Exploring Withania somnifera derived natural products as promising inhibitors of Mycobacterium tuberculosis Pantothenate Kinase-PanK: An integrated in silico and in vitro approach

Abstract

Background

Tuberculosis remains a pervasive and enduring global health challenge, with the alarming rise of drug-resistant variants. Mycobacterium tuberculosis (M. tuberculosis), bacterium responsible for tuberculosis, deploys a complex arsenal of virulence factors to evade the host’s immune defences. The quest for novel targets or compounds to combat drug-resistant M. tuberculosis strains is of paramount importance. PanK is an essential enzyme for Co-enzyme A (CoA) biosynthesis pathway, targeting inhibition of its activity by Withania somnifera phytochemicals may provide an effective therapeutic strategy against resistant strains.

Objective

The study aims to identify the potential of natural compounds derived from Withania somnifera as inhibitors of the PanK enzyme (novel target) in M. tuberculosis.

Methodology

In silico computational approach, includes steps-structure based virtual screening of 83 Withania compounds followed by molecular docking and dynamic simulations spanning 100 ns, to assess the binding affinity and stability between screen key compounds and PanK. In vitro anti-tuberculosis bioassays was also performed to validate the In silico experiments finding.

Result

Through in silico experiments, four key compounds of Withania somnifera were —Morkotin A, Rutin, Withaoxylactone, and 2,3-Dihydrowithanolide E were identified. They exhibited strong potential to inhibit PanK enzyme activity. The In silico as well as In vitro findings suggest that Withania somnifera-derived natural compounds could serve as effective candidates for targeting vital enzymes in M. tuberculosis.

Conclusion

Withania somnifera can be explored as valuable resource for developing novel drugs for PanK as a target to combat tuberculosis.

1. Introduction

Tuberculosis (TB) remains an important global health problem that stems from infection by Mycobacterium tuberculosis (M. tuberculosis) known as a difficult infectious disease. The global tuberculosis diagnosis number reached 10.6 million people during 2021 while showing a 4.5 % increase from the previous year. This disease resulted in the death of 1.6 million people according to WHO Report (2022) [1]. India, in particular, witnessed a 28 % surge in TB cases, reporting 1.9 million incidents in 2021 [2]. Extensively drug-resistant (XDR) and multidrug-resistant (MDR) strains of M. tuberculosis have become a serious threat to worldwide human health because of their emergence [3]. Drug-resistant variants of M. tuberculosis require extended treatments periods and expensive medications that produce worse side effects. The worldwide increase in drug-resistant tuberculosis (DR-TB) persists because, Rifampicin-resistant tuberculosis (RR-TB) will affect an additional 4, 50,000 people in 2021. Treatment of MDR-TB requires expensive second-line drugs although XDR-TB usually receives a label of incurable infection [4]. The development of new anti-TB drugs remains insufficient to handle this critical health problem. M. tuberculosis strains that develop resistance to currently utilized anti-mycobacterial agents are widely spreading which demonstrates the urgent need to develop new drug candidates for tuberculosis control.

The fight against drug-resistant TB requires immediate development of new therapeutic medicines. The de novo pyrimidine biosynthesis pathway represents a promising research point because it serves as the essential metabolic pathway that produces vital building blocks for cell growth. The performance of de novo pyrimidine biosynthesis holds promising potential for developing next-generation anti-TB medications based on research conducted by researchers in 2022 [5]. Within this context, Pantothenate Kinases (PanK), enzymes encoded by the panK gene, emerge as newly identified drug targets. PanKs assume a crucial role in catalyzing the rate-limiting step of the Coenzyme A (CoA) biosynthesis pathway, a mechanism wherein Pantothenate (Vitamin B5) is transformed into 4′-phosphopantothenate through the utilization of ATP as a cofactor, as elucidated by a study [6] and earlier research by Jackowski and Rock in 1981. Coenzyme A serves as an essential cofactor which acts by properly controlling different metabolic enzymes throughout various cellular pathways while influencing both lipid synthesis and degradation processes. Two PanKs occur in M. tuberculosis genome through coaA (type I) and coaX (type III) genes while researchers confirmed CoaA functions encoded by coaA gene as essential for bacterial in vitro and in vivo growth [6]. Scientists have evaluated different approaches to find new TB medications through investigation of CoA intracellular levels reduction. They have adopted inhibitor substances and gene disruption methods to inhibit enzymes responsible for CoA biosynthetic pathway production [7,8]. These studies demonstrated that blocking CoA synthesis represents a promising method to create new drug target against tuberculosis. Research experts have demonstrated how phytochemicals show therapeutic value for treating TB infections through multiple approaches that enhance current anti-infectious disease strategies.

