An artificial synthetic pathway for acetoin, 2,3-butanediol, and 2-butanol production from ethanol using cell free multi-enzyme catalysis

Abstract

Upgrading ethanol to higher order alcohols is desired but difficult using current biotechnological methods. In this study, we designed a completely artificial reaction pathway for upgrading ethanol to acetoin, 2,3-butanediol, and 2-butanol in a cell-free bio-system composed of ethanol dehydrogenase, formolase, 2,3-butanediol dehydrogenase, diol dehydratase, and NADH oxidase. Under optimized conditions, acetoin, 2,3-butanediol, and 2-butanol were produced at 88.78%, 88.28%, and 27.25% of the theoretical yield from 100 mM ethanol, respectively. These results demonstrate that this artificial synthetic pathway is an environmentally-friendly novel approach for upgrading bio-ethanol to acetoin, 2,3-butanediol, and 2-butanol.

---

Introduction

The production of biofuels from renewable biomass has been given increased attention because of the finite supply of fossil fuels.^1–3 It is estimated that the global biofuel market consists of approximately 15% biodiesel and 85% bioethanol.^4,5 Bioethanol, a direct product of biomass fermentation, is considered a renewable energy source that can replace or blend with gasoline for transportation use.^6,7 Gasoline blended with an appropriate bioethanol as a fuel additive may increase the octane level and reduce toxic emission of pollutants.⁸ However, several disadvantages including the low energy density, high hygroscopicity, and corrosivity to engine technology and fuel pipelines limit the broad implementation of bioethanol in global transportation.^7,9,10 This has shifted research interest to higher-order alcohols such as 2,3-butanediol and butanol.^7,11,12

2,3-Butanediol is an important platform chemical and potential aviation fuel with a heating value of 27.2 kJ g⁻¹ and can be used to produce butadiene (a monomer of synthetic rubber), acetoin (a volatile compound used in foods, plant growth promoters, and biological pest control), diacetyl (a flavor enhancer), and methyl ethyl ketone (2-butanone, an excellent organic solvent).^13–15 Furthermore, 2-butanone can be hydrogenated to produce 2-butanol, which shows the highest octane number and the lowest boiling point among the four stereoisomers of butanol.^1,16 Compared to bioethanol, butanol has a higher energy density and lower hygroscopicity and can be directly blended with gasoline without the need for modifying the current vehicle system.^17,18 In recent years, many studies have been conducted to engineer microbes for 1-butanol and isobutanol production through genetic modification.^19,20 However, bio-production of 2-butanol has not been widely explored. A few studies on 2-butanol production have been conducted by extending the terminal product of 2,3-butanediol or acetolactate using metabolic engineering methods.^1,16,21 The final 2-butanol concentration was still quite low because of metabolic flux limitations and its inevitable toxicity to microbial cells.

One of the potential solutions for overcoming these limitations is to construct a simplified pathway in vitro using a limited number of enzymes, a process known as cell-free metabolic engineering (CFME). CFME is a cell-free bio-system that uses in vitro ensembles of catalytic proteins prepared from purified enzymes or crude lysates of cells to produce target products.^22,23 It can efficiently eliminate cell-associated process barriers, such as substrate or product toxicity, intracellular flux balance that results in low target product yield and unwanted by-products, and product excretion constraints by intracellular transport barriers.^4,22,24 Compared with traditional metabolic engineering, CFME has many unique advantages including no requirement for cell growth, shorter synthetic pathway, faster reaction rate, higher theoretical yield and productivity, and easier manipulation of reaction conditions, although some challenges such as activity, stability, and cost of enzymes must be resolved.^25,26

The costs of acetoin (10 000–30 000 $ per ton), 2,3-butanediol (10 000–50 000 $ per ton), and 2-butanol (10 000–78 000 $ per ton) are over 10-fold that of ethanol (800–950 $ per ton), according to the global trade data. To upgrade ethanol to higher-order alcohols, we designed a completely artificial CFME-based reaction pathway composed of ethanol dehydrogenase, formolase, 2,3-butanediol dehydrogenase, diol dehydratase, and NADH oxidase for the conversion of ethanol into C₄ alcohols. Furthermore, protein engineering was conducted to improve the catalytic efficiencies of formolase and diol dehydratase to increase the conversion rate of this artificial pathway. In addition, a novel NAD(P)H purge valve regulatory node was developed to prevent NADH buildup in this artificial reaction. The results showed that C₄ compounds including acetoin, 2,3-butanediol, and 2-butanol can be efficiently produced from ethanol using the CFME-based artificial reaction pathway.

Experimental

Enzymes and chemicals

Restriction enzymes, DNA Polymerase High Fidelity and T₄ DNA ligase were purchased from TaKaRa Biotech (Shiga, Japan) and New England Biolabs (Ipswich, MA, USA), respectively. DNA and protein markers were obtained from Tiangen Biotech (Shanghai, China). Isopropyl-beta-D-thiogalactopyranoside (IPTG), dithiothreitol (DTT), and dimethyl sulfoxide (DMSO) were purchased from Sigma-Aldrich (St Louis, MO, USA) and Sinopharm (Shanghai, China), respectively. The standards, including (3S/3R)-acetoin, (2S,3S)-2,3-butanediol, (2R,3R)-2,3-butanediol, meso-2,3-butanediol, butanone, and 2-butanol, were obtained from Sigma-Aldrich. All other chemicals, unless otherwise indicated, were of analytical grade and commercially available.

Bacterial strains, plasmids, and bacterial growth conditions

The strains and plasmids used in this study are presented in Table 1. Escherichia coli DH5α and BL21(DE3) as the cloning and expression hosts were cultured at 37 °C. The plasmid pET28a was used to construct the expression vector. Luria-Bertani (LB) medium was used for strain cultivation and recombinant protein expression. Kanamycin was added to the LB medium for cultivation of recombinant strains at a final concentration of 50 μg mL⁻¹.

Table 1 Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant genotype and description	Source
Strains
E. coli DH5α	Host of plasmid for cloning	Lab stock
E. coli BL21(DE3)	Host of plasmid for expression, F-, ompT, hsdSB(rB-mB-), gal(λ c I 857, ind1, Sam7, nin5, lacUV-T7 gene1), dcm(DE3)	Lab stock
Plasmids
pET28a	Expression vector, Km^R	Novagen
pET-EtDH	pET28a carries EtDH gene	This study
pET-EtDH:D46G	pET28a carries EtDH mutant gene	This study
pET-FLS	pET28a carries FLS gene	This study
pET-FLS:L482S	pET28a carries FLS mutant gene	This study
pET-FLS:L482R	pET28a carries FLS mutant gene	This study
pET-FLS:L482E	pET28a carries FLS mutant gene	This study
pET-BDH	pET28a carries BDH gene	This study
pET-BDH:S199A	pET28a carries BDH mutant gene	This study
pET-DDH	pET28a carries DDH gene	This study
pET-DDH:S302A	pET28a carries DDH mutant gene	This study
pET-DDH:Q337A	pET28a carries DDH mutant gene	This study
pET-DDH:F375I	pET28a carries DDH mutant gene	This study
pET-DDH:S302A/Q337A	pET28a carries DDH mutant gene	This study
pET-DDH:S302A/F375I	pET28a carries DDH mutant gene	This study
pET-DDH:Q337A/F375I	pET28a carries DDH mutant gene	This study
pET-DDH:S302A/Q337A/F375I	pET28a carries DDH mutant gene	This study
pET-dhaR	pET28a carries dhaR gene	This study
pET-DDH-dhaR	pET28a carries DDH and dhaR genes	This study
pET-NOX	pET28a carries NOX gene	This study

