Scientific African 23 (2024) e02109

Contents lists available at ScienceDirect

Scientific African
journal homepage: www.elsevier.com/locate/sciaf

Computational bioprospection of selected plant secondary

metabolites against VP7 (capsid protein) of rotavirus A
Adedayo Ayodeji Lanrewaju , Abimbola Motunrayo Enitan-Folami , Saheed Sabiu *,
Feroz Mahomed Swalaha
Department of Biotechnology and Food Science, Faculty of Applied Science, Durban University of Technology, P. O. Box 1334, Durban 4000, South
Africa

A R T I C L E I N F O A B S T R A C T

Editor: DR B Gyampoh Despite the global implementation of prequalified oral vaccines by the World Health Organization
(WHO) in numerous countries, rotavirus A (RVA) have continued to be the principal cause of
Keywords: acute dehydrating diarrhoea in children under five years of age. Unfortunately, there are no
VP7 epitopes authorized anti-rotaviral medications. Hence, it is crucial to prioritize the development of
Capsid protein
specialized therapeutics to combat rotaviral infections. For the first time, Spondias mombin,
Binding free energy
Macaranga barteri and Dicerocaryum eriocarpum metabolites were screened using computational
Thermodynamic stability
Molecular dynamics simulation techniques to identify potential novel modulators with broad-spectrum activity against VP7
epitopes (capsid protein) of RVA. Compounds with poor pharmacokinetics and drug-likeness were
screened out from the initial top 20 metabolites identified using molecular docking. Thereafter,
molecular dynamics (MD) simulation was used to assess the ability of the resulting compounds to
modulate the selected VP7 epitopes. Remarkably, all the lead compounds had higher negative
binding free energy than tizoxanide across the three epitopes of VP7, with apigenin-4′-glucoside
having the highest affinity for VP7A (− 24.13 kcal/mol) and VP7C (− 43.67 kcal/mol) while the
highest affinity for VP7D was observed in 2SG (− 36.08 kcal/mol). Interestingly, 2SG (− 18.24
kcal/mol, − 31.21 kcal/mol, − 36.08 kcal/mol), apigenin-4′-glucoside (− 24.13 kcal/mol, − 43.67
kcal/mol, − 33.52 kcal/mol) and gnetin L (− 21.21 kcal/mol, − 27.56 kcal/mol, − 32.48 kcal/mol)
had better broad-spectrum affinities for VP7A, C and D relative to tizoxanide (− 10.36 kcal/mol,
− 18.32 kcal/mol, − 12.98 kcal/mol) respectively. Generally, the compounds are thermodynam­
ically stable with 2SG (1.88 Å), ellagic acid (2.45 Å) and sericetin (2.22 Å) being the most stable
against VP7A, C and D respectively. While this study unearths the lead compounds’ promising
ability to modulate the investigated VP7 epitopes, further confirmatory in vitro and in vivo
studies are imperative, and efforts are underway towards achieving this task.

Introduction

Rotavirus (RV) infection is responsible for an estimated 258 million cases of infectious diarrhoea in children under five years,
worldwide [1]. RV-diarrhoea was implicated in approximately 122,000–215,000 deaths of children annually between 2013 and 2017
[2]. Specifically, RVA remains a persistent cause of diarrhoea-mortality among children under five years [3,4]. Even though

* Corresponding author.
E-mail address: [email protected] (S. Sabiu).

https://doi.org/10.1016/j.sciaf.2024.e02109
Received 17 December 2023; Received in revised form 6 January 2024; Accepted 26 January 2024
Available online 2 February 2024
2468-2276/© 2024 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

implementing rotavirus immunization initiatives has resulted in a 60 % decrease in mortality, the vaccine’s effectiveness in developing
countries remains relatively low [5]. To date, the main approach for treating RVA-induced diarrhoea remains oral and intravenous
rehydration; nevertheless, this method only provides symptomatic therapy and does not mitigate the duration or intensity of diarrhoea
[6]. Unfortunately, there is currently no approved and effective antiviral medication to treat RVA infection; hence, there is a need to
prioritize the development of anti-RVA drugs.
Rotavirus, a non-enveloped icosahedral-shaped double-stranded RNA (dsRNA) virus, belongs to the family Reoviridae with a
multilayered virion composed of 11 segments coding for six structural proteins (VP1-4, VP6 and VP7) and six non-structural proteins
(NSP1-NSP6) [7]. The highly immunogenic outer capsid glycoprotein VP7 stimulates the production of neutralizing antibodies and
plays a crucial role in the attachment of the virus to the host cell [8]. Owing to the involvement of VP7 in the virus’ attachment,
pathogenesis and abundance on the viral surface, targeting and destabilizing it could be logical in deactivating RVA prior to its entry
into host cells [9] and makes it an attractive therapeutic target for drug bioprospection and development.
Nature has provided humanity with useful products with low toxicity and minimal side effects to treat illnesses [10]. The possible
antiviral capabilities of several plant species and constituents have been evaluated [11]. Studies have demonstrated the antiviral
properties of Rubia cordifolia [12] and Rindera lanata [13] against RV in vitro. Furthermore, the notable antiviral activity of Spondias
mombin against coxsackie and herpes simplex viruses [14,15] and Macaranga barteri against herpes simplex virus type-1viruses,
echoviruses E7 and E19 [16,17] have been reported. In addition, Dicerocaryum eriocarpum exhibits a wide range of applications in
traditional medicine, such as in treating wounds, facilitation of placenta expulsion, antibacterial activity [18], anti-inflammatory
effects and ethnoveterinary uses [19]. Interestingly, the modulatory potential of the secondary metabolites of the three plants
against VP5* and VP8* of RVA in silico has been reported [20]. Therefore, this study investigated the therapeutic potential of
S. mombin, M. barteri, and D. eriocarpum secondary metabolites against the VP7 protein of RVA to identify promising compounds that
could be further developed as drugs to manage/treat RVA-induced diarrhoea. The conventional approach to drug discovery involves
the identification of potential drug candidates through trial-and-error screening procedure which is often characterized by its lengthy
duration, substantial expenses, and significant likelihood of unsuccessful outcomes [21]. Consequently, by employing computational
techniques, such as modern computer-aided molecular screening tools, researchers have successfully found lead compounds derived
from plant secondary metabolites that exhibit noteworthy antiviral properties [8,22-24]. Therefore, this study forms part of the
continuing efforts to develop anti-RVA drugs and adopted a computational structure-activity relationship approach on the secondary
metabolites of the selected medicinal plants with proven antiviral activity through molecular docking, density functional theory (DFT),
and MD simulation strategies (Fig. 1) to identify broad-spectrum compounds against the VP7 epitopes of RVA.

Materials and methods

Target protein acquisition and optimization

The X-ray crystal structure of VP7 (3FMG) was downloaded from the Protein Data Bank https://www.rcsb.orgs/ (accessed on 26 April

Fig. 1. Workflow of the concept adopted in this study.

2
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

2023). The VP7 protein exhibits a propensity for undergoing structural reconfiguration, while four prominent conserved polypeptide
sequences (epitopes) have been identified within its chain A. Three epitopes (A, C and D) are potentially important for drug discovery and
development due to their involvement in the interactions between the VP7 chains and the VP6 of RVA. By utilizing these conserved se­
quences, it is possible to disrupt the stability of VP7, prevent its interaction with VP6, and circumvent the structural reconfiguration of VP7
[25]. The protein structure was optimized as previously reported [26] using the UCSF Chimera v1.15 software [27].

Optimization of ligand

The 3D structures of the ligands; 77 secondary metabolites reported from the three selected plants (S. mombin, M. barteri and
D. eriocarpum) [28] and the reference standard (tizoxanide) [29] were acquired from Pubchem (https://pubchem.ncbi.nlm.nih.gov/
(accessed on 11 September 2022)). The ligands and reference standard were optimized using the Open Babel plugin in Python Pre­
scription (PyRx) v 0.9.5. and the universal force field (UFF) in 200 steps [30] before being saved in PDBQT format.

Molecular docking and validation

The optimized ligands and prepared protein were docked using PyRx v 0.9.5 [30]. The docking procedure entailed the selection of
amino residues at the active site of the protein of the selected VP7 epitopes as previously reported [8,25] with the x-y-z coordinates;
VP7A [centre (X: 24.02; Y: 96.82; Z: 18.38)], VP7C [centre (X: 3.34; Y: 114.16; Z: − 3.73)] and VP7D [centre (X: − 2.95; Y: 114.16; Z:
19.54)]. The standard was used as a point of reference for comparing the ligands that exhibited the greatest binding affinity at the
receptor binding domain (RBD) of the protein and served as a criterion for choosing ligands with the greatest affinity for the respective
epitopes. Thereafter, the techniques of redocking and superimposition were used to evaluate the docking process, ensuring its pre­
cision and reliability by mitigating the occurrence of false positive results. Briefly, the top-ranked ligands and tizoxanide (reference
standard) were redocked into the RBD of the native protein to evaluate the ability of the AutoDock program to accurately replicate the
ligand’s orientation and location in the crystal structure [8].

Pharmacokinetic assessment

The physicochemical characteristics, pharmacokinetics, and drug-likeness of 20 compounds with the top docking score were
predicted by using SwissADME web (http://swissadme.ch/index.php (accessed on 18 August 2023)). In addition, the Protox II
webserver (https://tox-new.charite.de/protox_II/ (accessed on 18 August 2023)) was used to evaluate their toxicity profiles. The
results of these assessments were employed to reduce the numbers to six, four and six for VP7A, C and D respectively. Subsequently, the
docked complex of the top-scoring compounds with the lowest energy minimization was acquired from their corresponding PDB files
and used for MD simulation.

Quantum chemical calculations

The reactivity of the top-ranked compounds was evaluated by conducting quantum chemical calculations using DFT. The electronic
characteristics of the compounds were evaluated using the DFT approach in Gaussian 16 suite. Owing to its accuracy and lower
computational requirements, the Becke3-Lee–Yang–Parr (B3LYP) method combined with the 6–31 + G(d,p) basis set was employed to
optimize the top-ranked compounds [31]. Gauss View v 6.0 was used to view the output files. The energies of frontier highest occupied
molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO), taking into consideration the Parr and Pearson inter­
pretation [32], were calculated. Thereafter, other chemical descriptors such as energy gap (ΔE), ionization energy (I), electron affinity
(A), chemical hardness (ŋ), softness (δ), electronegativity (χ ), chemical potential (Cp), global electrophilicity (ω) were calculated using
the equations below as described previously [33].
ΔE = ELUMO − EHOMO (1)

I = − ELUMO (2)

A = − EHOMO (3)

(4)

(5)

(I + A)
χ= (6)
2

Cp = − χ (7)

χ2
ω= (8)
ΔE

3
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Fig. 2. The superimposition and interactions of the top-hit compounds at the active site of the crystal structure of (a) VP7A (2SG (green), apigenin-
4′-glucoside (red), glabrescin (brown), glabratephrin (yellow), gnetin L (pink), sericetin (blue) and reference standard (black)) (b) C (2SG (green),
apigenin-4′-glucoside (red), ellagic acid (brown), gnetin L (blue) and reference standard (black)) (c) D (2SG (green), apigenin-4′-glucoside (red),
edulisin I (brown), glabratephrin (yellow), gnetin L (pink), sericetin (blue) and reference standard (black)). (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)

MD simulation

The MD simulation was conducted as previously reported [34]. The AMBER 18 software suite was utilized for the simulation
conducted for 120 ns. The FF18SB variant of the AMBER 18 package was employed to depict the operating systems. The atomic partial
charges of the ligands were generated through the utilization of ANTECHAMBER, which used the GAFF (general amber force field) and
RESP (restrained electrostatic potential) methods for calculating measures, respectively. The systems were neutralized using the Leap
Module’s hydrogen atoms, Na+ and Cl− counter ions. The amino acid residues were numbered accordingly (1–312), and the systems
were contained within a TIP3P orthorhombic box of water molecules to ensure that all atoms were within 10 Å of the box edge. A
restriction potential of 500 kcal/mol was utilized for an initial minimization of 2000 steps in each simulation. The steepest descent
technique was used for 990 steps, followed by 990 steps using conjugate degrees. An additional 990 steps of full minimization without
any limitations were executed using the conjugate gradient method. The simulations involved gradually heating MD systems from 0 to
300 K for 50 ps while ensuring that the number of atoms and volume remained constant. The systems’ solutes were subjected to a
potential harmonic restraint of 10 kcal/mol and a collision frequency of 1.0 ps. Following this, every system was equilibrated for
approximately 500 ps, during which the temperature was maintained consistently at 300 K. The SHAKE algorithm was adopted to
impose restraints on the hydrogen bond atoms in every simulation. In each instance, the simulation was conducted with a 2 fs temporal
resolution, consistent with the NPT (isobaric-isothermal ensemble) using randomised seeding with temperature at 300 K, constant
pressure of 1 bar and a pressure-coupling constant of 2 ps (Langevin thermostat). Thereafter, the 120 ns MD simulation outcomes were
adopted as post-dynamic data.
The post-MD simulation was conducted as previously described [34]. After conducting the simulation, the coordinates of the
systems were combined and analyzed using the AMBER 18 PTRAJ module. The examination of root mean square fluctuation (RMSF),
radius of gyration (ROG), root mean square deviation (RMSD), solvent-accessible surface area (SASA) and hydrogen bonds (H-bonds)
analysis (number, distance and angle) were executed using the CPPTRAJ module and their plots were constructed with Origin v 6.0
[35]. Similarly, the binding free energy was determined using an average of 120,000 snapshots from a 120 ns MD simulation tra­
jectory. The molecular mechanics generalized born surface area (MMGBSA) approach was employed, using the equation ΔGbind =
Gcomplex − (Greceptor + Gligand).

