activity; however, no 2,3-BD was detected. BDH activity was measured through the detection of NADH oxidation (Li et al., 2012b). NADH itself is not very stable and some enzymes might be present in the host cell that could non-specifically and slowly oxidize NADH; therefore, E. coli BL21(DE3) seems to possess pseudo-BDH activity. As shown in Table 4, all recombinant E. coli strains exhibited relatively high ALS and ALDC activities. However, they had different BDH activities. The results indicated that the production of 2,3-BD increased from the wild-type strain to E. coli BL21/pETPT7-EcABC as the BDH activity increased (Table 3). For example, in E. coli BL21/ pETPT7-BlABC and E. coli BL21/pETPT7-SmABC, the yield of diol (AC plus BD) was similar to that of E. coli BL21/pETPT7-EcABC. However, due to the low activity of BDH, E. coli BL21/pETPT7-BlABC and E. coli BL21/pETPT7-SmABC produced low amounts of 2,3-BD and accumulated high amounts of AC. Thus, ALS and ALDC might be overexpressed under the control of the promoter PT7 in all recombinant strains and the bottleneck for 2,3-BD production in E. coli would be

the BDH activity. We also investigated the mRNA levels of these strains and found that all genes of the 2,3-BD gene clusters in these strains were transcribed, but their mRNA levels showed large differences (Table S3 and Fig. S2). The plasmid copy numbers of these vectors in the strains were analyzed (Fig. S3), and the strain harboring the vector with low copy number showed low mRNA levels. Previous reports also indicated that the distinct DNA sequences of the downstream genes of promoters could affect gene transcription, which might contribute to the differences (Lee et al., 1982; Mutalik et al., 2013). 3.4. Optimization of promoters for expression of 2,3-BD pathway gene clusters For high 2,3-BD production through appropriate expression of 2,3-BD biosynthetic gene clusters, the strength and regulatory features of promoters should be optimized. Because E. coli BL21/ pETPT7-EcABC showed the best performance with regard to 2,3-BD synthesis, the 2,3-BD pathway gene cluster from Enterobacter cloacae was expressed under the control of different types of promoters and 2,3-BD synthesis abilities of the recombinant strains were assayed. The commonly used IPTG-inducible promoter Ptac (de Boer, et al., 1983), the constitutive promoter Pc (Xu et al., 2013; Xu et al., 2007), and the native promoter of the 2,3-BD biosynthetic gene cluster of Enterobacter cloacae (Pabc) (Fig. 3) were used to regulate the expression of the 2,3-BD pathway gene cluster. The 2,3-BD pathway gene cluster of Enterobacter cloacae subsp. dissolvens SDM was inserted into pET28a under the control of the promoters mentioned above. The constructed expression vectors pETPtacABC, pETPc-ABC, and pET-RABC were transformed into E. coli BL21(DE3), respectively, and the 2,3-BD synthesis abilities of the recombinant strains were assayed. As shown in Table 5, the constitutive promoters Pc and Pabc showed a better performance than the inducible promoters Ptac and PT7, because E. coli BL21/ pETPc-ABC and E. coli BL21/pET-RABC produced higher amounts of 2,3-BD compared to E. coli BL21/pETPtac-ABC and E. coli BL21/ pETPT7-EcABC. E. coli BL21/pET-RABC had the best ability to produce 2,3-BD; this strain consumed 49.7 g/L glucose and produced 21.5 g/L 2,3-BD and 1.59 g/L AC within 24 h. Biomass and extracellular metabolic byproduct analyses also indicated that Pabc performed better than the inducible promoters Ptac and PT7. In particular, E. coli BL21/pET-RABC showed higher biomass growth and lower acetate production compared to E. coli BL21/pETPtacABC and E. coli BL21/pETPT7-EcABC. Carbon balance data also showed that E. coli BL21/pET-RABC had the highest 2,3-BD carbon ratio per glucose (Table S4). The activities of ALDC, ALS, and BDH in E. coli BL21/pETPtacABC, E. coli BL21/pETPc-ABC, and E. coli BL21/pET-RABC were also measured. As shown in Table 6, with the increasing activities of ALDC, ALS, and BDH in E. coli BL21/pETPtac-ABC, E. coli BL21/pETPc-ABC, and E. coli BL21/pET-RABC, higher 2,3-BD yields

Table 4 Enzyme activities of Escherichia coli strains harboring vectors carrying 2,3-butanediol (2,3-BD) pathway gene clusters from different bacteria in 12-h flask culturesa. Strain

ALS activity (U/mg)

Escherichia coli BL21(DE3) E. coli BL21/pETPT7-BsABC E. coli BL21/pETPT7-BlABC E. coli BL21/pETPT7-SmABC E. coli BL21/pETPT7-KpABC E. coli BL21/pETPT7-EcABC

1074.067 111.18 3425.87 7 257.82 304.817 36.19 154.017 15.97 138.46 7 7.26

14.447 14.83

ALDC activity (U/mg)

BDH activity (U/mg)

7.23 7 12.53 587.767 46.70 712.487 51.02 1223.377 99.04 2505.707 223.63 2381.20 7 146.06

0.54 7 0.15 0.75 7 0.15 0.87 7 0.09 0.62 7 0.09 17.30 7 0.11 24.287 1.81

ALS: α-acetolactate synthase; ALDC: α-acetolactate decarboxylase; BDH: 2,3-butanediol dehydrogenase. a

Data are the means 7 standard deviations (SDs) from three parallel experiments.

