The Benzopyrone Biochanin-A as a reversible, competitive, and selective monoamine oxidase B inhibitor

The tested pure PCS constituents (≥95%) of PS and PSD and the standard selective MAO-AI clorgyline (CLORG) were obtained from Santa Cruz Biotechnology Inc. (Dallas, TX, USA). Hank's Balanced Salt Solution (HBSS), HEPES, benzylamine HCl, IPS, 6-PN, NBI, BIO-A, the selective MAO-BI standard selegiline (Deprenyl®) (DEP), and other chemicals and chromatography supplies were purchased from Sigma-Aldrich (St Louis, MO, USA). Human MAO-A and MAO-B, derived from a recombinant baculovirus infected insect cells, and measured in active units (U) were purchased from Sigma-Aldrich and received in a buffer solution at pH 7.4. The China originated PCS were obtained from East Earth Trade Winds (Redding, CA, USA), and the MAO-Glo™ kits from Promega (Madison, WI, USA).

The dry PCS were milled and sieved into 30 g of fine seed powder and stored in a nitrogen-sealed container at 2°C. To prepare PCSEE, we followed the procedure mentioned in our previous study [17]. Briefly, the fine PCS powder was macerated for two days using 99.95% ethanol. The extract was filtered and the powder obtained was subject to 8-10 h Soxhlet reflux extraction at 60° to 70 °C with continuous solvent (ethanol) renewal every 2 h. The combined ethanol extracts were evaporated in a dark hood to obtain an oily non-homogenous PCSEE. All the collected dry crude extracts and prepared PCS ethanolic stock solutions were stored in glass containers at 2 °C until use.

A slightly modified continuous spectrophotometric assay [25] for measuring enzymatic H₂O₂ production was validated before the experiments. In this assay, the peroxidase chromogen reagent of 1 mM of vanillic acid and 500 μM 4-aminoantipyrine, and 4 U purpurogallin/mL of horseradish peroxidase (HRP) type-II in HBSS (pH 7.4) was prepared and validated with H₂O₂ standard curve. The linearity of reaction on the instrument was confirmed to reach up to 1.76 nmol per reaction volume with R² of 0.9999. Isozymes were diluted to 0.7 U/mL final concentration by 10 mM HEPES in HBSS (pH 7.4). Substrates final concentrations were optimized to be 0.5 and 3 mM tyramine HCl and benzylamine HCl for hMAO-A and hMAO-B reactions. The assay was run with and without different standard MAOIs with different selectivities (DEP, CLORG, pirlindole, and rasagiline) and monitored for few hours. Isozymes were found stable within the time of incubation. Stock solvents (ethanol or dimethyl sulfoxide) were kept ≤ 2% the final volume of all assays. IC₅₀s were measured at earliest readings with high signal to noise ratios possible were H₂O₂ formed < 10% for hMAO-A and hMAO-B. For the confirmatory luminescence assay, using MAO-Glo™ kit, the protocol was followed to determining the time of incubation at 60 min at RT guided by Valley’s method [26] and the preliminary experiments which we conducted.

The six benzopyrones of 5 × final concentrations between 0 to 400 μM and PCSEE from 0 to 250 μg/mL of at least ten data points were prepared in HBSS (pH 7.4). The diluted compounds (25 μL) were incubated with 50 μL of 2.5 × the final isozyme concentration (or buffer for blanks) in 96-96-good plates for 40 min at RT. Freshly prepared volumes of 50 μL of 2.5 × substrate and peroxidase chromogen mix (1:1) were added. Absorbance was immediately read at 490 nm by the Bio-Tek μQuant Monochromatic Microplate Spectrophotometer for time zero. H₂O₂ production progression for each enzyme was measured every 20 min for the duration of the assay. First read measurements and the blanks background were subtracted from all absorbance values.

