Stoichiometric Conversion of Maltose for Biomanufacturing by In Vitro Synthetic Enzymatic Biosystems

Maltose is a natural α-(1,4)-linked disaccharide with wide applications in food industries and microbial fermentation. However, maltose has scarcely been used for in vitro biosynthesis, possibly because its phosphorylation by maltose phosphorylase (MP) yields β-glucose 1-phosphate (β-G1P) that cannot be utilized by α-phosphoglucomutase (α-PGM) commonly found in in vitro synthetic enzymatic biosystems previously constructed by our group. Herein, we designed an in vitro synthetic enzymatic reaction module comprised of MP, β-phosphoglucomutase (β-PGM), and polyphosphate glucokinase (PPGK) for the stoichiometric conversion of each maltose molecule to two glucose 6-phosphate (G6P) molecules. Based on this synthetic module, we further constructed two in vitro synthetic biosystems to produce bioelectricity and fructose 1,6-diphosphate (FDP), respectively. The 14-enzyme biobattery achieved a Faraday efficiency of 96.4% and a maximal power density of 0.6 mW/cm2, whereas the 5-enzyme in vitro FDP-producing biosystem yielded 187.0 mM FDP from 50 g/L (139 mM) maltose by adopting a fed-batch substrate feeding strategy. Our study not only suggests new application scenarios for maltose but also provides novel strategies for the high-efficient production of bioelectricity and value-added biochemicals.

Maltose is a natural disaccharide consisting of two molecules of glucose joined by an α- (1,4) glycosidic linkage.As the key structural motif of starch, maltose can be readily produced from starch hydrolysis catalyzed by amylase [18][19][20].The practicability and relatively low cost of maltose render this disaccharide a preferred material for the industrial production of confectioneries, ice creams, pastries, and beverages.Maltose has also been used as a carbon source for fermentation [21][22][23], as well as a substrate for wholecell biocatalysis to produce trehalose [24].Despite that maltose is one of the most readily available carbohydrates, few in vitro synthetic enzymatic biosystems have been reported to utilize maltose as the substrate for biomanufacturing.The main difference of maltose from the other abovementioned disaccharides and polysaccharides is that its phosphorylation catalyzed by maltose phosphorylase (MP) yields β-glucose 1-phosphate (β-G1P), whereas the phosphorylation of sucrose, cellobiose, cellulose, and starch yields α-G1P.Correspondingly, the PGMs used in our previously established in vitro biosystems were all α-phosphoglucomutases (α-PGMs) which is a key enzyme linking the glycolysis and gluconeogenesis pathways for the interconversion of α-G1P and α-/β-G6P [25,26] and cannot convert β-G1P to α-/β-G6P [27].Noted that the interconversion between αand β-G6P occurs spontaneously, whereas the interconversion between αand β-G1P does not.These facts lead to the mismatch between MP and α-PGM which might account for the lack of maltose-based in vitro biosystems to date.
In this study, two in vitro synthetic enzymatic biosystems were designed for the stoichiometric utilization of maltose for biomanufacturing.At first, a three-enzyme reaction module consisting of MP, β-phosphoglucomutase (β-PGM), and PPGK was designed for the stoichiometric conversion of maltose to G6P.Then, more downstream enzymes were added to prepare two in vitro synthetic enzymatic biosystems for the high-yield generation of bioelectricity and FDP, respectively.Our study not only provides new application scenarios for maltose but also suggests novel strategies for the high-efficient synthesis of other products such as hydrogen, inositol, glucosamine, and rare sugars in the future.

Materials and Methods
2.1.Experimental Design.Aiming at the stoichiometric utilization of maltose for biomanufacturing, we first designed the enzymatic pathways capable of converting both glucose units of maltose into G6P, which could be further utilized by other enzymes for biomanufacturing.Next, enzymes in the pathways were expressed and purified individually.Proof-of-concept experiments were conducted to evaluate the feasibility of our designed pathway.The consumption of maltose and the production of bioelectricity and FDP as well as the accumulation of G6P were quantified.If necessary, the reaction systems would subsequently be optimized to improve their product yields.The schematic workflow is illustrated in Figure 1.