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

would be obtained. Although E. coli BL21/pETPT7-EcABC had higher ALS and ALDC activities than E. coli BL21/pETPc-ABC and E. coli BL21/pET-RABC, less 2,3-BD was produced. As shown in Table 5, in the system using E. coli BL21/pETPT7-EcABC, a high acetate concentration was observed. Over-expression of the 2,3-BD gene clusters under the control of a strong promoter might increase the metabolic burden to the host and result in lower production of the target chemical (Dueber et al., 2009). Thus, directly enhancing the expression of 2,3-BD pathway gene clusters might not result in high production of the target chemical. Therefore, it was very important to screen multiple promoters for the construction of an efficient 2,3-BD-producing recombinant E. coli strain.

29

metabolism from acid production to neutral compound production (Ji et al., 2011; Tsau et al., 1992). Thus, the pH is an important factor that influences 2,3-BD production (Celińska and Grajek, 2009). As shown in Table 7, biomass and extracellular metabolite concentrations were determined under different pH conditions. Under an alkaline condition (pH 8.0), formation of acetate was favored, accompanied by a simultaneous decrease in biomass and diol yield. Under an acidic condition (pH 6.0), lactate synthesis increased and biomass and diol synthesis decreased. The highest cell density of E. coli BL21/pET-RABC was 14.2 measured at 620 nm and pH 7.0. This pH was also found to be optimal for 2,3-BD production (Table 7). The amount of glucose consumed was 57.0 g/L, and 23.7 g/ L of 2,3-BD and 1.30 g/L of AC were produced within 18 h.

3.5. Optimization of fermentation conditions E. coli BL21/pET-RABC showed the best performance with regard to 2,3-BD production and used the constitutive promoter Pabc, which made it convenient to operate and economically advantageous because no induction was required. Thereafter, fermentation conditions, including pH and oxygen supply, were optimized for the efficient production of 2,3-BD using E. coli BL21/pET-RABC. 3.5.1. Optimization of pH In native 2,3-BD producers, pyruvate is channeled into a mixture of 2,3-BD, AC, acetate, lactate, succinate, and ethanol, through mixed acid-2,3-BD fermentation (Celińska and Grajek, 2009; Ji et al., 2011). Formation of diol (2,3-BD and AC) plays a role in preventing intracellular acidification by changing the

3.5.2. Optimization of oxygen supply In bacteria, BDH catalyzes the reversible transformation between AC and 2,3-BD, which is coupled with the NADH/NAD þ conversion (Celińska and Grajek, 2009). The synthesis of AC and 2,3-BD is regulated by the NADH/NAD þ ratio (Blomqvist et al., 1993). The oxygen supply, which can influence the intracellular NAD þ level and NADH/NAD þ ratio, is therefore another important factor that influences 2,3-BD fermentation (Blomqvist et al., 1993; Celińska and Grajek, 2009). Because the oxygen supply is mutually associated with agitation speed and aeration rate, the effects of these two factors on 2,3-BD production by E. coli BL21/pET-RABC were studied in this work. As shown in Table 8, the agitation speed had a strong effect on biomass and extracellular metabolite concentrations. The production

Table 5 Cell growth, glucose consumption, product and byproduct production analyses of Escherichia coli strains harboring vectors carrying 2,3-butanediol (2,3-BD) pathway gene cluster with different promoters in 24-h flask culturesa. Strain

E. E. E. E.

coli coli coli coli a

BL21/pETPT7-EcABC BL21/pETPtac-ABC BL21/pETPc-ABC BL21/pET-RABC

Glucose consumed (g/L)

2,3-BD (g/L)

Acetoin (g/L) Succinate (g/L)

Lactate (g/L)

5.98 7 0.09

32.337 0.58 16.007 1.00 45.677 1.15 49.677 0.58

12.7770.13 5.53 70.20 18.28 70.55 21.48 70.21

1.88 70.07 1.34 70.09 1.20 70.13 1.59 70.05

0.077 0.01 1.63 7 0.04 1.69 70.23 0.61 0.03 7 0.01 2.60 7 0.40 0.63 70.27 0.29 0.94 7 0.10 0.08 7 0.01 1.13 70.06 0.81 0.89 7 0.08 0.39 7 0.03 0.84 70.11 0.96

5.34 7 0.40 7.32 7 0.10 7.93 7 0.05

1.54 7 0.01 0.687 0.11 1.65 7 0.02 1.45 7 0.02

Acetate (g/L)

Ethanol (g/L)

Diol yield Diol productivity (g/g) (g/[L h])

Cell density (OD620)

0.45 0.43 0.43 0.46

Data are the means7 standard deviations (SDs) from three parallel experiments. OD620: optical density at 620 nm.

