Potential inhibitory activity of phytoconstituents against black fungus: In silico ADMET, molecular docking and MD simulation studies.

Mucormycosis or "black fungus" has been currently observed in India, as a secondary infection in COVID-19 infected patients in the post-COVID-stage. Fungus is an uncommon opportunistic infection that affects people who have a weak immune system. In this study, 158 antifungal phytochemicals were screened using molecular docking against glucoamylase enzyme of Rhizopus oryzae to identify potential inhibitors. The docking scores of the selected phytochemicals were compared with Isomaltotriose as a positive control. Most of the compounds showed lower binding energy values than Isomaltotriose (-6.4 kcal/mol). Computational studies also revealed the strongest binding affinity of the screened phytochemicals was Dioscin (-9.4 kcal/mol). Furthermore, the binding interactions of the top ten potential phytochemicals were elucidated and further analyzed. In-silico ADME and toxicity prediction were also evaluated using SwissADME and admetSAR online servers. Compounds Piscisoflavone C, 8-O-methylaverufin and Punicalagin exhibited positive results with the Lipinski filter and drug-likeness and showed mild to moderate of toxicity. Molecular dynamics (MD) simulation (at 300 K for 100 ns) was also employed to the docked ligand-target complex to explore the stability of ligand-target complex, improve docking results, and analyze the molecular mechanisms of protein-target interactions.

Introduction

Mucormycosis, is known as black fungus is an uncommon and deadly fungal infection, that was previously termed as ‘zygomycosis’ [1]. Mucormycosis is caused by fungi belonging to the Mucorales order and the family Mucoraceae, first reported in the history by Paultauf [2], [3]. Rhizopus oryzae is the primary fungus responsible for around 70 % of all disease manifestation and 90 % of all rhinocerebral cases [4]. Generally, mucormycosis has emerged in immunocompromised patients, including uncontrolled diabetes, bone marrow transplantation, other types of metabolic acidosis, corticosteroid therapy, and solid organ, malignant haematological disorders and deferoxamine therapy in patients receiving haemodialysis [5], [6]. Recently, the pandemic coronavirus disease-2019 (COVID-19) which is caused by SARS-CoV 2 is another fiery example of a patient’s immune system weakening whereby the patient becomes susceptible to such opportunistic fungal infections. There is no genetic link between COVID and mucormycosis, they are immunologically interconnected [7], [8], [9], [10], [11], [12], [13]. COVID-19 infection increases the level of pro-inflammatory cytokines like IL-1, IL-6, tumor necrosis factor alpha (TNF-α), and may decrease the number of immune cells that maintaining the immune system homeostasis such as CD4 + and CD8 + T cells, which may lead to increased susceptibility for those patients with bacterial and fungal co-infections [14].

Just after the COVID-19 s wave, significantly high number of black fungus cases were reported in India as post-COVID infections [15], [16], [17]. Incidences were found to be relatively high in the patients who received corticosteroid including patients admitted to intensive care units and longer duration of hospital stay. Corticosteroid was extensively used during COVID-19 in order to reduce inflammation in lungs and repairing damages caused by the overdriving of immune system of the body. Steroid therapy weakens the immune system of the host and both diabetic and non-diabetic patients lead to a rise in blood sugar levels which both factors increase the susceptibility of host toward mucormycosis. Moreover, diabetic condition is counted as a high-risk factor for black fungal infection [18], [19], [20], [21]. Therefore, if this condition of mucormycosis without any protection, it can readily enter the body and cause severe infection, leading the patient to inevitable death. The route of entry by this fungus is via inhalation of the spore, so it is advisable to keep the house mold-free [22], [23]. The aggressive form of coronavirus infects the lungs, which makes the patients more prone to get infections that are airborne, including mucormycosis [24].

Black fungus affects multiple organs, predominantly, brain, lungs and sinuses, and it has symptoms are similar to COVID-19 infections including fever, cough, etc., which both diseases can be distinguished by the clinical investigations. Other symptoms that are linked to black fungus infections are as follows: swelling in one side of the face, headache, fever, nasal congestion, black lesions on the nose or inside the mouth [7], [25], [26], [27]. Most common treatment strategies involve high doses of liposomal amphotericin B, and sometimes surgical resections whenever required. The mortality rate for mucormycosis is higher than 50 % and depends on the affiliated diseases [5], [22], [28], [29]. The study was done in India by Patel A. et al., whereby diabetes was a risk factor, with or without ketoacidosis occurring in 73.5 % of patients with mucormycosis [30].The death rate is 90–100 % in patients with disseminated disease, central nervous system infection or prolonged neutropenia [31], [32], [33], [34], [35]. Unfortunately, the unacceptably high mortality rate, limited options for therapy, the high cost of managing mucormycosis, and the highly disfiguring surgical therapy, etc., drive the urgent need of newer therapeutic drugs to treat the disease [6].

