Enzymatic Synthesis of Resveratrol α-Glucoside by Amylosucrase of Deinococcus geothermalis

Glycosylation of resveratrol was carried out by using the amylosucrase of Deinococcus geothermalis, and the glycosylated products were tested for their solubility, chemical stability, and biological activities. We synthesized and identified these two major glycosylated products as resveratrol-4′-O-α-glucoside and resveratrol-3-O-α-glucoside by nuclear magnetic resonance analysis with a ratio of 5:1. The water solubilities of the two resveratrol-α-glucoside isomers (α-piceid isomers) were approximately 3.6 and 13.5 times higher than that of β-piceid and resveratrol, respectively, and they were also highly stable in buffered solutions. The antioxidant activity of the α-piceid isomers, examined by radical scavenging capability, showed it to be initially lower than that of resveratrol, but as time passed, the α-piceid isomers’ activity reached a level similar to that of resveratrol. The α-piceid isomers also showed better inhibitory activity against tyrosinase and melanin synthesis in B16F10 melanoma cells than β-piceid. The cellular uptake of the α-piceid isomers, which was assessed by ultra-performance liquid chromatography (UPLC) analysis of the cell-free extracts of B16F10 melanoma cells, demonstrated that the glycosylated form of resveratrol was gradually converted to resveratrol inside the cells. These results indicate that the enzymatic glycosylation of resveratrol could be a useful method for enhancing the bioavailability of resveratrol.


Introduction
In this study, resveratrol glucosides were synthesized through transglycosylation by the amylosucrase from Deinococcus geothermalis to improve the bioavailability of resveratrol. The glycosylation was successfully conducted with resveratrol as the acceptor and sucrose as the donor. The molecular structures of the glycosylated compounds, that is, the resveratrol glucosides, were identified, and their water solubility, oxidative stability, and biological activities were examined compared with resveratrol and β-piceid, a natural resveratrol glucoside.

Enzymatic Synthesis of Resveratrol Glucosides with DGAS
The substrate solution in 50 mM Tris-HCl buffer (pH 8.0) containing 5 mM resveratrol, 10 mM sucrose, and 20% dimethylsulfoxide (DMSO) was preincubated at 35°C for 10 min, followed by the addition of DGAS (0.1 mg/ ml). The incubation of the reaction mixture was continued for 12 h [16]. The reaction was stopped by boiling for 10 min and placing the mixture tube in ice. The amount of synthesized resveratrol glucosides was determined by thin-layer chromatography (TLC) and ultra-performance liquid chromatography (UPLC) analyses.

TLC and UPLC Analyses
TLC was performed using silica gel 60 and silica gel RP-18 F254S (Merck KGaA, Germany) to identify the respective resveratrol glucosides, and the plates were developed in a developing solution containing n-butanol/ ethanol/water (5:3:2, v/v/v) and methanol/water (5:5, v/v) for silica gel 60 and silica gel RP-18 F254S, respectively. The developed TLC plates were dried completely at room temperature and then visualized by dipping in a solution containing 1% (w/v) N-(1-naphtyl)-ethylenediamine and 20% (v/v) sulfuric acid in methanol followed by heating at 110°C for 10 min.
UPLC analysis was performed on an Acquity H-Class UPLC system (Waters, Ireland), which comprises a photo diode array eλ detector. The separation of resveratrol and resveratrol glucosides was carried out by reversed-phase LC using a BEH C 18 (2.1 mm × 50 mm id, particle size: 1.7 μm) column (Waters). The column temperature was 40°C, the flow rate was 0.6 ml/min, and the injection volume was 2 μl. The resveratrol and its glucosides were observed at 306 nm, at which resveratrol showed an absorbance maxima. Solvent A was 0.1% (v/v) formic acid in water, and solvent B was 0.02% (v/v) formic acid in 80% (v/v) acetonitrile. The gradient conditions were as follows: 82% A over 0.68 min, 82% to 77% A over 0.48 min, 77% to 75.5% A over 0.27 min, 75.5% to 68.5% A over 0.41 min, 68.5% to 0% A over 0.2 min, 0% A over 0.34 min, and then 82% A for 2.62 min.