---

Recombinant protein expression and purification

The genes for ethanol dehydrogenase (EtDH), formolase (FLS), 2,3-butanediol dehydrogenase (BDH), diol dehydratase (DDH), and NADH oxidase (NOX) from Cupriavidus necator,²⁷Pseudomonas fluorescens,²⁸Clostridium autoethanogenum,²⁹Lactobacillus brevis,^1,30 and Lactobacillus rhamnosus³¹ were synthesized by General Biosystems, Inc. (Anhui, China) and cloned into the expression plasmid pET28a (the details of enzyme sequences are presented in the ESI†). The protein expression plasmids were introduced into E. coli BL21(DE3). Recombinant E. coli BL21(DE3) harboring pET-EtDH, pET-FLS, pET-BDH, pET-DDH, pET-dhaR, pET-DDH-dhaR, and pET-NOX were cultured at 37 °C in LB medium and induced by adding 0.5 mM IPTG when the optical density was 0.6 at 600 nm. After induction for 24 h at 18 °C, the cells were harvested by centrifugation and disrupted by sonication in an ice bath. The cell lysate was centrifuged at 8000g for 10 min to remove the cell debris. For the EtDH, FLS, BDH, dhaR, and NOX enzymes, the soluble fraction was subjected to purification with a HisTrap HP column according to the purification protocol (GE Healthcare, Little Chalfont, UK). DDH purification was carried out as described previously.³² The purified enzymes were concentrated and desalted by ultrafiltration, and then detected by SDS-PAGE.

Development of the enzyme variants

EtDH:D46G and BDH:S199A variants were generated by site-directed mutagenesis, which was performed using the primers EtDH1/EtDH2 and BDH1/BDH2 as shown in Table S1.† The recombinant plasmids pET-EtDH and pET-BDH containing the wild-type EtDH and BDH genes were used as DNA templates for PCR-amplification, respectively. Recombinant plasmids harboring the correct mutant genes were transformed into E. coli BL21(DE3) and colonies selected via kanamycin resistance were used for protein expression. After purification, the activities and kinetic parameters of the EtDH and BDH variants were determined.

To improve the catalytic efficiency of the FLS enzyme, the HotSpot Wizard 2.0 server was used to analyze hot spots by inputting the FLS structure (PDB no.: 4QPZ).^28,33 Six residuals as hot spots (T396, T446, M473, S477, L482, and L499) were identified and subjected to site-saturated mutagenesis using the primers FLS1–FLS12 (Table S1†). The recombinant plasmid pET-FLS containing the wild-type FLS gene was used as a DNA template. Recombinant plasmids containing the mutant genes were transformed into E. coli BL21(DE3). FLS variants were screened using the whole-cell biocatalytic method with acetaldehyde as a substrate. Briefly, the colonies were inoculated in LB medium and protein expression was induced by adding 0.5 mM IPTG for 24 h at 18 °C when the optical density at 600 nm reached 0.6. The cells were harvested by centrifugation and used to carry out whole-cell biocatalysis in a reaction mixture containing 50 mM phosphate buffer (pH 8.0), 100 mM acetaldehyde, and 40 g L⁻¹ wet cell weight (WCW) at 30 °C for 6 h. The product acetoin was quantified using the Voges–Proskauer (VP) reaction (see the ESI†). Moreover, positive variants were verified by enzyme activity and kinetic parameter assays after purification. Variants were also analyzed by commercial sequencing.

DDH variants including S302A, Q337A, F375I, S302A/Q337A, S302A/F375I, Q337A/F375I, and S302A/Q337A/F375I were developed to assess their catalytic efficiencies compared with the wild-type DDH enzyme. Site-directed mutagenesis was performed using the primers DDH1–DDH6 as shown in Table S1.† The recombinant plasmid pET-DDH-dhaR harboring the wild-type DDH and its reactivating factor dhaR genes were used as DNA templates for PCR amplification. The PCR products were transformed into E. coli BL21(DE3) for protein expression, which was conducted in LB medium containing 0.5 mM IPTG at 18 °C for 24 h. These variants were used to evaluate the catalytic activities using whole-cell biocatalysis with meso-2,3-butanediol as a substrate. The reaction mixture consisted of 50 mM HEPES buffer (pH 7.0), 50 mM meso-2,3-butanediol, 20 μM coenzyme B₁₂, and 40 g L⁻¹ wet cell weight, and whole-cell biocatalysis was carried out at 30 °C for 6 h in the dark. The product butanone was analyzed and quantified by gas chromatography.

Enzyme activity assays

EtDH activity. The activities of EtDH and its variant were determined in a reaction mixture containing 100 mM glycine–NaOH buffer (pH 9.5), 5 mM Mg²⁺, 3 mM NAD⁺/NADP⁺, and 10 mM ethanol at 25 °C. The activity was defined by the reduction rate of NAD⁺/NADP⁺ at 340 nm using a spectrophotometer (UV-1800, MAPADA, Shanghai, China). One unit of EtDH activity was defined as the amount of enzyme required to reduce 1 μmol of NAD⁺/NADP⁺ per minute.

FLS activity. The activities of FLS and its variants were assayed in a reaction mixture containing 100 mM phosphate buffer (pH 8.0), 1 mM Mg²⁺, 0.1 mM TPP, and 20 mM acetaldehyde. After the reaction was conducted at room temperature for 1 h, acetoin concentration from acetaldehyde was determined by the VP reaction and calculated from the calibration curves of standard acetoin. One unit of FLS activity was defined as the amount of enzyme that produced 1 μmol acetoin from acetaldehyde per minute.

BDH activity. The activities of BDH and its variant were measured in a reaction mixture containing 50 mM Tris-HCl buffer (pH 7.5), 0.2 mM NADPH, 1 mM DTT, and 20 mM acetoin or 5 mM butanone at room temperature. The activity was defined by the oxidation rate of NADPH at 340 nm using a spectrophotometer (UV-1800, MAPADA). One unit of BDH activity was defined as the amount of enzyme required to oxidize 1 μmol of NADPH per minute.

DDH activity. The activities of DDH and its variants with its reactivating factor dhaR were determined in a reaction mixture containing 50 mM phosphate buffer (pH 7.0), 1 mM coenzyme B₁₂, 100 mM ATP, 1 mM Mg²⁺, and 50 mM meso-2,3-butanediol. After conversion at room temperature in the dark for 1 h, the reaction was stopped by adding an equal volume of citrate buffer (100 mM, pH 3.6). The product butanone was detected and quantified by gas chromatography. One unit of DDH activity was defined as the amount of enzyme that produces 1 μmol butanone from meso-2,3-butanediol per minute.

NOX activity. The NOX activity was determined in a reaction mixture containing 50 mM HEPES–NaOH buffer (pH 8.0) and 0.2 mM NADH at room temperature. The activity was defined by the oxidation rate of NADH at 340 nm using a spectrophotometer (UV-1800, MAPADA). One unit of NOX activity was defined as the amount of enzyme required to oxidize 1 μmol of NADH per minute.

Determination of kinetic parameters

The kinetic parameters of EtDH and EtDH:D46G were determined in a reaction mixture containing 100 mM glycine–NaOH buffer (pH 9.5), 5 mM Mg²⁺, 3 mM NAD⁺/NADP⁺, and 0.5–100 mM ethanol at room temperature. The K_m and k_cat values were obtained by a nonlinear regression fitting of the Michaelis–Menten equation. All assays were carried out in triplicate.

The kinetic parameters of FLS and its variants were assayed in a reaction mixture containing 100 mM phosphate buffer (pH 8.0), 1 mM Mg²⁺, 0.1 mM TPP, and 0.5–20 mM acetaldehyde at room temperature. The K_m and k_cat values were obtained by a nonlinear regression fitting of the Michaelis–Menten equation. All assays were carried out in triplicate.

