Tyrosine (98%), dopa (98%), and mushroom tyrosinase (1000U/mg, EC 1.14.18.1) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Mercuric chloride (HgCl₂, 99.5%) and kojic acid (99%) was obtained from Aladdin Bio-technology (Shanghai, China). Histidine (His), valine (Val), glutamate (Glu), and alanine (Ala) were of analytical grade and bought from Xiya Chemical Industry (Shandong, China). All other chemicals and reagents used were of analytical grade or pharmaceutical grade. A Milli-Q-Plus ultra-pure water system from Millipore (Sartorius 611, Germany) was used throughout the study to obtain water used during the experiments.

The method was modified based on the previous description (Chen et al., 2016). In brief, the assays were carried out by using 3 mL of reaction system containing 0.5 mmol/L substrate (tyrosine) in 50 mmol/L Na₂HPO₄-NaH₂PO₄ buffer solution (PBS, pH = 6.8) at 305 K. A 100 μL aliquot of HgCl₂ solutions with various concentrations (0, 1.25, 5 and 10 mmol/L) was added to the reaction system to give concentrations ranged from 0 to 333.33 μmol/L. The final concentration of tyrosinase was 33.33 U/mL. The optical densities (OD) at wavelength of 475 nm were recorded using UV spectrophotometer (UV 2550, SHIMADZU) every 30 s in the first 2 mins and then every 1 min until 30 mins. Kojic acid was used as a positive control. The appropriate blanks (containing both tyrosine and HgCl₂ but without tyrosinase) were tested and subtracted to correct the absorbance.

The method was performed as previously described (Chen et al., 2017). The assays were carried out by using 3 mL of reaction system containing 0.5 mmol/L substrate (dopa) in 50 mmol/L PBS (pH = 6.8) at 305 K. The final concentration of tyrosinase was 13.33 U/mL. A 100 μL aliquot of HgCl₂ solutions with various concentrations (0.15, 0.31, 0.625, 1.25, 5, 10, and and 20mmol/L) was added to the reaction system to give concentrations ranged from 10.42 to 666.67 μmol/L. The OD values of the reaction system in the presence of HgCl₂ solutions (A₁) were recorded at the wavelength of 475 nm using the UV spectrophotometer. The OD values of the reaction system in the absence of HgCl₂ and tyrosinase were used as a positive (A₂) and a negative (A₃) control, respectively. Relative enzymatic activity (%) = (A₁-A₃)/A₂×100%. The experiment was repeated three times. The IC₅₀ value was calculated as the concentration (leading to achieve 50% inhibition) according to the equation, which is obtained by curve fitting of the enzyme activity versus the concentrations of inhibitor.

Experiments were conducted to confirm the inhibitory mechanism of HgCl₂ on tyrosinase furtherly by using immobilized tyrosinase. In brief, we added 200 µL of diluted tyrosinase (2500 U/mL) to each well of a 24-well polystyrene plate and then covered the plate and incubated it at 4°C overnight to produce immobilized tyrosinase. The plate was brought to room temperature, and we thoroughly decanted the solution from wells and washed the wells two times with PBS to remove the unfixed enzyme. We added 3 mL of inhibitor solution to each well and incubated them at room temperature for 30 mins to assure the fully reaction between the inhibitor and the enzyme, discarding the liquid following by washing the wells two times with PBS to remove the excess inhibitor. Then, 3 mL of PBS was added to each well, and they stood for 10 mins to make the dissociation of inhibitor-tyrosinase complex. WE thoroughly decanted the solution from wells to remove the dissociated inhibitor (the procedure was repeated three times). Finally, 3 mL of substrate (dopa) was added to each well and react at room temperature for 30 mins to test the residual activity of tyrosinase. The tested samples included PBS (negative control), 95% ethanol (positive control of irreversible inhibitor), kojic acid (positive control of reversible inhibitor), and HgCl₂, and each sample was tested three times. The concentration of Kojic acid and HgCl₂ was both 666.67 μmol/L.

The non-competitive inhibition type of tyrosinase was assayed by a Lineweaver–Burk plot, which can be described as the following equation:

and a secondary plot is presented below

where ν is velocity of the enzyme reaction calculated by a change in absorbance at the wavelength of 475 nm per minute in the absence and presence of HgCl₂. K _m and K _i are the Michaelis–Menten constant and inhibition constant, respectively. The values of K _m can be calculated from the equation (1). [I] and [S] are the concentrations of inhibitor and substrate, respectively. In. is the intercept value of the Lineweaver–Burk curve. K _i should be calculated by the secondary plot of In. vs. [I].