Purification of Resveratrol Glucosides
The resveratrol glucosides were purified in accordance with the method described by Moon et al. [17]. In brief, the resveratrol glucosides were separated by Strata C 18 -T cartridge (Phenomenex, USA) and preparative LC (Prep LC) equipped with an ultraviolet (UV) detector (JAI, Japan). A C 18 -T cartridge, which was previously activated by methanol and water, was used to absorb the resveratrol glucosides in the reaction mixture and to eliminate sugar and salts. The reaction mixture was filtered by a 0.45 μm syringe filter (Germany) and subjected to a C 18 -T cartridge. After washing with water, the elution of resveratrol glucosides was carried out with methanol. The main resveratrol glucosides in methanol were purified using a W252 polymeric gel filtration column (2 cm × 50 cm, JAI) in the Prep LC. The mobile phase was 100% methanol at a flow rate of 3 ml/min. The fractions corresponding to the detected peaks were collected and lyophilized. The purity of each sample was confirmed using UPLC analysis.

Nuclear Magnetic Resonance (NMR) Analysis
The 1 H and 13 C NMR spectra of purified resveratrol glucosides were obtained with a Varian Inova AS 400 MHz NMR spectrometer (Varian, USA). The sample was dissolved in DMSO-d 6 at 24°C with tetramethylsilane as the chemical shift reference.

Water Solubility Determination
To maximize the solubility of each compound, we suspended the excess amounts of resveratrol, β-piceid, and resveratrol glucosides in 200 μl distilled water in a microcentrifuge tube and then sonicated the tube for 1 h at room temperature using an ultrasonic cleaner (JAC-4020, Kodo, Korea) [17]. Each sample was centrifuged at 16,000 g for 30 min. Subsequently, the supernatant of each sample was filtered, and the concentration of the compound in the supernatant as a water-soluble component was estimated by UPLC analysis. Resveratrol and βpiceid were used as standards, and the water solubility was expressed as g/l.

Stability against Oxidative Degradation
Resveratrol, β-piceid, and resveratrol glucosides were dissolved in buffer solutions of various pH values (hydrochloric acid, pH 2.0; phosphate-buffered saline (PBS), simulated body fluid (SBF), pH 7.4; Tris-HCl, pH 9.0, and DMEM to reach a final concentration of 10 mM. The solutions were then incubated at 37°C, and an aliquot (300 μl) of the reaction mixture was extracted at different time points up to 72 h, filtered, and analyzed by UPLC.

Determination of Antioxidant Activities
DPPH free-radical scavenging capability and ferric reducing antioxidant power (FRAP) assay results were evaluated as described previously by Moon et al. [24]. Hydroxy radical-scavenging assay was measured as described by Su et al. [3]. Vitamin C was used as a positive control. Each assay was measured from 1 min to 30 h. The DPPH radical-scavenging activity was calculated as follows: DPPH radical-scavenging activity (%) = (1-A sample /A blank ) × 100, where A sample is the absorbance in the presence of the sample, and A blank is the absorbance of the blank. Hydroxy radical-scavenging activity was calculated in accordance with the following formula: Hydroxy radical scavenging activity (%) = (A sample -A control )/(A blank -A control ) × 100, where A sample is the absorbance in the presence of the sample, A control is the absorbance of the blank with H 2 O 2 , and A blank is the absorbance of the blank without H 2 O 2 .

Determination of Anti-Melanogenesis Activity
Cell culture and cell viability assay. B16F10 melanoma cells were maintained in DMEM supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin. The cells were cultured at 37°C under 5% CO 2 in a humidified culture chamber. Cell viability was determined by 3-(4,5-dimethythiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay. The cells (6 × 10 3 cells/well) were seeded into 96-well plates and cultured for 24 h. The cells were then treated with resveratrol, β-piceid, or resveratrol glucosides at various concentrations. After 24 h, the cells were washed with PBS, MTT reagent (0.5 mg/ml) was added to each well, and the cells were incubated at 37°C for 1 h. The supernatant in each well was discarded for the treatment with DMSO, which was used to dissolve the formazan crystals formed from viable cells. After the DMSO treatment, the absorbance of each well was measured at 595 nm using a microplate reader (Epoch, Biotek, USA).