Table 6 Enzyme activities of Escherichia coli strains harboring vectors carrying 2,3-butanediol (2,3-BD) pathway gene cluster with different promoters in 12-h flask culturesa. Strain E. E. E. E.

coli coli coli coli

BL21/pETPT7-EcABC BL21/pETPtac-ABC BL21/pETPc-ABC BL21/pET-RABC

ALS activity (U/mg)

ALDC activity (U/mg)

BDH activity (U/mg)

138.46 7 7.26 53.767 13.59 114.357 6.30 154.977 1.24

2381.20 7 146.06 288.87 7 27.00 478.767 10.16 484.087 28.34

24.28 7 1.81 2.577 0.13 35.197 2.27 42.007 4.29

ALS: α-acetolactate synthase; ALDC: α-acetolactate decarboxylase; BDH: 2,3-butanediol dehydrogenase. a

Data are the means 7standard deviations (SDs) from three parallel experiments.

Table 7 Cell growth, glucose consumption, product and byproduct production analyses of Escherichia coli BL21/pET-RABC during pH optimization in batch fermentation. pH

Cell density (OD620)

Glucose consumed (g/L)

2,3Butanediol (g/L)

Acetoin (g/L)

Succinate (g/L)

Lactate (g/L)

Acetate (g/L)

Ethanol (g/L)

Diol productivity (g/[L h])

Diol yield (g/g)

5.0 6.0 7.0 8.0

8.55 12.15 14.15 10.25

32.00 54.00 57.00 45.00

13.47 22.12 23.74 19.39

1.24 1.28 1.30 1.19

0.36 1.45 1.66 0.79

1.21 4.68 2.45 1.19

ND ND 0.36 1.29

0.29 0.50 0.61 1.01

0.82 1.30 1.39 1.14

0.46 0.43 0.44 0.46

ND: not detected; OD620: optical density at 620 nm.

30

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

Table 8 Cell growth, glucose consumption, product and byproduct production analyses of Escherichia coli BL21/pET-RABC during oxygen supply optimization in batch fermentation. Glucose consumed (g/L)

2,3Butanediol (g/L)

Acetoin (g/L)

Succinate (g/L)

Lactate (g/L)

Acetate (g/L)

Ethanol (g/L)

Diol productivity (g/[L h])

Diol yield (g/g)

Agitation speed (rpm) 200 8.12 300 12.66 400 15.08 500 13.98

23.00 38.00 54.00 43.00

8.31 14.74 23.72 13.04

0.33 0.41 1.17 3.87

0.43 0.82 2.15 2.16

3.26 3.89 1.78 0.03

0.73 0.58 0.40 0.25

0.84 0.82 1.25 1.63

0.51 0.89 1.46 0.99

0.38 0.40 0.46 0.39

Aeration rate (vvm) 0.5 13.82 1.0 14.80 1.5 15.30 2.0 15.18

51.00 56.00 57.00 57.00

21.77 23.54 24.56 23.84

0.20 0.37 0.56 1.21

1.56 2.12 2.43 2.59

2.79 2.02 1.01 1.74

0.44 0.36 0.31 0.40

1.66 1.44 1.76 1.35

1.29 1.41 1.48 1.47

0.43 0.43 0.44 0.44

Fermentation condition

Cell density (OD620)

OD620: optical density at 620 nm.

Fig. 4. 2,3-Butanediol (2,3-BD) production from glucose using Escherichia coli BL21/ pET-RABC in fed-batch fermentation. ■, cell density, optical density at 620 nm (OD620); ●, glucose concentration; ▲, AC concentration; ▼, 2,3-BD concentration. Cultivation was carried out at an initial pH of 7.0; during the fermentation processes, the pH was maintained at 7.0 by automatic addition of 4 M H3PO4 or 5 M NaOH using a program-controlled peristaltic pump. Agitation speed was 400 rpm and aeration rate was 1.5 vvm.

of compounds that require NAD þ for their synthesis, e.g., AC, increased as the agitation speed increased. The production of compounds that require NADH for their synthesis, e.g., lactate, decreased as the agitation speed increased. Biomass and 2,3-BD production increased as the agitation speed increased up to 400 rpm and decreased thereafter. At an agitation speed of 400 rpm, 23.7 g/L of 2,3-BD and 1.17 g/L of AC were produced from 54.0 g/L of glucose. The aeration rate also clearly influenced biomass and extracellular metabolite concentrations during 2,3-BD fermentation. It was indicated that the ratio between AC and 2,3-BD increased from 0.01:1 to 0.05:1 as the aeration rate increased from 0.5 vvm to 2.0 vvm (Table 8). Biomass and 2,3-BD production increased as the aeration rate increased up to 1.5 vvm and decreased thereafter. Although lowering the aeration rate would decrease AC production, the amount of biomass would also decrease, which would consequently result in lower 2,3-BD production. With an optimal aeration rate of 1.5 vvm, 24.6 g/L of 2,3-BD and 0.56 g/L of AC were produced from 57.0 g/L of glucose within 17 h. 3.6. Fed-batch bioconversions Fed-batch fermentation was conducted with the optimized fermentation conditions to achieve high 2,3-BD production. As

