Wells–Dawson phosphotungstates as mushroom tyrosinase inhibitors: a speciation study

In order to elucidate the active polyoxotungstate (POT) species that inhibit fungal polyphenol oxidase (AbPPO4) in sodium citrate buffer at pH 6.8, four Wells–Dawson phosphotungstates [α/β-PV2WVI18O62]6− (intact form), [α2-PV2WVI17O61]10− (monolacunary), [PV2WVI15O56]12− (trilacunary) and [H2PV2WVI12O48]12− (hexalacunary) were investigated. The speciation of the POT solutions under the dopachrome assay (50 mM Na-citrate buffer, pH 6.8; L-3,4−dihydroxyphenylalanine as a substrate) conditions were determined by 183W-NMR, 31P-NMR spectroscopy and mass spectrometry. The intact Wells–Dawson POT [α/β-PV2WVI18O62]6− shows partial (~ 69%) disintegration into the monolacunary [α2-PV2WVI17O61]10− anion with moderate activity (Ki = 9.7 mM). The monolacunary [α2-PV2WVI17O61]10− retains its structural integrity and exhibits the strongest inhibition of AbPPO4 (Ki = 6.5 mM). The trilacunary POT [PV2WVI15O56]12− rearranges to the more stable monolacunary [α2-PV2WVI17O61]10− (~ 62%) accompanied by release of free phosphates and shows the weakest inhibition (Ki = 13.6 mM). The hexalacunary anion [H2PV2WVI12O48]12− undergoes time-dependent hydrolysis resulting in a mixture of [H2PV2WVI12O48]12−, [PV8WVI48O184]40−, [PV2WVI19O69(H2O)]14− and [α2-PV2WVI17O61]10− which together leads to comparable inhibitory activity (Ki = 7.5 mM) after 48 h. For the solutions of [α/β-PV2WVI18O62]6−, [α2-PV2WVI17O61]10− and [PV2WVI15O56]12− the inhibitory activity is correlated to the degree of their rearrangement to [α2-PV2WVI17O61]10−. The rearrangement of hexalacunary [H2PV2WVI12O48]12− into at least four POTs with a negligible amount of monolacunary anion interferes with the correlation of activity to the degree of their rearrangement to [α2-PV2WVI17O61]10−. The good inhibitory effect of the Wells–Dawson [α2-PV2WVI17O61]10− anion is explained by the low charge density of its protonated forms Hx[α2-PV2WVI17O61](10−x)− (x = 3 or 4) at pH 6.8.

Polyphenol oxidases (PPOs) are copper-containing proteins omnipresent in animals, fungi, plants and bacteria [1][2][3][4][5] , with tyrosinases and catechol oxidases being prominent members of this enzyme family. Tyrosinases exhibit cresolase activity (EC 1.14.18.1; ortho-hydroxylation of monophenols to ortho-diphenols, monophenolase activity) and catecholase activity (EC 1.10.3.1; oxidation of ortho-diphenols to ortho-quinones; diphenolase activity), give rise to the rate determining step in melanogenesis and are involved in pigment coating and browning 6 . For its negligible lag phase and generally higher reaction velocity, the diphenolase activity of tyrosinases is usually assayed using L-DOPA (L-3,4−dihydroxyphenylalanine, Figure S1A), as the substrate for the dopachrome assay 7 . The polyoxotungstate (POT) inhibition parameters were determined by fitting the kinetic data to a generalized Michaelis-Menten 8 model and Lineweaver-Burk plots 9 , accompanied by POT speciation studies applying 31 P-NMR and 183 W-NMR spectroscopy and mass spectrometry under physiological conditions (50 mM Na-citrate buffer, pH 6.8) with or without the substrate L-DOPA. A good mushroom PPO inhibitor is the structurally related kojic acid (Fig. S1B), which holds inhibition constant (K i )-values in the µM range 10 and acts as a competitive inhibitor. 10 Mushroom (Agaricus bisporus; abbr: Ab) PPO is present in great quantity in fruiting bodies 11 . The enzyme has been thoroughly characterized in its structure 12,13 and activity 14 . Following the established protocol by Pretzler et al. 15 , AbPPO4 was recombinantly expressed in E. coli and purified in its active form for the here presented inhibition studies.
Polyoxometalates (POMs) are metal-oxygen clusters commonly built up by W, Mo or V addenda ions, which are usually in their highest oxidation states exhibiting the electronic configuration d 0 or d 116, 17 . POMs show a widespread biological applicability such as antibacterial 18 and anti-tumor activity 19  www.nature.com/scientificreports/ successfully been applied as additives in co-crystallization experiments with proteins [20][21][22][23][24] . Apart from the Anderson-Evans archetype 25 6− is stable in aqueous solution between pH 4.5 and 7.5) most POTs show low stability at physiological conditions, requiring solution NMR-measurements to reveal the composition under the applied conditions. As POTs possess a higher chemical stability than polyoxomolybdates (POMos), tungsten is usually selected as the addenda ion, when applying this compound class in biological investigations 26 .
Recently, a systematic approach varying the charge density in a series of Keggin POTs was reported by Breibeck et al. 27 to characterize their inhibitory effects against recombinant AbPPO4 15 . A detailed assignment of the active Keggin POT species was undertaken applying NMR spectroscopy 27