Results and discussion

Molecular docking

The molecular docking approach is often employed to assess the binding interactions of ligands and proteins. This technique
evaluates compounds by their docking scores and discerns the crucial interactions between ligands and the binding domain of a target
protein towards drug discovery and development [36]. In this study, the highest and lowest docking score against VP7A, VP7C and
VP7D was observed in (2S)-6-(gamma,gamma-dimethylallyl)-3′,4′-dimethoxy-6″,6″-dimethylpyran[2″,3″:7,8]flavanone (2SG) (− 7.2
kcal/mol) and 2-Isopropylmalic acid (− 3.9 kcal/mol, mallotusinic acid (− 8.8 kcal/mol) and bis(ethoxycarbonyloxymethyl) dodec­
anedioate (− 4.6 kcal/mol), 3-(2,2-Dimethylpropoxymethoxy)-2-oxopropyl] 2-methylprop-2-enoate; [3-(3-methylbutan-2-ylox­
y)-2-oxopropyl] 2-methylprop-2-enoate; [3-(oxan-2-yloxy)-2-oxopropyl] 2-methylprop-2-enoate (3DO) (− 8.1 kcal/mol) and
2-Isopropylmalic acid (− 3.9 kcal/mol) respectively (Supplementary Table S1). The top twenty compounds with higher docking
scores ranging from − 7.2 to − 6.2 kcal/mol (VP7A), − 8.8 to − 7.4 kcal/mol (VP7C) and − 8.1 to 6.8 kcal/mol (VP7D) compared to the

4
 A.A. Lanrewaju et al.
Table 1
The cDFT parameters of top-ranked compounds against VP7A, C and D.
cDFT parameters (eV)

Ligands LUMO HOMO Energy gap Ionization energy Electron affinity Hardness Softness Electronegativity Chemical potential Global electrophilicity
5

Edulisin I − 1.89 − 6.07 4.19 1.89 6.07 2.09 0.48 3.98 − 3.98 3.78
Glabratephrin − 1.99 − 6.52 4.52 1.99 6.52 2.26 0.44 4.26 − 4.26 4.01
Glabrescin − 1.88 − 5.51 3.64 1.88 5.51 1.82 0.55 3.69 − 3.69 3.76

Scientific African 23 (2024) e02109

A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

standard [tizoxanide (− 5.5 kcal/mol), (− 5.5 kcal/mol) and (− 5.8 kcal/mol)] targeted against VP7A, C and D were selected respec­
tively (Supplementary Table S2). The compounds’ higher docking scores compared to the reference standard may indicate their po­
tential affinity for the respective epitopes.

Verification of the docking protocol

The docking procedure pose was validated by superimposing the top-ranked compounds on the crystal structure of the native VP7
protein, though it lacks a co-crystallized ligand. Fig. 2 illustrates the overlapping and interplay between the highest-ranked compounds
and the reference standard at the binding domain of the examined VP7 epitopes. All the top-hit compounds shared similar binding
orientations against VP7A epitope except tizoxanide on the VP7 crystal structure (Fig. 1(a)). Of all the residues at the active site of the
VP7A epitope (Lys194-Cys207), the top-ranked compounds shared a similar binding pose and interacted with Val195, Pro197, Leu
198, Asn199 and Thr200. Conversely, tizoxanide was observed to have partial occupancy at the binding pocket of the VP7A epitope,
and interacted with Pro197, Ile205 and Gly206. Similarly, against VP7C, 2SG, apigenin-4′-glucoside and gnetin L shared a similar
binding orientation interacting with Gln280, Glu282, Met284, Met285, Arg286 of all the residues at the active site of VP7C epitope
(Ala273–Arg286) while ellagic and tizoxanide having a partial occupancy at the binding pocket of the VP7C interacted with Gln280,
Gln282 and Met284 (Fig. 1(b)). Remarkably, all the top-ranked compounds and the reference standard had similar binding pose
interacting with Tyr302, Gln305, Ile306, Ala309 and Lys312 of all the amino acids at the binding site of VP7D (Val300–Lys312) (Fig. 1
(c)). Interestingly, the observed amino acids interactions in VP7A, C and D against the investigated compounds are consistent with
those reported in previous studies [8,25]. This observation establishes that the docking scores exhibited by the compounds and the
reference standard could be a function of their interactions with the residues at the active site of the examined VP7 epitopes and
consequently confirms the reliability of the employed docking procedure.

Prediction of pharmacokinetics, drug-likeness, physicochemical and toxicity properties

Apart from establishing favourable interaction with the intended target, an efficacious therapeutic molecule should possess good
pharmacokinetic attributes while minimizing potential toxicities and adverse effects [37]. The identification of molecules with sig­
nificant biological activity and favourable pharmacokinetic characteristics constitutes a paramount concern in drug design [38].
Despite the significant biological activity shown by several compounds, their therapeutic use is hindered by inadequate absorption,
distribution, metabolism, excretion and toxicity (ADMET) characteristics. Therefore, ADMET prediction has emerged as a prominent
research focus within the realm of computer-aided drug molecular design [39], owing to its significance in meeting the demands of
drug design and reduction of several probable safety issues, thereby accelerating the drug development procedure.
According to Lipinski’s rule of five (Ro5), bioactive molecules are required to adhere to limitations on their molecular weight (500
g/mol), hydrogen donor (5), hydrogen acceptor (10), and octanol co-efficient (5) as molecules with less than two violations would be
orally accessible [40]. Of all the twenty top-scoring compounds against VP7A selected for ADMET analysis, 2SG, 6 MM, glabratephrin,
gnetin L, sericetin, apigenin-4′-glucoside and glabrescin passed the Ro5 with 0, 1, 0, 0, 0, 1 and 0 violations respectively (Table 1) while
gnetin L, 2SG, apigenin-4′-glucoside and ellagic acid passed the Ro5 with 0, 0, 1 and 0 violations respectively. Similarly, gnetin L, 6
MM, edulisin I, sericetin, 2SG, glabratephrin and apigenin-4′-glucoside passed the Ro5 with 0, 1, 0, 0, 0, 0 and 1 violations respectively,
of all the top-scoring compounds against VP7D (Table 1) while tizoxanide (reference standard) passed the Ro5 with 0 violation.
Therefore, these compounds with less than 2 violations display their potential to be orally bioavailable, indicating their capacity to
reach target cells, tissues or organs and exert their pharmacological effects.
The assessment of a drug’s bioavailability score is a crucial pharmacokinetic characteristic that must be considered when deter­
mining the right dosage for a medication. The quantity of the unaltered drug quantity that reaches the systemic circulation, based on a
specific rate of absorption, is determined by the bioavailability score of a medication [41]. All the top-scoring compounds against the
investigated targets that passed the Ro5 and tizoxanide (reference standard) had a bioavailability score of 0.55, indicating their po­
tential to be developed as oral drugs [42].
To evaluate the toxicity profile of drugs, cytochrome P450 (CYP450), a key isoenzyme in drug metabolism, is crucial [43]. In this
study, only apigenin-4′-glucoside is not an inhibitor of all the CYP450 isoenzymes, suggesting its inability to initiate any drug-drug
interactions when co-administered with other drugs normally metabolised by the CYP450 isoenzymes of all the top-ranked com­
pounds across the investigated targets. Conversely, other top-ranked compounds inhibited at least one or more CYP450 isoenzymes,
indicating that they could interact with other medications processed by such enzymes. Furthermore, all the top-ranked compounds
were active for at least one toxicity endpoint or the other (Supplementary Table S2).
In addition, the top-ranked compounds belonged to drug toxicity classes 4 and 5, suitable for drug development [44], however
6MM belonged to drug toxicity class 2, which is inappropriate for drug development; hence, it was excluded from subsequent in­
vestigations. These findings indicate the need for structural alterations of the top-ranked compounds to improve their druggability. In
general, the top-ranked compounds exhibited favourable medicinal and pharmacokinetic characteristics, possess drug-like properties,
and could be modified to enhance their suitability into being developed as drugs while reducing their toxicity.

Frontier molecular orbital properties

The cDFT parameters of the top-ranked compounds were calculated to investigate their molecular properties that could be of
therapeutic significance. The HOMO and LUMO orbitals are widely acknowledged as significant indicators for predicting the chemical

6
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

and biological reactivity of chemical species. The HOMO is the electron-filled outermost orbital, which donates an electron to the
protein to make a bond that blocks the active site of the protein in pathogens. Conversely, the LUMO is the first empty innermost orbital
that is not filled with electrons, which acts as an electron acceptor or positive charge carrier to heavier parts where the nucleophilic
part of a protein can be attacked [45]. A compound exhibiting a larger energy gap (LUMO-HOMO) indicates the less reactive nature of
the compound, while a smaller energy gap corresponds to increased reactivity and heightened biological activity of the compound
[46]. The cDFT parameters of 2SG, sericetin, apigenin-4′-glucoside, ellagic acid, gnetin L and tizoxanide have been previously reported
[20], while that of edulisin I, glabratephrin and glabrescin are presented in Table 1.
The lowest energy gap was observed in sericetin (3.57 eV) and could have contributed to its reactivity, while glabratephrin (4.52
eV) exhibited the largest energy gap (Fig. 3), and could have been responsible for its low binding free energy after tizoxanide against
VP7A and D. Chemical hardness linked to reactivity, measures the ability of a compound to withstand alterations in electron distri­
bution. This parameter serves as a dependable measure for assessing chemical stability and holds significant importance in drug design
[47]. An elevated level of chemical softness and a diminished level of chemical hardness are indicative of increased compound
reactivity [48]. Therefore, the reactivity of sericetin against VP7A and D could be attributed to its chemical softness and hardness being
the highest and lowest, respectively.
Another quantitative parameter is electronegativity, which measures the relative distribution of electrons in a molecule. The
highest electronegativity value was observed in tizoxanide (5.07 eV), while ellagic acid (4.37 eV), glabratephrin (4.26 eV) and api­
genin-4′-glucoside (4.23 eV) are next to tizoxanide against VP7 (Table 1). Remarkably, the lead compounds possess elevated elec­
tronegativity levels, indicating their effectiveness in accepting electrons. In addition, the electrophilicity index characterizes the
propensity of an electrophile to attract a specific quantity of electron density and the capacity of a molecule to resist the transfer of
electron density [48]. Molecules are classified into low (<0.8 eV), moderate (0.8–1.5 eV) and heavy (>1.5 eV) electrophiles based on
their electrophilicity values [49]. Remarkably, all the top-hit compounds and tizoxanide are heavy electrophiles, thus implying a
substantial presence of electrophiles.

