Article pubs.acs.org/JAFC
Structure−Activity Relationship Study of Hydroxycoumarins and Mushroom Tyrosinase Shailendra Asthana,∥,† Paolo Zucca,∥,§,# Attilio V. Vargiu,⊥ Enrico Sanjust,§ Paolo Ruggerone,⊥ and Antonio Rescigno*,§ †
Drug Discovery Research Center (DDRC), Translational Health Science and Technology Institute (THSTI), NCR Biotech Science Cluster, Third Milestone, Faridabad-Gurgaon Expressway, Haryana 121001, India § Dipartimento di Scienze Biomediche, Università degli Studi di Cagliari, Cittadella Universitaria, 09042 Monserrato, Cagliari, Italy # Consorzio UNO Università Oristano, 09170 Oristano, Italy ⊥ Dipartimento di Fisica, Università degli Studi di Cagliari, Cittadella Universitaria, 09042 Monserrato, Cagliari, Italy S Supporting Information *
ABSTRACT: The structure−activity relationships of four hydroxycoumarins, two with the hydroxyl group on the aromatic ring of the molecule and two with the hydroxyl group replacing hydrogen of the pyrone ring, and their interactions with mushroom tyrosinase were studied. These compounds displayed different behaviors upon action of the enzyme. The two compounds, arhydroxylated 6-hydroxycoumarin and 7-hydroxycoumarin, were both weak substrates of the enzyme. Interestingly, in both cases, the product of the catalysis was the 6,7-hydroxycoumarin, although 5,6- and 7,8-isomers could also theoretically be formed. Additionally, both were able to reduce the formation of dopachrome when tyrosinase acted on its typical substrate, L-tyrosine. Although none of the compounds that contained a hydroxyl group on the pyrone ring were substrates of tyrosinase, the 3hydroxycoumarin was a potent inhibitor of the enzyme, and the 4-hydroxycoumarin was not an inhibitor. These results were compared with those obtained by in silico molecular docking predictions to obtain potentially useful information for the synthesis of new coumarin-based inhibitors that resemble the structure of the 3-hydroxycoumarin. KEYWORDS: tyrosinase, polyphenol oxidase, hydroxycoumarins, docking, structure−activity, inhibition, melanin
■
INTRODUCTION Polyphenol oxidases are copper enzymes that are widespread among microorganisms, plants, and animals. They catalyze the ortho-hydroxylation of monophenols, which leads to the corresponding o-diphenols (catechols), as well as the oxidation of catechols, which leads to the corresponding o-quinones.1 Under many conditions, these then react further to result in the formation of various pigments. It is important to note that the enzyme nomenclature for tyrosinase varies, depending on the enzyme sources, and includes phenolase, phenol oxidase, and polyphenol oxidase. Whereas polyphenol oxidase (PPO) is perhaps the most generic, we use the acronym MT to specifically differentiate mushroom tyrosinase from Agaricus bisporus. Within its active site, MT contains two nonidentical cupric ions, CuA and CuB, that are each coordinated by three imidazole rings from three histidine residues (see Figure 1A,B). This dicupric cluster is well-defined as a spectroscopic type III dicopper center (electron paramagnetic resonance (EPR) silent in both cuprous and cupric states). Recently, the threedimensional (3D) structure of the A. bisporus enzyme has been reported. 2,3 MT, by far the most studied PPO, is a heterotetrameric protein that is formed by two H subunits (392 amino acid residues) and two L subunits (150 amino acid residues) and has a molecular weight of approximately 120 kDa. The role of the two L subunits is not defined, whereas the two H subunits contain the two dicupric clusters that are responsible for catalytic activity. The orientation of the coordinating histidines depends on the presence of hydrogen © XXXX American Chemical Society
