Identification of bio-active secondary metabolites from Actinobacteria as potential drug targets against Porphyromonas gingivalis in oral squamous cell carcinoma using molecular docking and dynamics study

Abstract

Chronic periodontitis caused by the bacteria Porphyromonas gingivalis is thought to be a risk factor for the advancement of oral squamous cell carcinoma (OSCC). Virulence factors of P. gingivalis include gingipains, outer membrane surface lipoproteins, and fimbriae contribute to the activation of oncogenic pathways in OSCC by up-regulating different cytokines. Gingipains (Arg and Lys) proteases have an important role in the activation of proMMP-9, which promotes cellular invasion and metastatic ability of OSCC. Thus gingipains and MMP-9 were actively investigated as potential therapeutic targets in OSCC therapy. Various natural bioactive compounds from Actinobacteria have been explored for their anticancer potential in a variety of cancers, but very few studies have been reported in OSCC. Therefore, the current study is focused to identify potential actinobacterial compounds that can be considered as a therapeutic target against gingipains and inflammatory proteins in OSCC through high-throughput virtual screening, Molecular Docking (MD), and Molecular Dynamics Simulation (MDS) approaches. A total of 179 bioactive secondary metabolites of Actinobacteria were explored for their binding affinity against six virulence proteins of P. gingivalis. The Molecular Docking studies revealed that among 179 metabolites screened, Actinosporin G showed a highly acceptable binding affinity of −7.9 kcal/mol with RgpB (1CVR), and exhibited multi-protein targeting and drug-likeness property and passed level of toxicity. Comprehensive docking interaction of the best top-ranked Actinosporin G with OSCC-related protein targets illustrated high binding affinity towards MMP-9 and JAK-1 proteins among all targeted receptor proteins. The molecular dynamic (MD) simulation has been executed for the metabolite Actinosporin G for both bacterial gingipain (RgpB) and MMP-9 & JAK-1 showed stable intermolecular binding with both hydrogen and hydrophobic interactions. In conclusion, this work suggests that the bioactive secondary metabolite of Actinosporin G from Actinobacteria genera may serve as a promising therapy for P. gingivalis-induced OSCC.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s40203-024-00209-0.

Introduction

Oral squamous cell carcinoma (OSCC) is widely recognized as the predominant form of head and neck squamous cell carcinoma, and it ranks in the top ten most prevalent cancer forms on a global scale (Perner and Idel 2021). Colonization by pathogenic microorganisms is strongly associated with inflammation and the progression of cancer. Poor oral health is associated with dysbiosis of the oral microbiome which drives the development of head and neck cancer (HNC) (Galvão-Moreira and Da Cruz 2016). The dysregulation of the microbiome can contribute to the development of cancer by stimulating the proliferation of cells, activating oncogenes, inducing chronic inflammation, and impairing the local and systemic immune responses, leading to the breakdown of the mucosal barrier. There have been reports indicating that chronic inflammation has a role in the promotion of several malignancies, including HNC. Previous studies have demonstrated that microbes can activate pro-inflammatory immune pathways and induce immunosuppressive reactions (Grivennikov et al. 2010). OSCC has often been found to be positively correlated with the presence of traditional periodontal infections, including Porphyromonas gingivalis, Fusobacterium nucleatum, and Streptococcus mutans. The presence of periodontal infections in lesions of oral squamous cell carcinoma is correlated with a worse prognosis (Lamont et al. 2022). P. gingivalis, a gram-negative periodontal pathogen involved in oral carcinogenesis, has been intensively examined among oral bacteria. Numerous studies have comprehensively documented the robust correlation between P. gingivalis and OSCC (Chattopadhyay et al. 2019). The molecular mechanism by which P. gingivalis promotes the progression of OSCC involves several key processes, including cellular invasion and proliferation, chronic inflammation, inhibition of cell apoptosis, and suppression of immune response (Zhang et al. 2019). These effects are mediated by the presence of various virulence factors, such as gingipains R and K (cysteine proteases), fimbriae, and outer membrane surface antigens. Based on the critical literature mining these virulence factors were chosen for our study. The gingipains RgpB and Kgp play a significant role in the degradation of the host extracellular cellular matrix, hence facilitating efficient oral invasion (Kadowaki 2021; Bostanci and Belibasakis 2012). The FimA1 proteins, which are characterized by their elongated structure, are known to bind to β1 integrins located on the surface of cells, facilitating adherence (Cai et al. 2019; Enersen et al. 2013). P. gingivalis infiltrates gingival epithelial cells via FimA, which has an epithelial binding domain. FimA induces invasiveness of tumor through Zinc finger E-box-binding homeobox 1 (Zeb1) dependent activation of N‐cadherin, matrix metalloproteinase 9 (MMP9), and vimentin (Chattopadhyay et al. 2019). On the other hand, the Mfa1 and Mfa3 proteins, which are shorter in length, play a vital role in enhancing the adhesive and invasive capacities observed in periodontal illnesses (Lee et al. 2018). These virulence factors induce the production of different cytokines, including interleukin-1 (IL-1), interleukin-6 (IL-6), interleukin-8 (IL-8), and tumor necrosis factor-α (TNF-α), which in turn promote cell proliferation and apoptotic progression, leading to the development of malignancies (Chattopadhyay et al. 2019). The present study used computational-based high-throughput virtual screening to identify novel therapeutic compounds that may aid in the development of medications targeting OSCC.

Although several treatment options are available for OSCC, a promising option is yet to be identified, as the already available options are either ineffective or cause side effects (Scully and Bagan 2009). Hence, it is imperative to pursue the advancement of secure therapeutic agents and innovative efficacious anticancer medications for the therapy of OSCC. The utilization of natural products as a reservoir of pharmacologically active compounds has a rich historical background with considerable potential for the development of future therapeutic medications (Jayawardene et al. 2021). There have been reports of around 23,000 bioactive secondary metabolites produced by microbes, with actinomycetes accounting for over 10,000 of these compounds, constituting 45% of the total number of bioactive microbial metabolites that have been identified (Bérdy 2005). Actinobacteria have been found to synthesize a broad spectrum of bioactive chemicals with therapeutic uses, including antifungal, antiviral, antibacterial, anticancer, and neuroprotective compounds (Azman et al. 2019). The advancements in biotechnology and bioinformatics have facilitated the emergence of novel methodologies for drug design and enhancement (Uesaka et al. 2022). To enhance cost-effectiveness, risk mitigation, and operational efficiency in the manufacturing process, the pharmaceutical sector has progressively adopted in silico analysis. This computational approach facilitates the execution of diverse analytical tasks, including the evaluation of the quantitative structure–activity relationship of a drug, the identification of pharmacophores (regions within a ligand molecule that exhibit strong binding affinity to receptors), and other modelling techniques (Readhead and Dudley 2013). Hence, computer-aided drug design (CADD) employs advanced computational techniques to facilitate the creation and refinement of molecules to develop novel pharmaceutical compounds. The utilization of CADD in the realm of pharmaceutical development enhances the efficiency of the development procedure, leading to heightened success rates, reduced expenditures on laboratory resources and employees, and expedited outcomes (Warr 2017). This study aimed to identify a potent actinobacteria-based bioactive secondary metabolite as an antimicrobial and anticancer candidate against the virulence factors of P. gingivalis and proteins activated in OSCC by P. gingivalis infection.

