Screening and Characterizing Tyrosinase Inhibitors from Salvia miltiorrhiza and Carthamus tinctorius by Spectrum-Effect Relationship Analysis and Molecular Docking

Tyrosinase (TYR) is a rate-limiting enzyme in the synthesis of melanin, while direct TYR inhibitors are a class of important clinical antimelanoma drugs. This study established a spectrum-effect relationship analysis method and high-performance liquid chromatography-mass spectrometry (LC-MS) analysis method to screen and identify the active ingredients that inhibited TYR in Salvia miltiorrhiza–Carthamus tinctorius (Danshen–Honghua, DH) herbal pair. Seventeen potential active compounds (peaks) in the extract of DH herbal pair were predicted, and thirteen of them were tentatively identified by LC-MS analysis. Furthermore, TYR inhibitory activities of five pure compounds obtained from the DH herbal pair were validated in the test in which kojic acid served as a positive control drug. Among them, three compounds including protocatechuic aldehyde, hydroxysafflor yellow A, and tanshinone IIA were verified to have high TYR inhibitory activity (IC50 value of 455, 498, and 1214 μM, resp.) and bind to the same amino acid residues in TYR catalytic pocket according to the results of the molecular docking test. However, the other two compounds lithospermic acid and salvianolic acid A had a weak effect on TYR, as they do not combine with the active amino acid residues or act on the active center of TYR. Therefore, the developed methods (spectrum-effect relationship analysis and molecular docking) could be used to effectively screen TYR inhibitors in complex mixtures such as natural products.


Introduction
Tyrosinase (TYR) belongs to the type 3 copper protein family containing dinuclear copper ions and widely exists in nature from microorganisms to humans [1]. It is a critical enzyme in the synthesis process of melanin pigments, catalyzing orthohydroxylation of monophenols to o-diphenols and then to the corresponding o-quinones [2]. However, overproduction of melanin pigments becomes a problem in the cosmetic and clinical points of view, such as melasma, freckles, and melanosis [3,4]. Due to the importance of TYR during the synthesis of melanin, blocking the activity of TYR is one of the ideal strategies to treat melanin pigment diseases currently. To date, arbutin and kojic acid are the most commonly used tyrosinase inhibitors, which often serve as positive control drugs [5,6]. However, traditional synthetic or microbial origin tyrosine inhibitors have some drawbacks, such as long-term contact of hydroquinone can lead to skin cancer, dermatitis, and other diseases [7,8]. In reality, it is reported that the polyphenols and flavonoids isolated from natural plants have significant tyrosinase inhibition effect and less potential side effects [9,10]. From this point of view, it is of great importance to screen TYR inhibitors from natural products.
Salviae Miltiorrhizae Radix et Rhizoma (Danshen in Chinese, DS) is the dried root or rhizome of Salvia miltiorrhiza Bunge, in which the main components are phenolic acids and diterpenes [11]. Carthami flos (Honghua in Chinese, HH), the dried flower of Carthamus tinctorius L., is generally composed of flavonoids, fatty acids, volatile oils, and polysaccharides [12]. In previous reports, the inhibitory effect of Salvia miltiorrhiza and Carthamus tinctorius on tyrosinase has been validated [13][14][15]; however, the active constituents with tyrosinase inhibition activity have not been clearly reported yet. erefore, in this study, a spectrum-effect analysis method is developed to screen the active constituents that inhibit tyrosinase in Danshen-Honghua (DH) herbal pair. e spectrum-effect relationship analysis combines the chemical compositions of the fingerprint of natural products with the results of the efficacy, and is originally used to develop control standards that can truly reflect the inherent quality of products [16]. Furthermore, spectrum-effect analysis is also used to screen the active components from natural products [17]. In reality, spectrum-effect analysis shows some positive features such as reliability, time-saving capacity, and simple operation [18,19].
In this study, the inhibition effect of DH herbal pair and single drug on tyrosinase was compared first. en, the components in the DH herbal pair are analyzed and identified by HPLC analysis. ird, the active components in DH herbal pair were predicted by spectrum-effect analysis, and their structures were identified by LC-MS analysis. Furthermore, the TYR inhibition activities of the predicted compounds were evaluated in an in vitro model. Finally, molecular docking, which is a method of drug design through the characterization of the receptor and the interaction between the receptor and the drug molecules, and binding mode and affinity prediction [20], was used to confirm the binding sites of compounds with tyrosinase and to predict several possible TYR inhibitors which possess similar structure to the screened active compounds by molecular docking.

