Action of tyrosinase on alpha and beta-arbutin: A kinetic study

The known derivatives from hydroquinone, α and β-arbutin, are used as depigmenting agents. In this work, we demonstrate that the oxy form of tyrosinase (oxytyrosinase) hydroxylates α and β-arbutin in ortho position of the phenolic hydroxyl group, giving rise to a complex formed by met-tyrosinase with the hydroxylated α or β-arbutin. This complex could evolve in two ways: by oxidizing the originated o-diphenol to o-quinone and deoxy-tyrosinase, or by delivering the o-diphenol and met-tyrosinase to the medium, which would produce the self-activation of the system. Note that the quinones generated in both cases are unstable, so the catalysis cannot be studied quantitatively. However, if 3-methyl-2-benzothiazolinone hydrazone hydrochloride hydrate is used, the o-quinone is attacked, so that it becomes an adduct, which can be oxidized by another molecule of o-quinone, generating o-diphenol in the medium. In this way, the system reaches the steady state and originates a chromophore, which, in turn, has a high absorptivity in the visible spectrum. This reaction allowed us to characterize α and β-arbutin kinetically as substrates of tyrosinase for the first time, obtaining a Michaelis constant values of 6.5 ± 0.58 mM and 3 ± 0.19 mM, respectively. The data agree with those from docking studies that showed that the enzyme has a higher affinity for β-arbutin. Moreover, the catalytic constants obtained by the kinetic studies (catalytic constant = 4.43 ± 0.33 s-1 and 3.7 ± 0.29 s-1 for α and β-arbutin respectively) agree with our forecast based on 13 C NMR considerations. This kinetic characterization of α and β-arbutin as substrates of tyrosinase should be taken into account to explain possible adverse effects of these compounds.


Introduction
Tyrosinase (EC 1.14.18.1) is a copper enzyme widely distributed in nature. It is involved in the production of melanin, which causes pigmentation of the skin, protecting it from UV-induced a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Kubo et al. showed that tyrosinase can hydroxylate arbutin, generating 3,4-dihydroxphenyl-o-β-D-glucopyranoside in the presence of catalytic amounts of L-dopa [27,28]. Some studies on the activity of tyrosinase from human malignant melanoma cells demonstrate that αarbutin is more potent than β-arbutin as an inhibitor of the enzyme [29]. Subsequently, the same authors show that α-arbutin does not inhibit the growth of cultured human melanoma cells, HMV-II, but it does inhibit melanin synthesis, meaning that the use α-arbutin in cosmetics is effective and safe for treating hyperpigmentation disorders [30]. Moreover, it was described that α and β-arbutin inhibit the formation of melanin in B16 cells induced by α-MSH and decrease the tyrosinase activity in a cell free system [31]. On the other hand, arbutin derivatives such as deoxyarbutin [32,33] or arbutin undecylenic acid ester [34] were demonstrated to be more potent than α and β-arbutin.
In addition to the applications of arbutins in cosmetics, they also have therapeutic applications such as in the treatment of infections of the urinary tract, and for their antioxidant properties, anti-inflammatory properties and antitumor activity [18].
Regarding the safety of α and β-arbutin in cosmetics, the Scientific Committee on Consumer Safety (SCCS) has stated that the limit in cosmetics should be 2% for face creams and 0.5% in body lotions in the case of α-arbutin, and 7% in face creams for β-arbutin [35,36]. Therefore, although α and β-arbutin are used in cosmetics, their action mechanism needs to be fully understood.
Recently, a study of the effect of α-arbutin on the monophenolase and diphenolase activities of tyrosinase concluded that this compound inhibits monophenolase activity and activates diphenolase activity [37]. In light of the kinetic mechanism for the monophenolase and diphenolase activities of tyrosinase proposed in the bibliography [1], this double effect led us to carry out a deeper study of α and β-arbutin.

Materials
Mushroom tyrosinase (3130 U/mg) was obtained from Sigma (Madrid, Spain) and purified as previously described [38]. Bradford's method was used to determine the protein content using bovine serum albumin as standard [39].

