The radix of G. uralensis was collected from Gansu province. Saccharomyces cerevisiae (2.3880) was purchased from China General Microbiological Culture Collection Center (CGMCC). 1,1-Diphenyl-2-picrylhydrazyl radical (DPPH), mushroom tyrosinase, and Folin-Ciocalteu reagent were obtained from Biodee Biotechnology Co., Ltd. (Beijing, China). Matrix metalloproteinase-1 (MMP-1) was purchased from (CUSABIO， No. I21032323). Gallic acid (GAE) and L-DOPA were purchased from Sigma-Aldrich (USA). HPLC-grade methanol was purchased from Merck (Germany). NaNO₂, HCl, Na₂CO₃, AlCl3 and NaOH were purchased from Beijing Chemical Plant （Beijing, China）. Ultra-pure water (18.2 MΩ) was prepared with a Milli-Q water purification system (Millipore, USA). Liquiritigenin (1), isoliquiritigenin (2), liquiritoside (3), isoliquiritoside (4), glabridin (5), glycyrrhizicacid (6), glycyrrhetinic acid (7), and licochalcone A (8), rutin (9) were purchased as reference standards from Beijing Simianti Tech. Co., Ltd. (China). All of the reference compounds were determined to be greater than 98% pure by HPLC analysis. All other reagents were of analytical grade.

Air-dried samples (0.1 kg) were sieved and extracted with 1.5 L of 75% ethanol (v/v) by ultrasonic extraction for 60 min. The extraction was then concentrated and successively partitioned between water and organic solvents to obtain 4 fractions: petroleum (PeF), ethyl acetate (EaF), n-butanol (nBF), and water (WtF). The concentrated extraction was freeze dried and was diluted with water (NFL). Next, 50 mL of each diluted extract was poured into a 250 mL D-flask (pH was adjusted to 9.0) and sterilized for 15 min. The diluted extracts were then added with 1.0 mL inoculated Saccharomyces cerevisiae (1×10⁶ CFU/mL) and incubated at 30℃ for 40 h with mild shaking. The fermenting liquor (FL) was sterilized for 30 min, filtered and freeze dried to the following chemical and bioactivity assays.

Human dermal fibroblasts (HDFs) and B16F10 mouse melanoma cells were purchased from the Cell Bank of the Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences. The cells were cultured in DMEM supplemented with 10% FBS (BioWhittaker, Walkersville, USA), 1% penicillin streptomycin (Gibco BRL, USA), at 37℃ in a humidified atmosphere containing 5% CO₂.

Rates of radical scavenging by the samples were measured by the DPPH method (Rahman et al., 2010). For each sample, an aliquot of 0.5 mL at different concentrations was added to 1.0 mL DPPH solution (100 μM) for 30 min. Methanol was used as a blank solution. The decrease in absorbance at 517 nm was measured. The optical absorption module uses a UV silicon photodiode. DPPH radical scavenging activity was expressed as a percentage (%) of the original absorbance as follows:

DPPH radical scavenging activity=[(ADPPH–AS)/ADPPH]×100

Where AS is the absorbance of the solution after the sample extract has been added, whereas ADPPH is the absorbance of the DPPH solution.

The effects of the extracts on free mushroom tyrosinase activity were determined spectrophotometrically as described previously (Mosmann et al., 1983) with minor modifications. Tyrosinase activity was determined using L-DOPA as a substrate. In brief, 250 µL of 250 units/mL of mushroom tyrosinase in 0.1M phosphate buffer (pH 6.8), 400 µL of the samples at various concentrations and 350 µL of 0.1 M phosphate buffer (pH 6.8) were added to each well of a 96-well plate (TPP, Switzerland) and mixed. The assay mixture was preincubated at 37℃ for 10 min. Next, 400 µL of 2.5 mM L-DOPA in 0.1 M phosphate buffer (pH 6.8) was added to each well and incubation was continued for 25 min at 37℃. The amount of dopachrome formed in the reaction mixture was determined against the blank solution at 475 nm on a microplate reader (Infinite® 200 PRO Tecan Sunrise, Switzerland). The percentage of tyrosinase activity was calculated as follows:

Tyrosinase activity (%)=[(A−B)/(C−D)]×100

where A is the absorbance of the reaction mixture containing the test sample and mushroom tyrosinase; B is the absorbance of the blank sample containing the test sample, but without mushroom tyrosinase; C is the absorbance of the reaction mixture without the test sample and with mushroom tyrosinase; and D is the absorbance of the well without the test sample or mushroom tyrosinase (L-DOPA alone).