Materials and methods

Preparation of databases

The database containing the secondary metabolites extracted from the various sourced Actinobacteria genera was constructed by us based on the comprehensive literature survey of research studies from the year 2018 to March 2023. The actinobacteria-derived secondary metabolites with a strong correlation with several biological activities, including antimicrobial, anti-biofilm, antioxidant, anticancer, and anti-inflammatory properties have been considered for screening. This in-house database contains over 300 bioactive compounds with a focus on natural products and structural analogs. This library will be used further for the virtual screening of different targets. In the context of high throughput virtual screening performed in this study, we employed an internally developed database of around 179 compounds with previously reported 3D crystal structures of compounds designated as "AM-1 to AM-179" (Table S1).

Computational Methods

The binding mode and affinity between the ligands and targeted receptors were virtually screened using PyRx tool (https://pyrx.sourceforge.io/), which is a free open-source package with improved speed and accuracy for docking studies (Dallakyan and Olson 2015).Molecular docking involves three steps: (i) target protein preparation, (ii) ligand preparation, and (iii) molecular docking.

Protein preparation

Six virulence proteins derived from P. gingivalis and eight regulatory proteins associated with OSCC have been selected as targets in order to dock with specific ligands. The virulence proteins targeted in this study were the Arginine (R)-specific gingipain R (RgpB) (PDB ID: 1CVR), Lysine (K)-specific gingipain K (KgpB) (PDB ID: 6I9A), Immunodominant outer-membrane surface receptor antigens B lipoprotein (RagB) (PDB ID: 5CX8), Fimbrial shaft protein (Mfa1) (PDB ID: 5NF3), Tip fimbrial protein (Mfa3) (PDB ID: 5NF4), and Type V pili FimA type 1 (PDB ID: 6JZK). The host cellular receptors for the corresponding virulence protein include Tumor necrosis factor (TNF-α) (PDB ID: 1TNF), Interleukin-6 (IL-6) (PDB ID: 1ALU), Interleukin-8 (IL-8) (PDB ID: 6NZU), Matrix metalloproteinase-9 (MMP-9) (PDB ID: 4XCT), Matrix metalloproteinase-13 (MMP-13) (PDB ID: 3ZXH), p53 (PDB ID: 3ZME), E-cadherin (PDB ID: 4ZTE) and tyrosine kinase protein (JAK1) (PDB ID: 4E4N) of OSCC. The three-dimensional crystal structures of all proteins were acquired from the RCSB Protein Data Bank (https://www.rcsb.org/). The structure was determined using the X-Ray diffraction method with a resolution of 2.10. The proteins were produced by the established methodology using MGL tools 1.5.7. The missing residues were supplemented, and the water molecules around the receptor were eliminated. The crystal structure, active site, and grid generation were modified through the use of co-crystallized molecules. The ligands that were previously bound to the protein were dissociated, and the protein was formed by introducing polar hydrogen atoms. Subsequently, the Kollman charges were computed. The proteins that were prepared were saved in a file format known as Protein Data Bank, Partial Charge (Q), and Atom Type (T) (PDBQT) format before being analyzed to optimize docking efficiency. Protein Contact Atlas interactive web-based database (http://pca.mggroup.bio/) allows users to visualize the interactions among ligands, small molecules, proteins (single or complex), and nucleic acids. This tool utilizes a network of residue-residue interactions to aid in visualizing molecule structure at atomic resolutions. The protein contact atlas is a precomputed database of crystal structures determined through experimental or computational approaches.

Ligand preparation

PubChem (https://pubchem.ncbi.nlm.nih.gov/) was used to download the three-dimensional structures of bioactive secondary metabolites from the in-house compound library containing 179 compounds. During the optimization method, the OpenBabel software in PyRx was employed, and all parameters were selected to achieve a stable structure with the least amount of energy (O'Boyle et al. 2011). The structural optimization approach was used to estimate the lowest global energy of the title chemicals.

Control drug

In addition to the in-house compound library, the control compound Dequalinium (DrugBank Accession Number: DB04209) was also subjected to screening against the target proteins. Dequalinium is a multifunctional antibacterial agent that exhibits a range of targeted activities. Additionally, it has antifungal, antiparasitic, antiviral, anticancer, and neuroprotective characteristics. Dequalinium continues to be present in several over-the-counter medications utilized for the treatment of oral infections and inflammation (Lin Shiau 2020; Bailly 2021). The control drug was also optimized as similar to the aforementioned ligand preparation.

Auto-dock vina analysis

AutoDock Tools (ADT) (https://autodock.scripps.edu/), a free visual user interface (GUI) for AutoDockVina software 1.5.6. (https://vina.scripps.edu/) was used for molecular docking research (Di Muzio et al. 2017). Depending on where the protein and crystalized ligand interacted, the grid size was set by targeting the active site with a size of (x = 40, y = 40, z = 40), and the grid centre was designated at dimensions (x, y, and z) of the co-crystallised ligand of the respective protein. AutoDockVina 1.5.6 was used to dock the grid-box properties in the configuration file. These parameters were applied to all the proteins and ligands. A Discovery Studio Visualizer v21.1.0.20298 (BIOVIA, San Diego, CA, USA) was used to obtain a PDB complex file (Biovia 2017). The AutoDock-Vina significantly improves the average accuracy of binding mode predictions compared to AutoDock V4. In our study, we employed 179 bioactive secondary metabolite compounds against 6 different virulence proteins from P. gingivalis thus employing AutoDockVina exhibits faster operating orders of magnitude than the AutoDock V4 (Eberhardt et al. 2021).

Docking visualization

The analysis involved examining the interactions between the target receptor and ligands. This was done by selecting conformations, wherein nine different conformations were generated for each ligand. These conformations were then prioritized based on their binding energies, with the most favorable (least) free binding energy being prioritized.

Absorption, distribution, metabolism, excretion, (ADME) properties, and toxicity prediction

ADMET profiling was performed on the ligands with the highest binding affinity to the receptor protein and the lowest negative score (Top 3). Swiss ADME (http://www.swissadme.ch/) web server was used to predict drug ADME parameters (adsorption, distribution, metabolism, and excretion) (Daina et al. 2017; Lucas et al. 2019). Structure SMILES were imported via the website's interface to analyze the pharmacokinetic features of compounds. The ADMET properties/parameters were derived using the Swiss ADME drug design study. To calculate toxicity, a computational approach was employed to obtain the safety profile of the targeted chemicals. The toxicity profile can evaluate and determine the mutagenicity, carcinogenicity, Immunotoxicity, and LD50 value. ProTox-II (http://tox.charite.de/protox_II) server was used to determine the toxicity of the selected top-scored compounds including the analysis of various toxicological pathways such as nuclear receptor signalling, and stress response pathways. This evaluation was conducted to determine the drug-like characteristics of the compounds under consideration. Eventually, the most optimal compound was selected for molecular dynamics (MD) simulation investigations after successfully fulfilling the criteria of ADME and Toxicity filtering, as well as satisfying 'Lipinski's Rule of Five' during the molecular docking analysis against bacterial virulence proteins.

The PASS programme (https://www.way2drug.com/passonline/predict.php) functions as an online server that assesses the comprehensive biological potential of a compound by analyzing its structure–activity relationship. By analyzing the desired structure alongside a vast training set of over 205,000 compounds, it can anticipate the pharmacological effects accurately, uncovering over 7200 biological activities. The outcomes of PASS prediction were compiled as a list of potential biological activities, along with the likelihood of being active (Pa) and the likelihood of being inactive (Pi). Furthermore, the Molinspiration tool (https://molinspiration.com/) was used to calculate the pharmacokinetic properties and bioactivity scores. Predicted bioactivity scores were assigned to the compounds for various drug targets, such as enzymes, nuclear receptors, kinase inhibitors; G-protein coupled receptor ligands, and ion channel modulators. The structural activity relationship of the best compound has been carried out using a webserver admetSAR3.0 (http://lmmd.ecust.edu.cn/admetsar2) to determine the drug-likeness property of the selected bioactive compound.