Enzyme
Preparation.E. coli BL21(DE3) transformed with the constructed plasmid was cultivated at 37 °C in LB medium containing either 100 μg/mL ampicillin or 50 μg/mL kanamycin.Recombinant protein expression was induced by adding a final concentration of 0.01-0.1 mM of isopropyl-β-D-thiogalactopyranoside (IPTG) when the absorbance of the bacterial culture at 600 nm reached 0.8-1.2.The bacterial culture was further incubated at 37 °C for 4 h or at 16 °C for 20 h.The cells were harvested by centrifugation at 4 °C, washed twice with 50 mM HEPES buffer (pH 7.0), and resuspended in 50 mM HEPES buffer (pH 7.5) containing 100 mM NaCl.The suspended cells were then placed in an ice bath and lysed by a Fisher Scientific Sonic Dismembrator Model 500 ultrasonicator.Three approaches, including nickel affinity chromatography with Ni Sepharose 6 Fast Flow medium (GE Healthcare, USA), heat precipitation at 70 °C for 20 min, and affinity adsorption of carbohydrate-binding module (CBM) on regenerated amorphous cellulose followed by selfcleavage of intein [36,37], were adopted for enzyme purification (Table S2).Each enzyme was expressed and purified individually.The purities of the recombinant enzymes were analyzed by SDS-PAGE.The concentrations of proteins were determined using the Bradford method with bovine serum albumin as standard.
2.5.Enzyme Activity Assay.The activity of β-PGM was determined based on the generation of G6P.Reactions were performed at 37 °C in 100 mM HEPES (pH 7.0) containing 10 mM β-G1P and 10 mM MgCl 2 .The reaction was initiated by the addition of β-PGM and stopped by cooling in an icewater bath.G6P was determined by mixing 50 μL of the sample with 200 μL of 100 mM HEPES buffer (pH 7.0) containing 10 mM MgCl 2 , 1 mM NAD + , and 1 U/mL G6PDH.After incubation at 37 °C for 15 min, the increase of absorbance at 340 nm was measured.One enzyme unit was 2 BioDesign Research defined as the amount of enzyme that produced 1 μmol of G6P from β-G1P per minute.Optimal temperature of β-PGM was determined in 100 mM HEPES (pH 7.0) within the temperature range of 25-90 °C.Optimal pH of β-PGM was tested at 37 °C in 100 mM Bis-Tris buffer (pH 6.0-7.0) and 100 mM HEPES buffer (pH 7.0-8.0).The activity of MP was determined at 37 °C according to the generation of β-G1P from maltose.Reactions were performed in 100 mM HEPES buffer (pH 7.0) containing 50 mM NaH 2 PO 4 -Na 2 HPO 4 and 10 mM maltose.The reaction was initiated by the addition of MP, stopped with HClO 4 , and then neutralized with KOH.G6P was determined by mixing 50 μL of sample with 200 μL of 100 mM HEPES buffer (pH 7.0) containing 10 mM MgCl 2 , 1 mM NAD + , 1 U/mL β-PGM, and 1 U/mL G6PDH.After incubation at 37 °C for 15 min, the increase of absorbance at 340 nm was measured.One enzyme unit was defined as the amount of enzyme that produced 1 μmol of β-G1P from maltose per minute.
The activities of the rest of the enzymes were measured at 37 °C using the method described previously, as listed in Table S2.
2.6.Electrochemical Measurement.All electrochemical experiments were conducted at 37 °C using a CHI660E Potentiostat (CH Instruments Inc.).The enzymatic reaction solution contained 100 mM HEPES buffer (pH 7.3), 100 mM NaCl, 8 mM NAD + , 10 mM MgCl 2 , 0.5 mM MnCl 2 , 0.5 mM sodium hexametaphosphate, 0.5 mM thiamine pyrophosphate, 2 mM potassium phosphate, 100 mg/L ampicillin, 50 mg/L kanamycin, enzymes, and 0.5 mM maltose.For electrochemical measurements using a 10 mL three-electrode system, the working electrode, the reference electrode, and the counter electrode were prepared as previously described [11] and were immersed into the enzymatic reaction solution.For the determination of Faraday efficiency, chronoamperometry measurements at 0.16 V were performed to record the gener-ated current as a function of the reaction time.The applied potential of 0.16 V for oxidation was determined by cyclic voltammetry (CV).Nitrogen flushing was used to remove oxygen in this three-electrode system.The total charge was calculated based on time and the measured current.The Faraday efficiency (ƞF) was calculated based on a previously described method [38].Theoretically, 48 electrons can be generated from each maltose molecule.This value was used for the calculation of ƞF.