The six benzopyrones, PCSEE, and DEP inhibitory activities of hMAO-A and hMAO-B isozymes were confirmed using MAO-Glo™ kit luminescence assay. Briefly, in white opaque 96-well microplates, 12.5 μL of 4 × 10 μg/mL final concentrations of each test compound, standard DEP, or PCSEE in reaction buffer of pH 7.4 were added (n = 3 per isozyme). A buffer of 12.5 μL was used to make equivalent blank volumes of wells without enzyme. Equivalent amounts of used solvent volumes (ethanol or DMSO) were also added to separate wells. A volume of 25μL of 2 × 0.9 U/mL final concentration of each of freshly thawed hMAO-A and hMAO-B isozymes were incubated with the test substances at RT. Freshly prepared luciferin derivative substrate volume of 12.5 μL of 4 × 40 and 4 μM the final concentrations were incubated for 60 min with hMAO-A and hMAO-B and inhibitors, respectively at RT. Reporter luciferase detects reagent of 50 μL was incubated with each reaction for 20-30 min to inhibit the enzyme and produce luminescence. Arbitrary light units (ALU) signal changes were measured by Synergy HTX Multi-Reader from Bio-Tek and converted to percent control change.

The H₂O₂ scavenging activity of PCSEE and BIO-A at their highest used MAO inhibitory concentrations was investigated and compared to the standard scavenger L-ascorbic acid. In 96 well plates, 25 μL of 5 × final concentration of PCSEE (62.5, 125, and 250 μg/mL), BIO-A (6.25, 12.5, and 25 μg/mL), and ascorbic acid (62.5, 125, and 250 μg/mL) in HBSS (7.4) were prepared. A volume of 75 μL of 1.67 × final concentration of 15 μM H₂O₂ in HBSS (7.4) was plated in test wells (n = 4). Buffer substituted the H₂O₂ (n = 4) to make blank wells. Immediately, peroxidase chromogen-containing vanillic acid, 4-aminoantipyrine, and HRP type-II were prepared as in the MAO spectrophotometric assay and added as 25 μL per well. The developing color at 490 nm was optimally read within 15 min at RT using Bio-Tek Synergy HTX Multi-Reader absorbance.

PCSEE and BIO-A were concurrently run on a silica gel matrix (5 × 10 cm²) containing a fluorescent indicator with 254 nm excitation on aluminum foil (Sigma-Aldrich) for thin-layer chromatography (TLC) identification. On replicate TLC matrices, 5 μL of a 1 cm short band of 20 mg/mL PCSEE, and 1 and 5 mM BIO-A (Sigma-Aldrich) ethanol stock solution were directly deposited at 0.5 cm or more from TLC plate lower edge. After ethanol evaporation, TLC plate was slowly placed and developed in a clean glass developing chamber saturated with a miscible solvent system (water: acetone: ethanol of 5: 3: 1 ratio), 0.5 cm deep or less. Plates were taken out after 50-60 min incubation at RT, or when solvent front reached 1 cm from the upper TLC edge (approximately 8.5 cm height). Dried developed TLC plates were then visualized under short (254 nm) and long (366 nm) UV lights. Retention factor (R _f ) of BIO-A band and a matching band were quantified.

hMAO-A and hMAO-B were tested for their activity recovery after preincubation with BIO-A, utilizing a slightly modified method reported by Legoabe et al [27] and using the continuous spectrophotometric assay [25]. Briefly, in HBSS medium (pH 7.4), BIO-A of 10 × and 100 × IC₅₀ and DEP at 10 × IC₅₀ were preincubated separately with 70 U/mL hMAO isozymes for 40 min at RT. BIO-A of 4 × IC₅₀ was concurrently preincubated with 0.7 U/mL. Buffers substituted inhibitors and isozymes for negative controls and blanks, respectively. All reactions were diluted to 20-fold. The 100 × dilution was immediately accomplished by adding 50 μL of the isozyme-inhibitor mixture to 75 μL of related substrates-peroxidase chromogen reagent mix (a 25 μL buffer and the original 50 μL previously used a mixture) in a 96-well plate. By reaching 0.1 × and 1 × IC₅₀ BIO-A final concentration, developing color was immediately monitored using Bio-Tek μQuant Spectrophotometer readings at RT.

The effects of BIO-A on Michaelis-Menten kinetics parameters of hMAO-A and hMAO-B were investigated using the same spectrophotometric assay mentioned above. Standard DEP was concurrently tested to validate the method. Determined DEP kinetics for hMAO-A and hMAO-B in the conditions of this experiment were determined and used as a standard for the Prism program validation for Inhibitor constant (K_i) determination of BIO-A. Briefly, seven serially diluted tyramine and benzylamine concentrations were prepared in a range between 0 to 1 mM for hMAO-A and 0 to 6 mM for hMAO-B. BIO-A for a final concentration of 0, 0.5, 1, and 2 × IC₅₀ concentrations were incubated with each isozyme for a final concentration of 0.7 U/mL in a 2:1 ratio (0, 1.7, 3.4 and 6.9 μM for hMAO-A, and 0, 0.05, 0.09, 0.19 μM for hMAO-B). HBSS (pH7.4) substituted inhibitors and isozymes in the negative controls and the blanks, respectively. After 40 min, 75 μL of the enzyme-BIO-A mixture were added to 50 μL of the related substrate-chromogen reagent mix (1: 1) in 96-well plates. Developing color was monitored at RT using Bio-Tek μQuant Spectrophotometer. Values of maximum velocity (V_max), Michaelis constant (K_m), and consequently Ki were determined.