Preparation of DH Extracts and Stock
Solutions. All the dried raw DS and HH were pulverized and griddled through 50 mesh sieves (about 0.29 mm) prior to extraction. Seven different proportions of the herbs were prepared with ratios of 1 : 1, 2 : 1, 3 : 1, 5 : 1, 1 : 5, 1 : 3, and 1 : 2 (g/g) DS to HH, respectively. 20 g of DS and HH mixed powder was extracted with 200 mL water in a glass-stoppered conical flask at 75°C for 1.5 h. After extraction, the mixture was filtered through gauze, and the residue was collected and extracted with the above process for a second time.
e two filtrates were combined and evaporated in a rotary evaporator (ZFQ 85A, Shanghai Medical Instrument Special Factory, Shanghai, China) at 55°C under reducing pressure to remove the solvent.
e extracts were further dried by lyophilization with freezing-drying system (DZF-6050, Shanghai Jing Hong Laboratory Instrument Co., Ltd., Shanghai, China) to obtain the DH extracts at a yield of about 25% (w/w, dried extract/crude herb). All pre-and postdilution solutions were stored at 4°C. Before HPLC analysis, the sample solutions were filtered through a 0.22 μm nylon membrane filter (Shanghai Titan Scientific Co., Ltd., Shanghai, China). e reference substances protocatechuic aldehyde, hydroxysafflor yellow A, tanshinone IIA, lithospermic acid, and salvianolic acid A and a positive control (kojic acid) were all prepared by dissolving the respective substance in methanol solution and diluted with PBS (50 mM, pH 6.8) to the required concentrations for TYR inhibitory and binding assay, respectively. All the solutions were stored at 4°C in dark before use.
Shimadzu LC/MS-MS 8060 electrospray ionization-mass spectrometer (ESI-MS), consisting of a Triple Quadruple Detector (TSQ) as the mass detector (Shimadzu, Kyoto, Japan) and coupled with HPLC, was used for LC-MS identification.
e LC conditions were the same as described previously. e ESI-MS conditions were as follows: the ESI was used in both positive and negative mode; nitrogen gas was used for desolvation at a flow rate of 3 L/min at 250°C; the temperature and flow rate of drying gas were set under 400°C and 10 L/min, respectively; the cone voltage was (+) 20 and (−) 20 V; MS data were recorded in the full-scan mode (m/z 50-1500), and MS 2 data were recorded in the range of m/z 50-1200.

TYR Inhibitory Activity
Assay. e enzyme assay was performed in 96-well Corining Costar plates (Corning Incorporated, USA). 50 μL test solution and 50 μL TYR solution (800 U/mL) were mixed and incubated for 10 min at room temperature. After the incubation, 100 μL (1 mg/mL) of L-tyrosine in PBS (pH 6.8) buffer as chromogenic substrate was added to start the reaction. e absorbance was monitored at 490 nm every 30 s for 10 min with an iMark ™ Microplate Absorbance Reader (Bio-Rad Laboratories, Inc., USA). PBS (50 mM, pH 6.8) buffer was prepared as blank control, and kojic acid was used as positive control. TYR inhibition activity was expressed as the inhibitory percentage of TYR: where(dA/dt) blank and (dA/dt) sample are the reaction rate of the blank and sample group, respectively. All trials were independently performed in triplicates, and the results were shown with mean value of the triplicate observations.

Spectrum-Effect Relationship Analysis.
Spectrum-effect analysis was performed by transferring DH fingerprint peak area and TYR inhibition activity test results into SPSS software for canonical correlation analysis (CCA). e optimized HPLC fingerprints of seven ratios DH samples were calculated and generated by professional software named "Similarity Evaluation System for Chromatographic Fingerprint of Traditional Chinese Medicine" composed by Chinese Pharmacopoeia Committee (Version 2012). CCA was used to assess the spectrum-effect relationships between the areas of 86 peaks in fingerprint and the TYR inhibition ratios.