MMGBSA-based binding free energy of the top-ranked compounds

Generally, the molecular docking approach ranks the feasible poses of a ligand in a binding pocket based on a scoring function, and
the accuracy of these scoring functions has been improved using the MMGBSA approach [50]. This approach facilitates the integration
of high-throughput MD simulations with the determination of binding free energies associated with protein-ligand interactions [51].
The thermodynamic binding free energy is a measure of the energy difference between a complex and its unbound components, as a
more negative value indicates a stronger affinity of the ligand towards the protein [52]. In this study, all the lead compounds had
higher negative binding free energy than tizoxanide across the three epitopes of VP7 with apigenin-4′-glucoside having the highest
affinity for VP7A (− 24.13 kcal/mol) and VP7C (− 43.67 kcal/mol) while the highest affinity for VP7D was observed with 2SG (− 36.08
kcal/mol). Interestingly, 2SG (− 18.24 kcal/mol, − 31.21 kcal/mol, − 36.08 kcal/mol), apigenin-4′-glucoside (− 24.13 kcal/mol,
− 43.67 kcal/mol, − 33.52 kcal/mol) and gnetin L (− 21.21 kcal/mol, − 27.56 kcal/mol, − 32.48 kcal/mol) had better broad-spectrum
affinities for VP7A, C and D relative to tizoxanide (− 10.36 kcal/mol, − 18.32 kcal/mol, − 12.98 kcal/mol) respectively (Table 2). This
observation is in tandem with our previous study, where the secondary metabolites of selected medicinal plants exhibited different
capacities at modulating both the VP5* and VP8* of RVA and specifically with sericetin and 2SG having a broad-spectrum affinity
against both targets [20]. In general, the binding free energy results demonstrated the superiority of the top-ranked compounds over
tizoxanide towards the management of RVA-induced diarrhoea.

Fig. 3. Frontier molecular orbitals for (a) glabrescin (b) edulisin I (c) glabratephrin.

7
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Stability, fluctuation compactness, protein surface conformation and H-bonds analysis of the simulated complexes

The RMSD is an essential metric used in the assessment of structural variations in protein structures, providing insights into the
stability of the protein upon ligand binding over a certain period [53]. A reduced average RMSD value indicates enhanced stability of a
protein-ligand combination during simulation, which implies a favourable capacity of the ligand to modulate the protein [54]. In this
study, each complex exhibited unique pathways and varied patterns of variation, resulting in varying mean RMSD values after
equilibrating at 4 ns, 10 ns and 3 ns for VP7A, C and D complexes, respectively (Fig. 4). Sericetin-VP7A and glabrescin-VP7A exhibited
the highest fluctuation at 37 ns (4.9 Å) and 68 ns (5.8 Å) against VP7A while gnetin L-VP7C and 2SG-VP7C exhibited the highest
fluctuation at 84 ns (6.1 Å) and 116 ns (5.4 Å) against VP7C (Fig. 3(a) and (b)). Also, the highest fluctuation against VP7D was observed
in gnetin L-VP7D at 26 ns (6.0 Å) and edulisin I-VP7D at 75 ns (5.8 Å) (Fig. 4(c)). Expectedly, the observed fluctuations resulted in
increased mean RMSD values in the highlighted ligands against VP7A (sericetin-3.56 Å and glabrescin-3.66 Å), VP7C (gnetin L-3.87 Å)
and VP7D (gnetin L-4.14 Å and edulisin I-3.66 Å) except 2SG-VP7C (2.69 Å) which could be due to the short fluctuation moment
exhibited by its complex during the simulation as opposed to other ligands (Table 3). Unfortunately, the observed increased average
RMSD values were marginally above the acceptable 3 Å limit [55], suggesting probable reduced stability of those complexes. Other
complexes with slightly higher average RMSD values above the acceptable limit of 3 Å but with considerable fluctuation patterns
include VP7A (apigenin-4′-glucoside and glabratephrin) and VP7D (2SG, apigenin-4′-glucoside and tizoxanide). The lowest RMSD
value against VP7A was observed in 2SG-VP7A (1.88 Å), followed by gnetin L-VP7A (2.33 Å), tizoxanide-VP7A (2.82 Å),
apigenin-VP7A (3.09 Å) and glabratephrin-VP7A (3.11 Å) all of which are lower than the apo-VP7A (3.42 Å) suggesting their ability to
confer reasonable stability on the protein upon binding (Table 3). Against VP7C, ellagic acid-VP7C (2.45 Å) had the lowest average
RMSD value followed by 2SG-VP7C (2.69 Å), which are lesser than that of the unbound VP7C (2.87 Å) indicating their potential to
enhance the stability of the protein. In addition, both glabratephrin-VP7D (2.82 Å) and sericetin-VP7D (2.22 Å) have lower mean
RMSD values relative to the reference standard-tizoxanide (3.20 Å) and unbound VP7D (2.99 Å) (Table 3). This observation cor­
roborates that of Aribisala et al. [21], in which ligand binding unto ParE resulted in reduced average RMSD value relative to the
unbound ParE. This finding further suggests that these metabolites are better inhibitors compared to the reference standard and
enhance the stability of the protein compared to the apo-protein. Conversely, apigenin-4′-glucoside-VP7C (2.91 Å) had a higher
average RMSD value than the apo-protein (2.87 Å) (Table 3), but it is still within the 3.0 Å acceptable limit. This finding aligns with a
previous study that revealed an increase in the mean RMSD value due to the binding of verbascoside and the SARS-CoV-2 spike protein
[52]. Generally, the average RMSD values of the complexes are in-consistent with the binding free energy results against the three
investigated epitopes however, most of the reported RMSD values were within the acceptable 3 Å limit while the complexes with RMSD
values above the 3 Å limit show minor deviations from their respective unbound proteins.
The RMSF was evaluated to comprehend the structural flexibility of the amino acids at the RBD of the VP7 epitopes upon the
binding of ligands (Table 3). This metric considers the mean volatility of atoms and residues in a receptor during a simulation period

Table 2
MMGBSA-based binding free energy framework (kcal/mol) of the top-ranked compounds against VP7A, C and D.
Energy components (kcal/mol)

Ligands ΔEvdW ΔEelec ΔGgas ΔGsolv ΔGbind

VP7A

2SG − 23.36 ± 6.95 − 3.74 ± 4.30 − 27.09 ± 8.39 8.85 ± 4.27 − 18.24 ± 6.31
Apigenin-4′-glucoside − 29.38 ± 5.30 − 24.58 ± 11.80 − 53.96 ± 12.11 29.83 ± 9.02 − 24.13 ± 5.54
Glabrescin − 25.87 ± 5.15 − 6.94 ± 8.01 − 32.81 ± 9.85 13.34 ± − 19.47 − 19.47 ± 4.54
Glabratephrin − 21.35 ± 3.97 − 2.21 ± 4.74 − 23.57 ± 6.28 8.07 ± 4.55 − 15.50 ± 3.52
Gnetin L − 26.78 ± 3.57 − 11.25 ± 10.56 − 38.05 ± 11.47 16.83 ± 8.75 − 21.21 ± 4.57
Sericetin − 24.13 ± 8.11 − 5.56 ± 5.24 − 29.69 ± 11.58 11.26 ± 5.06 − 18.42 ± 7.15
Tizoxanide − 14.05 ± 6.37 − 240.47 ± 70.59 − 254.53 ± 74.45 244.17 ± 70.64 − 10.36 ± 5.10

VP7C

2SG − 39.72 ± 3.69 − 3.57 ± 3.69 − 43.29 ± 5.20 12.09 ± 3.20 − 31.21 ± 3.63
Apigenin-4′-glucoside − 39.65 ± 3.68 − 59.91 ± 10.27 − 99.56 ± 9.91 55.89 ± 7.67 − 43.67 ± 3.85
Ellagic acid − 22.16 ± 4.95 − 57.33 ± 12.80 − 79.50 ± 10.80 47.80 ± 7.91 − 31.69 ± 4.30
Gnetin L − 33.55 ± 8.47 − 20.63 ± 10.47 − 54.20 ± 11.78 26.64 ± 7.90 − 27.56 ± 7.28
Tizoxanide − 20.71 ± 5.40 − 276.39 ± 29.00 − 297.10 ± 28.76 278.79 ± 26.72 − 18.32 ± 5.71

VP7D

2SG − 43.40 ± 5.36 − 7.36 ± 3.42 − 50.76 ± 6.66 14.67 ± 2.84 − 36.08 ± 5.46
Apigenin-4′-glucoside − 30.41 ± 4.72 − 60.02 ± 16.94 − 90.43 ± 16.76 56.91 ± 11.44 − 33.52 ± 6.46
Edulisin I − 37.02 ± 2.86 − 9.87 ± 6.65 − 46.90 ± 6.61 19.48 ± 5.52 − 27.42 ± 2.83
Glabratephrin − 29.54 ± 4.65 − 17.65 ± 5.24 − 47.19 ± 8.63 24.71 ± 5.25 − 22.48 ± 4.44
Gnetin L − 37.23 ± 3.46 − 22.27 ± 8.95 − 59.52 ± 8.58 27.04 ± 6.55 − 32.48 ± 3.73
Sericetin − 30.89 ± 4.87 − 2.34 ± 4.34 − 33.24 ± 7.02 9.26 ± 3.88 − 23.98 ± 5.13
Tizoxanide − 18.80 ± 7.35 − 181.63 ± 44.02 − 200.42 ± 47.86 187.45 ± 43.88 − 12.98 ± 6.10

ΔEvdW = van der Waals energy; ΔEelec = electrostatic energy; ΔEgas = gas phase free energy; ΔGsol = solvation free energy; ΔGbind = total binding free
energy.

8
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Fig. 4. Comparative RMSD plots of alpha-carbon, top-ranked compounds and reference standards against (a) VP7A (b) C and (c) D after 120 ns MD
simulation period.

