Keggin-type polyoxotungstates as mushroom tyrosinase inhibitors - A speciation study

Mushroom tyrosinase abPPO4 is a commercially relevant polyphenol oxidase and has been being targeted for numerous inhibition studies including polyoxometalates (POMs). In the present work, its diphenolase activity was inhibited at pH 6.8 by a series of structurally related polyoxotungstates (POTs) of the α-Keggin archetype, exhibiting the general formula [Xn+W12O40](8−n)− in order to elucidate charge-dependent activity correlations. Kinetic data were obtained from the dopachrome assay and 183W NMR was applied to obtain crucial insights into the actual Keggin POT speciation in solution, facilitating a straightforward assignment of inhibition effects to the identified POT species. While [PW12O40]3− was completely hydrolyzed to its moderately active lacunary form Hx[PW11O39](7−x)− (Ki = 25.6 mM), [SiW12O40]4− showed the most pronounced inhibition effects with a Ki of 4.7 mM despite of partial hydrolysis to its ineffective lacunary form Hx[SiW11O39](8−x)−. More negative Keggin cluster charges of 5− and 6− generally resulted in preclusion of inhibitory efficacy as well as hydrolysis, but with the Ni-substituted cluster [PW11O39{Ni(H2O)}]5− enzymatic inhibition was clearly restored (Ki = 9.7 mM). The inhibitory capacity of the structurally intact Keggin POTs was found to be inversely correlated to their net charge. The here applied speciation strategy is of utmost importance for any biological POM application to identify the actually active POM species.


Preparation of abPPO4 in its active form
. SDS-PAGE of the chromatographic purification of recombinant abPPO4 in its active form according to Pretzler et al. [1] . For preparation of this figure, the gel was slightly cropped at the edges (see Figure S24 for the original image). Lane M: Precision Plus Protein TM Standard Dual Color (Bio-Rad Laboratories), 1: GST-PPO4(latent) fusion protein after first GST chromatography, 2: GST-PPO4(latent) cleavage reaction with HRV 3C protease, 3 and 4: flowthrough with free PPO4(latent) after second GST chromatography, 5: eluate of second GST chromatography, 6: PPO4(latent) cleavage reaction with Proteinase K, 7 and 8: PPO4(active) after SEC chromatography. Calculated molecular weights: GST-PPO4(latent), 91.5 kDa; PPO4(latent), 65.1 kDa; PPO4(active), 44.5 kDa; GST-Tag, 26.4 kDa. Figure S2. Summary of the NMR speciation of non-substituted Keggin POT anions performed in this study. The higher the negative net charge of the POT anion, the more stable against hydrolysis it appears to be at the given pH 6 The pH-dependent speciation of the non-substituted Keggin POTs under investigation according to the extensive NMR analyses shown in section 2.2.2 below is presented in Figure S2. The oxidation state of the incorporated heteroatom determines the negative net charge of the POT anion. With increasing negative charge, hydrolytic decomposition is prevented more efficiently. Figure S3. Five structural isomers of the Keggin archetype of polyoxometalates. The α-isomer is the energetically most favoured and synthetically best accessible isomer and was therefore used as a common scaffold for all the POTs in this study. In accordance with the tetrahedral symmetry of the α-Keggin polyoxometalate archetype, the cluster can be grouped into four {M 3 O 13 } triads. Rotation of one triad about 60° results in another isomeric structure. The γ-, δ-and ε-Keggin isomers are usually reported only for mixed addenda atom compositions or for structures stabilized by covalent modification with organic moieties or 3d-and 4-f metals. There are synthetic procedures for the preparation of β-isomers, but only for a limited selection of different heteroatoms. The α-Keggin isomer is the most stable form for its highly symmetrical architecture and is accessible for a large array of heteroatoms, which is why it served the basic POT structure throughout this study [2] .

Characterization of α-Keggin POTs
Previous to the enzymatic inhibition studies, the Keggin POTs were verified by IR spectroscopy (see section 2.2.1). In the course of NMR speciation (see section 2.2.2), all the POT anions were confirmed to correspond to the α-isomer. Only the [BeW 12 ] 6preparation was revealed as a mixture of the α-and β-isomer, however, leaving the experimental consistency and data interpretation unaffected due to direct comparability with [H 2 W 12 ] 6-. As neither 183 W nor 31 P NMR provided unambiguous evidence for the substituted POT [PW 11 Ni] 5-, additional ESI-MS was performed to directly prove the intact cluster.  Figure S4. IR spectra of [XW 12 ] n-(X = H 2 2+ , Be 2+ , B 3+ , Al 3+ , Si 4+ , P 5+ ) and [PW 11 Ni] 5in the region of W-O-W and W=O bonds vibration (1000 -400 cm -1 ).  Figure S6. 27 Figure S7. 31 P NMR spectra of [PW 12 ] 3-A) at pH 1.5 and B) at pH 6.8 with 10% D 2 O. No intact Keggin anion [PW 12 ] 3was detectable at pH 6.8 (cf. Figure 4D). The lacunary anion [PW 11 ] 7is depicted in colours representing the symmetry-equivalent W positions in order to illustrate the less symmetric environment experienced by the P center. C) 31 Figure S9. 9 Be NMR spectrum of [BeW 12 ] 6at pH 6.8 with 10% D 2 O. The spectrum features two distinguishable signals, in accordance with the two Keggin isoforms determined by 183 W NMR (cf. Figure S14). Signal assignment is arbitrary due to the unprecedented β-isomer and the close proximity of the signals. The β-isomer is depicted in colours representing the symmetry-equivalent W positions in order to illustrate the less symmetric environment experienced by the Be center. Color code: {WO 6 }, blue, green, dark-blue, orange; Be, yellow; O, red. Figure S10. 183

W NMR spectrum of [WO 4 ]
2solution at pH 6.8 with 10% D 2 O. The NMR signals are assigned to W nuclei sharing the same chemical environment according to literature data [23] . The total recording time is about 60 hours for the measurement, chemical shifts were measured relative to external 1 M Na 2 WO 4 . Color code: {WO 6 }, blue, green, red; O, red.    W NMR signals are assigned to W nuclei sharing the same chemical environment according to literature data [27] and can be assigned to the α-and β-Keggin isomer.    Figure S16).

