2.1 Production of BgpA

GST-BgpA gene was expressed in E. coli under the control of the IPTG-inducible promoter Ptac. Induction with 0.5 mM IPTG at 20 ℃ for 6 h produced the maximum level of soluble active enzyme, resulting in 1,000 mL of crude enzyme solution. The protein concentration was 3.886 ± 0.116 mg/mL and specific activity of BgpA in crude enzyme solution was 82.988 ± 3.581 U/mg. The results of SDS-PAGE were shown in Fig. 1a (Lane 1) and the size of the target protein band at ca. 70 kDa were in good agreement with the estimated size based on the translated polypeptide sequence (71.14 kDa) [32], which suggested that the glucosidase was physiologically active as a trimeric protein. However, some protein impurities were also observed, which originated from endogenous proteins of E. coli. and need to be removed by conventional column chromatography and precipitation. In this study, recombinant BgpA was purified and immobilized from the crude enzyme in one step by using the mesoporous silica SBA-15 after surface modification, so there was no extra step to remove protein impurities in crude enzyme by routine operations, and the removal of protein impurities was achieved at the time of immobilization of BgpA on R-SBA-15.

Fig. 1

Immobilization of BgpA onto SBA-15 or X-SBA-15. a SDS-PAGE profiles. Lane L: protein marker; Lane 1: crude BgpA; Lane 2: unbound proteins of SBA-15; Lane 3: unbound proteins of R-SBA-15; Lane 4: unbound proteins of S-SBA-15; Lane 5: unbound proteins of P-SBA-15; Lane 6: unbound proteins of Q-SBA-15; Lane 7: unbound proteins of U-SBA-15. b activity recovery and loading efficiency of SBA-15 or X-SBA-15. BgpA recombinant β-glucosidase from Terrabacter ginsenosidimutans, SBA-15 Santa Barbara Amorphous 15, X-SBA-15 different group modified SBA-15, R-SBA-15 N-aminoethyl-γ-aminopropyl trimethoxy modified SBA-15, S-SBA-15 (3-Aminopropyl) triethoxy modified SBA-15, P-SBA-15 Trimethyloxy phenyl modified SBA-15, Q-SBA-15 N-octyl triethoxy modified SBA-15, U-SBA-15 Anilinomethyl triethoxy modified SBA-15

2.2 Immobilization of BgpA onto the carriers

To achieve the highest activity recovery, a total of six different carriers, namely SBA-15, Anilinomethyl triethoxy, (3-Aminopropyl) triethoxy, N-aminoethyl-γ-aminopropyl trimethoxy, N-octyl triethoxy or Trimethyloxy phenyl modified SBA-15 (U-SBA-15, S-SBA-15, R-SBA-15, Q-SBA-15 or P-SBA-15) were compared for BgpA immobilization. The characteristics of the carrier could affect the properties of the immobilized enzyme. For the X-SBA-15 carriers, different performance in loading efficiency and activity recovery was observed when an equal amount of BgpA was added. As shown in Fig. 1b, the activity recovery of U-SBA-15@BgpA, P-SBA-15@BgpA and Q-SBA-15@BgpA was all below 10%, not much higher than SBA-15@BgpA. Meanwhile, except for U-SBA-15 exhibiting an extremely low loading efficiency at ca. 5%, the efficiency of the other five carriers were around 50%. In addition, S-SBA-15@BgpA exhibited an activity recovery above 10% and the highest loading efficiency among those six carriers, while R-SBA-15@BgpA showed a significant advantage with activity recovery of 23.5%. S-SBA-15 and R-SBA-15 demonstrated their good performance for immobilizing BgpA and it implied that the enzyme has strong affinity to some specific carriers. The similarity between these two carriers was that they both carry amino groups on their functionalized surfaces, which could be responsible for the high activity recovery of S-SBA-15@BgpA and R-SBA-15@BgpA.

