Engineering Escherichia coli for autoinducible production of n-butanol

Background: Escherichia coli does not produce n-butanol naturally, but can be butanologenic when related enzymes were expressed using inducible elements on plasmids. In this study we attempted to confer E. coli strain capability of automatic excretion of the chemical by employing a native anaerobic promoter. Also, a novel DNA kit was designed for PCR preparation of linear DNA fragments to perform strain modi ﬁ cation. The kit is primarily composed of two mother vectors, co-transformation of linear DNAs into E. coli can simultaneously introduce two butanol synthetic operons into the chromosome and create two in-frame gene deletions at targeted native loci. Results: E.coli strainBw2VcarriesplasmidpCNA-PHCand pENA-TA,bothutilizesnativeanaerobic promoterP hya for the expression of butanol synthetic enzymes. When Bw2V was subjected in anaerobic fermentation using medium containing extra glucose, the accumulated n-butanol in the broth was up to 2.8 g/L in bioreactor; as the genetic element expressing the same pathway was introduced into the genome, the titer of butanol was 1.4 g/L. Conclusions: TheexpressionsystemusingP hya iseffectiveinapplicationsthatinvolveexpressionplasmidsasalso applicable in ectopic expression as single copy on the chromosome. Results imply that P hya can be subjected for broader application in bioproduction of more feedstock chemicals.


Introduction
Biobutanol (microbial produced n-butanol) is regarded as one of potential clean biofuels, and has been the renewed focus of bioenergy studies in recent years due to incessantly fermented public concerns on energy crisis and the environmental problems incurred by excessive use of fossil oil. The industrial biobutanol is usually produced by the Clostridia genus through acetone-butanol-ethanol (ABE) fermentation process [1,2].
Novel butanol producing micro-organisms other than Clostridia were in rapid development in anticipation of overcoming the negative aspects of the ABE fermentation technology. Among of them, Escherichia coli has become an ideal candidate because of the abundant genetic tools, its rapid growth, and relatively rich knowledge on the regulation of its metabolic network [3]. Unlike Clostridia, E. coli is not a native producer of n-butanol, the formation of this chemical would not be accomplished before the heterologous pathway is introduced into host cells; with expression plasmids, E. coli was able to accumulate significantly high amount of n-butanol [4]. However, the risk of plasmid loss or copy number reduction is high, bringing cautious concerns in strains' long-term genetic stability. Therefore, chromosomal expression, if possible, is always favored during construction of industrial strains.
In addition, the promoter chosen always played pivotal roles in production of proteins or biochemicals, therefore, utilization of strong promoters in plasmid expression system is frequently reported due to their contribution in increased yields of products [5]. In E. coli, strong inducible promoters, such as T7/tac, are specially favored in the laboratory production of proteins for biochemical characterization [6]; however, the use of expensive inducer, such as IPTG is not desirable for industrial practice. In this study, P hya , a native promoter of hydrogenase I cluster, was tested for anaerobic production of metabolites, and the aim of present study is to develop a comprehensive solution to confer E. coli to be a potential industrial strain that is able to produce butanol efficiently during anaerobic growth, through renewed genomic technology that enabled rapid genetic modification in host cells' chromosome.

Bacterial strains, media and grow conditions
Main strains, plasmids and the primers in this study are listed in Table 1. BW25113 (Yale CGSC) and its derivate mutants were made Electronic Journal of Biotechnology 18 (2015) 138-142 competent for the transformation of vectors expressing butanol synthetic pathway [7]. Bw2V carries two plasmids for the expression of the pathway; BuS1 is a butanologenic recombinant derived from BW25113, null in adhE and ldhA, and has the synthetic butanol pathway integrated at those sites with selective markers remained; both drug resistant markers were subsequently removed in strain BuS2 with the aid of helper plasmid pCP20. Luria broth (LB) and Terrfic Broth (TB) with 0.5% glucose were used for growth of E. coli cells, ampicillin (100 mg L -1 ), kanamycin (50 mg L -1 ), streptomycin (50 mg L -1 ) and chloramphenicol (50 mg L -1 ) were added as needed for selection. Anaerobic fermentation in shaking flasks was performed in sealed serum bottles containing 100 or 200 mL of aerobically pregrown cultures, with shaking at 100 rpm at 37°C.