Molecular dynamics simulation

A molecular dynamics simulation analysis of a protein–ligand complex was conducted using the Desmond modules of the Schrodinger 2020-2 suite software (Schrödinger Release 2023; Opo et al. 2021). The objective of this study was to analyze the consistency of the ligand and the stability of the binding mode within the protein's binding pocket. The most promising candidate, based on favorable docking results, was selected for molecular dynamics (MD) simulation to evaluate the stability of the chemical. The cubic box was filled with TIP3P water molecules, which are simple point charges. The system was then solvated, and the OPLS3 force field, which is optimized for liquid simulations, was employed to prepare the entire complex. The interactions within the system were afterward analyzed. To maintain a neutral state of the system, the lowest amount of Na+ ions, in conjunction with salt atoms, was introduced. The molecule was introduced by energy minimization, employing a conjugate approach, with a convergence threshold set at 1 kcal mol⁻¹ Å. After the chemical achieved energy minimization, it underwent equilibrium employing an NPT ensemble for duration of 2 ns. The relaxed complete system was then subjected to 150 ns molecular dynamics (MD) simulations. These simulations were conducted under the NPT ensemble, employing a Marlyna-Tobias-Klein barostat to maintain a pressure of 1 bar and a Nose–Hoover thermostat to maintain a temperature of 300 K. The trajectories obtained from the MD simulation were subsequently subjected to analysis using the Simulation Interaction Diagram (SID) tool inside the Desmond module of the Schrödinger programme. The stability of the ligand–protein complexes was assessed based on the trajectories obtained from the simulation. The present study investigated the protein–ligand complexes system by analyzing many parameters, including root mean square deviation (RMSD), root mean square fluctuations (RMSF), protein–ligand interaction, and energy potential stability.

Molecular mechanics-generalized born surface area (MM-GBSA) calculations

MM-GBSA calculations were performed using web server-based tool, the fastDRH (http://cadd.zju.edu.cn/fastdrh/overview) (Wang et al. 2022). This tool was used to compute the binding energies using the MM-GBSA method which involves below mathematical equations

$$\mathrm{\Delta }{G}_{\text{bind}}=G_{\text{Complex}}-\left(G_{\text{Receptor}}+G_{\text{Ligand}}\right)$$

The MM-GBSA free energy of binding (∆G_bind) is calculated as the difference between the free energy of the complex (G_Complex), and the sum of the free energies of the receptor (G_Receptor), and the ligand (G_Ligand), in solution

$$G_{\text{Molecule}}=\mathrm{\Delta }{E}_{\mathit{MM}}+\mathrm{\Delta }{G}_{\mathit{GB}}+\mathrm{\Delta }{G}_{\mathit{SA}}-T\mathrm{\Delta }S$$

This equation represents the free energy of the molecule. It considers the change in internal energy of the molecule (∆E_MM) calculated using molecular mechanics, the change in solvation energy of the molecule (∆G_GB) calculated using the generalized Born surface area method, the change in surface area of the molecule (∆G_SA), and the change in entropy of the molecule (T∆S) (Genheden and Ryde 2015).

Results

The current study employed a high-throughput integrative computational analysis to identify the most potential treatment candidate among bioactive compounds derived from actinobacteria, with a focus on targeting the virulence proteins of P. gingivalis. The schematic workflow in Fig. 1 illustrates the methodological pipeline employed in this research investigation. Porphyromonas gingivalis is a significant etiological factor in the occurrence of oral gingivitis, which subsequently contributes to the progression of oral squamous cell carcinoma (OSCC). A selection of 179 compounds, encountered by literature mining, were selected to perform molecular docking investigations against bacterial proteins RgpB, Kgp, RagB, Mfa1, Mfa3, and FimA1 (Fig. 2).

Fig. 1.

Methodology workflow. Workflow of molecular dynamics, docking, and high-throughput virtual screening based on structure study of actinobacteria based bioactive secondary metabolites with anti-bacterial and anti-cancer properties against the P. gingivalis virulence proteins in oral squamous cell carcinoma (OSCC)

Fig. 2.

Porphyromonas gingivalis virulence proteins: the three-dimensional crystal structures of P. gingivalis virulence proteins with their respective Protein Data Bank (PDB) IDs

Molecular docking against P. gingivalis virulence proteins

The docking analysis of 179 bioactive secondary metabolites against pathogenic bacterial proteins yielded astoundingly high scores (Table S2). The conformations of ligands with the highest negative affinity scores were selected as the most favorable binding positions. The interactions between the highest-scoring metabolites and each target protein were briefly described (Fig. S1). AM-158 displayed a binding score of −7.9 kcal/mol towards the protein RgpB, as a consequence of its superior binding affinity compared to the other compounds evaluated. Hydrogen bonding, hydrophobic interactions, and van der Waals forces facilitated this bonding. The formation of hydrogen bonds was avidly promoted by amino acid residues, particularly ASP328 and SER156. The residues TRP284, TYR283, and TYR325 were discovered to have a crucial role in van der Waals interactions with the protein RgpB (Fig. 3A). As measured by a binding score of −9.4 kcal/mol, K-gingipain (Kgp) displayed a higher affinity for AM-102. Hydrogen bonds exist between the amino acid residues Leu270 and GLY298 in this interaction. ASP271, LYS294, and ALA303 interact with ligand atoms via halogen bonding, whereas TYR269, VAL272, VAL635, ARG243, and PRO304 participate in van der Waals interactions (Fig. 3B). Regarding the lipoprotein RagB, it was found that the metabolite AM-179 had a binding affinity of -8.6 kcal/mol towards the active residues ARG53, GLN119, ASN44, ASP30, and PHE292 primarily via hydrogen bond interactions. Alkyl van der Waals interactions with the protein RagB are mediated by the residues ALA50 and LEU228 (Fig. 3C). The interaction between fimbriae proteins Mfa1 and AM-157 and Mfa3 and AM-154 resulted in binding scores of −8.5 and −7.3 kcal/mol, respectively (Fig. 3D, E). The FimA1 protein, a member of the type V pilin family, interacted with the ligand AM-67 with a binding score of approximately −7.9 kcal/mol. The formation of hydrogen bonds, hydrophobic interactions, and van der Waals forces facilitated this interaction (Fig. 3F). Table 1 and Fig. 3 illustrate the various modes of interaction between proteins and their high negative affinity scored respective metabolites, taking distance and binding affinity into consideration. Based on their higher binding affinity, the top three metabolites coupled with each virulence protein has been the focus of further studies.

Fig. 3.

Molecular interaction of P. gingivalis virulence proteins with respective top scored compound: binding modes and LigPlot + residues of best docked molecules with their target receptors: A RgpB_AM-158, B Kgp_AM-102, C RagB_AM-67, D Mfa1_AM-157, E Mfa3_AM-154, and F FimA1_AM-179. The residues that contribute to hydrophobic interactions are shown as red arcs with spikes, and the hydrogen bonds are denoted by green dashed lines with the bond distance

Table 1.