To evaluate the current and power densities, a mediated electron transfer (MET) type of two-chamber enzymatic fuel cell comprised of the cathode, bioanode, and a Nafion 212 membrane was constructed.A 1 × 1 cm 2 carbon felt was placed in the middle of the cathode chamber containing ferricyanide solution and connected with a titanium wire.Another 1 × 1 cm 2 carbon felt adsorbed with VK 3 was placed in the middle of the anodic chamber containing the enzymatic reaction solution as described above and connected to an external circuit through the titanium wire.The volumes of the anode and cathode solutions in the container were both 4 mL.Linear sweep voltammetry at a scan rate of 1 mV/s was then conducted.The current density was calculated based on the surface area of the bioanode.The power density was calculated from the current density and the recorded potential.

Production of FDP from
Maltose.Proof-of-concept production of FDP from 5 g/L (13.9 mM) maltose was carried out at 37 °C in 100 mM HEPES buffer (pH 7.0) containing 5 mM MgCl 2 , 100 mg/L ampicillin, 50 mg/L kanamycin, 20 mM potassium phosphate, 60 mM sodium pyrophosphate, 8 mM sodium hexametaphosphate, and 3 U/mL of each enzyme.The reaction volume was 1 mL.Optimization of enzyme concentrations was performed under the same conditions as described above, except that the concentration of a single enzyme was varied while the concentrations of the other four enzymes were fixed.The reaction was carried out

BioDesign Research
for 5 h.To compromise between high product titer and low cost of enzymes, the enzyme concentration that resulted in a relatively high FDP titer was chosen as the optimal value.This optimized enzyme concentration was then used for the subsequent enzyme optimization processes.Enzymes were optimized in the order of MP, β-PGM, PPGK, PGI, and PP i -PFK through this systematic titration method.
Fed-batch production of FDP was carried out at 37 °C in 100 mM HEPES buffer (pH 7.0) containing 50 mM MgCl 2 , 100 mg/L ampicillin, 50 mg/L kanamycin, 20 mM potassium phosphate, 80 mM sodium hexametaphosphate, 5 U/mL MP, 20 U/mL β-PGM, 10 U/mL PPGK, 5 U/mL PGI, and 10 U/mL PP i -PFK.The reaction volume was 2 mL.At 0 h, a final concentration of 20 g/L (55.6 mM) maltose was added together with 224 mM sodium pyrophosphate to initiate the reaction.An aliquot of the reaction mixture was taken every hour to determine the concentration of FDP.When the reaction reached equilibrium, another 15 g/L (41.7 mM) maltose and 168 mM sodium pyrophosphate were supplemented to the system.This operation was repeated when the reaction reach equilibrium again.
2.8.Quantification of Maltose and FDP.Maltose was quantified by HPLC equipped with a refractive index detector.An Aminex HPX-87H column (Bio-Rad) was used for separation, and the column temperature was maintained at 80 °C. 5 mM H 2 SO 4 was used as mobile phase at a flow rate of 0.6 mL/min.The injection volume was set as 20 μL.The concentrations of maltose in the samples were calculated based on the peak intensities with commercial maltose monohydrate (molecular weight: 360.31) as a reference.The quantification of FDP followed the method of Wang et al. [14].10 μL of the sample was mixed with 300 μL of FDP assay solution which contained 200 mM triethanolamine buffer (pH 7.6), 0.1 mM NADH, 0.135 U/mL ALD (purchased from Sigma-Aldrich), 2.5 U/mL TIM from Thermus thermophilus, and 0.4 U/mL glycerol 3-phosphate dehydrogenase (GPDH, purchased from Sigma-Aldrich).The reaction was conducted at 30 °C for 30 min.The amount of FDP in the sample was calculated based on the consumption of NADH which could be measured at 340 nm by a spectrophotometer.Results were means ± standard deviation of three parallel replicates.

