Bioassay-Guided Isolation, Metabolic Profiling, and Docking Studies of Hyaluronidase Inhibitors from Ravenala madagascariensis.

Hyaluronidase enzyme (HAase) has a role in the dissolution or disintegration of hyaluronic acid (HA) and in maintaining the heathy state of skin. Bioassay-guided fractionation of Ravenala madagascariensis (Sonn.) organ extracts (leaf, flower, stem, and root) testing for hyaluronidase inhibition was performed followed by metabolic profiling using LC-HRMS. Additionally, a hyaluronidase docking study was achieved using Molecular Operating Environment (MOE). Results showed that the crude hydroalcoholic (70% EtOH) extract of the leaves as well as its n-butanol (n-BuOH) partition showed higher HAase activity with 64.3% inhibition. Metabolic analysis of R. madagascariensis resulted in the identification of 19 phenolic compounds ranging from different chemical classes (flavone glycosides, flavonol glycosides, and flavanol aglycones). Bioassay-guided purification of the leaf n-BuOH partition led to the isolation of seven compounds that were identified as narcissin, rutin, epiafzelechin, epicatechin, isorhamnetin 7-O-glucoside, kaempferol, and isorhamnetin-7-O-rutinoside. The docking study showed that narcissin, rutin, and quercetin 3-O-glucoside all interact with HAase through hydrogen bonding with the Asp111, Gln271, and/or Glu113 residues. Our results highlight Ravenala madagascariensis and its flavonoids as promising hyaluronidase inhibitors in natural cosmetology preparations for skin care.

major bioactive compounds was carried out leading to the rediscovery of multiple known compounds. Moreover, the interaction of isolated compounds to the HAase via docking studies was carried out.

Results
The crude extracts of different organs (leaf, flower, stem, and root) were subjected to hyaluronidase inhibition activity (Table 1), where the crude extract of the leaves was the most active HAase inhibitor (78.9%). These findings of the crude extracts of leaves and stems were much better that previously reported when compared with the aqueous and methanolic extract of Pleurotus tuber-regium that showed inhibiton by 22.19% and 3.94%, respectively at 100 µg/mL while it showed comparable activities with those of Trametes lactinea aqueous and acetonic extract that inhibited hyaluronidase by 88.6% and 88.3%, respectively [6]. Additionally, partitions of the leaves were evaluated for their HAase inhibition activity, where hydro-alcoholic as well as its n-BuOH extracts of R. madagascariensis showed higher inhibition (90% and 64.3%, respectively). The n-BuOH was subjected to chromatographic purification, yielding seven compounds identified as narcissin, rutin, epicatechin, isorhamnetin 7-O-glucoside, isorhamnetin 7-O-rutinoside, epiafzelechin, and kaempferol. The isolated compounds were tested for their inhibition percentage (Table 1). Results showed that epiafzelechin and epicatechin the highest inhibition value with 36.5% and 34.4%, respectively. While the other isolated ones are less, these results were going with previously reported that increasing the number of hydroxyl groups, especially in 2,3 and 5 (myricetin) and decreased after glycosylation or substitution of hydroxyl groups. [13,14]. The Docking studies were carried out to describe the interaction properties between the isolated ones and the enzyme, showed that narcissin, rutin, and Quercetin 3-O-glucoside interact with HAase through the formation of hydrogen bonds with the Asp111, Gln271, and/or Glu113 residues.

Identification of Purified Compounds
The isolated compounds ( Figure 2) were structurally elucidated established on their physiochemical, chromatographic properties, and spectroscopic analysis (UV, ESI-MS, 1 H-NMR and 13 C-NMR), as well as comparison to standards. The isolated compound 1 was obtained as yellow amorphous powder, dark purple color under UV light which changed to bright yellow with ammonia vapor. R f = 0.38 by using CH 2 Cl 2 / MeOH (8:2) as solvent system for TLC. 1 [28,32].
Compound 2 was the same as 1 except the absence of OCH 3 signals at δ 3.83 and 55. 6 and was identified as rutin [31], compound 15 was obtained as yellow amorphous powder, dark purple color under UV light which changed to bright yellow with ammonia vapor. R f = 0.18 by using CHCl 3 / MeOH (9.5:0.5) as solvent system for TLC. 1 [33,34], while compound 16 was obtained as yellow amorphous powder, dark purple color under UV light which changed to bright yellow with ammonia vapor. R f = 0.14 by using CHCl 3 / MeOH (9.8:0.2) as solvent system for TLC. 1 [37], and compound 19 was the same with only one missing aromatic proton at position 3' at δ 6.8 and was identified as epicatechin [38,39].

