RESEARCH ARTICLE

Flavonoids as magical bullets: In silico ADMET profile and molecular docking study against hMPX Virus

Dhananjay Tandon¹*, Surendra Kuamr Gautam¹, Madhavi Tiwari¹, Vikram Singh²

¹Department of Applied Science, Shri Rawatpura Sarkar University, Raipur, Chhattisgarh, 492016, India.

²Department of Botany, Sant Guru Ghasidas Govt PG College Kurud College Dhamtari, Chhattisgarh, 493663, India.

Author’s Email- dhanajaytandon@sruraipur.ac.in, surendrakumargautam@sruraipur.ac.in, madhavitiwari5@gmail.com, vickys.singh11@gmail.com

*Corresponding Author Email- dhanajaytandon@sruraipur.ac.in

Graphical Abstract

Introduction

The human monkeypox virus (hMPXV) was identified in Denmark (1958) on captive imported monkeys. In 1966, it was implicated in a large epidemic at a Rotterdam zoo, thought to have originated from South America before spreading to other ape and monkey species (von Magnus et al. 1959). From 1970 until 2003, the virus was endemic in Central and Western African rainforests. Increased trade, tourism, and travel have since facilitated its spread beyond Africa (Ligon 2004).

The UK Health Security Agency confirmed the first case in a person returning from Nigeria on May 7, 2022 (Sharma 2022). By June 11^th, 2022, 45 countries had reported over 1500 suspected or confirmed cases (Kraemer et al. 2022). Thus, the most recent outbreak was noticed this year due to the lack of an epidemiological relationship and the likelihood of sexual transmission (Lansiaux et al., 2022). Research on its natural reservoir suggests that, while monkeys are unlikely reservoirs, species like rope squirrels and sooty mangabeys may be involved (Moss 2024).

hMPXV has two genomic clades: the Central African/Congo Basin (CA) and West African (WA) clades. The WA clade is linked to milder disease, lower death rates, and less clear human-to-human transmission, while the CA clade is more likely to cause disease (Shchelkunov et al. 2002). Genomic comparisons reveal a nucleotide difference of 0.55–0.56% between the two clades (Chen et al. 2005). The CA clade contains 173 functionally distinct genes, compared to 171 in the WA clade. Despite these differences, both clades share 170 protein-level orthologs and are 99.4% similar (Weaver and Isaacs 2008).

The monkeypox virus is a large member of the Poxviridae family. The virus uses double-stranded DNA as its genetic material. Rodents, rabbits, and nonhuman primates act as potential reservoirs (Joklik et al. 1966). It has virulence genes that help it hide from the host's immune system by making molecules that change how it works (Petersen et al. 2019; Okyay et al. 2022). Transmission occurs via respiratory droplets, direct contact with skin lesions, or contaminated surfaces (Simpson et al. 2020). Although person-to-person transmission is rare, the virus primarily infects humans through animal contact. Once the virus gets in through open mucosa, it multiplies at the site of infection and spreads to lymph nodes during primary viremia. During secondary viremia, it eventually reaches organs far away (Kumar et al. 2022).

The pathogenesis of hMPXV involves entry, fusion, replication, and release. The virus produces external enveloped virions and intracellular mature virions. The release of mature viruses involves a triple membrane with specific antigens (Realegeno et al. 2020). The Conserved Oligomeric Golgi (COG) and Golgi-Associated Retrograde Protein (GARP) complexes are very important to the lifecycle of the virus. They keep the Golgi structure and make it easier for cells to move around inside the intra-Golgi traffics (Realegeno et al. 2017).  The R1 and R2 open reading frames of the CA clade genome are associated with pathogenesis. Deleting these ORFs reduces pathogenicity, slowing viral replication and decreasing mortality in mice (Lopera et al. 2015).