Stoichiometric Production of G6P from Maltose.
In this study, we designed an in vitro synthetic enzymatic reaction scheme to utilize maltose for biomanufacturing (Figure 2).The key of this scheme is a three-enzyme reaction module aiming at the stoichiometric conversion of maltose into G6P.In this module, (1) maltose is phosphorylated by maltose phosphorylase (MP) to yield β-G1P and glucose, (2) β-G1P requires β-phosphoglucomutase (β-PGM) for its conversion to G6P, and (3) glucose is phosphorylated by polyphosphate glucokinase (PPGK) to produce G6P.In this way, one maltose molecule theoretically yields two G6P molecules which can subsequently be used for the enzymatic biomanufacturing of biochemicals or bioelectricity.
To construct the three-enzyme maltose-to-G6P reaction module, a known MP from B. subtilis, a known engineered T. fusca PPGK, and a putative thermostable β-PGM from P. horikoshii OT3 (NCBI reference sequence: WP_ 010884842.1)were cloned, expressed, and purified.β-PGM was able to utilize β-G1P but showed no activity for α-G1P, thus classifying this enzyme as EC 5.4.2.6.β-PGM functioned optimally at around 70 °C in HEPES buffer at pH 7.0 (Figure S1), with a specific activity of 34.7 U/mg.Meanwhile, it also had catalytic activity at 37 °C.Because MP was from a mesophilic source, the in vitro reaction temperature was set as 37 °C.At this temperature and pH 7.0, the specific activities of the purified MP, β-PGM, and PPGK were 7.0, 1.6, and 40.0 U/mg, respectively.Then, the three-enzyme in vitro reaction module was tested for its ability to generate G6P, with each enzyme loaded at 1 U/mL.When 0.5 mM maltose was used as the substrate, a double-enzyme module comprised of MP and β-PGM produced 0.48 mM G6P after 120 min of reaction, while the double-enzyme module comprised of MP and PPGK produced 0.47 mM G6P.In comparison, the threeenzyme module produced 1.0 mM G6P within 120 min (Figure 3), indicating the stoichiometric production of G6P from maltose as we expected.
3.2.Production of Bioelectricity from Maltose via G6P.After confirming that G6P could be stoichiometrically generated from maltose, this three-enzyme reaction module was extended for bioelectricity generation by adding more downstream enzymes.The reason for using bioelectricity generation to demonstrate the stoichiometric utilization of maltose is that the signal of generated electricity can be precisely captured by the instruments even at a low maltose concentration.In this designed biobattery, G6P produced from maltose was oxidized by two cascade enzymes, glucose 6-phosphate dehydrogenase (G6PDH) and 6-phosphogluconate dehydrogenase (6PGDH), generating CO 2 and ribulose 5-phosphate (Ru5P).Accompanying with the oxidation of one G6P molecule, two NAD + molecules were reduced to NADH.Subsequently, each NADH was reoxidized into NAD + by diaphorase (DI) and produced two electrons which could be transferred to the anode via immobilized vitamin K 3 (VK 3 ).In this case, the oxidation of one maltose molecule could stoichiometrically produce 8 electrons by the 6-enzyme biobattery (Figure S2).Another 8 enzymes, which were ribose 5-phosphate isomerase (RPI), ribulose 5-phosphate 3-epimerase (RPE), transketolase (TK), transaldolase (TAL), triose-phosphate isomerase (TIM), fructose-bisphosphate aldolase (ALD), fructose bisphosphatase (FBP), and phosphoglucose isomerase (PGI), constituted a C5-to-C6 recycling module for the conversion of every Ru5P into 5/6 molecule of G6P (Figure 4).Theoretically, this 14-enzyme biobattery could generate 48 electrons from the complete oxidation of each maltose molecule.