Interaction between Actinobacterial metabolite and P. gingivalis virulence proteins

Protein – PDB ID	Ligand ID - PubChem ID	Residues	Ligand	Interaction type	Distance (Å)	Binding Score (kcal/mol)
RGP -1CVR	AM-158- 132576269	ASP328 SER156 TRP284 TYR325 TRP284 TRP284	H O π π π π	Conventional Hydrogen Bond Conventional Hydrogen Bond Pi-Donor Hydrogen Bond Pi-Pi T-shaped Pi-Alkyl Pi-Alkyl	2.6297 2.39064 3.13524 4.90558 3.71789 3.80294	−7.9
KGP-6I9A	AM-102-1200113	LEU270 GLY298 ASP271 ASP271 LYS294 ALA303 ASP271 TYR269 TYR269 VAL272 VAL635 LYS294 PRO305 ARG243 PRO305	O H H H O O π π π π π π π π π	Conventional Hydrogen Bond Carbon Hydrogen Bond Halogen Halogen Halogen Halogen Pi–Anion Pi–Sulfur Pi–Pi T–shaped Alkyl Alkyl Alkyl Alkyl Pi–Alkyl Pi–Alkyl	2.26122 3.46086 3.26228 3.07323 3.08594 3.66211 4.29728 5.09331 4.95133 3.8418 4.97499 4.44332 4.19153 4.46509 5.0525	−9.4
RagB – 5CX8	AM-179-156582121	ARG53 GLN119 ASN44 ASP30 PHE292 ALA50 LEU228	H H H H HN π	Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Pi–Donor Hydrogen Bond Pi–Alkyl Pi–Alkyl	2.23758 2.51015 2.75881 2.76402 2.96442 5.08371 5.4065	−8.6
Mfa1 – 5NF3	AM-157-118715856	GLN175 ARG179 ARG30 LEU178 LEU178	O C π π π	Conventional Hydrogen Bond Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl	2.62581 3.62474 5.41721 5.46621 4.68316	−8.5
Mfa3–5NF4	AM-154-76310274	SER123 VAL136 VAL136 PHE137 HIS216	O π π π π	Carbon Hydrogen Bond Alkyl Alkyl Pi–Alkyl Pi–Alkyl	3.14305 4.37454 3.79871 4.93472 4.50735	−7.3
FimA1-6JZK	AM-67- 128369	ALA102 MET103 LYS108 THR117 LYS108 THR118 MET103 MET103 LEU116 MET103 LEU116 ALA115 LEU116	O O O O O O C π π π π π π	Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Carbon Hydrogen Bond Carbon Hydrogen Bond Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl	2.22999 2.65708 3.0479 1.85576 2.34582 2.8248 4.60163 5.10562 5.39049 4.22861 4.6865 5.40629 5.22129	−7.9

ADME and toxicity test

The ADME/T properties of the three ligands exhibiting the highest affinity scores for all bacterial virulence proteins were determined using SWISSADME and the ProToX II web server (Table 2). Among the metabolites that have been listed, it has been observed that the metabolite AM-158, which is also referred to as actinosporin-G, demonstrates interactions with over five specific bacterial proteins associated with virulence exhibiting an optimal binding score. The ligand AM-158, which satisfied Lipinski's rule of five (Ro5), exhibited adherence to the established standards. We used the ProTox-II online server to assess toxicity and side effects, which allowed us to evaluate numerous toxicological parameters such as acute toxicity, hepatotoxicity, cytotoxicity, carcinogenicity, immunotoxicity, and mutagenicity (Table S3). These evaluations were based on the projected median lethal dosage (LD50) in mg/kg of body weight (Banerjee et al. 2018). The metabolite AM-158 has been classified as class 5 based on the study performed. Furthermore, it has an LD50 value of 3000 mg/kg, showing that it is non-toxic (Table 3). Exploring the potential biological activity of specific bioactive components was done using the PASS online server. PASS predictions provided the range of effects, actions, and potential harm that a compound can cause to biological systems. The values of Pa and Pi range from 0 to 1, and their sum is less than 1 because they are calculated separately. Pa and Pi are indicators of the compound being part of either active or inactive classes, respectively. When the value of Pa is greater than 0.7, there is a high probability of achieving the desired experimental activity for a compound.

Table 2.

ADME properties of Top 3 rank docking scored Actinomycetes metabolites against Porphyromonas ginigivalis virulence proteins

Ligand ID	SA	GI	BBB	Pgp	BS	MW (g/mol)	nHBA	nHBD	TPSA (Å)	iLOGP	WLOGP	nLV
AM-67	3.99	Low	No	No	0.11	474.42	9	5	169.43	1.01	2088	0
AM-102	4.08	Low	No	Yes	0.55	555.59	8	1	139.38	4.24	7.38	1
AM-106	3.7	Low	No	No	0.55	521.54	9	1	143.466	3.32	6.56	1
AM-126	2.02	High	Yes	Yes	0.55	260.33	0	2	31.58	2.1	4.8	0
AM-146	2.57	High	Yes	No	0.55	268.26	4	1	63.6	2.4	2.48	0
AM-154	5.86	High	Yes	Yes	0.55	421.57	3	3	69.56	4.31	4.41	0
AM-157	6.37	Low	No	Yes	0.17	612.58	13	7	212.67	1.8	0.04	3
AM-158	5.15	High	No	Yes	0.55	450.44	8	1	133.52	2.96	1.84	0
AM-163	6.38	High	No	Yes	0.55	391.55	4	2	53.17	3.85	6.01	0
AM-178	5.51	Low	No	Yes	0.11	502.47	11	6	191.05	1.71	0.14	3
AM-179	5.83	Low	No	No	0.17	591.61	10	6	182.85	3.2	1.22	3

SA synthetic accessibility, GI gastrointestinal absorption, BBB blood–brain barrier permeant, Pgp P-glycoprotein substrate, MW molecular weight, nHBD number of hydrogen bond donor, nHBA number of hydrogen bond acceptor, BS Bioavailability Score, TPSA topological polar surface area, WLOGP water partition coefficient, nLV number of Lipinski violation

Table 3.

Pharmocokinetics and toxicity properties of top 3 rank docking scored Actinomycetes metabolites against Porphyromonas ginigivalis virulence proteins

Ligand ID		AM-67	AM-102	AM-106	AM-126	AM-146	AM-154	AM-157	AM-158	AM-163	AM-178	AM-179
Classification	Target
Toxicity class	Tox	4	4	4	5	4	3	5	5	6	5	5
Lethal dosage	LD50 (mg/kg)	2000	1000	560	2300	650	196	3000	3000	7000	3000	3000
Organotoxicity	Hepatoxicity	Inactive	Active	Active	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive
Toxicity endpoints	Carcinogenicity	Inactive	Active	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive
	Immunotoxicity	Active	Inactive	Inactive	Inactive	Inactive	Active	Active	Active	Active	Active	Active
	Mutagenicity	Inactive	Inactive	Inactive	Active	Inactive	Inactive	Active	Active	Inactive	Inactive	Inactive
	Cytotoxicity	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Active
Tox 21- Nuclear receptor signalling pathways	nr_ahr	Inactive	Inactive	Inactive	Active	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive
	nr_ar	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive
Tox21- Stress response pathway	sr_hse	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive
	sr_mmp	Active	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Active	Inactive	Active	Inactive
	sr_p53	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Inactive	Active	Inactive

ToxPredicted toxicity class, LD50 Predicted Lethal Dosage; nr_ahr Aryl hydrocarbon Receptor; nr_ar Androgen Receptor; sr_hse Heat shock factor response element; sr_mmp Mitochondrial Membrane Potential; sr_p53 Phosphoprotein (Tumor suppressor)p53