Modeling Study Exhibiting the Binding Ability of Polyphenolic Compounds to HAase
To predict the precise binding sites on HAase and explore the interaction between the target compounds and HAase systematically, a docking study was run using Molecular Operating Environment (MOE; Chemical Computing Group Inc., Montreal, Canada). From the docking calculation, the smallest energy-ranked results of the target compounds -HAase conformations are outlined in Table 2. By comparing, of the data from Table 1 with the data from Table 2, the noticed free energy change (∆G • ) for the target compounds -HAase systems was consistent with the results of biological studies. As shown in Figure 3, all the compounds containing the flavonoid nucleus were situated in the hydrophobic cavity of HAase and were encircled by hydrophilic and hydrophobic amino acids. Therefore, mainly the interaction between flavonoids and HAase is due to electrostatic forces and hydrophobic interactions. However, due to more hydrophobic amino acids lining quercetin in the binding site, hydrophilic interaction was more conspicuous than electrostatic force in the quercetin -HAase system. As shown in Figure 3, many amino acid residues, such as Ser303, Gln271, Glu113, Tyr227, Asp111, Tyr55 and Ser304, manifested in the binding of each compound with HAase [40]. These results indicate that these residues would perform a significant role in the interaction between the illustrated compounds and HAase and might constitute the catalytic site of HAase. Furthermore, due to the presence of different ionic and polar groups, there are also large numbers of hydrogen bonds in HAase binding site. As shown in the following figures, one of the hydrogen bonds is formed with the Asp111, Gln271 and/or Glu113 residues in the flavonoid (narcissin, rutin and quercetin-3-O-glucoside) -HAase systems (Figure 3). The compounds that exhibited the lowest binding free energy values were rutin and narcissin with values -7.119 Kcal/mol and -6.853 Kcal/mol, respectively, which agrees with their inhibitory effect on HAase. Therefore, according to docking studies and the biological screening, it could be suggested that flavonoid or flavonoid-like compounds have the ability bind to the HAase catalytic site, which would suppress the activity of the enzyme. The displayed score is the mean of 3 sequential runs. The docking technique was validated by successful pose-retrieval docking trial of the ligand (score: −5.370). Structure activity relationship showed that: The C2,3 double bonded flavonoids have high potency [13,14], the flavonoid glycosides were more potent in their inhibitory effects than aglycone and the ortho-dihydroxyl substitution exhibited greater inhibition than those singly hydroxylated, while the methoxy substitution decreased the inhibitory effects. The following flavonoid structure conferred potent inhibitory effect: the C2,3 double bond and ortho-dihydroxyl substitution.

Plant Material
Ravenala madagascariensis (Sonn.) leaves, flowers, stem and root were collected from the El Orman garden, Egypt, in August 2015. The plant was authenticated by Professor Ibrahim Ahmed El-Garf (Department of Botany and Microbiology, Faculty of Science, Cairo University, Giza, Egypt). The leaves were air-dried, powdered, and kept in amber-colored, air-tight glass containers at low temperature for preservation of chemical materials for future testing.

Extraction and Isolation
Dried plant material was detached into leaves, flowers, stem and root and then ground. Subsequently, each part was extracted with 70% ethanol (EtOH) by cold maceration, and then concentrated under vacuum. The powdered leaves (1.7 kg) were extracted exhaustively at room temperature then concentrated to yield 400 g. Approximately half of this extract (~200 g) was resuspended in distilled H2O and successively fractionated with solvents of increasing polarity: n-hexanes, dichloromethane (DCM), ethyl acetate (EtOAc) and finally n-BuOH.

Preparation of Hyaluronidase
Following the procedure laid out by Smith et al., 1968 [41], 100 mL of brain heart infusion media was prepared with 1 g of agar, autoclaved for 15 min at 121 • C, and cooled to 46 • C; then, a 2 mg/mL solution of sodium hyaluronidate, was added to the cooled medium to a final concentration of 400 µg/mL. 5% human albumin was added under constant mixing to give a final concentration of 1%. Plates were then poured to a depth of 3-4 mm and left to solidify at 4 • C. The isolates of Staphylococcus aureus were grown on sheep blood agar until growth was clearly observed. A single colony was streaked onto the hyaluronic acid medium, incubated at 37 • C, and observed daily until growth was detected (24 to 72 h). On the day growth was first observed, the plate was immersed with 2N glacial acetic acid which binds hyaluronic acid and albumin, forming a white precipitate. Hyaluronidase production was confirmed if a clear zone was observed.