All enzymes used in this biobattery were purified and analyzed by SDS-PAGE for their purities (Figure S3).We first tested the performance of the 6-enzyme biobattery via a proof-of-concept trial using 1 U/mL of MP, β-PGM, PPGK, and 5 U/mL of G6PDH, 6PGDH, and DI.After adding 0.5 mM maltose to initiate the reaction, a sharp 4 BioDesign Research increase of the current was observed (Figure S4).The current reached its peak of 0.09 mA at around 1.5 h and then declined over time.Within a reaction time of 20 h, the cumulative electric charges generated from 10 mL of the reaction mixture were 1.63 C, whereas the theoretical electric charges generated from 0.5 mM maltose would be 3.86 C. In this case, the Faraday efficiency was only 42.2%.Enhancing the loading concentrations of MP, β-PGM, and PPGK from 1 U/mL to 3 U/mL resulted in an enhancement of cumulative electric charges generated within 20 h to 3.85 C, corresponding to a near-theoretical Faraday efficiency of 99.7% (Figure 5(a), black curve).The maximum current density and power density of this 6-enzyme biobattery were 3.6 mA/cm 2 and 0.5 mW/cm 2 , respectively (Figure 5(b), black curve).We then added the C5-to-C6 recycling module to construct a 14-enzyme biobattery.Within the reaction period of 65 h, the 14-enzyme biobattery produced 22.32 C of electricity, corresponding to a near-theoretical Faraday efficiency of 96.4% (Figure 5(a), red curve).
Compared with the 6-enzyme system, the 14-enzyme system displayed enhancements of both maximum current density and power density, which were 4.5 mA/cm 2 and 0.6 mW/cm 2 , respectively (Figure 5(b), red curve).These results suggested that the addition of the C5-to-C6 recycling module facilitated the complete oxidation of maltose and enhanced the discharge rate of the biobattery.
In addition, the maximal power density of the 14-enzyme biobattery at 37 °C was comparable with the performance of the maltodextrin-based biobattery (0.35 mW/cm 2 at 23 °C and 0.8 mW/cm 2 at 50 °C) [38] and was higher than the power density of another enzymatic biobattery using xylose as input (0.36 mW/cm 2 at 37 °C) [11].
3.3.Production of FDP from Maltose via G6P.After confirming that the three-enzyme in vitro reaction module could achieve stoichiometric utilization of maltose, another in vitro biosystem was designed for the biomanufacturing of chemicals using relatively higher concentrations of maltose.Fructose 1,6-diphosphate (FDP) was chosen as the product for the following two reasons: (1) FDP plays a cytoprotective role in many pathological situations, making it a useful therapeutic agent for ischaemic injury [39] and a potential treatment agent in scenarios such as convulsions [40], UV-induced oxidative skin damage [41], and diabetic testicular complication [42]; (2) FDP is a high-energy chemical with two phosphate groups per module; therefore, the accumulation of phosphate ions in the in vitro biosystem for FDP production will be less severe than that in a similar system for phosphate-free products such as inositol.In this study, based on the three-enzyme maltose-to-G6P module, we designed an in vitro synthetic biosystem for the ATPfree synthesis of FDP from maltose.In addition to MP, β-PGM, and PPGK, this system also contained PGI which isomerized G6P into fructose 6-phosphate (F6P) and pyrophosphate-dependent phosphofructokinase (PP i -PFK) which phosphorylated F6P into FDP in the presence of pyrophosphate (PP i ) (Figure 6).Half of the orthophosphate ions released by PP i -PFK could be reused by MP.Theoretically, each maltose molecule would be capable of the synthesis of two FDP molecules.