Determination of Cellular Tyrosinase Activity and Melanin Content
Cellular tyrosinase activity was measured following the method described by Park et al. with slight modifications [25]. The cells (4 × 10 4 cells/well) were seeded into 24-well plates and cultured for 24 h. The cells were exposed to resveratrol, β-piceid, or resveratrol glucosides in the presence or absence of 100 nM α-melanocyte stimulating hormone (α-MSH) for 48 h. The cells were then washed with PBS and lysed with a lysis buffer. The lysates were then clarified by centrifugation at 16,000 g for 15 min at 4°C. The protein concentration was determined by Bradford method (Bio-Rad Laboratories Inc., USA) using bovine serum albumin as a standard. The reaction mixture consisting of 50 μg protein (adjusted to 100 μl with 0.1 M sodium phosphate, pH 6.8) and 100 μl 10 mM L-3,4-dihydroxyphenylalanine was added to each well of the 96-well plates. After incubation at 37°C for 1 h, the absorbance was measured at 490 nm using a microplate reader.
Cellular melanin content was determined by the method of Park et al. with minor modifications [25]. The cells (5 × 10 4 cells/well) were seeded into 6-well plates and cultured for 24 h. The cells were treated with resveratrol, βpiceid, or resveratrol glucosides for 72 h. Then, the cells were washed with PBS and lysed with lysis buffer, and the supernatant was removed by centrifugation at 16,000 g for 15 min at 4°C. The protein concentration of the supernatant was determined by the Bradford method. The pellet was dissolved in 1 M NaOH and measured at 405 nm using a microplate reader.

Cellular Uptake Assay
The cellular uptake assay was conducted following the method described by Moon et al. with minor modification [17]. The B16F10 cells (2 × 10 5 cells/well) were seeded into 6-well plates. The cells were cultured for 24 h, and then, resveratrol (20 μM), β-piceid, or resveratrol glucosides (100 μM) was added. After 24 h of incubation, the medium was removed, and the cells were washed with PBS and then scraped off. After centrifugation, the cell pellet was resuspended with 100 μl of 10 mM HCl and lysed by freezing and thawing using liquid nitrogen three times. The lysate was centrifuged at 16,000 g for 10 min and filtered by a 0.2 μm syringe filter. The cell extracts were analyzed by UPLC.

Statistical Analysis
All experiments were performed in triplicate. Data were analyzed using SigmaPlot software (version 12.5, Systat, USA) and expressed as mean ± SD. The data were analyzed by one-way analysis of variance, and the mean value was considered to be significantly different at p < 0.05, p < 0.005, and p < 0.001.

