Inorganic phosphate self-sufficient whole-cell biocatalysts containing two co-expressed phosphorylases facilitate cellobiose production

Abstract Cellobiose, a natural disaccharide, attracts extensive attention as a potential functional food/feed additive. In this study, we present an inorganic phosphate (Pi) self-sufficient biotransformation system to produce cellobiose by co-expressing sucrose phosphorylase (SP) and cellobiose phosphorylase (CBP). The Bifidobacterium adolescentis SP (BASP) and Cellvibrio gilvus CBP (CGCBP) were co-expressed in Escherichia coli. Escherichia coli cells containing BASP and CGCBP were used as whole-cell catalysts to convert sucrose and glucose to cellobiose. The effects of reaction pH, temperature, Pi concentration, and substrate concentration were investigated. In the optimum biotransformation conditions, 800 mM cellobiose was produced from 1.0 M sucrose, 1.0 M glucose, and 50 mM Pi, within 12 hr. The by-product fructose and residual substrate (sucrose and glucose) were efficiently removed by treatment with yeast, to help purify the product cellobiose. The wider applicability of this Pi self-sufficiency strategy was demonstrated in the production of laminaribiose by co-expressing SP and laminaribiose phosphorylase. This study suggests that the Pi self-sufficiency strategy through co-expressing two phosphorylases has the advantage of great flexibility for enhanced production of cellobiose (or laminaribiose).


Introduction
Oligosaccharides are important for a wide range of applications in the food, medicine, nutrition, and cosmetic industries. One interesting example of disaccharides with physiological functions or potential prebiotic effects is cellobiose ( van Zanten et al., 2015 ) . Cellobiose, a zero-calorie functional sweetener, attracts considerable interest because of its biological functions ( Schwaiger et al., 2020 ) . As a dietary fiber, cellobiose is a biologically functional feed additive that is well tolerated by animals ( Nakamura et al., 2004 ) . For example, the observed increase of Clostridium in the feces of horses may indicate beneficial and potentially prebiotic effects of cellobiose ( Paßlack et al., 2020 ) . The effect of cellobiose supplementation on the performance, health status, and digestive traits of growing rabbits has been evaluated ( Ocasio-Vega et al., 2019 ;Zhong et al., 2019 ) . Moreover, Bifidobacterium infantis remains viable during 4 weeks of storage in milk supplemented with cellobiose, suggesting the usefulness of cellobiose as a prebiotic ingredient in fermented products involving bifidobacteria ( Basholli-Salihu et al., 2013 ) . When cellobiose is ingested by humans, there is no increase in blood glucose or insulin secretion, and orally ingested cellobiose is well fermented in the large intestine ( Ruiz-Matute et al., 2011 ) . Consequently, cellobiose is an emerging food and feed ingredient, large-scale production of which by biocatalytic methods has received considerable attention in recent years ( Brucher & Häßler, 2019 ;Ubiparip et al., 2020 ) .
Cellobiose is predominately obtained on an industrial scale by acid hydrolysis of cellulose, the abundant organic compound in nature  . However, it is difficult to isolate cellobiose effectively because of the by-products cellooligosaccharides and glucose ( Homma et al., 1993 ) . Cellulases cleave the second 1,4-linkages from reducing or nonreducing ends of cellulose to release cellobiose, but the activities of these enzymes are not sufficient for the industrial production of cellobiose ( Ubiparip et al., 2020 ) . Cellobiose has been enzymatically synthesized from starch using glucan phosphorylase and cellobiose phosphorylase ( CBP; EC 2.4.1.20 ) , but the process involves three steps and has a yield of just 23.7% ( Suzuki et al., 2009 ) . Furthermore, the production of cellobiose from sucrose using three enzymes, sucrose phosphorylase ( SP; EC 2.4.1.7 ) , xylose isomerase, and CBP, has been reported, but only 62.3 mM cellobiose was obtained from 100 mM sucrose ( Zhong et al., 2017 ) . Previous studies have indicated that the key precursor for cellobiose production is α-glucose-1-phosphate ( G-1-P ) . G-1-P can be produced from two primary carbohydrates, sucrose and starch, by phosphorylases ( Winter et al., 2011 ;Zhong et al., 2017 ) . SP is a key enzyme that catalyzes the phosphorolysis of sucrose. In the presence of inorganic phosphate ( Pi ) , SP can convert sucrose to produce G-1-P and the by-product fructose ( Goedl et al., 2007 ;Kitaoka et al., 1992a ) . CBP is classified as disaccharide phosphorylase, which catalyzes reversible phosphorolysis of the corresponding disaccharides into G-1-P and glucose ( Kitaoka et al., 2012 ;Yoshida et al., 1998 ) . CBP belongs to the CAZy ( a database of carbohydrate-active enzymes ) family GH94, together with chitobiose phosphorylase and cellodextrin phosphorylase. The natural role of CBP is the energyefficient metabolism of cellobiose, producing G-1-P in which much Fig. 1. Inorganic phosphate ( Pi ) self-sufficient biotransformation system to produce cellobiose from sucrose and glucose by co-expressing sucrose phosphorylase and cellobiose phosphorylase. The by-product ( fructose ) and residual substrate ( sucrose and glucose ) were removed by yeast fermentation.  ( Cagnin et al., 2019 ;de Groeve et al., 2011 ;Yuan & Zhao, 2013 ) .
In the presence of Pi, the cascade reaction of SP and CBP converts sucrose and glucose into cellobiose through a twostep glucosyl transfer via G-1-P ( Brucher & Häßler, 2019 ;Zhong et al., 2017 ) . Therefore, the primary factors influencing cellobiose biosynthesis are the high activity of SP and CBP, and the concentration of Pi. However, a high Pi concentration has an important influence on cellobiose synthesis due to product inhibition and causes phosphorus pollution of the environ-ment ( Ubiparip et al., 2020 ;Zhong et al., 2017 . As such, a Pi self-sufficient biotransformation method to produce cellobiose from sucrose and glucose by co-expressing SP and CBP was developed. The by-product fructose and residual substrate were removed by yeast fermentation ( Fig. 1 ) . We also developed an effective synthesis of laminaribiose ( β-1,3-linked glucobiose ) from sucrose and glucose using this Pi self-sufficient strategy, by co-expressing SP and laminaribiose phosphorylase ( LBP; EC 2.4.1.31 ) . The strategy of Pi self-sufficient biotransformation by co-expressing two phosphorylases is a promising method to

