Isolation, structural elucidation and molecular docking studies against SARS-CoV-2 main protease of new stigmastane-type steroidal glucosides isolated from the whole plants of Vernonia gratiosa

Abstract Phytochemical investigation of the whole plants of Vernonia gratiosa Hance. led in the isolation and identification of two new stigmastane-type steroidal glucosides (1–2), namely vernogratiosides A (1), and B (2). Their chemical structures were fully elucidated based on 1 D/2D NMR spectroscopic, HR-ESI-MS data analyses, and by producing derivatives by chemical reactions. The binding potential of the isolated compounds to replicase protein − main protease of SARS-CoV-2 were examined using the molecular docking simulations. Our results show that the isolated steroidal glucosides (1–2) bind to the substrate‐binding site of SARS-CoV-2 main protease with binding affinities of −7.2 and −7.6 kcal/mol, respectively, as well as binding abilities equivalent to N3 inhibitor that has already been reported (–7.5 kcal/mol). Graphical Abstract


Introduction
Natural products are rich sources of drug discovery, and the development of bioactive constituents from medicinal herbs continues to be a major resource of therapeutic agents for a variety of diseases (Cao et al. 2021). The Vernonia genus is belonging to the family Asteraceae with over a thousand species, many of which are utilized as food and medicine (Toyang and Verpoorte 2013). The genus is distributed both in the New and Old Worlds, although it is usually found in tropical areas. Vernonia species may be found in many different habitats with a wide range of biological diversity and climatic conditions, including tropical forests, marshes and wet places, dry plains, tropical savannahs, desert xeric or dry locations, and even frigid eastern North American regions (Keeley and Jones 1979). Previous studies have revealed the presence of diterpenes, flavonoids, sesquiterpene lactones, and steroidal saponins in various Vernonia species , Jepkoech et al. 2021, Toyang and Verpoorte 2013. These secondary metabolites from the Vernonia genus have been reported to have a variety of pharmacological properties, including antioxidant, anti-inflammatory, anti-malarial, anti-tumor, and anti-diabetic activities (Atangwho et al. 2013, Bordignon et al. 2018, Dogra and Kumar 2015, Nguyen et al. 2021, Zhao et al. 2021. Despite the widespread use of plants of the Vernonia genus in food and medicine, the chemical components and biological activities of Vernonia gratiosa Hance. have not been reported to date.

