Transglycosylation of gallic acid by using Leuconostoc glucansucrase and its characterization as a functional cosmetic agent

Materials

Gallic acid, deuterium oxide (D₂O), 1,1-diphenyl-2-picrylhydrazyl (DPPH), 3-(3,4-dihydroxylphenyl)-l-alanine (l-DOPA), mushroom tyrosinase, and β-arbutin were obtained from Sigma-Aldrich (St. Louis, MO, USA). All chemical reagents were commercially available and of analytical reagent grade.

Enzyme preparation

Dextransucrase (EC 3.2.1.11) was obtained from L. mesenteroides B-512 FMCM (KCCM 11728P), which were cultured on LM medium with 2% (w/v) glucose, as previously described (Moon et al. 2007a). The fermented culture was harvested, centrifuged, and concentrated with 30 K hollow fibers (Millipore, Bedford, MA, USA). The enzyme activity was measured at 28 °C with 0.1 M sucrose in 20 mM sodium acetate (pH 5.2) for different reaction periods. The reactants were spotted on a thin-layer chromatography (TLC) silica gel 60 plate (Merck, Darmstadt, Germany) and developed twice in an acetonitrile–water (85:15, v/v) solution. The TLC plate was visualized by spraying with N-(1-naphthyl)-ethylenediamine-H₂SO₄ solution and heating at 121 °C for 10 min. The content of fructose released from sucrose was measured by the evaluation of its density using the NIH Image Program (http://rsb.info.nih.gov/nih-image) with a standard compound. One unit (U) was defined as the amount of enzyme that caused the release of 1 μmol of fructose per minute at 28 °C in 20 mM sodium acetate buffer (pH 5.2).

Synthesis, purification, and identification of gallic acid glucoside

The reactants (1 L), which consisted of 325 mM gallic acid, 355 mM sucrose, and B-512 FMCM dextransucrase (0.55 mU/mL), were incubated in 20 mM sodium acetate (pH 5.2) at 28 °C for 6 h and boiled for 10 min to stop the enzyme action. Glucosylated gallic acid was confirmed by using TLC plate analysis (Merck, Darmstadt, Germany) at 25 °C. The reaction mixtures were placed on TLC plates and developed twice in the following solvent systems: (1) nitromethane/1-propanol/water (2:5:1.5, v/v/v) or (2) ethyl acetate/acetic acid/water (3:1:1, v/v/v) with gallic acid, fructose, and sucrose as the standard materials. Subsequently, the developed plate was visualized by spraying with N-(1-naphthyl)-ethylenediamine-H₂SO₄ solution and heating at 121 °C (Moon et al. 2007a) or UV exposure, as previously described (Seo et al. 2005).

The reaction mixture (1 L) was partitioned with n-butanol to obtain the modified gallic acid products from the upper layer. The modified products were further concentrated under vacuum to 50 mL by using a rotary evaporator (EYELA, Tokyo, Japan) at 47 °C. The sample was applied to a 4.0 × 75 cm silica gel column. After the removal of the remaining sugars with distilled water (total, 2.5 L; flow rate, 1 mL/min), gallic acid glucoside was extracted with 85% (v/v) acetonitrile in water. The compound was purified by high-pressure liquid chromatography (HPLC) under the following conditions: column TSK-GEL amide-80, 5 μm (Waters, Milford, MA, USA); 80% (v/v) acetonitrile in water mobile phase; 1.0 mL/min flow rate; RID-10A RI detector (Shimadzu, Tokyo, Japan).

Purified gallic acid glucoside (2 mg/mL) was mixed with 2,5-dihydroxybenzoic acid (1 mg/mL) in a ratio of 1:1 (v/v), loaded, and dried on a stainless-steel plate at 25 °C. The molecular mass of the sample was measured by MALDI-TOF (Voyager DE-STR, Applied Biosystems, Poster, CA, USA) in a linear mode with delayed extraction (75 laser shots) and an acceleration voltage of 65 kV.

