key: cord-0912364-b03df2sv
authors: Elshamy, Abdelsamed I.; Mohamed, Tarik A.; Ibrahim, Mahmoud A. A.; Atia, Mohamed A. M.; Yoneyama, Tatsuro; Umeyama, Akemi; Hegazy, Mohamed-Elamir F.
title: Two novel oxetane containing lignans and a new megastigmane from Paronychia arabica and in silico analysis of them as prospective SARS-CoV-2 inhibitors
date: 2021-06-04
journal: RSC Adv
DOI: 10.1039/d1ra02486h
sha: e118ccf38fefb68846df010099bce0e6217dfcd1
doc_id: 912364
cord_uid: b03df2sv

The chemical characterization of the extract of the aerial parts of Paronychia arabica afforded two oxetane containing lignans, paronychiarabicine A (1) and B (2), and one new megastigmane, paronychiarabicastigmane A (3), alongside a known lignan (4), eight known phenolic compounds (5–12), one known elemene sesquiterpene (13) and one steroid glycoside (14). The chemical structures of the isolated compounds were constructed based upon the HRMS, 1D, and 2D-NMR results. The absolute configurations were established via NOESY experiments as well as experimental and TDDFT-calculated electronic circular dichroism (ECD). Utilizing molecular docking, the binding scores and modes of compounds 1–3 towards the SARS-CoV-2 main protease (M(pro)), papain-like protease (PL(pro)), and RNA-dependent RNA polymerase (RdRp) were revealed. Compound 3 exhibited a promising docking score (−9.8 kcal mol(−1)) against SARS-CoV-2 M(pro) by forming seven hydrogen bonds inside the active site with the key amino acids. The reactome pathway enrichment analysis revealed a correlation between the inhibition of GSK3 and GSK3B genes (identified as the main targets of megastigmane treatment) and significant inhibition of SARS-CoV-1 viral replication in infected Vero E6 cells. Our results manifest a novel understanding of genes, proteins and corresponding pathways against SARS-CoV-2 infection and could facilitate the identification and characterization of novel therapeutic targets as treatments of SARS-CoV-2 infection.

Medicinal plants and their products have been used for the prevention and/or treatment of several diseases. [1] [2] [3] According to the World Health Organization (WHO), around 80% of people around the worldwide have used herbal medicinal plants, comprising about 21 000 plant species, as primary health care.

According to WHO, around 21 000 plant species have potential for being used as medicinal plants. 4 The Paronychia genus, one of the genera of the family Caryophyllaceae, includes more than 100 plant species all around the world, especially in warm temperature regions such as Africa, the Mediterranean, North and South America and Eurasia. 5 Several traditional uses have been documented for Paronychia plants in different areas all over the world such as for the treatment of prostate, bladder, and abdominal ailments, kidney stones, eczema, as a febrifuge and digestive, diabetes, heart pains, as a gastric analgesic and for ulcers, hypoglycemia, as an aperitif and as a diuretic. [6] [7] [8] [9] Extracts and isolated compounds from Paronychia plant species have been reported to exhibit signicant biological activities such as cytotoxic 10 and antioxidant activities. 9, 11, 12 Numerous phytochemical constituents from plants belonging to the Paronychia genus have been documented, such as gypsogenic acid and polygalacic acid-type saponins, oleanane-type glycosides, avonoids and tocopherols. 10, 11, [13] [14] [15] [16] A recent respiratory infectious disease (coronavirus disease (COVID- 19) ) has been attributed to the novel Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2). 17 Up to this date, there are no reports concerning the chemical constituents and/or biological activities of P. arabica although the medicinal importance of these plants has been shown. Continuing our work using natural resources for the isolation and identication of bioactive metabolites, [18] [19] [20] [21] [22] [23] we described here the isolation and identication of two new oxetane containing lignans 1 and 2, one new megastigmane 3, along with a known lignan (4), eight known phenolic compounds (5) (6) (7) (8) (9) (10) (11) (12) , a known elemene sesquiterpene (13) and one steroid glycoside (14) (Fig. 1 ) from the aerial parts.