Synthesis and Identification of Resveratrol Glucosides by DGAS
Resveratrol was glycosylated by DGAS using resveratrol as the acceptor and sucrose as the donor molecule, and the reaction products were detected by TLC and UPLC analyses. On the TLC plate, the spots corresponding to sucrose and fructose were visible, whereas the other spots were possibly the produced glucoside derivatives (Fig. S1). The sugars included in the reaction products were first removed by a reversed-phase C 18 -T cartridge. Then, the mixture of resveratrol glucosides was separated by using a W252 polymeric gel-filtration column through Prep LC. The separated resveratrol glucosides were analyzed by TLC and UPLC.
The three resveratrol glucosides were detected by TLC analysis (Fig. S1, lanes 6-8). The major compound (lane 8) among three resveratrol glucosides was isolated, and the silica gel RP-18 F254S revealed it to be two compounds. The molecular structure of these two compounds was confirmed by NMR analysis. The 1 H and 13 C NMR spectra of these compounds were compared with those of resveratrol. Two anomer proton signals were detected, and the bonds between the glucose moiety and aglycon resveratrol were determined to be α-glycosidic linkages based on the coupling constant ( 1 H, d, J = 3.6 Hz) in the 1 H NMR spectrum. The carbon signal of the NMR results indicated that DGAS transferred a glucose residue of sucrose to the C-4′ or C-3 position in the resveratrol skeleton. Based on the 1 H and 13 C spectra of the major resveratrol glucosides, two isomers were detected and determined to be resveratrol-4′-O-α-glucoside (R1) and resveratrol-3-O-α-glucoside (R2) (Fig. 1). The presence of two resveratrol-α-glucoside isomers (α-piceid isomers) was confirmed by TLC silica gel 60 F254S plate, which is known as an octadecyl silica TLC plate showing two TLC spots (Fig. S1, lane 9). Likewise, the other resveratrol glucosides synthesized by DGAS were determined to be resveratrol-4′-O-α-maltoside and resveratrol-3-O-αmaltoside with α-1,4-glucosidic linkage (data not shown).
The major resveratrol-α-glucoside isomers (R1 and R2) were analyzed by UPLC analysis. Under the reversedphase UPLC condition, two isomer peaks were detected by a UV detector (Fig. S2), and R1 was obtained as the major resveratrol glucoside. The ratio of the peak areas for R1 and R2 in the UPLC profile was approximately 5:1. The conversion yields of the two compounds from resveratrol were 38.7% (R1) and 6.8% (R2).

Physicochemical Properties of α-Piceid Isomers
Phenolic compounds have been studied to increase their applicability in pharmaceutical or cosmetic preparations by improving their low solubility or chemical stability [16][17][18][19]21]. The relative polarity of the αpiceid isomers was evaluated by comparing their retention times with those of β-piceid and resveratrol by UPLC analysis combined with a C 18 column. The retention times for R1, R2, β-piceid, and resveratrol were 0.950, 1.254, 1.255, and 1.904 min, respectively (Fig. S2). Thus, R1 is a more polar compound than R2, β-piceid, and resveratrol. The retention times of R2 and β-piceid were almost the same, and no difference was observed due to the α-or β-  206 mM), respectively. The water solubility of the α-piceid isomers was about 3.6 and 13.5 times higher than that of β-piceid and resveratrol, respectively.
In general, stilbenes, which make up the skeleton of resveratrol, undergo rapid autoxidative degradation. Thus, their half-lives (t 1/2 ) in buffered aqueous solutions are as short as 10 h [26,27]. The oxidative stability of resveratrol and piceid isomers was determined by their half-lives in various solutions. Resveratrol decomposed rapidly at a high pH, whereas it was very stable in acidic conditions, with a half-life of >72 h. By contrast, both α-piceid isomers (R1 and R2) and β-piceid were stable in all the tested solutions ( Table 1). The improved chemical stability of the piceid isomers was due to the glycosylation of the oxidative group of resveratrol, which prevented its oxidative degradation after incubation for 72 h at alkaline pH [17,19].
Vitamin C, resveratrol, and β-piceid removed the FRAP radical quickly (Figs. 2B, S3B). Similar to the DPPH radical-scavenging activity, the FRAP radical-scavenging activity of α-piceid isomers was initially lower than that of β-piceid but increased over time. The radical-scavenging activity of the α-piceid isomers was lower than that of β-piceid at 30 min, attained a value similar at 8 h, and was higher at 24 h. Interestingly, the reaction of FRAP radical scavenging by β-piceid showed no increase over time, unlike other samples.