Cell viability and proliferation were evaluated using the 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Lee et al., 2006). B16F10 cells （Purchased from China Agricultural Microbial Culture Collection and Management Center）were pretreated with the samples at various concentrations. After incubation for 48 h, MTT solution (final concentration: 5 mg/mL) was added. Next, the cells were incubated at 37℃ for 3 h. Finally, the absorbance of each sample was measured on a microplate reader at 570 nm to obtain the percentage of viable cells.

Cellular tyrosinase activity was measured using a previously described method (Hosoi et al., 1985) with some modifications. Six-well plates (TPP, Switzerland) were seeded with B16F10 melanoma cells (2×10⁵ cells/well in 3 mL medium). The plates were incubated overnight in a humidified CO₂ incubator with a 5% CO₂ atmosphere at 37℃ to allow cells to adhere. Next, the cells were exposed to increasing doses of the test extracts or arbutin for 48 h. After the treatment, the cells were washed with sodium phosphate buffered saline (pH 6.8) and lysed. The lysates were clarified by centrifugation（5424; Eppendorf, Germany） at 12,000 rpm for 30 min at 4℃. The protein concentration of each sample was determined using bovine serum albumin as the standard. The reaction mixture, consisting of 40 µg protein (adjusted to 100 µL with 0.1 M PBS (pH 6.8) and 50 µL L-DOPA (1%)), was added to each well of a 96-well plate. After incubation at 37℃ for 1 h, the absorbance of each well was measured at 475 nm using a microplate reader. Tyrosinase activity was calculated by the following formula:

Tyrosinase activity (%)=OD475_sample/OD475_control×100

Melanin content was measured as described previously with some modifications (Chang et al., 2002). B16F10 melanoma cells were seeded (2×10⁵ cells/well in 3 mL medium) in 6-well culture plates and incubated overnight to allow cells to adhere. The cells were exposed to various concentrations of the test extracts or arbutin for 48 h. At the end of the treatment, the cells were washed with PBS and lysed with 1 M NaOH containing 10% DMSO for 30 min at 80℃. The absorbance was measured at 475 nm using a microplate reader. Melanin content was calculated by the following formula:

Melanin content (%)=OD475_sample/OD475_control×100

Enzyme-linked immunosorbent assays (ELISA) were performed to measure MMP-1 level. HDFs were cultured in 6-well plates (5×10⁵/mL cells) with DMEM containing 10% fetal bovine serum for 24 h, after which the medium was replaced with serum-free medium containing the test samples. After incubation for 24 h, the supernatant was collected from each well and MMP-1 level was measured using an ELISA kit (CUSABIO, China).

The total flavonoid content of the extracts was determined using a previously reported colorimetric method (Singleton et al., 1965). Briefly, 0.4 mL of NaNO₂ (5%, v/v) was added to 1.0 mL of each appropriately diluted sample. To this mixture, 0.4 mL of AlCl3 (10%, v/v) was added after 6 min. Six minutes later, 4.0 mL of NaOH (4%, v/v) and ethanol (60%, v/v) was added to 10 mL of the resulting solution. The absorbance was measured at 508 nm using a microplate reader（TECAN, Shanghai, China after 15 min. A standard curve was constructed using rutin. Total flavonoid content was expressed as mg of rutin equivalent per g of the samples. The samples were analyzed in triplicate.

The total polyphenol content of the extracts was measured over 6 days by Folin-Ciocalteu colorimetry (Fang et al., 2011). Folin-Ciocalteu reagent was prepared by diluting the commercial reagent concentrate with water (1:2). The supernatant of each sample (0.1 mL) was added to 0.1 mL methanol/0.3% HCl (6:4, v/v), after which the resulting solution was added to 10% Na₂CO₃ (2.0 mL). Folin-Ciocalteu reagent (0.1 mL) was added to the mixture, which was then kept in the dark for 30 min. The samples were shaken thoroughly and the absorbance using was measured at 750 nm a microplate reader. The standard curve was constructed using gallic acid (mg GAE g^-1).