bonds and, in the case of His85, on an uncommon thioether bridge formed by the residue Cys83. The available information about the detailed 3D structure of the H subunit allows for an in-depth study of the interactions between the active site and its substrates/inhibitors. This could result in applications to food technology (such as vegetable browning),4 skin pathologies (such as pigmentation upsets),5 technological and pharmaceutical processes,6,7 and other biotechnology fields. For these reasons, we performed a structure−activity relationship study that included kinetics experiments and computational docking between MT and some selected hydroxy-substituted coumarins. Some coumarin derivatives show bioactive properties.8−10 Hydroxycoumarins (hydroxy1,2-benzopyrones) are common among higher plants, including edible vegetables and fruits, and are well-known to be bioactive.11,12 We chose and tested some commercially available monohydroxycoumarins as putative substrates or inhibitors for MT; their names and chemical structures are summarized in Scheme 1. For each compound, kinetic data of their activity as a substrate or inhibitor were collected and compared with the molecular docking studies. This study aimed to better understand the mechanism of the interactions between monohydroxycoumarins and MT, thereby identifying Received: May 28, 2015 Revised: July 20, 2015 Accepted: July 30, 2015
A
DOI: 10.1021/acs.jafc.5b02636 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
The monophenolase activity using L-tyrosine as its substrate was measured at 468 nm by following the formation of dopachrome with an Ultrospec 2100 Pro UV−vis (Amersham Biosciences) spectrophotometer in the presence of various putative substrates or inhibitors. A control reaction was also conducted. Diphenolase activity was monitored spectrophotometrically by tracking the adduct formation between 4-tert-butyl-1,2-benzoquinone and 4-amino-N,N-diethylaniline at 625 nm (ε625 = 11120 M−1 cm−1).18,19 The percentage of inhibition (I %) of the enzyme activity was also calculated according to the equation I% = (ΔA − ΔB)/ΔA × 100, where ΔA is the difference in the absorbance of the control sample between two incubation times when the absorbance increased linearly and ΔB is the difference in the absorbance of the test sample calculated in the same range of time.20 The results were the average of at least three readings. Other details are given in the text and in the figure captions. HPLC Analysis and Sample Preparation. Identification of the products was performed by HPLC analysis using a Beckman System Gold apparatus that was equipped with an UV−vis detector module. An RPC18 column (Phenomenex, Luna, 250 mm × 3 mm i.d., 5 μm), purchased from Chemtek Analytica (Bologna, Italy), was equilibrated with 0.085% phosphoric acid and 5% methanol in water, v/v (solvent A). Solvent B was 0.085% phosphoric acid and 95% methanol in water, v/v. After an initial isocratic elution (10% B for 10 min), a gradient phase of 10−90%, using solvent B, was applied for 10 min (1 mL/min flow rate). Before the HPLC run was performed, aliquots of the samples were withdrawn at a fixed time, deproteinized by adding phosphoric acid up to 0.085% v/v, and centrifuged at 8000g for 10 min. The supernatants were then filtered using a 0.22 μm pore size membrane. Molecular Docking. Protein Modeling. The X-ray crystal structures of MT (PDB entries 2Y9X and 2Y9W)3 were used as templates for molecular docking studies of the enzyme in complex with our compounds. Note that the crystal structure denoted with the PDB entry 2Y9X contained tropolone, a well-known inhibitor that binds to the active site of the enzyme. The protein structures were prepared using the Amber tool Xleap,21 which included all hydrogen atoms. The orientation of polar hydrogen atoms was optimized, the protein protonation state was adjusted, and each structure was minimized with harmonic restraints on the heavy atoms. Successively, all hetero groups and water molecules were deleted. The structures of the enzyme were compared through rootmean-square distance (RMSD) analysis by superimposing the 2Y9W and 2Y9X structures. The RMSD for Cα is low (∼0.3 Å), which indicates that the structures are similar (Figure 1C). Because no significant difference was observed in the protein structures, particularly in the binding site, the prepared cocrystal structure (after removal of tropolone) was used for docking studies. Ligand Modeling. The lead molecules were first drawn with ACD Chem Sketch 11.0 (http://www.acdlabs.com/resources/freeware/ chemsketch/). The resulting molecular geometry was then optimized at the B3 LYP/6-31G (d, p) level until a convergence, in energy, of 10−5 au using the Gaussian 03 package.22 Docking Protocol. Because no explicit information concerning the interaction between MT and the compounds under investigation has been reported previously, we searched for putative binding sites on the enzyme using two different protocols. These protocols were modified from our previous