Biocatalysis for Rare Ginsenoside Rh2 Production in High Level with Co-Immobilized UDP-Glycosyltransferase Bs -YjiC Mutant and Sucrose Synthase At SuSy

: Rare ginsenoside Rh2 exhibits diverse pharmacological effects. UDP-glycosyltransferase (UGT) catalyzed glycosylation of protopanaxadiol (PPD) has been of growing interest in recent years. UDP-glycosyltransferase Bs -YjiC coupling sucrose synthase in one-pot reaction was successfully applied to ginsenoside biosynthesis with UDP-glucose regeneration from sucrose and UDP, which formed a green and sustainable approach. In this study, the his-tagged UDP-glycosyltransferase Bs -YjiC mutant M315F and sucrose synthase At SuSy were co-immobilized on heterofunctional supports. The afﬁnity adsorption signiﬁcantly improved the capacity of speciﬁc binding of the two recombinant enzymes, and the dual enzyme covalently cross-linked by the acetaldehyde groups signiﬁcantly promoted the binding stability of the immobilized bienzyme, allowing higher substrate concentration by easing substrate inhibition for the coupled reaction. The dual enzyme amount used for ginsenoside Rh2 biosynthesis is Bs -YjiC-M315F: At SuSy = 18 mU/mL: 25.2 mU/mL, a yield of 79.2% was achieved. The coimmobilized M315F/ At SuSy had good operational stability of repetitive usage for 10 cycles, and the yield of ginsenoside Rh2 was kept between 77.6% and 81.3%. The high titer of the ginsenoside Rh2 cumulatively reached up to 16.6 mM (10.3 g/L) using fed-batch technology, and the ﬁnal yield was 83.2%. This study has established a green and sustainable approach for the production of ginsenoside Rh2 in a high level of titer, which provides promising candidates for natural drug and development.


Co-Immobilization of Dual Enzyme onto Heterofunctional Carriers
Macroporous resin LX1000HG was selected as the support for modification and bienzyme adsorption. Resin LX1000HG modified with two kinds of functional groups was prepared for the dual enzyme co-immobilization. Heterofunctional support with glyoxylmetal-chelate bifunctional groups was inspired for his-tagged enzyme immobilization [26][27][28]. The two-step immobilization was firstly finished as the metal-chelate groups were able to promote the specific adsorption of the proteins, and then glyoxyl groups enabled the adsorbed enzyme to covalently anchor to the resin (Scheme 2). The structural flexibility of the immobilized enzymes was critical for substrate binding and its catalytic function [27]. Glyoxyl-metal-chelate supports involved coordination bonding with Ni 2+ , Cu 2+ , and Co 2+ were respectively applied for individual adsorption of his-tagged M315F and AtSuSy. In the case of the Ni 2+ -charged support, the highest activity expression, recovery, and specific activity were observed by both M315F and AtSuSy (Table 1). Another kind of heterofunctional resin, amino-glyoxyl resin, was also used for M315F and AtSuSy adsorption. No further improved results were obtained (data not shown).

Scheme 2.
Absorption procedures for two-step immobilization of his-tagged M315F and AtSuSy. Scheme 1. One-pot reaction catalyzed by M315F and AtSuSy for biosynthesis of Rh2 coupled with UDPG regeneration.

Co-Immobilization of Dual Enzyme onto Heterofunctional Carriers
Macroporous resin LX1000HG was selected as the support for modification and bienzyme adsorption. Resin LX1000HG modified with two kinds of functional groups was prepared for the dual enzyme co-immobilization. Heterofunctional support with glyoxyl-metal-chelate bifunctional groups was inspired for his-tagged enzyme immobilization [26][27][28]. The two-step immobilization was firstly finished as the metal-chelate groups were able to promote the specific adsorption of the proteins, and then glyoxyl groups enabled the adsorbed enzyme to covalently anchor to the resin (Scheme 2). The structural flexibility of the immobilized enzymes was critical for substrate binding and its catalytic function [27]. Glyoxyl-metal-chelate supports involved coordination bonding with Ni 2+ , Cu 2+ , and Co 2+ were respectively applied for individual adsorption of his-tagged M315F and AtSuSy. In the case of the Ni 2+ -charged support, the highest activity expression, recovery, and specific activity were observed by both M315F and AtSuSy (Table 1). Another kind of heterofunctional resin, amino-glyoxyl resin, was also used for M315F and AtSuSy adsorption. No further improved results were obtained (data not shown).