Discovery and development by de nava design is a time-consuming and costly process, so, repurposing can be an appropriate strategy to search potent molecules from the pre-existing databases. The computational approaches such as molecular docking and dynamics simulations, in silico ADMET and drug-likeness predictions are mostly used to discovery and development of drug candidates from several databases. Considering literature data, many active agents against fungal infections have been documented in the herbal medicine. The natural products comprising of a diverse range of secondary metabolites including flavonoids, coumarins, limonoids, terpenoids, phenolic compounds, tannins, quinones, saponins, xanthones, alkaloids, peptides, and biosurfactants were reported having in vitro antifungal properties [36], [37], [38], [39], [40], [41], [42], [43], [44], [45]. Therefore, in recent year research on these phytochemicals and their derived compounds have been the subject of increased investigation due to their importance in drug design. Plants have a variety of active compounds having antifungal activity against various fungus strains, these are important to humans. For instance, there are currently-two major drug classes in use (amphotericin B, the gold standard and armamentarium and the lipopeptide caspofungin) are the important antifungal drugs, which are derived from natural products [46], [47], [48], [49].

In this work, 158 phytochemicals contain of terpenoids, limonoids, phenolic compounds, tannins, flavonoids, coumarins, quinones, saponins, xanthones, alkaloids, peptides and other classes of compounds were screened based on their reported bioactivity to discovery of hit for therapeutic against protein of mucormycosis. In this study, we selected the glucoamylase enzyme of Rhizopus oryzae (RoGA) for screening of phytochemical. Glucoamylase (EC 3.2.1.3), also known as amyloglucosidase; glucan 1,4-α-glucosidase; amylase; 1,4-α-d-glucanglucohydrolase is amylolytic enzyme, that yieldsβ-d-glucose from the starch and polysaccharide by hydrolyzing α-1,4 and α-1,6 linkages, produced by the different sources such as fungi (Aspergillus, Penicillium spp. Rhizopus oryzae), yeast and bacteria [50], [51]. Glucoamylase (GA) belongs to the family of the glycoside hydrolase 15 (GH15) due to its structural similarity to that enzyme group or family. Glycoside hydrolase family 15 comprises enzymes with several known activities; glucoamylase; alpha-glucosidase; glucodextranase [52]. RoGA is comprising of starch binding domain at N-terminal and catalytic domain at C-terminal terminal, connected by an o-glycosylated linker. It can hydrolyze both amylose and amylopectin by breaking α-1,4 as well as α-1,6 glycosidic bonds and produce glucose units [53].

Glucose is an essential nutrient for the growth of fungal infections because these metabolites serve as carbon skeletons for the biosynthesis of other molecules as well as serving precursors for energy production and being involved in cellular signaling pathways in fungi [54]. GA among these, Rhizopus spp is well known for significant production of glucoamylase enzyme, which plays an important role in fungal growth and life cycle [55]. Glucoamylase is also produced within the human intestinal body as an essential enzyme besides carbohydrates and long-chain starches are broken down into sugars by this enzyme, which the body uses as fuel. The primary function of all essential enzymes is to break down most starches and carbohydrates to their purest forms so that the body can obtain nutrients from these foods [56].

Several studies have been reported that Glucoamylase inhibitors exhibited antimicrobial, antifungal, antibacterial activities and anti-diabetic for type II diabetes [57], [58], [59]. There are many studies have been revealed that Rhizopus oryzae would use its spore coat protein CotH3 of Mucorales to interact with glucose-regulated protein 78 (GRP78) as a receptor on nasal epithelial cells to penetrate and destroy the cells. The ketone body levels (hallmark features of Diabetic ketoacidosis (DKA)), iron, and high glucose, increase expression of the CotH3 and GRP78 proteins, potentially leading to frequently fatal cerebral mucormycosis and rhinoorbital [60], [61], [62].Increasing expression of the GRP78 receptor in COVID-19 patients may promote Mucorales spore binding, resulting in increased endothelial invasion and damage [63]. Teclegiorgis et al., found that inhibiting Rhizopus CotH3 expression blocked invasion even in mammalian cells overexpressing GRP78, showing the importance of this fungus protein in pathogenicity. It was also found that the Polyclonal antibodies produced toward both peptides in CotH3′s GRP78-binding domain inhibited fungal invasion and adhesion of endothelial cells in vitro and protected DKA mice from Rhizopus oryzae associated pulmonary mucormycosis [62]. Likewise, repression of GRP78 and CotH3 expression by GA inhibitors is probably due to their ability to block glucose production, which would protect endothelial cells from Rhizopus oryzae -induced endocytosis and subsequent damage.