Bacterial Strains, Plasmids, and Chemicals
All strains and plasmids used in this work are listed in Table 1 . Escherichia coli Trans T1 was used for the construction and propagation of the plasmid, while E. coli BW25113 was the host for enzyme expression. The vector pYB1s ( derived from the pBAD/HisB vector ) was used to provide the pBAD promoter and the corresponding rrnB terminator. All genes encoding SP, CBP, and LBP were artificially synthesized by General Biosystems Co., Ltd ( Anhui, China ) . Sucrose, glucose, and fructose were purchased from Shanghai ShengGong Bio-chemical Co. Ltd ( Shanghai, China ) . Standards for G-1-P, cellobiose, laminaribiose, and other chemicals were purchased from Aladdin ( Shanghai, China ) .

Heterologous Expression of Phosphorylases
Engineered E. coli strains containing different phosphorylases were cultured in 5 ml of LB medium with appropriate antibiotics at 37°C for 6-7 hr. One milliliter of seed culture was then transferred to a 500-ml flask containing 100 ml of ZYM autoinduction medium. Cultures were incubated in a rotary shaker at 220 rpm and 37°C until the absorbance at 600 nm ( OD 600 nm ) was between 0.6 and 0.8, at which point phosphorylase expression was induced at 30°C for 12 hr. The phosphorylases used in this study were SP from Bifidobacterium adolescentis ( BASP ) , CBP from Cellvibrio gilvus ( CGCBP ) , and LBP from Paenibacillus sp. YM1 ( PSLBP ) . The