Results and discussion
Activity plots of AbPPO4 inhibited by Wells-Dawson POTs. AbPPO4 was purified and activated according to Pretzler et al. 15 . The ESI-MS of active AbPPO4 is presented in Fig. S2 and Table S1, the protein sequence of AbPPO4 in Fig. S3 4,15 . The diphenolase activity of tyrosinase was monitored using 1 mM L-DOPA (Fig. S1A) as the substrate. Considering that the Wells-Dawson POT stability is pH and possibly time dependent, a speciation study was performed which is detailed in the paragraph "Speciation of Wells-Dawson POT by 183 W-NMR and 31 P-NMR analyses at pH 6.8". The concentration of the investigated Wells-Dawson phosphotungstates was in the range between 0 and 5 mM. If enzymatic inhibition was observed, data were taken in triplicates and fitted with a hyperbolic function (cf. SI Eq. (5)) from a mixed inhibition model to evaluate the K i and α-parameters. All four solutions [P 2 W 18 ] 6 Fig. 2). As a positive control for the inhibition of AbPPO4 diphenolase activity and for validation of the kinetic methodology, the well-characterized natural PPO-inhibitor kojic acid 10 (Fig. S1B) was additionally tested and evaluated applying exactly the same mathematical model. The organic inhibitor kojic  , Be II ), which were obtained under the same conditions, are also in the mM range and vary from 4.7 mM (X = Si IV ) to 25.6 mM (X = P V ) 27 . The K i values obtained using the Michaelis-Menten model, which is the generally accepted method for calculating the enzyme kinetic parameters, are used as primary values to compare the POTs' activities.
To evaluate the inhibitory effect, the activity curves for four Wells-Dawson phosphotungstates were plotted in the concentration range 0-5 mM (Fig. 2). Among the Wells-Dawson clusters, [P 2 W 17 ] 10− showed the greatest inhibitory effect (K i = 6.5 mM) and the fitted α-parameter suggests a mixed mode of inhibition (Table 1). According to their K i values [P 2 W 17 ] 10− (K i = 6.5 mM) and [P 2 W 12 ] 12− (K i = 7.5 mM) exhibited nearly identical inhibitory activity and [P 2 W 18 ] 6− (K i = 9.7 mM) and [P 2 W 15 ] 12− (K i = 13.6 mM) showed a lower inhibition capacity.

Lineweaver-Burk evaluation of inhibition types.
The type of enzymatic inhibition is usually investigated by linear plots according to Lineweaver-Burk, also allowing for further validation of the inhibitory constant K i . 9 For [P 2 W 18 ] 6− , [P 2 W 17 ] 10− , [P 2 W 15 ] 12− and [P 2 W 12 ] 12− as well as for the kojic acid control measured in our previous study 27 , the dopachrome assay was repeated at five different substrate concentrations (varying from 0.4 to 1.5 mM) and three different inhibitor concentrations, respectively (SI section 7, Figs. S14-S17, Tables S4, S5). Each POT analysis yielded a set of three lines intersecting in a common point. The slopes (SI Eq. (13), insets in Figs. S14-S17) and ordinate intercepts of these regression lines were further evaluated to validate the inhibitory constant K i and the α-parameter from the non-linear regression procedure. Therefore, the slopes were plotted against the used inhibitor concentrations to give lines intersecting the abscissa at -K i . Similarly, the Lineweaver-Burk ordinate intercepts were evaluated for the α-parameter (cf. Table S5 and Table 1). In good accordance with their structural similarity, the respective intersection point of the three Lineweaver-Burk lines is the third quadrant for all four Wells- Dawson 10− was detected based on 31 P-NMR peaks integration (Fig. 3A), whereas 48 h after dissolution the concentration of monolacunary anion increased to 69%. The 183 W-NMR spectrum of [P 2 W 18 ] 6− (Fig. S8B) is in agreement with 31 P-NMR data and demonstrates 9 signals for the mono-lacunary anion and 6 additional signals related to a mixture of intact [α/β-P V 2 W VI 18 O 62 ] 6 10− as the only species present in the freshly prepared solution (Fig. 3B), and as the dominant anion (92%) after 48 h of solution aging. The 31 P-NMR spectrum of tri-lacunary POT [P 2 W 15 ] 12− (Fig. 3C) demonstrates the fast rearrangement of the tri-lacunary anion to 62% of monolacunary [α 2 -P V 2 W VI 17 O 61 ] 10− in both fresh and in 48 h aged solutions with 35% of free phosphate anions H x PO 4 x−3 and 3% remaining unidentified phosphotungstates. The 183 W-NMR spectra of monolacunary [P 2 W 17 ] 10− (Fig. S8A) and trilacunary [P 2 W 15 ] 12− (Fig. S8C) are in agreement with 31 P-NMR data confirming the presence of only the monolacunary anion by 9 signals. The 31 P-NMR spectrum of fresh hexalacunary POT [P 2 W 12 ] 12− (Fig. 3D)  . The presence of four different POTs in [P 2 W 12 ] 12− 48 h after preparation renders the detection of all species' signals in the 183 W-NMR spectrum impossible due to their low concentrations and the low abundancy (14%) of the 183 W isotope (Fig. S8D) from the POT clusters as a consequence of partial decomposition or rearrangement (Fig. 3). The weak signals between − 12 and − 7 ppm in the 31 P-NMR spectra (Fig. 3) with a total amount of no more than 19% of all POT signals, correspond to solution intermediates of unknown identity that have been reported previously 44 6− displays notable proton-coupled electron redox activity, which allows this molecule to reversibly accept up to 18 electrons in aqueous solution at pH 4 46 . In contrast, lacunary anions (Fig. 1B-D)