Interestingly, seen from the results of SDS-PAGE in Fig. 1a, both R-SBA-15 (Lane 3) and S-SBA-15 (Lane 4) among the six carriers were found to specifically immobilize target protein BgpA (ca. 70 kDa), meanwhile other protein bands did not significantly vary. Especially, the recombinant BgpA in the crude enzyme was completely captured by R-SBA-15 to achieve immobilization and purification of BgpA in one step, demonstrating a specific interaction between the carrier and the enzyme. The rationale for this finding could be that organic modification has changed the micro-environment of original SBA-15, and different organic functional groups led to differences in the micro-environment, which could potentially alter the characteristics of the immobilization agent [36]. In this application, the modification by N-aminoethyl-γ-aminopropyl trimethoxy on SBA-15 has achieved better specific affinity with BgpA than (3-Aminopropyl) triethoxy, Trimethyloxy phenyl, N-octyl triethoxy, and Anilinomethyl triethoxy. After recombinant BgpA has been adsorbed onto R-SBA-15, the amount of protonated species becomes dominant as a result of the formation of -NH3+•••−OOC- as a result of the ionic interaction between surface amine groups and carboxyl groups of BgpA [37].

2.3 Optimization of R-SBA-15@BgpA preparation conditions

There are several key factors affecting immobilization were studied by one-factor-at-a-time experiments to achieve the best performance of R-SBA-15@BgpA. It was seen from Fig. 2 that the amount of crude enzyme solution added, adsorption duration, pH of buffer solution, and temperature have differently affected loading efficiency and activity recovery, respectively.

Fig. 2

Effect of the a the amount of crude enzyme solution added, b adsorption duration, c pH and d temperature on the activity recovery and loading efficiency of R-SBA-15@BgpA. R-SBA-15@BgpA immobilized BgpA on N-aminoethyl-γ-aminopropyl trimethoxy group modified Santa Barbara Amorphous 15

As shown in Fig. 2a, the loading efficiency was at the highest level when 0.8 mL of crude enzyme solution was added, and the activity recovery continued to increase as more enzyme solution was added. When the enzyme solution volume reached 2.4 mL, the activity recovery arrived at its maximum. However, with further addition of enzyme solution, both the activity recovery and loading efficiency showed a decreasing trend. This could be due to saturated protein immobilization on the carrier. In addition, the enzyme became oversaturated on the carrier, probably masking the active sites of the enzyme immobilized on the carrier.

Seen from Fig. 2b, the loading efficiency had already reached 50% after the enzyme solution was poured into the carrier prior to the mixing and sonication for 1 min, indicating an extremely fast immobilization of BgpA on R-SBA-15. As the adsorption lasted for a longer time, the maximum activity recovery was achieved 25.4% at 15 min. Then the loading efficiency still slowly increased to 64.9% even if the activity recovery kept decreasing. High protein loading created steric hindrance between protein molecules, retarding the diffusion of substrates and products; or it led to a high degree of enzyme cross-linking, which could block the active sites on the enzyme and cause a decrease in the activity recovery.

As shown in Fig. 3c, the difference in the effects of pH on the activity recovery and loading efficiency of immobilized enzymes was not significant. The highest enzyme activity recovery was achieved when the pH was set at 7.0.

Fig. 3

Effect of a pH and b temperature on the activity of the free BgpA and R-SBA-15@BgpA. BgpA recombinant β-glucosidase from Terrabacter ginsenosidimutans, R-SBA-15@BgpA immobilized BgpA on N-aminoethyl-γ-aminopropyl trimethoxy modified Santa Barbara Amorphous 15

Seen from Fig. 3d, with the increase of temperature, the activity recovery of immobilized enzyme showed an increasing trend followed by a decreasing trend. There is no significant difference in the loading efficiency, and the optimal activity recovery 24.2% was achieved at 37℃. This suggested that the temperature could influence the distribution of enzyme molecules in the carrier, thus affecting the exposure of the active site.