shown in Fig. 4, 73.8 g/L of 2,3-BD and 9.34 g/L of AC were produced within 62 h by E. coli BL21/pET-RABC. The 2,3-BD productivity was 1.19 g/[L h] with a yield of 0.41 g 2,3-BD/g glucose. Genetically engineered E. coli strains possess a high potential for biofuel production because of the well-studied metabolic pathways and established genetic manipulation methods (Woolston et al., 2013). However, as shown in Table 9, the productivities and final concentrations of many biofuels produced by recombinant E. coli strains were relatively low. In this work, the recombinant strain E. coli BL21/pET-RABC was developed for the production of 2,3-BD, a biofuel that exhibits low toxicity to microbial systems. To construct an efficient 2,3-BD production system, we hypothesized that screening of multiple gene clusters from different native producers is necessary to achieve optimal 2,3-BD production in recombinant E. coli strains. We also proposed that screening multiple promoters could improve the pathway function for 2,3-BD production by recombinant E. coli strains. The effects of fermentation conditions and operation modes on 2,3-BD formation is vital in the establishment of an optimal process design. For the efficient production of 2,3-BD, biomass and extracellular metabolite profiling analyses of the recombinant E. coli strains were used for the optimization of the key fermentation conditions. After completing the systematic optimization process mentioned above, 73.8 g/L of 2,3-BD were produced from a recombinant E. coli strain constructed in this work (Fig. 4). This result demonstrates that developing a strong production system through a systematic approach could increase the production of chemicals in recombinant E. coli strains. To further improve the system, understanding the engineered strains at the system level and simultaneous engagement of multiple genes would be desirable. The stereoisomeric composition of 2,3-BD formed by E. coli BL21/pET-RABC was analyzed by GC (Agilent GC6820) with a flame ionization detector and a fused silica capillary column (Supelco Beta DEXTM 120, 0.25 mm  30 m) (Wang et al., 2012). As shown in Fig. S4, a mixture of 2,3-BD was obtained, which contained 95.5% of meso-2,3-BD, 3.9% of (2S,3S)-2,3-BD, and 0.5% of (2R,3R)2,3-BD. The BDH of Enterobacter cloacae subsp. dissolvens SDM has been reported to catalyze the interconversion of (3S)-AC/(2S,3S)2,3-BD and (3R)–AC/meso-2,3-BD (Li et al., 2012b). Therefore, the meso-2,3-BD and (2S,3S)-2,3-BD were converted from AC by BDH of Enterobacter cloacae subsp. dissolvens SDM. It is reported that glycerol dehydrogenase has the ability to catalyze the interconversion of (3R)–AC/(2R,3R)-2,3-BD (Li et al., 2013), and (2R,3R)-2,3-BD may be transformed from AC by the glycerol dehydrogenase of E. coli BL21(DE3). In addition to being used as a liquid fuel or fuel additive (Xiao et al., 2012), meso-2,3-BD has also been used as a

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

31

Table 9 Biofuel production using recombinant Escherichia coli strains. Biofuel

Host strain

Substrate

Method

Concentration (g/L)

Productivity (g/[L h])

Reference

Acetone α-Pinene 1-Butanol Ethanol Isobutanol Isoprenol Prenol Fatty alcohol 2,3-Butanediol

E. E. E. E. E. E. E. E. E.

Glucose Glucose Glycerol Xylose Glucose Glucose Glucose Glucose Glucose

Fed-batch Fed-batch Shake flask Batch Fed-batch Shake flask Shake flask Fed-batch Fed-batch

7.1 1.0 0.6 63.8 50.8 1.3 0.2 0.6 73.8

0.15 0.03 0.02 0.89 0.71 0.04 0.01 0.02 1.19

May et al. (2013) Yang et al. (2013a) Atsumi et al. (2008) Yomano et al. (1998) Baez et al. (2011) Zheng et al. (2013)

coli coli coli coli coli coli coli coli coli

HB101 BL21(DE3) BW25113 KO11 BW25113 BL21(DE3) BL21(DE3) BL21(DE3) BL21(DE3)

Zheng et al. (2012) This work

Table 10 Meso-2,3-Butanediol (2,3-BD) production using recombinant Escherichia coli strains. Substrate

Host strain

2,3-BD stereoisomer

Method

Concentration (g/L)

Productivity (g/[L h])

Reference

Seaweed hydrolysate

E. coli MG1655

Fed-batch

14.1

0.20

Cellodextrin Crude glycerol Glucose Glucose Glucose

E. coli E. coli E. coli E. coli E. coli (DE3) E. coli E. coli

MG1655 W3110 W3110 JM109 YYC202

Mixed of meso-2,3-BD and (2S,3S)-2,3-BD Meso-2,3-BD Meso-2,3-BD Meso-2,3-BD Meso-2,3-BD Meso-2,3-BD

SSF* Batch Batch Batch Shake Flask

4.7 6.9 15.7 14.5 1.1

0.03 0.14 0.33 0.30 0.01

Mazumdar et al. (2013) Shin et al. (2012) Lee et al. (2012)

JM109 BL21(DE3)

Meso-2,3-BD Meso-2,3-BD

Shake Flask Shake Flask Batch Fed-batch

17.7 21.5 24.6 73.8

0.31 0.90 1.44 1.19

Glucose Glucose

n

Li et al. (2010) Nielsen et al. (2010) Ui et al. (1997) This work This work This work