The disease progresses through distinct stages. During the prodromal stage, the virus spreads through secondary viremia, causing swollen lymph nodes and skin rashes. The infected person becomes highly contagious (Okyay et al. 2022). The rash develops in various forms, including papules, blisters, and pustules, eventually crusting and peeling off. Histopathological studies have shown that in the early stages of the rash, there are wounds that are only partially healed, swelling, and the loss of sebaceous glands in the early rash stages (Reynolds et al. 2017). Severe facial rashes can be treated using moist, sealed bandages (Patel et al. 2009).

Similar to COVID-19, a cytokine storm can exacerbate hMPXV infections (Ragab et al. 2020). Elevated levels of interleukins (IL-4, -5, -6, and -10) and suppressed levels of IL-2, -12, TNF-α, and IFN-γ have been observed. When you get the orthopoxvirus, your body's B cells can start making cross-species reactive antigen clones (Johnston et al. 2015).

Phytochemicals are being investigated for their antiviral properties against hMPXV due to the limitations of synthetic antivirals, such as side effects and resistance (Denaro et al. 2020). Plant-based studies have identified several bioactive compounds with antiviral potential (Noor et al. 2022). Drug discovery involves multiple steps, including target identification, compound synthesis, pharmacokinetics, and clinical trials. Computer-aided approaches have streamlined the drug screening process, predicting lead compounds more efficiently (Parikh et al. 2023; Das et al., 2024). In this study, an in-silico approach has been used to discover new compounds targeting hMPXV.

Materials and methods

Receptor preparation

We got the OPG071, B7R, B13R, C2L, COG4, COG7, OPG027, OPG198, OPG103, VPS52, and VPS54 from NCBI. The NCBI annotation service turned the gene code into a protein sequence. The proteins were modelled through the Swiss-Model server homology modelling pipeline (Waterhouse et al. 2018 relies on ProMod3 (Studer et al. 2021), an in-house comparative modelling engine based on OpenStructure (Biasini et al. 2013). The final structure was selected based on a higher value of GMQE and a lower value of X-ray resolution.

Phytochemical data extraction

We obtained 4000 phytochemicals under the flavonoids class from the IMPPAT 2.0 database. We downloaded each phytochemical's Mol2 file format for further use.

Modelling of pharmacophore properties

The modelling was performed using the PharmaGist online tool to filter the phytochemicals based on their physiological properties. The compound having a high score was identified as a hit compound (Schneidman-Duhovny et al. 2008).

Analysis of Drug Likeliness

Screened compounds were passed by the five rules of LIPINSKI (RO5) to screen potential drug-like compounds (Lipinski 2004).

Pharmacokinetic study of ligands

ADMET analysis of the selected compound was performed through the Swiss-ADME web server to analyze pharmacokinetic properties (Daina et al. 2017).

Active site prediction

The target active site was predicted through the GASS Web server (Moraes et al. 2017). The 10 residues were selected to detect the appropriate site. The predicted active sites have been given in Table 1.

Table 1: Active site prediction of target proteins by GASS Web server.

Multiple docking between viral proteins and ligands

A computational ligand-target docking approach was applied to analyze structural complexes of the given viral proteins with phytochemicals to understand the structural basis of this protein target specificity. Docking was performed by the PyRx tool (Dallakyan and Olson 2015). PyRx uses AutoDock 4.2.6 and Auto-Dock Vina as docking software, implying the Lamarckian Genetic Algorithm and Empirical Free Energy Scoring Function. The energy of the protein and the ligands was minimized by CHARMM force field minimization and converted into PDBQT format. The docking was conducted after generating a grid box with coordinates along the X, Y, and Z axes and dimensions. The values of coordination are given in Table 2. The docked structure was analyzed for lowest energy conformer, & hydrogen bond interaction. 3D and 2D interactions profiles were obtained by using BIOVIA Discovery Studio (Sharma et al. 2021).

Table 2: The size measurement of grid box used during molecular docking.

Results and Discussion

Drug-likeness

After pharmacophore modeling, 63 compounds were chosen based on the physiochemical properties, such as their ability to donate and accept hydrogen, their molecular weight, their LogP value, and their topological surface area (TSA), which can be seen in Table 3. 