Co-Immobilization of Dual Enzyme onto Heterofunctional Carriers
Macroporous resin LX1000HG was selected as the support for modification and bienzyme adsorption. Resin LX1000HG modified with two kinds of functional groups was prepared for the dual enzyme co-immobilization. Heterofunctional support with glyoxylmetal-chelate bifunctional groups was inspired for his-tagged enzyme immobilization [26][27][28]. The two-step immobilization was firstly finished as the metal-chelate groups were able to promote the specific adsorption of the proteins, and then glyoxyl groups enabled the adsorbed enzyme to covalently anchor to the resin (Scheme 2). The structural flexibility of the immobilized enzymes was critical for substrate binding and its catalytic function [27]. Glyoxyl-metal-chelate supports involved coordination bonding with Ni 2+ , Cu 2+ , and Co 2+ were respectively applied for individual adsorption of his-tagged M315F and AtSuSy. In the case of the Ni 2+ -charged support, the highest activity expression, recovery, and specific activity were observed by both M315F and AtSuSy (Table 1). Another kind of heterofunctional resin, amino-glyoxyl resin, was also used for M315F and AtSuSy adsorption. No further improved results were obtained (data not shown).

Scheme 2.
Absorption procedures for two-step immobilization of his-tagged M315F and AtSuSy. Scheme 2. Absorption procedures for two-step immobilization of his-tagged M315F and AtSuSy.
For the M315F adsorption by heterofunctional resin, the expressed activity was 718.5 mU/g wet support as the collaborative coordination bonding of the His tag of M315F and IDA to the Ni 2+ . The specific activity of the M315F binding on glyoxyl-Ni 2+ -chelate carriers was 412.7 mU/mg bound protein, and it produced the recovery of 47.9% and enzyme loading of 1.74 mg/g wet carriers ( Table 1). The enzyme activity expression by the specific bound AtSuSy onto glyoxyl-Ni 2+ -chelate resin was 1959.5 mU/g wet support, and the specific activity of the adsorbed AtSuSy was 1660.5 mU/mg bound protein, resulting in the recovery of 39.1% and enzyme loading of 1.18 mg/g wet carriers (Table 1). Moreover, the conditions for enzyme adsorption were also optimized, the temperature of 25 • C, pH 7.0, and adsorption time of 6 h ( Figure S1) was selected for the his-tagged M315F and AtSuSy co-binding onto heterofunctional resin. For the one-pot reaction of PPD glycosylating by coupling M315F to AtSuSy from sucrose and UDP, the portion of the co-immobilized dual enzyme was critical for its catalytic activities. The portion of M315F and AtSuSy after co-immobilization was regulated by the initial addition of two enzymes in different amounts, and the optimized portion of co-immobilized M315F/ AtSuSy onto the resin at the activity ratio of 1:1.4 was selected for effective production of Rh2, which reached 69.3% conversion at high substrate PPD (6 mM) and high regioselectivity (99%) for Rh2 formation ( Figure 1A). Free enzymes M315F-AtSuSy displayed a conversion of 60% at 2 mM PPD in our previous study [21], suggesting the inhibition on the catalytic yield under high PPD concentration, and similar substrate inhibition was also revealed by the UGT73C5-SuSy coupled reaction [14]. The dual enzyme ratio after coimmobilization for the M315F-AtSuSy coupled reaction was much lower than that of 1:2.1 in our previous report [21]. Then, the bound M315F and AtSuSy were further conducted by the covalent cross-linking immobilization between the glyoxyl groups and adsorbed enzymes. The recovery of 50.9% by M315F and 48.3% by AtSuSy was resulted from the above adsorption optimization and covalent cross-linking coimmobilization. After coimmobilization, the co-localized multiple enzymes in close proximity of the active sites to one another with a suitable orientation by first step specifical adsorption may be beneficial to control the covalent cross-linking in the next step, which leads to high activity expression, facilitating the access of substrates between one enzyme and another in an adequate concentration to perform improved reaction [27].  The binding stability of the immobilized enzymes was investigated ( Figure 1). The M315F and AtSuSy directly adsorbed by specifical binding with Ni 2+ -chelate groups were easy to lose the total enzyme activity using buffer wash over twice, which both M315F and AtSuSy retained less than 60% of activities ( Figure 1B,C) by desorption test, suggesting the enzymes through direct affinity adsorption were easy to leach from the resin and 2+ Figure 1. (A) Effect of the ratio of co-immobilized bienzyme M315F/AtSuSy for Rh2 production. The yield of products Rh2 and F12 was recorded. (B) Desorption of immobilized M315F, his-tagged M315F directly specifical binding with Ni 2+ -chelate groups ( ), co-immobilized M315F by affinity adsorption and covalently cross-linking (•). (C) Desorption of co-immobilized AtSuSy, his-tagged M315F directly specifical binding with Ni 2+ -chelate groups ( ), co-immobilized M315F by affinity adsorption and covalently cross-linking ( ). Enzyme desorption was conducted by measuring the enzyme activity in supernatant and suspension with 0.5 M NaCl solution (pH 7.0), the suspension was maintained under very gentle stirring for 60 min.
The binding stability of the immobilized enzymes was investigated ( Figure 1). The M315F and AtSuSy directly adsorbed by specifical binding with Ni 2+ -chelate groups were easy to lose the total enzyme activity using buffer wash over twice, which both M315F and AtSuSy retained less than 60% of activities ( Figure 1B,C) by desorption test, suggesting the enzymes through direct affinity adsorption were easy to leach from the resin and showed the low ability to bear mild conditions. The M315F and AtSuSy onto glyoxyl-Ni 2+ -chelate supports by two-step of co-immobilization retained high activities (nearly 99%) after seven times of desorption ( Figure 1B,C), nearly no enzyme activity was leached. Most importantly, the co-immobilization of M315F and AtSuSy with the two-step ways significantly improved the binding stability of the co-immobilized bi-enzyme, though histagged enzymes specifical binding with Ni 2+ -chelate groups and covalently cross-linking between the glyoxyl groups and the adsorbed proteins [29].