High-level Production of Cellobiose and Laminaribiose by Co-Expressing Two Phosphorylases
Under the optimized reaction conditions, cellobiose was produced from sucrose and glucose by using the recombinant strain CB01, which co-expressed BASP and CGCBP. Cells of strain CB01 were harvested by centrifugation, adjusted to OD 600 nm = 30, and then resuspended in 50 mM phosphate buffer ( pH 6.5 ) containing 2 mM MgCl 2 , 1 M sucrose, and 1 M glucose. The bioconversion was performed in a 1-l bioreactor with 500 ml working volume.
For laminaribiose production, cells of strain LR01, which coexpressed BASP and PSLBP, were harvested by centrifugation at 6,000 × g for 10 min. The conversion experiments were carried out in a 1-l bioreactor with 500 ml working volume. The whole-cell biocatalyst LR01 ( OD 600 nm = 30 ) and 2 mM MgCl 2 were added to the reaction system, the buffer was 50 mM phosphate buffer ( pH 6.5 ) , and the substrates were 1 M sucrose and 1 M glucose.

Analytical Methods
Cell density was estimated by measuring the optical density at 600 nm. The expression of recombinant phosphorylases was analyzed by SDS-PAGE. The concentrations of cellobiose, glucose, G-1-P, Pi, fructose, and glycerol were measured by HPLC ( Agilent 1260 series, Hewlett Packard ) with an Aminex HPX-87H column ( 7.8 × 300 mm ) . The analysis was performed at 50°C with a mobile phase comprising 5 mM H 2 SO 4 at a flow rate of 0.3 ml/min by a refractive index detector ( RID ) . The HPLC chromatograms of cellobiose, glucose, G-1-P, Pi, fructose, and glycerol are shown in Fig. S1.

Construction of Whole-Cell Biocatalysts Co-Expressing BASP and CGCBP
Whole-cell biotransformation using recombinant strains containing a cascade of enzymes is an attractive method to produce many chemicals, and the biocatalytic reactions using whole-cell biocatalysts have the advantages of low cost, better process control, environmental friendliness, and greater efficiency ( de Carvalho, 2017 ; Morgado et al., 2016 ;Xu et al., 2021 ) . As such, BASP and CGCBP were co-expressed under the control of the promoter pBAD in E. coli using the plasmid pYB1s. Plasmids pYB1s-BASP-CGCBP and pYB1s-CGCBP-BASP were constructed to co-express the two phosphorylases to determine the effects of the order of the genes ( Fig. 3 a ) . The plasmids pYB1s-BASP-CGCBP and pYB1s-CGCBP-BASP were respectively transformed into E. coli to obtain whole-cell biocatalysts named CB01 and CB02. The expression of BASP and CGCBP was analyzed by SDS-PAGE ( Fig. 3 b ) . To determine the catalytic efficiency of the two whole-cell biocatalysts, cells of CB01 and CB02 were grown, harvested by centrifugation, and adjusted to OD 600 nm = 20 in 50 mM phosphate buffer ( pH 6.0 ) . The cellobiose synthesis reaction was performed at 40°C, and the substrate concentrations were 0.5 M sucrose and 0.5 M glucose. CB01 exhibited high performance in producing cellobiose from sucrose and glucose; ∼330 mM cellobiose was obtained from 0.5 M sucrose and 0.5 M glucose with a conversion rate of 65% ( Fig. 3 c ) .
BASP ( from B. adolescentis ) and CGCBP ( from C. gilvus ) were selected for co-expression in E. coli BW25113 for the production of cellobiose. The BASP protein showed a higher molecular weight in SDS-PAGE ( Fig. 3 b, lanes 4-5 ) when the genes encoding BASP and CGCBP were placed in pYB1s in the order BASP-CGCBP ( lanes 4-5 ) than when they were in the order CGCBP-BASP ( lanes 2-3 ) . In the former arrangement, the N -terminus of BASP encoded in pYB1s included a sequence that translates into the extra amino acids MGTSSSGLVPRGSLE. Although BASP and CGCBP co-expressed in one plasmid in either order exhibited high expression levels ( Fig. 3 b ) , strain CB01 with plasmid pYB1s-CGCBP-BASP was more effective in the production of cellobiose ( Fig. 3 c ) . The low production of cellobiose by using strain CB02 might be due to the incomplete folding of enzymes during transcription and translation, suggesting that the order of the two genes might affect the apparent activities of the two enzymes. Hence, the order of the genes should be considered when multiple enzymes are co-expressed from one plasmid. In addition, the CGCBP protein showed a higher expression level in SDS-PAGE ( Fig.  3 b, lanes 2-5 ) when two genes were placed in pYB1s in the order CGCBP-BASP ( strain CB01 ) than when they were in the order BASP-CGCBP ( strain CB02 ) . These results showed that CB01 ( gene order CGCBP-BASP ) exhibited high performance in producing cellobiose ( Fig. 3 c ) . Therefore, CGCBP is a key enzyme for the Pi selfsufficient biotransformation method to produce cellobiose from sucrose and glucose.