and can be associated with their ability to form stable intra and intermolecular bonds [56]. A reduction in volatility at the active site
where ligand binding and catalysis take place corresponds to an increase in the strength of the binding and affinity between the ligand
and the protein [54]. In this study, a relatively higher fluctuation between residues 180 and 245, 265 and 280 was observed in VP7A
(Fig. 5(a)), suggesting the limited capacity of these residues to form stable bonds. Unfortunately, all the residues at the active site of
VP7A were situated in the region of higher fluctuation (Fig. 5(a)); hence, the average RMSF values at the active site were higher than
the whole VP7A (Tables 3 and 4). The unbound VP7A, tizoxanide, sericetin, gnetin L, glabratephrin, glabrescin, apigenin-4′-glucoside
and 2SG had RMSF values of 3.05 and 1.74 Å, 2.22 and 1.67 Å, 2.01 and 1.83 Å, 2.39 and 1.74 Å, 1.97 and 1.57 Å, 4.81 and 2.53 Å, 3.19
and 2.11 Å, 2.09 and 1.70 Å respectively at the active site and in the whole protein (Tables 3 and 4). During the 120 ns simulation
period, majority of the protein’s active site fluctuation could be attributed to residues Asn199, Thr200, Gln201, Thr202, Leu203,
Gly204, Ile205, Gly206 and Cys207 (Table 4).
Residues with high levels of fluctuations indicate increased flexibility and the presence of unstable bonds, while residues with
minimal fluctuations are linked with less flexibility and enhanced structural stability [57]. The lesser RMSF value observed at the
active site and whole VP7A protein upon ligand binding in gnetin L-VP7A, glabratephrin-VP7A and 2SG-VP7A compared to the un­
bound VP7A suggests the potential affinity of these compounds for VP7A. This observation implies that these compounds could reduce
flexibility and form stable interactions with the amino acid residues of VP7A. However, both apigenin-4′-glucoside-VP7A and
glabrescin-VP7A had higher mean RMSF values at the active site of the protein compared to the unbound VP7A, while higher mean
RMSF values were observed in sericetin-VP7A and glabrescin-VP7A upon binding to the whole protein relative to the apo-VP7A and
relatively consistent with the RMSD results. Glabratephrin-VP7A had the least RMSF value in the whole protein (1.57 Å) and at the
active site (1.97 Å) of VP7A respectively, indicating its ability to form stable bonds and could be a better inhibitor of the epitope than
other compounds. This observation opposes the binding free energy and RMSD results in which the lowest binding free energy was
observed in apigenin-4′-glucoside-VP7A and 2SG-VP7A, respectively. Nonetheless, the mean RMSF value of 2SG-VP7A in the whole
protein (1.70 Å) and at the active site (2.0 Å) is within the 3.0 Å satisfactory boundary, while apigenin-4′-glucoside-VP7A has 3.09 Å
and 3.19 Å (Tables 3 and 4) which does not deviate much from the 3.0 Å satisfactory value indicating their potential as a good inhibitor
of the VP7A epitope. Relatively reduced fluctuations in both the active site residues and the whole protein due to ligand binding
suggest improved intra- and intermolecular bond stability. They are in tandem with the RMSD results, thereby suggesting the potential
of the investigated compounds as VP7A epitope inhibitors. Leu198 had the lowest fluctuation of the residues at the active site of VP7A
(Table 4), hence an indication of its significance in forming stable bonds with the studied compounds during the simulation.
Furthermore, a higher fluctuation was observed between residues 195 and 220, 260 and 280 in VP7C. Of the residues at the active
site of VP7C, Ala273-Gln280 were situated in the region of high fluctuation (Fig. 5(b)), and this could have led to the higher average
RMSF values at the active site relative to the whole VP7C. The unbound VP7C, tizoxanide, gnetin L, ellagic acid, apigenin-4′-glucoside
and 2SG had mean RMSF values of 2.47 and 1.75 Å, 2.91 and 1.82 Å, 4.37 and 2.05 Å, 2.82 and 1.75 Å, 1.96 and 1.83 Å, 4.05 and 1.70
Å at the active site and in the whole protein respectively (Tables 3 and 5). Except for 2SG-VP7C complex, ligand binding unto VP7C
resulted in increased RMSF value relative to the unbound VP7C; however, the average RMSF values recorded in the whole protein are
all within the acceptable limit of 3.0 Å, indicating that the fluctuations in residues were conducive to maintaining stability. The lowest
average RMSF value at the active site of VP7C exhibited by apigenin-4′-glucoside is consistent with its highest negative binding free
energy, suggesting its potential as a better inhibitor compared to other compounds. The least fluctuation in the residues at the active
site of VP7C was observed in Arg283, which portray its importance in the formation of stable bond with the metabolites.
Similar to VP7A, a relatively higher fluctuation was observed between residues 195 and 220, 260 and 280 in VP7D; however, none
of the residues at the RBD is in the region of higher fluctuation (Fig. 5(c)) and consistent with expectations, the RMSF values of the
amino acid residues at the active site were lesser compared to the values obtained for the whole protein. This finding indicates
enhanced intra and intermolecular bonding at the binding domain of VP7D throughout the simulation. The unbound VP7D, tizoxanide,
sericetin, gnetin L, glabratephrin, edulisin I, apigenin-4′-glucoside and 2SG had RMSF values of 1.38 and 1.59 Å, 1.43 and 2.04 Å, 1.51
and 2.06 Å, 1.38 and 1.98 Å, 1.47 and 2.17 Å, 1.44 and 1.77 Å, 1.41 and 1.73 Å, 1.72 and 2.04 Å, at the active site and in the whole
protein respectively (Tables 3 and 6). The lowest mean RMSF value in the whole protein after the apo-protein was recorded in

9
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Table 3
Mean RMSD, RMSF, ROG, SASA and hydrogen bonds of apo-VP7A, C and D proteins, top-ranked compounds, and the standard against VP7A, C and D.
Ligands Mean RMSD Mean RMSF Mean ROG Mean SASA (Å) Mean Number of H- Mean distance (Å) of Mean angle (◦ ) of
(Å) (Å) (Å) bonds H-bonds H-bonds

VP7A

2SG 1.88 ± 0.34 1.70 ± 1.02 20.84 ± 12,598.49 ± 104.81 ± 7.04 2.85 ± 0.06 151.89 ± 7.67
0.20 306.47
Apigenin-4′- 3.09 ± 1.02 2.11 ± 0.83 20.64 ± 12,759.30 ± 107.25 ± 7.74 2.85 ± 0.06 152.27 ± 7.83
glucoside 0.30 241.39
Glabrescin 3.66 ± 1.18 2.53 ± 1.20 20.42 ± 12,978.32 ± 102.52 ± 6.81 2.85 ± 0.06 152.27 ± 7.58
0.44 324.86
Glabratephrin 3.11 ± 0.28 1.57 ± 0.60 20.98 ± 12,796.71 ± 102.81 ± 6.97 2.85 ± 0.06 152.17 ± 7.54
0.13 302.97
Gnetin L 2.33 ± 0.57 1.74 ± 0.64 21.16 ± 12,904.61 ± 103.32 ± 6.90 2.85 ± 0.06 152.17 ± 7.54
0.22 260.74
Sericetin 3.56 ± 0.45 1.83 ± 1.27 20.90 ± 12,958.84 ± 104.93 ± 6.98 2.85 ± 0.06 151.79 ± 7.69
0.17 309.23
Tizoxanide 2.82 ± 0.37 1.67 ± 1.35 20.68 ± 12,535.76 ± 109.51 ± 6.90 2.85 ± 0.06 151.89 ± 7.58
0.17 280.15
VP7A 3.42 ± 0.77 1.74 ± 1.11 20.85 ± 12,864.65 ± 102.85 ± 7.12 2.85 ± 0.06 152.05 ± 7.80
0.17 301.64

VP7C

2SG 2.69 ± 0.96 1.70 ± 1.14 20.93 ± 13,101.72 ± 103.24 ± 6.94 2.85 ± 0.06 151.77 ± 7.74
0.18 307.33
Apigenin-4′- 2.91 ± 0.56 1.83 ± 0.68 20.71 ± 12,799.93 ± 107.73 ± 7.23 2.85 ± 0.06 151.88 ± 7.57
glucoside 0.18 314.38
Ellagic acid 2.45 ± 0.51 1.75 ± 0.67 20.85 ± 12,585.80 ± 105.23 ± 6.98 2.85 ± 0.06 151.98 ± 7.68
0.17 270.79
Gnetin L 3.87 ± 1.07 2.05 ± 1.43 20.91 ± 12,985.99 ± 104.37 ± 7.02 2.85 ± 0.06 152.30 ± 7.56
0.23 299.64
Tizoxanide 3.55 ± 0.55 1.82 ± 0.80 20.76 ± 12,839.12 ± 107.65 ± 7.25 2.85 ± 0.06 152.17 ± 7.64
0.16 270.76
VP7C 2.87 ± 0.56 1.75 ± 0.61 20.92 ± 12,791.76 ± 101.25 ± 7.04 2.85 ± 0.06 152.13 ± 7.60
0.21 253.44

VP7D

2SG 3.07 ± 0.64 2.04 ± 0.91 20.87 ± 13,005.91 ± 103.73 ± 7.16 2.85 ± 0.06 152.04 ± 7.60
0.18 290.79
Apigenin-4′- 3.12 ± 0.45 1.73 ± 0.79 20.67 ± 12,622.74 ± 110.51 ± 7.50 2.85 ± 0.06 152.24 ± 7.60
glucoside 0.20 280.81
Edulisin I 3.66 ± 0.86 1.77 ± 0.89 20.99 ± 12,957.04 ± 100.97 ± 7.36 2.85 ± 0.06 152.09 ± 7.72
0.19 255.83
Glabratephrin 2.82 ± 0.85 2.17 ± 0.83 20.87 ± 12,992.45 ± 105.76 ± 6.90 2.85 ± 0.06 152.07 ± 7.76
0.23 251.45
Gnetin L 4.14 ± 0.50 1.98 ± 1.13 20.74 ± 12,995.75 ± 105.65 ± 6.98 2.85 ± 0.06 152.20 ± 7.62
0.15 255.13
Sericetin 2.22 ± 0.51 2.06 ± 0.78 20.94 ± 12,904.16 ± 105.14 ± 7.23 2.85 ± 0.07 151.76 ± 7.53
0.21 293.79
Tizoxanide 3.20 ± 0.89 2.04 ± 1.42 20.90 ± 12,972.23 ± 106.90 ± 7.44 2.85 ± 0.07 151.82 ± 7.77
0.21 378.50
VP7D 2.99 ± 0.41 1.59 ± 0.60 20.67 ± 12,773.99 ± 105.75 ± 6.94 2.85 ± 0.06 151.87 ± 7.61
0.14 226.46

apigenin-4′-glucoside; however, at the active site, gnetin L had the lowest RMSF value (Table 6). However, the average RMSF value in
both the whole protein and at the binding domain is still below the acceptable border of 3.0 Å, indicating the possibility of the
compounds under study as good inhibitors. The lowest fluctuation in the active site of VP7D was found in Ile306, hence its contribution
towards the stability of the investigated metabolites (Table 6). Among the lead compounds, glabratephrin-VP7A, 2SG-VP7C and
apigenin-4′-glucoside-VP7D had the lowest average RMSF values in the whole protein against VP7A, C and D, respectively, indicating
their potential to induce minimal distortion of the targets during the simulation however, it contradicts their free binding energy
results and average RMSD values. Generally, ligand binding resulted in increased average RMSF values across the three studied
epitopes and this contradicts Aribisala et al. [54], in which phenolics binding to PBP3 and PBP5 yielded reduced average RMSF values
relative to the unbound protein.
Apart from the RMSF, ROG assesses the level of compactness and folding shown by a complex during a simulation. The presence of
high compactness and folding in a complex indicates a higher degree of thermodynamic orderliness and, in some cases, instability. The
VP7A system exhibited reduced fluctuations in compactness during the 120 ns simulation period (Fig. 6(a)), suggesting that the
binding of the lead compounds and tizoxanide to apo-VP7A had little impact on its folding except in glabrescin-VP7A complex with
noticeable protein unfolding between 60 and 85 ns indicating lower compactness within this period. Furthermore, only glabratephrin-

10
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Fig. 5. Comparative RMSF plots of alpha-carbon, top-ranked compounds and reference standards against (a) VP7A (b) C and (c) D after 120 ns MD
simulation period.

Table 4
Fluctuations of active residues within the VP7A epitope during the simulation.
KBS residues of VP7A 2SG Apigenin-4′-glucoside Glabrescin Glabratephrin Gnetin L Sericetin Tizoxanide VP7A Total RMSF

Lys194 1.34 1.79 2.25 1.23 1.48 1.42 1.37 1.24 1.52
Val195 1.29 1.68 2.49 1.26 1.39 1.26 1.23 1.20 1.48
Cys196 1.34 1.85 2.86 1.34 1.48 1.38 1.39 1.22 1.61
Pro197 1.39 2.56 3.82 1.26 1.78 1.46 1.59 1.35 1.90
Leu198 1.46 2.77 4.09 1.44 1.96 1.62 1.69 1.50 1.18
Asn199 1.95 3.99 5.85 2.00 2.64 2.17 2.32 2.09 2.88
Thr200 2.56 4.67 6.68 2.62 3.34 2.65 2.79 3.02 3.54
Gln201 2.91 4.91 7.12 2.89 3.46 3.00 3.04 3.41 3.84
Thr202 2.25 3.57 5.25 2.09 2.59 2.14 2.23 2.63 2.84
Leu203 2.38 3.75 5.59 2.10 2.57 2.30 2.43 3.57 3.09
Gly204 2.02 3.15 4.88 1.66 2.18 1.91 2.08 4.07 2.74
Ile205 2.37 3.87 5.51 2.07 2.65 2.11 2.51 4.91 3.25
Gly206 2.99 3.45 5.87 2.74 3.01 2.42 3.21 6.53 3.78
Cys207 2.95 2.63 5.01 2.92 2.97 2.31 3.26 6.00 3.51
Total RMSF 2.09 3.19 4.81 1.97 2.39 2.01 2.22 3.05

Table 5
Fluctuations of active residues within the VP7C epitope during the simulation.
KBS residues of VP7C 2SG Apigenin-4′-glucoside Ellagic acid Gnetin L Tizoxanide VP7C Total RMSF