In Silico Molecular Docking of TYR and Identified Active
Compounds. Auto Dock 4.2 program ( e Scripps Research Institute, La Jolla, CA, USA) was employed for in silico molecular docking study to validate the binding potency of the compounds to TYR [21]. e docking operation was performed according to the following steps. First, the crystal structure file of TYR (Agaricus bisporus mushroom tyrosinase) complex (PDB ID � 2y9x) was downloaded [22]. e dimension grid box (90Å × 90Å × 102Å) and the grid spacing of 0.619Å were defined to enclose the active site. Second, the ligand was deleted using UCSF Chimera, and unnecessary water molecules were removed, and hydrogen atoms were added [23]. ird, the 3D chemical structure of investigated compounds was drawn by using Microsoft office 3D and output in PDB format with minimized energy.
With the aim of docking with Autodock Vina, the grid size was set to (x, y, z) � (50, 50, 50) and the grid center was set to (x, y, z) � (10.044, 28.706, 43.443). In each simulation process, progress with default parameters run from Autogrid and Autodock. Lamarckian genetic algorithm (LGA) was used to find the most favorable ligand binding orientations, and the number of LGA runs is equal to 50. e interaction figures were generated, and the results of docking were recorded with binding energies and bonded residues.  Table 1 and Figure 1, PBS served as a blank control, while kojic acid served as a positive control. Some DH herbal pair extracts (1 : 5 and 1 : 3) showed a weaker inhibitory effect than single herbal extracts when the concentrations of tyrosinase and the sample were kept constant. However, other DH herbal pair extracts (1 : 1, 3 : 1, 5 : 1, and 1 : 2) displayed a stronger inhibitory effect than single herbal extracts, which indicated that a synergistic effect of the herbal pair may occur on the inhibition of tyrosinase activity. erefore, DH herbal pair was used as the research object for screening their tyrosinase inhibitors.

Effects of DH Extracts on Tyrosinase Activity. As shown in
In order to obtain the best screening performance for active compounds in the complex matrix, some important parameters of the method including incubation time, TYR concentration, and sample concentration should be optimized. According to previous research, the incubation time might be controlled at the range of 30-120 min because a short incubation time (less than 30 min) might prevent the identification of target molecules which are not firmly bound to TYR, while a long incubation time (about 120 Journal of Analytical Methods in Chemistry 3 minutes) would not make significant influence on the screening result [24,25]. After investigation, a proper TYR concentration (800 U/mL) increased the sensitivity and number of bioactive constituents detected in the sample; meanwhile, the inhibition effect of tyrosinase is most stable by the incubation time of 40 min (data not shown). erefore, a sufficient incubation time (40 min) and a sufficient TYR concentration (800 U/mL) were used in this study.

Spectrum-Effect Relationship
Analysis. e optimized HPLC fingerprints of DH samples with seven ratios are shown in Figure 2. A total of 86 peaks involved were detected in the   Figure 2: HPLC calibration fingerprints of DH herbal pair with different ratios. e chromatograms of S1-S7 are represented as follows: DH 1 : 1 (S1); DH 2 : 1 (S2); DH 3 : 1 (S3); DH 5 : 1 (S4); DH 1 : 5 (S5); DH 1 : 3 (S6); DH 1 : 2 (S7); control map (R).   calculation of spectrum-effect relationship. CCA (canonical correlation analysis) was used to assess the spectrum-effect relationship between the areas of 86 peaks and the main parameters (inhibition rate), and the results are shown in Table 2 Figure 3. Five pure reference compounds including protocatechuic aldehyde, tanshinone IIA, hydroxysafflor yellow A, lithospermic acid, and salvianolic acid A were obtained for further in vitro activity tests.