Optimization of Reaction Parameters for Cellobiose Production
In this study, the one-pot biotransformation of sucrose and glucose to produce cellobiose involves two reversible reactions: ( 1 ) phosphorylation of sucrose to G-1-P and fructose catalyzed by BASP and ( 2 ) synthesis of cellobiose from G-1-P and glucose catalyzed by CGCBP. Pi plays an important role because it is a substrate of SP for G-1-P generation and a by-product of cellobiose synthesis by CBP. A high Pi concentration promotes the generation of G-1-P by SP, but Pi also decreases the reaction rate of cellobiose synthesis because of product inhibition ( Zhong et al., 2017 . Moreover, suitable catalytic conditions enable wholecell biocatalysts to perform effectively, and pH and temperature are important factors that greatly affect such bioconversions ( Xu et al., 2021 ) . As such, the effects of the substrate molar ratio, Pi concentration, pH, and temperature were investigated to obtain the maximum production of cellobiose.
In this system developed in this work, the supply of Pi was maintained at a nearly constant level, and glucose is needed to start the cellobiose production reaction, and the kinetic requirements of CBP mean that glucose is added at the beginning of the two-step reaction ( Suzuki et al., 2009 ) . We firstly carried out the effects of substrate molar ratio and Pi concentration for cellobiose production. The substrate molar ratio ( sucrose:glucose ) was set to 1:0.4, 1:0.6, 1:0.8, 1:1, 0.8:1, 0.6:1, and 0.4:1. The maximum amount of cellobiose was obtained with a molar ratio of 1:1 ( [sucrose] and [glucose] = 0.5 M ) ( Fig. 4 a ) . This condition was used for subsequent screening of the best Pi concentration, and the reaction pH and temperature. A Pi concentration of 50 mM ( in sodium phosphate buffer ) resulted in the maximum yield of cellobiose ( Fig. 4 b ) . Sodium phosphate buffers ( pH 6.0-8.0 ) were used to test the effects of the reaction pH; as shown in Fig. 4 c, the optimum pH for cellobiose production was 6.5. The reaction was performed at 30-70°C to investigate the effect of temperature. The maximum cellobiose concentration ( 350 mM ) was obtained from 0.5 M sucrose and 0.5 M glucose at 50°C ( Fig. 4 d ) .

Enhanced Cellobiose Production by Adjusting Substrate Concentrations and Biomass of CB01
Prior studies largely focused on enzymatic production of cellobiose by using multienzyme transformations or immobilized bienzymatic reactions ( Brucher & Häßler, 2019 ;Schwaiger et al., 2020 ) . However, these titers and the substrate conversion rates are not approaching the levels required for industrialization. Therefore, it is necessary to enhance cellobiose production to meet industrial requirements.
In this study, we develop a Pi self-sufficient whole-cell catalytic system for cellobiose production from sucrose and glucose. The substrate ( sucrose and glucose ) concentrations were increased from 0.5 to 1.0 M to enhance cellobiose production by the wholecell biocatalyst CB01. The concentration of cellobiose obtained reached 750 mM ( conversion rate 75% ) ( Fig. 5 a ) . Under this condition, the molar ratios of these reactions were all 1:1, in which the concentrations of sucrose and glucose were from 0.5 to 1.0 M, and the appropriate reaction time was 12 hr ( Fig. S2 ) . The optimum biomass of CB01 for cellobiose production was investigated. Keeping the sucrose and glucose concentrations at 1.0 M, the biomass of CB01 was set at OD 600 nm = 10, 20, or 30. Finally, the conversion rate reached 80% with CB01 biomass of OD 600 nm = 30, and the concentration of cellobiose obtained was 800 mM ( Fig. 5 b ) .