Experimental

Ligand preparation

A list of 158 bioactive compounds was collected from various herbals that were reported as anti-fungal agents [64], [65], [66], [67], [68], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83], [84], [85], [86], [87], [88], [89], [90], [91], [92], [93], [94], [95], [96], [97], [98], [99], [100], [101], [102], [103], [104], [105], [106], [107], [108], [109], [110], [111], [112], [113]. The 3D structure of phytochemicals were downloaded from Pubchem database (https://pubchem.ncbi.nlm.nih.gov/) in SDF format. The structures that were not available in PubChem were drawn using the Avogadro tool (version 1.2.0) (https://avogadro.cc/) [114]. The structures were further converted into PDBQT format using Open Babel software (version 2.3.1) (https://openbabel.org/wiki/Main_Page) [115]. MMFF94 force field and conjugate gradient optimization algorithm were used to minimize the ligand energy in 200 steps using PyRx-Python prescription 0.8 [116].

Receptor preparation

The 3D crystal structure of Glucoamylase enzyme(PDB ID: 4BFO) was retrieved from RCSB protein data bank (https://www.rcsb.org/structure/4BFO) [117]. The resolution of the obtained structures was 1.18 Å. The file was opened using BIOVIA Discovery Studio Visualizer (v17.2.0.16349). The protein preparation was carried out by eliminating the native inhibitor and other heteroatoms, including water molecules. The receptor was then saved as pdb format. AutoDock Tool [118] was used to add polar hydrogen atoms to the receptor, to determine Kollman charges, and Partial charges were assigned. Then, the prepared file was saved as pdbqt format.

Molecular docking

The original ligand was re-docked on the target receptor to validate the docking method, BIOVIA Discovery Studio Visualizer (DSV) was used to eliminate this ligand from the receptor. The grid box parametric dimension values were adjusted as X  = -21.75700, Y = -0.386824 and Z = 11.718029. The exhaustiveness value was set as 8 to obtain an efficient binding conformation pose of the protein–ligand complex. Afterward, the retrieved phytochemicals were docked against the protein of target with PDB ID:4BFO. AutoDock Vina [119]. was used to perform the docking simulation. Finally, AutoDock Vina generated a docked complex for each ligand with various conformation and affinity scores (in kcal/mol) and ranked them based on the lowest binding energy theory (kcal/mol) of docked complexes (whereby more negative value means greater binding affinity). The best protein–ligand complex docked pose were analyzed and graphically visualized by DSV (https://www.3ds.com/products-services/biovia/).

In-silico predicted ADMET profiles

In-silico prediction of ADME properties of the selected phytochemicals was carried out using SwissADME online server [120], while toxicity was predicted using the Organ toxicity and Genomic toxicity model from admetSAR 2.0 [121] as reported in literature [122]. The organ toxicity is classified under the category: Drug-induced liver injury, Human ether-a-go-go-related gene (hERG) inhibition, acute toxicity, Eye injury and Eye corrosion. The genomic toxicity is classified: ames mutagenesis, carcinogenesis and micronucleus assay. The smile notation for all the selected 10 compounds was considered as the starting point and input to the SwissADME and admetSAR 2.0 webserver and thereby ADME and toxicity predictions were carried out.

Molecular dynamics (MD) simulations

The best docked poses of Dioscin and Piscisoflavone C obtained from the docking was used further for MD simulation and evaluating their conformational space and inhibitory potential. The GROMACS package (Ver. 2019.2) with GROMOS96 43a1 force field was carried out to analysis the MD simulation. Parameter files and topology of ligands were prepared utilizing the latest CGenFF via CHARMM-GUI [123], [124]. The SPC water models that extended 10 from the protein was used to solvate the structures of both complexes in a triclinic box [125]. In order to mimic physiological salt concentration, 9 Na+ and 9 Cl− ions (0.15 M salt) were added to neutralize the systems for both cases, as shown in Fig. 1 . In the NPT/NVT equilibration run, both systems were introduced to periodic boundary conditions (1.0 bar, 300 K) utilizing a Leap-frog MD integrator for a 100 ns [126]. For electrostatic interactions, the particle-mesh Ewald protocol was applied [127]. The force-based switching algorithm was applied to truncate interactions (non-bonded) above 10 and 12 Å [128]. Minimization of energy was used to eliminate bad contact inside the system via the steepest descent with 5000 steps. Additionally, to avoid the cold solute-hot solvent problem, the indexing system applied temperature coupling into both water and non-water components. Furthermore, for NVT/NPT equilibrations run, a modified Berendsen thermostat and a Parrinello-Rahmanbarostat were performed. Trajectory analysis was carried out using GROMACS analysis. Calculation of the root mean square deviation (RMSD) and root mean square fluctuations (RMSF) of receptor were carried out utilizing gmxrms and gmx rmsf, respectively. The gmx sasa and gmx gyrate tools were used to calculate the solvent accessible surface area (SASA) and radius of gyration (RG), respectively. Hydrogen bonds were analyzed by gmxh bond tool. Grace Software was utilized to generate the plots. The complex structure was visualized by PyMol and VMD [129], [130]. The simulation time for the ligand-receptor was 10.4 and 10.3 h for Dioscin and Piscisoflavone C, respectively, for 100 ns.