Results and discussion
The methanolic extract of V. gratiosa whole plants was suspended in water and then partitioned with n-hexane, CH 2 Cl 2 , and EtOAc to afford four layers including n-hexane, CH 2 Cl 2 , EtOAc, and water layer. Combined column chromatography techniques (Silica gel,  of the CH 2 Cl 2 residue from the whole plants of V. gratiosa led to the isolation of 2 new compounds (1 À 2) ( Figure 1). . The 13 C-NMR spectrum of 1 showed 38 carbon atoms, consisting of six carbons belonging to a glucopyranosyl unit at d C 102.37 (CH, C-1 0 ), 75.14 (CH, C-2 0 ), 78.10 (CH, C-3 0 ), 71.69 (CH, C-4 0 ), 77.87 (CH, C-5 0 ), and 62.81 (CH 2 , C-6 0 ) which were also elucidated by COSY, HSQC, and HMBC correlations. Analysis of 13 C-NMR signals was indicated an isopropyl unit at d with the aid of HSQC spectra. The aglycone of 1 was identified as D 7,9(11) dienstigmastane glycoside previously reported from V. amygdalina by clear agreement of its 1 H and 13 C NMR data (Quasie et al. 2016 (Ponglux et al. 1992) suggested the orientation of H-21 of 1 was b-configuration. Moreover, the coupling constant (J ¼ 8.0 Hz) between H-1 0 and H-2 0 indicated the b configuration of the sugar, and acid hydrolysis of compound 1 produced ᴅ -glucose using TLC solvent systems CHCl 3 -MeOH-H 2 O (8:5:1, R f 0.30) and HPLC analysis in comparison with those of authentic samples (Supplementary material). The glucose unit was linked at C-3 of the aglycone was confirmed by HMBC correlation between the anomeric Glc H-1 0 (d H 4.42) and C-3 (d C 78.9). Consequently, the structure of 1 was established as 16a-acetoxy-21b-methoxy-24a-hydroxy-21,23:22,28-diepoxy-5a-stigmastane-7(8),9(11)-dien-3-O-b-D-glucopyranoside, and named vernogratioside A. Vernogratioside B (2), ½a 20 D þ30.6 (c 0.1, MeOH), was afforded as a white amorphous powder. The molecular formula of 2 was determined to be C 37 H 56 O 12 by HR-ESI-TOF-MS analysis with a chlorinated molecular ion peak at m/z 727.   Figure  S14, Supplementary material). Therefore, the structure of 2 was established as 16a-acetoxy-21a,24a-dihydroxy-21,23:22,28-diepoxy-5a-stigmastane-7(8),9(11)-dien-3-O-b-D-glucopyranoside, named vernogratioside B. The genome of SARS-CoV-2 is around 30,000 nucleotides in length, and the replicase gene encodes two overlapping polyproteins, pp1a and pp1ab, that are needed for viral replication and transcription (Wu et al. 2020, Zhou et al. 2020. Extensive proteolytic processing releases functional polypeptides from polyproteins, primarily by the main protease M pro (also known as 3 C-like protease) (Arafet et al. 2020). M pro digests the polyprotein at least 11 conserved sites, beginning with the autolytic cleavage of this enzyme from pp1a and pp1ab (Pillaiyar et al. 2016). The functional importance of M pro in the viral life cycle, as well as the absence of closely similar homologues in humans, make it a potential target for antiviral drug development (Jin et al. 2020). In a recent study, N3 was found to have significant antiviral action against the infectious bronchitis virus in an animal model . Furthermore, N3 can bind to the substrate-binding pocket of SARS-CoV-2 M pro , which is highly conserved among all coronaviruses, using a combination of structure-based virtual and high-throughput screening. The crystal structure of SARS-CoV-2 main protease M pro in complex with N3 inhibitor, which was recently released in Protein Data Bank database (PDB ID: 6LU7), revealed the molecular mechanism of N3 inhibitor against the novel coronavirus (Jin et al. 2020). The crystal structure of the main protease M pro protein from SARS-CoV-2 is highly similar to that of other coronaviruses (Su arez and D ıaz 2020, Zhang et al. 2020). The protein is composed of three domains: domains I (residues 8 À 101) and II (residues 102-184) have an antiparallel b-barrel structure, while domain III (residues 201-303) forms five a-helices arranged into a largely antiparallel globular cluster, connected to domain II by a long linker loop (residues 185-200). The substrate-binding site is located in a cleft between domains I and II, and it holds a histidine/cysteine catalytic dyad (Hilgenfeld 2014, Jin et al. 2020).
The molecular docking study was described with Autodock Vina 1.1.2 and N3 inhibitor was re-docked as a native ligand into the substrate-binding site of SARS-CoV-2 M pro (PDB ID: 6LU7) to validate and optimize docking produce ( Figure S16, Supplementary data). The results of molecular docking including the binding affinities of tested compounds (1-2), root mean square deviation (RMSD) values, and their interactions (hydrogen bond and Van der Waals interaction) were listed in Table S2 (Supplementary data). A full view of two new stigmastane-type steroidal glucosides docking pose to the active site of SARS-CoV-2 M pro was exhibited in Figure S16 (Supplementary data).
The binding affinity of N3 inhibitor was found to be À7.5 kcal/mol and was used to compare the binding of the isolated compounds with SARS-CoV-2 M pro . It formed five hydrogen bond interactions with Gly143, Ser144, Cys145, His163, Gln189 amino acid residues. A grid box was generated around the conserved residues of the substrate-binding site with the major focus on the amino acid residues making polar contacts with the N3 inhibitor. Following that, the isolated compounds (1-2) were then docked using the same grid box. From our docking results, two new steroidal glucosides (1-2) displayed a significant effect on the protein of SARS-CoV-2 with the binding affinity of À7.2 and À7.6 kcal/ mol, respectively. Compound 1 and N3 inhibitor shared the same residues Gly143, Ser144, and Cys145 from the domain II via H-bond interaction; Thr25, Thr26, Leu27, and Met49 from domain I, Phe140, Leu141, Asn142, and Glu166 from domain II via Van der Waals interaction with the SARS-CoV-2 protein. Besides, the hydroxyl group at the sugar moiety of compound 1 established an H-bond interaction with that of Thr190 from the linker loop between domain II and domain III. Compound 2 and N3 inhibitor shared the same residues Gly143, Ser144 from domain II, and Gln189 from the domain II-domain III linker loop via H-bond interaction. Moreover, these inhibitors also established interaction with Thr26, Leu27, and Met49 from the domain I, Phe140, Asn142, Met165, Glu166 from the domain II via Van der Waals interaction with the main protease of SARS-CoV-2. The carbonyl group in both compounds can bind in the oxyanion hole of SARS-CoV-2 M pro (residues 138-146). Compound 2 that differs from compound 1 by a hydroxy group instead of the methoxy group at the position of C-21, exhibited higher activity against SARS-CoV-2 M pro than compound 1. This result revealed that steroidal glucosides from the whole plants of Vernonia gratiosa Hance. had a strong inhibitory effect on SARS-CoV-2 main protease.