The kinetic parameters of BDH and BDH:S199A were measured in a reaction mixture containing 50 mM Tris-HCl buffer (pH 7.5), 0.2 mM NADPH, 1 mM DTT, and 0.5–100 mM acetoin or 0.5–10 mM butanone at room temperature. The K_m and k_cat values were obtained by a nonlinear regression fitting of the Michaelis–Menten equation. All assays were carried out in triplicate.

Cell-free multi-enzyme catalysis system

The synthesis of acetoin, 2,3-butanediol, and 2-butanol from ethanol by cell-free multi-enzyme catalysis was conducted in a 0.5 mL reaction mixture containing the substrate, coenzymes, metal ions, and the corresponding enzymes. The reaction conditions including temperature, pH, coenzyme, and metal ions were optimized to improve the flux of the artificial reaction pathway (see the ESI†). The optimal reaction conditions for acetoin, 2,3-butanediol, and 2-butanol from ethanol were as follows.

Acetoin production was carried out in a 0.5 mL reaction mixture containing 50 mM HEPES buffer (pH 8.0), 1 mM NAD⁺, 0.1 mg mL⁻¹ EtDH, 0.2 mg mL⁻¹ FLS:L482S, 0.1 mg mL⁻¹ NOX, 0.1 mM TPP, 1 mM Mg²⁺, 1 mM DTT, 20% DMSO, and 100 mM ethanol. The reaction was conducted at 30 °C.

2,3-Butanediol was produced in a 0.5 mL reaction mixture containing 50 mM HEPES buffer (pH 8.0), 1 mM NAD⁺, 1 mM NADP⁺, 0.1 mg mL⁻¹ EtDH:D46G, 0.2 mg mL⁻¹ FLS:L482S, 0.1 mg mL⁻¹ NOX, 0.1 mg mL⁻¹ BDH:S199A, 0.1 mM TPP, 1 mM Mg²⁺, 1 mM DTT, 20% DMSO, and 100 mM ethanol. The reaction was conducted at 30 °C.

2-Butanol was produced in a 0.5 mL reaction mixture containing 50 mM HEPES buffer (pH 8.0), 1 mM NAD⁺, 1 mM NADP⁺, 0.1 mg mL⁻¹ EtDH:D46G, 0.2 mg mL⁻¹ FLS:L482S, 0.1 mg mL⁻¹ NOX, 0.1 mg mL⁻¹ BDH:S199A, 0.2 mg mL⁻¹ DDH:Q337A/F375I, 0.2 mg mL⁻¹ dhaR, 0.1 mM TPP, 1 mM Mg²⁺, 1 mM DTT, 20% DMSO, 1 mM coenzyme B₁₂, 100 mM ATP, and 100 mM ethanol. The reaction was carried out at 30 °C.

All reactions were carried out for 6 h and the products were identified by gas chromatography (GC). The product percentage yields were calculated using the following equation: percentage yield (%) = product yield (mM)/theoretical yield (mM). Theoretically, two moles of ethanol can generate one mole of acetoin or 2,3-butanediol or 2-butanol.

Recyclability of cascade reactions

Purified enzymes were mixed with activated silicon oxide nanoparticles and incubated for 12 h at 4 °C. Before immobilization, the silicon oxide nanoparticles (4830HT; Nanostructured & Amorphous Materials, Houston, TX, USA) were activated by treating the nanoparticles with glutaraldehyde (Sigma). Immobilization yield (%) and immobilization efficiency (%) were calculated for the immobilized enzymes as follows: immobilization efficiency = (α_i/α_f) × 100, immobilization yield = [{P_i − (P_w + P_s)}/P_i] × 100, where α_i is the total activity of the immobilized enzyme and α_f is the total activity of the free enzyme. P_i is the total protein content of the crude enzyme preparation and P_w and P_s are the protein concentrations in the wash solution and supernatant after immobilization, respectively.

For acetoin production, the reaction was performed in a 0.5 mL reaction mixture containing 50 mM HEPES buffer (pH 8.0), 1 mM NAD⁺, 1.06 U mL⁻¹ EtDH, 0.05 U mL⁻¹ FLS:L482S, 0.98 U mL⁻¹ NOX, 0.1 mM TPP, 1 mM Mg²⁺, 1 mM DTT, 20% DMSO, and 100 mM ethanol. Similarly, 2,3-butanediol production was performed in a 0.5 mL reaction mixture containing 50 mM HEPES buffer (pH 8.0), 1 mM NAD⁺, 1 mM NADP⁺, 0.1 mg mL⁻¹ EtDH:D46G, 0.2 mg mL⁻¹ FLS:L482S, 0.1 mg mL⁻¹ NOX, 0.1 mg mL⁻¹ BDH:S199A, 0.1 mM TPP, 1 mM Mg²⁺, 1 mM DTT, 20% DMSO, and 100 mM ethanol. The reaction was conducted at 30 °C for 6 h. The reusability of immobilized enzymes was examined under the same reaction conditions. After each reaction cycle, the immobilized enzyme was removed by centrifugation at 4000g for 30 min. The immobilized enzyme was collected and washed with deionized water and buffer. For the second reaction cycle, the immobilized enzyme was dissolved in fresh buffer, added to the substrate, and processed as in the first cycle.

Analytical methods

Cell growth was determined by measuring the optical density at 600 nm with a spectrophotometer (UV-1800, MAPADA). Protein concentration was determined using the Bradford method, and bovine serum albumin served as the standard protein. The concentrations of ethanol, acetaldehyde, acetoin, 2,3-butanediol, butanone, and 2-butanol were determined with the addition of isoamylol as an internal standard and quantified using a gas chromatograph system (Agilent GC9860, Santa Clara, CA, USA) equipped with a chiral column (Supelco β-DEX™ 120, 30 m length, 0.25 mm inner diameter). The operating conditions were as follows: N₂ was used as the carrier gas at a flow rate of 1.2 mL min⁻¹; the injector temperature and detector temperature were 215 and 245 °C, respectively; and the column temperature was maintained at 50 °C for 1.5 min, and then increased to 180 °C at a rate of 15 °C min⁻¹.

Results and discussion

Construction of an artificial pathway for C₄ compound production from ethanol

We designed a novel artificial reaction pathway for the conversion of ethanol into the C₄ compounds acetoin, 2,3-butanediol, and 2-butanol, which required cascade enzymes (Fig. 1). Ethanol is first dehydrogenated by the NAD(P)H-dependent EtDH to afford acetaldehyde, which undergoes a condensation reaction to yield acetoin by FLS, a computationally designed enzyme. Subsequently, acetoin is reduced by the NADPH-dependent BDH to produce 2,3-butanediol. Finally, 2-butanol is obtained via a dehydration and hydrogenation reaction by DDH and BDH, respectively. NOX is used to regenerate NAD⁺ throughout the reaction process. Considering that oxygen is used as a substrate for NOX, aeration may be required for a large-scale reaction. By stoichiometry, 1 M ethanol is converted completely into acetoin or 2,3-butanediol and generates 1 or 0.5 M NADH, respectively, which requires 0.5 M oxygen for acetoin production or 0.25 M oxygen for 2,3-butanediol production (1 M NADH oxidized by NOX requires 0.5 M oxygen as a substrate). Thus, 1 L of the reaction mixture containing 1 M ethanol as a substrate requires 11.2 or 5.6 L oxygen for acetoin or 2,3-butanediol production throughout the reaction process. For 2-butanol production, the oxygen demand in the reaction was lower than that of 2,3-butanediol production. Compared to traditional fermentation, the aeration demand in the current reaction system is much lower supporting the scale-up potential of the artificial reaction pathway. In this artificial reaction pathway, we chose EtDH from C. necator,²⁷ BDH from C. autoethanogenum,²⁹ DDH from L. brevis^1,30 with its reactivating factor dhaR, and NOX enzymes from L. rhamnosus³¹ as candidate enzymes because of their relatively high catalytic efficiencies for their corresponding substrates as demonstrated in previous studies. However, the FLS enzyme was only reported to catalyze three molecules of formaldehyde into one molecule of dihydroxyacetone.²⁸ To verify the catalytic ability of acetaldehyde conversion to acetoin by the FLS enzyme, the catalytic reaction was carried out using whole-cell biocatalysts over-expressing FLS with acetaldehyde as a substrate. The VP test and GC/GC-MS analysis indicated that racemic acetoin ((3S)-acetoin and (3R)-acetoin) was produced from acetaldehyde, demonstrating that the FLS enzyme can catalyze the conversion of acetaldehyde into acetoin (Fig. S1, S14 and S15†). These results show that C₄ compounds including acetoin, 2,3-butanediol, and 2-butanol can be produced from ethanol using the artificial reaction pathway.