Enzyme kinetics
(The kinetic models with according mathematical expressions applied in this study are based on Copeland's guide for enzyme kinetics [28] .)

Kinetic evaluation according to general mixed inhibition model
The classical Michaelis-Menten equation [28] relates the velocity v of an enzymatic reaction to the applied substrate concentration [S]: In this equation, the Michaelis constant K M represents the substrate concentration effecting the half-maximum reaction rate v max .
For the reported inhibition studies, we chose an expanded version of this equation (1) which accounts for all possible inhibition mechanisms and combinations thereof [28] . In general, both Therefore, values of α < 1 often correspond to mixed-type inhibition with contributions from uncompetitive and non-competitive inhibition modes. (1)

Insertion of expressions (2) and (3) in equation
The relative enzymatic activity upon inhibition can be expressed as the ratio of the observed reaction velocity v app to the non-inhibited reaction performance v through dividing equation (4) by equation (1) This corresponds to a normalization of the reaction rates with respect to the non-inhibited rate. As we were interested in the dependence of the reaction velocity on the inhibitor concentration, the substrate concentration was fixed to 1 mM L-DOPA for all activity plots. This led to a further simplification of equation (1): Equation (6) was used as the final curve fit expression for kinetic evaluation of those POM compounds showing inhibitory activity. K M was set to 26.1 mM as reported for abPPO4 by Pretzler et al. [1] . The fitted α-parameters were smaller than 1, but not close to zero, suggesting a mixed-type inhibition as a combination of uncompetitive and non-competitive contributions.

Kinetic evaluation using Lineweaver-Burk plots
The inhibition mode was determined for each compound with inhibitory effects by analysis of linear Lineweaver-Burk [29] extrapolations. The double-reciprocal plots of 1/v app against 1/[S] yielded straight lines that were fitted accordingly to the correspondingly rearranged form of equation (  The ordinate intercepts t of the Lineweaver-Burk graphs can be evaluated for the inhibition parameter K is [30] , which corresponds to α • K i in Cope's nomenclature: With K i known from the plots according to equation (8), another estimate for the parameter α can be extracted from the Lineweaver-Burk data.

Kinetic evaluation using Dixon plots
Finally, the inhibition mode was further assessed by Dixon plots [31] . Therefore, equation (7 In a similar fashion as performed with the Lineweaver-Burk plots, the intersection behaviour of the Dixon lines is often used for determination of inhibition types. An ideal competitive inhibitor leads to an intersection point in the second quadrant, as demonstrated for kojic acid (see Figure S18B). The K i value was verified from the found intersection point at (-K i , v max ). Perfect non-competitive inhibition results in Dixon plots intersecting on the abscissa, and for uncompetitive inhibition parallel lines are obtained, which strongly suggests a mixed-type inhibition for the tungstate species as evident from the plots depicted here (Figures S19-23B and 3B in the main text). In accordance with the proposed mixed-type inhibition mode, the five lines obtained for each inhibitor did not intersect in a single common point, but rather within a common region. Nevertheless, we calculated an average intersection from the means of the intersections and used its x-coordinate as an estimate for -K i (cf. Table 1 in the main text). Fitting Dixon plots for mixed-type inhibition curves can be performed with nonlinear equations [31] requiring more than three data points, which is why we used the presented linear simplification only to obtain a third estimate for the K i value. Figure S17. Activity plot for abPPO4 inhibition by kojic acid. The dopachrome assay was performed with 1 mM L-DOPA (in 50 mM Na-citrate pH 6.8) and 0-14 mM POTs. The initial linear reaction rates were normalized with respect to the non-inhibited reaction velocity to be plotted as relative enzymatic activities and fitted according to equation (6). For fit parameters, see Table 1 in the main text. Figure S18. Kinetic evaluation of abPPO4 inhibition by kojic acid using A) Lineweaver-Burk and B) Dixon plots. Fit parameters can be found in Table S3.  Table S3.     Table S3.   Figure S24. Original gel photography of the SDS-PAGE analysis shown in Figure S1. The image was taken with a Biorad Molecular Imager Gel Doc XR documentation system using the instrument software Image Lab (version 5.2.1) with standard settings for Coomassie-stained SDS-PAGE. Figure S25. Timeline of selected milestones in the structural characterization of Keggin POTs by 183 W NMR. Since the 1980's, 183 W NMR analysis facilitates distinction of α-and β-isomers of the intact Keggin anion XW 12 , the lacunary form α-XW 11 and even monosubstituted clusters α-XW 11 M after insertion of a transition metal into the lacunary site. X depicts the tetrahedral heteroatom center of Keggin anions. Symmetry-equivalent W centers are depicted in the same colours in corresponding structure schemes. The Baker group started with the first NMR analysis of α-Keggin POTs in 1979 (Acerete et al. 1979 [10]). They extended their findings to homologous structures (Acerete et al. 1982 [7]) and transition metal substituted anions (Jorris et al. 1987 [21]). Lefebvre et al. (1981 [27]) contributed signal assignments to the β-Keggin isomer and Maksimovskaya & Burtseva (1985 [22]) identified the characteristic heptatungstate signal pattern which is abundant in acidified orthotungstate solutions.