Plant material
The whole plants of V. gratiosa was collected in Huong Hoa, Quang Tri province, Vietnam in April 2019. The plant material was identified by Dr. Tran Thi Phuong Anh, a botanist from the Institute of Ecology and Biological Resources, VAST, Vietnam. A voucher specimen (Ver.03.2020) was deposited at the Center of Research and Technology Transfer, and Mientrung Institute for Scientific Research, VAST, Vietnam.

Extraction and isolation
The dried whole plants of V. gratiosa (5.0 kg) was cut into pieces and extracted with MeOH (20 L Â 3 times) at room temperature. The methanolic layer was concentrated using a rotatory evaporator to yield a residue of the MeOH extract (400 g), which was further suspended in H 2 O and successively partitioned with n-hexane, CH 2 Cl 2 , and EtOAc to afford n-hexane extract (H, 70 g), CH 2 Cl 2 extract (DCM, 40 g), EtOAc extract (EA, 42 g), and water layer, respectively.

Molecular docking simulation in SARS-CoV-2 main protease m pro inhibition
Molecular docking simulations were carried out with AutoDock Vina 1.1.2 to understand the interaction and binding of the isolated compounds (1 À 2) with the SARS-CoV-2 main protease M pro protein, following the previously described protocols . The X-ray crystallographic structure of SARS-CoV-2 main protease M pro in complex with an inhibitor N3 was obtained from the RCSB Protein Data Bank database at the resolution of 2.16 Å. The 3 D structures of the isolated compounds (1 À 2) were built using Spartan'18 (Wavefunction Inc., CA, USA). The docking results were analyzed using Discovery Studio 21.1 (Dassault Systemes BIOVIA, CA, USA).

Conclusion
In conclusion, the chemical study on the whole plants of V. gratiosa led to the isolation and structural elucidation of two new stigmastane-type steroidal glucosides, named vernogratiosides A (1), and B (2). The isolated compounds were compared for the SARS-CoV-2 inhibitory activity through SARS-CoV-2 main protease M pro protein in complex with N3 inhibitor by using the molecular docking studies. Two new compounds (1-2) displayed remarkable binding abilities into the substrate-binding site of SARS-CoV-2 M pro with docking affinities of À7.2 and À7.6 kcal/mol, respectively. This is the first report of the chemical investigations on the whole plants of V. gratiosa. Our results and in silico simulations revealed that natural products in general and especially secondary metabolites isolated from V. gratiosa can be used as potential candidates for preventing and treating SARS-CoV-2.

Disclosure statement
All of the authors declare no competing financial interest.