Fig. 1

Protein-ligand complex(a) Dioscin and (b) Piscisoflavone C in triclinic box solvated with water molecules and neutralized with 9 Na+ and 9 Cl− ions (0.15 M salt).

Results

Selected 158 anti-fungal phytochemicals were docked against target enzyme. The native ligand, Isomaltotriose was used as the control for comparative purpose. Molecular docking of these phytochemicals revealed binding affinity ranging from −4.1 to −9.4 kcal/mol. Among the 158 phytochemicals screened, 123 phytochemicals showed docking energy values higher than Isomaltotriose (-6.4 kcal/mol). The interactions of the best 10 compounds (Fig. 2 ) is depicted in Table 1 . These compounds that have higher potential making hydrogen bond and hydrophobic interactions with Tyr32, Lys34, Lys35, Phe58 and Ser57 were common interacting residues.

Fig. 2

The molecular structure of the best 10 screened phytochemicals.

Table 1

The docking score of the best 10 screened phytochemicals and their interactions with the enzyme.

Entry	Compounds	B.E (kcal/mol)	Interactions
			H-Bond	Hydrophobic	Others
1	Hopeanolin	−8.8	Tyr32, Lys34, Phe58	π-π stacked: Tyr32, Phe58; π-alkyl:Phe58	Glu68 (π-anion)
8	Vaticanol E	−8.7	Lys34, Lys35, Ser57	π-alkyl:Phe58
11	Vaticanol B	−8.5	Tyr32, Lys34, Ser57, Phe58	π-π stacked: Tyr32, Phe58, Trp70; π-alkyl: Tyr32, Phe58	Glu68 (π-anion)
12	alpha-Viniferin	−8.6	Asn29, Tyr32, Tyr67, Glu68	π-π stacked: Tyr32, Phe58; alkyl: Phe58, Pro61	Glu68 (π-anion)
13	Pauciflorol A	−8.7	Asn29, Tyr32, Ser65	π-π stacked: Tyr32, Phe58; π-alkyl: Tyr32, Phe58	Glu68 (π-anion)
31	Dioscin	−9.4	Lys34, Lys35, Ala56, Ser57, Phe58	π-alkyl: Tyr32, Phe58	Phe58 (π-sigma)
114	8-O-methylaverufin	−8.3	Lys34, Lys35, Phe58	π-alkyl: Tyr32, Phe58, Trp70	Ala55 (π-sigma)
118	Piscisoflavone C	−8.3	Asn29, Lys34	π-π stacked: Tyr32, Phe58; alkyl:Pro61	Glu68 (π-anion)
139	Punicalin	−8.3	Tyr32, Lys34, Lys35, Ala56, Ser57, Phe58
145	Punicalagin	−8.2	Lys34, Lys35, Ser57	π-π stacked: Tyr32, Phe58; π-alkyl: Ala55
	Isomaltotriose	−6.4	Tyr32, Lys34, Ser57, Phe5		Tyr32(π-sigma)

B.E: Binding Energy.

In silico predicted ADMET profiles

Predicted ADME profiles

The physico-chemical properties of the top ten compounds were given in the Table 2 .

Table 2

Results of predicted ADME properties of the 10 best compounds.

Entry	Formula	M.W	FC	HBA	HBD	MR	TPSA	X3	WS	GI.A	B.P	Pg	LV	B.S	S.A
1	C42H28O10	692.67	0.14	10	5	186	162.98	4.9	P.S	Low	No	Yes	1	0.11	7.05
8	C42H32O9	680.7	0.14	9	8	189.88	171.07	6.84	P.S	Low	No	Yes	2	0.17	6.24
11	C56H42O12	906.93	0.14	12	10	252.25	220.76	9.12	Ins	Low	No	Yes	3	0.17	7.55
12	C42H30O9	678.68	0.14	9	6	187.09	149.07	6.83	P.S	Low	No	Yes	2	0.17	6.41
13	C42H32O9	680.7	0.14	9	8	189.88	171.07	6.84	P.S	Low	No	Yes	2	0.17	6.24
31	C45H72O16	869.04	0.96	16	8	216.42	235.68	1.34	M.S	Low	No	Yes	3	0.17	10
114	C21H18O7	382.36	0.33	7	2	97.87	102.29	3.3	M.S	High	No	Yes	0	0.55	4.85
118	C21H18O5	350.36	0.19	5	1	100.56	68.9	3.72	M.S	High	Yes	No	0	0.55	3.83
139	C34H22O22	782.53	0.18	22	13	180.45	385.24	−0.29	M.S	Low	No	Yes	3	0.17	6.74
145	C20H16O5	336.34	0.15	5	2	96.09	79.9	3.94	M.S	High	No	No	0	0.55	3.76

M.W: Molecular weight (g/mol), FC: Fraction Csp, HBA: Number of H-bond acceptors, HBD: Number of H-bond donors, MR: Molar Refractivity, TPSA: Topological polar surface area (Å2), X3: XLOGP3,WS: Water Solubility Class, GI.A: Gastrointestinal Absorption, B.P: Blood Brain-Barrier permeant, Pg: P-glycoprotein substrate, LV: Lipinski Violation, B.S: Bioavailability Score, S.A: Synthetic accessibility.