Ala273 8.41 2.37 3.64 10.12 3.61 3.31 5.24

Asp274 7.97 2.39 3.83 9.16 4.36 3.24 5.16
Pro275 8.21 2.34 3.30 8.86 4.33 3.01 5.03
Thr276 8.05 2.53 3.88 8.43 4.87 3.43 5.20
Thr277 5.55 2.31 3.62 5.72 4.31 2.97 4.08
Ala278 4.05 2.03 2.94 3.82 3.21 2.54 3.10
Pro279 3.17 2.20 2.78 3.19 2.95 2.68 2.83
Gln280 3.46 3.24 4.59 2.35 3.78 3.19 3.44
Thr281 1.61 1.47 2.70 1.47 1.88 1.92 1.84
Glu282 1.60 1.70 2.56 1.78 1.75 1.88 1.88
Arg283 1.31 1.26 1.53 1.10 0.98 1.33 1.25
Met284 1.22 1.13 1.23 1.53 1.62 1.63 1.39
Met285 1.01 1.14 1.41 1.46 1.40 1.55 1.33
Arg286 1.10 1.32 1.46 2.14 1.69 1.86 1.60
Total RMSF 4.05 1.96 2.82 4.37 2.91 2.47

VP7A, gnetin L-VP7A and sericetin-VP7A have mean ROG values higher than that of the apo-VP7A (Table 3); however, a minimal
overall difference of 0.74 Å was observed, which indicates that the ligand binding induced a minimal effect on the protein folding.
Therefore, it can be inferred that ligand binding sustained the thermodynamic entropy of the protein, and this is partly corroborated by
the RMSD values indicating their thermodynamic stability towards VP7A. The interaction between the highest-ranking compounds
and tizoxanide with VP7C had little impact on its folding process, as revealed by the limited fluctuations in compactness (Fig. 6(b))
over the 120 ns simulation period. Expectedly, the corresponding ROG values across all the complexes corroborated this finding with
an overall difference of 0.22 Å and the least value (20.71 Å) was exhibited by apigenin-4′-glucoside-VP7C (Table 3). Likewise, the
binding of the investigated compounds and tizoxanide with VP7D had a negligible impact on the protein’s folding process (Fig. 6(c))
throughout the simulation. This was evidenced in the overall difference of average ROG value of 0.32 Å across all the complexes, which

11
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Table 6
Fluctuations of active residues within the VP7D epitope during the simulation.
KBS residues of VP7D 2SG Apigenin-4′-glucoside Edulisin I Glabratephrin Gnetin L Sericetin Tizoxanide VP7D Total RMSF

Val300 1.31 1.15 1.19 1.27 1.47 1.55 1.53 1.25 1.34
Asp301 1.46 1.24 1.32 1.59 1.66 1.86 1.57 1.47 1.52
Tyr302 1.24 1.26 1.06 1.50 1.25 1.92 1.63 1.57 1.43
Val303 1.26 1.27 1.05 1.14 1.07 1.13 1.22 1.07 1.15
Asn304 1.60 1.48 1.32 1.47 1.41 1.60 1.66 1.48 1.50
Gln305 1.66 1.49 1.25 1.56 1.23 1.59 1.80 1.82 1.55
Ile306 1.04 1.17 0.92 0.94 0.92 1.07 1.03 1.10 1.02
Ile307 1.19 1.05 0.97 1.00 0.88 1.09 1.08 1.00 1.03
Gln308 1.98 1.64 1.66 1.69 1.34 1.55 1.67 1.62 1.64
Ala309 1.37 1.05 1.13 1.24 0.97 1.11 0.99 1.02 1.11
Met310 1.25 1.05 1.04 1.21 0.98 1.08 0.99 0.89 1.06
Ser311 2.30 1.54 1.63 1.66 1.36 1.41 1.32 1.32 1.57
Lys312 4.73 2.98 4.23 2.80 3.37 2.68 2.15 2.30 3.16
Total RMSF 1.72 1.41 1.44 1.47 1.38 1.51 1.43 1.38

Fig. 6. Comparative ROG plots of alpha-carbon, top-ranked compounds and reference standard against (a) VP7A (b) C and (c) D after 120 ns MD
simulation period.

is similar to the findings of Aribisala et al. [34] in which an overall variation of 0.3 Å was observed in the binding of SARS-CoV-2
wild-type spike protein with secondary metabolites among all the studied systems. Interestingly, the lowest mean ROG value was
observed in apigenin-4′-glucoside-VP7D, just as in VP7C (Table 4). Generally, the structural compatibility observed in this study could
be associated with various factors such as the topology, size, and length of the protein with the ligand’s molecular weight and its
contact with the target protein [56,58].
Another important thermodynamic stability index for evaluating protein folding and surface area changes throughout the simu­
lation is the SASA, with higher values indicative of an expansion in the volume of the protein [57]. The degree of change in the SASA
value is determined by the physicochemical characteristics of the amino acid residues that have been rearranged or altered [59] upon
exposure to solvent or water [60] and the observed variation in exposure levels might potentially be ascribed to changes in the tertiary
conformation of the protein [61]. During the simulation period, minor swaying was observed in the VP7A, C and D SASA plots (Fig. 7),
suggesting a minimal effect owing to the binding of the studied metabolites to the targets. Against VP7A, an overall minimal difference
of 442.56 Å was observed across all the complexes, with the lowest mean SASA value observed in tizoxanide-VP7A (12,535.76 Å)

Fig. 7. Comparative SASA plots of alpha-carbon, top-ranked compounds and reference standard against (a) VP7A (b) C and (d) D after 120 ns MD
simulation period.

12
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

followed by 2SG-VP7A (12,598.49 Å) (Table 3). The lowest SASA value observed in tizoxanide-VP7A complex suggests its potential to
reduce the surface exposure of the protein to surrounding solvent compared to other investigated compounds; however, tizoxanide,
2SG, apigenin-4′-glucoside, and glabratephrin have lower SASA values relative to the unbound VP7A (Table 3) indicating that their
binding to the protein reduced its volume though with a little effect. Furthermore, a minimum overall difference of 515.92 Å was
observed in VP7C across all the complexes. The lowest average SASA value was found in ellagic acid-VP7C complex (12,585.80 Å), and
this is consistent with the mean RMSD value, indicating that it could be a better inhibitor of VP7C compared to other compounds
(Table 3). In addition, only ellagic acid-VP7C has lower mean SASA value relative to the apo-VP7C, indicating its capacity to reduce the
target’s surface area and limit its interaction with surrounding solvent molecules consistent with the RMSD finding. Similarly, a little
overall difference of 383.17 Å was observed across all systems against VP7D with apigenin-4′-glucoside (12,622.74 Å) exhibiting the
lowest average SASA value consistent with the RMSF and ROG findings. Also, only the apigenin-4′-glucoside-VP7D complex mean
SASA value is lesser than that of VP7D unbound protein, suggesting its potential to decrease the protein’s target surface area and limit
exposure to solvent molecules. Generally, the minimal influence of ligand binding on the mean SASA values of the resulting complexes
exhibits a partly similar trend to the findings of the average ROG values observed in this study, as the ligand binding did not
significantly affect the folding of the protein, suggesting that the binding of the compounds and tizoxanide on VP7A, C and D does not
disturb its thermodynamic entropy. This agrees with the findings of Moussaoui et al. [62], in which the mean SASA values for each
complex did not vary significantly from one another.
Hydrogen bonds (H-bonds) indicators are essential for a protein’s structural stability and compactness [63]. Consequently, H-bonds
studies may be utilized to evaluate protein and protein-ligand complex stability during simulation [64]. Throughout the simulation, a
steady variation in the form of the number of H-bonds formed in each system was observed against VP7A, C, and D (Fig. 8(a)–(c)). The
observed stable fluctuation in the number of H-bonds formed upon ligand binding suggests that the thermodynamic entropy of the
respective systems was unaffected, thereby retaining the protein geometry. Specifically, for glabrescin-VP7A and glabratephrin-VP7A
complexes, ligand binding resulted in reduced intramolecular hydrogen bonds formation against VP7A, and a similar trend was
observed in 2SG, edulisin I, gnetin L and sericetin with reduction of hydrogen bonds formation against VP7D relative to the apo-protein
(Table 3). The observed reduction in the mean number of hydrogen bonds formed relative to the apo-protein could indicate breakage in
some intramolecular hydrogen bonds formed owing to the ligand binding. This observation is similar to that of S’thebe et al. [65], in
which the binding of the top three compounds and reference standards with the investigated target led to a reduction in the average
number of intramolecular H-bonds. However, this observation contrasted the number of hydrogen bonds formed upon the binding of
other metabolites to VP7A and VP7D. Interestingly, in VP7C, ligand binding resulted in the increased formation of intramolecular
hydrogen bonds across all systems. This aligns with the findings of Lanrewaju et al. [28], where ligand binding to SARS-CoV-2 main
protease resulted in the rise of hydrogen bonds compared to the unbound main protease. The observed increase in intramolecular
H-bonds suggests that the ligands have occupied the intramolecular space of the proteins, as the rise in the number of H-bonds may
imply the occurrence of H-bond interactions facilitated by the intermolecular bindings of ligands to the protein.
Evaluations of the intramolecular H-bond distance in VP7A, C and D during the simulation revealed a continuous reduction across
all systems (Fig. 9(a)–(c)). Interestingly, an average H-bond distance of 2.85 Å was recorded in all the systems across VP7A, C and D
(Table 3). This finding is consistent with our previous study [28], in which a consistent reduction in the distance of hydrogen bonds
with an average hydrogen bond distance of 2.85 Å across all systems was observed. In addition, previous studies have documented an
acceptable intramolecular H-bonds distance of 3.5 Å, with a lower value indicating a stronger hydrogen bonding interaction [54].
Thus, in the present investigation, a hydrogen bond distance below 3.5 Å indicates a favourable hydrogen bonding state throughout the
simulation. Exhibiting identical average H-bonds distance across the apo-proteins and bonding systems in the investigated epitopes
implies a state of thermodynamic orderliness upon ligand binding, and this provides further support to other thermodynamic metrics
evaluated in this study.
Furthermore, as the simulation continued, this study showed a consistent variation in the H-bonds angle in VP7A, C and D (Fig. 10
(a)–(c)). At the outset of the simulation, a significant variation ranging from 138◦ to 166◦ was observed in the plots. Nevertheless, as
the simulation proceeded, the variation significantly decreased to 180◦ in the systems across the targets (Fig. 10(a)–(c)). The

Fig. 8. Time progression of the number of H-bonds in (a) VP7A (b) C and (c) D upon binding with the top-ranked compounds and reference
standards during the 120 ns MD simulation period.

13
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Fig. 9. Time progression of the H-bonds distance in (a) VP7A (b) C and (b) D upon binding of the top-ranked compounds and reference standards
during the 120 ns MD simulation period.

diminishing degree of variation towards 180◦ which has been reported to be the ideal hydrogen bond angle [66], during the simulation
signifies the enhanced strength of the bonding, and this aligns with the results obtained from the analysis of the H-bonds distance in
this study. Generally, the observed average bond angles across the three targets are closely related, ranging from 151.79◦ –152.27◦ ,
151.77◦ –152.30◦ and 151.76◦ –152.24◦ in VP7A, C and D respectively (Table 3). This observation suggests that the target protein VP7
is not in a disordered state upon ligand binding, and this is relatively consistent with other thermodynamic metrics examined in this
study for stability, compactness, and flexibility. The fact that the H-bonds distance and angle values are similar across board for both
targets suggests their negligible effect on the binding free energy reported in this study.