works.23,24 First, we exploited the available empirical data from the guided docking runs concerning the binding of tropolone to the enzyme.3 To assess the validity of our results, we performed blind docking calculations. More details about the docking steps and binding site detection programs are provided in the next sections. From the ligand clusters that were obtained from blind docking on the protein surface, the top five clusters were chosen for further analysis. We then compared these poses with those obtained from guided docking. Once consensus sites were identified, we performed focused docking (“redocking”) calculations that centered the docking grid on the center of mass of our compounds in the top pose. All docking runs were performed with AutoDock version 4.2.25 This program uses a semiempirical, free energy force field to evaluate
Figure 1. Structural comparison between 2Y9W and 2Y9X: (A) superimposition of 2Y9W (blue) and 2Y9X (white) (binding site residues are rendered in licorice and in an atom-wise type: C, white; O, red; N, blue; S, yellow; whereas tropolone is shown in licorice and in a cyan color); (B) inset view of tropolone binding site; (C) RMSD per residue of the two structures under study.
Scheme 1. Chemical Structure of the Hydroxy-Substituted Coumarins
lead compounds for the synthesis of new potent PPO inhibitors that are suitable for biotechnological applications.13−15
■
MATERIALS AND METHODS
Mushroom Tyrosinase and Enzymatic Assay. MT (5771 U/ mg) was purchased from Sigma-Aldrich (Milan, Italy). A stock solution of the enzyme contained 1 mg/mL. Before use, the enzyme solution was accurately inspected by means of syringaldazine, a typical and sensitive laccase substrate, which allowed us to exclude any laccase contamination. This was necessary because some lots of commercial tyrosinase contain laccase activity.16,17 This could lead to measurements with ambiguous results as contaminant laccases could share the same substrates with tyrosinase (i.e., o-diphenols). A tyrosinase unit was defined according to the manufacturer’s definition and calculated such that 1 unit will cause an increase in A280 of 0.001 per minute at pH 6.5 and 25 °C in a 3 mL reaction mix containing L-tyrosine. All experiments were performed in 100 mM potassium phosphate buffer, pH 6.5, at 25 °C. B
DOI: 10.1021/acs.jafc.5b02636 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry conformations during docking simulations. More details about the formulation of the scoring function and its testing can be found in the manual for the program (http://autodock.scripps.edu/faqs-help/ manual/autodock-4-2-user-guide/AutoDock4.2_UserGuide.pdf, pp 8−11). The binding site is composed of two copper ions that are coordinated by three histidine residues (for CuA, H61, H85, and H94; and for CuB, H259, H263, and H296). The other residues involved in the stability of compound bonding are as follows: C83, F90, E256, N260, H263, M280, G281, S282, and F292 (see Figure 1B). Therefore, in all docking runs, we modeled the flexibility of these residues. In an additional protocol for defining the binding sites, we defined the center of the binding site as the center of mass of the residues surrounding the tropolone (Figure 1B), that is, H61, H85, H259, N260, H263, P264, M280, S282, V283, and A286. Guided Docking. Guided docking is a biased docking procedure, for which we used previously acquired information about the binding of the substrate on the target protein (MT). The availability of the cocrystal structure of MT with tropolone allows us to perform this guided docking. Thus, to check the reproducibility of our docking protocol, we first performed the docking with the tropolone molecule [the complexed structure (PDB-ID 2Y9X)]. In step 1, the tropolone molecule was removed and then redocked. In step 2, the modeled and optimized structure of tropolone (drawn by Chemsketch and optimized through the Gaussian 03 package) was used to perform the docking. The binding pocket of the tropolone was chosen as the grid center in both steps. Grids with dimensions of 68 × 80 × 80 and a spacing of 0.375 Å were used. The Lamarckian genetic algorithm was used for conformational sampling of the compound. For each docking simulation, 200 runs were carried out with 300 random individuals in the first population and 2.5 × 106 energy evaluations. For a local search, the