Optimization of gallic acid glucoside production

Antioxidant activity

The antioxidant activities of gallic acid and gallic acid glucoside were detected by using a DPPH scavenging assay (Abe et al. 2000). Samples at each concentration (0.01–2.0 mM) were dissolved in ethanol, reacted with a 0.1 M DPPH reagent for 10 min at 25 °C, and monitored at 517 nm by using a microplate reader (Molecular Devices, Sunnyvale, CA, USA). The radical scavenging activity was expressed as the percentage inhibition of DPPH radical concentration against the reference compound of ascorbic acid. The IC₅₀ value was designated as the concentration of sample that resulted in a 50% reduction in DPPH radicals.

Anti-lipid peroxidation activity

The anti-lipid peroxidation effect was analyzed by ARA-L kit (ABCD GmbH, Berlin, Germany) with an HP-CLA chemiluminescence-measuring device (Tohoku Electronic Industrial, Tokyo, Japan). In accordance with the TIC (thermo-initiated chemiluminescence) method, the antioxidant species in the sample (gallic acid or gallic acid glucoside) were incubated with ample free radical-attached luminol to delay photon generation until the antioxidant species were consumed. The lag time(s) was proportional to the amount of antioxidant species in sample. Each sample (20 μL of 0.5, 1.0, and 5.0 mM) or α-tocopherol (20 μL of 10, 25, 50, and 100 μM/mL) was mixed with a reaction buffer (1.0 mL) at 37 °C, and the effects were measured (Sreejayan et al. 1997).

Tyrosinase inhibition

The incubation mixture consisted of l-DOPA (0–5 mM, l-β-3,4-dihydroxyphenyl alanine) and mushroom tyrosinase (10 U/mL), in the presence or absence of gallic acid or gallic acid glucoside (0–10 mM) as the inhibitor. Ten units of mushroom tyrosinase were used to find the Ki value. The nature of the inhibition was determined by a Dixon plot of the relationship between the reciprocal of the velocity of the reaction and the concentration of the inhibitor (gallic acid or gallic acid glucoside at 0–10 mM) at various substrate concentrations (0.1, 0.5, 1.0, and 5.0 mM l-DOPA). After the mixture was incubated at 37 °C for 10 min, the absorbance was measured at 475 nm by using a microplate reader (Molecular Devices, Sunnyvale, CA, USA), which allowed the calculation of the tyrosinase inhibition (Kim et al. 2010).

MMP-1 production and type 1 procollagen production by enzyme-linked immunosorbent assay (ELISA)

Human newborn foreskin fibroblast cells (HS68) were cultured in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum (Sigma-Aldrich, St. Louis, MO, USA) and 1% antibiotic–antimycotic (Sigma-Aldrich, St. Louis, MO, USA) in a humidified atmosphere with 5% CO₂ at 37 °C. The HS68 cells (ATCC CRL 1635; Rockville, MD, USA) were subcultured in a 1:5 ratio for several passages until they reached 80–90% confluence (Ho et al. 2005; Watanabe et al. 2004). The serum-starved confluent cells were washed twice with phosphate-buffered saline (PBS), left in a thin layer of PBS, and exposed to UVB (100 mJ/cm²) from a UVB lamp (312 nm, Spectroline Model EB-160C, New York, NY). Immediately after irradiation, the cells were washed in a serum-free medium and the response was detected after incubation for 24 h. Prior to UVB irradiation, the cells were pretreated with arbutin (standard, Sigma-Aldrich, St. Louis, MO, USA), gallic acid, and gallic acid glucoside (10–100 μM/mL). The negative control (Control) comprised cells that did not receive UVB exposure and the positive control (UVB) comprised cells that received UVB exposure in the absence of an antioxidant compound (Ho et al. 2005; Watanabe et al. 2004). MMP-1 production was measured by using an ELISA kit (Merck & Co. Inc., Whitehouse Station, NJ, USA), as previously described (Ho et al. 2005). Type 1 procollagen content was measured by using a procollagen type I C-peptide ELISA kit (MK101, Takara, Tokyo, Japan), as previously described (Watanabe et al. 2004). Each sample was measured in triplicate.