Removal of By-product and Residual Substrate by Treatment with Yeast
In the whole-cell biotransformation process, the by-product fructose ( ∼180 g/l ) originated from sucrose, and residual glucose ( ∼40 g/L ) was retained in the reaction system. The by-product fructose and residual substrate ( glucose and sucrose ) could be efficiently removed by treatment with yeast ( Winter et al., 2011 ) . Therefore, the reaction solutions were treated with yeast to remove the by-product fructose and residual substrate to help purify the cellobiose. We screened the ability of different yeast strains to consume fructose. Strain SC01 effectively consumed 100 g/L fructose in 12 hr ( Fig. S4A ) . Next, consumption of cellobiose, sucrose, and glucose by strain SC01 was tested; sucrose ( 100 g/L ) or glucose ( 100 g/l ) was consumed completely within 24 hr, but cellobiose ( 100 g/l ) was not degraded by SC01 ( Fig. S4B ) . As such, the reaction solution from whole-cell biotransformations of glucose and sucrose to produce cellobiose was treated with SC01 to aid with the purification of cellobiose.
After treatment with SC01, the by-product fructose and residual substrate glucose were degraded, there was no residual sucrose, and trace glycerol was observed ( Fig. S5 ) . The yeast cells were removed by centrifugation, and the supernatant was treated using cation exchange resin JK800 and anion exchange resin D392 ( Xi'an Lanxiao Technology New Material Co., Ltd ) to remove ions. Then, the solution was decolored by adding activated charcoal ( 30 g/l ) in an 80:20 water/ethanol solution. The trace glycerol adsorbed onto the activated charcoal. The water/ethanol solutions were pooled and concentrated under a vacuum at 50°C to obtain mother liquor, which was crystallized to obtain cellobiose crystals. As shown in Table S3, the final purity of cellobiose crystals reached 95%.

Production of Laminaribiose from Sucrose and Glucose by Pi Self-Sufficient System Containing Two Phosphorylases
Laminaribiose has attracted widespread attention as a soluble dietary fiber, a powerful germination agent, and a widely used antiseptic ( Kitaoka et al., 1992b ;Müller et al., 2017a ;Sun et al., 2019 ) . Laminaribiose is also valuable because of its promising roles in biotechnology, including as a precursor for building blocks in the pharmaceutical industry, and as an inducer substrate for developing a regulatable gene expression system in C. thermocellum ( Mearls et al., 2015 ) . Conventional laminaribiose production involves chemical hydrolysis of laminarin, but this has disadvantages, including low product yields, high consumption of energy and water, and high separation costs ( Kitaoka et al., 1992b ) . A previous study reported that 31.9 g/l laminaribiose was obtained from sucrose and glucose in bienzymatic batch experiments with reaction-integrated product separation by adsorption, but the low Fig. 7. Summary of whole-cell biotransformation procedures for cellobiose production from sucrose and glucose by a Pi self-sufficient whole-cell biocatalyst CB01 containing two phosphorylases. yield and the cumbersome procedure limit the application of this approach ( Müller et al., 2017b ) .

Conclusion
In conclusion, a Pi self-sufficient whole-cell catalytic system was developed for cellobiose production by co-expressing the phosphorylases SP and CBP in E. coli . The resulting whole-cell biocatalyst CB01 was used to produce cellobiose from sucrose and glucose; after optimization, the titer of cellobiose reached 275 g/l ( an 80% substrate conversion rate ) ( Fig. 7 ) . To demonstrate the wider applicability of this strategy, we also produced laminaribiose using the whole-cell biocatalyst LR01, which co-expressed BASP and PSLBP. A high titer of laminaribiose, up to 260 g/l, was obtained, corresponding to a substrate conversion rate of 75% ( Fig. 6 ) . To the best of our knowledge, these are the highest reported cellobiose and laminaribiose titers from sucrose and glucose. This study demonstrates a promising strategy for meeting industrial requirements for large-scale cellobiose and laminaribiose production. Furthermore, this work could assist the sugar industry to diversify the production of value-added products from sucrose and glucose by using different phosphorylases.