The fluorescence emission spectra were obtained using Spectrofluorometer FS5 (Edinburgh Instruments, England) equipped with a 150 W xenon lamp and a thermostat bath. We added a 2.0 mL aliquot of tyrosinase solution (400 U/mL) to the quartz cuvette (1 cm path length) and gradually added 2 mmol/L solution of HgCl₂ to give concentrations ranged from 0 to 400 μmol/L. After five-minute incubation periods, fluorescence in the range of 290–500 nm was measured while the excitation wavelength was fixed at 280 nm. All the fluorescence data collected from these experiments were corrected for absorption of excitation light and re-absorption of emitted light according to the following equation: (Wang et al., 2015)

where F _c and F _m represent the corrected and measured fluorescence, respectively. A₁ and A₂ are the absorbance of HgCl₂ at excitation and emission wavelengths, respectively.

Molecular docking study was performed by using AutoDock 4.2.6. The X-ray crystal structure of Agaricus bisporus tyrosinase (PDB ID: 2Y9W) and the 3D structure of HgCl₂ were both retrieved from the RCSB Protein Data Bank (Ismaya et al., 2011). All input files were prepared using an AutoDockTools (ADT) 1.5.4 package, and a charge of +2 was assigned for copper ions. In order to carry out the docking simulation, a 60 Å × 60 Å × 60 Å -point grid with a spacing of 0.375 Å centered at 4.827, 28.489, 92.878 Å was defined, which fully enclosed the catalytic center of tyrosinase. The AutoGrid program was used to construct the grid maps for energy scoring. The three-dimensional location and orientation of the HgCl₂ was explored using the Lamarckian genetic algorithm (LGA). The docking results generated 100 conformations of HgCl₂, which were grouped into clusters by a root mean square (RMS) deviation tolerance of 2.0. The lowest energy (optimal conformation) and the largest number cluster (suboptimal conformation) were chosen for further analyses with the PyMOL molecular graphics system.

According to the results of docking study, we furtherly checked the interactions between HgCl₂ and amino acid. We assumed that if HgCl₂ indeed binds to the amino acid residues of the enzyme, such as His, Val, Glu, and Ala, premixing the amino acid with HgCl₂ would affect the inhibitory effects. Thus, the monophenolase and diphenolase activity of tyrosinase both were determined by adding the mixture of HgCl₂ and amino acid. The mixture was prepared just before use by adding an aliquot of 200 μL HgCl₂ solution to the equal volume of amino acid solution, and which was allowed to react for 10 mins at room temperature. Furtherly, to prove the interaction between His and HgCl₂, the UV-Vis spectra of His, HgCl₂, and the mixture of both were measured. In brief, 5 and 10 mmol/L solutions of His, HgCl₂ were prepared beforehand, and the mixture was produced by mixing the His and HgCl₂ solutions (10 mmol/L) in the volume ratio of 1:1. The spectra of the samples at the wavelengths between 200 and 800 nm were scanned by using UV spectrophotometer (UV 2550, SHIMADZU).The experiment was repeated three times.

Each data point of the experimental results, including tyrosinase activity assay, kinetic analysis for non-competitive type inhibition, fluorescence quenching test, and interactions between amino acid and HgCl₂, was repeated at least three times. The data are presented as the mean ± SD (standard deviation). The statistical significance was determined at the level of P-value (< 0.05) by one-way analysis of variance.

Figures 1A, B showed the kinetic progression of tyrosine oxidation by tyrosinase in the presence of various concentrations of kojic acid and HgCl₂, respectively. With the increase of kojic acid concentration, the lag time was prolonged from 1.44 min to approximately 10.33 min. Similarly, HgCl₂ also exhibited a marked inhibitory effect with significant prolongation of the lag period from 2.00 to 13.4 mins. The inhibitory effects of HgCl₂ on monophenolase were activity dependent on the concentrations because the steady rate (the slope of linear part of the kinetic equation) decreased with the increasing concentration of the inhibitors (Park et al., 2005). The IC₅₀ values of kojic acid and HgCl₂ were 13.10 and 29.97μmol/L, respectively, suggesting that HgCl₂ had less inhibitory capability than kojic acid on monophenolase activity.

Both kojic acid and HgCl₂ significantly inhibited the diphenolase activity in a dose-dependent manner, as shown in Figure 2 . With an increase in the concentration of kojic acid and HgCl₂, the diphenolase activities decreased rapidly with IC₅₀ values of 73.05 and 77.93 μmol/L, respectively. HgCl₂ exhibited significant inhibitory effect on the monophenolase activity also showed the same inhibitory effect on diphenolase activity, indicating it could inhibit the melanin synthesis in the early enzymatic stages. It is well-known that mercury replaces the copper required for tyrosinase activity, but once copper becomes attached to the enzyme it is replaced by ions of mercury with difficulty (Lerner, 1952b). That is to say, mercury was commonly considered to inhibit the melanin production by preventing the formation of the activated tyrosinase via replacement of copper ions. However, in this study, we found that mercury could inhibit both the monophenolase and diphenolase activity of tyrosinase by binding to the enzyme directly instead of replacing the copper ions required for tyrosinase.