SSF: simultaneous saccharification and fermentation.

monomer to produce renewable polyesters, which is a good alternative to the monomers produced from non-renewable fossil fuel (Gubbels et al., 2012). Several recombinant E. coli strains have been used to produce meso-2,3-BD (Table 10). Here, we found that E. coli BL21/pET-RABC was able to produce meso-2,3-BD of relatively high enantiomeric purity. The values of concentration and yield of meso-2,3-BD produced using the novel process were the highest ever obtained in meso-2,3-BD production. Therefore, the process presented in this work would not only be a promising alternative for the fuel bio-chemical production, but could also be a good example for the efficient production of other important chemicals.

Acknowledgments The authors would like to thank the partial financial support of the National Basic Research Program of China (2011CBA00800). This work was supported by the Chinese National Program for High Technology Research and Development (2011AA02A207), Program for High Technology Research and Development of Shandong province (2012GSF12119). We acknowledge Yu Wang for his experimental assistance on the paper revision. We also thank anonymous reviewers for very helpful comments on the manuscript.

Appendix A. Supporting information 4. Conclusions To summarize, a systematic approach was used to optimize the 2,3-BD production through recombinant E. coli strains. High yield (73.8 g/L of 2,3-BD) and productivity (1.19 g/[L h]) were achieved using the optimized process. The systematic approach used in this work might also be applicable to improve the production of other important chemicals through recombinant E. coli strains.

Authors' contributions YX, CM, CG, and PX designed and conceived the study. YX, HC, ZZ, and KL executed the experimental work. YX, HC, LL, and CG analyzed the data. CM, CG, and PX contributed reagents and materials. YX, CG, FT, CM, and PX wrote and revised the manuscript.

Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ymben.2014.02.004.

References Alper, H., Fischer, C., Nevoigt, E., Stephanopoulos, G., 2005. Tuning genetic control through promoter engineering. Proc. Natl. Acad. Sci. U.S.A. 102, pp. 12678– 12683. Atsumi, S., Cann, A.F., Connor, M.R., Shen, C.R., Smith, K.M., Brynildsen, M.P., Chou, K.J., Hanai, T., Liao, J.C., 2008. Metabolic engineering of Escherichia coli for 1-butanol production. Metab. Eng. 10, 305–311. Baez, A., Cho, K.M., Liao, J.C., 2011. High-flux isobutanol production using engineered Escherichia coli: a bioreactor study with in situ product removal. Appl. Microbiol. Biotechnol. 90, 1681–1690. Blattner, F.R., Plunkett , G., Bloch, C.A., Perna, N.T., Burland, V., Riley, M., Collado-Vides, J., Glasner, J.D., Rode, C.K., Mayhew, G.F., Gregor, J., Davis, N.W., Kirkpatrick, H.A., Goeden, M.A., Rose, D.J., Mau, B., Shao, Y., 1997. The complete genome sequence of Escherichia coli K-12. Science 277, 1453–1462. Blomqvist, K., Nikkola, M., Lehtovaara, P., Suihko, M.L., Airaksinen, U., Stråby, K.B., Knowles, J.K., Penttilä, M.E., 1993. Characterization of the genes of the