Predicted toxicity results

The organ toxicity and genome toxicity were summarized in Table 3 .

Table 3

Results of predicted toxicity of the 10 best compounds.

Entry	1	8	11	12	13	31	114	118	139	145
Organ Toxicity
Acute Oral Toxicity	II	III	III	III	III	III	III	III	III	III
Eye corrosion	–	–	–	–	–	–	–	–	–	–
Eye irritation	–	–	–	–	–	–	–	+	–	+
Hepatotoxicity	+	+	+	+	+	–	+	+	+	+
Human either-a-go-go inhibition	+	+	+	+	+	+	–	–	+	–
Genomic toxicity
Ames mutagenesis	+	–	–	–	–	–	+	–	+	–
Carcinogenicity (trinary)	Danger	N.R	N.R	Warning	N.R	N.R	N.R	Danger	N.R	N.R
micronuclear	+	+	+	+	+	–	–	+	+	+

N.R: Non-required.

Molecular dynamics simulation

The best docked conformations of each Dioscin and Piscisoflavone C, which exhibited significant results based on in-silico predicted ADMET analysis, were selected for 100 ns MD simulation. The best conformation was utilizing to set up dynamics simulation process in a high-throughput manner for analyzing the mechanism of ligands binding dynamics with target receptor under clearly expressed water conditions.

Discussion

Molecular docking studies investigating for mucormycosis inhibitors exhibited that the natural compounds are already reported for inhibiting the growth of Rhizopus oryzae. The fungal cell walls and enzymes have been previously founding as an attractive target for the discovery of new antifungal agents. According to the results obtained, it has been shown that the higher dock score was against the receptor of Rhizopus oryzae, and even comparatively higher than the native ligand, Isomaltotriose. Evidence from several in vivo, in vitro, and animal studies suggests that phenolic compounds such as phenols, coumarins, flavonoids, tannins, quinines, xanthones and stilbenoids from natural sources have antifungal properties and, it’s also shown that their activity is mainly due to the site (s) and the number of hydroxyl groups in a compound are assumed to be related to their relative toxicity to microorganisms. As previously stated, increasing toxicity is related to more hydroxylation [80], [108], [113]. Moreover, phenol compounds with higher degree of oxidation were shown more inhibitory activities [131]. Studies also indicated that the antifungal activities of flavonoids like isopiscerythrone, allolicoisoflavone A, piscisoflavones A, B and C, licoflavone A, papyriflavonol A, quercetin, hesperidin, neohesperidin and naringin, because of their ability to bind to extracellular and soluble proteins, as well as to bind to fungal cell walls. The more lipophilic nature of flavonoids may also account for the disrupted fungal membranes [69], [80], [90], [96], [99], [107]. A large number of studies have established that terpenoids such as estragole, encelin and panicutine exhibit antifungal activity. Terpenes' mechanism of action is not fully explained, but it is thought to include membrane disruption due to their lipophilic nature. The addition of a methyl group to the hydrophilicity of kaurene diterpenoids drastically decreased their antimicrobial activity as reported by Mendoza et al., [132]. Encelin, a known sesquiterpene lactone extracted from the Mexican species Montanoa speciosa, has a significant effect on fungal cell development and morphogenetic processes [133]. Limonoids are highly oxygenated and modified triterpenes and to date, more than 300 limonoids are isolated, which include: obacunone, corosolic acid, maslinic acid, hydroxyseneganolide, oleanane, epoxyazadiradione,7-deacetylgedunin, oleanolic acid, ursolic acid, limoninglucoside, azadiradionolide, eugenol,1,3-dideacetyl-7-deacetoxy-7-oxokhivorin, seneganolide A, B., methyl 6-hydroxyangolensate and gedunin. Limonin is the first isolated limonoids found in medicinal plants, such as citrus, neem, tulsi and licorice. Limonoids have been identified as having various biological properties including antibacterial, antifungal, antiviral, antimalarial, and anticancer [72], [78], [80], [106]. Saponins are steroidal or terpenoid-based glycosides with surface-active properties such as saponin and sapindoside B. CAY-1, a triterpene saponin from the Capsicum frutescens, was reported to be potent against 16 different fungal strains, including C. neoformans, A. fumigatus, and Candida spp. It has been shown that the antifungal properties of CAY-1 can be attributed to the breakdown of fungal cell membrane integrity [91], [134], [135]. Nowadays, an alkaloid namely, 2-(3,4-dimethyl-2,5-dihydro-1H-pyrrol-2-yl)-1-methylethyl pentanoate was extracted from the Daturametel and displayed in vivo and in vitro activities against both species (Candida and Aspergillus). The antifungal alkaloids such as β-carboline, derived alkaloids including two phenylethylamine along with tryptamine, and N-methyl-N-formyl-4-hydroxy-beta-phenylethylamine from Cyathobasis fruticulosa, haloxylines A and B, new piperidine from Haloxylon salicornium all displayed antifungal potentials. The reduced fungal growth of Aspergillus flavus has been shown to correlate with reduced aflatoxin production in laboratory studies [136]. Jatrorrhizine isolated from Mahonia aquifolium was the most potent towards all fungal species [68], [80]. Fig. 3 shows the experimental part of this research.