To understand the possible molecular interactions between BIO-A and isozymes active sites, we used the HYBRID docking method of OEDocking (v 3.0.1) (OpenEye Scientific Software Inc.; Santa Fe, NM, USA) [28]. Both X-ray crystal structures of human (h) MAO-A in complex with harmine (PDB ID: 2Z5X) and human (h) MAO-B in complex with 2-(2-benzofuranyl)-2-imidazoline (2-BFI) (PDB ID: 2XFN) were downloaded from the RCSB Protein Data Bank (PDB). Both proteins structures were imported into Sybyl -X 1.3 environment of Tripos International (St. Louis, MO) where the protein’s chain-A were extracted and refined using the Biopolymer Structure Preparation Tool. Atom types were corrected, amino acid residues were repaired, active water molecules were retained, and all hydrogen atoms were added to both proteins. The cofactor flavin adenine dinucleotide (FAD) and its covalent linkage to the cysteine residues of hMAO-A (CYS: 406: A), and hMAO-B (CYS: 397: A) were retained during the docking process. To be ready for the docking, the energy of each protein was minimized using the MMFF94 charges with the MMFF94s force fields. For docking the compound of interest (BIO-A), the low energy 3D conformers were generated using OMEGA version 2.4.6 of the same software [28]. The bound standard ligands of 2Z5X and 2XFN were sketched using Sybyl sketch and were re-docked to the crystal structures of the proteins for method validation. The root mean square deviation values between retrieved and re-docked poses of the ligands were less than 2 Å. Top ten poses were examined. BIO-A was then docked as a test ligand to both MAO proteins at their Ligand Binding Domain. The H-bonds formed between BIO-A, and amino acid residues of the top poses were measured.

Statistical analyses were performed using GraphPad Prism program version 6.02 for Windows, GraphPad Prism Software Inc. (San Diego, CA, USA). Data points and parameters were expressed as the mean ± SEM of all data. All parameters were calculated from non-linear regressions with the best R² values. Inhibitory concentrations of 50% of isozyme (IC₅₀) were obtained by the interpolation of a normalized variable slope logarithmic curve. RS values of all benzopyrones and PCSEE were calculated from their mean IC₅₀s as follows:

The V_max and K_m calculated from Michaelis-Menten kinetics model. K_i was calculated according to the competitive behavior model as the best fit and Cheng-Prosuff’s equation (K_{i (comp)} = IC₅₀/(1+ [S]/K_m)). SI of BIO-A was calculated from mean K_i values (SI = hMAO-A K_i/hMAO-B K_i). The significance of differences between the different groups was determined using one-way ANOVA followed by Dunnett's multiple comparisons test or two-way ANOVA followed by Sidak's multiple comparisons test.

To investigate the hMAO-A and hMAO-B inhibitory effects of the six PCS benzopyrones constituents, we determined their enzymatic H₂O₂ IC₅₀s using a spectrophotometric assay (Fig. 2 Table) and verified their inhibition by a luminescence assay (Fig. 2 a and b). In both assays, solvents used in compounds stocks showed no significant difference from the controls without solvents. From the first glance at the Table of Fig. 2, benzopyrones of flavonoids and coumarins results clearly indicate an array of hMAO-A and hMAO-B inhibition potencies with hMAO-A and hMAO-B IC₅₀s less than ~100 and ~200 μM, respectively. Compared with standards CLORG and DEP, hMAO-A IC₅₀s of the benzopyrones showed BIO-A > IPS > 6-PN > NBI > PS > PSD (no detectable inhibition), with IPS, PS, and NBI having a relative selectivity to inhibit hMAO-A (RS_A). Meanwhile, hMAO-B IC₅₀s showed BIO-A > 6-PN > IPS > NBI > PS > PSD (partial inhibition, p < 0.001), with BIO-A and 6-PN having a relative selectivity in favor to inhibit hMAO-B (RS_B). In this assay, BIO-A was, by far, the most potent and selective hMAO-B inhibitor among the six tested benzopyrones.