Physicochemical Properties of Heterofunctional Carriers
The SEM images of bifunctional macroporous resin revealed that the surface of supports was smooth and uniform with some large holes on the surface (Figure 2A), and the internal morphology of the resin also showed the shape of various pore volumes ( Figure 2B,C). After enzyme co-immobilization, pores on the surface of resin were coarse and rough as covered with many enzyme particles, meaning that the out surface of carriers had been utilized for the enzyme adsorption ( Figure 2D). Moreover, the internal morphologies of resin showed that it contained various smaller spores as the binding of the enzymes ( Figure 2E,F). The carriers and dual enzymes co-immobilized onto resin were analyzed by FT-IR, the results can be found in Figure S2. The successful dual enzyme immobilization was further confirmed by the measurements of specific surface and void analyzer. Compared with heterofunctional resin, the carriers adsorbed with enzymes showed a decrease in the pore volumes (from 0.47 to 0.37 cm 3 /g) ( Table 2). The pore volume decrease may be attributed to the enzyme occupation in the pore channels, suggesting that enzymes had been immobilized on resin [30], while the specific surface area increased from 41.34 to 58.93 m 2 /g ( Table 2) as the enzymes immobilized onto the support. The average pore diameter of heterofunctional resin was 33.7 nm, which matched the requirement of pore diameter for the UGTs and AtSuSy binding to prevent restrictions of the enzyme access [31]. After immobilization, the average pore diameter of supports dropped to 27.6 nm, suggesting resin was successfully binding with enzymes.  The carriers and dual enzymes co-immobilized onto resin were analyzed by FT-IR, the results can be found in Figure S2. The successful dual enzyme immobilization was further confirmed by the measurements of specific surface and void analyzer. Compared with heterofunctional resin, the carriers adsorbed with enzymes showed a decrease in the pore volumes (from 0.47 to 0.37 cm 3 /g) ( Table 2). The pore volume decrease may be attributed to the enzyme occupation in the pore channels, suggesting that enzymes had been immobilized on resin [30], while the specific surface area increased from 41.34 to 58.93 m 2 /g ( Table 2) as the enzymes immobilized onto the support. The average pore diameter of heterofunctional resin was 33.7 nm, which matched the requirement of pore diameter for the UGTs and AtSuSy binding to prevent restrictions of the enzyme access [31]. After immobilization, the average pore diameter of supports dropped to 27.6 nm, suggesting resin was successfully binding with enzymes.