Over the past decade, the analysis of pathways and networks has provided a deep understanding of the interactions between genetic variations and therapeutic responses to a large number of drugs in terms of their biological framework. 24 Meanwhile, pathways are known as groups of biological objects connected to specic functions or targets; biological networks are generally assembled in a systems manner, comprising many pathways concurrently. 25 Numerous studies have employed pathway and network-based approaches in targeted therapies to predict drug side effects, explain hazardous toxicity issues and, moreover, to disclose drug resistance mechanisms. 26 To better understand the effects of targeted therapies in patients, a package of soware tools can be used to visualize patient-specic variations and drug targets followed by developing pathways, which are process-oriented representations of biological reactions or biological networks that predict interactions among genes, proteins and other biological entities. 24, 27 To combat COVID-19, prevention of SARS-CoV-2 replication could be achieved by targeting the viral main protease (M pro ), papain-like protease (PL pro ) and RNA-dependent RNA polymerase (RdRp) enzymes. Therefore, the binding modes and affinities of the isolated compounds were predicted against the three SARS-CoV-2 targets using the molecular docking technique. Furthermore, this study aims to identify targets and pathways enriched in response to megastigmane in terms of SARS-CoV-2 infection.

The chemical description of the hydro-methanolic extract of P. arabica aerial parts provides fourteen metabolites, comprising two new oxetane containing lignans, The analysis of the 1D NMR of 1 (Table 1 ) as well as the 2D NMR ( Fig. 2 ) displayed three aromatic rings: (i) a tetrasubstituted ring (ring A), (ii) a tri-substituted ring (ring B) and (iii) a tri-substituted ring (ring C). The HMBC correlations of H-1 

were assigned. These HMBC correlations construct the 3/8 0 attachment of ring A by the oxetane moiety (ring E) as well as the 1 0 /7 0 of ring C by the oxetane moiety (ring E).

The absolute conguration of 1 was conrmed from the J coupling constants as well as the NOESY experiments (Fig. 3 ). Numerous reports have described how the low J coupling constants between both H-7 0 and H-8 0 (6.9-7.6 Hz) can be used to deduce the trans orientation of diphenyloxetanes. [28] [29] [30] In 1, To elucidate the absolute conguration of 1, a Boltzmannweighted TDDFT-ECD spectrum was generated and compared to the experimental one (Fig. 4) . The TDDFT-simulated ECD spectrum of 7 0 R,8 0 R was in a good agreement with the corresponding experimental ECD spectrum of 1 (Fig. 4) . From all of the above-described spectroscopic data, 1 was shown to be a novel oxetane containing lignan, paronychiarabicine A.

Paronychiarabicine B (2), a yellow amorphous powder, showed a negative optical rotation in methanol [a] + 6.6 (C 0.1, MeOH). The molecular formula of 2 was established as C 30 H 32 O 9 from the observed HR-CIMS molecular ion peak at 536.2030 (M) + (calc.: 536.2046; C 30 H 32 O 9 ), which revealed 15 of unsaturation. The FT-IR spectrum of 2 revealed the characteristic absorption bands of hydroxyl (at 3346 cm À1 ) and keto (at 1723 cm À1 ) functional groups. Carefully analysis of the 1 H and 13 C NMR of 2 (Table 1 ) allowed us to deduce that this compound has same skeleton of 1 except for (i) the down eld shi of C-15 29 For 2, the R f (0.62) was found to be lower than that of 1 (0.67) and the coupling constant of H-7 0 was assigned at 6.4 Hz (<6.9 Hz), which is in full agreement with the above mentioned reports. 28, 29 The NOESY correlations of 2 exhibited the same correlations of 1 except for some different correlations that (Fig. 3) . Using the same method, the absolute conguration of 2 was conrmed based upon comparing the Boltzmann-weighted TDDFT-ECD spectra with the experimental one (Fig. 4) . The TDDFT-simulated ECD spectrum of the 7 0 S,8 0 R isomer was in good agreement with the experimental ECD of 2 (Fig. 4) . From all of the mentioned data, 2 was established as a novel oxetane containing lignin, namely paronychiarabicine B. Twenty-seven carbon resonances were characterized in the 13 C NMR of 3 ( Table 2 ). The complete examination of the 13 From the careful assignment of the 1D NMR, the structure of 3 was established as a megastigmane skeleton. 33 This structure was deduced via 2D NMR using 1 H 1 H COSY and HMBC (Fig. 2) To conrm the absolute conguration of 3, the TDDFT-simulated ECD spectra for 3-6S,9R, 3-6R,9R and 3-6R,9S were generated and compared to the experimental spectrum (Fig. 4) . The observed positive Cotton effect at 238 nm in the ECD of 3 demonstrated the S-conguration of C-6 34 as well as the TDDFT-calculated ECD of 3-6S,9R being in a good agreement with the experimental spectrum (Fig. 4) . From all of these data, 3 was identied as Paronychiarabicastigmane A.