According to the PASS prediction results, the selected compounds evaluated in Table S4 showed a higher Pa value than the Pi value for biological properties, indicating their potential for anticancerous effects. AM-67, AM-157, AM-158, AM-178, and AM-179 exhibited anti-carcinogenic properties within the range of 0.396–0.838. Out of the five active compounds, AM-158 (Actinosporin G) showed a high Pa value of approximately 0.838 in comparison to the rest. These compounds might inhibit metastasis by obstructing protein regulators of OSCC, as indicated by docking studies. The drug likeliness property of the top-scoring compounds against various targets was investigated and compiled in Table S4. If a molecule has a bioactivity score above 0.00, it is likely to have significant biological effects. Scores between −0.50 and 0.00 are considered moderately active, while scores below −0.50 are assumed to be inactive (Srivastava 2021). The study's findings revealed that the compounds under investigation exhibited biological activity and induced physiological effects by interacting with various types of ligands and enzymes. AM_158 displayed a bioactivity score of −0.12 for the GPCR ligand. The bioactivity score for ion channel modulator activity and kinase inhibitors ranged from −0.05 to −0.2. The bioactivity scores for the nuclear receptor ligand, protease inhibitor, and enzyme inhibition were determined to be −0.13, 0.01, and −0.01 respectively.

Molecular docking of OSCC proteins

The primary therapeutic targets for OSCC that were examined in this study are TNF-α, IL-6, IL-8, MMP-9, matrix metalloproteinase-13 (MMP-13), p53, E-cadherin, and Janus kinase-1 (Jak-1). The three-dimensional conformation of the aforementioned proteins was acquired from the Protein Data Bank (PDB) using the following identification codes: 1TNF, 1ALU, 6N2U, 4XCT, 3ZXH, 3ZME, 4ZTE, and 4E4N (Fig. 4A). The present investigation involved the docking of the compound AM-158, also known as Actinosporin G, with protein targets associated with oral squamous cell carcinoma (OSCC) (Fig. 4B). The ligand conformation of AM-158 was determined based on its binding affinities with various receptor targets, including TNF-α, IL-6, IL-8, MMP-9, MMP-13, p53, E-cadherin, and Jak-1. In the present evaluation, it was shown that the compound AM-158 had the most acceptable docking confirmation, with a binding affinity of −10.9 kcal/mol towards JAK-1. The compound AM-158 primarily interacted with the amino acid residues LYS908 and GLY884 through hydrogen bonding and serves as a critical binding site of the ATP molecule (Fig. 5). The inhibition of ATP binding to the enzyme Jak-1 protein kinase hurdle the phosphorylation process. The inhibition of phosphorylation limits the binding of interferons IFNAR1-IFNAR2 heterodimers which creates the docking site of STAT proteins to simulate the JAK/STAT signaling pathway during the cancer progression (Gruber et al. 2020). Vander Waals interactions were observed between the residues VAL889, LEU1010, ALA906, LEU881, LEU959, and PHE958. The protein MMP-9, which was targeted in this study, exhibited a binding score of −10.2 kcal/mol. This strong binding was facilitated by active hydrogen bonds formed with several residues, namely LEU188, HIS226, HIS230, HIS236, TYR248, GLY186, and ALA189 (Fig. 6). The receptor protein MMP-9 utilizes the amino acid residues HIS226, HIS230, and HIS236 to bind zinc, which is essential for the functional activity of MMP-9 (Mathpal et al. 2022). The interaction between the ligand AM-158 and the FN2 domain involves zinc-binding residues, which play a crucial role in collagen binding. This interaction leads to the degradation of type IV collagen, ultimately contributing to the advancement of cancer (Quintero-Fabián et al. 2019). The ligand's docking at this specific site functions as a Zinc inhibitor, thereby reducing the progression of malignancy. Table 4 represents the binding affinities −9.5, −7.3, −7.1, −8.3, −9, and −7.4 kcal/mol of the proteins TNF-α, IL-6, IL-8, MMP-13, p53, and E-cadherin. The compound called AM-158 demonstrated significant bioactivity and had significant drug analogy attributes. The optimal docking conformation was also determined to elucidate the residues with the highest degree of interaction within the active sites of TNF-α, IL-6, IL-8, MMP-9, MMP-13, p53, E-cadherin, and Jak-1. Figure 7 illustrates the interaction between the AM-158 compound and eight therapeutic cancer targets, displaying the posture view along with their respective active pockets.

Fig. 4.

A OSCC Target Proteins: The three-dimensional crystal structures of Oral Squamous Cell Carcinoma (OSCC) protein targets with their respective Protein Data Bank (PDB) IDs. B The three-dimensional structural conformation of metabolite Actinosporin G (AM-158)

Fig. 5.

Molecular interaction of the AM-158 with JAK-1: A 3D interaction; B Binding mode of the AM-158 and Jak-1; C The 2D interaction between the ligand and amino acids residues of protein. The residues that contribute to hydrophobic interactions are shown as red arcs with spikes, and the hydrogen bonds are denoted by green dashed lines with the bond distance

Fig. 6.

Molecular interaction of the AM-158 with MMP-9: A 3D interaction; B Binding mode of the AM-158 and MMP-9; C The 2D interaction between the ligand and amino acids residues of protein. The residues that contribute to hydrophobic interactions are shown as red arcs with spikes, and the hydrogen bonds are denoted by green dashed lines with the bond distance

Table 4.

Molecular docking results of compound AM-158 (Actinosporin G) and OSCC proteins interactions

Protein	PDB ID	Residues	Ligand	Interaction type	Distance (Å)	Binding Score (kcal/mol)
TNF-α	1TNF	ARG103 ARG103 GLN102 ARG103 GLU104 PRO106 GLU104 ARG103 ARG103	O O H O O O π π π	Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Carbon Hydrogen Bond Carbon Hydrogen Bond Carbon Hydrogen Bond Pi–Anion Pi–Alkyl Pi–Alkyl	2.69317 2.40224 1.92755 3.38975 3.13335 3.79955 4.4462 4.46026 4.47579	−9.5
IL-6	1ALU	GLN176 ARG179 ARG30 LEU178 LEU178	O C π π π	Conventional Hydrogen Bond Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl	2.62581 3.62474 5.41721 5.46621 4.68316	−7.3
IL-8	6N2U	GLU46 CYS48 CYS48 GLU46 CYS48 CYS48 LYS9 TYR11 LEU47 LYS9 LEU47	O O O O π π π π π π π	Conventional Hydrogen Bond Conventional Hydrogen Bond Sulfur-X Pi–Anion Pi–Sulfur Pi–Sulfur Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl	2.55933 2.45234 1.53537 4.00982 5.36216 5.63568 4.22806 4.63721 5.37828 4.32783 4.81363	−7.1
p53	3ZME	VAL147 THR150 VAL147 CYS220 GLU221;PRO222 LEU145 VAL157 CYS220 LEU257 PRO223 VAL147 PRO151 PRO223 PRO151	O O π π C,O;N C C C C π π π π π π	Conventional Hydrogen Bond Conventional Hydrogen Bond Pi–Sigma Pi–Sulfur Amide–Pi Stacked Alkyl Alkyl Alkyl Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl Pi–Alkyl	2.76435 2.57694 3.56943 5.63373 4.91502 4.05163 5.0372 4.10281 5.05911 4.26195 3.90178 4.8844 5.32339 4.7638 5.31059	−8.3
MMP-9	4XCT	LEU188 HIS226 HIS230 HIS236 HIS236 TYR248 GLY186 ALA189 TYR179 HIS226 TYR179 HIS230 PHE192	O O O O O H H H C C π π C	Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Pi–Sigma Pi–Sigma Pi–Pi T-shaped Pi–Pi T-shaped Pi–Alkyl	2.65883 2.82472 2.42342 2.44145 3.04127 2.52673 3.00442 2.08997 3.71759 3.74981 5.3668 5.31688 4.32092	−10.2
MMP-13	3ZXH	GLY183 GLY183 GLY183 HIS226 LEU185 LEU185	O O H H π π	Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Pi–Sigma Pi–Alkyl	2.77719 2.87023 2.66196 2.12658 3.77788 5.18109	−9
E-cadherin	4ZTE	ASP100 ASP100 ASP137 ARG68 THR99 ALA72	O O H O O π	Conventional Hydrogen Bond Conventional Hydrogen Bond Conventional Hydrogen Bond Carbon Hydrogen Bond Carbon Hydrogen Bond Pi–Alkyl	2.77416 1.88624 2.35878 3.73834 3.09962 4.82374	−7.4
JAK-1	4E4N	LYS908 GLY884 VAL889 LEU1010 ALA906 LEU881 LEU959 PHE958	O O π π C C C π	Conventional Hydrogen Bond Carbon Hydrogen Bond Pi−Alkyl Pi−Sigma Alkyl Alkyl Alkyl Pi−Alkyl	3.04094 3.35709 4.83763 4.914 3.49773 5.13599 5.01985 4.83763	−10.9