Effect of pH and Temperature on Activity and Stability of Enzymes
Both free M315F and co-immobilized M315F had the highest activity at pH 8.0 ( Figure S3A). Free sucrose synthase AtSuSy had the highest activity at pH 7.0, and coimmobilized AtSuSy showed the highest activity at pH 6.0. Both free AtSuSy and coimmobilized AtSuSy were most active under the pH range from 6.0 to 8.0, which showed 90% activity ( Figure S3B). After co-immobilization, both enzyme M315F and AtSuSy exhibited pH stability improvement ( Figure S4). The immobilized AtSuSy was stable under the pH range from 6.0 to 9.0, retaining 90% initial activities. The immobilized AtSuSy still kept nearly 80% of initial activity at pH 10.0 for 12 h incubation. The immobilized M315F displayed improved pH stability as it maintained 90% initial activities (12 h) at a pH range from 6.0 to 9.0 ( Figure S4).
Both free and immobilized AtSuSy had the highest activity at 40 • C and were showed high activities at a temperature range from 35 to 40 • C ( Figure S5). Free enzyme M315F had the highest activity at 35 • C ( Figure S5), while the immobilized M315F were most active at 40 • C, and also showed high activity at a temperature range from 35 to 45 • C. The immobilized enzyme M315F showed the improvement of thermal stability. The immobilized enzyme M315F possessed good thermal stability below 35 • C as it retained nearly 100% initial activity. However, free M315F just retained <70% initial activity when incubation at 35 • C for 12 h ( Figure 3A). There is a certain improvement of thermal stability by the immobilized enzyme of AtSuSy. The co-immobilized AtSuSy was stable when incubated under the temperature below 35 • C for 12 h, and it retained nearly 100% initial activity, while free AtSuSy just showed about 80% activity at 35 • C ( Figure 3B). Such improved properties of co-immobilized AtSuSy and M315F were beneficial to the cascade reaction for the ginsenoside Rh2 production. had the highest activity at 35 °C ( Figure S5), while the immobilized M315F were most active at 40 °C, and also showed high activity at a temperature range from 35 to 45 °C. The immobilized enzyme M315F showed the improvement of thermal stability. The immobilized enzyme M315F possessed good thermal stability below 35 °C as it retained nearly 100% initial activity. However, free M315F just retained <70% initial activity when incubation at 35 °C for 12 h ( Figure 3A). There is a certain improvement of thermal stability by the immobilized enzyme of AtSuSy. The co-immobilized AtSuSy was stable when incubated under the temperature below 35 °C for 12 h, and it retained nearly 100% initial activity, while free AtSuSy just showed about 80% activity at 35 °C ( Figure 3B). Such improved properties of co-immobilized AtSuSy and M315F were beneficial to the cascade reaction for the ginsenoside Rh2 production.

Conditions for the Co-Immobilized Enzymes Catalyzed Reactions
AtSuSy could biosynthesize UDPG using sucrose as the expedient sugar donor [13,14,19,21], and the UGTs-AtSuSy coupled one-pot reaction are a useful protocol for glycosides production with continuous UDPG regeneration [14,19]. In our previous report, Bs-YjiC mutant M315F-AtSuSy coupled one-pot reaction was applied for UDPG regeneration from UDP and cheap sucrose, and it was demonstrated to produce ginsenoside Rh2 up to 3.7 g/L using PPD as substrate [21]. The reaction conditions by the co-immobilized