so-called pseudo-Solis and Wets algorithms25 were applied using the default parameter. The low RMSD difference (0.6 Å) between the complex docked-pose of tropolone obtained from the above-described steps 1 and 2 (data not shown) confirmed the reliability of the docking protocol. Furthermore, the optimized structures of our compounds were docked in the same binding site by using the same input parameters. The binding poses of compounds were ranked and clustered on the basis of their energy scores and populations. Blind Docking. This docking protocol is an unbiased docking approach, whereby a grid covers the whole surface of the protein to locate the most likely binding site of lead molecules. AutoGrid 4.0 was employed to build a 126 × 126 × 126 grid map covering the whole RdRp, with grid spacing of 0.586 Å and a center on the center of mass of the whole MT. The Lamarckian genetic algorithm was used for conformational sampling of the compounds. For each docking simulation, 200 runs were carried out with 300 random individuals in the first population, 2.5 × 107 energy evaluations, and 2.7 × 107 generation numbers. For each run, the lowest energy conformation of the compounds was chosen for clustering according to an RMSD cutoff of 2 Å. Clusters were then ranked and clustered on the basis of the AutoDock energy scores and populations. The goal of conducting blind docking was to validate the guided docking outcome: whether the lead substrate candidate sat only in the tropolone-binding site or in some other binding location. Indeed, the top cluster representatives obtained from the blind docking (i.e., the clusters having the lowest docking energy and maximum number of confirmations/population) confirmed that each lead occupied the same binding site.
hydroxylated on the aromatic ring. Of these, only one has a reported interaction with PPOs.26−28 In particular, the inhibitory action of 7-hydroxycoumarin (7-HC, also known as umbelliferone) toward L-tyrosine hydroxylation was noted.27 The MT-catalyzed transformation of L-tyrosine was monitored through the detection of dopachrome formation. As shown in Figure 2, dopachrome is strongly inhibited by 7-HC in a
Figure 2. Tyrosinase-catalyzed oxidation of L-tyrosine in the presence of 7-hydroxycoumarin (7-HC). The reaction mixture contained 3 mM L-tyrosine, 95 enzyme units, and 7-HC substrate at a concentration that ranged from 0.3 to 0.9 mM. Other experimental conditions were 100 mM potassium phosphate buffer, pH 6.5, in a 1 mL final volume and 25 °C temperature.
concentration range of 0.3−0.9 mM. At the maximum 7-HC concentration, the inhibition reached 88%. This phenomenon is due to the specific inhibition of L-tyrosine conversion, which is confirmed by HPLC analysis of the reaction products (see the Supporting Information, Figure SM1) and not due to any interference by a putative quinone arising from 7-HC hydroxylation and oxidation, as previously suggested.26 A further confirmation of the reduced conversion of L-tyrosine can be observed in the decreased appearance of L-3,4dihyroxyphenylalanine (DOPA), which is the product of the reaction (see the Supporting Information, Figure SM2). Moreover, the hydroxylation of 7-HC by MT, which yielded 6,7-dihydroxycoumarin (also known as esculetin), has been observed; however, this occurred only under particular experimental conditions and with a very low kcat.26 Although these observations appear contradictory, they are not surprising. A substantial reduction of the conversion speed of a certain substrate is normally observed in the presence of another substrate, the affinity of which is much higher and having a kcat much smaller than those observed for the first substrate. The second substrate then behaves as a “competitive substrate”, as its enzymatic transformation is too slow to be observed during the planned reaction time for the first enzymatic reaction. However, if the favorable substrate is absent from the reaction mixture, then even the less favorable substrate can be converted by the enzyme. Recently, a method to discriminate between alternative substrates and inhibitors of tyrosinase has been proposed.29 Unlike 7-HC, 6-hydroxycoumarin (6-HC) is rare in nature and has been found in only a few plants30 and as a secondary
■