Determination of monophenolase and diphenolase activities
Spectrophotometric assays were carried out with a PerkinElmer Lambda-35 spectrophotometer, online interfaced with a compatible PC 486DX microcomputer controlled by UV-Winlab software, where the kinetic data were recorded, stored, and analyzed.
The diphenolase activity of tyrosinase on L-dopa [40][41][42] and the monophenolase activity on L-tyrosine [43] were measured at 475 nm, the maximum absorption wavelength of dopachrome. To measure the activity of tyrosinase on L-tyrosine, the quantity of o-diphenol necessary to reach the steady state at time t = 0 was added. In this way, the characteristic lag period of the monophenolase activity, which complicates the measurement of the V 0 , was eliminated. This quantity is given by the equation R = [D] ss / [M] ss [41,43] We determined spectrophotometrically the monophenolase and diphenolase activities of tyrosinase acting on substrates that originate o-quinones and do not evolve with a defined stoichiometry. This was done by using MBTH [40,44], which is a potent nucleophile through its amino group which attacks enzyme-generated o-quinones, giving rise to an adduct. This adduct is oxidized by another molecule of o-quinone, leading to the accumulation of o-diphenol in the medium, so, the system reach the steady state. This assay method is highly sensitive, reliable, and precise [45]. MBTH traps the enzyme-generated o-quinones to render a stable MBTH-quinone adduct with a high molar absorptivity. The stability of the MBTH-quinone adducts and the rapidity of the kinetic assays makes this a suitable method for determining the monophenolase and diphenolase activities of tyrosinase [40,41,44,45] (S1A Fig). The sequence of reactions is described in S1B Fig.
All of the assays were carried out at 25˚C, using 30 mM phosphate buffer at pH 7.0. Three repetitions of each experiment were made.
Action of tyrosinase on α and β-arbutin in the presence of hydrogen peroxide Low concentrations of tyrosinase do not show catalytic activity on α and β-arbutin, so, these compounds are described as inhibitors. However, taking into account the action mechanism of the enzyme on monophenols, the action of tyrosinase can be facilitated in the presence of an o-diphenol or H 2 O 2 , which transform E m into E ox . Therefore, the possible reaction of tyrosinase on α and β-arbutin in the presence of H 2 O 2 must be taken into account to confirm the nature of these substrates [46].

Determination of kinetic parameters
Initial rate values (V 0 ) were calculated at different substrate concentrations. The assays were carried out in saturating conditions of O 2 [47][48][49]. The data for V 0 vs. [arbutin] 0 were represented and fitted to the Michaelis−Menten equation using the Sigma Plot 9.0 program for Windows [50], providing the maximum rate (V max ) and the Michaelis constant (K M ).
The degrees of inhibition (i) were calculated using tyrosinase, L-dopa, L-tyrosine and the following formula: i (%) = [(V 0 -V i )/V 0 ] x 100, where V 0 is the initial rate of the control and V i the initial rate in the presence of the target molecule. The initial rates were obtained by linear regression fitting of the initial portions of each experimental recording.

HPLC analysis
The high performance liquid chromatography assays were made using an Agilent 1200 Rapid Resolution coupled with a photodiode detector (UHPLC-DAD).
The samples were filtered to remove possible particles, and injected (20 μl) in a Kinetex Core Shell C-18 column (Phenomenex, Torrance) for reversed phase of 100 x 4.60 mm, 2.6 μm particle size and 100 Å pore size with a flow rate of 1 mL/min. The mobile phase was composed of water (A) and acetonitrile (B), both with formic acid 0.1%, and a multistep linear gradient: 0-30 min, 5-8% B; 30-23 min, 8-95% B; 32-35 min, 95% B. The temperature of the column was maintained at 25˚C.

Computational docking
Molecular docking was carried out around the active site of mushroom tyrosinase with α and β-arbutin as ligands. The chemical structures for α and β-arbutin are available in the PubChem Substance and Compound database [52] through the unique chemical structure identifier CID: 158637 for α-arbutin [53], and CID: 346 for β-arbutin [54]. The molecular structure of tyrosinase was taken from the Protein Databank (PDB ID:2Y9W, Chain A) [55], corresponding to the deoxy form of tyrosinase from Agaricus bisporus. The input protein structure was prepared by adding hydrogen atoms and removing non-functional water molecules. The met and oxy forms of tyrosinase were built by a slight modification of the binuclear copper-binding site as previously described [56]. Rotatable bonds in the ligands and Gasteiger's partial charges were assigned by AutoDockTools4 program [57,58].
The AutoDock 4.2.6 [58] package was used for docking. Lamarkian Genetic Algorithm was chosen to explore the space of active binding to search for the best conformers. The maximum number of energy evaluations was set to 2,500,000, the number of independent dockings to 200 and the population size to 150. Grid parameter files were built using AutoGrid 4.2.6 [59]. The grid box was centred close to the copper ions with a grid size set to 35x35x35 grid points (x, y and z), with grid points spacing kept at 0.375 Å. Other AutoDock parameters were used with default values. PyMOL 1.8.2.1 [60] and AutoDockTools4 [57,58] were used to edit and inspect the molecule structures and docked conformations.