2.4 Optimum pH and temperature

Figure 3a showed the relative activities of free BgpA and the immobilized enzyme R-SBA-15@BgpA in buffers of different pH values ranging from 3.0 to 9.0. It can be observed that R-SBA-15@BgpA had the highest relative activity at pH 7.0, and showed no significant advantage over the free enzyme. This indicated that the carrier R-SBA-15 did not have a significant beneficial effect on the acid–base tolerance of BgpA. On the other hand, Fig. 3b showed the relative activities of free BgpA and the immobilized enzyme R-SBA-15@BgpA at different temperatures ranging from 27 to 82 °C. It can be seen that R-SBA-15@BgpA had a 20 °C higher optimal temperature than the free enzyme, which was increased from 42 to 62 °C.

2.5 Thermal stability

To further confirm the advantage of the immobilized enzyme R-SBA-15@BgpA, thermal stability experiment was subsequently performed. The relative activity of immobilized and free BgpA at different temperatures after incubation was shown in Fig. 4. After continuous incubation at 42 °C for 8.0 h, immobilized BgpA still retained relative activity above 70%, while free enzyme lost 60% of its initial enzyme activity. However, after incubation at 47 °C for 6 h, immobilized enzyme retained one-third of its initial enzyme activity, while the activity of free enzyme was only about one-tenth. When the temperature was raised to 52 °C, the activity of free enzyme decreased sharply and the relative activity dropped to below 10% after 2-h incubation, while that of immobilized enzyme remained at around 40%. When the temperature continued to rise to 57 °C or above, the relative activity of both free and immobilized enzymes decreased in a rapid manner. Within one-hour incubation, the relative activity of immobilized enzyme decreased to ca. 10%, while the activity of free enzyme almost disappeared. The results confirm the role of immobilization in improving the thermal stability of BgpA. The good thermal stability could be attributed to the suitable microenvironment provided by the functionalized mesoporous silica, which prevented BgpA from thermal denaturation [36]. Strong thermal stability is a crucial feature, which could equip the enzyme with a good prospect in its industrial applications.

Fig. 4

Thermal stability of free BgpA and R-SBA-15@BgpA at 42 ℃, 47 ℃, 52 ℃, 57 ℃, 62 ℃. BgpA recombinant β-glucosidase from Terrabacter ginsenosidimutans, R-SBA-15@BgpA immobilized BgpA on N-aminoethyl-γ-aminopropyl trimethoxy modified Santa Barbara Amorphous 15

2.6 Organic solvent resistance

The relative activity of free and immobilized BgpA in different proportions of various organic reagents was determined, and the results were shown in Fig. 5a. From the degree of influence of organic reagents on relative activity, the immobilized enzymes have an advantage in tolerating organic solvent over free BgpA. When the buffer contains 15% (v/v) organic solvent, methanol and DMSO have little effect on the relative activity of free and immobilized enzymes, while DES have a greater impact and decrease the activity quickly. When the buffer contains 60% (v/v) organic solvent, methanol has the greatest impact on relative activity among the solvents. As a matter of fact, the relative activity of immobilized enzymes does not significantly decrease at an organic reagent volume ratio of 15%. To further determine the capability of immobilized enzymes to tolerate organic reagents, they were incubated at different time intervals in 0–15% methanol, DES, and DMSO, and the changes in relative activity were calculated.

Fig. 5

Organic solvents resistance of free BgpA and R-SBA-15@BgpA: a different proportions of different organic reagents; b different proportions of methanol; c different proportions of DMSO; d different proportions of DES. BgpA recombinant β-glucosidase from Terrabacter ginsenosidimutans, R-SBA-15@BgpA immobilized BgpA on N-aminoethyl-γ-aminopropyl trimethoxy modified Santa Barbara Amorphous 15, DES Deep eutectic solvent

It was seen from Fig. 5b that the relative activity of both free and immobilized enzymes declined to varying degrees after prolonged incubation in buffers containing methanol. Especially, under the condition of 15% methanol, the relative activity was found to decline the fastest among all the concentration, and recent research suggests that methanol alters the molecular structure of proteins [38], thereby affecting enzyme catalytic activity. The moderate decline in the relative activity of immobilized enzyme was also observed, suggesting that the carrier was beneficial to protect the structure of BgpA.