RESULTS AND DISCUSSION In this study, we explored the effects of four hydroxycoumarins (3-, 4-, 6- and 7 isomers, see Scheme 1) on the activity of MT under different conditions. This will ultimately provide us with mechanistic insights from kinetic data. Hydroxycoumarins with a Hydroxy Substituent on the Aromatic Ring. Two of the four hydroxy-substituted coumarins studied in this work (6- and 7-isomers) are C
DOI: 10.1021/acs.jafc.5b02636 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
of tyrosinase.31 Indeed, when the reaction was conducted in the presence of sufficient concentrations of esculetin (Figure 3B, line b), the disappearance of 6-HC and the formation/ disappearance of esculetin (Figure 3B, line d) accelerated. This confirmed the ability of this compound to play the dual role of both a substrate and an activator of tyrosinase. Hydroxycoumarins with an Unsubstituted Aromatic Ring. 3-Hydroxycoumarin (3-HC) is also rare in nature, but is sometimes found in plants.32 Interestingly, 3-HC was recently demonstrated to protect sea urchin reproductive cells against ultraviolet B damage.33 It has been chosen for the present study because MT should not hydroxylate it. This is because the hydroxyl-bearing ring is not aromatic and the aromatic ring of the compound is unsubstituted. MT is only able to orthohydroxylate phenols and aromatic amines.31,34,35 Furthermore, the close proximity of the carbonyl moiety in 3-HC renders the putative hydroxylation site, the 4-position, relatively electrophilic and therefore not suitable for hydroxylation by MT,36 which performs ortho-hydroxylation through an electrophilic mechanism that requires relatively high electron density on the target carbon atom.37 The HPLC analysis of the products of the reaction of tyrosinase with 3-HC as a possible substrate shows that there was no consumption of 3-HC (see the Supporting Information, Figure SM3). Finally, 3-HC shows the same structural motif (i.e., an enol−ketone) that is typical of some well-known PPO inhibitors, such as tropolone, kojic acid, and mimosine.38 On the basis of these considerations, we tested 3HC as a possible inhibitor of MT and compared its effectiveness with that of tropolone. Figure 4 shows the trend of catalytic consumption of Ltyrosine by MT in the presence of 3-HC or tropolone in comparison to a control. Figure 4A shows that the consumption of L-tyrosine is negligible when the reaction is carried out in the presence of tropolone, whereas in the presence of 3-HC, only approximately 3% of the substrate is consumed after 30 min of incubation. This behavior of the two inhibitors was confirmed by the concentration of L-DOPA in the same samples (Figure 4B), which is below the limit of measurement in the presence of 3-HC and tropolone. This sharp inhibitory behavior of 3-HC on MT follows from the general trend of the structure−activity relationship between MT and phenolics (and phenolic analogs). In contrast, nonsubstituted coumarin (1,2-benzopyrone) was reported to inhibit apple PPO.39 Docking studies performed with this enzyme revealed that the compound is able to enter the enzyme active site, where it is anchored to a specific Ltyrosine residue by a hydrogen bond.39 The alleged complexing ability of coumarin toward Cu2+ is not presently well evidenced and, therefore, should not be considered as a driving force for the observed PPO and coumarin interaction(s). However, we could not find any interaction between the MT and the unsubstituted coumarin. This observation highlights the importance of the hydroxy substituent(s) for a specific interaction with the enzyme active site to occur. We can conclude that the chelating ability of tropolone toward the cupric ions of the dicopper cluster is not important to explain its inhibitory properties: the molecule lays too far from the metal center and is most likely connected to one cupric ion by means of a hydrogen-bond-engaged water molecule. The kinetics of the inhibition of MT by tropolone has been already studied,40 and a specific interaction with the oxy form of the enzyme was proposed to explain the particular behavior of the inhibitor. To establish the true mechanism of
metabolite of an unsubstituted coumarin in certain animals. Yet, like 7-HC, it bears its hydroxyl substituent on the aromatic ring; for this reason, it could be considered a true monophenol derivative. Therefore, given the monophenolase/hydroxylase activity of the enzyme, this compound could behave as a substrate. This was confirmed by the experimental results, with the consumption of 6-HC in the presence of the enzyme. Therefore, 6-HC was found to be a substrate for MT (Figure 3A, line a). 6-HC behaved as a monophenol substrate and was