Conditions for the Co-Immobilized Enzymes Catalyzed Reactions
AtSuSy could biosynthesize UDPG using sucrose as the expedient sugar donor [13,14,19,21], and the UGTs-AtSuSy coupled one-pot reaction are a useful protocol for glycosides production with continuous UDPG regeneration [14,19]. In our previous report, Bs-YjiC mutant M315F-AtSuSy coupled one-pot reaction was applied for UDPG regeneration from UDP and cheap sucrose, and it was demonstrated to produce ginsenoside Rh2 up to 3.7 g/L using PPD as substrate [21]. The reaction conditions by the co-immobilized enzymes for the biosynthesis of ginsenoside Rh2, such as pH 8.0, temperature 35 • C, and 0.4 M sucrose and 5 mM Mg 2+ were generally the same as in our previous report [21].
For the co-immobilized dual enzyme M315F and AtSuSy catalyzed cascade reaction, the other conditions were further optimized to reach a higher level of Rh2 production. Our previous study disclosed that the conversion by free M315F-AtSuSy coupled reaction was about 60% using 2 mM substrate PPD [21]. The concentration of PPD showed a significant impact on the conversion for ginsenoside Rh2 production by the co-immobilized enzymes ( Figure 4A). Using 2 mM PPD as substrate, it showed a higher conversion (~99%) by coimmobilized M315F and AtSuSy than that by free enzymes, indicating the co-immobilized bienzyme could improve the conversion rate under higher PPD concentration and allowing higher substrate concentration by easing substrate inhibition for the coupled reaction [24,27]. The co-immobilized enzymes catalyzed reaction showed the decreased conversion from 99% to 50% as the addition of PPD concentration increased from 2 to 10 mM, and the co-immobilized enzymes displayed higher activity than that of free enzymes at high PPD substrate concentration [22,23]. Interestingly, the titer of produced ginsenoside Rh2 was gradually raised as the PPD concentration increased. As a result, about 5.05 mM ginsenoside Rh2 (conversion 63.1%) was obtained using an initial concentration of 8 mM PPD ( Figure 4A), the titer of ginsenoside Rh2 by the co-immobilized dual enzyme was higher than that by free enzymes. A higher concentration of PPD (above 10 mM) may not be beneficial for the reaction. Thus, an initial PPD concentration of 8 mM was selected for the following optimization of reaction conditions by co-immobilized enzymes. mM PPD ( Figure 4A), the titer of ginsenoside Rh2 by the co-immobilized dual enzyme was higher than that by free enzymes. A higher concentration of PPD (above 10 mM) may not be beneficial for the reaction. Thus, an initial PPD concentration of 8 mM was selected for the following optimization of reaction conditions by co-immobilized enzymes. The addition of hydrophilic co-solvent DMSO could show the effect of substrate PPD solubility increasing in the reaction media, resulting in a higher PPD initial concentration [14]. The conversion (above 60%) by co-immobilized enzymes in the hydrophilic media contained 5%-20% (v/v) DMSO was higher than that in the buffer system (yield 22.4%) ( Figure 4B). The conversion dramatically increased as the reaction media with the addition of 5% (v/v) DMSO and 10% (v/v) DMSO. The highest conversion was 72% using 10% (v/v) DMSO as the reaction media, which was about three times the yield by using buffer media. The yield slightly decreased using 15% (v/v) DMSO (yield 68.1%) and 20% (v/v) DMSO (yield 63.7%) as the reaction media, suggesting the drawback of activity inhibition by higher a content of DMSO addition. There was no significant influence on the yield of ginsenoside Rh2 by increasing the addition of UDP, only a catalytic amount of UDP (0.5 mM) was necessary for the co-immobilized M315F-AtSuSy coupled reaction ( Figure S6). The conversion was greatly improved with the increased addition of co-immobilized biocatalysts. Using the co-immobilized enzymes at Bs-YjiC-M315F: AtSuSy = 18 mU/mL: 25.2 mU/mL, a high yield of 79.2% was achieved ( Figure 4C). The biocatalyst usage for ginsenoside Rh2 production was much lower than that reported in other studies [14,19].

Reusability of Co-Immobilized Enzymes and Fed-Batch Strategy for Ginsenoside Rh2
The co-immobilized glycosyltransferase Bs-YjiC mutant M315F and sucrose synthase AtSuSy exhibited good operational stability after repetitive 10 cycles of dual enzyme coupled reaction for ginsenoside Rh2 production. The co-immobilized sucrose synthase AtSuSy could maintain high catalytic activities of UDPG regeneration using sucrose and UDP, and co-immobilized M315F equally kept high activity of ginsenoside Rh2 synthesis, yielding a total conversion between 77.6% and 81.3% ( Figure 5A) during the 10 consecutive cycles usage. This implied the good operation stability of the M315F-AtSuSy co-immobilized onto heterofunctional resin. The addition of hydrophilic co-solvent DMSO could show the effect of substrate PPD solubility increasing in the reaction media, resulting in a higher PPD initial concentration [14]. The conversion (above 60%) by co-immobilized enzymes in the hydrophilic media contained 5%-20% (v/v) DMSO was higher than that in the buffer system (yield 22.4%) ( Figure 4B). The conversion dramatically increased as the reaction media with the addition of 5% (v/v) DMSO and 10% (v/v) DMSO. The highest conversion was 72% using 10% (v/v) DMSO as the reaction media, which was about three times the yield by using buffer media. The yield slightly decreased using 15% (v/v) DMSO (yield 68.1%) and 20% (v/v) DMSO (yield 63.7%) as the reaction media, suggesting the drawback of activity inhibition by higher a content of DMSO addition. There was no significant influence on the yield of ginsenoside Rh2 by increasing the addition of UDP, only a catalytic amount of UDP (0.5 mM) was necessary for the co-immobilized M315F-AtSuSy coupled reaction ( Figure S6). The conversion was greatly improved with the increased addition of coimmobilized biocatalysts. Using the co-immobilized enzymes at Bs-YjiC-M315F: AtSuSy = 18 mU/mL: 25.2 mU/mL, a high yield of 79.2% was achieved ( Figure 4C). The biocatalyst usage for ginsenoside Rh2 production was much lower than that reported in other studies [14,19].