In Vitro Activity Tests for the Predicted Compounds.
To confirm the ability of the hit compounds with TYR inhibitory activity, in vitro enzymatic activity assays were performed. Five concentrations of each compound were tested, and the results are shown in Table 4. As a well-known TYR inhibitor [5], kojic acid showed strong inhibition effect with a IC50 value of 127 μM. From the results shown in Figure 4, among the five identified hit compounds, protocatechuic aldehyde, hydroxysafflor yellow A, and tanshinone IIA possessed strong TYR inhibition effects in a dose-dependent manner, with the IC50 values of 455, 498, and 1214 μM, respectively. However, lithospermic acid and salvianolic acid A did not show   significant inhibitory effect on tyrosinase at a relatively high concentration (1.6 mM). e reason may be that they do not combine with the active amino acid residues or do not act on the active center of TYR, but further studies are required.

Molecular Docking of TYR and Identified Active
Compounds. Molecular docking can be used to study the binding mechanism of compounds interacting with proteins. In this study, Autodock 4.2 was selected as the docking software to check out the active site of those components screened from DH extracts combined with TYR. e crystal structure of tyrosinase includes four identical parts, and one of them was used as the crystal structure of tyrosinase for computational docking analysis. Interestingly, as shown in Figures 5(a) and 5(b), protocatechuic aldehyde and hydroxysafflor yellow A could bond to the same catalytically active amino acid residues (THR324, ASN81, CYS83, GLU322, and HIS85) of TYR, which could explain the similar TYR inhibitory activity of these two compounds. Moreover, three hydrogen bonds (green line) between protocatechuic aldehyde, hydroxysafflor yellow A, and the amino acid residues were observed. As shown in Figure 5(c), tanshinone IIA was found bonding into the hydrophobic cavity of tyrosinase (blue region) and surrounded by amino acid residues VAL248, HIS244, OTR410, VAL283, SER282, PRO277, PHE264, and ARG268 of tyrosinase. e amino acid residues of the tyrosinase to which the active compound binds are shown in Table 5. e reason for the weak activity may be that lithospermic acid and salvianolic acid A do not combine with the active amino acid residues (such as GLU322 and THR324) or do not act on the active center (Cu 2+ ) of tyrosinase. As tyrosinase is a copper-containing enzyme, it is expected that potential tyrosinase inhibitors should show high binding affinity for copper ions [26].
Along with the screening result, molecular docking was also carried out on compounds (absence of pure reference substances) having similar structure to protocatechuic aldehyde, lithospermic acid, and salvianolic acid A (the structure of these compounds are shown in Figure 4), attempting to explore other active TYR inhibitors by perspective of structure-activity relationship. As shown in Figure 5(a), the important region of protocatechuic aldehyde for copper chelation was the catechol structure. Compounds that functionally chelate copper ions at tyrosinase active sites have been frequently reported as effective tyrosinase inhibitors because they are analogous to the phenolic hydroxyl substrates of tyrosinase [27,28]. In addition, danshensu, caffeic acid, rosmarinic acid, and salvianolic acid E having a catechol structure might also be potential TYR inhibitors. e docking result is shown in Table 5. With a portion of the same active sites with screened active compounds bound to TYR, it could be found that danshensu, caffeic acid, rosmarinic acid, salvianolic acid E, anhydrosafflor yellow B, 6-hydroxykaempferol-3,6,7-O-β-D-glucoside, and 6hydroxykaempferol 3-O-rutinoside-6-O-glucoside similar to protocatechuic aldehyde, hydroxysafflor yellow A, and tanshinone IIA might also be potential TYR inhibitors.

Conclusions
In this study, the TYR inhibitor screening methods were established, and the potential TYR inhibitory components from DH extract were screened. Combining the results of spectrum-effect analysis, LC-MS analysis, and enzymatic activity assay, three active compounds including protocatechuic aldehyde, hydroxysafflor yellow A, and tanshinone IIA were discovered as inhibitors targeting TYR. Meanwhile, docking results showed that these compounds might bind to the same amino acid residues in TYR catalytic pocket. Additionally, other potential active TYR inhibitors such as danshensu, caffeic acid, rosmarinic acid, and salvianolic acid E, which gained similar structures with the hit compounds, might also be identified. ese results proved that the proposed method could effectively screen TYR inhibitors in complex mixtures and provided a reference for the discovery of other active TYR inhibitors. Data Availability e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declared that they have no conflicts of interest. Journal of Analytical Methods in Chemistry 9