Figure 3. Tyrosinase-catalyzed oxidation of 6-hydroxycoumarin (6HC) and 6,7-dihydroxycoumarin (esculetin). HPLC measurement was performed in a time course of 6-HC (A) and esculetin (B) concentration in a reaction mixture containing mushroom tyrosinase (95 enzyme units), 0.3 M 6-HC (lines a, c), and 0.3 mM 6-HC plus 10 μM esculetin (lines b, d) as the initial concentrations. Other experimental conditions were 100 mM potassium phosphate buffer, pH 6.5, in a 1 mL final volume and 25 °C temperature.
therefore o-hydroxylated to yield the corresponding o-diphenol. Two different compounds could be the products of 6-HC ohydroxylation: esculetin and 5,6-dihydroxycoumarin. From our experiments, this latter unnatural compound was not detected at all, indicating that 6-HC exclusively was converted into esculetin, (Figure 3B, line c), which is a substrate for the enzyme.27 Indeed, after 60 min, the esculetin concentration decreased as it was enzymatically converted to the corresponding o-quinone. It is already known that low concentrations of diphenols can operate by activating the monophenolase activity D
DOI: 10.1021/acs.jafc.5b02636 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 5. Tyrosinase-catalyzed oxidation of L-tyrosine in the presence of 4-hydroxycoumarin (4-HC). The reaction mixture contained 3 mM L-tyrosine and 95 enzyme units. The 4-HC concentration ranged from 0.3 to 0.6 mM. Other experimental conditions were 100 mM potassium phosphate buffer, pH 6.5, in a 1 mL final volume and 25 °C temperature.
not substantially affected by the presence of 4-HC. In its presence, the formation of dopachrome appears to be slightly increased between the fourth and sixth minutes and then persists for a longer period before it decreases because of nonenzymatic reactions that lead to the formation of melanin. The origin of this behavior is still unknown, but we can assume that, at concentrations much higher than the amount of dopachrome, 4-HC can create an effect of crowding at the catalytic site of the enzyme, thereby slowing the release of the product of the reaction. Yet, a reaction between 4-HC and nonenzymatic transformation products of dopachrome is also possible. A comparison of the overall effect of the examined hydroxycoumarins can be observed in Figure 6. The four compounds, 3-, 4-, 6-, and 7-hydroxycoumarin, were incubated in the presence of MT. We then used L-tyrosine as a substrate and measured the formation of melanin over time. Both 4-HC and 6-HC had no inhibitory effect on enzymatic activity, as the formation of melanin did not decrease. The presence of 7-HC, however, slowed the formation of dopachrome in the early stages of the reaction but failed to prevent melanin formation after an adequate period of time. In contrast, only 3-HC persisted in its inhibitory activity, as the formation of melanin was not observed in the 2 h following the start of the experiment. Moreover, the inhibitory activity of 3-HC is strictly dose-dependent (see the Supporting Information, Figure SM5). Molecular Docking Study. Prompted by these findings, we searched for the putative positions of three compounds within the active site of the enzyme (3-HC, 6-HC, and 7-HC) via in silico molecular docking (4-HC was not considered in the molecular docking investigation, being neither a substrate nor an inhibitor). The results of the docking runs are summarized in Table 1. Concerning the comparatively bulky 3-HC, the molecular modeling investigation resulted in two clusters, of which the representative Ori1 and Ori2 clusters are shown in Figures 7A,B. The two clusters have very different populations (46% of the configurations are in Ori1, whereas 14% of the configurations are in Ori2; see Table 1), and Ori1 has a
Figure 4. Tyrosinase-catalyzed oxidation of L-tyrosine in the presence of 3-hydroxycoumarin (3-HC) and tropolone. HPLC measurements of L-tyrosine (A) and L-DOPA (B) concentrations were performed over time in the presence of 0.5 mM 3-HC and 0.5 mM tropolone and in their absence (Ctrl). Other experimental conditions were 2 mM Ltyrosine, 100 mM potassium phosphate buffer, pH 6.5, in a 1 mL final volume and 25 °C temperature.