32

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

2,3-butanediol operons from Klebsiella terrigena and Enterobacter aerogenes. J. Bacteriol. 175, 1392–1404. Bokinsky, G., Peralta-Yahya, P.P., George, A., Holmes, B.M., Steen, E.J., Dietrich, J., Lee, T.S., Tullman-Ercek, D., Voigt, C.A., Simmons, B.A., Keasling, J.D., 2011. Synthesis of three advanced biofuels from ionic liquid-pretreated switchgrass using engineered Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 108, pp. 19949–19954. Celińska, E., Grajek, W., 2009. Biotechnological production of 2,3-butanediol – current state and prospects. Biotechnol. Adv. 27, 715–725. Chart, H., Smith, H.R., La Ragione, R.M., Woodward, M.J., 2000. An investigation into the pathogenic properties of Escherichia coli strains BLR, BL21, DH5alpha and EQ1. J. Appl. Microbiol. 89, 1048–1058. Chen, C.Y., Yeh, K.L., Aisyah, R., Lee, D.J., Chang, J.S., 2011. Cultivation, photobioreactor design and harvesting of microalgae for biodiesel production: a critical review. Bioresour. Technol 102, 71–81. Contiero, J., Beatty, C., Kumari, S., DeSanti, C.L., Strohl, W.R., Wolfe, A., 2000. Effects of mutations in acetate metabolism on high-cell-density growth of Escherichia coli. J. Ind. Microbiol. Biotechnol. 24, pp. 421–430. de Boer, H.A., Comstock, L.J., Vasser, M., 1983. The tac promoter: a functional hybrid derived from the trp and lac promoters. Proc. Natl. Acad. Sci. U.S.A. 80, pp. 21–25. Domínguez, de María, P., 2011. Recent developments in the biotechnological production of hydrocarbons: paving the way for bio-based platform chemicals. ChemSusChem 4, 327–329. Dueber, J.E., Wu, G.C., Malmirchegini, G.R., Moon, T.S., Petzold, C.J., Ullal, A.V., Prather, K.L., Keasling, J.D., 2009. Synthetic protein scaffolds provide modular control over metabolic flux. Nat. Biotechnol. 27, 753–759. Fürste, J.P., Pansegrau, W., Frank, R., Blöcker, H., Scholz, P., Bagdasarian, M., Lanka, E., 1986. Molecular cloning of the plasmid RP4 primase region in a multi-hostrange tacP expression vector. Gene 48, 119–131. Gross, M., 2012. Looking for alternative energy sources. Curr. Biol. 22, R103–106. Gubbels, E., Jasinska-Walc, L., Koning, C.E., 2012. Synthesis and characterization of novel renewable polyesters based on 2,5-furandicarboxylic acid and 2,3-butanediol. J. Polym. Sci 51, 890–898. Ha, S.J., Galazka, J.M., Kim, S.R., Choi, J.H., Yang, X., Seo, J.H., Glass, N.L., Cate, J.H., Jin, Y.S., 2011. Engineered Saccharomyces cerevisiae capable of simultaneous cellobiose and xylose fermentation. Proc. Natl. Acad. Sci. U.S.A 108, 504–509. Horton, R.M., Ho, S.N., Pullen, J.K., Hunt, H.D., Cai, Z., Pease, L.R., 1993. Gene splicing by overlap extension. Methods Enzymol. 217, 270–279. Howard-Flanders, P., Theriot, L., 1966. Mutants of Escherichia coli K-12 defective in DNA repair and in genetic recombination. Genetics 53, 1137–1150. Ji, X.J., Huang, H., Ouyang, P.K., 2011. Microbial 2,3-butanediol production: a stateof-the-art review. Biotechnol. Adv. 29, 351–364. Keasling, J.D., 2012. Synthetic biology and the development of tools for metabolic engineering. Metab. Eng. 14, 189–195. Keasling, J.D., 2010. Manufacturing molecules through metabolic engineering. Science 330, 1355–1358. Kim, D.K., Rathnasingh, C., Song, H., Lee, H.J., Seung, D., Chang, Y.K., 2013a. Metabolic engineering of a novel Klebsiella oxytoca strain for enhanced 2,3butanediol production. J. Biosci. Bioeng. 116, 186–192. Kim, S.J., Seo, S.O., Jin, Y.S., Seo, J.H., 2013b. Production of 2,3-butanediol by engineered Saccharomyces cerevisiae. Bioresour. Technol. 146, 274–281. Kunst, F., Ogasawara, N., Moszer, I., Albertini, A.M., Alloni, G., Azevedo, V., Bertero, M.G., Bessières, P., Bolotin, A., Borchert, S., Borriss, R., Boursier, L., Brans, A., Braun, M., Brignell, S.C., Bron, S., Brouillet, S., Bruschi, C.V., Caldwell, B., Capuano, V., Carter, N.M., Choi, S.K., Codani, J.J., Connerton, I.F., Danchin, A., Denizot, F., Devine, K.M., Düsterhöft, A., Ehrlich, S.D., Emmerson, P.T., Entian, K. D., Errington, J., Fabret, C., Ferrari, E., Foulger, D., Fritz, C., Fujita, M., Fujita, Y., Fuma, S., Galizzi, A., Galleron, N., Ghim, S.Y., Glaser, P., Goffeau, A., Golightly, E.J., Grandi, G., Guiseppi, G., Guy, B.J., Haga, K., Haiech, J., Harwood, C.R., Hénaut, A., Hilbert, H., Holsappel, S., Hosono, S., Hullo, M.F., Itaya, M., Jones, L., Joris, B., Karamata, D., Kasahara, Y., Klaerr-Blanchard, M., Klein, C., Kobayashi, Y., Koetter, P., Koningstein, G., Krogh, S., Kumano, M., Kurita, K., Lapidus, A., Lardinois, S., Lauber, J., Lazarevic, V., Lee, S.M., Levine, A., Liu, H., Masuda, S., Mauël, C., Médigue, C., Medina, N., Mellado, R.P., Mizuno, M., Moestl, D., Nakai, S., Noback, M., Noone, D., O'Reilly, M., Ogawa, K., Ogiwara, A., Oudega, B., Park, S.H., Parro, V., Pohl, T.M., Portetelle, D., Porwollik, S., Prescott, A.M., Presecan, E., Pujic, P., Purnelle, B., Rapoport, G., Rey, M., Reynolds, S., Rieger, M., Rivolta, C., Rocha, E., Roche, B., Rose, M., Sadaie, Y., Sato, T., Scanlan, E., Schleich, S., Schroeter, R., Scoffone, F., Sekiguchi, J., Sekowska, A., Seror, S.J., Serror, P., Shin, B.S., Soldo, B., Sorokin, A., Tacconi, E., Takagi, T., Takahashi, H., Takemaru, K., Takeuchi, M., Tamakoshi, A., Tanaka, T., Terpstra, P., Tognoni, A., Tosato, V., Uchiyama, S., Vandenbol, M., Vannier, F., Vassarotti, A., Viari, A., Wambutt, R., Wedler, E., Wedler, H., Weitzenegger, T., Winters, P., Wipat, A., Yamamoto, H., Yamane, K., Yasumoto, K., Yata, K., Yoshida, K., Yoshikawa, H.F., Zumstein, E., Yoshikawa, H., Danchin, A., 1997. The complete genome sequence of the gram-positive bacterium Bacillus subtilis. Nature 390, 249–256. Lee, D.C., Roeder, R.G., Wold, W.S., 1982. DNA sequences affecting specific initiation of transcription in vitro from the EIII promoter of adenovirus 2. Proc. Natl. Acad. Sci. U.S.A. 79, pp. 41–45. Lee, S., Kim, B., Park, K., Um, Y., Lee, J., 2012. Synthesis of pure meso-2,3-butanediol from crude glycerol using an engineered metabolic pathway in Escherichia coli. Appl. Biochem. Biotechnol. 166, 1801–1813. Li, L., Su, F., Wang, Y., Zhang, L., Liu, C., Li, J., Ma, C., Xu, P., 2012a. Genome sequences of two thermophilic Bacillus licheniformis strains, efficient producers of platform chemical 2,3-butanediol. J. Bacteriol. 194, 4133–4134.