In addition to compounds 1-3, eleven metabolites were isolated and identied as matairesinol 4 0 -O-glucoside (4), 36 leonuriside A (5), 37 arbutine (6), 38 isotachioside (7) (2 0 -hydroxymethyl-2 0 -butenoyloxy) derivative of dehydromelitensin (13) 45 and daucosterin (14) . 46 Megastigmane (3) protein targets associated with SARS diseases were predicted using the SwissTargetPrediction-DisGeNET online tools. One hundred and seventeen genes were identied with genes classied using Venn diagram comparison. Based on the PPI network proles, a STRING webtool and Cytoscape 3.8.0 for visualization were used to explore the predicted gene targets of megastigmane (3) (Fig. 5) . The top 20 genes responding to megastigmane (3) included HRAS, MAPK1, MMP9 and MTOR.

For further investigation of the megastigmane target-function interactions, pathway enrichment analysis and Boolean Network modeling were carried out. The reactome hierarchy map of megastigmane (3) identied disease pathways which were inuenced by the top 20 gene targets responding to megastigmane (3) in terms of SARS-CoV-2 infection (Fig. 6) . Out of the top ten biological pathways resulting from the reactome pathway enrichment analysis, four major biological pathways for megastigmane (3) of cytokine signaling in the immune system, signaling by receptor tyrosine kinases, signaling by interleukins and axon guidance with a high signicance (FDR <0.00001%) were identied (Table 3) .

Investigating the reactome pathway enrichment analysis results highlighted a set of two genes (GSK3 and GSK3B) that represent signicant biological targets in the action of megastigmane (3) as a potent SARS-CoV-2 inhibitor. Additionally, the reactome pathway enrichment analysis for SARS-CoV-2 revealed that GSK3B gene interacts with other genes/interactors including AXIN1, CTNNB1, FRAT1, AKT1, APC, GSKIP, PRKACA, MAPT, GYS1 and IPP2 (Fig. 7) .

Many additional host factors are involved in the SARS-CoV-2 genome transcription/replication. For instance, the coronavirus N protein plays an important role as a RNA chaperone which enables template switching. 47 Importantly, the N protein of SARS-CoV-1 is phosphorylated by the host glycogen synthase kinase 3 (GSK3), and its inhibition was found to signicantly inhibit viral replication. 48

To combat COVID-19, the potentials of the isolated compounds 1-3 as SARS-CoV-2 M pro , PL pro and RdRp inhibitors were predicted. The investigated compounds were docked into the active sites of the SARS-CoV-2 targets using AutoDock4.2.6 soware. 49 The predicted docking scores and the 2D representations of the binding modes of compounds 1-3 inside the active site of SARS-CoV-2 targets are depicted in Fig. 8 .

According to the molecular docking calculations, compounds 1-3 demonstrated better binding affinities towards SARS-CoV-2 M pro compared to those with PL pro and RdRp. Comparing the molecular docking results towards M pro revealed that compound 3 exhibited an outstanding binding affinity towards M pro with a docking score of À9.8 kcal mol À1 , forming seven hydrogen bonds with the key amino acids inside the active site (Fig. 8) . Compounds 1 and 2 showed moderate binding affinities with docking scores of À7.6 and À8.2 kcal mol À1 with M pro , forming three and four hydrogen bonds, respectively.

Optical rotations were measured on a JASCO P-2300 polarimeter (Tokyo, Japan). 1 H (500 MHz), and 13 C NMR (125 MHz) spectra were recorded on a Bruker 500 NMR spectrometer (USA). The chemical shis were given in d (ppm), and coupling constants were reported in Hz. HR-MS spectra were obtained on a JEOL JMS-700 instrument (Tokyo, Japan). Column chromatography (CC) was carried out on polyamide 6L and Sephadex LH 20. Precoated silica gel plates (Merck, Kieselgel 60 F 254 , 0.25 mm, Merck, Darmstadt, Germany) and precoated RP-18 F 254S plates (Merck, Darmstadt, Germany) were used for TLC analysis. A semi-preparative reversed-phase column (Supelco C18 column 250 Â 10 mm, 5 mm) was used for HPLC. 

Following our previous protocol, 50 a black gum (75.6 g) of P. arabica dry material (850 g) was further chromatographed using silica gel column chromatography eluted with a mixture of CHCl 3 /MeOH step gradient. Eight major fractions (PA 1-8) were afforded aer the nal collection, according to the TLC prole.