Fig. 7.

Molecular interaction of the ligand AM-158 (Actinosporin G) with OSCC protein targets: 3D interaction pocket of protein receptor and AM-158 with the binding energy score and the 2D interaction plot of AM-158_receptor protein constructed using Ligplot

The Protein Connections Atlas tool was utilised to analyse the visualisation of protein connections at the atomic level. This enables the investigation of non-covalent interactions within protein structures. This technique is mostly employed to combine external data to comprehend the protein's interaction with the chosen drug. The investigation of interactions at the atomic level allows for the analysis of protein behaviour, individual subunits, interfaces, and protein–ligand connections, hence facilitating the process of drug development for the creation of novel therapeutics. The nodes of the chord plot describe the intricate network that represents the entities that ultimately represent nucleic acids or proteins. The relationships between the nodes in the chain subunits illustrate the areas where interactions take place. The asteroid plot and scatter plot of the protein–ligand complexes of RgpB, Jak-1, and MMP-9 proteins with AM-158 are constructed using Protein Contact Atlas mapping tools are depicted in Figs. S4B and S4C, respectively. The asteroid plot with the view centred on the ligand provides a visual representation of the proximity of a certain residue or ligand during an interaction. The interaction with other residues is determined by the size of each residue node. The inner circle delineates the amino acid residues that are closely linked to the ligand moiety, as depicted in Fig. S4B. The complex RgpB_AM-158 exhibits closeness between the amino acid residues F319 and E323 of the receptor. The residues H226 and L1010 in the OSCC target proteins MMP-9 and Jak-1 complexes are nearby and also engage in interactions with the ligand moiety. The scatter plot matrix provides a comprehensive overview of the network metrics associated with each residue, encompassing variables such as proximity, distance, degree centrality, and solvent-accessible area. The grid view of scatter plots displays the plot of each pair of matrices, as depicted in Fig. S4B.

The diagrams of the docked protein complexes demonstrate the protein interactions at the secondary level (Fig. 6A).

Molecular dynamic simulation

The present investigation focused on investigating the binding modalities of the most suitable docked compound, AM-158, against protein targets: the bacterial protein RgpB and the target proteins MMP-9 and Jak-1 associated with OSCC. A molecular dynamics simulation was conducted in an aqueous system, with simulation duration of 100 ns. The stability of the protein–ligand combination was assessed by conducting a molecular dynamic simulation, which involved analyzing the atomic motion of the complex. The stability of the protein targets-AM-158 complexes was assessed by examining the geometric characteristics of the complexes, including the root mean square deviation (RMSD), the radius of gyration (Rg), and root mean square fluctuation (RMSF). The RMSD is utilized to assess the accuracy of atomic position measurements to establish the mean distance between protein and ligand structures that have been superimposed. The calculated RMSD values for the Cα atoms of RgpB and AM-158 in the complex were determined to be 1.3788 ± 0.1634 Å and 4.538 ± 0.2338 Å, respectively. The stability of the protein–ligand complex found throughout the simulation, however, is consistent from 60 ns. Although mild deviations occurred in the stability of the protein–ligand complex around 50 and 80 ns of the simulation run, the interaction of compound AM-158 is consistent (Fig. 8A). The control drug Dequalinium exhibited RMSD values of about 1.7876 ± 0.134 Å. On comparing with our compound AM-158 targeted towards the virulence protein RgpB, the control drug showed stability only after 60 ns of simulation (Fig. S5). The RMSD values of the Cα atoms in the Jak-1 and AM-158 proteins are 1.180 ± 0.3192 Å and 1.610 ± 0.3438 Å, respectively (Fig. 9A). RMSD graph showed higher fluctuations between the ligand and protein from 30 to 70 ns. However, consistent RMSD was obtained from 70 ns to the end of the simulation. The average RMSD for the Cα atoms of MMP-9 and the compound AM-158 in the OSCC target proteins was found to be 2.499 ± 0.3916 Å and 2.0676 ± 0.2134 Å, respectively (Fig. 10A). The term "Rg" can be defined as the root mean square distance between the centre of mass of each protein atom and its corresponding protein atom in a given system, as described by Lobanov et al. (2008). The Rg plot of ligand assessed the capacity, morphology, and folding dynamics at each time step of the trajectory during the simulation. The ligand AM-158 exhibits a consistently acceptable range of variations in its Rg values when bound to RgpB, MMP-9, and Jak-1, varying from 4.5 to 5.4 Å, 4.73 to 5.45 Å, and 4.54 to 5.34 Å, respectively (Fig. S2).

Fig. 8.

Molecular dynamics simulation trajectories (150 ns) analysis for the potential docked complex of Actinosporin G and RgpB A RMSD profile, B RMSF fingerprint, C, D protein–ligand interactions of the complex

Fig. 9.

Molecular dynamics simulation trajectories (150 ns) analysis for the potential docked complex of Actinosporin G and Jak-1 A RMSD profile, B RMSF fingerprint, C, D protein–ligand interactions of the complex

Fig. 10.