Reusability of Co-Immobilized Enzymes and Fed-Batch Strategy for Ginsenoside Rh2
The co-immobilized glycosyltransferase Bs-YjiC mutant M315F and sucrose synthase AtSuSy exhibited good operational stability after repetitive 10 cycles of dual enzyme coupled reaction for ginsenoside Rh2 production. The co-immobilized sucrose synthase AtSuSy could maintain high catalytic activities of UDPG regeneration using sucrose and UDP, and co-immobilized M315F equally kept high activity of ginsenoside Rh2 synthesis, yielding a total conversion between 77.6% and 81.3% ( Figure 5A) during the 10 consecutive cycles usage. This implied the good operation stability of the M315F-AtSuSy co-immobilized onto heterofunctional resin. To further alleviate the possible inhibition by high substrate concentration, a fedbatch strategy by stepwise addition of 2.0 mM fresh PPD was adopted to the cascade reaction with the co-immobilized M315F-AtSuSy. The initial concentration of PPD (8 mM) was used, a titer of the ginsenoside Rh2 cumulatively reached up to 16.6 mM (10.3 g/L) ( Figure 5B) by using fed-batch technology and the final yield was about 83.2%. To our acknowledgment, this titer of ginsenoside Rh2 produced by the present biotechnological way was the highest level, and it was 2.78-3.2 times of free enzymes [21], and much higher than other reports [14]. Using co-immobilized M315F-AtSuSy coupled reaction with fedbatch protocol, the overall STY of ginsenoside Rh2 was about 0.396 g/L/h, which was 20fold higher than that of the report [13]. The co-immobilized M315F-AtSuSy catalyzed the transglycosylating reaction afforded a new way for efficient manufacturing of rare ginsenoside Rh2, and formed a green, cost-effective and sustainable approach with UDP-glucose regeneration from sucrose and catalytic amount of UDP.

Materials and Chemicals
The C-terminal 6*his-tagged glycosyltransferase Bs-YjiC (from Bacillus subtilis 168) mutant M315F and C-terminal 6*his-tagged sucrose synthase AtSuSy (from Arabidopsis thaliana) were obtained and prepared respectively as in our previous report [21]. Macroporous resin LX1000HG (contained hydroxyl group) was from Sunresin New Materials Co., Ltd. (Xi'an, China). PPD, Rh2, and F12 were purchased from Nanjing Spring & Autumn Biological Engineering Co., Ltd. (Nanjing, China). Uridine 5′-diphosphate glucose (UDPG) and uridine 5′-diphosphate sodium salt (UDP) were obtained from Nanjing Duly Biotech Co., Ltd. (Nanjing, China). Sucrose, epichlorohydrin, iminodiacetic acid (IDA), and other chemicals used were all analytical grade. Solvents used for high-performance liquid chromatography (HPLC) analysis were of HPLC grade. Bradford Protein Assay Kit and bovine serum albumin for Protein Concentration Assay were from Sangon To further alleviate the possible inhibition by high substrate concentration, a fedbatch strategy by stepwise addition of 2.0 mM fresh PPD was adopted to the cascade reaction with the co-immobilized M315F-AtSuSy. The initial concentration of PPD (8 mM) was used, a titer of the ginsenoside Rh2 cumulatively reached up to 16.6 mM (10.3 g/L) ( Figure 5B) by using fed-batch technology and the final yield was about 83.2%. To our acknowledgment, this titer of ginsenoside Rh2 produced by the present biotechnological way was the highest level, and it was 2.78-3.2 times of free enzymes [21], and much higher than other reports [14]. Using co-immobilized M315F-AtSuSy coupled reaction with fedbatch protocol, the overall STY of ginsenoside Rh2 was about 0.396 g/L/h, which was 20-fold higher than that of the report [13]. The co-immobilized M315F-AtSuSy catalyzed the transglycosylating reaction afforded a new way for efficient manufacturing of rare ginsenoside Rh2, and formed a green, cost-effective and sustainable approach with UDPglucose regeneration from sucrose and catalytic amount of UDP.