The total polysaccharide content of the extracts was determined by the improved phenol-sulfuric acid method (Fang et al., 2011). Measurements were taken at a wavelength of 490 nm using a microplate reader (SpectraMax 190; Molecular Devices LLC, USA). The total polysaccharide contents of the PeF, EaF, nBF, and WtF fractions were calculated as glucose/g using the calibration curve. The samples were analyzed in triplicate.

A small portion of different samples were passed through a 0.45 µm filter, after which a 5.0 µL volume was injected for HPLC/DAD quantitative analysis. Chromatographic separation was carried out using an Agilent 1260LC series system (Agilent Technologies, USA) equipped with an online vacuum degasser, quaternary pump, autosampler, thermostatted column compartment, and diode array detector (DAD). Agilent Technologies Chemstation software for LC (B.02.01) was used. HPLC separation was performed using an Agilent ZORBAX SB-C18 column (4.6 mm×250 mm, 5 µm). The detection wavelength was set to 254 nm. The mobile phase consisted of water/acetic acid (99.8:0.2, v/v) (A) and acetonitrile (B). The gradient program consisted of 20%-30% (B) for 0-10 min, 30%-50% (B) for 10-25 min, and 50%-90% (B) for 25-35 min. The flow rate was 1.0 mL min^-1. The column temperature was set to 30℃. Peak assignments were made for each single-compound injection by comparing their retention times and characteristic absorption spectra from the DAD with those of the authentic standards. The HPLC/DAD method was used to simultaneously determine 8 compounds in the extract fractions, FL, and NFL. The content of each compound of interest was calculated using an appropriate standard curve. Representative chromatograms of the samples are shown in Figure 1.

The statistical significance of the differences between the mean measurements of each treated group and that of the control group were determined using Dunnett's t-test. Results of p<0.05 were considered statistically significant. The data was also analyzed by one-way analysis of variance (ANOVA) using SPSS 17.0 (IBM Corp., USA).

The antioxidant capacities of the 75% ethanol extract of G. uralensis, as well as those of the EaF, nBF, WtF, PeF, FL, and NFL are shown in Figure 2. The DPPH radical scavenging activities of the EaF (IC₅₀=200 µg/mL) and PeF (IC₅₀=370 µg/mL) were significantly higher than those of the total extract (IC₅₀=630 µg/mL), nBF, and WtF. While the yield of the PeF (0.5%) from the total extract was significantly lower than that of the EaF (20.4%). The DPPH radical scavenging activity of the FL (IC₅₀=192 µg/mL) was significantly higher than that of the NFL (IC₅₀=298 µg/mL).

Tyrosinase is an important enzyme in the process of melanin synthesis. The inhibitory activities of the EaF and PeF were significantly greater than those of the nBF and WtF (Figure 3). The extraction yield of the PeF (0.5%) from the 75% extraction was very low, it was concluded that most of the active constituents were extracted into the EaF. In addition, the mushroom tyrosinase inhibition activities of the FL and NFL solutions were not significantly different.

The cell viability of B16F10 cells exposed to the EaF and nBF at concentrations of 50, 100, 200 and 300 µg/mL was tested. As the detection concentrations of EAF and NBF increase, the cell viability of B16F10 melanoma cells gradually decreased, and when the EAF detection concentration was 200 µg/mL, the cell viability decreased to 79.5%; When the concentration of NBF detection was 600 µg/mL, the cell viability decreased to 56.0%.

The EaF showed obvious inhibition of cellular tyrosinase and melanogenesis in B16F10 melanoma cells with concentrations of 50, 100, and 200 µg/mL. EaF significantly reduced tyrosinase activity in B16F10 cells in a dose-dependent manner, with activity of 39.7% at a concentration of 50 µg/mL, 10.4% at 100 µg/mL, and 10.7% at a concentration of 250 µg/mL. Compared with arbutin, EaF at the same concentration produces less melanin and has a better inhibitory effect on tyrosinase activity. At concentrations of 200 and 300 µg/mL, nBF showed no significant inhibitory effect on cell tyrosinase or melanin synthesis (Table 1). The effects of the test samples on the B16F10 cells were ranked similarly to their effects on mushroom tyrosinase and DPPH radical scavenging; therefore, the EaF was determined to be the most active fraction on skin-whitening activities. In HDFs, FL showed more significant inhibition on MMP-1 level in comparison with that of NFL (Table 2).