Effects of α-Piceid Isomers on Melanogenesis
Melanoma cells have been used to evaluate the inhibition of melanin synthesis and cell toxicity [29]. To investigate the structure-activity relationship of resveratrol glucosides for the antiproliferative activity on B16F10 melanoma cells, we examined resveratrol and the piceid isomers by MTT assay. B16F10 cells were incubated with the indicated concentration of the compounds for 1 day. Resveratrol inhibited the growth of B16F10 cells in a dose-dependent manner, but neither β-piceid nor the α-piceid isomers affected the proliferation of B16F10 cells up to 100 μM. However, the α-piceid isomers inhibited the growth of B16F10 cells at concentrations above 100 μM (Fig. S4).
Similar to the results of tyrosinase activity, melanin synthesis was best suppressed by resveratrol in the   piceid, or α-piceid isomers for 48 h and then cellular tyrosinase activity was measured. Significant differences were compared with α-MSH treated cells: **p < 0.005 and ***p < 0.001, significant differences were also observed between β-piceid and α-piceid isomers: # p < 0.05 and ### p < 0.001. compounds used (Fig. 4). Arbutin, β-piceid, and the α-piceid isomers suppressed melanin synthesis by 51.2%, 36.5%, and 50.4% at a concentration of 100 μM, respectively. The values were 74.5% (arbutin), 46.3% (β-piceid), and 72.7% (α-piceid isomers) at a concentration of 300 μM. The α-piceid isomers suppressed melanin synthesis more effectively than β-piceid (Fig. 4). The cellular uptake of each compound was investigated in cell lysates by UPLC. Resveratrol or resveratrol glucosides were added to B16F10 cells, and the concentration of each compound present in the DMEM taken up by the B16F10 cells was then determined after 24 h (Fig. 5). The amount of piceid isomers remaining in the extracellular medium during incubation was similar with the initial amount of each compound after 24 h, whereas the amount of remaining resveratrol was only 59% of the initial amount (Fig. S5). During 24 h exposure of B16F10 cells to piceid isomers, an increase in the intracellular resveratrol content rather than that of piceid isomers was observed indicating that the piceid isomers were converted to resveratrol within the cell. In the case of resveratrol treatment, intact resveratrol was observed in the culture medium and cell lysates [17,19].