Hyaluronidase Inhibition Assay
The strains of Staphylococcus aureus were expanded overnight in TSB at 37 • C with shaking, then sub-cultured 1:1000 and grown at 37 • C for the time specified below. Spent culture medium was isolated with Spin-X filters (pore size, 0.22 µm) then frozen at −20 • C until used in biological assays. 50 µL of spent culture medium was incubated for 15 min at 37 • C with 50 µL of the used test inhibitor (1 mg/mL dissolved in 5% DMSO). Hyaluronic acid (HA) (100 µL at 1 mg/mL) was blended and permitted to react at 37 • C for 15 min. To stop the reaction 25 µL of sodium tetraborate solution (0.8 M, pH 9.1) was added and then the reaction mixture was vortexed and boiled for 3 min. The positive control (tannic acid, was used as HAase inhibitor) [42][43][44] and negative control were made using the identical procedure but without the test or standard inhibitor (50 µL of 5% DMSO). In parallel, 50 µL of 5% DMSO and 50 µL spent medium were added to 125 µL of stop solution (1 mg/mL hyaluronic acid, 0.8 M potassium tetraborate, pH 9.1) at time zero, vortexed, and boiled for 3 min. The samples were distributed into a 96-well microtiter plate in quadruplicate. Freshly prepared PDMAB solution (10% [w/v] p-dimethyl amino benzaldehyde, 12.5% [v/v] 10M HCl, and 87.5% [v/v] glacial acetic acid) was added to each well. The plate was incubated at 37 • C for 20 min to let the color change. The absorbance at 590 nm was evaluated by using a microplate reader where A1 is the absorbance of the standard/ extracts, A o is the absorbance of the control.

Metabolic Profiling
The crude extracts of R. madagascariensis (leaves, flowers, stem, and root) were undergone to metabolomic using LC-HRESIMS analytical methods as stated by Abdel Mohsen et al. [45]. Each extract (1 mg/mL dissolved in MeOH) was analyzed separately on an Accela HPLC with an Accela UV-Vis coupled to an Exactive (Orbitrap) MS spectrometer from Thermo Fisher Scientific (Berman, Germany). Purified water (a) and acetonitrile (b) used as mobile phase with 0.1% formic acid in both. The gradient began at 10% (b) to 100% (b) in 30 min and kept isocratic elution for the next 5 min before back to 10% for 1 min (flow rate of 300 µL/ min). Then mobile phase was re-equilibrated for 9 min previous to the following injection. By using untargeted Higher-energy Collision Dissociation (HCD) mechanism, the range of mass was set from m/z 50-1000 for MS/MS and m/z 100-2000 for ESI-MS using in source collision-induced dissociation (CID). The raw data was derived in MZmine 2.12, a framework for differential analysis of MS data. The individual peaks were detected by chromatogram deconvolution. The retention time normalizer was applied for chromatographic alignment and gap-filling. An Excel macro was applied on the positive ionization mode data files produced by MZmine to extract the peaks from each samples and dereplicated each ion peak with compounds in the customized database (using RT and m/z threshold of ±5 ppm), which supplied details on the assumed identities of metabolites in all extracts. By comparison with LipidMaps, DNP and METLIN databases, 19 metabolites were identified.

Molecular Docking Investigation
Docking calculations were achieved using (MOE) Molecular Operating Environment on a HAase model (PDB code 1FCV, http://www.rcsb.org/pdb/home/home.do). The structures of the compounds were generated by Chemdraw Ultra 8.0. Docking emulation was achieved on the identified compounds with the next protocol: (1) Enzyme structures were examined for missing atoms, bonds and contacts, as well as (2) hydrogen atoms were joined to the enzyme structure. (3) Then by using the builder module, the ligand molecules were established and were energy minimized. (4) The active site was created using the MOE Alpha Site Finder. (5) Ligands were docked within the HAase active site using MOE Dock with simulated annealing utilized as the search protocol and CHARMm molecular mechanics force field. (6) The smallest energy conformation of the docked ligand complex was chosen and undergone to more energy minimization using CHARMm force field. Determination of precision of this docking protocol was carried out by redocking the co-crystallized ligand into the HAase vital site. This method was rerun triple times and the best ranked solutions of the ligand displayed RMSD values of 1.84 A • from the site of the co-crystallized ligand for HAase. Generally, RMSD values smaller than 2.0 A • show that the docking protocol is efficient to accurately promising the binding orientation of the co-crystallized ligand [46]. This protocol was deemed to be convenient for the docking of the test compounds into the active site model of HAase.

Conclusions
This work showed the importance of natural products specially phenols and polyphenols as a competitive inhibitor of hyaluronidase as a key enzyme in healthiness of skin. These findings may assist for the application of plant under investigation in future in cosmetic and in phototherapy as a dietary supplement with fewer side effects. On the other hand, it gave insight about the proposed interactive sites between compounds and the enzyme that may help in optimization of these compounds for production of more active compounds.