Fig. 3

Flowchart of the research work.

The docking studies revealed that the Dioscin binds to the target enzyme with highest dock score of −9.4 kcal/mol. The illustration of molecular docking for the interactions of Dioscin and receptor was depicted in Fig. 4 a, b. The analysis of the best docked pose of Dioscin exhibited that the amino acid residues including Lys34, Lys35, Ala56 and Ser57 had formed hydrogen bond interactions. Additionally, Phe58 established two interactions including hydrogen bond, π-alkyl and π-sigma. Moreover, Dioscin interacted with residue Tyr32 via π-alkyl interaction. The Piscisoflavone C which exhibited significant in silico ADMET profile was selected for further docking studies. The interactions between Piscisoflavone C and residues in 4BFO were shown in Fig. 5 a, b. The binding energy formed was −8.3 kcal/mol, which was higher compared to the energy recorded by the Isomaltotriose interaction (Table 1). As presented in Fig. 4a, b, this phytochemical made two hydrogen bonds with amino acid residues Asn29 and Lys34. Furthermore, the Piscisoflavone C exhibited three hydrophobic interactions with Tyr32 and Phe58 via π-π stacked interaction and Pro61involved in alkyl interaction. It also established π-anion with Glu68.

Fig. 4

 The putative binding site of Dioscin on target protein (PDB ID: 4BFO). (A) Three-dimensional representation of important interactions. (B) 2D schematic diagram showing amino acids residues involved in interactions.

Fig. 5

 The putative binding site of Piscisoflavone C on target protein (PDB ID: 4BFO). (A) Three-dimensional representation of important interactions. (B) 2D schematic diagram showing amino acids residues involved in interactions.

As illustrated in Table 2, Piscisoflavone C is eligible for further studies based on Lipinski filter i.e. drug-likeness and has 55 % of predictive bioavailability with a synthetic accessibility score of 3.83. This compound is predicted for moderately soluble with a high gastrointestinal (GI) absorption rate, permeable to BBB (Blood-brain barrier) and a non-effluxable compound as it is non-substrate to P-glycoprotein (Pgp,) but might be not readily metabolized by Cytochrome P450 (CYP) class enzymes due to the inhibitory activity predicted for CYP2D6 and predictively can take a longer time to be excreted from the system. Compound 8-o-methylaverufin is eligible for further studies based on Lipinski filter i.e. drug-likeness with 55 % of predictive bioavailability score, synthetic accessibility score of 4.85. 8-o-methylaverufin is predicted for moderately soluble with a high GI absorption rate, non-permeable to BBB (Blood-brain barrier) and can be effluxed out as it is predicted to be a substrate for Pgp. Moreover, the prediction as an inhibitor to CYP1A2 and CYP2C19 may affect the metabolizing rate and process of metabolism and excretion. Meanwhile, Punicalagin is eligible for further studies based on the Lipinski filter i.e., drug-likeness with 55 % of predictive bioavailability with a synthetic accessibility score of 3.76. This compound is moderately soluble with a high GI absorption rate, non-permeable to BBB (Blood-brain barrier) and non-effluxable compound as it is non-substrate to Pgp, but might not be readily metabolized by CYP class enzymes and predictively can take a longer time to be excreted due to its predictive inhibitory properties to CYP2C19 and CYP2D6 enzymes. However, Hopeanolin violates the Lipinski filter i.e., drug-likeness with a higher molecular weight of more than 500 kDa. Compounds Vaticanol E, alpha-Viniferin and Pauciflorol A violate the Lipinski filter with a molecular weight of more than 500 kDa and more than 5 hydrogen bond donors. Compounds Vaticanol B, Dioscin and Punicalin all violate the Lipinski filter with a molecular weight of more than 500-kilodaltons, more than 5 hydrogen bond donors and more than 10 hydrogen bond acceptors. Hopeanolin, Vaticanol E, Vaticanol B, alpha-Viniferin, Pauciflorol A, Dioscin and Punicalin have a poor predictive bioavailability score ranging between 11 % and 17 % with high synthetic accessibility scores ranging from 6.24 to 10.0. Furthermore, Hopeanolin, Vaticanol E, alpha-Viniferin and Pauciflorol A predicted for a poor solubility, low GI absorption rate, non-permeable to BBB (Blood-brain barrier) and can be effluxed out as it is predicted as a substrate to Pgp, but might be readily metabolized by CYP class enzymes and predictively can easily be excreted through the system, because they fall under the prediction of the substrate to CYP class of enzyme. The compound Vaticanol B exhibited insolubility, low GI absorption rate, non-permeability to BBB (Blood-brain barrier), but it can be effluxed out as it is predicted to be a substrate to Pgp, but might be readily metabolized by CYP class enzymes and predictively can easily be excreted from the system, due to its prediction as a substrate to CYP class of enzyme. The Dioscin and Punicalin predicted for a moderately soluble, low GI absorption rate, non-permeable to BBB (Blood-brain barrier), the substrate to CYP class of enzyme hence might be readily metabolized and easily be excreted from the system. But and can be effluxed out as it is predicted as a substrate to Pgp.