Fig. 1 Bio-system based on the artificial reaction cascade for the conversion of ethanol to acetoin (A), 2,3-butanediol (B), and 2-butanol (C). The enzymes are ethanol dehydrogenase (EtDH and its variant EtDH:D46G from C. necator), formolase (FLS:L482S, a newly designed enzyme originate from P. fluorescens), 2,3-butanediol dehydrogenase (BDH:S199A from C. autoethanogenum), diol dehydratase (DDH:Q337/F375I from L. brevis), and NADH oxidase (NOX from L. rhamnosus).

Characterization of enzymes in the artificial pathway

Furthermore, to determine the catalytic efficiencies of these enzymes in the artificial pathway, we systematically determined the kinetic parameters of EtDH, FLS, BDH, and DDH after purification (the results of enzyme expression and purification are presented in ESI, Fig. S2 and S3†). As shown in Table 2, the k_cat/K_m values of EtDH for ethanol with NAD⁺, FLS for acetaldehyde with TPP, and BDH for acetoin/butanone with NADPH were 17.09, 7.69 × 10⁻³, 18.50, and 19.31 s⁻¹ mM⁻¹, respectively. However, the purified DDH enzyme with its reactivating factor dhaR showed a low activity (0.02 U mg⁻¹) towards meso-2,3-butanediol as a substrate, partly because of its low stability.³² A whole-cell catalytic method was used to determine the conversion of three 2,3-butanediol isomers (meso-2,3-butanediol, (2R,3R)-2,3-butanediol, and (2S,3S)-2,3-butanediol) to butanone by E. coli/pET-DDH-dhaR cells. The results showed that 20.56 mM butanone was produced from 50 mM meso-2,3-butanediol after 6 h (Fig. S4†), indicating that the DDH enzyme with dhaR had a high catalytic activity towards meso-2,3-butanediol in vivo. In contrast, no butanone was detected during whole-cell catalysis in the presence of (2R,3R)-2,3-butanediol and (2S,3S)-2,3-butanediol as substrates (Fig. S4†). These results are consistent with those of a previous study showing that only meso-2,3-butanediol was dehydrated by the DDH enzyme.³⁴ The enzyme thermostability assay showed that the three enzymes had relatively higher stabilities and retained 87.91% (EtDH), 70.43% (FLS), and 91.30% (NOX) of their initial activities after incubation at 30 °C for 6 h, while all enzymes except NOX showed a decreased stability at 37 and 45 °C (Fig. S5†).

Table 2 Kinetic parameters of EtDH, FLS, BDH, and their variants

Enzyme	Substrate/coenzyme	K m (mM)	k cat (s^−1)	k cat/Km (s^−1 mM^−1)	Ref.
EtDH	Ethanol/NAD^+	0.37 ± 0.05	6.28 ± 0.11	17.09	This study
	Ethanol/NADP^+	0	0	0	This study
EtDH:D46G	Ethanol/NAD^+	0.57 ± 0.03	5.65 ± 0.09	9.97	This study
	Ethanol/NADP^+	0.60 ± 0.02	0.99 ± 0.03	1.65	This study
FLS	Acetaldehyde/TPP	58.46 ± 2.32	0.45 ± 0.03	7.69 × 10^−3	This study
FLS:L482S	Acetaldehyde/TPP	47.45 ± 1.20	0.63 ± 0.01	1.33 × 10^−2	This study
FLS:L482R	Acetaldehyde/TPP	50.27 ± 1.44	0.53 ± 0.02	1.06 × 10^−2	This study
FLS:L482E	Acetaldehyde/TPP	50.95 ± 1.51	0.49 ± 0.03	9.66 × 10^−3	This study
BDH	Acetoin/NADPH	84.86 ± 7.98	157.0 ± 9.0	18.50	This study
	Acetoin/NADH	0	0	0	This study
	Butanone/NADPH	1.94 ± 0.05	29.90 ± 1.02	15.41	This study
	Butanone/NADH	0	0	0	This study
BDH:S199A	Acetoin/NADPH	116.1 ± 8.7	224.3 ± 9.7	19.31	This study
	Acetoin/NADH	0	0	0	This study
	Butanone/NADPH	1.08 ± 0.03	40.78 ± 1.42	37.76	This study
	Butanone/NADH	0	0	0	This study
DDH	2,3-Butanediol/B12	10.4	35	3.4	36
NOX	O2/NADH	5.8 × 10^−3	218.7	3.77 × 10^4	31

---

Acetoin production from ethanol via the artificial pathway

Rate limiting step for acetoin production from ethanol. To test the viability of the artificial pathway, we first investigated acetoin production from ethanol using three enzymes, EtDH, FLS, and NOX, as presented in Fig. 1A. The initial reaction was carried out in a 0.5 mL reaction mixture containing 50 mM HEPES buffer (pH 7.0), 0.1 mg mL⁻¹ EtDH, 0.2 mg mL⁻¹ FLS, 0.1 mg mL⁻¹ NOX, 4 mM NAD⁺, 0.1 mM TPP, 1 mM Mg²⁺, 1 mM DTT, 20% DMSO, and 100 mM ethanol as the starting substrate. The reaction was conducted at 30 °C for 6 h and 17.98 mM of acetoin at 35.96% of the theoretical yield was generated in the reaction solution after 6 h, suggesting that the reaction pathway for upgrading ethanol to acetoin is feasible. Furthermore, the reaction conditions including temperature, pH, coenzymes (NAD⁺ and TPP), and metal ions were optimized to improve the reaction flux from ethanol to acetoin. The optimal reaction conditions including temperature (30 °C), pH (8.0), NAD⁺ (1 mM), TPP (0.1 mM), and Mg²⁺ (1 mM) were determined (Fig. S6†). Under the optimized conditions, 22.75 mM of acetoin at 45.50% of the theoretical yield was produced from ethanol (100 mM) after 6 h (Fig. S7†). To determine the rate-limiting step in the reaction, the reaction pathway was systematically reconstituted with 1/10 of the starting amount of each enzyme, while the other enzymes were kept constant. The results are shown in Fig. 2. For EtDH and NOX, lowering the concentration to 10% had minimal effects on the production of acetoin from ethanol. The exception was FLS, indicating that FLS represents potential bottlenecks.

Fig. 2 Analysis of the rate-limiting step in the conversion of ethanol to acetoin. Results are the means ± SD of three parallel replicates.

Engineering of FLS to improve its activity. To improve the catalytic efficiency of the FLS enzyme, mutational hot spots in the FLS amino acid sequence were analyzed using the HotSpot Wizard 2.0 server.³³ Six hot spot residues (T396, T446, M473, S477, L482, and L499) were identified and altered by site-saturated mutagenesis (Fig. S8†). The results revealed that site 482 in FLS plays an important role in enzyme activity. The FLS variants L482S, L482R, and L482E showed a specific activity increase of 59.03%, 36.89%, and 34.12%, respectively (Fig. 3 and Table S2†). Furthermore, the kinetic parameters of the three FLS variants were assayed using the substrate acetaldehyde with TPP as a coenzyme. The k_cat/K_m values of the FLS variants L482S, L482R, and L482E for acetaldehyde were 1.33 × 10⁻², 1.06 × 10⁻², and 9.66 × 10⁻³ with improvements of 72.95%, 37.84%, and 25.62%, respectively, compared to wild-type FLS (Table 2).