We filtered the compounds using drug-likeness properties, a set of parameters that indicate the compound's drug-like behavior. The parameters include the Lipinski rule of five, the Ghose rule, the Veber rule, the Egan rule, the GSK 4/400 rule, and the Pfizer 3/75 rule. Among the 63 compounds, 26 passed this test, while the remaining compounds failed (Table 4). The screened compounds are salvigenin, luteolin, chrysoeriol, 6-hydroxyluteolin, 3-O-methylquercetin, agehoustin G, and isokaempferide. 6-Hydroxy-7,8,3',4'-tetramethoxyisoflavanquinone, 7,3',4'-Trihydroxy-3,8-dimethoxyflavone, Chrysin, Chalcone, Cyanidin, Epigallocatechin, Naringetol, Taxifolin, Bergapten, Galanganal, 5-Epidihydroyashabushiketol, 5-Hydroxy-7-(4'-hydroxyphenyl)1-phenyl-3-heptanone, 7-Methoxycoumarin, Genistein, Cirsimaritin, Psoralen, Hesperetin, Isoscopoletin, and 5,6,7,3'-Tetramethoxy-4'-hydroxyflavone.

Salvigenin, Agehoustin G, Chrysin, Chalcone, Bergapten, Galanganal, 5-Epidihydro Yashabushiketol, 5-Hydroxy-7-(4'-hydroxyphenyl)1-phenyl-3-heptanone, 7-Methoxycoumarin, Psoralen, Isoscopoletin, and 5,6,7,3'-Tetramethoxy-4'-hydroxyflavone were unable to pass the Pfizer rule. These molecules are considered drugs because they follow the maximum parameters. 

The pharmacokinetic properties of molecules

To measure the availability and safety of drugs, the pharmacokinetic study analyzed the ADMET of drugs. In the present study, a total of 26 compounds were filtered among 63 compounds based on their ADMET profile. Salvigenin, Luteolin, Chrysoeriol, 6-Hydroxyluteolin, 3-O-Methylquercetin, Agehoustin G, Isokaempferide, 6-Hydroxy-7,8,3',4'-tetramethoxyisoflavanquinone, 7,3',4'-trihydroxy-3,8-dimethoxyflavone, Chrysin, Chalcone, Cyanidin, Epigallocatechin, Naringenin, Taxifolin, Bergapten, Galangin, 5-Epidihydroyashabushiketol, 5-Hydroxy-7-(4'-hydroxyphenyl)1-phenyl-3-heptanone, 7-Methoxycoumarin, Genistein, Cirsimaritin, Psoralen, Hesperetin, Isoscopoletin, and 5,6,7,3'-Tetramethoxy-4'-hydroxyflavone have a good record in ADME profiling, but some of the compounds exhibit toxicity against different cytochrome P450 (CYP) proteins. The result is shown in Table 5. The chemicals that did not follow the drug-likeness property had shown no toxic effects. The reason behind this may be the unavailability of the molecule within the cell.

Binding energy analysis between monkeypox virus proteins and phytochemicals

We used molecular docking to find out how much EEV membrane glycoprotein, early protein, kelch-like protein, host-range protein, part of oligomeric Golgi complex 4 and 7, ser/thr kinase, DNA-directed RNA polymerase, and vacuolar protein sorting-associated protein 52 and 54 bind to 26 different phytochemicals. Table 6 provides the binding energy. Figure 1 provides an image of the best-docked drugs.

There was a drug called Chrysoeriol (IMPHY004433) that fit best with DNA polymerase (8HG1), EEV membrane glycoprotein (B7R), early protein (B13R), COG4 and COG7 components of oligomeric Golgi complexes, DNA-directed RNA polymerase, and VPS52. Drug Galanganal (IMPHY000021) exhibited the lowest binding energy with Ser/Thr kinase (OPG198). The drug Genistein (IMPHY004643) fits with vacuolar protein sorting-associated protein 54 (VPS54) the best.

Table 3: Physiochemical properties of compounds.

Table 4: Drug-likeliness properties of compounds.

Table 5: ADMET properties of compounds.