Materials and Chemicals
The C-terminal 6*his-tagged glycosyltransferase Bs-YjiC (from Bacillus subtilis 168) mutant M315F and C-terminal 6*his-tagged sucrose synthase AtSuSy (from Arabidopsis thaliana) were obtained and prepared respectively as in our previous report [21]. Macroporous resin LX1000HG (contained hydroxyl group) was from Sunresin New Materials Co., Ltd. (Xi'an, China). PPD, Rh2, and F12 were purchased from Nanjing Spring & Autumn Biological Engineering Co., Ltd. (Nanjing, China). Uridine 5 -diphosphate glucose (UDPG) and uridine 5 -diphosphate sodium salt (UDP) were obtained from Nanjing Duly Biotech Co., Ltd. (Nanjing, China). Sucrose, epichlorohydrin, iminodiacetic acid (IDA), and other chemicals used were all analytical grade. Solvents used for high-performance liquid chromatography (HPLC) analysis were of HPLC grade. Bradford Protein Assay Kit and bovine serum albumin for Protein Concentration Assay were from Sangon Biotech (Shanghai, China).

Resin Modified with Bifunctional Groups and Carries Characterization
Macroporous resin LX1000HG (contained hydroxyl group) modification was according to ref. [29,32], and the procedure was also shown in Figure S7. It was firstly modified with epichlorohydrin to obtain epoxy-activated resin LX1000HG. Epoxy-resin LX1000HG was used in the treatment with iminodiacetic acid (IDA), then produced diol (glyceryl) groups and IDA groups activated resin. The support was washed with distilled water, and further oxidation of glyceryl groups was achieved using 100 mL of sodium periodate (NaIO 4 ), producing glyoxyl-IDA-activated resin [29,32]. To completely block the epoxy group, the resin was incubated with 3.0 M glycine for 16 h at 20 • C [33]. Glyoxyl-IDA-activated resin was added to the metal chelate solutions (30 mg/mL, CoCl 2 ·6H 2 O, NiCl 2 ·6H 2 O, or CuSO 4 ·5H 2 O) at pH 7.0 for 1.0 h at 25 • C [28]. Finally, the glyoxyl-IDA metal chelate bifunctionalized resin was washed with distilled water and vacuum dried for enzyme adsorption.
The morphology of bifunctional resin and support with co-immobilized enzymes of M315F-AtSuSy were examined by a scanning electron microscope (SEM, S-4800, Hitachi, Tokyo, Japan). During imaging, the scanning electron microscope had an acceleration voltage of 3.0 kV. The samples were freeze-dried and prepared for microscope imaging.
The FTIR spectra were obtained on samples in KBr pellets using a Shimadzu FTIR-8400S (Tokyo, Japan) spectrometer in the frequency range of 400 to 4000 cm −1 . The samples were mixed with 1% (w/w) KBr, and the analysis was employed at 10 scans per second with a resolution of 4 cm −1 .
The specific surface area, pore volumes, and average pore diameters of resin and support with co-immobilized enzymes were measured by obtaining N 2 adsorption-desorption (ASAP 2020 V4.00, Micromeritics Instrument Co., Ltd., Shanghai). The specific areas were estimated by the standard Brunauer-Emmett-Teller (BET) method, and the distributions of pore diameters were estimated by the desorption branches of the isotherms with the Barrett-Joyner-Halenda (BJH) model.

Dual Enzyme Co-Immobilization with Adsorption and Covalent Attachment
E. coli BL21(DE3) crude extract was used as a multi-protein model system during the immobilization tests with different supports. For the co-immobilization of engineered M315F and AtSuSy onto heterofunctional resin with two kinds of groups, the two-step immobilization protocol was accomplished according to Mateo et al. [32,33]. His tag affinity method is generic and can be easily applied to specific adsorption [28,34]. The crude two enzymes from E. coli BL21(DE3) crude cell extract (M315/AtSuSy total activity addition of 4000 mU/g wet resin) were in 10 mL solution (25 mM Tris-HCl buffer, pH 7.0) at 25 • C and adsorbed by heterofunctional resin (1.0 g wet resin) under mild stirring for 6 h. In the first step, the optimal ratio of M315F/AtSuSy adsorbed by resin for the biosynthesis of Rh2 was determined using the various ratios of M315F/AtSuSy adjustment, next, adsorption time, pH, and temperature were studied, respectively.
The final step was incubated with 50 mM Gly-NaOH buffer pH 10.0 for 3 h (intermittent shaking) to allow an intramolecular multipoint covalent attachment between the glyoxyl groups on the support and the amino/other groups on the enzyme surface. Finally, the immobilized enzyme was washed with excess distilled water. The resin with co-immobilized enzymes was separated, and samples of the supernatants and resins were withdrawn. The enzyme activity was measured and the protein concentration of the solution was also analyzed. The biocatalyst was stored at 4 • C until further use.
The immobilization yield IY imob (%) was calculated via the difference between the activity of the free enzyme solution (A tf ) and the remaining activity of the supernatant (A ts ) using Equation (1): The amount of enzyme offered for each gram of resin (At off − U/g) and the immobilization yield were individually collected to calculate the theoretical activity (At t ) of immobilized enzyme M315F or AtSuSy, as shown in Equation (2): The co-immobilized enzymes activity (At d ) individually assayed by the theoretical activity (At t ) in the solution was defined as the enzymatic activity recovery according to Equation (3):

Enzyme Activity and Protein Assay
The enzyme activities of glycosyltransferase Bs-YjiC mutant M315F and sucrose synthase AtSuSy were measured as in our previous report [21]. Protein concentration in solution was examined by Bradford's method using bovine serum albumin as Ref. [35].