Molecular dynamics simulation trajectories (150 ns) analysis for the potential docked complex of Actinosporin G and MMP-9 A RMSD profile, B RMSF fingerprint, C, D protein–ligand interactions of the complex

The use of RMSF analysis allows for the examination and assessment of local conformational alterations within both protein chains and ligand molecules. The RMSD concerning the central axis of rotation were calculated to produce the RMSF. The variations produced by residue index Cα were utilized to assess the local structural fluctuations of the proteins RgpB, MMP-9, and Jak-1 in the presence of AM-158. It has been seen that the RgpB_AM-158 complex system shows an approximate RMSF value of 0.8–2.8 Å with the least fluctuation in the region of 150–170, 200, 280–290, and 325 regions (Fig. 8B). The complex RgpB_Dequalinium exhibited an approximate RMSF value of about 0.5 to 3.5 Å. Similarly, the complex system Jak-1_AM-158 shows a fluctuation in the region of 10–25, 40–58, 82–90, 130–140 and 145–150 with fluctuations within the range of 0.8 to 5.8 Å. The maximum fluctuations were found to be in the region of 40–60 residues index (Fig. 9B). According to the RMSF analysis, AM-158_MMP-9 complexes exhibited the most notable oscillations within the range of 0.5–3.0 Å (Fig. 10B). In comparison, the average RMSF data indicated a higher level of interactions between the AM-158 and the receptor protein MMP-9 among the target proteins of OSCC with lower fluctuations in the region of 70–75, and 110–130 residues. Therefore, the examination of RMSF and RMSD values for each protein–ligand complex provided evidence for the collective screened potential of compound (AM-158) Actinosporin G in targeting therapeutic indicators associated with gingivitis-related OSCC. The protein–ligand interaction is facilitated by several types of bonding, including hydrogen bonding, ionic bonding, hydrophobic bonding, and van der Waals forces. These interactions are crucial in the rational design of molecules with drug-like properties. Protein–ligand interactions were observed in all the targeted proteins, and a specific compound, AM-158 from the molecular dynamics (MD) trajectories. The default parameters of the Desmond module were employed for this analysis. The compound AM-158 exhibited significant interaction with a majority of the protein amino acid residues. The results of the simulation runs demonstrated a consistent number of hydrogen bonds formed, which aligns with the findings of the molecular docking study. The creation and breakage of hydrogen bonds is the leading segment of the molecular dynamic simulation. The binding affinity increases when there is a greater H-bond between the ligand and protein. The complex RgpB_AM-158 system exhibited a higher number of hydrogen bond interactions with the residue ASP159 followed by TRP284, TRP283, SER156, HIS211, ALA285 and LYS324. A higher H-bonding interaction of 0.8 nm was found for the residue ASP1021 in the Jak1_AM-158 complex. For the complex MMP-9_AM-158, the residue GLU227 exhibited a hydrogen bond of about 1 nm (Figs. 8C, 9C, and 10C). The present study showcases the ligand atom interactions with the protein residues by depicting the interactions that transpire for duration above 30% of the simulation time in the trajectory of 150 ns. The complex RgpB_AM-158 system shows a hydrogen bond with ASN159 residue which was observed in 41% of trajectories, while the complex RgpB_Dequalinium system shows hydrogen bonds with varied residues with less than 30% duration of the simulation. Similarly, the complex Jak1_AM-158 system shows a hydrogen bond with SER963 in 39% of the trajectory and ASP1021 residue which was reported to be 45% and 39%. The system MMP-9_AM-158 complex shows a hydrogen bond with HIS226 at 54%, GLU227 at 77% and LEU188 at 41% respectively. These interactions are visually represented in Figs. 8D, 9D, and 10D, providing a comprehensive illustration of the precise interplay between the ligand atoms and the protein residues throughout the 150 ns molecular dynamic simulation.

The binding free energy value is a pivotal parameter for evaluating the intensity of the interaction between the ligand and protein. A lower value signifies a more advantageous alignment during docking and a more robust interaction between molecules. A negative number signifies that the free energy of the protein–ligand complex is lower than that of the individual components in a solution, indicating enhanced stability and a greater probability of natural existence. The results revealed that the MM-GBSA binding energy, measured using fastDRH, between the receptor RgpB and AM_158 was −31.62 kcal/mol. The Jak1-AM_158 and MMP9-AM_158 complexes exhibited decreased binding free energy values of −51.55 and −36.12 kcal/mol, respectively. The Jak1-AM_158 complex displayed the most negative value, whereas the MMP9-AM_158 complex showed the least negative value (Figure S3). The negative binding energy serves as an indicator of the ligands’ affinity towards the target, wherein larger values signify a more pronounced attraction. The protein structures exhibited a significant affinity for the ligand AM-158, as evidenced by the findings derived from a meticulous investigation of the binding free energy. The analysis of hotspot residue prediction, which determines the residues most likely to interact with ligands for the complex RgpB_AM-158, revealed that amino acid residues ALA149, PRO155, and VAL169 had higher negative energy, whereas GLY887, ARG893 & CYS944 and PHE181 and HIS175 are shown in OSCC target proteins JAK-1 and MMP-9. Protein hotspot residue predictions, which are obtained by examining several docking poses, are essential for ligand binding.

Discussion

The therapy and management of OSCC is a significant concern due to its frequent diagnosis in advanced stages, which is associated with an increased likelihood of metastasis (Li et al. 2022). The development of OSCC entails a multifaceted series of events that are governed by genetic control influenced by many carcinogenic factors (Li et al. 2020). Numerous studies have comprehensively documented the robust correlation between P. gingivalis and OSCC (Chattopadhyay et al. 2019). The molecular mechanism by which P. gingivalis promotes the progression of OSCC involves several key processes, including cellular invasion and proliferation, chronic inflammation, inhibition of cell apoptosis, and suppression of immune response (Zhang et al. 2019). These effects are mediated by the presence of various virulence factors, such as gingipains R and K (cysteine proteases), fimbriae, and outer membrane surface antigens. The gingipains RgpB and Kgp play a significant role in the degradation of the host extracellular cellular matrix, hence facilitating efficient oral invasion (Kadowaki 2021; Bostanci and Belibasakis 2012). The FimA1 proteins, which are characterized by their elongated structure, are known to bind to β1 integrins located on the surface of cells, facilitating adherence (Cai et al. 2019; Enersen et al. 2013). On the other hand, the Mfa1 and Mfa3 proteins, which are shorter in length, play a vital role in enhancing the adhesive and invasive capacities observed in periodontal illnesses (Lee et al; 2018). These virulence factors induce the production of different cytokines, including IL-6, IL-8, and TNF-α, which in turn promote cell proliferation and apoptotic progression, leading to the development of malignancies (Chattopadhyay et al. 2019). The present study used computational-based high-throughput virtual screening to identify novel therapeutic compounds that may aid in the development of medications targeting OSCC. Singh et al (2023) conducted a recent study that focused on the in silico screening of natural compounds against the P. gingivalis FimA1 pathogenicity protein about OSCC. The study revealed that the natural compound protocatechuic acid showed significant molecular docking binding energy (Singh et al. 2023).

The present study aimed to investigate the binding affinity between a collection of 179 Actinobacterial-derived bioactive secondary metabolites and specific proteins, namely RgpB, Kgp, RagB, Mfa1, Mfa3, and FimA1, found in P. gingivalis. The in-house ligand library containing 179 compounds screened against Kgp receptor protein displayed the compound AM-102 referred to as 2-[[5-(furan-2-yl)-4-(2-methyl phenyl)-1,2,4 triazol-3-yl]sulfanylmethyl]-N-[[3-(trifluoromethyl)phenyl]methyl]-1,3-thiazole-4-carboxamide involved in cell cycle arrest in colorectal cancer along with the other metabolites isolated from Streptomyces sp. 801 (Kouroshnia et al. 2022).