Results
Apparent inhibitory effect of α and β-arbutin on the monophenolase and diphenolase activities of tyrosinase  [1,61]), when α-arbutin is added, the activity rate of the enzyme on L-tyrosine varies (Fig 2A Inset). When the degree of inhibition (i) was calculated, a hyperbole was obtained (Fig 2A). Analogous experiments with β-arbutin were made (Fig 2B Inset), obtaining similar results and a different degree of inhibition ( Fig 2B). Note that big difference between the apparent inhibition values, as is shown in Table 1.
Experiments with the diphenolase activity show similar inhibition ( Table 1, the difference between α and β-arbutin again being of note. Moreover, the degree of inhibition did not reach 100% in either case. The fact that the degrees of inhibition for the monophenolase and diphenolase activities were not the same and that inhibition was not total in either case (α and β-arbutin) leads us propose that arbutins are probably not inhibitors, but alternative substrates [62]. Moreover, experiments with HPLC, as described in Materials and Methods, were made to confirm that there are no secondary reactions, peaks being obtained for α and β-arbutin without mixing with hydroquinone (S2 Fig).
Graphical representations of the Lineweaver-Burk equation for the inhibition of the monophenolase activity by α-arbutin (Fig 4)    Total oxygen consumption test A total oxygen consumption test was made in order to confirm that α and β-arbutin are alternative substrates of tyrosinase.

Action of tyrosinase on α and β-arbutin in the presence of hydrogen peroxide
The action of the enzyme on α and β-arbutin, respectively, in the presence of H 2 O 2 [46] is shown in S6 and S6 Inset Fig. It can be observed that there is catalytic activity on these compounds, in the same way as happens with other alternative substrates [2,9,10]. It must be taken into account that although the activity of the enzyme is almost zero at these concentrations at short times, the addition of hydrogen peroxide transforms E m to E ox (S7 Fig), which is able to hydroxylate arbutin, although the o-quinone that is originated is unstable.
Action of tyrosinase on α and β-arbutin at long measurement times Spectra of the action of tyrosinase on α and β-arbutin are shown in S8 and S9 Figs, respectively, and the formation of an unstable o-quinone can be seen in each case [28]. At short times, there is barely no activity, but a self-activation of the system due to the release of The o-quinones generated by the action of tyrosinase on α and β-arbutin are unstable, as mentioned above. However, they can be attacked by a hydrazone such as MBTH, giving rise to adducts, which are oxidized to become into stable chromophores with a high molar absorptivity (S1A Fig), absorbing between 350 nm and 600 nm. This formation of adducts with MBTH has been used as method to characterize many monophenols and o-diphenols [40,41,44,45]. In the case of α and β-arbutin, unstable adducts are originated at pH = 7, which, as they evolve, give rise to an isosbestic point (Fig 6 and  S10 Fig).  of the corresponding adduct, using different amounts of MBTH. The stoichiometry of the reaction is established from a monophenol as described in S1B Fig. According to the stoichiometry described in S1B Fig, the rate equation for the accumulation of the cromophore with time is [43]: where V A 0 is the initial rate for the accumulation of the cromophore originated by the action of tyrosinase on arbutin, and the kinetic parameters are: K A M = Michaelis constant for α and βarbutin and V A max is the maximum rate, which is equivalent to: Kinetic characterization. V 0 values were calculated taking into account the increase of absorbance with time at λ = 480 nm for α-arbutin and λ = 490 nm for β-arbutin (isosbestic points of the respective adducts) and the molar absorptivity values, which were 22400 M -1 cm -1 and 21200 M -1 cm -1 respectively. K M and k cat values were obtained fitting by non-linear

Molecular docking
Docking complexes between the oxy form of mushroom tyrosinase and α and β-arbutin at the binuclear copper active site of tyrosinase were analyzed. Fig 8 and S11 Fig show the docking poses corresponding to the lowest binding energies at the active site of tyrosinase where catalysis can take place, as mentioned above.
It is interesting that the phenolic groups of α-arbutin (Fig 8) and β-arbutin (S11 Fig) show similar interactions at the catalytic site of tyrosinase. The hydroxyl group could establish hydrogen bonds with the peroxide ion and polar contacts with a copper ion as well as with H259 and H263. However, the aromatic ring position cannot be stabilized by π-π-interactions   Action of tyrosinase on alpha and beta-arbutin with H263 as occurs in many other aromatic ligands [10,11,63]. The ortho carbon is found 3.7 Å from an oxygen atom of the peroxide ion, close enough to allow substrate hydroxylation by the monophenolase activity. Conversely, a clear difference can be seen in the glycosyl moiety orientation of both ligands (Fig 8 and S11 Fig), which form hydrogen bonds with E322 from hydroxyl groups of C4 and C6 of the glucopyranose ring. Moreover, in both of them, there is a polar interaction of the glycosidic oxygen atom of the glucopyranose ring with H244. Nevertheless, only β-arbutin produces an additional polar interaction between the ether oxygen atom of the glucopyranose ring and the same histidine residue.
Docking of β-arbutin to mushroom tyrosinase has previously been reported to occur at the active site, where it interacts with E256 and N260. The discrepancy with our results is due to the different tyrosinase form used by the authors [6]. They used the met form of tyrosinase in contrast with this work where the oxy form was selected for docking purposes since the monophenolase activity resides in this form. The presence of the peroxide ion in the binuclear copper centre requires a different arrangement of the substrate [10].
The dissociation constants, K d , calculated for the docking conformations shown in   (Table 2).