The relationship between enzyme activity (expressed as the rate of the oxidation reaction) and its concentration (6.67, 13.33, 20, 26.67, and 33.33 U/mL) in the presence of different concentrations of HgCl₂ (0, 4.17, 16.67, 33.33, and 66.67 μmol/L) was investigated. A series of parallel (or roughly parallel) lines with the same slope and different abscissa intercepts ( Figure 3A ) were observed from the plots of tyrosinase activity versus the concentrations of enzyme in the presence of various concentrations of HgCl₂, suggesting the inhibition of HgCl₂ on tyrosinase was irreversible (Qiu et al., 2005). The increase of abscissa intercept indicated that the number of effective enzyme molecules decreased due to the irreversible inhibition.

In order to prove the irreversible inhibition furtherly, immobilized tyrosinase was used to study the effects of HgCl₂ on tyrosinase, the results of which were shown in Figure 3B . As a negative control, the activity of tyrosinase treated by PBS was defined as 100%. The activity of tyrosinase treated by kojic acid returned to 96.55% because of the reversible inhibition, whilst treated by 95% ethanol only returned to 3.34% because of the irreversible inhibition. The activity of tyrosinase treated by HgCl₂ returned to 28.32%, indicating HgCl₂ indeed inhibits the tyrosinase in an irreversible way compared to kojic acid. However, the enzyme activity is much higher than the one treated by 95% ethanol, which may be due to the fact that the amount of the inhibitor (HgCl₂) is not enough to inactivate the whole enzyme molecules coated on the plate.

The inhibitory effect of HgCl₂ on the diphenolase activity was studied by the Lineweaver–Burk double reciprocal plots. As shown in Figure 4A , the horizontal axis intercept (-1/K _m) keeps the value unchanged and the vertical axis intercept ( 1 + [ I ] K i V max ) increases with the increasing inhibitor concentration, indicating that HgCl₂ induced a non-competitive inhibition (Chen et al., 2005). Therefore, HgCl₂ could bind both free enzyme and the enzyme-substrate complex with the same inhibition constants (K _i). The value of K _i (29.41 μmol/L) was obtained from a plot of the vertical axis intercept (intercept of Lineweaver–Burk double reciprocal plots) versus the inhibitor concentration (Mu et al., 2013), as shown in Figure 4B .

From the result of the intrinsic fluorescence spectra, we observed that HgCl₂ bound to the tyrosinase resulted in conformational changes of tyrosinase: the spectra were gradually decreased in a dose-dependent manner ( Figures 5A, B ). The concentrations of HgCl₂ for the curves (a–f) were 0, 95.24, 181.82, 260.87, 333.33, and 400.00 μmol/L, respectively. Although the decrease in the fluorescence intensity were caused by quenching, there were no significant wavelength shift with the accumulation of HgCl₂, indicating that tyrosinase does not undergo conspicuous overall structural changes. Moreover, the fluorescence intensity decreased with the increasing concentration of the inhibitor because of the tryptophan-masking properties of HgCl₂, which implied that the tyrosinase became disagglomerated, and its structure was loosened (Donghyun et al., 2006).

To identify the interaction mechanism of HgCl₂ on tyrosinase, the data of fluorescence quenching were presented in the form of a Stern-Volmer plot, using the following equation:

where, F ₀ and F represent the steady-state fluorescence intensities in the absence and presence of HgCl₂ (quencher), respectively, [Q] denotes the concentration of HgCl₂, K _q is the quenching rate constant of the biomolecule (K _q = K _sv/τ₀), K _sv is the Stern–Volmer dynamic quenching constant, and τ₀ is the average lifetime of the fluorophore in the absence of quencher (τ₀ = 10^-8 s). The value of K _sv determined by the linear regression plot of F₀/F vs. [Q] at 298 and 304 K were (2.14 ± 0.05) × 10⁴ and 1.83 ± 0.03 × 10⁴ L/mol, respectively. The plot ( Figure 5C ) showed a good linear relationship, indicating that a single type of quenching process (static or dynamic quenching) occurred during the formation of HgCl₂-tyrosinase complex. K _sv values decreased gradually with the increase of the temperatures, which manifested that the fluorescence quenching of tyrosinase by HgCl₂ was a static quenching process (Lin et al., 2018). The corresponding K _q values of (2.14 ± 0.05) × 10¹² and (1.83 ± 0.03) × 10¹² L/mol/s at 298 and 304 K were much higher than the maximum scattering collision quenching constant (2.0×10¹⁰ L/mol/s) (Abouzied and Alshihi, 2008), suggesting that static quenching is dominant in the HgCl₂–tyrosinase interaction (Wang et al., 2014).