Fig. 3 Screening of positive FLS variants by whole-cell biocatalysis using acetaldehyde (100 mM) as a substrate (A) and activity assays of FLS and its variants after purification (B). (A) The product acetoin from acetaldehyde through whole-cell biocatalysis was determined using the VP test. (B) Percentage ratios were calculated based on specific activities (specific activity of FLS was 0.16 U mg⁻¹). Results are the means ± SD of three parallel replicates.

The mutant FLS:L482S showed a higher activity than the wild-type FLS. Computational studies for FLS variants at atomic resolution using molecular dynamics simulation for 100 ns provided insight into the structural changes for the mutation L482S and its correlation with enzymatic activity. Computational analysis revealed that W480 in the mutant FLS:L482S made stronger contacts (2.1 Å) with the substrate acetaldehyde than the wild-type (2.8 Å) (Fig. S9A and B†). Siegel et al. (2015) reported W480 as an active site residue in FLS (4QPZ), supporting the higher activity of FLS:L482S.²⁸ The 100 ns molecular dynamics analysis also shows that the FLS:L482S retained more hydrogen bond contacts than the wild-type enzyme (Fig. S9C and D†).

Acetoin production from ethanol using mutant FLS:L482S. The FLS:L482S rather than the wild-type enzyme was used to conduct the reaction from ethanol to acetoin under optimized reaction conditions. As shown in Fig. 4A, a maximum concentration of acetoin (44.39 mM) was obtained from ethanol (100 mM) at 4 h, with an acetoin yield of 88.78% of the theoretical yield. In addition to acetoin, the substrate ethanol and intermediate acetaldehyde were measured during the reaction. The substrate ethanol was rapidly consumed and 6.25 mM residual ethanol remained in the reaction solution at 4 h. Acetaldehyde accumulated up to 9.58 mM during the first 2 h of the reaction, and then decreased to 5.65 mM at 6 h. No accumulation of other undesired by-products was detected, indicating that the artificial synthetic pathway was specific for the substrate. Subsequently, the effects of substrate concentrations (200, 300, 400, and 500 mM) on the acetoin production were investigated using the optimized reaction mixture. As shown in Fig. 4B, the acetoin concentration was increased by increasing the initial substrate concentration. The maximum acetoin concentration of 117.5 mM was achieved at an ethanol concentration of 400 mM after 6 h. However, the acetoin yield was only 58.71% of its theoretical yield. The intermediate acetaldehyde showed a significant accumulation when the initial ethanol concentration was increased, suggesting that the FLS:L482S enzyme limits the productivity of this CFME-based artificial pathway.

Fig. 4 Time-course of acetoin production from 100 mM ethanol (A) and effects of different substrate concentrations on acetoin production (B) by the artificial reaction cascade. (A) The reaction was performed in a 0.5 mL reaction mixture containing 50 mM HEPES buffer (pH 8.0), 1 mM NAD⁺, 1.06 U ml⁻¹ EtDH, 0.05 U mL⁻¹ FLS:L482S, 0.98 U mL⁻¹ NOX, 0.1 mM TPP, 1 mM Mg²⁺, 1 mM DTT, 20% DMSO, and 100 mM ethanol. (B) The reactions were carried out under the same conditions, except that the substrate concentration was 200–500 mM ethanol. These reactions were conducted at 30 °C for 6 h. Results are the means ± SD of three parallel replicates.

2,3-Butanediol production from ethanol via the artificial pathway

NAD(P)H purge valve regulatory node to prevent NADH buildup. Subsequently, we investigated the 2,3-butanediol production from ethanol with the enzyme BDH added to the CFME system. However, the artificial pathway from ethanol to 2,3-butanediol generates more reducing equivalents than is required to produce 2,3-butanediol (1 NADH per ethanol, with 0.5 NADH required). Herein, we designed a NAD(P)H purge valve regulatory node to prevent the buildup of NADH. As shown in Fig. 1B, we developed the mutant enzyme EtDH (D46G) to simultaneously utilize NAD⁺ and NADP⁺ as coenzymes in the conversion of ethanol to acetaldehyde by site-directed mutagenesis according to the results of the sequence alignment (Table 2 and Fig. S10†). Additionally, NADPH-dependent BDH^1,29 from C. autoethanogenum was used to catalyze acetoin into 2,3-butanediol in the presence of NADPH. During the reaction process, NADH produced was rapidly recycled back to NAD⁺via the H₂O-forming NOX, while NADPH generated was used to produce 2,3-butanediol from acetoin. When the NADP⁺ levels were high (NADPH low), the mutant enzyme EtDH:D46G regenerated NADPH via the oxidation of ethanol. However, when the NADPH levels were high (NADP⁺ low), BDH generated 2,3-butanediol from acetoin and regenerate NADP⁺ without the buildup of additional reducing equivalents (NADH produced from conversion of ethanol to acetaldehyde was immediately recycled for purging the excess). To achieve high catalytic efficiency, the mutant BDH enzyme (S199A) was developed via site-specific mutagenesis as reported previously.³⁵

Characterization of mutants EtDH and BDH in the artificial pathway. As shown in Table 2 and Table S2,† the wild-type EtDH enzyme only employed NAD⁺ as a coenzyme for ethanol oxidation with 17.09 s⁻¹ mM⁻¹ of the k_cat/K_m value, whereas its variant EtDH:D46G simultaneously used NAD⁺ and NADP⁺ as coenzymes and the k_cat/K_m values for ethanol with NAD⁺ and NADP⁺ were 9.97 and 1.65 s⁻¹ mM⁻¹, respectively. Compared with the wild-type NADPH-dependent BDH enzyme, the k_cat/K_m value of its variant BDH:S199A for acetoin with NADPH was up to 19.31 s⁻¹ mM⁻¹ (Table 2 and Table S2†). In addition, the enzyme thermostability of EtDH:D46G and BDH:S199A was assayed at 30, 37, and 45 °C. The EtDH:D46G activity levels of 86.53% and 86.67% were retained after incubation at 30 °C for 6 h when acetaldehyde was used as a substrate with the coenzymes NAD⁺ and NADP⁺, respectively (Fig. S5†). For the BDH:S199A enzyme, 80.81% of the initial enzyme activity for acetoin with the coenzyme NADPH was retained at 30 °C for 6 h (Fig. S5†).