Li, L., Wang, Y., Zhang, L., Ma, C., Wang, A., Tao, F., Xu, P., 2012b. Biocatalytic production of (2S,3S)-2,3-butanediol from diacetyl using whole cells of engineered Escherichia coli. Bioresour. Technol. 115, 111–116. Li, L., Zhang, L., Li, K., Wang, Y., Gao, C., Han, B., Ma, C., Xu, P., 2013. A newly isolated Bacillus licheniformis strain thermophilically produces 2,3-butanediol, a platform and fuel bio-chemical. Biotechnol. Biofuels 6, 123. Li, Z.J., Jian, J., Wei, X.X., Shen, X.W., Chen, G.Q., 2010. Microbial production of meso2,3-butanediol by metabolically engineered Escherichia coli under low oxygen condition. Appl. Microbiol. Biotechnol. 87, 2001–2009. Ma, C., Wang, A., Qin, J., Li, L., Ai, X., Jiang, T., Tang, H., Xu, P., 2009. Enhanced 2,3butanediol production by Klebsiella pneumoniae SDM. Appl. Microbiol. Biotechnol. 82, 49–57. May, A., Fischer, R.J., Thum, S.M., Schaffer, S., Verseck, S., Peter, Dürre, P., Bahl, H., 2013. A modified pathway for the production of acetone in Escherichia coli. Metab. Eng. 15, 218–225. Mazumdar, S., Lee, J., Oh, M.K., 2013. Microbial production of 2,3-butanediol from seaweed hydrolysate using metabolically engineered Escherichia coli. Bioresour. Technol. 136, 329–336. Mutalik, V.K., Guimaraes, J.C., Cambray, G., Lam, C., Christoffersen, M.J., Mai, Q.A., Tran, A.B., Paull, M., Keasling, J.D., Arkin, A.P., Endy, D., 2013. Precise and reliable gene expression via standard transcription and translation initiation elements. Nat. Methods 10, 354–360. Nielsen, D.R., Yoon, S.H., Yuan, C.J., Prather, K.L., 2010. Metabolic engineering of acetoin and meso-2,3-butanediol biosynthesis in E. coli. Biotechnol. J. 5, 274–284. Oliver, J.W., Machado, I.M., Yoneda, H., Atsumi, S., 2013. Cyanobacterial conversion of carbon dioxide to 2,3-butanediol. Proc. Natl. Acad. Sci. U.S.A. 110, pp. 1249– 1254. Peretti, S.W., Bailey, J.E., 1987. Simulations of host-plasmid interactions in Escherichia coli: Copy number, promoter strength, and ribosome binding site strength effects on metabolic activity and plasmid gene expression. Biotechnol. Bioeng. 29, 316–328. Phalip, V., Monnet, C., Schmitt, P., Renault, P., Godon, J.J., Diviès, C., 1994. Purification and properties of the alpha-acetolactate decarboxylase from Lactococcus lactis subsp. lactis NCDO 2118. FEBS Lett. 351, 95–99. Ragauskas, A.J., Williams, C.K., Davison, B.H., Britovsek, G., Cairney, J., Eckert, C.A., Frederick , W.J., Hallett, J.P., Leak, D.J., Liotta, C.L., Mielenz, J.R., Murphy, R., Templer, R., Tschaplinski, T., 2006. The path forward for biofuels and biomaterials. Science 311, 484–489. Reese, M.G., 2001. Application of a time-delay neural network to promoter annotation in the Drosophila melanogaster genome. Comput. Chem. 26, 51–56. Renna, M.C., Najimudin, N., Winik, L.R., Zahler, S.A., 1993. Regulation of the Bacillus subtilis alsS, alsD, and alsR genes involved in post-exponential-phase production of acetoin. J. Bacteriol. 175, 3863–3875. Sambrook, J., Russell, D.W., 2001. Molecular Cloning: A Laboratory Manual, 3rd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Shin, H.D., Yoon, S.H., Wu, J., Rutter, C., Kim, S.W., Chen, R.R., 2012. High-yield production of meso-2,3-butanediol from cellodextrin by engineered E. coli biocatalysts. Bioresour. Technol. 118, 367–373. Steinmetz, A., Vyazmensky, M., Meyer, D., Barak, Z.E., Golbik, R., Chipman, D.M., Tittmann, K., 2010. Valine 375 and phenylalanine 109 confer affinity and specificity for pyruvate as donor substrate in acetohydroxy acid synthase isozyme II from Escherichia coli. Biochemistry 49, 5188–5199. Stormer, F.C., 1975. 2,3-Butanediol biosynthetic system in Aerobacter aerogenes. Methods Enzymol. 41, 518–532. Tsau, J.L., Guffanti, A.A., Montville, T.J., 1992. Conversion of pyruvate to acetoin helps to maintain pH homeostasis in Lactobacillus plantarum. Appl. Environ. Microbiol. 58, 891–894. Ui, S., Okajima, Y., Mimura1, A., Kanai, H., Kudo, T., 1997. Molecular generation of an Escherichia coli strain producing only the meso-isomer of 2,3-butanediol. J. Ferment. Bioeng. 84, 185–189. Wang, A., Xu, Y., Ma, C., Gao, C., Li, L., Wang, Y., Tao, F., Xu, P., 2012. Efficient 2,3butanediol production from cassava powder by a crop-biomass-utilizer, Enterobacter cloacae subsp. dissolvens SDM. PLoS One 7, e40442. Woolston, B.M., Edgar, S., Stephanopoulos, G., 2013. Metabolic engineering: past and future. Annu. Rev. Chem. Biomol. Eng. 4, 259–288. Xiao, Z., Wang, X., Huang, Y., Huo, F., Zhu, X., Xi, L., Lu, J.R., 2012. Thermophilic fermentation of acetoin and 2,3-butanediol by a novel Geobacillus strain. Biotechnol. Biofuels 5, 88. Xu, H., Davies, J., Miao, V., 2007. Molecular characterization of class 3 integrons from Delftia spp. J. Bacteriol. 189, 6276–6283. Xu, Y., Tao, F., Ma, C., Xu, P., 2013. New Constitutive vectors: useful genetic engineering tools for biocatalysis. Appl. Environ. Microbiol. 79, 2836–2840. Xu, Y., Wang, A., Tao, F., Su, F., Tang, H., Ma, C., Xu, P., 2012. Genome sequence of Enterobacter cloacae subsp. dissolvens SDM, an efficient biomassutilizing producer of platform chemical 2,3-butanediol. J. Bacteriol. 194, 897–898. Yang, J., Nie, Q., Ren, M., Feng, H., Jiang, X., Zheng, Y., Liu, M., Zhang, H., Xian, M., 2013a. Metabolic engineering of Escherichia coli for the biosynthesis of alphapinene. Biotechnol. Biofuels 6, 60. Yang, T.W., Rao, Z.M., Zhang, X., Xu, M.J., Xu, Z.H., Yang, S.T., 2013b. Effects of corn steep liquor on production of 2,3-butanediol and acetoin by Bacillus subtilis. Process Biochem. 48, 1610–1617. Yomano, L.P., York, S.W., Ingram, L.O., 1998. Isolation and characterization of ethanol-tolerant mutants of Escherichia coli KO11 for fuel ethanol production. J. Ind. Microbiol. Biotechnol. 20, 132–138.