Bond analysis of interaction plots of the top-ranked compounds after 120 ns MD simulation

The binding affinity between a ligand and a receptor and the stability of the resulting complex are intricately linked to the kind of
interactions, bond length, and the number of interactions occurring within the complex [57]. Therefore, this study evaluated the
features in the binding of the highest-ranked compounds with VP7A, C, and D after a 120 ns simulation period. Figs. 11–16 illustrate
the plot of interactions between tizoxanide and the top-ranked compounds which exhibited the highest negative binding free energy
with VP7A, C and D. On the other hand, Supplementary Table S3 displays the plot of interactions for the remaining top-ranked
compounds. Interactions observed in this study between the VP7 epitopes, and the top-ranked ligands include hydrogen bonds
(conventional, π-donor, and carbon), van der Waals, amide π-stacked, π-anion, π-sigma, π-cation, π-π t-cation, π-π t-shaped, π-π
t-stacked, π-alkyl, alkyl, π-Sulphur, unfavourable donor-donor interactions among others. Apigenin-4′-glucoside-VP7A complex with
the highest negative binding free energy against VP7A had the highest number of interactions (17) with hydrogen bonds contacts with
Ile149 and Asp151 (2) (Table 7 and Fig. 11). These observations established a correlation between the higher binding free energy in the
apigenin-4′-glucoside-VP7A complex to the number, nature, and distance of interactions within the complex. The next highest negative
binding free energy against VP7A was observed in gnetin L-VP7A complex but did not correlate with the observed sum of interactions
(15). Analysis of the bond interactions revealed that gnetin L-VP7A made three hydrogen bond contacts with Pro120, Cys130 and
Val118 and with shorter bond length compared to apigenin-4′-glucoside-VP7A (Table 7), and could be connected with the observed
high binding free energy. Hydrogen bonds have been recognized as a significant class of non-covalent bonds in drug discovery due to
their exceptional intermolecular interactions [67]. Glabrescin-VP7A complex has the highest number of interactions (16) after api­
genin-4′-glucoside-VP7A but with no correlation with its binding free energy. Also, just one hydrogen bond contact with Leu121 was
made and could be liable for the observed lesser binding free energy (Tables 7 and S3). A similar binding free energy was observed in
sericetin-VP7A and 2SG-VP7A; unfortunately, a lesser number of interactions was observed in 2SG relative to sericetin (Tables 2 and

Fig. 10. Time progression of the H-bonds angle in (a) VP7A (b) C and (c) D upon binding of the top-ranked compounds and reference standards
during the 120 ns MD simulation period.

14
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Fig. 11. Interaction plots of apigenin-4′-glucoside + VP7A at 0, 40, 80 and 120 ns.

Fig. 12. Interaction plots of tizoxanide + VP7A at 0, 40, 80 and 120 ns.

7). A closer look at the interaction plot of 2SG-VP7A at 0, 40, and 80 ns (Supplementary Table S3) revealed a larger number of in­
teractions compared to 120 ns, which could be associated with the free binding energy that was detected. In the same vein, the
observed interactions in tizoxanide-VP7A relative to glabratephrin-VP7A does not justify its binding free energy, but a closer exam­
ination of tizoxanide-VP7A at 40 and 80 ns revealed lesser and no interactions respectively and could be liable for its binding free
energy (Fig. 12).
Remarkably, all investigated top-ranked compounds against VP7C exhibited a higher number of interactions and hydrogen bonds
contacts than tizoxanide which is consistent with the binding free energy results reported in this study. The higher number of

15
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Fig. 13. Interaction plots of apigenin-4′-glucoside + VP7C at 0, 40, 80 and 120 ns.

Fig. 14. Interaction plots of tizoxanide + VP7C at 0, 40, 80 and 120 ns.

interactions against VP7C observed in gnetin L-VP7C (20) and 2SG-VP7C (19) is not in sync with its binding free energy (Table 7)
relative to apigenin-4′-glucoside-VP7C (18) with the highest negative binding free energy. The observed higher binding free energy in
apigenin-4′-glucoside relative to other top-ranked compounds and the reference standard could be due to its higher number of
hydrogen bonds (9) with Gln185, Leu192, Val191, Asp190 (2), Arg209 (3) and Glu205 (Fig. 13). As highlighted earlier, hydrogen
bonds are one of the most vital non-covalent bonds in drug discovery and, in addition, are important in facilitating molecular
recognition, enhancing the stability of protein-ligand interactions, and promoting the overall stability of drug-target complexes [67].
Therefore, the lesser number of hydrogen bonds observed in gnetin L (4) and 2SG (2), despite their higher number of interactions

16
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Fig. 15. Interaction plots of 2SG + VP7D at 0, 40, 80 and 120 ns.

Fig. 16. Interaction plots of tizoxanide + VP7D at 0, 40, 80 and 120 ns.

relative to apigenin-4′-glucoside (9) (Table 7) could be responsible for their reduced binding free energy. Interestingly, the number of
interactions and hydrogen bond contacts observed in ellagic acid-VP7C and tizoxanide+-VP7C (Fig. 14) is consistent with the binding
free energy (Tables 3 and 7). Therefore, these observations indicate the significance of the number, length and nature of contacts in the
calculation of binding free energy, as previously shown in the interactions with VP7A in this investigation.
The lowest and highest number of interactions against VP7D was exhibited by 2SG-VP7D (22) and tizoxanide-VP7D (3), and is
consistent with their binding free energy (Figs. 15 and 16). Furthermore, gnetin L-VP7D (14) and sericetin-VP7D (14) shared equal
numbers of interactions; however, they are different in terms of their binding free energy (Tables 3 and 7). The hydrogen bonds

17
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

Table 7
Bond interactions between the top-ranked compounds against VP7A, C and D after 120 ns simulation.
Top-ranked Total number of interactions and Number of hydrogen bonds [mean distance Other vital interactions and residues
compounds mean bond length (Å) (Å)] and interaction residues

VP7A
2SG 4 (− ) None None
Apigenin-4′- 17 (5.15) 3 (4.70) [Ile149, Asp151 (2)] 5 (5.43) [Val152, Val153, Pro120 (2), Ile128]
glucoside
Glabrescin 16 (5.49) 1 (5.77) [Leu121] 9 (5.46) [Val153 (3), Val152, Pro120, Val156 (2),
Ile128 (2)]
Glabratephrin 9 (4.69) 1 (4.67) [Asn122] 1 (4.71) [Val156]
Gnetin L 15 (4.75) 3 (4.34) [Pro120, Cys130, Val118] 3 (5.15) [Ile128, Val141 (2)]
Sericetin 10 (4.85) 1 (3.45) [Trp34] 4 (5.20) [Pro35, Trp34 (2), Gly33]
Tizoxanide 8 (4.19) 3 (3.61) [Pro90, Arg170, Asp53] 1 (5.95) [Arg170]

VP7C

2SG 19 (4.85) 2 (4.98) [Gln185, Met208] 11 (4.83) [Arg209 (2), Val191, Leu192, Met207 (2),
Met208, Tyr225 (2), Ile229, Ala232]
Apigenin-4′- 18 (4.74) 9 (4.30) [Gln185, Leu192, Val191, Asp190 4 (5.74) [Met208 (2), Ile229, Ala232]
glucoside (2), Arg209 (3), Glu205]
Ellagic acid 14 (4.33) 5 (4.06) [Thr195, Thr204, Asp80 (2), 4 (4.67) [Pro202, Arg206, Ile194 (2)]
Lys174]
Gnetin L 20 (4.70) 4 (4.61) [Ile210, Ala196, Tyr225, Glu205] 7 (4.76) [Ile210, Arg209, Ala196 (2), Met208,
Pro198, Tyr225]
Tizoxanide 7 (5.62) 1 (5.70) [Thr195] 2 (5.58) [Val191, Pro198]

VP7D

2SG 22 (5.00) 2 (3.88) [Asn211 (2)] 11 (5.20) [Met208 (2), Gln228 (2), Tyr225 (1),
Ala232, Ile229, Ala182 (2), Arg206, Val62]
Apigenin-4′- 13 (4.88) 6 (4.88) [Met207 (2), Met208, Arg209, 2 (4.88) [Met208, Tyr225]
glucoside Ile210, Asn211]
Edulisin I 15 (5.24) 2 (3.87) [Val222, Thr221] 6 (5.70) [Met208, Ile210 (2) Tyr225 (3)]
Glabratephrin 12 (4.99) 4 (4.87) [Asn211, Arg209 (2), Met208] 4 (4.83) [Tyr225, Met208, Arg209, Ile210]
Gnetin L 14 (5.48) 2 (4.57) [Lys235, Glu205] 4 (5.95) [Ile184, Ile210, Tyr225, Met208]
Sericetin 14 (5.67) None 7 (5.67) [Met208 (2), Tyr225 (2), Ile184, Ile210,
Ala232]
Tizoxanide 3 (4.7) None 3 (4.7) [Tyr57, Gln55]

analysis of the two complexes revealed that sericetin does not make hydrogen bond contact with any of the residues, while gnetin L was
observed to make hydrogen bond contact with Lys235 and Glu205. The observed hydrogen bond contact, as highlighted earlier, could
be associated with the higher binding free energy exhibited by gnetin L. In addition, a shorter bond length in gnetin L-VP7D compared
to sericetin-VP7D (Table 7) could have resulted in its higher binding free energy against VP7D relative to sericetin. Moreover, the sum
of interactions in edulisin I-VP7D (15) is insufficient to account for its binding free energy relative to apigenin-4′-glucoside-VP7D (13);
however, a shorter bond length was observed in apigenin-4′-glucoside compared to edulisin I-VP7D. Shorter bond lengths have been
reported to result in a stronger force between intermolecular or intramolecular entities owing to the proximity of the molecules, hence
a higher affinity for the receptor [64]. This observation could be connected to the higher negative binding free energy in apige­
nin-4′-glucoside-VP7D coupled with a higher number of hydrogen bond contacts relative to edulisin I-VP7D. Conversely, the number of
interactions in glabratephrin-VP7D (12) aligns with its binding free energy (Tables 3 and 7).
Overall, a conspicuous periodic shift was observed during the simulation period due to the inherent flexibility of the protein and
ligand in the MD simulation, which is known for its dynamic nature [58,68]. Consequently, the amino acid residues in the interaction
plots for the complexes in this study exhibit variations at regular intervals during the 120 ns simulation period (Figs. 11–16). Inter­
estingly, no adverse donor-donor interactions, which might potentially result in a repulsive force [69], were detected in the molecular
interactions between the top-ranked compounds against VP7A, C and D at 120 ns of simulation (Table 7) and such interactions were
rarely identified over the whole simulation period (Figs. 11–16). Summarily, the enhanced broad-spectrum activity demonstrated by
2SG, apigenin-4′-glucoside, and gnetin L relative to tizoxanide against the VP7 epitopes can be ascribed to the number, type and bond
lengths of contacts with the receptor.

Conclusion

This study has demonstrated the abilities of the lead compounds as potential inhibitors of the investigated VP7 epitopes. The lead
compounds cut across all the three investigated medicinal plants: 2SG (M. barteri), apigenin-4′-glucoside and ellagic acid (M. barteri
and D. eriocarpum), glabrescin, sericetin and glabratephrin (D. eriocarpum), gnetin L (S. mombin), and edulisin I (M. barteri) which
further validates the antiviral abilities of these plants. The top-ranked compounds exhibited satisfactory pharmacokinetics and toxicity
properties utilized in this investigation for screening, enhancing their potential to be developed as drug candidates with further im­
provements and structural alterations. Remarkably, all the lead compounds had higher negative binding free energy than tizoxanide

18
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

across the three epitopes of VP7 ascribed to the number, type, and bond length formed with the targets as apigenin-4′-glucoside had the
highest affinity for VP7A and C while the highest affinity for VP7D was observed in 2SG indicating their better potential at inhibiting
the protein relative to other compounds. Interestingly, 2SG, apigenin-4′-glucoside and gnetin L had better broad-spectrum affinities for
VP7A, C and D relative to tizoxanide (reference standard) respectively. This observation denotes that 2SG, apigenin-4′-glucoside, and
gnetin L exhibit superior advantages compared to other lead compounds in the management of RVA-induced diarrhoea, which any of
the VP7 epitopes can cause. Furthermore, the thermodynamic information obtained from a 120 ns simulation confirmed the stability,
flexibility, and compactness of the hit compounds with 2SG, ellagic acid and sericetin being the most stable against VP7A, C and D
respectively. Although this study’s findings unearth the lead compounds’ promising ability to modulate the investigated VP7 epitopes,
further confirmatory in vitro and in vivo studies are recommended to validate the findings and efforts are being made to pursue this
approach.

Funding

The study is based on the financial assistance received from the Directorate of Research and Postgraduate Support, DUT, and the
South African Medical Research Council (SAMRC) under a Self-Initiated Research Grant as well as the National Research Foundation
(NRF) Research Development Grant for Rated Researchers (Grant number 120433) and the Competitive Programme for Rated Re­
searchers Support (SRUG2204193723) to S. Sabiu. We also acknowledge Water Research Commission (WRC) of South Africa for the
grant (K5/C2020-2021-00181) awarded to A.M. Enitan-Folami.