As shown in Fig. 5c, an unexpected decrease in relative activity was observed after the free enzyme was incubated in buffer without organic solvent. Meanwhile after incubation in buffers composed of DMSO for 2 h, both the free and immobilized enzymes maintained relative activity at 100% or even above. Among them, in the buffer containing 15% DMSO, the relative activity of immobilized enzyme was much higher than that of free enzyme, indicating that immobilized enzyme can remain stable against 15% DMSO.

From Fig. 5d, it was seen that the relative enzyme activity of both free BgpA and R-SBA-15@BgpA in buffers containing DES were increased within the initial phase and then returned to around 100% when the incubation lasted for 2 h. It is particularly noteworthy that the relative enzyme activity of the free enzyme can reach as high as 300% in the buffer composed of 5% DES, followed by a rapid decline with prolonged incubation time. However, the relative activity of immobilized enzyme did not show an intensive increase or decrease, this is pertaining to the fact that DES are solvents enhancing solubility but of high viscosity and the higher solubility of substrates in the DES. Thus, for both free and immobilized enzymes, the increase in the solubility helps diffuse substrates and products, thereby accelerating contact between substrates and active sites and subsequently enhancing catalyst’s performance [39, 40]. However, the high viscosity could have greatly blocked the reaction between substrates and the immobilized enzyme, so no significant advantage of the immobilized enzyme over the free enzyme was observed when DES were incorporated into the buffer.

The results have confirmed the role of immobilization in improving BgpA resistance to organic solvents. Good organic solvent tolerance means that the organic groups offered by R-SBA-15 have created enzymes with a suitable microenvironment to prevent enzyme denaturation [41]. DES and DMSO, as co-solvents, were expected to play a helpful role in enzyme hydrolysis of natural products. The results suggested that DES were suitable for hydrolysis reactions using free enzymes, while DMSO was more suitable for hydrolysis reactions by immobilized enzymes.

2.7 Leaching test and storage stability

The results of the leaching test were shown in Fig. 6. From the SDS-PAGE results, the protein bands were not observed in the solution of immobilized enzyme on the 1st, 15th, and 30th day, indicating negligible enzyme leakage from R-SBA-15@BgpA. Gascón et al. [42] proposed that the limited amount of protein desorption was due to pore restrictions. Additionally, Yiu et al. [43] pointed out that the functional groups of modified SBA-15 prevent enzyme leaching, which was related to the interaction between the modified SBA-15 surface and the enzyme's functional groups. The firm ionic interaction between the enzyme and carrier prevents protein leaching. And, due to the presence of ionic interaction between carboxyl groups in BgpA and amine groups on the surface of R-SBA-15, the release rate of BgpA from SBA-15 functionalized by aminoethyl-γ-aminopropyl trimethoxy could be very limited.

Fig. 6

Leaching test and storage stability of R-SBA-15@BgpA. a leakage rate and storage stability of BgpA and R-SBA-15@BgpA. b SDS-PAGE profiles. Lane L. protein ladder; Lane 1. crude BgpA. Lane 2. unbound proteins of R-SBA-15; Lane 3. storage for 1 day. Lane 4. storage for 15 days. Lane 5. storage for 30 days. BgpA recombinant β-glucosidase from Terrabacter ginsenosidimutans, R-SBA-15 N-aminoethyl-γ-aminopropyl trimethoxy group modified Santa Barbara Amorphous 15, R-SBA-15@BgpA immobilized BgpA on R-SBA-15

The storage stability of both free and immobilized BgpA was determined, and the results were shown in Fig. 6. The immobilized enzyme R-SBA-15@BgpA retained over 50% of its relative activity even after continuous storage at 4 °C for 33 days. However, the relative activity of free BgpA almost lost, which could be due to the autolysis of BgpA in the phosphate buffer. The results confirmed the crucial role of immobilization in enhancing the storage stability of BgpA.