All the six benzopyrones activities against hMAO-A and hMAO-B were further confirmed at a fixed concentration using MAO-Glo™ kit luminescence assay, a highly sensitive and an H₂O₂-independent technique (Fig. 2 a and b). Standard DEP and PCSEE were used as positive controls. All benzopyrones at 10 μg/mL are differently affected hMAO-B hMAO-A (p ≤ 0.05 to 0.0001), confirming the inhibitory array pattern of the spectrophotometric assay with even higher inhibitory effects. Consistent with our spectrophotometric assay selectivity data, IPS, PS, and NBI, were more effective against hMAO-A, and 6-PN was more effective against hMAO-B, while BIO-A selectivity was masked by its highly potent effects using luminescence. Thus, the relative selectivity of benzopyrones matching the spectrophotometric assay is an indication that these benzopyrones possess hMAO-A and hMAO-B inhibitory effects with different selectivities.

To compare the BIO-A inhibitory potencies of hMAO-A and hMAO-B with PCSEE inhibitory potencies, we elucidated PCSEE effects on both isozymes H₂O₂ production using the same spectrophotometric assay (Fig. 3). Ethanol concentrations did not have effects on control isozymes activities. The PCSEE treatments significantly inhibited both isozymes (p < 0.0001) (Fig. 3 a). PCSEE showed RS_B (17.3-fold) with an IC₅₀ of 3.03 μg/mL (hMAO-B) compared to 52.36 μg/mL (hMAO-A). Nevertheless, the left shifted BIO-A curves compared to PCSEE curves display BIO-A significantly higher potency and RS_B than PCSEE (Fig. 3 b).

H₂O₂ scavenging activities were tested to reveal BIO-A and PCSEE direct effects on the enzymatic H₂O₂ or other side interactions with the hMAOs assays (Fig. 3 c). The ascorbic acid standard showed a complete scavenging activity at all tested concentrations (p < 0.001). However, the high tested concentrations of BIO-A and PCSEE did not show any scavenging activities. Moreover, PCSEE increased the H₂O₂ signal (p ≤ 0.05) which may indicate other autoxidation reactions that occur only at a very high concentration of 250 μg/mL. The results suggest that the inhibition of enzymatic H₂O₂ by BIO-A and PCSEE were solely by the mechanism of inhibiting hMAO-A and hMAO-B, and not by H₂O₂ scavenging activities.

BIO-A identification was verified in the PCSEE extract using the fluorescent silica gel TLC and our optimized developing system (Fig. 3 d). BIO-A’s non-fluorescent dark band on the TLC plate matched a non-fluorescent dark band in PCSEE fingerprint at both UV excitation waves. At 366 nm, BIO-A band (\( Rf \) = 0.36 ± 0.002) was visibly matching one of the PCSEE dark bands (\( Rf \) = 0.37 ± 0.002) at similar conditions (p > 0.05). At 254 nm with BIO-A increased concentration, BIO-A dark band matched a non-fluorescent dark band in the PCSEE fingerprint developed bands. Thus, in addition to its identification of the used seeds, the chromatographic method supported the reported presence of BIO-A as one of PCS phytochemicals.

To determine the isozymes mode(s) of inhibition by BIO-A, reversibility of inhibition, Michaelis-Menten parameters changes, K_i values for hMAO-A and hMAO-B, are presented in Figs. 4, and 5, respectively.

To determine if BIO-A is a reversible hMAO-A and hMAO-B inhibitor, we tested the recovery of both isozymes from BIO-A inhibition after dilution (Fig. 4 a, and b), with time monitoring (Fig. 4 c and d). Compared to controls, the 0.1 × IC₅₀ of standard DEP reduced both isozymes activities to 7% (with hMAO-A), and 51% (with hMAO-B) instead of 90%, respectively. That is an indication of possible slow, partial, or no recovery of the isozymes from DEP inhibition at the experimental conditions. With hMAO-B, repeatedly unexpected observed high signal with DEP could be due to interactions with hMAO-B sample contaminants at very elevated levels of drug and enzyme. Conversely, in similar preincubation conditions of BIO-A with hMAO-A and hMAO-B, dilution allowed a complete or an almost complete recovery of activities. The hMAO-A (Fig. 4 a) and hMAO-B (Fig. 4 b) catalytic rates with 1 and 0.1 × IC₅₀ BIO-A were ~50% and ~90%, respectively. Both isozymes constant recovery with time (Fig. 4 c and d) is a character of a reversible inhibitor, as the activities of recovered enzymes were similar to the enzymes subjected directly to 1or 0.1 × IC₅₀ of BIO-A. The obtained results demonstrate evidence that BIO-A is a reversible inhibitor of hMAO-A and hMAO-B.