the FDP concentration was 21.73 mM, corresponding to 78.2% of the theoretical product yield and suggesting the production of 1.6 FDP molecules from each maltose.To improve the product yield, we first investigated the effect of pH on the reaction system and found that no further pH optimization was necessary because the system produced the highest amount of FDP at pH 7.0 (Figure S8).Next, we optimized the loading amounts of the five enzymes by systematic titration.The optimal concentrations of MP, β-PGM, PPGK, PGI, and PP i -PFK were 1, 4, 2, 1, and 2 U/mL, respectively (Figure S9).At the optimized enzyme concentrations, one-pot production of FDP from maltose by the in vitro biosystem containing MP, β-PGM, PPGK, PGI, and PP i -PFK was carried out again (Figure 7).The reaction reached equilibrium at 4 h when 0.64 mM maltose, 2.02 mM G6P, and 24.65 mM FDP were detected.It was observed that optimization of enzyme loading amounts facilitated the increase of FDP yield from 78.2% to 88.7% of the theoretical value.The consumption of each maltose molecule yielded around 1.9 FDP molecules, indicating the near-stoichiometric synthesis of FDP.
To evaluate the industrialization potential of this system for FDP production, it is necessary to test a higher substrate concentration such as 50 g/L.We adopted a step-wise strategy for the addition of substrates which was necessary due to the limited solubility of pyrophosphate at 37 °C.Besides, fed-batch addition of pyrophosphate was specifically required because high concentrations of inorganic phosphate could precipitate magnesium ions [14], which would in turn affect the activities of magnesium-dependent enzymes such as β-PGM, PPGK, PGI, and PP i -PFK.Meanwhile, maltose was added in a fed-batch manner to avoid the accumulation of glucose in the system which might inhibit the activity of MP.Considering that a portion of pyrophosphate might chelate with magnesium ions, pyrophosphate and maltose were added at a constant molar ratio of 4 : 1 which was higher than their stoichiometric molar ratio of 2 : 1.A high concentration of MgCl 2 of 50 mM was also used to ensure the presence of sufficient magnesium ions in the reaction system.The loading amount of each enzyme was five times as the previously optimized concentration for 5 g/L maltose.This was a compromise strategy, because scaling up of enzyme loadings by tenfold would result in a total enzyme concentration of 47 g/L which might not only increase the viscosity of the reaction solution but also drastically raise the cost of enzymes.The reaction was carried out at 37 °C, pH 7.0.At 0 h, 20 g/L (55.6 mM) maltose and 224 mM sodium pyrophosphate were added to initiate the reaction.After 6 h when the FDP titer reached approximately 86.5 mM, 15 g/L (41.7 mM) maltose and 168 mM pyrophosphate were supplemented to the system (Figure 8.The FDP titer kept increasing within the next 3 h to 137.0 mM.Then, another 15 g/L (41.7 mM) maltose and 168 mM pyrophosphate were added.At 14 h, the final titer of FDP was 187.0 mM, corresponding to an overall product yield of 67.3% of the theoretical value.Our system outperformed a previously reported in vitro synthetic biosystem which produced 125 mM FDP from 50 g/L maltodextrins [14].

Discussion
In this study, we chose maltose, the economic and most widely distributed α-linked glucose disaccharides in nature, as a Figure 7: One-pot synthesis of FDP from maltose under optimized conditions.The reaction was performed at 37 °C in 100 mM HEPES buffer (pH 7.0) containing 5 g/L (13.9 mM) maltose, 1 U/mL MP, 4 U/mL β-PGM, 2 U/mL PPGK, 1 U/mL PGI, 2 U/mL PP i -PFK, 5 mM MgCl 2 , 100 mg/L ampicillin, 50 mg/L kanamycin, 20 mM potassium phosphate, 60 mM sodium pyrophosphate, and 8 mM sodium hexametaphosphate.7 BioDesign Research substrate for in vitro biomanufacturing.A three-enzyme in vitro synthetic reaction module was constructed for the ATP-free, stoichiometric conversion of each maltose molecule to two G6P molecules.Then, downstream enzymes were added to this module to construct a 14-enzyme in vitro biobattery as well as a 5-enzyme FDP-producing in vitro biosystem.After system optimization, both the biobattery and the FDP-producing biosystem demonstrated near-theoretical performances.The 14-enzyme biobattery resulted in a Faraday efficiency of 96.4% and a maximal power density of 0.6 mW/cm 2 at 37 °C.This power density was comparable with the performance of a previously reported maltodextrin-based biobattery [38] and was higher than the power density of another enzymatic biobattery using xylose as input [11].The 5-enzyme biosystem for FDP production also achieved a near-stoichiometric yield by producing 1.9 FDP molecules per maltose when 5 g/L of maltose was used as input.These results suggested the potential of the three-enzyme reaction module for the stoichiometric manufacturing of more value-added products from maltose via G6P, including biohydrogen, D-allulose, inositol, and glucosamine.