2.8 Catalyst characterization

The N2 adsorption–desorption isotherms and pore size distribution of SBA-15, R-SBA-15, and R-SBA-15@BgpA were shown in Fig. 7. According to the IUPAC classification, all particles displayed typical IV-type isotherms with a well-defined H4-type hysteresis loop and sharp capillary condensation [36]. The H4-type hysteresis loop was a typical feature of ordered mesoporous materials with a two-dimensional hexagonal structure. Additionally, the shape of the hysteresis loop remained unchanged after organic modification and enzyme immobilization, indicating that the ordered mesoporous structure of SBA-15 remained intact. The pore size distributions of the parent material, modified SBA-15, and R-SBA-15@BgpA were also presented in Fig. 7 for comparison. The pore size, the pore volume and BET surface area of SBA-15 was 12.1 nm, 0.967 cm3/g and 568 m2/g, respectively. R-SBA-15 had pore size of 6.7 nm, BET surface area of 432 m2/g and pore volume of 0.665 cm3/g, and the resulting immobilized enzyme R-SBA-15@BgpA has a smaller pore size (6.5 nm), BET surface area (417 m2/g) and pore volume (0.568 cm3/g). As expected, the pore size, pore volume, and surface area of SBA-15 decreased after chemical modification. In particular, the incorporation of N-aminoethyl-γ-aminopropyl trimethoxysilane reduced the surface area of SBA-15, and the pore size distribution of R-SBA-15@BgpA indicated further reduction in pore size after successful enzyme immobilization in this application [44,45,46].

Fig. 7

N2 adsorption–desorption isotherms and pore size distributions of the a SBA-15, b R-SBA-15 and c R-SBA-15@BgpA (BgpA: recombinant β-glucosidase from Terrabacter ginsenosidimutans; SBA-15 Santa Barbara Amorphous 15, R-SBA-15 N-aminoethyl-γ-aminopropyl trimethoxy modified SBA-15, R-SBA-15@BgpA immobilized BgpA on R-SBA-15

2.9 Determination of epimedin A and sagittatoside A

HPLC–UV chromatograms of epimedin A, sagittatoside A and sample solution were shown in Figs. 8 and 9. Standard curves of two references were drawn as Y = 5.776X + 16.8 (r2 = 0.9994) and Y = 5.7671X + 26.263 (r2 = 0.9994), separately, indicating good linear relationships within their concentration ranges.

Fig. 8

Typical HPLC–UV chromatograms of epimedin A (a) and sagittatoside A (b)

Fig. 9

Typical HPLC–UV chromatograms of epimedin A (a); incomplete hydrolysis (b) and complete hydrolysis (c) (1: epimedin A, 2: sagittatoside A)

2.10 Conversion of epimedin A and catalyst reusability tests of R-SBA-15@BgpA

For large-scale applications of an enzyme in industry, reusability is crucial in terms of efficiency and economy. Reusability of the R-SBA-15@BgpA was studied and results were presented in Fig. 10. Under the condition of 15% DMSO in phosphate buffer, the relative conversion rate of epimedin A was kept nearly 100% even after the R-SBA-15@BgpA has been used for ten cycles, and the R-SBA-15@BgpA showed an excellent stability against the reaction and remained 76.1% of its initial activity after fourteen cycles. It indicated that immobilized enzyme on R-SBA-15 were potential in practical application for preparation of natural products.

Fig. 10

Reusability of R-SBA-15@BgpA for hydrolysis of epimedin A (R-SBA-15@BgpA: immobilized BgpA on N-aminoethyl-γ-aminopropyl trimethoxy modified Santa Barbara Amorphous 15)