Discussion
Resveratrol has various beneficial effects, but its medicinal use is limited because of its low bioavailability [3][4][5][6][7]10]. To synthesize highly soluble resveratrol derivatives as drugs, we reacted DGAS with resveratrol in the presence of sucrose as a donor molecule. When DGAS was used to transglycosylate salicin, baicalein, and catechin, the major transfer products were salicin glycoside, baicalein glycoside, and catechin glycoside, respectively [16,19,21]. Likewise, in our study, resveratrol glucosides, which are α-piceid isomers, were detected by TLC and UPLC analyses, and the chemical structures were identified by NMR analysis. The ratio of the peak areas for R1 and R2 was 5:1. Previously, a glucosyltransferase from Phytolacca americana (PaGT3) expressed in E. coli synthesized resveratrol glucosides with a ratio of trans-resveratrol-4′-O-β-glucoside to trans-resveratrol-3-Oβ-glucoside of 10:3 [35]. From these data, the C-4′ position of resveratrol is the preferable site to conjugate a glucosyl residue by the transglycosylation of glycosyltransferase.
In general, glycosylated compounds have higher water solubility and chemical stability than non-glycosylated were treated with α-MSH alone or together with arbutin, resveratrol, β-piceid, or α-piceid for 72 h and then cellular melanin content was measured. Significant differences were compared with α-MSH treated cells: *p < 0.05 and ***p < 0.001, significant differences were also observed between β-piceid and α-piceid isomers: # p < 0.05 and ### p < 0.001. compounds [16,[19][20][21]. The α-piceid isomers showed 3.6 and 13.5 times higher water solubility than β-piceid and resveratrol, respectively. Interestingly, although R2 and β-piceid showed the same retention time, R2 showed higher water solubility than β-piceid. In the UPLC analysis using the C 18 column, R1 had a shorter retention time than R2 and β-piceid, suggesting that the former is more polar than the latter. However, in the present study, R1 and R2 showed similar water solubilities, suggesting that the difference between α-and β-glucosidic bonds has a greater effect on water solubility than the position of the glucosidic linkage. The stability of polyphenols in buffered aqueous solutions at different pH and in biological fluids, such as plasma, is important for gut absorption given the sharp increase in pH from the acidic stomach to the slightly alkaline intestine. In our study, glycosylation significantly extended the half-life of resveratrol in various buffered solutions, including PBS, SBF, and DMEM. The increased stability observed in our study is consistent with the observed increased stability of baicalein with the addition of one unit of glucose [19]. Therefore, glycosylation helps to protect resveratrol from chemical and enzymatic oxidation. Stilbenes, including resveratrol, exhibit excellent antioxidant activity, and the position of the hydroxyl residue of stilbene affects its antioxidant activity [3,[36][37][38]. The IC 50 values of trans-resveratrol, β-piceid, and resveratrol-4′-O-β-glucoside were 72, 198, and 1,000 μM, respectively, indicating that the C-4′ position of stilbene plays an important role in antioxidant activity [37]. The IC 50 value of the cis-form was also similar to that of the trans-form. Compared with the result of Waffo Teguo et al., the α-piceid isomers were similar to resveratrol, although the initial activities of the α-piceid isomers were low in DPPH and FRAP radical-scavenging activities. These results were due to the differences in the reaction time of antioxidant activity, the location of glycosidic linkage, or conformational differences due to α-or β-glucosidic bonds. Henríquez et al. reported that the extract of blackberries formed the Fe 2+ -TPTZ complex more slowly than the apple pulp [39]. In this study, β-piceid formed the Fe 2+ -TPTZ complex within 30 min, whereas the α-piceid isomers formed the complex very slowly. Whether these results are due to the type or position of the glycosidic linkage is unclear.
α-Piceid and β-piceid, which are glycoside derivatives of resveratrol, showed lower tyrosinase and melanin synthesis inhibitory capability than resveratrol. The hydroxyl group at the C-3, C-4′, and C-5 positions of resveratrol plays a critical role in tyrosinase and melanogenesis inhibitory activity [40,41]. Therefore, the lower melanogenesis inhibitory activity of α-piceid isomer and β-piceid compared with resveratrol possibly occurred because resveratrol loses a hydroxyl group at the C-3 or C-4′ position and gains a glycosyl group. Meanwhile, the α-piceid isomers showed slightly higher activities in inhibiting tyrosinase and melanin synthesis than β-piceid. The stereochemistry of α-and β-linkage also affected the melanogenesis inhibitory activity. Funayama et al. examined the effect of α-and β-arbutin on the activity of tyrosinase from mushroom and mouse melanoma [42]. β-Arbutin inhibited both tyrosinase activity in mushroom and B16 mouse melanoma, whereas α-arbutin only inhibited the tyrosinase activity in the B16 mouse melanoma. The inhibitory activity of α-arbutin was 10 times as strong as that of β-arbutin. The inhibitory mechanism of α-arbutin was speculated to be a mixed-type inhibition, whereas that of β-arbutin was non-competitive. The α-piceid isomers inhibited tyrosinase activity and melanin synthesis less than resveratrol, but the isomers inhibited these processes better than β-piceid in α-MSHstimulated B16F10 cells. The configuration of the anomeric carbon in a sugar of polyphenol glycoside may also affect the tyrosinase activity. To understand the higher tyrosinase and melanin synthesis inhibitory activity of the α-piceid isomers compared to β-piceid, we should separate each compound from the α-piceid isomer mixtures and then investigate which one is more involved in these inhibitions.
The cellular uptake analysis of resveratrol and resveratrol glucosides implied that the glycosylated form of piceid isomers was converted to resveratrol within the cells. The deglycosylation of glycosides in cells is a common phenomenon in the cellular metabolism of various glycoside compounds [17,19]. The previous pharmacokinetic parameter results revealed that glycosylated derivatives were absorbed more rapidly and efficiently than their aglycones [43,44].
In conclusion, resveratrol was converted to resveratrol-4′-O-α-glucoside and resveratrol-3-O-α-glucoside by DGAS. We propose that glycosylated resveratrol increased the bioavailability of resveratrol by protecting this vital molecule from chemical oxidation, thereby extending its half-life in the cell and maintaining antioxidant and melanin synthesis inhibitory properties. We anticipate that resveratrol glucosides may serve as a model for developing analogues for phytochemicals with therapeutic potential in the cosmetic industry.