Biochemical Properties of Free and Co-Immobilized Enzymes
For the characterization of free and co-immobilized M315F and AtSuSy, the effects of pH on the activity and pH stability (35 • C, 12 h) of enzymes were determined in various buffer systems (pH 5.0-10.0). The effects of temperature on the activity and thermal stability (pH 8.0, 12 h) of free and immobilized enzymes were determined at different temperatures (from 20 to 55 • C).

Optimization of the Co-Immobilized Enzymes Catalyzed Reaction Condition
Given the results of Rh2 synthesized by free enzymes M315F-AtSuSy coupled reaction in our previous report [21], some reaction conditions for the co-immobilized enzymes catalyzed reactions were the same as free enzymes. The reaction contained sucrose (0.4 M) and MgSO 4 (5 mM), Tris-HCl buffer (pH 8.0), and was conducted at a temperature of 35 • C and 200 rpm.

Reusability of Co-Immobilized Enzymes and Fed-Batch Strategy
The operational stability for the co-immobilized M315F and AtSuSy coupled reaction was examined by repeated usage of co-immobilized enzymes. At the end of each batch, the immobilized enzymes were removed and washed with fresh buffer (pH 8.0), and then added to the initial reaction mixture of the next cycle. The reusability was measured as the yield of Rh2 produced by each time.
For the co-immobilized M315F and AtSuSy coupled reaction (20 mL) with fed-batch method, the optimal conditions of initial concentration of PPD (8 mM), UDP (0.5 mM), sucrose (0.4 M), 50 mM Tris-HCl (pH 8.0), 10%(v/v) DMSO and MgSO 4 (5 mM) in Tris-HCl buffer (pH 8.0), and conducted at a temperature of 35 • C and 200 rpm. During the proceed reaction, the sample was monitored by HPLC [21], fresh PPD (0.2 mL) dissolved in DMSO (100 mM) was stepwise added to the reactants at 2, 4, 6, 8,10, and 12 h, and a total concentration of PPD was added to the final reaction mixture.

Conclusions
In summary, glycosyltransferase Bs-YjiC mutant M315F and sucrose synthase AtSuSy was exploited to biosynthesize rare ginsenoside Rh2 in one-pot reaction from UDP and sucrose. This study disclosed that resin LX1000HG modified with glyoxyl-metal-chelate bifunctional groups were useful for co-immobilization of the engineered enzymes with his-tags. Using the two-step process of specific adsorption and multipoint covalent attachment, the co-immobilized enzymes showed the advantages of improved binding stability, improved pH and thermal stabilities, and good operation stability. Interestingly, a higher yield was achieved using co-immobilized enzymes through allowing the bioreaction at a high initial concentration of PPD by alleviating inhibition, and the noticeable titer of ginsenoside Rh2 reached the high level of 10.3 g/L by periodic feeding of PPD. This study established a green and sustainable approach for the production of ginsenoside Rh2, and it may provide promising candidates for natural drug research and development.

Patents
There are no patents resulting from the work reported in this manuscript.

Supplementary Materials:
The following are available online at https://www.mdpi.com/2073-4 344/11/1/132/s1, Figure S1: Optimization of the adsorbed condition for enzymes of M315F and AtSuSy bound, Figure S2: FTIR spectra of resin LX1000HG, LX1000HG-glyoxyl-Ni 2+ and LX1000HGglyoxyl-Ni 2+ with enzymes, Figure S3: Effect of pH on the activities of free and immobilized enzyme M315F/AtSuSy, Figure S4: pH stability of free and immobilized enzyme M315F/AtSuSy, Figure S5: Effect of temperature on the activities of free and immobilized enzyme M315F/AtSuSy, Figure S6: Effect of UDP concentration on the immobilized M315F and AtSuSy coupled reactions for Rh2 synthesis, Figure S7: Activation procedure for heterofunctional resin preparation.