In common with several previously designed in vitro synthetic enzymatic biosystems using starch, cellulose, cellobiose, or sucrose as the substrate, the maltose-powered biosystems in this study start with the orthophosphatedependent phosphorylation of substrate into G1P, followed by the isomerization of G1P to G6P.To construct a maltose-powered in vitro system, however, it is insufficient to merely replace the sugar substrate and the corresponding sugar phosphorylase of the established in vitro systems with maltose and MP.This is because the phosphorylation of starch, cellulose, cellobiose, and sucrose yielded α-G1P whereas the phosphorylation of maltose produced β-G1P.Unlike G6P and glucose which can undergo mutarotation to an equilibrium mixture of the α and β configurations due to their free anomeric carbons (C1), G1P has a phosphate group at its C1 position and hence is not able to spontaneously interconvert between its αand βforms.As a result, in addition to the utilization of MP, the α-PGMs (EC 5.4.2.2) which are specific for α-G1P in the previous in vitro systems should be replaced with a β-PGM (EC 5.4.2.6) for the conversion of maltose to G6P.Meanwhile, maltose and MP can be used for the in vitro enzymatic production of rare disaccharides such as kojibiose and nigerose which can be synthesized from β-G1P and glucose by the corresponding phosphorylases [43].
Our design of the biobattery was inspired by a previous study using maltodextrin for stoichiometric bioelectricity generation via G6P [38].Therefore, we adopted the previously reported loading concentrations of the downstream enzymes (5 U/mL for G6PDH, 6PGDH, and DI) for our proof-of-concept test.However, a low Faraday efficiency of only 42.2% was obtained.Considering that the concentrations of upstream G6P-producing enzymes in the maltodextrin-based biobattery were also 5 U/mL, while MP, β-PGM, and PPGK in our system were loaded at only 1 U/mL, it was suspected that our low Faraday efficiency might be due to the low G6P-producing rate which did not match with the downstream bioelectricity generation process.Indeed, a simple system optimization by enhancing the loading concentration of each enzyme to 3 U/mL resulted in a near-theoretical Faraday efficiency.A main issue for further improvement of this biobattery is to prolong its lifetime.This may be achieved by methods such as enzyme immobilization [44] or the addition of BSA and Triton X-100 to the enzyme solution [38].
For the FDP-producing in vitro biosystem, optimization of enzyme concentrations resulted in an improvement of product yield from 78.2% to 88.7% of the theoretical value.The decrease of MP concentration, together with the increase of β-PGM concentration, might have mitigated the possible substrate inhibition of β-PGM, which was reported on the E. coli homologue [45].Nevertheless, when increasing the input of maltose from 5 g/L to 50 g/L, the overall product yield dropped to 67.3% of the theoretical value despite a fed-batch strategy was adopted.Thermodynamic analysis using the eQuilibrator calculator [42] showed that the Gibbs free energy changes (ΔG) of reactions catalyzed by MP, β-PGM, PPGK, PGI, and PP i -PFK were 3.0, -7.4,-6.4,2.5, and -0.6 kJ/mol, respectively, indicating the reaction catalyzed by PP i -PFK was reversible.Therefore, the accumulation of FDP at high concentrations might have caused PP i -PFK to cease to produce FDP.Another shortcoming of this FDP-producing system is the accumulation of inorganic phosphate during the reaction, as illustrated in Figure 6.This might inhibit the activities of magnesiumdependent enzymes through the chelation of phosphate with magnesium.