To determine the mode of BIO-A reversible inhibition on both MAOs, we investigated BIO-A effects on Michaelis-Menten kinetics of hMAO-A and hMAO-B at their initial velocities (Fig. 5). In Fig. 5 a and b, Lineweaver-Burk plot linear regressions for the BIO-A increased IC₅₀ fold co-intersected at the Y-axis indicating a competitive mode of inhibition in both isozymes. Compared to control in Table of Fig. 5 c, BIO-A showed no significant change in hMAO-A V_max while K_m values significantly (p < 0.001) geometrically doubled (9.54, 19.66, 40.38-fold, respectively) with increasing BIO-A IC₅₀s fold of concentrations. Similarly, hMAO-B V_max showed no significant changes while benzylamine K_m values were significantly (p < 0.01) nearly linearly increased (3.33-, 6.76-, 10.18- fold respectively) with the BIO-A doubled IC₅₀s. The obtained data indicate the competitiveness of BIO-A to inhibit both isozymes.

To confirm competitiveness of BIO-A for hMAO-A and hMAO-B inhibitions, and consequently determine BIO-A K_i and its hMAO-B selectivity index (SI), we analyzed Michaelis-Menten data by best-fit enzyme inhibitory models using GraphPad Prism. (Fig. 5 d; Table). The mixed enzyme inhibition model for standard DEP indicated a competitive inhibitory mode for hMAO-A (alpha > 1) and a non-competitive mode for hMAO-B inhibition (alpha ~1). Meanwhile, the same model supported that BIO-A impedes substrate-binding affinity to both isozymes (alpha > > 1). Thus, the analysis confirms BIO-competitiveness for both isozymes and rejects the non-competitive, uncompetitive behavior. Consequently, BIO-A K_i values showed that BIO-A SI has a tight affinity to bind to hMAO-B more than hMAO-A (26-fold; p < 0.001).

To recognize the molecular level of the interactions of BIO-A with the amino acids in the active site involved in its MAOs reversible and competitive inhibitions; we conducted a molecular docking study on both isozymes (Fig. 6). BIO-A was well docked to the same human MAO-A and hMAO-B crystal structure active sites at which 2Z5X and 2XFN interacted, respectively (Fig. 6). In the MAO-A compact active site cavity, BIO-A matched the 2Z5X pose with its C3-phenyl ring and the chromone core structure embedded in the hydrophobic zone of the entrance site (Fig. 6 a, Brown zone). Also, BIO-A was predicted to be a donor for a relatively tight H-bond (1.99 Å) formed between the hydrogen of its C7-OH group and the nitrogen FAD of the enzyme (Fig. 6 c Table). Therefore, the reversible H-bond with FAD and the hydrophobic interactions at the entrance cavity may contribute to BIO-A reversible competitive behavior in effectively inhibiting human MAO-A.

Nonetheless, BIO-A showed stronger interactions at the active site of MAO-B than MAO-A. All of the BIO-A i.e., the lipophilic chromone core structure, the C3-phenyl ring, and possibly the methyl group of C4’ positions, were completely embedded in the MAO-B active site lipophilic zones, away from the FAD (Fig. 6 b, Brown zone). Notably, BIO-A was predicted to be a donor and an acceptor for two distant H-bond. One at the entrance cavity between the oxygen acceptor of its C4’-OCH₃ group and the active water molecule, and the other, is closer to the substrate cavity between the carbon donor of its C7-OH and the Leucine 164 (Fig. 6 c Table). The BIO-A H-bond distances were close to the 2XFN H-bond distances. The predicted reversible hydrophobic interactions and distant H-bonds formations might consistently reflect the high affinity, reversibility, and competitiveness extent of BIO-A towards its MAO-B mode of inhibition.