Charge density dependence of inhibitory effects of Wells-Dawson POT species.
The charge density of POMs, expressed in number of charges (q) per addenda metal ions (m), is a criterion to characterize the chaotropic behavior of POMs (Table S6) [49][50][51][52][53] . Recently, it was shown that the affinity of POMs towards biomolecules is attributable to their superchaotropic character, and POMs with moderate charge densities (q/m = 0.33) interact considerably strong with surfaces of different or mixed polarities, which are present in protein molecules 54 10− in Na-citrate buffer at pH 6.8 (Fig. 3, Scheme S1).   10− the q/m ratio is 0.59 (Table 2), however, the protonation of this anion in neutral solutions has previously been shown by electrochemical analysis in combination with theoretical calculations 47,55 . For POMs the proton affinity difference between the terminal and bridging oxygen is around 11 kcal mol −156 , proving that the four oxygen atoms surrounding the lacuna are much more basic than the bridging and terminal O atoms (Fig. 1B), and thus prone to protonation. Electrospray-ionization mass spectrometry (ESI-MS) has already been successfully used for the protonation states assignment of Wells-Dawson POTs containing different heteroatoms 57 . Consequently, ESI-MS has been applied to investigate the protonation state of monolacunary [α 2 -P V 2 W VI 17 O 61 ] 10− . When recording ESI-MS for POMs in Na-citrate buffer, the relative intensity of citrate anion signals is almost 100%, suppressing POM signals. Hence, the measurements had to be carried out in water at pH 6.4 (adjusted with HCl), where [α 2 -P V 2 W VI 17 O 61 ] 10− is the predominant species (Fig. S19), as well as in a CH 3 CN/CH 3 OH/H 2 O mixture. The spectra recorded in both solvents show signals for NaH 4 [α 2 -P V 2 W VI 17 O 61 ] 5at 838.0 and NaKH 3 [α 2 -P V 2 W VI 17 O 61 ] 5at 845.6 m/z, while species with lower protonation states have not been detected (Fig. S19). Therefore, the monolacunary anion [α 2 -P V 2 W VI 17 O 61 ] 10 24−58 q/m = 0.5. Due to the compositional complexity of the hexalacunary anion it is impossible to accurately correlate its activity with the POTs' charge density.

Conclusions
The inhibitory effects of four Wells-Dawson phosphotungstates starting from intact, mono-, three-and hexalacunary forms against AbPPO4 were investigated with a focus on speciation under the dopachrome assay conditions. During the investigation of [P 2 W 18 ] 6

Methods
Protocols for synthesis, spectroscopic assignments and kinetic curves are discussed in the supplementary information. All chemicals have been purchased from Sigma-Aldrich (Vienna, Austria) and Carl-Roth (Karlsruhe, Germany) and were at least of analytical grade. They were used without further purification.
Preparation of AbPPO4 in its active form. For the preparation of active AbPPO4, the procedure published by Pretzler et al. 15 was used. Briefly, latent (inactive) AbPPO4 was expressed with an N-terminal glutathione-S-transferase tag in E. coli BL21(DE3). For the protein expression, ZYM-5052 medium 59 without addition of trace elements was used for 20 h at 20 °C, before 0.5 mM CuSO 4 was added to the medium and incubation was continued for another 20 h. The cells were resuspended in Tris/HCl buffer.
The cell lysis was done via French press. After centrifugation, the supernatant was loaded onto an affinity chromatography GSTrap Fast Flow column and GST-AbPPO4 was obtained after elution with glutathione. The GST tag was cut by digestion with HRV-3C protease. After another affinity chromatography step the latent AbPPO4 was obtained in the flow-through. For activation of latent AbPPO4 for inhibition studies the C-terminal cap was removed from the PPO with Proteinase K and the activated form was purified via size-exclusion chromatography in 50 mM Na-citrate buffer at pH 6.8. www.nature.com/scientificreports/ and with the spectrometer calibrated with the standard tune-mix to give an accuracy of better than 5 ppm in the region of m/z 100-1900.
For a general 1 mL reaction setup in a polystyrene cuvette, 1 µg/mL AbPPO4 (~ 23 nM) was used to catalyze the reaction. The relative inhibition was investigated using 1 mM L-DOPA as the substrate and applying an inhibitor concentration range from 0 to 5 mM to obtain activity curves. For the Lineweaver-Burk plots, a L-DOPA concentration range of 0.4-1.5 mM was assayed for three different POT concentrations, respectively.