Several works can be done in the future to improve our current in vitro systems: (1) The MP used in this study was from B. subtilis and was mesophilic.Low thermostability of the enzymes in in vitro biosystems not only increases the production cost of enzymes but may also affect the longterm stability of the reaction system.The thermostability of MP thus needs to be improved, which can be achieved through protein engineering and the discovery of new thermostable MPs; (2) β-PGM from P. horikoshii has a low specific activity (1.6 U/mg) under the in vitro reaction conditions in this study, and hence, a relatively large amount of this enzyme was required for the optimal performance of the in vitro synthetic enzymatic biosystems.One solution to this issue is to increase the activity of β-PGM by protein engineering and gene mining.Another solution is to modify MP or to design novel enzymes to produce α-G1P through the phosphorylation of maltose, so that α-PGMs with high activities (such as the α-PGM from Clostridium thermocellum which is of around 885 U/mg at 37 °C [46]) can be used instead of β-PGM.A careful mechanistic investigation on MP would benefit the latter idea.(3) The main drawback of the biosystems in this study is the accumulation of inorganic phosphates.To solve this problem, metal ions can be added during the reaction to chelate with the accumulated inorganic phosphates in the system.Another solution to mitigate the accumulation of phosphate is to replace the polyphosphate-dependent PPGK in our systems with some microorganisms to consume glucose for the production of single-cell proteins in parallel with the in vitro synthesis of biochemicals.( 4) The present reaction pathway can be further extended for the high-yield synthesis of other downstream biochemicals such as 2-deoxy-ribose, allulose, tagatose, and ribulose [47].
In conclusion, we successfully designed and constructed an in vitro synthetic enzymatic reaction module which stoichiometrically produced two G6P molecules per maltose.Based on this reaction module, we constructed an enzymatic biobattery and an FDP-producing in vitro biosystem.The biobattery achieved a near-theoretical Faraday efficiency of 96.4% and a maximal power density of 0.6 mW/cm 2 .The FDP-producing in vitro biosystem generated 187.0 mM FDP from 50 g/L (139 mM) maltose through step-wise feeding of substrates.Our results provide a promising strategy for the complete utilization of maltose in both enzymatic fuel cells and in vitro biosynthesis of value-added chemicals in the future.

Figure 1 :
Figure 1: Schematic workflow of the experimental design.

[Figure 2 :Figure 3 :
Figure 2: Schematic representation of the in vitro synthetic enzymatic reactions for the stoichiometric utilization of maltose.The threeenzyme module for the conversion of maltose to G6P was marked by a rectangular frame.P i : orthophosphate; [P i ] n : polyphosphate.Enzymes and the other compounds are abbreviated as described in the text.

Figure 5 :
Figure5: Bioelectricity generation from the oxidation of maltose at near-theoretical efficiency: (a) profiles of the current produced versus time; (b) profiles of power density versus current density.3 U/mL MP, 3 U/mL β-PGM, 3 U/mL PPGK, 5 U/mL G6PDH, 5 U/mL 6PGDH, and 5 U/mL DI were used.For the 14-enzyme biobattery, the reaction solution also contained a C5-to-C6 recycling module comprised of RPI, RPE, TK, TAL, TIM, ALD, FBP, and PGI.Each of these C5-to-C6 enzymes was loaded at 1 U/mL.

Figure 6 :
Figure 6: Pathway for the ATP-free synthesis of FDP from maltose by an in vitro synthetic biosystem.Enzymes and compounds are abbreviated as described in the text.
Figure8: Fed-batch reaction for the synthesis of FDP from high concentrations of substrates.The reaction was carried out at 37 °C in 100 mM HEPES buffer (pH 7.0) containing 50 mM MgCl 2 , 100 mg/L ampicillin, 50 mg/L kanamycin, 20 mM potassium phosphate, 80 mM sodium hexametaphosphate, 5 U/mL MP, 20 U/mL β-PGM, 10 U/mL PPGK, 5 U/mL PGI, and 10 U/mL PP i -PFK.At 0 h, a final concentration of 20 g/L (55.6 mM) maltose was added together with 224 mM sodium pyrophosphate to initiate the reaction.Red arrows indicated the time when another 15 g/L (41.7 mM) maltose and 168 mM sodium pyrophosphate were supplemented to the system.