Y. Xu et al. / Metabolic Engineering 23 (2014) 22–33

Zeng, A.P., Sabra, W., 2011. Microbial production of diols as platform chemicals: recent progresses. Curr. Opin. Biotechnol. 22, 749–757. Zhang, C., Liu, L., Teng, L., Chen, J., Liu, J., Li, J., Du, G., Chen, J., 2012. Metabolic engineering of Escherichia coli BL21 for biosynthesis of heparosan, a bioengineered heparin precursor. Metab. Eng. 14, 521–527. Zhang, L., Xu, Q., Zhan, S., Li, Y., Lin, H., Sun, S., Sha, L., Hu, K., Guan, X., Shen, Y., 2013. A new NAD(H)-dependent meso-2,3-butanediol dehydrogenase from an industrially potential strain Serratia marcescens H30. Appl. Microbiol. Biotechnol., http://dx.doi.org/10.1007/s00253-013-4959-x

33

Zheng, Y., Liu, Q., Li, L., Qin, W., Yang, J., Zhang, H., Jiang, X., Cheng, T., Liu, W., Xu, X., Xian, M., 2013. Metabolic engineering of Escherichia coli for high-specificity production of isoprenol and prenol as next generation of biofuels. Biotechnol. Biofuels 6, 57. Zheng, Y.N., Li, L.L., Liu, Q., Yang, J.M., Wang, X.W., Liu, W., Xu, X., Liu, H., Zhao, G., Xian, M., 2012. Optimization of fatty alcohol biosynthesis pathway for selectively enhanced production of C12/14 and C16/18 fatty alcohols in engineered Escherichia coli. Microb. Cell. Fact. 11, 65.