Construction of plasmids and butanologenic strains
The biosynthetic pathway for n-butanol production in this study is composed of phaA (acetoacetyl-CoAthiolase/synthase), hbd (3-hydroxybutyryl-CoAdehydrogenase), crt (crotonase), ter (crotonyl-CoA reductase) and adhE1 (butanol dehydrogenase) [8,9]. We assembled phaA, hbd and ter as an artificial operon (designated as PHC) through DNA synthesis, ter and adhE1 in the other operon (TA). P hya was cloned into pACYCduet-1 and pETduet-1, generating pCNA and pENA respectively (supplementary file); then as seen in Fig. 1a, PHC/TA were inserted behind P hya in pCNA or pENA to get pCNA-PHC and pENA-TA, both of them were transformed into wild type E. coli strain BW25113 to generate Bw2V (information on construction of pCNA-PHC and pENA-TA as well as the sequences of the operons are available in the Supplementary materials). Next, to facilitate genomic modification, we attempted to grant the expression vectors with removable selective markers. One popular choice was kan, encoding kanamycin resistance enzyme, which can be obtained through PCR using pKD4 as template, DNA fragment containing kan and frt sites was inserted into pCNA-PHC, generating pCNA-PHCF (illustrated construction routes are available in the Supplementary materials); to speed up the strain modification process, we planned to put another selective marker into plasmid pENA-TA, therefore we inserted a chemically synthesized DNA fragment containing streptomycin phosphotransferase gene (sm) with frt into the unique Eco RI site on plasmid pENA-TA, resulting in plasmid pENA-TAF (Fig. 1b). pCNA-PHCF and pENA-TAF are working as the templates for the generation of linear DNAs through PCR using primers containing short homologous DNA fragments of 40-50 nucleotides for Red/ET aided recombination at targeted loci [10]. This strategy has been proved to be very efficient in strain modification, it was immediately applied in in-frame deletion of adhE and ldhA, in order to reduce the production of ethanol and lactic acid during anaerobic growth. As demonstrated in Fig. 1c, linear DNAs of TA-sm-ΔldhA and PHC-kan-ΔadhE were prepared by primer pair TA-F/TA-Rand PHC-F/PHC-R, and used to co-transformed into BW25113 bearing pKD46, the resultant BuS1 was selected on LB agar plates containing both kanamycin and streptomycin, and confirmed as null in adhE and ldhA; ensuing the introduction of pCP20 into BuS1 removed kan and sm, finally generated butalogenic strain BuS2, thus, BuS2 owns an intact copy of butanol biosynthetic pathway on its genome by displacing native adhE and ldhA with the operon PHC and TA, respectively.

In vivo production of n-butanol
n-Butanol fermentation was performed either through 2-stage flask shaking culturing, or 5-L stirred tank bioreactor (BIOSTAT-A plus Sartorius) with a working volume of 2.5 L. In flask experiments, 4 mL of overnight pre-culture was inoculated into 80 mL of fresh medium with antibiotics, cells were grown at 37°C with shaking at 200 rpm, with the cell density monitored by measuring OD 600 . When OD 600 reached 1.0, cells were transferred into 100 mL or 250 mL of serum bottles with stoppers for anaerobic growth, extra glucose (20 g/L) was added into the cultures. In Bioreactor experiments, the broth was inoculated with 200 mL of overnight pre-culture, dissolved oxygen (DO) was maintained above 25% during aerobic stage by raising stirrer speed and using air pump. The cells were switched to anaerobic growth after OD 600 reached 5.0, anaerobic conditions were created by shutting off the air pump and sealing the outlets of the fermenter with water, pH was adjusted and maintained at 5.5 automatically using 1 M NaOH as base.

Quantification of n-butanol and other metabolites
The fermented products in the broth including n-butanol were quantified by gas chromatograph (GC, Shimadzu 2010 plus) equipped with an Inert Cap Pure War (30 m × 0.25 mm × 0.25 μm) and a flame-ionized detector. Cell free samples were collected from the cultures by centrifugation at 12,000 × g for 2 min; the supernatant was then mixed in a 9:1 ratio with an aqueous solution containing isobutanol internal standard (10 g/L); samples were quantified relative to a standard curve of 250, 500, 1000, 2000, 4000, 8000 mg/L n-butanol, standard curves were prepared freshly during each batch run and normalized for injection volume with isobutanol as standard. The oven temperature was held at 50°C for 4 min and subsequently programmed to increase from 50°C to 220°C at the rate of 20°C/min, then held temperature at 220°C for 3 min, the injector was set at 200°C and the detector temperature was set at 230°C.