2,3-Butanediol production from ethanol via the improved artificial pathway. The four enzymes, EtDH:D46G, FLS:L482S, BDH:S199A, and NOX, were reacted with ethanol in the presence of 1 mM NAD⁺ and/or 1 mM NADP⁺. As shown in Fig. S11,† no 2,3-butanediol was detected when NAD⁺ was used as a coenzyme in the reaction system. This was primarily because BDH is an NADPH-dependent enzyme in the conversion of acetoin to 2,3-butanediol. In contrast, 18.20 mM 2,3-butanediol was produced from 100 mM ethanol as a substrate in the presence of NADP⁺, indicating that NADP⁺ is required for 2,3-butanediol production in the artificial cascade reaction. When 1 mM NAD⁺ and 1 mM NADP⁺ were simultaneously used in the reaction, a maximum 2,3-butanediol concentration of 44.14 mM at 88.28% of the theoretical yield was produced from 100 mM ethanol after 5 h (Fig. 5A). The result showed that the use of the NAD(P)H purge valve regulatory node for preventing the buildup of NADH was feasible and efficient. Throughout the reaction process, low levels of acetoin accumulated in the reaction solution, demonstrating a high catalytic efficiency from acetoin to 2,3-butanediol by BDH:S199A (Fig. 5A). Furthermore, different substrate concentrations were used to investigate the effects on the 2,3-butanediol production in the reaction pathway. The maximum 2,3-butanediol concentration was 127.3 mM when the initial ethanol concentration in the reaction mixture was 500 mM (Fig. 5B). However, the 2,3-butanediol yield decreased with the increase in initial ethanol concentration, although a higher 2,3-butanediol concentration was obtained. Analysis of the intermediates in the reaction solution revealed accumulation of a large amount of acetaldehyde in the reaction system, indicating that the conversion of acetaldehyde to acetoin by the FLS:L482S enzyme remained the rate-limiting step. Compared to the results for the acetoin production from ethanol, the concentration of acetaldehyde in the 2,3-butanediol production system was significantly decreased (Fig. 4B and 5B), suggesting that the FLS:L482S enzyme activity is partially inhibited by the product acetoin. However, the relatively low catalytic efficiency of the FLS:L482S enzyme remained the major cause of acetaldehyde accumulation.

Fig. 5 Time-course of 2,3-butanediol production from 100 mM ethanol (A) and effects of different substrate concentrations on 2,3-butanediol production (B) by the artificial reaction cascade. (A) The reaction was carried out in a 0.5 mL reaction mixture containing 50 mM HEPES buffer (pH 8.0), 1 mM NAD⁺, 1 mM NADP⁺, 0.88 U mL⁻¹ EtDH:D46G, 0.05 U mL⁻¹ FLS:L482S, 0.98 U mL⁻¹ NOX, 5.11 U mL⁻¹ BDH:S199A, 0.1 mM TPP, 1 mM Mg²⁺, 1 mM DTT, 20% DMSO, and 100 mM ethanol. (B) The reactions were carried out under the same conditions except that the substrate concentrations were 200–500 mM ethanol. These reactions were conducted at 30 °C for 6 h. Results are the means ± SD of three parallel replicates.

2-Butanol production from ethanol

Construction of an artificial pathway for 2-butanol production from ethanol. Finally, we employed an additional B₁₂-dependent DDH to convert meso-2,3-butanediol into butanone,^30,32,36 which can further be transformed into 2-butanol by BDH:S199A from C. autoethanogenum as shown in Fig. 1C.³⁵ Previous studies showed that the DDH enzyme with its reactivating factor dhaR from L. brevis was the best candidate for catalyzing meso-2,3-butanediol into butanone in vivo.¹ To determine its reaction conditions in vitro, the effects of four factors including coenzyme B12, ATP, Mg²⁺, and dhaR on the conversion of meso-2,3-butanediol to butanone by DDH were investigated. The results suggested that the coenzyme B12 and ATP were required for the catalytic reaction, whereas dhaR and Mg²⁺ efficiently enhanced the butanone production from meso-2,3-butanediol (Table S3†). Therefore, coenzyme B12, ATP, Mg²⁺, and dhaR were supplemented into the artificial reaction pathway for the efficient butanone production from meso-2,3-butanediol. To convert butanone to 2-butanol, the BDH:S199A enzyme was used because it can reduce acetoin and butanone into 2,3-butanediol and 2-butanol, respectively.³⁵ As shown in Table 2 and Table S2,† the BDH:S199A enzyme showed a high catalytic efficiency for the substrate butanone with NADP⁺ (k_cat/K_m = 37.76 s⁻¹ mM⁻¹), which was improved by 145.0% compared to wild-type BDH. Considering that 2-butanol exists in two forms (R-2-butanol and S-2-butanol), we also analyzed the configuration of the product from butanone using BDH:S199A. As shown in Fig. S16,†R- and S-2-butanol in a ratio of 1.23 : 1 were produced from butanone using BDH:S199A at 30 °C after 6 h.

Theoretically, two moles of ethanol are oxidized to generate two moles of reducing equivalents, which satisfies the need for the reduction of acetoin and butanone into 2,3-butanediol and 2-butanol in the artificial pathway (Fig. 1C). However, the DDH enzyme only catalyzed meso-2,3-butanediol into butanone, whereas (2R,3R)-butanediol and (2S,3S)-butanediol are not substrates for DDH (Fig. S4†). Thus, reducing equivalents would be excessive if the products from acetoin by BDH were other forms of 2,3-butanediol except meso-2,3-butanediol. Therefore, the configurations of acetoin and 2,3-butanediol were determined using a GC system equipped with a chiral column. Chiral analysis indicated that 2,3-butanediol generated from ethanol contained meso-2,3-butanediol (65%) and (2R,3R)-butanediol (35%) from (3S)-acetoin and (3R)-acetoin, which were produced from acetaldehyde using FLS:L482S (Fig. S14 and S15†). Considering that only meso-2,3-butanediol was converted into butanone by DDH,³⁴ we used the same strategy for the 2-butanol production as the 2,3-butanediol production from ethanol by employing an NAD(P)H purge valve regulatory node to prevent the buildup of NADH (Fig. 1C).

Rate limiting step for 2-butanol production from ethanol. The reaction for 2-butanol production was carried out to determine the rate-limiting step using EtDH:D46G, FLS:L482S, BDH:S199A, DDH with its reactivating factor dhaR, and NOX enzymes. The rate-limiting experiment showed that only 1.84 mM 2-butanol with a significant decrease of 79.16% was obtained from 100 mM ethanol when the 1/10 concentration of the DDH enzyme was used, whereas lowering the concentration of BDH:S199A to 10% had little effect on the 2-butanol production (Fig. 6A). This demonstrates that DDH became a potential rate-limiting factor in the artificial reaction for the 2-butanol production. Based on previous studies,^1,30 S302A, Q337A, F375I, and their combinatorial variants in the DDH enzyme were developed by site-directed mutagenesis. The results of whole-cell catalysis indicated that the mutant Q337A and F375I DDH enzymes with dhaR produced 27.20 and 25.63 mM butanone from 50 mM meso-2,3-butanediol, while 19.76 mM butanone was obtained using the wild-type DDH enzyme with dhaR. Furthermore, a maximum butanone concentration of 30.35 mM, which was an increase of 53.60%, was observed by the combinatorial variant of DDH:Q337A/F375I with its reactivating factor dhaR (Fig. 6B and Fig. S12†).

Fig. 6 Analysis of the rate-limiting step in the conversion of ethanol to 2-butanol (A). Catalytic efficiency assays of the DDH enzyme and its variants with meso-2,3-butanediol (50 mM) as a substrate using the whole-cell biocatalytic method (B). Results are the means ± SD of three parallel replicates.

Computational analysis showed that the proton donor residue E171 of the double mutant DDH (Q337A/F375I) (1.9 Å) comes into contact with the substrate 2,3-butanediol more strongly than the wild-type (2.6 Å) (Fig. S13A and B†). E171 showed more hydrogen bond contacts with the substrate in the double mutant compared to the wild-type. Long-range simulation (100 ns) results also showed that the substrate had stronger interaction networks including hydrogen bonding and water bridges with E171 in the double mutant compared to that in the wild-type (Fig. S13C and D†).