CRediT authorship contribution statement

Adedayo Ayodeji Lanrewaju: Methodology, Validation, Formal analysis, Investigation, Data curation, Writing – original draft.
Abimbola Motunrayo Enitan-Folami: Conceptualization, Writing – review & editing, Supervision, Funding acquisition. Saheed
Sabiu: Conceptualization, Software, Validation, Resources, Data curation, Writing – review & editing, Supervision, Funding acqui­
sition. Feroz Mahomed Swalaha: Writing – review & editing, Supervision, Funding acquisition.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Data availability

The data is contained within the article or supplementary material.

Acknowledgments

The Durban University of Technology (DUT) Doctoral Scholarship Scheme administered by the Directorate of Research and
Postgraduate Support to A. A. Lanrewaju is duly acknowledged. The Centre for High-Performance Computing (CHPC), South Africa, is
equally acknowledged for granting access to the computing systems used in this study.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.sciaf.2024.e02109.

References

[1] C. Troeger, I.A. Khalil, P.C. Rao, S. Cao, B.F. Blacker, T. Ahmed, G. Armah, J.E. Bines, T.G. Brewer, D.V. Colombara, Rotavirus vaccination and the global burden
of rotavirus diarrhea among children younger than 5 years, JAMA Pediatr. 172 (10) (2018) 958–965, https://doi.org/10.1001/jamapediatrics.2018.1960.
[2] C.A. Omatola, A.O. Olaniran, Rotaviruses: from pathogenesis to disease control—A critical review, Viruses 14 (5) (2022) 875, https://doi.org/10.3390/
v14050875.
[3] H.U. Ugboko, O.C. Nwinyi, S.U. Oranusi, J.O. Oyewale, Childhood diarrhoeal diseases in developing countries, Heliyon 6 (4) (2020) e03690, https://doi.org/
10.1016/j.heliyon.2020.e03690.
[4] A.A. Lanrewaju, A.M. Enitan-Folami, S. Sabiu, J.N. Edokpayi, F.M. Swalaha, Global public health implications of human exposure to viral contaminated water,
Front. Microbiol. 13 (2022) 981896, https://doi.org/10.3389/fmicb.2022.981896.
[5] U.D. Parashar, H. Johnson, A.D. Steele, J.E. Tate, Health impact of rotavirus vaccination in developing countries: progress and way forward, Clin. Infect. Dis. 62
(suppl_2) (2016) S91–S95, https://doi.org/10.1093/cid/civ1015.
[6] L. Jiang, A. Tang, L. Song, Y. Tong, H. Fan, Advances in the development of antivirals for rotavirus infection, Front. Immunol. 14 (2023), https://doi.org/
10.3389/fimmu.2023.1041149.
[7] M.M. Nyaga, S. Sabiu, V.N. Ndze, F.E. Dennis, K.C. Jere, Report of the 1st African enteric viruses genome initiative (AEVGI) data and bioinformatics workshop
on whole-genome analysis of some African rotavirus strains held in bloemfontein, South Africa, Vaccine 38 (34) (2020) 5402–5407, https://doi.org/10.1016/j.
vaccine.2020.06.010.

19
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

[8] T. Garuba, R. Govender, H.A. Isah, S. Sabiu, Metabolites profiling and molecular docking identification of putative leads from endophytic Phyllosticta capitalensis
as modulators of key druggable structural targets of rotavirus A, Trans. R. Soc. South Africa 77 (3) (2022) 207–217, https://doi.org/10.1080/
0035919X.2022.2158387.
[9] D. Asensio-Cob, J.M. Rodríguez, D. Luque, Rotavirus particle disassembly and assembly in vivo and in vitro, Viruses 15 (8) (2023) 1750, https://doi.org/
10.3390/v15081750.
[10] A.E. Fadiji, I. Omomowo, O. Omomowo, A. Lanrewaju, D. Akande, Antimicrobial efficacy and phytochemical screening of aqueous and ethanolic extracts of
Ocimum gratissimum (scent leaf) leaf against some clinical isolates, J. Basic Pharmacol. Toxicol. 2 (2) (2018) 19–24.
[11] R. Sharma, M. Bhattu, A. Tripathi, M. Verma, R. Acevedo, P. Kumar, V.D. Rajput, J. Singh, Potential medicinal plants to combat viral infections: a way forward
to environmental biotechnology, Environ. Res. (2023) 115725, https://doi.org/10.1016/j.envres.2023.115725.
[12] Y. Sun, X. Gong, J.Y. Tan, L. Kang, D. Li, Vikash, J. Yang, G Du, In vitro antiviral activity of Rubia cordifolia aerial part extract against rotavirus, Front.
Pharmacol. 7 (2016) 308, https://doi.org/10.3389/fphar.2016.00308.
[13] A. Civra, R. Francese, D. Sinato, M. Donalisio, V. Cagno, P. Rubiolo, R. Ceylan, A. Uysal, G. Zengin, D. Lembo, In vitro screening for antiviral activity of Turkish
plants revealing methanolic extract of Rindera lanata var. lanata active against human rotavirus, BMC Complement. Altern. Med. 17 (2017) 1–8, https://doi.org/
10.1186/s12906-017-1560-3.
[14] J. Corthout, L. Pieters, M. Claeys, D.V. Berghe, A. Vlietinck, Antiviral ellagitannins from Spondias mombin, Phytochemistry 30 (4) (1991) 1129–1130, https://
doi.org/10.1016/S0031-9422(00)95187-2.
[15] E.M.d.S. Siqueira, T.L. Lima, L. Boff, S.G. Lima, E.M. Lourenço, É.G. Ferreira, E.G. Barbosa, P.R. Machado, K.J. Farias, L.d.S. Ferreira, Antiviral potential of
Spondias mombin L. leaves extract against herpes simplex virus type-1 replication using in vitro and in silico approaches, Planta Med. 86 (07) (2020) 505–515,
https://doi.org/10.1055/a-1135-9066.
[16] O.O. Ogbole, T.E. Akinleye, P.A. Segun, T.C. Faleye, A.J. Adeniji, In vitro antiviral activity of twenty-seven medicinal plant extracts from Southwest Nigeria
against three serotypes of echoviruses, Virol. J. 15 (1) (2018) 1–8, https://doi.org/10.1186/s12985-018-1022-7.
[17] P.A. Segun, O.O. Ogbole, T.E. Akinleye, T.O. Faleye, A.J. Adeniji, In vitro anti-enteroviral activity of stilbenoids isolated from the leaves of Macaranga barteri,
Nat. Prod. Res. (2019) 1–5, https://doi.org/10.1080/14786419.2019.1644505.
[18] D. Luseba, E. Elgorashi, D. Ntloedibe, J. Van Staden, Antibacterial, anti-inflammatory and mutagenic effects of some medicinal plants used in South Africa for
the treatment of wounds and retained placenta in livestock, S. Afr. J. Bot. 73 (3) (2007) 378–383, https://doi.org/10.1016/j.sajb.2007.03.003.
[19] D. Van der Merwe, G. Swan, C. Botha, Use of ethnoveterinary medicinal plants in cattle by Setswana-speaking people in the Madikwe area of the North West
Province of South Africa, J. S. Afr. Vet. Assoc. 72 (4) (2001) 189–196.
[20] A.A. Lanrewaju, A.M. Enitan-Folami, M.M. Nyaga, S. Sabiu, F.M. Swalaha, Cheminformatics bioprospection of selected medicinal plants metabolites against
trypsin cleaved VP4 (spike protein) of rotavirus A, J. Biomol. Struct. Dyn. (2023) 1–20, https://doi.org/10.1080/07391102.2023.2258405.
[21] J.O. Aribisala, R.A. Abdulsalam, Y. Dweba, K. Madonsela, S. Sabiu, Identification of secondary metabolites from Crescentia cujete as promising antibacterial
therapeutics targeting type 2A topoisomerases through molecular dynamics simulation, Comput. Biol. Med. 145 (2022) 105432, https://doi.org/10.1016/j.
compbiomed.2022.105432.
[22] F.O. Shode, J.O.-o. Uhomoibhi, K.A. Idowu, S. Sabiu, K.K. Govender, Molecular dynamics study on selected bioactive phytochemicals as potential inhibitors of
HIV-1 subtype C protease, Metabolites 12 (11) (2022) 1155, https://doi.org/10.3390/metabo12111155.
[23] J.O.-O. Uhomoibhi, F.O. Shode, K.A. Idowu, S. Sabiu, Molecular modelling identification of phytocompounds from selected African botanicals as promising
therapeutics against druggable human host cell targets of SARS-CoV-2, J. Mol. Graph. Model. 114 (2022) 108185, https://doi.org/10.1016/j.
jmgm.2022.108185.
[24] J.J. Oloche, B.B. Oluremi, C.E. Aruwa, S. Sabiu, Molecular modeling identification of key secondary metabolites from Xylopia aethiopica as promising
therapeutics targeting essential measles viral proteins, Evidence-Based Complement. Altern. Med. (2023) 1575358, https://doi.org/10.1155/2023/1575358.
[25] A. Ghosh, S. Chattopadhyay, M. Chawla-Sarkar, P. Nandy, A. Nandy, In silico study of rotavirus VP7 surface accessible conserved regions for antiviral drug/
vaccine design, PLoS One 7 (7) (2012) e40749, https://doi.org/10.1371/journal.pone.0040749.
[26] S. Sabiu, F.O. Balogun, S.O. Amoo, Phenolics profiling of Carpobrotus edulis (L.) NE Br. and insights into molecular dynamics of their significance in type 2
diabetes therapy and its retinopathy complication, Molecules 26 (16) (2021) 4867, https://doi.org/10.3390/molecules26164867.
[27] E.F. Pettersen, T.D. Goddard, C.C. Huang, G.S. Couch, D.M. Greenblatt, E.C. Meng, T.E. Ferrin, UCSF Chimera—A visualization system for exploratory research
and analysis, J. Comput. Chem. 25 (13) (2004) 1605–1612, https://doi.org/10.1002/jcc.20084.
[28] A.A. Lanrewaju, A.M. Enitan-Folami, M.M. Nyaga, S. Sabiu, F.M. Swalaha, Metabolites profiling and cheminformatics bioprospection of selected medicinal
plants against the main protease and RNA-dependent RNA polymerase of SARS-CoV-2, J. Biomol. Struct. Dyn. (2023) 1–21, https://doi.org/10.1080/
07391102.2023.2236718.
[29] J.-F. Rossignol, M. Abu-Zekry, A. Hussein, M.G. Santoro, Effect of nitazoxanide for treatment of severe rotavirus diarrhoea: randomised double-blind placebo-
controlled trial, Lancet 368 (9530) (2006) 124–129, https://doi.org/10.1016/S0140-6736(06)68852-1.
[30] S. Dallakyan, A.J. Olson, Small-molecule library screening by docking with PyRx, Chem. Biol. (2015) 243–250, https://doi.org/10.1007/978-1-4939-2269-7_
19.
[31] H. Kruse, L. Goerigk, S. Grimme, Why the standard B3LYP/6–31G* model chemistry should not be used in DFT calculations of molecular thermochemistry:
understanding and correcting the problem, J. Org. Chem. 77 (23) (2012) 10824–10834, https://doi.org/10.1021/jo302156p.
[32] Calais, J.-L. (1993). Density-Functional Theory of Atoms and Molecules. R.G. Parr and W. Yang, Oxford University Press, New York, Oxford, 1989. IX+ 333 pp.
Price£ 45.00. International Journal of Quantum Chemistry, 47(1), 101-101. 10.1002/qua.560470107.
[33] C.U. Ibeji, D.C. Akintayo, H.O. Oluwasola, E.O. Akintemi, O.G. Onwukwe, O.M. Eziomume, Synthesis, experimental and computational studies on the anti-
corrosion performance of substituted Schiff bases of 2-methoxybenzaldehyde for mild steel in HCl medium, Sci. Rep. 13 (1) (2023) 3265, https://doi.org/
10.1038/s41598-023-30396-35.
[34] J.O. Aribisala, C.E. Aruwa, T.O. Uthman, I.O. Nurain, K. Idowu, S. Sabiu, Cheminformatics bioprospection of broad spectrum plant secondary metabolites
targeting the spike proteins of omicron variant and wild-type SARS-CoV-2, Metabolites, 12 (10) (2022) 982, https://doi.org/10.3390/metabo12100982.
[35] L. Deschenes, D. Bout, Origin 6.0: scientific data analysis and graphing software origin lab corporation. University of Texas, Austin, J. Am. Chem. Soc. 122 (39)
(2000) 9567–9568, https://doi.org/10.1021/ja004761d.
[36] N.S. Pagadala, K. Syed, J. Tuszynski, Software for molecular docking: a review, Biophys. Rev. 9 (2017) 91–102, https://doi.org/10.1007/s12551-016-0247-1.
[37] T. Jian-Bo, X. Zhang, L. Ding, B. Shuai, Molecular design, molecular docking and ADMET study of cyclic sulfonamide derivatives as SARS-CoV-2 inhibitors,
Chin. J. Anal. Chem. 49 (12) (2021) 63–73, https://doi.org/10.1016/j.cjac.2021.09.006.
[38] C. Prashantha, K. Gouthami, L. Lavanya, S. Bhavanam, A. Jakhar, R. Shakthiraju, V. Suraj, K. Sahana, H. Sujana, N. Guruprasad, Molecular screening of
antimalarial, antiviral, anti-inflammatory and HIV protease inhibitors against spike glycoprotein of coronavirus, J. Mol. Graph. Model. 102 (2021) 107769,
https://doi.org/10.1016/j.jmgm.2020.107769.
[39] Y. Wang, J. Xing, Y. Xu, N. Zhou, J. Peng, Z. Xiong, X. Liu, X. Luo, C. Luo, K. Chen, In silico ADME/T modelling for rational drug design, Q. Rev. Biophys. 48 (4)
(2015) 488–515, https://doi.org/10.1017/S0033583515000190.
[40] C.A. Lipinski, F. Lombardo, B.W. Dominy, P.J. Feeney, Experimental and computational approaches to estimate solubility and permeability in drug discovery
and development settings, Adv. Drug Deliv. Rev. 64 (2001) 4–17, https://doi.org/10.1016/s0169-409x(00)00129-0.
[41] R.P. Heaney, Factors influencing the measurement of bioavailability, taking calcium as a model, J. Nutr. 131 (4) (2001) 1344S–1348S, https://doi.org/
10.1093/jn/131.4.1344S.
[42] O.A. Ojo, A.E. Adegboyega, G.I. Johnson, N.L. Umedum, K. Onuh, M.N. Adeduro, V.O. Nwobodo, A.O. Elekan, T.E. Alemika, T.O. Johnson, Deciphering the
interactions of compounds from Allium sativum targeted towards identification of novel PTP 1B inhibitors in diabetes treatment: a computational approach,
Inform. Med. Unlocked 26 (2021) 100719, https://doi.org/10.1016/j.imu.2021.100719.
[43] C.C. Ogu, J.L. Maxa, Drug interactions due to cytochrome P450, Baylor Univ. Med. Center Proc. (2000), https://doi.org/10.1080/08998280.2000.11927719.