Synthesis, purification, and identification

Gallic acid glucoside was detected as a reaction product of dextransucrase with gallic acid and sucrose by TLC and HPLC (Fig. 1). The acceptor reaction mixture was purified by butanol partitioning, which removed the unreacted or hydrolyzed carbohydrates or enzymes present in the lower layer. The upper layer was enriched with gallic acid and the reaction product, gallic acid glucoside. Subsequently, gallic acid and gallic acid glucoside were separated by preparative HPLC using 80% (v/v) acetonitrile in water (Fig. 1c).

The purified gallic acid glucoside was obtained as a brownish-yellow powder with a purified yield of 12.0 g (62% of the total product synthesized). The number of glucose units attached to compounds was verified by using MALDI-TOF MS. The molecular ions of gallic acid glucoside were observed at m/z 355 (M+Na)⁺ (Fig. 1d and Additional file 1: Figure S1). The molecular weight of the compound was increased by adding a glucose moiety to the expected structure via a glycosidic linkage, as shown in Fig. 1d. Our studies revealed that dextransucrase from L. mesenteroides B-512 FMCM could synthesize α-1,3 glycoside linkage with phenolic compound like caffeic acid or astragalin (Nam et al. 2017; Kim et al. 2012).

Optimum gallic acid glucoside synthesis

The production of gallic acid glucoside was optimized by a CCD matrix by using the actual and predicted values, as shown in Table 1. The average amounts of gallic acid glucoside (mM) produced in 20 different experiments are presented in Table 1. The interactions of these variables were evaluated by RSM within the range from − 1.682 to + 1.682 (Fig. 2) and described by a second-order polynomial equation. The response (gallic acid glucoside production) was contained in the following regression equation:

Run no.	Coded levels			Gallic acid glucoside synthesis (mM)
	X1	X2	X3	Actual	Predicted
1	150	300	150	29.0	13/9
2	560	300	150	22.4	10.9
3	150	1000	150	86.4	68.0
4	560	1000	150	80.9	69.0
5	150	300	500	43.5	24.5
6	560	300	500	32.5	19.9
7	150	1000	500	82.9	63.4
8	560	1000	500	78.7	62.8
9	10.2	650	325	21.6	49.5
10	699.8	650	325	30.7	46.5
11	355	61.4	325	0.2	19.8
12	355	1238.6	325	77.4	101.4
13	355	650	30.7	6.2	25.0
14	355	650	619.3	3.9	28.7
15	355	650	325	122.7	103.9
16	355	650	325	92.9	103.9
17	355	650	325	104.6	103.9
18	355	650	325	98.2	103.9
19	355	650	325	114.8	103.9
20	355	650	325	98.2	103.9

Y = − 158.9 + 0.32X₁ + 0.24X₂ + 0.63X₃ − 0.00047X ₁ ²  − 0.00013X ₂ ²  − 0.00089X ₃ ²  + 0.0000139X₁X₂ − 0.000011X₁X₃ − 0.000062X₂X₃

$$ \begin {aligned} {\text{Y}} &= - 1 5 8. 9 + 0. 3 2 {\text{X}}_{ 1} + 0. 2 4 {\text{X}}_{ 2} + 0. 6 3 {\text{X}}_{ 3} \\ & \quad - 0.000 4 7 {{\text{X}}_{ 1}}^{ 2} - 0.000 1 3 {{\text{X}}_{ 2}}^{ 2} - 0.000 8 9 {{\text{X}}_{ 3}}^{ 2} \\ & \quad + 0.0000 1 3 9 {\text{X}}_{ 1} {\text{X}}_{ 2} - 0.0000 1 1 {\text{X}}_{ 1} {\text{X}}_{ 3} - 0.0000 6 2 {\text{X}}_{ 2} {\text{X}}_{ 3} \end {aligned}$$

where X₁ was the sucrose concentration (mM), X₂ was the dextransucrase unit (mU/mL), and X₃ was the gallic acid concentration (mM). The results of the second-order response surface model fitting in ANOVA are presented in Table 1, Additional file 1; Tables S1 and S2. The goodness fit of the model was evaluated by the coefficient R ² . The multiple correlation coefficient was 0.82, which explained 82% of the variation in the response. As “adequate precision value” was indicative of the signal-to-noise ratio index, a value greater than 4 indicated the proper prerequisites for a good fitting model. The adequate precision value of this model was 4.95, which indicated that the model was capable of navigating the design space. The predicted value for gallic acid glucoside production was 104.0 mM; the observed data was 105.2 ± 11.4 mM at 355 mM sucrose, 650 mU/mL dextransucrase, and 325 mM gallic acid, which showed the similarity between the predicted and experimental syntheses of gallic acid glucoside (Table 1 and Fig. 2). The optimum yield for gallic acid glucoside was 114 mM or 35.7% by the reaction of 930 mU/mL dextransucrase with 319 mM gallic acid and 355 mM sucrose.