2-Butanol production from ethanol by the improved artificial pathway. DDH:Q337A/F375I with dhaR rather than wild-type DDH was used to perform the cascade reaction from ethanol to 2-butanol (Fig. 1C). As shown in Fig. 7, 13.62 mM 2-butanol at 27.24% of the theoretical yield was produced from 100 mM ethanol after 6 h in the cell-free bio-system. Furthermore, the high levels of 2,3-butanediol (up to 32.55 mM) accumulated in the reaction solution, whereas butanone was not detected during the reaction process. These results indicate that the conversion of 2,3-butanediol to butanone by DDH:Q337A/F375I was a rate-limiting step. Although the whole-cell biocatalyst co-expressing the DDH:Q337A/F375I enzyme and its reactivating factor dhaR exhibited a higher catalytic efficiency for the conversion of meso-2,3-butanediol (50 mM) to butanone (30.35 mM), the in vitro enzyme activity assay showed a low specific activity of 0.05 U mg⁻¹ for meso-2,3-butanediol (Table S2† and Fig. 6B). The thermal instability of the DDH:Q337A/F375I enzyme also limited the conversion of 2,3-butanediol to butanone. As shown in Fig. S5,† the DDH:Q337A/F375I enzyme only retained 1.21% of its initial activity after incubation at 30 °C for 6 h, and no activity was observed when the enzyme was incubated at 37 and 45 °C for 6 h. During the reaction process, 2-butanol production rapidly increased during the first 4 h and then stopped increasing after 5 h, likely because of inactivation (Fig. 7). Therefore, further improving the enzyme activity and stability of DDH by protein engineering or searching for a new enzyme with high catalytic efficiency to convert 2,3-butanediol to butanone would improve 2-butanol yield in the artificial synthetic pathway.

Fig. 7 Conversion of ethanol into 2-butanol by the artificial reaction cascade. Ethanol (squares), acetaldehyde (inverted triangles), acetoin (circles), 2,3-butanediol (diamonds), butanone (stars), and 2-butanol (upright triangles). The reaction was carried out in a 0.5 mL reaction mixture containing 50 mM HEPES buffer (pH 8.0), 1 mM NAD⁺, 1 mM NADP⁺, 0.88 U mL⁻¹ EtDH:D46G, 0.05 U mL⁻¹ FLS:L482S, 0.98 U mL⁻¹ NOX, 5.11 U mL⁻¹ BDH:S199A, 0.01 U mL⁻¹ DDH:Q337A/F375I, 0.2 mg mL⁻¹ dhaR, 0.1 mM TPP, 1 mM Mg²⁺, 1 mM DTT, 20% DMSO, 1 mM coenzyme B₁₂, 100 mM ATP and 100 mM ethanol. The reaction was conducted at 30 °C for 6 h. Results are the means ± SD of three parallel replicates.

Recyclability of cascade reactions

The recyclability of the catalytic system is an attractive parameter that dictates the economic feasibility of using these enzymes in the cell-free bio-system. To evaluate the recyclability of the catalytic system, the enzymes for acetoin production (EtDH, FLS:L482S, and NOX) and 2,3-butanediol production (EtDH:D46G, FLS:L482S, BDH:S199A, and NOX) were immobilized on glutaraldehyde-functionalized silicon oxide nanoparticles.^37,38 Immobilized enzymes showed more than 90% immobilization efficiency. Acetoin production by the immobilized enzymes was evaluated and a recyclability test of the catalytic system was performed (Fig. 8). Enzymes immobilized on silicon oxide nanoparticles exhibited 94% of their initial product concentrations, even after 10 cycles of reuse to produce acetoin. This indicates that there was no significant decrease in the acetoin production using the immobilized enzymes during repeated use. Additionally, the enzymes immobilized on silicon oxide nanoparticles were stable because of the covalent linkages between the enzymes and silicon oxide nanoparticles. Similarly, the 2,3-butanediol production by immobilized enzymes and recyclability of the catalytic system were evaluated (Fig. 8). Here, the immobilized enzymes exhibited 73% of their initial 2,3-butanediol production activity after 10 cycles of reuse. Upon immobilization, biocatalysts catalyzed the production of acetoin and 2,3-butanediol for many cycles without significant loss of activity. Hence, the immobilized catalytic system can be efficiently recycled for industrial production of acetoin and 2,3-butanediol. Unfortunately, we were not able to produce 2-butanol under the same experimental conditions because of the low stability of DDH.

Fig. 8 Recyclability of the catalytic system. (A) Conversion of ethanol into acetoin by the artificial cascade reaction using immobilized enzymes. Ethanol (squares), acetaldehyde (upright triangles), and acetoin (circles). (B) Reusability of the immobilized enzymes to produce acetoin from ethanol. (C) Conversion of ethanol into 2,3-butanediol by the artificial reaction cascade using immobilized enzymes. Ethanol (squares), acetaldehyde (upright triangles), acetoin (circles), and 2,3-butanediol (inverted triangles). (D) Reusability of the immobilized enzymes to produce 2,3-butanediol from ethanol. The product concentration of the immobilized enzyme catalyzed reaction after the first cycle was set at 100%. Acetoin and 2,3-butanediol concentration were determined using a gas chromatograph system equipped with a chiral column (Supelco β-DEX™ 120, 30 m length, 0.25 mm inner diameter).

In recent years, acetoin, 2,3-butanediol, and 2-butanol as platform chemicals have gained more attention because of their increasing market demands. The derivatives of 2,3-butanediol have a potential global market of approximately 32 million tons per annum.²⁹ As a precursor of 2,3-butanediol, acetoin has been classified as one of the 30 platform chemicals that are given priority for development and utilization by the US Department of Energy, and its market demand has reached more than 10 kilotons per annum,³⁹ while the market capacity of bio-butanol is approximately 5 million tons per annum.⁴⁰ Previous studies have shown that acetoin and 2,3-butanediol could be produced by microbial fermentation or chemical synthesis.^39,41 Compared to chemical synthesis, microbial fermentation for acetoin and 2,3-butanediol production has attracted great attention due to its mild conditions. However, during the fermentation process, many metabolic byproducts are inevitably produced, resulting in difficulty in product isolation and high purification cost. In addition, the resulting microbial biomass causes environmental problems. Commercially available 2-butanol is primarily obtained industrially by the hydration of 1-butene using sulfuric acid as a catalyst. However, the harsh reaction conditions result in a higher cost for the 2-butanol production and potential environmental pollution. Compared to chemical synthesis and microbial fermentation, the artificial multi-enzyme pathway in the current work provides an environmentally-friendly, fewer byproduct-forming, and recyclable process with high conversion efficiency, thus supporting the viability of producing acetoin, 2,3-butanediol, and 2-butanol from ethanol by CFME.

Conclusions

In conclusion, we developed an artificial multi-enzyme pathway capable of upgrading ethanol to C₄ compounds (Table 3). High activity (>88% conversion) was obtained for acetoin or 2,3-butanediol production from ethanol. Additionally, the artificial synthetic pathway showed potential for butanone and 2-butanol production, although their yields were relatively low because of the low activity and stability of the DDH enzyme. This environmentally friendly novel approach can be used to upgrade bio-ethanol to acetoin, 2,3-butanediol, and 2-butanol. Ongoing efforts are focused on developing modified FLS and DDH enzymes with high catalytic efficiency as well as identifying new enzymes. Overall, this artificial CFME approach is a promising and highly selective strategy for ethanol-to-C₄ compound conversion using our novel cascade enzyme system.

Table 3 Summary of acetoin, 2,3-butanediol, and 2-butanol production from 100 mM ethanol by cell free multi-enzyme catalysis

Product (mM)	Enzymes (U mL^−1)	Chemicals (mM)	Yield (%)	Reaction time (h)
All reactions were performed in 50 mM HEPES buffer (pH 8.0) and 20% DMSO using 100 mM EtOH as a substrate at 30 °C.
Acetoin (44.39)	EtDH (1.06), FLS:L482S (0.05), NOX (0.98)	NAD^+ (1), TPP (0.1), Mg^2+ (1), DTT (1)	88.78	4
2,3-BD (44.14)	EtDH:D46G (0.88), FLS:L482S (0.05), NOX (0.98), BDH:S199A (5.11)	NAD^+ (1), NADP^+ (1), TPP (0.1), Mg^2+ (1), DTT (1)	88.28	5
2-Butanol (13.62)	EtDH:D46G (0.88), FLS:L482S (0.05), NOX (0.98), BDH:S199A (5.11), DDH:Q337A/F375I (0.01)	NAD^+ (1), NADP^+ (1), TPP (0.1), Mg^2+ (1), DTT (1), coenzyme B12 (1), ATP (100)	27.24	6

---

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (no. 81673542), the New Century Excellent Talents Supporting Plan of the Provincial Education Department of Fujian Province of China (no. K8015056A), the Development Platform of Edible Fungi Industry in Fujian Province (no. K5114001A), and the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT & Future Planning (2017R1A2B3011676, 2017R1A4A1014806, 2013 M3A6A8073184).