The acute oral toxicity was calculated using admetSAR, following compounds exhibited a grade III oral acute toxicity (Table 3): Vaticanol E, Vaticanol B, alpha-Viniferin, Pauciflorol A, Dioscin, 8-o-methylaverufin, Piscisoflavone C, Punicalin and Punicalagin. Grade II oral toxicity was reported only for Hopeanolin. No compound was reported for Eye corrosion, whereas Piscisoflavone C and Punicalagin were reported for Eye irritation. Liver injury/hepatotoxicity was predicted and found only Dioscin to be safe with no hepatotoxicity. The inward rectifying voltage gated potassium channel in the heart (IKr), which is important in cardiac repolarization, is encoded by the human ether-a-go-go-related gene (hERG). Inhibition of the hERG current promotes the QT interval, resulting in Torsade de Pointes, a potentially fatal ventricular tachyarrhythmia. 8-o-methylaverufin, Piscisoflavone C and Punicalagin were predicted as non-inhibitor to hERG. On the other hand, AMES toxicity test was used to know whether a compound is mutagenic or not, excluding Hopeanolin, 8-o-methylaverufin and Punicalin, all others are predicted as non-mutagenic in nature. As per carcinogenicity prediction, Hopeanolin, alpha-Viniferin and Piscisoflavone C gave positive signals for carcinogenesis, whereas others were marked safe. The micronuclear assay prediction reported a genotoxicity positive score for Hopeanolin, Vaticanol E, Vaticanol B, alpha-Viniferin, Pauciflorol A, Piscisoflavone C, Punicalin and Punicalagin. The compounds that were marked safe are Dioscin and 8-o-methylaverufin (Table 3).

The complexes simulations were showed no loss of structural integrity and compactness, which suggests good performance of the applied force field and adequate quality of the experimental structures were used as starting states for the simulations [137], [138], [139]. The conformational changes of the complex were studied during 1, 10, 20, 50 and 100 ns MD simulation run was depicted in Fig. 6 (a,b). The different structures represented in Fig. 6 (a,b) provide a visualization of the dynamics of the process and the sequence of events. Only the ligand is displayed as space-filling for better exhibition of the occurring motions, while the receptor is shown by its backbone. Molecular dynamics data are processed by calculating the RMSD (from the starting structure) to analyze structural stability. It is observed from Fig. 7 Piscisoflavone C complex exhibits the lowest RMSD than Dioscin complex. Even RMSD of the unbound-protein is slightly higher than the Piscisoflavone C, which indicates the greater stability of Piscisoflavone C. From the RMSD diagram the stable complex conformations of Dioscin and Piscisoflavone C were observed after ∼ 60 and 85 ns, respectively, having RMSD values of ∼ 2.8 and 2.2 Å, respectively (Fig. 7). According to the literature, RMSD value of less than 3.0 Å is the most acceptable, as the lower the RMSD value shows greater stability of complex [140], [141]. The obtained results indicated that Piscisoflavone C forms a more stable protein–ligand complex than Dioscin and does not make any considerable conformational change in the protein structure during simulation study. The flexibility of each residue is computed to gain a better understanding of how the binding of a ligand changes the protein flexibility since RMSF aids in understanding the area of the protein that is being varied during the simulation. To assess the average fluctuation and flexibility of individual residues, the RMSF of protein and complexes were plotted using 100 ns MD trajectory, as shown in Fig. 8 . The RMSF plot shows that the fluctuations of residues are occurring in the receptor at several times during the ligand-bound state and the binding makes the protein most flexible in all areas in contrast to unbound -protein and the other complex. The protein structure is found to have the lowest RMSF at some key residues, which indicates that even in unliganded state, the protein is not very much flexible. Overall, the residues are found flexible for both of unbound -protein and ligand-bound complexes. The lowest RMSF value of ∼ 0.067 Å was found in the complex of Piscisoflavone C-receptor. The findings show that ligand–protein interaction brings protein chains closer together and reduces the space between them, shown in Fig. 9 . The standard deviation and average distance for all pairs of residue between two conformations are represented by RR (residue-residue) distance map (two-dimensional representations of protein 3D structure) and it was used for comparing and analyzing the structures of protein. The RR distance maps are represented in Fig. 10 , which plots patterns of spatial interactions [142], [143]. In the map, the white diagonal shows the zero distance between two residues, while the blue and red elements represent the residue pairings with the longest distance variances in the two conformations. A radius of gyration (Rg) plot was used to analyze the compactness of the ligand-bound protein. The lowest Rg value of ∼ 12.75 Å was shown by Piscisoflavone C; on the other hand, Dioscin shows an Rg value of ∼ 13.15 Å. Therefore, the order of compactness and rigidness should be Piscisoflavone C > unbound -protein > Dioscin. A decrease in Rg for complex along with the simulation time was found means an increase in the compactness of the structure (Fig. 11 ). A minute reduce in the Rg value of protein was observed upon interacting with Piscisoflavone C. As a result of the interacting of Piscisoflavone C, the microenvironment of receptor was altered, resulting in conformational changes in the structure of receptor.