Antioxidant activity and anti-lipid peroxidation

The antioxidant activities of gallic acid and its glucoside were determined by a DPPH scavenging assay (Fig. 3). The IC₅₀ value of gallic acid glucoside was 0.94 mM, which was sevenfold higher than that of gallic acid (IC₅₀ = 0.13 mM). In the terms of anti-peroxidation activity, gallic acid and gallic acid glucoside exhibited different magnitudes of inhibition of lipid peroxidation (Fig. 4). The use of α-tocopherol (10–100 μM) as a positive control resulted in a dose-dependent increase in chemiluminescent absorbance units (Y = 0.9671x + 2.8027, R² = 0.9889). Gallic acid glucoside showed a significantly (P < 0.05) higher dose-dependent inhibitory effect (19–31%) than gallic acid (Fig. 4).

Tyrosinase inhibition effects

The kinetic studies of both gallic acid and its glucoside were performed by using tyrosinase-inhibitor complexes and different substrate concentrations (Fig. 5). The initial value was measured at various concentrations of substrate [l-DOPA ([S] = 0.1–5.0 mM)] and gallic acid and gallic acid glucoside (0–10 mM). The slope, s, and the vertical axis intercept, i, increased with an increase in gallic acid or gallic acid glucoside content (Fig. 5a and b). The corresponding reciprocal plot was linear and gallic acid and gallic acid glucoside was identified as a mixed noncompetitive inhibition type (Sugimoto et al. 2005). The secondary plots of the slopes and vertical axes intercepts from the Dixon plots against the gallic acid and gallic acid glucoside concentrations resulted in a straight line. The Ki values of gallic acid and its glucoside were calculated to be 1.98 and 1.23 mM, respectively; furthermore, the whitening effect of gallic acid glucoside was greater than that of gallic acid, which had a lower Ki value (Fig. 5).

MMP-1 production and collagen content induced by UV irradiation

Matrix metalloproteinases (MMPs) play an important role in photo-aging through the facilitation of the degradation of extracellular matrix proteins. In particular, MMP-1 and interstitial collagenases have been reported to initiate the degradation of type I and III collagen levels after single or repeated UV exposure in human skin in vivo (Kim et al. 2011). The human-skin fibroblast cells were pretreated with arbutin, gallic acid, or gallic acid glucoside (10 μM/mL) prior to UVB irradiation. MMP-1 production was determined by ELISA kit and its content was normalized to the negative control (100%) or the positive control (200%). Gallic acid glucoside (153%) showed 22% stronger inhibition of MMP-1 than gallic acid (175%) (Fig. 6a). Arbutin showed a slightly higher inhibitory effect (170%) than gallic acid, with lower MMP-1 production. These results indicated that gallic acid glucoside displayed stronger inhibition of MMP-1 production either gallic acid or arbutin.

Collagen alteration has been reported as the primary explanation for skin aging and wrinkle formation (Jung et al. 2015). In addition to MMP-1, type I procollagen levels have also been measured in cells treated with gallic acid, gallic acid glucoside, or arbutin (10–100 μM/mL), as shown in Fig. 6b. When the collagen content was normalized to that of the positive control (100%), the collagen contents of three samples were positively correlated with their treatment concentrations (Fig. 6b). Among the samples, gallic acid glucoside resulted in higher collagen production than that induced by gallic acid or arbutin between 10 and 100 μM/mL. At 100 μM/mL, the treatment with gallic acid glucoside (176%) resulted in higher type I collagen production (46% or 27%), which was higher than that after treatment with gallic acid (130%) or arbutin (149%) (Fig. 6b).