Actinokineospora spheciospongiae sp. nov. (DSM 45935T) produces Fridamycin I (AM-179), which inhibits growth of Trypanosoma brucei strain TC221 (Tawfike et al; 2019). Actinokineospora spheciospongiae sp. nov. (DSM 45935T) also synthesizes Actinosporin C (AM-157) (Tawfike et al. 2019) and Streptomyces sp. SCSIO 03032 synthesizes Heronamide F (AM-154) (Zhang et al. 2014). Heronamide F (AM-154) which is isolated from Streptomyces marincola showed antimicrobial activity (Shi et al. 2022). In our study, RagB interacted with Fridamycin I (AM-179) having an affinity score of −8.6 kcal/mol. The short fimbrial associated proteins Mfa1 docked with Actinosporin C (AM-157) with a score of −8.5 kcal/mol and Mfa3 docked with Heronamide F (AM-154) with a score of −7.3 kcal/mol. The compound AM-67 1, 7, 9, 11-tetrahydroxy-8, 13-dioxo-3-(2-isopropyl)-5, 6-dihydrobenzo[a]tetracene-2-carboxylic acid which is isolated from Streptomyces sp. showed antibacterial activity against Staphylococcus aureus MRSA strains (Fatima et al. 2019). AM-67 interacted with fimbriae protein FimA1 with a binding score of −7.9 kcal/mol in our study. From the in-house ligand library, the compound AM-158, referred to as Actinosporin G isolated from the Actinokineospora spheciospongiae sp. nov. found to be a common bioactive compound interacting with a major of virulent proteins with a higher binding affinity (negative) score. For instance, it exhibited −7.9 kcal/mol binding score to RgpB, −8.7 kcal/mol to Kgp, −7.5 kcal/mol to RagB, −7.1 kcal/mol to Mfa1, −8.2 kcal/mol to Mfa3 and −7.5 kcal/mol to FimA1 which was higher affinity score than the control drug molecule Dequalinium (Shi et al. 2022) (Supplementary Table S2). The ligand AM-158 attaches to the active site of the catalytic domain of the RgpB protein, which is responsible for the degradation of periodontal tissues and likely, plays a role in evading the host's defense mechanism (de Diego et al. 2014). The interactions between AM-158 and the amino acid residues in the catalytic domain result in the reduction of functional proteolysis activity, which in turn suppresses the pathogenicity of P. gingivalis. Furthermore, it should be noted that this particular domain plays a crucial role in facilitating the maturation process of fimbrial subunits enhancing the bacterial adhesion towards host epithelial tissues (Ciaston et al. 2022). Consequently, the binding of the drug AM-158 to this domain has a synergistic effect, serving as an inhibitory agent against bacterial infections.

Actinosporin G was selected because it obeys Lipsinki's rule of five (Ro5) based on the criteria that the ligand's molecular weight is less than 450.44, it has fewer 8 hydrogen bond acceptors, 4 hydrogen bond donors, and logP value of 1.81, all of which fall within acceptable ranges to be a potential drug candidate. Additionally, it is noteworthy to add that the pharmacokinetic characteristics of the compound, such as its oral absorption in humans, partition coefficient, and water solubility (as shown by QPlogS values), were observed to align with Lipinski's Rule of Five (Ro5). The Drug-likeness of the compound AM-158, Actinospoin G is depicted in Fig. S6. Before commencing clinical trials for possible pharmaceutical candidates, it is crucial to conduct toxicity evaluation using in-silico methodologies. The increasing popularity of computational techniques for toxicity evaluation can be attributed to their notable attributes, including high accuracy, rapidity, and ease of access. This is due to their ability to facilitate the collection of valuable insights about a diverse range of chemical compounds. Nevertheless, it is important to bear in mind that AM-158 exhibits immunotoxicity as well as active toxicity towards MMPs. To enhance therapeutic results, the ligand AM-158, alternatively referred to as Actinosporin-G, was investigated as a prospective therapeutic intervention for several proteins linked to OSCC.

The interaction and general molecular stability of the protein–ligand complex were demonstrated by the highest negative binding affinity score, RMSD, RMSF, and Rg plot in the molecular docking and MD simulation experiments between the AM-158, Actinosporin -G, and RgpB complex. Throughout the 150 ns simulation duration, the receptor protein RgpB's GLU258 and LYS228 protein residues showed 35 and 38% interaction, respectively. The interaction of P. gingivalis gingipains, which increases the levels of host-defensins, results in enhanced cellular proliferation during malignancy (Hoppe et al. 2016). These gingipains stimulate β-catenin, which then moves to the nucleus, binds with TCF/LEF, and induces IL-6 and cell cylin-D1, all of which lead to cell proliferation (Binder Gallimidi et al. 2015). They also reduce the level of p53 activity, which in turn encourages the proliferation of gingival epithelial cells through the activation of the PI3K pathway (Cho et al. 2014). Gingipains induced the expression of MMP9 and PAR receptors through activation of p38MAPK and ERK1/2 signalling pathways (Inaba et al. 2016). P. gingivalis also activated the JAK pathway to increase the levels of IL-1 and IL-6 and the MAPK and NF-B pathways to produce the cytokine IL-8 (Fujita et al. 2014). These inflammatory mediators cause oncogene activation and angiogenesis, which may lead to immunosuppression and the growth of OSCC cells (Karpiński 2019).

Actinosporin G was molecularly docked with the OSCC target receptor proteins TNF-α, IL-6, IL-8, p53, MMP-9, MMP-13, E-cadherin, and Jak1 in our investigation. The bioactive compound AM-158's molecular docking interaction with MMP-9 and Jak1 receptor proteins demonstrated binding scores of −10.2 and −10.9 kcal/mol with exhaustiveness 8 and an average RMSD of zero, in contrast to the control drug dequalinium's binding scores of −7.4 and −8.9 and −8.9 kcal/mol, respectively. Dequalinium is a quaternary ammonium chloride chemical that has been utilized as an over-the-counter throat lozenge for oral and throat infections as well as vaginal candidiasis. Additionally, it has anti-parasitic, anti-viral, anti-cancer, anti-fungal, and neuroprotective effects (Bailly 2021). Many bonding processes, including hydrogen bonding, ionic bonding, water bridges, and hydrophobic bonding, aid in protein–ligand interaction. These interactions are critical in the rational design of compounds with improved therapeutic effectiveness. The Desmond module's molecular trajectories were used to analyze protein–ligand interactions involving MMP-9_Actinosporin G and Jak1_Actinosporin G. In this investigation, the default settings were used. Throughout the 150 ns simulation period, there was considerable interaction between the protein residues GLU227 of MMP-9 (57%) and ASP1021 of Jak1 (55%). Furthermore, Actinosporin G exhibited considerable intermolecular interaction. These findings show that Actinosporin G can specifically target proteins involved in cellular invasion and apoptosis during the evolution of OSCC. Our findings emphasize the significance of secondary metabolites produced by Actinobacterial genera as a promising resource for the development of natural pharmaceutical alternatives with low side effects.

Conclusion

This in-silico study suggests a treatment intervention for this target Porphyromonas gingivalis virulence proteins and OSCC regulatory proteins. The ADMET results revealed that Actinosporin G has a high LD50 value (3000 mg/kg) and does not cause toxicity. According to the findings of this investigation, Actinosporin G has the inhibitor's potential against the virulence proteins and tumor cell proteins. To the best of our knowledge, our study is the foremost in reporting the Actinobacterial-based secondary metabolites as therapeutical agents against P. gingivalis-associated OSCC. Our structure-based drug design strategy yields natural chemotherapeutic drug compounds that may be exploited to generate strong medicines for the treatment of P. gingivalis-related OSCC. Before the clinical trial, more in vitro and in vivo testing is needed to evaluate the potential of Actionosporin G as a therapeutic medication for oral cancer treatment.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Methodology/ Formal analysis/Writing: ZT and Conceptualization/ editing/ Supervision: IC.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Conflict of interest

No potential conflict of interest was reported by the author(s).