MT inhibition by 3-HC, we conducted spectrophotometric experiments to compare the kinetics of tropolone and 3-HC as inhibitors of MT under identical experimental conditions. Whereas tropolone behaved as a slow-binding inhibitor, which confirmed the findings of Espin and Wichers,40 3-HC behaved as a “normal” inhibitor (data not shown). This finding suggests that the two molecules are accommodated differently within the enzyme active site. More specifically, 3-HC most likely interacts with the most abundant form of MT, namely, the met form.1 Remarkably, even the addition of 0.2 mM CuSO4 was not able to remove the inhibition.40 Due to the bonding pattern described above, 4-hydroxycoumarin (4-HC) also should not behave as a substrate for MT. HPLC analysis confirmed that 4-HC did not behave as a substrate for MT (see the Supporting Information, Figure SM4). In Figure 5, the reaction of MT with L-tyrosine in the presence of increasing concentrations of 4-HC is reported. Note that in the first minutes, the formation of dopachrome is E
DOI: 10.1021/acs.jafc.5b02636 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
(flip orientation). In Figure 7C, the two orientations are superimposed upon one another. The H atoms of the hydroxyl groups of the two conformations are directed in an opposing manner, whereas, in both conformations, the oxygen atom of the OH group occupies a position along the location of the oxygen atom of a crystallographic water molecule. Therefore, 3HC accommodates within the MT active site in a manner quite similar to that proposed for mimosine41 and kojic acid.42 With these inhibitors, which are transition state analogues, the oxygen of the hydroxyl group acted as a bridge between the two cupric ions, thus displacing the pre-existing water (or hydroxide) already present in the met form. As a consequence, a stable adduct between the dicupric cluster and the inhibitor molecule is formed, preventing the approach of the substrate to the catalytic site. In principle, both Ori1 and Ori2 poses fit well with such an inhibitory mechanism. However, the distances of the O atom of the OH group from the Cu ions differ in the two poses. The O atom of Ori1 is roughly at the same distance from both Cu ions (2.4 vs 2.6 Å), whereas the O atom of Ori2 is located 2.0 and 3.0 Å away from CuA and CuB, respectively. Ori1 is characterized by the position of the O atom of the OH group, which is more symmetric with respect to the two Cu ions when compared with Ori2. The different population of the two clusters associated with Ori1 and Ori2 suggests that the hydroxyl group interacts with the two Cu ions, thus confirming the bridging role of the oxygen atom of the hydroxyl group as depicted above. Both kinetic and docking data confirm the strict analogy between the inhibitory behavior of mimosine and kojic acid and that of 3-HC. However, despite the structural analogy, tropolone behaves differently, due to its peculiar positioning within the active MT site that follows from the water molecule bridging the inhibitor to the oxy form of the enzyme.40 The Ori2 pose is less symmetrical than the Ori1 and could be less effective as an inhibitor; however, due to the above-described structural and mechanistic reasons, the potential substrate role for this pose is not possible. For the 6-HC substrate, two clusters have been identified that have the lowest docking scoring functions. Interestingly, the highest affinity site obtained from the blind docking protocol is similar to the site obtained from the guided docking protocol (i.e., Ori1 and Ori2 are representatives of the two clusters that are reported in Figure 7D,E, respectively, with Figure 7F displaying both poses). The difference in the scoring function of the two orientations was small, and also the populations of the two clusters did not differ remarkably (29.5 and 38.0% over the total population, respectively, as reported in