Fig. 6

(a). Protein-ligand structure of Dioscin at (a)1 ns, (b) 10 ns, (c) 20 ns, (d) 50 ns and (e) 100 ns MD run, giving a visual impression of the sequence of events and the dynamics of the process. Fig. 6 (b). Protein-ligand structure of Piscisoflavone C at(a) 1 ns, (b) 10 ns, (c) 20 ns, (d) 50 ns and (e) 100 ns MD run, giving a visual impression of the sequence of events and the dynamics of the process.

Fig. 7

The root mean square deviation (RMSD) plot of solvated protein backbone and ligand complex during 100 ns MD simulation for unbound protein (blue), Dioscin (brown color) and Piscisoflavone C (green color). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 8

The root mean square fluctuation (RMSF) plot of solvated unbound protein (blue), protein–ligand complex (Dioscin, brown color) and (Piscisoflavone C, green color) plotted against residue numbers. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 9

Superimposed structure of unbounded protein (red) and protein after simulation (green)(a) Dioscin and (b) Piscisoflavone C. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 10

RR distance map patterns of spatial interactions of the protein(a) Dioscin and (b) Piscisoflavone C showing the average distance and standard deviation for all amino acid pairs.

Fig. 11

Radius of gyration (Rg) of unbound protein (blue), protein–ligand complex (Dioscin, brown color) and (Piscisoflavone C, green color) during 100 ns simulation time. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

As shown in Fig. 12 , Grid-search on a 16x16x19 grid, rcut = 0.35, was used to calculate the hydrogen bond interactions established between ligand and protein, which was then plotted versus time. On calculating hydrogen bonds between ligands (32 atoms for Dioscin and 63 atoms for Piscisoflavone C) and protein (1100 atoms), 164donors and 303 acceptors were observed for Dioscin, whereas, 171 donors and 318 acceptors were observed for Piscisoflavone C. The average number of hydrogen bonds per timeframe was found to be 0.038 and 0.770out of 25,174 and 27,189 possible for Dioscin and Piscisoflavone C, respectively. Overall, it was observed that the ligand interaction with receptor dramatically enhanced the number of hydrogen bonds. During MD simulation in ligand-receptor bound conditions, the SASA of receptor was calculated. Interactions of the ligand-receptor leads to change of the SASA values (Fig. 13 ). The analysis displays the receptor folds and its stability when it binds to a ligand. The decrease of SASA vale of protein upon binding of ligand suggests that the structure of the receptor has altered, with a decreased pocket size and more hydrophobicity around it. The result of the MD simulation indicates the Piscisoflavone C complex is more stable and compact as compared to the Dioscin.

Fig. 12

Number of average hydrogen bonding interactions between protein–ligand complex (Dioscin, brown color) and (Piscisoflavone C, green color) during 100 ns simulation time. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 13

Solvent accessible surface area (SASA) analysis for unbound protein (blue), protein–ligand complex (Dioscin, brown color and Piscisoflavone C, green color) during 100 ns simulation time. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Conclusion

Mucormycosis is a rare opportunistic fungal infection, that attack individuals with weak immune system. Considering its high prevalence as post COVID-19 infection, the development of new antifungal agents, preferably with a novel mechanism of action, is an urgent medical need. In the present study, virtual screening of natural products was performed against the glucoamylase enzyme of Rhizopus oryzae through molecular docking. Docking studies revealed that most of the tested compounds showed good binding affinity with target enzyme. As per the in-silico predicted, ADMET analysis of compounds, Piscisoflavone C, 8-o-methylaverufin and Punicalagin exhibited positive outcomes considering no violation to Lipinski filter and drug-likeness properties and mild to moderate of toxicity. Molecular dynamics simulation of 100 ns run unveils that ligand Piscisoflavone C formed a more stable complex with receptor throughout the simulation time in comparison with Dioscin.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.