References

1. Z. Chen, Y. Wu, J. H. Huang and D. H. Liu, Bioresour. Technol., 2015, 197, 260.
2. H. Latif, A. A. Zeidan, A. T. Nielsen and K. Zengler, Curr. Opin. Biotechnol., 2014, 27, 79.
3. A. S. Heeres, C. S. Picone, L. A. van der Wielen, R. L. Cunha and M. C. Cuellar, Trends Biotechnol., 2014, 32, 221.
4. C. Gao, Z. Li, L. J. Zhang, C. Wang, K. Li, C. Q. Ma and P. Xu, Green Chem., 2015, 17, 804.
5. M. Simoes, S. Baranton and C. Coutanceau, ChemSusChem, 2012, 5, 2106.
6. A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek, J. Cairney, C. A. Eckert, W. J. Frederick, J. P. Hallett, D. J. Leak, C. L. Liotta, J. R. Mielenz, R. Murphy, R. Templer and T. Tschaplinski, Science, 2006, 311, 484.
7. K. N. Tseng, S. Lin, J. W. Kampf and N. K. Szymczak, Chem. Commun., 2016, 52, 2901.
8. S. Manzetti and O. Andersen, Fuel, 2015, 140, 293.
9. P. Dürre, Biotechnol. J., 2007, 12, 1525.
10. V. García, J. Päkkilä, H. Ojamo, E. Muurinen and R. L. Keiski, Renewable Sustainable Energy Rev., 2011, 2, 964.
11. E. I. Lan and J. C. Liao, Bioresour. Technol., 2013, 135, 339.
12. P. P. Peralta-Yahya, F. Zhang, S. B. Del Cardayre and J. D. Keasling, Nature, 2012, 7411, 320.
13. L. X. Li, K. Li, Y. Wang, C. Chen, Y. Q. Xu, L. J. Zhang, B. B. Han, C. Gao, F. Tao, C. Q. Ma and P. Xu, Metab. Eng., 2016, 28, 19.
14. L. Y. Zhang, Y. L. Yang, J. A. Sun, Y. L. Shen, D. Z. Wei, J. W. Zhu and J. Chu, Bioresour. Technol., 2010, 101, 1961.
15. N. Li, W. J. Yan, P. J. Yang, H. X. Zhang, Z. J. Wang, J. F. Zheng, S. P. Jia and Z. P. Zhu, Green Chem., 2016, 18, 6029.
16. P. Ghiaci, J. Norbeck and C. Larsson, PLoS One, 2014, 9, e102774.
17. D. Cai, Y. Wang, C. J. Chen, P. Y. Qin, Q. Miao, C. W. Zhang, P. Li and T. W. Tan, Bioresour. Technol., 2016, 211, 704.
18. P. S. Nigam and A. Singh, Prog. Energy Combust. Sci., 2011, 37, 52.
19. A. Baez, K. M. Cho and J. C. Liao, Appl. Microbiol. Biotechnol., 2011, 90, 1681.
20. C. R. Shen, E. I. Lan, Y. Dekishima, A. Baez, K. M. Cho and J. C. Liao, Appl. Environ. Microbiol., 2011, 77, 2905.
21. B. R. Oh, S. Y. Heo, S. M. Lee, W. K. Hong, J. M. Park, Y. R. Jung, D. H. Kim, J. H. Sohn, J. W. Seo and C. H. Kim, Biotechnol. Lett., 2014, 36, 57.
22. Q. M. Dudley, A. S. Karim and M. C. Jewett, Biotechnol. J., 2015, 10, 69.
23. W. D. Fessner, New Biotechnol., 2015, 32, 658.
24. C. You, H. Chen, S. Myung, N. Sathitsuksanoh, H. Ma, X. Z. Zhang, J. Li and Y. H. Zhang, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 7182.
25. J. S. Martín del Campo, J. Rollin, S. Myung, Y. Chun, S. Chandrayan, R. Patiño, M. W. Adams and Y. H. Zhang, Angew. Chem., Int. Ed., 2013, 52, 4587.
26. S. Myung, J. Rollin, C. You, F. Sun, S. Chandrayan, M. W. Adams and Y. H. Zhang, Metab. Eng., 2014, 24, 70.
27. T. Y. Wu, C. T. Chen, J. T. Liu, I. W. Bogorad, R. Damoiseaux and J. C. Liao, Appl. Microbiol. Biotechnol., 2016, 100, 1.
28. J. Siegel, A. L. Smith, S. Poust, A. Wargacki, A. Bar-Even, C. Louw, B. Shen, C. Eiben, H. Tran, E. Noor, J. Gallaher, J. Bale, Y. Yoshikuni, M. Gelb, J. Keasling, B. Stoddard, M. Lidstrom and D. Baker, Proc. Natl. Acad. Sci. U. S. A., 2015, 112, 3704.
29. M. Köpke, M. L. Gerth, D. J. Maddock, A. P. Mueller, F. M. Liew and S. D. Simpson, Appl. Environ. Microbiol., 2014, 80, 3394.
30. M. Yamanishi, K. Kinoshita, M. Fukuoka, T. Saito, A. Tanokuchi, Y. Ikeda, H. Obayashi, K. Mori, N. Shibata, T. Tobimatsu and T. Toraya, FEBS J., 2012, 279, 793.
31. Y. W. Zhang, M. K. Tiwari, H. Gao, S. S. Dhiman, M. Jeya and J. K. Lee, Enzyme Microb. Technol., 2012, 50, 255.
32. M. Seyfried, R. Daniel and G. Gottschalk, J. Bacteriol., 1996, 178, 5793.
33. J. Bendl, J. Stourac, E. Sebestova, O. Vavra, M. Musil, J. Brezovsky and J. Damborsky, Nucleic Acids Res., 2016, 44, 479.
34. Z. Chen, H. Sun, J. H. Huang, Y. Wu and D. H. Liu, PLoS One, 2015, 10, e0140508.
35. D. J. Maddock, W. M. Patrick and M. L. Gerth, Protein Eng., Des. Sel., 2015, 28, 251.
36. S. Kwak, Y. C. Park and J. H. Seo, Bioresour. Technol., 2013, 135, 432.
37. S. K. S. Patel, S. H. Choi, Y. C. Kang and J. K. Lee, Nanoscale, 2016, 8, 6728.
38. S. K. S. Patel, S. H. Choi, Y. C. Kang and J. K. Lee, ACS Appl. Mater. Interfaces, 2017, 9, 2213.
39. L. J. Zhang, Q. Y. Liu, Y. S. Ge, L. X. Li, C. Gao, P. Xu and C. Q. Ma, Green Chem., 2015, 18, 1560.
40. G. R. Ding, B. Meng, R. L. Chen and J. H. Chen, Chem. Ind., 2016, 34, 37.
41. Y. Q. Xu, H. P. Chu, C. Gao, F. Tao, Z. K. Zhou, K. Li, L. X. Li, C. Q. Ma and P. Xu, Metab. Eng., 2014, 23, 22.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7gc02898a

‡ These authors contributed equally to this work.

---