MMP-1 is considered to be a key cellular regulator that plays a prominent role in the breakdown of the dermal extracellular matrix, namely of collagen types I and III, during the photoaging process, which results in collagen deficiency and leads to wrinkling. Inhibitors of the major collagen-degrading enzymes like MMP-1 could be useful anti-aging agents (Lu et al., 2013). The results of these experiments show that FL has potential as a raw material for anti-aging cosmetics.

The contents of total flavonoids, total polyphenols, and total polysaccharides in the extracts are shown in Table 3.

Flavonoids, phenolic acids, and polysaccharides have been isolated from licorice (Xie et al., 2009). In order to elucidate the main bioactive constituents responsible for the activity of licorice, total flavonoid content, total polyphenol content, and total polysaccharide content were determined. The EaF showed the highest total flavonoid content and total polyphenol content. In addition, FL showed total flavonoid content and total polyphenol content obviously higher than those of NFL. However, the total polysaccharide content of licorice was low. Previous reports have indicated that total flavonoids from licorice have anti-tumor (Zhao et al., 2006), antioxidant (Cui et al., 2007), anti-aging (Ye et al., 2004), skin-whitening, and anti-chloasma effects (Wang et al., 2003). We further quantified the main chemicals in the PeF, EaF, nBF, WtF, FL, and NFL by HPLC/DAD.

Liquiritigenin (1), isoliquiritigenin (2), liquiritoside (3), isoliquiritoside (4), glabridin (5), glycyrrhizic acid (6), glycyrrhetinic acid (7), and licochalcone A (8) were quantified in each fraction, FL, and NFL (Table 4).

Obvious differences were observed among the PeF, EaF, nBF, and WtF. The assayed chemicals were mainly distributed in the EaF and nBF. With the exception of glycyrrhetinic acid (7), all of the chemicals were detected in the EaF. Flavonoids liquiritigenin, isoliquiritigenin, liquiritoside, isoliquiritoside, glabridin, and licochalcone A in the EaF were determined to be the main contributors to the skin-whitening activities of the EaF. Glabridin (5, 46.9 mg g^-1) and licochalcone A (8, 64.9 mg g^-1) were the main chemical constituents of the EaF. Glabridin (5) and licochalcone A (8) have been reported to possess a wide range of biological activities. Glabridin (5) is widely used in cosmetics and has antioxidant (Ablise et al., 2007), melanin synthesis inhibitory, and anti-inflammatory (Yokota et al., 1998) activities. Licochalcone A (8) also has antioxidant and antiinflammatory properties, which may be mediated by inhibition of PGEs, COX-1, and COX-2 (Kolbe et al., 2006). Licochalcone A (8) also inhibited PGE release from HaCaT cells treated with UV radiation (Song et al., 2015). The results described above suggest that glabridin (5) and licochalcone A (8) were the main contributors to the antioxidant, tyrosinase inhibitory, and melanin synthesis inhibitory activities of the EaF. Therefore, the EaF, with high concentrations of glabridin (5) and licochalcone A (8), may have good anti-inflammatory properties.

Obvious differences were observed between FL and NFL. The glabridin (5) and licochalcone A (8) contents were significantly decreased after fermentation, while compounds liquiritigenin (1), isoliquiritigenin (2), liquiritoside (3), isoliquiritoside (4), andglycyrrhizic acid (6) were enriched after fermentation. It could be concluded that the enriched compounds maybe the main contributors to the samples against MMP-1 level. The content of isoliquiritigenin (2), which has been reported to have antitumor (Hsu et al., 2004), anti-oxidant (Chin et al., 2007), and anti-inflammation activities (Kwon et al., 2007), was increased by 38.6% in FL solution. The content of glycyrrhizic acid (6), an anti-inflammatory compound, was increased by 36.8% in FL (Takei et al., 2007). During the process of fermentation, S. cerevisiae expressed some enzymes, including phenylalanine ammonia-lyase (PAL), cinnamate-4-hydroxylase (C4H), 4-coumarate: coenzyme A ligase (4CL), chalcone synthase (CHS), chalconereductase (CHR), chalconeisomerase (CHI), and flavonoid glycosyltransferase (FGT), which participated in the biosynthesis of compounds 1-4 (Figure 4).

The fermented extract of G. uralensis has significant activities on antioxidant and MMP-1 level, and is a promising raw material for anti-aging medicines and cosmetics.