Formation of n-butanol in E. coli transformants containing expression vectors
To test the capacity of anaerobic promoter P hya for the production of n-butanol, the synthetic butanol pathway (PHC and TA) were put behind P hya , through the construction of expression plasmids pCNA-PHC and pENA-TA. Both of plasmids were transformed into Bw25113, resulting in butanologenic strain Bw2V. The transformant was grown aerobically and switched to anaerobic fermentation in TB, or LB (with or without 20 g/L glucose supplemented). As shown in Fig. 2a, after 24 h of anaerobic incubation in sealed serum bottles, the accumulated butanol from Bw2V was detected to be 0.32 g/L in LB or 0.57 g/L in LBG, the titer further increased to 0.78 g/L in TB with extra glucose. To probe the maximum of the production level, Bw2V was cultured in bioreactor with controlled pH and DO as seen in Fig. 2b, the maximum of 2.77 g/L of butanol was observed in the bioreactor. However, ethanol was also produced unavoidably at a high concentration of 4.2 g/L (Fig. 2b), suggesting that the deletion of native adhE would play beneficial role in increasing the productivity of butanol; besides, lactic acids were also excreted and maintained stably at a titer of about 1.5 g/L (Fig. 2b), indicating that the disruption of E. coli ldhA should be performed to further activate butanol synthesis.

Butanol and other metabolites produced by E. coli strain BuS2
To investigate the full capacity of E. coli strain BuS2 in butanol production, fermentation in 5 L bioreactor using TB medium was performed with extended incubation time. Optimal density at wavelength of 600 nm (OD 600 ), consumption of glucose, pH and growth temperature were all monitored, the data of related metabolic chemicals synthesized by the cells were collected during after anaerobic switch. As shown in Fig. 2c, n-butanol titer was found to be constantly increasing in newly 100 h, the maximum of accumulated butanol concentration reached 1.43 g/L, which was compatible or greater than the level obtained by other E. coli strains expressing polycistronic genes using inducible promoters [11]. As seen in Fig. 2d, the disruption of adhE and ldhA led to the decreased production of ethanol (b1 g/L) and lactic acid (less than 0.1 g/L), two of the main fermentation products from this E. coli host. However, the titer of acetic acid produced by BuS2 was rather high, over 8.0 g/L (Fig. 2d). Overproduction of acetate might explained why the broth pH dropped quickly during anaerobic fermentation, and that will consume lot of base solution if we controlled the broth pH at 6.0 or above, therefore the fermentation pH was maintained at 5.5. Also, it is expected that by shunting the carbon flux from the formation of acetic acid in BuS2, butanol productivity can be further elevated in the future by simultaneous gene deletion of pta and poxB if the genetic manipulation doesn't disturb the intracellular redox balance.

Discussion
In the present work we investigated the possibility of employing P hya as an auto-inducible promoter for the production of n-butanol in engineered E. coli. As a growth phase inducible element, P hya has been previously applied in heterologous expression of Pyrococcus furiosus hydrogenase [12], but its capacity on transcriptional initiation has never been evaluated upon genomic integration. This study demonstrated that the expression system using P hya is not only effective in applications that involve expression plasmids, but also applicable in ectopic expression as single copy on the chromosome. The insertion of single copy of synthetic pathway driven by P hya was able to convert E. coli a b c d model strain BW25113 into n-butanol excreting strain (BuS2), and the accumulated butanol yield reached 1.4 g/L, far exceeding the yields achieved by strains employing promoters other than T7 [11,13], which implies that P hya can be subjected for broader application in the bioproduction of more feedstock chemicals.
To perform rapid engineering in E. coli, we linked the optimized gene clusters with necessary elements coming from Red/ET recombination technology, and innovatively constructed a DNA kit composed by two plasmids templates for easy handling of gene deletion and chromosomal insertion in E. coli. Butanologenic strain BuS2 was obtained instantly by using this strategy. In comparison with routine process, this route saves time and labor costs. Also, because the generated strains are free of selective marker, the mutants can be used in the next round of genetic engineering for further strain improvement.
It is also worthy to note that, Clostridium acetobutylicum AdhE2 was reported to be more efficient in enzymatic conversion during butanol production than AdhE1 when they were expressed in E. coli [14], it would then be promising to increase butanol productivity by using adhE2 in the next generation of DNA kit instead of adhE1 in the present study.

Financial support
This work was supported by a general program (No. 31270122) from the National Natural Science Foundation of China, and the Scientific Research Foundation for the Returned Overseas Chinese Scholars (State Education Ministry supported program).