Discussion
The glycosylated derivatives of hydroquinone α and β-arbutin are used as depigmenting agents and their concentrations are regulated by law. In this work, the possible behaviour of these compounds as substrate of tyrosinase is studied.
The experiments described in Figs 2 and 3 demonstrate that α and β-arbutin always act as apparent inhibitors of tyrosinase on the monophenolase and diphenolase activities. The results of these experiments do not agree with those described in the literature, which propose that αarbutin only inhibits the monophenolase activity and activates the diphenolase activity [37]. Such results could be explained if the α-arbutin is partially hydrolyzed and the hydroquinone acts as activator of the diphenolase activity of the enzyme, as, indeed, has been described recently [67]. However, experiments with HPLC show that the samples of α and β-arbutin were not hydrolyzed, so, there is no possibility that this kind of activation occurred in our case (S2 Fig). Furthermore, the IC50 values for monophenolase (Fig 2) and diphenolase (Fig 3) activities were not the same, and neither were the apparent inhibition constants (Fig 4 and S3  Fig), as can be seen in Table 1. These observations suggest that the compounds are alternative substrates of the enzyme. This was lent weight by the oxygen consumption test with TBC ( Fig  5 and 5  The following mechanisms are proposed to explain the action of α and β-arbutin on the monophenolase and diphenolase activities of tyrosinase (S12 and S13 Figs, respectively), based on the above results. The schemes show how α and β-arbutin act as competitive substrates and alternatives to L-tyrosine and L-dopa.
The kinetic analysis of the mechanism is shown in Supporting Information. Despite the complexity of the mechanisms, an equation for the formation rate of dopachrome can be obtained. This equation agrees with the characteristics of the action of a competitive inhibitor: V max does not vary when the concentration of substrate, L-dopa or L-tyrosine, saturates the enzyme. S7 and S12 equations of Supporting Information demonstrate that the same maximum rate is obtained when the concentration of L-tyrosine or L-dopa increases and the concentration of the apparent inhibitor (α and β-arbutin) remains stable (Fig 4 and S3 Fig).
The experiments shown in Figs 5, 5 Inset, S4, S4 Inset, S5 and S5 Inset demonstrate that α and β-arbutin are alternative substrates of tyrosinase and, so, enzymatic activity is originated in the presence of hydrogen peroxide, since this compound gives rise to the formation of oxytyrosinase (S6 and S7 Figs).
Tyrosinase hydroxylates monophenols to o-diphenols through the action of E ox on A, originating the E ox A complex, which becomes E m AOH, which, in turn, can be oxidized giving rise to E d and o-quinone (P) or E m and o-diphenol (AOH). When the initial concentration of enzyme is high, there is sufficient E ox to generate the o-diphenol of the arbutins, which is consumed with time and, so, the decay rate of these o-quinones is greater than the rate of formation and the process stops. In this way, catalytic amounts of released o-diphenol are able to activate the system for long period of time, as can be seen in S8 and S9 Figs, where α and βarbutin, respectively, are consumed by tyrosinase. Note that the respective insets show the instability of the o-quinones generated, again demonstrating that they behave as substrates of the enzyme, as demonstrated previously in the presence of catalytic amounts of L-dopa [28].
The measurements of the initial rate in the isosbestic point (corresponding to the adduct originated by the reaction of α and β-arbutin with MBTH ( Fig 6 Inset and S10 Inset Fig)) were fitted by non-linear regression to Eq 1, thus obtaining the K A M and k A cat values for these substrates ( Table 2).
The distance between the oxygen of the peroxide group and the carbon with the hydroxyl group in ortho position facilitates hydroxylation. Both values are almost the same, although αarbutin has a slightly higher δ 4 value ( Table 2) [68]. β-arbutin has a lower Michaelis constant than α-arbutin ( Table 2). The docking results agree with these values.
The docking results agree quite well with our experimental values for the K M values ( Table  2): β-arbutin exhibits higher binding affinity than α-arbutin. However, both ligands are hydroxylated at essentially the same velocity, since the catalytic rate constants, k cat , were found to be similar ( Table 2). This result also agrees with a similar arrangement of the phenolic group at the binuclear copper centre and with the distance from the ortho carbon to the peroxide ion.
In conclusion, this work demonstrate that α and β-arbutin act as substrates of tyrosinase, since the enzyme is able to hydroxylate them and, subsequently, to oxidize the originated o-diphenol, as well as the hydroquinone, whose quinones are cytotoxic especially when they act on thiol compounds in the melanosome. Such possible adverse effects of α and β-arbutin should be studied in the future.