20
A.A. Lanrewaju et al. Scientific African 23 (2024) e02109

[44] A.K. Verma, S.F. Ahmed, M.S. Hossain, A.A. Bhojiya, A. Mathur, S.K. Upadhyay, A.K. Srivastava, N.K. Vishvakarma, M. Barik, M.M. Rahaman, Molecular
docking and simulation studies of flavonoid compounds against PBP-2a of methicillin-resistant Staphylococcus aureus, J. Biomol. Struct. Dyn. 40 (21) (2022)
10561–10577, https://doi.org/10.1080/07391102.2021.1944911.
[45] F.M. Ahamed, S. Chinnam, M. Challa, G. Kariyanna, A. Kumer, S. Jadoun, A. Salawi, A. G Al-Sehemi, U. Chakma, M.A.A Mashud, Molecular dynamics
simulation, QSAR, DFT, molecular docking, ADMET, and synthesis of ethyl 3-((5-bromopyridin-2-yl) imino) butanoate analogues as potential inhibitors of
SARS-CoV-2, Polycycl. Aromat. Compd. (2023) 1–19, https://doi.org/10.1080/10406638.2023.2173618.
[46] E.O. Akintemi, K.K. Govender, T. Singh, A DFT study of the chemical reactivity properties, spectroscopy and bioactivity scores of bioactive flavonols, Comput.
Theor. Chem. 1210 (2022) 113658, https://doi.org/10.1016/j.comptc.2022.113658.
[47] B.Y.G. Mountessou, A.W. Ngouonpe, A.S.W. Mbobda, E.O. Akintemi, H.-G. Stammler, S.F. Kouam, J.C. Tchouankeu, B.N. Lenta, N. Sewald, T. Singh, Structural
analysis and molecular docking study of pachypodostyflavone: a potent anti-onchocerca, J. Mol. Struct. (2023) 136003, https://doi.org/10.1016/j.
molstruc.2023.136003.
[48] A. Rampadarath, J.O. Aribisala, N.P. Makunga, S. Mazibuko-Mbeje, S. Sabiu, Molecular bioprospection of Helianthus annuus L.(sunflower) cypsela for
antidiabetic therapeutics through network pharmacology, density functional theory and molecular dynamics simulation, S. Afr. J. Bot. 162 (2023) 72–95,
https://doi.org/10.1016/j.sajb.2023.08.045.
[49] S. Lakhera, K. Devlal, A. Ghosh, P. Chowdhury, M. Rana, Modelling the DFT structural and reactivity study of feverfew and evaluation of its potential antiviral
activity against COVID-19 using molecular docking and MD simulations, Chem. Pap. 76 (5) (2022) 2759–2776, https://doi.org/10.1007/s11696-022-02067-6.
[50] N. Forouzesh, N. Mishra, An effective MM/GBSA protocol for absolute binding free energy calculations: a case study on SARS-CoV-2 spike protein and the
human ACE2 receptor, Molecules 26 (8) (2021) 2383, https://doi.org/10.3390/molecules26082383.
[51] U. Dasmahapatra, C.K. Kumar, S. Das, P.T. Subramanian, P. Murali, A.E. Isaac, K. Ramanathan, B. Mm, K. Chanda, In-silico molecular modelling, MM/GBSA
binding free energy and molecular dynamics simulation study of novel pyrido fused imidazo [4,5-c] quinolines as potential anti-tumor agents, Front. Chem. 10
(2022) 991369, https://doi.org/10.3389/fchem.2022.991369.
[52] A. Acharya, R. Agarwal, M.B. Baker, J. Baudry, D. Bhowmik, S. Boehm, K.G. Byler, S. Chen, L. Coates, C.J. Cooper, Supercomputer-based ensemble docking drug
discovery pipeline with application to COVID-19, J. Chem. Inf. Model. 60 (12) (2020) 5832–5852, https://doi.org/10.1021/acs.jcim.0c01010.
[53] E. Singh, R.K. Jha, R.J. Khan, A. Kumar, M. Jain, J. Muthukumaran, A.K. Singh, A computational essential dynamics approach to investigate structural
influences of ligand binding on Papain like protease from SARS-CoV-2, Comput. Biol. Chem. 99 (2022) 107721, https://doi.org/10.1016/j.
compbiolchem.2022.107721.
[54] J.O. Aribisala, K. Idowu, T.R. Makhanya, S. Sabiu, Cheminformatics identification of phenolics as modulators of key penicillin—binding proteins of Escherichia
coli towards interventive antibacterial therapy, Mol. Simul. (2023) 1–17, https://doi.org/10.1080/08927022.2023.2228423.
[55] D. Ramírez, J. Caballero, Is it reliable to use common molecular docking methods for comparing the binding affinities of enantiomer pairs for their protein
target? Int. J. Mol. Sci. 17 (4) (2016) 525, https://doi.org/10.3390/ijms17040525.
[56] M.H. Baig, D.R. Sudhakar, P. Kalaiarasan, N. Subbarao, G. Wadhawa, M. Lohani, M.K.A. Khan, A.U. Khan, Insight into the effect of inhibitor resistant S130G
mutant on physico-chemical properties of SHV type beta-lactamase: a molecular dynamics study, PLoS One 9 (12) (2014) e112456, https://doi.org/10.1371/
journal.pone.0112456.
[57] S.S. Mousavi, A. Karami, T.M. Haghighi, S.G. Tumilaar, R. Idroes, S. Mahmud, I. Celik, D. Ağagündüz, T.E. Tallei, T.B. Emran, In silico evaluation of Iranian
medicinal plant phytoconstituents as inhibitors against main protease and the receptor-binding domain of SARS-CoV-2, Molecules 26 (18) (2021) 5724, https://
doi.org/10.3390/molecules26185724.
[58] L. Stella, S. Melchionna, Equilibration and sampling in molecular dynamics simulations of biomolecules, J. Chem. Phys. 109 (23) (1998) 10115–10117, https://
doi.org/10.1063/1.477703.
[59] D. Gilis, M. Rooman, Stability changes upon mutation of solvent-accessible residues in proteins evaluated by database-derived potentials, J. Mol. Biol. 257 (5)
(1996) 1112–1126, https://doi.org/10.1006/jmbi.1996.0226.
[60] S. Khan, I. Bjij, R.M. Betz, M.E. Soliman, Reversible versus irreversible inhibition modes of ERK2: a comparative analysis for ERK2 protein kinase in cancer
therapy, Future Med. Chem. 10 (9) (2018) 1003–1015, https://doi.org/10.4155/fmc-2017-0275.
[61] D. Zhang, R. Lazim, Application of conventional molecular dynamics simulation in evaluating the stability of apomyoglobin in urea solution, Sci. Rep. 7 (1)
(2017) 44651, https://doi.org/10.1038/srep44651.
[62] M. Moussaoui, M. Baassi, S. Baammi, H. Soufi, M. Salah, R. Daoud, A. El Allali, M. Belghiti, S. Belaaouad, In silico design of novel CDK2 inhibitors through
QSAR, ADMET, molecular docking and molecular dynamics simulation studies, J. Biomol. Struct. Dyn. (2023) 1–17, https://doi.org/10.1080/
07391102.2023.2212304.
[63] J. Kataria, P. Rani, P. Devi, Evaluation of molecular interactions in binary mixtures comprising ethylene and di-ethylene glycol with ethyl lactate through
thermophysical and spectroscopic studies, J. Mol. Liq. 343 (2021) 117626, https://doi.org/10.1016/j.molliq.2021.117626.
[64] X. Du, Y. Li, Y.-L. Xia, S.-M. Ai, J. Liang, P. Sang, X.-L. Ji, S.-Q. Liu, Insights into protein–ligand interactions: mechanisms, models, and methods, Int. J. Mol. Sci.
17 (2) (2016) 144, https://doi.org/10.3390/ijms17020144.
[65] N.W. S’thebe, J.O. Aribisala, S. Sabiu, Cheminformatics bioprospection of sunflower seeds’ oils against quorum sensing system of Pseudomonas aeruginosa,
Antibiotics 12 (3) (2023) 504, https://doi.org/10.3390/antibiotics12030504.
[66] R. Kretschmer, D. Kinzel, L. González, The role of hydrogen bonds in protein–ligand interactions. DFT calculations in 1, 3-dihydrobenzimidazole-2 thione
derivatives with glycinamide as model HIV RT inhibitors, Int. J. Quantum Chem. 112 (7) (2012) 1786–1795, https://doi.org/10.1002/qua.23001.
[67] R. Patil, S. Das, A. Stanley, L. Yadav, A. Sudhakar, A.K. Varma, Optimized hydrophobic interactions and hydrogen bonding at the target-ligand interface leads
the pathways of drug-designing, PLoS One 5 (8) (2010) e12029, https://doi.org/10.1371/journal.pone.0012029.
[68] J.O. Aribisala, S. Sabiu, Cheminformatics identification of phenolics as modulators of penicillin-binding protein 2a of Staphylococcus aureus: a structure–activity-
relationship-based study, Pharmaceutics 14 (9) (2022) 1818, https://doi.org/10.3390/pharmaceutics14091818.
[69] J.O. Aribisala, S. Sabiu, Cheminformatics identification of phenolics as modulators of penicillin-binding protein-3 of Pseudomonas aeruginosa towards
interventive antibacterial therapy, J. Biomol. Struct. Dyn. (2023) 1–16, https://doi.org/10.1080/07391102.2023.2192808.

21