Table 1). The two poses are characterized by the opposite orientation of the hydroxyl group. In Ori1, the pose that belonged to the cluster with a better scoring function but a slightly lower population, the hydroxyl group is close to the two Cu ions with the OH group at distances of 1.9 and 2.8 Å from CuA and CuB, respectively. The asymmetrical positioning of the hydroxyl group in Ori1 is in agreement with the substrate nature of 6-HC, which can be ortho-hydroxylated to esculetin by the oxy-MT. The hydroxyl group of Ori2 points toward the opening of the active site, and the distances between the CuA and CuB and the closest oxygen atom of Ori2 are 2.1 and 2.9 Å, respectively. It is readily apparent that this pose is unproductive as a MT substrate; however, it could act, in principle, as a slight inhibitor by obstructing the active site. Interestingly, the stability of the poses of this compound appears to be unaffected
Figure 6. Effect of the four studied hydroxycoumarins on the formation of melanin via mushroom tyrosinase in the presence of Ltyrosine as the substrate. The following enzymatic reactions have been performed on a porcelain spot plate: Ctrl, incubation of tyrosinase (46 EU) and L-tyrosine (2 mM) as the control; 3-HC, 4-HC, 6-HC, and 7HC, incubation of tyrosinase (46 EU) and L-tyrosine (2 mM) in the presence of 0.6 mM of each hydroxycoumarin. The enzymatic reaction was buffered with 0.1 M potassium phosphate, pH 6.5. Pictures were taken over the reported time.
Table 1. Results of Molecular Dockinga compound
orientation
scoring function (kcal/mol)
population (%)
3-HC
Ori1 Ori2 Ori1 Ori2 Ori1 Ori2
−7.88 −7.36 −8.29 −8.15 −6.52 −6.28
46.0 14.0 29.5 38.0 17.5 21.5
6-HC 7-HC a
The population is given over the total population of the run associated with each compound.
slightly better scoring function. The two configurations are rotated with respect to one another by an angle of approximately 180° around the major axis of the molecule F
DOI: 10.1021/acs.jafc.5b02636 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
Figure 7. Characterization of the binding site and of representatives of the most populated clusters of the lead molecules (3-HC, upper panels; 6HC, lower panels). Panels A and B show the two representatives (Ori1 and Ori2) of 3-HC (in thick licorice), and panel C shows the combination of the two poses (in solid Ori1, in transparence Ori2). In panel A, the residues of the binding site are shown in licorice and are labeled. In all panels, the Cu ions are colored dark orange and the residues of the binding site are rendered in an atom type manner, such that C is white, O is red, N is blue, and S is yellow. Panels D and E contain the two representatives of the most populated clusters of 6-HC, and panel F collects both poses. The binding site residue’s name and their positions are the same in all of the panels as well as the color and rendering codes.
by the position of the hydroxyl group relative to the active site groups. The docking runs for the 7-HC substrate also yielded two main clusters with populations that were not dissimilar (17.5 and 21.5% for Ori1 and Ori2, respectively, over the total population; Table 1). The two clusters were characterized by the worst scoring functions among the three compounds. This might suggest a poor affinity of the compound for the enzyme. Yet, the 7-HC substrate behaves as an effective “competitive substrate” during MT-catalyzed L-tyrosine conversion to melanin, even though its affinity for the MT active site appears from the docking calculations to be comparatively low in comparison with that of 6-HC. Additionally, the 6-HC substrate is a distinctly more effective MT substrate than the 7-HC substrate (see the Supporting Information, Figure SM6). Moreover, it is unable to reduce the L-tyrosine conversion into melanin (Figure 6). The answer could be that the two top poses found for the 7-HC substrate account for