The elution of fraction PA-4 (1.18 g) over silica gel CC afforded 4 (68.7 mg) and 14 (11.3 mg) along with the subfraction PA-4A-B. The subfraction PA-4B (106.5 mg) was chromatographed on Sephadex LH-20 eluted with CHCl 3 /MeOH (1 : 1) as the solvent system, and afforded 5 (9.8 mg) and 6 (13.6 mg). The subfraction PA-4A (82.6 mg) was eluted by CHCl 3 /MeOH (1 : 1) over Sephadex LH-20, and afforded 7 (11.9 mg), 8 (17.2 mg) and 9 (10.5 mg). The fraction PA-5 (1.32 g) was chromatographed on silica gel CC and eluted with a CHCl 3 /MeOH step gradient, and afforded 10 ( Table 1 . Table 2 .

Two milligrams of 3 in 1 ml 2% H 2 SO 4 solution was heated under reux for 2 h, then the reaction mixture was dried followed by dissolving in distilled H 2 O and neutralizing with NaOH. The neutralized mixture was tested with some sugar moieties such as glucose, rhamnose and others using numerous elution systems over silica gel TLC. 32,33,51

To simulate the circular dichroism (ECD) spectra, a conformational analysis was rst carried out to generated all possible conformations of compounds 1-3 using Omega2 soware. 52 The generated conformations within the energy window value of 10 kcal mol À1 were optimized at the B3LYP/6-31G* level of theory, followed by frequency calculations to estimate the Gibbs free energies. Time-dependent density functional theory (TDDFT) calculations with incorporating a polarizable continuum model (PCM) using methanol as a solvent were performed at the B3LYP/6-31+G* level of theory to calculate the rst y excitation states. SpecDis 1.71 was used to generate ECD spectra for the investigated compounds. 53 Gaussian band shapes with a sigma value of 0.20-30 ev were applied for ECD spectra generation. The generated ECD spectra were Boltzmann-averaged. All quantum mechanical calculations were performed using Gaussian09 soware. 54 

The crystal structures of the SARS-CoV-2 main protease (M pro ; PDB code: 6LU7, 55 papain-like protease (PL pro ; PDB code: 6W9C 56 and RNA-dependent RNA polymerase (RdRp; PDB code: 6M71 57 were taken as templates for the molecular docking calculations. The docking calculations were carried out using AutoDock4.2.6 soware 58 according to our previously described protocol, against SARS-CoV-2 targets. 49, 58, 59 The AutoDock protocol was followed to prepare the pdbqt les for the three investigated SARS-CoV-2 targets. 60 For the three SARS-CoV-2 targets, the binding site was realized by a docking box around the active site with XYZ dimensions of 60Å Â 60Å Â 60 A and a spacing value of 0.375Å. The atomic charges of the isolated compounds were assigned using the Gasteiger method. 61 The built-in clustering analysis with 1.0Å RMSD tolerance was utilized to process the predicted binding positions.

For the pathway enrichment, we initially predicted all of the biological targets for megastigmane (3) as a SARS-CoV-2 inhibitor using the SwissTargetPredicition online tool (http:// www.swisstargetprediction.ch). The DisGeNET database (https://www.disgenet.org) was used to collect the available information on human gene-disease associations (GDAs) for SARS diseases. The protein-protein interaction (PPI) network was developed using the STRING database for the top 20 predicted gene targets. 62 To investigate all possible target-function relations based on the network proles, pathway enrichment analysis was performed for the top 20 SARS diseases-related genes using Cytoscape soware V.3.8.2. 63 Finally, the Cytoscape-based ReactomeFIViz tool was integrated to visualize and model all possible drug-target interactions. 64

The chemical characterization of the aerial parts of P. arabica led to the isolation and identication of two novel oxetane containing lignans, paronychiarabicine A (1) and B (2), one new megastigmane, paronychiarabicastigmane A (3), in addition to eleven known metabolites. The absolute congurations of the new compounds were achieved based on NOESY experiments as well as experimental and TDDFT-calculated electronic circular dichroism. Molecular docking calculations demonstrated the competitive binding affinity of the isolated compound 3 as a prospective SARS-CoV-2 M pro inhibitor.

The authors declare that there are no conicts of interest.

Computational Methods for Processing and Analysis of Biological Pathways

Dr Elshamy gratefully acknowledges the subsidy from the Takeda Science Foundation, Japan. Additionally, this work was supported by the Tokushima Bunri University, Japan and the National Research Centre, Egypt. The computational work was completed with resources supported by the Science and Technology Development Fund, STDF, Egypt, grant nos 5480 & 7972 (granted to Dr Mahmoud A. A. Ibrahim). Prof. Mohamed Hegazy acknowledges the nancial support from the Alexander von Humboldt Foundation "Georg Foster Research Fellowship for Experienced Researchers". The National Research Centre, Egypt is gratefully acknowledged.