For static quenching interactions, if there are similar and independent sites in biological molecules, the apparent binding constants (K _a) and the number of binding sites (n) can be obtained from the following equation:

According to the intercept and slope value of the regression curve ( Figure 5D ), the values of K _a and n for metal–tyrosinase interaction were calculated based on equation (5). The K _a value of (2.87 ± 0.02) × 10⁴ L/mol achieved the order of magnitude of 10⁴ L/mol, indicating the moderate binding of HgCl₂ to tyrosinase (Fan et al., 2017). The results of the section (INHIBITION MECHANISM OF HGCL₂ ON TYROSINASE) showed that HgCl₂ had a significant inhibitory effect on tyrosinase that was irreversible. We suggested that HgCl₂ might be irreversibly coordinated with the amino acid residues of tyrosinase to form the HgCl₂-tyrosinase complexes, resulting in the conformation change of the catalytic center of tyrosinase. Furthermore, the n values (0.96 ± 0.01) were close to one, suggesting that there was a single binding site or a single class of binding sites in tyrosinase for HgCl₂.

Molecular docking was furtherly performed to explore the interaction between HgCl₂ and tyrosinase. We cluster all the conformations (100) with ADT by a root mean square (RMS) deviation tolerance of 2.0, resulting in four clusters, as shown in Figure 6A . The X position of each bar is plotted by the energy of the lowest energy conformation in the cluster, and the height of the bar represents the number of docked conformations in each cluster. The best docking result (optimal conformation) can be recognized as the lowest energy conformation, and can also be selected according to the RMS deviation from the reference structure (usually the crystallographic bonding mode) (Morris et al., 2008). Thus, a lowest energy (-2.87 kJ/mol) conformation in the first cluster is recognized as the optimal conformation, as shown in Figure 6B . A lowest energy (-2.27 kJ/mol) conformation in the fourth cluster was recognized as the suboptimal conformation because the conformation-like one with the ligand (tropolone) binds to the catalytic center of tyrosinase (Ismaya et al., 2011), as shown in Figure 6C . For the optimal conformation, HgCl₂ is surrounded by amino acid residues (His 94, His 296, and Glu 98), whilst the suboptimal conformation is located adjacent to amino acid residues (His 263, Ala 286, and Val 283). The detailed information of interaction between the amino acid residues and HgCl₂ is summarized in Table 1 . As we all know, the catalytic center of tyrosinase is like a pocket, which consists of two Cu ions and six histidine residues, including His 61, His 85, His 94, His 259, His 263, and His 296 (Ismaya et al., 2011). Therefore, it can be inferred that the HgCl₂ may change the morphology of catalytic center by binding with the amino acid residues, such as His 94, His 296, and His 263. Some experiments (in next section) were furtherly designed to support this point.

Amino acids, including His, Val, Glu, and Ala, were checked to see whether they have any impact on the inhibitory effect of HgCl₂ on tyrosinase. It was found that only His exhibited the significant effect on protecting the diphenolase activity in the presence of HgCl₂, as shown in Figure 7A . We also found that His took effects in a concentration-dependent manner, and the protecting capability increased from 48.83 ± 1.89% to 83.43 ± 2.56% with the increasing concentrations ranged from 10.42 to 666.67 μmol/L, which is illustrated in Figure 7B . The effects of His on monophenolase activity in the presence of HgCl₂ were shown in Figure 7C , which demonstrated that His also protect the monophenolase activity when compared to Figure 1B . Val, Glu, and Ala still exhibited no effects on protecting the monophenolase activity (Data no shown). The results demonstrated that HgCl₂ preferred to bind with His rather than other amino acid residues, such as Val, Glu, and Ala. Once HgCl₂ bound to free His, it was difficult to bind it to the residual His, which resulted in the protective effects. In order to further prove the interaction between His and HgCl₂, the UV-Vis spectra of His, HgCl₂, and the mixture of both is presented in Figure 7D ; this shows the significant difference between the mixture and His (or HgCl₂), suggesting the formation of His-HgCl₂ complexes. Thus, based on this indirect and direct evidence (results of docking study), we believe that HgCl₂ exhibits inhibitory effects on tyrosinase by binding to the amino residual (His) of the catalytic center of tyrosinase.