Human history have always turned to medicinal plants for controlling infectious medical conditions throughout human history. Various plants have served as essential sources to create new pharmaceutical compounds. Indian medical systems Yoga, Homeopathy, Unani, Naturopathy and Siddha along with Ayurveda depend on plant medicines for their therapeutic practices to treat multiple health problems [9]. Natural drug manufacturing rests in plant derivatives, as demonstrated by WHO research, which shows that they constitute 30 % of pharmaceutical products while half of the planet uses plant-based medicines to maintain wellness. Medicinal plants represent excellent resources of phytochemicals that display remarkably diverse structural features while pursuing bacterial infection prevention. The chemical agents extracted from plants demonstrate the ability to suppress disease recurrence by delivering promising tuberculosis medication [10,11]. Numerous studies have highlighted the antibacterial properties of various medicinal plants and their natural constituents, often employed in tuberculosis treatment. One such noteworthy plant is Ashwagandha, scientifically known as Withania somnifera (L) Dunal, a member of the Solanaceae family. Ashwagandha holds a significant place in Ayurveda due to its therapeutic attributes [12]. It boasts a diverse array of phytochemicals, including lactones, sapogenins, and withaferin A [13], alongside a variety of alternative molecular entities [14]. These compounds showcase a broad spectrum of pharmacological properties, encompassing anti-diabetic, anti-inflammatory, anti-microbial, anti-stress, anti-tumour, cardio-protective, and neuro-protective effects [15]. Consequently, the primary objective of the present study is to delve into the therapeutic potential of the phytochemicals found in Withania somnifera concerning their ability to combat M. tuberculosis infection. This investigation employs an array of molecular simulation techniques, to explore the inhibitory effects of these phytochemicals on the M. tuberculosis-PanK enzyme.

2. Material and methods

2.1. Receptor and ligand optimization

The target protein structure (M. tuberculosis’s Pantothenate Kinase) (PanK) was manually prepared as a receptor for molecular docking. The initial structure, obtained from the Research Collaboratory for Structural Bioinformatics Protein Data Bank (PDB ID: 4BFW) with a resolution of 2.27 Å [6], method: X-ray diffraction, subjected to several processing steps. First, we began by eliminating all heteroatoms from the structure. Subsequently, we assigned bond orders and introduced polar hydrogen atoms to the protein. This process was carried out following the default configuration within the Protein Optimization Wizard, using Maestro Suite (Schrödinger Release 2018-3). Following this step, the protein’s residues were protonated using the PROPKA Program at a pH of 7.0. The final stage involved energy minimization, which was performed using force (OPLS3e) field, employing default commands. This comprehensive processing ensured that the protein structure was ready for subsequent molecular docking studies.

To identify potential phytochemical inhibitors of PanK derived from Withania somnifera, we assembled a library of 83 phytochemicals through an extensive literature review (supplementary file, Table S1). Subsequently, the 3D conformations of these phytochemicals were obtained from the PubChem database in SDF format and treated as ligands for subsequent computational analysis. For the preparation of the ligands, In the course of our research, we employed the LigPrep module derived from the Maestro suite, specifically the Schrödinger software package version 2018-3. LigPrep was utilized with default parameters to generate tautomeric conformations of the ligands. These conformations were generated using the EPIK module at a pH of 7 ± 2, and the OPLS3e force field was applied for energy minimization [16,17]. To facilitate virtual screening, a receptor grid was prepared, focusing on a 12 Å region around the active site of PanK. This region was selected to encompass key active site residues, including Asp129, His179, Tyr182, Leu203, Tyr235, Arg238, Phe239, Met242, Phe254, Tyr257, Ile276, and Asn277, ensuring that all critical interactions could be captured during the screening process The generation of the receptor grid was accomplished through the utilization of the Receptor Grid Generation panel within the Schrödinger software package version 2018-3, utilizing minimized protein receptor as a basis [17]. This meticulous process ensured the preparation of ligands and the receptor grid for subsequent computational analysis in our quest to identify potential PanK inhibitors among the Ashwagandha-derived phytochemicals.

2.2. Performing computational screening integrated with ADMET evaluation

To isolate the most promising compounds with strong binding potential, we performed a default parameter Glide XP computational screening in the Schrödinger suite [18]. This procedure allowed us to prioritize compounds based on their negative binding scores, indicative of favorable binding interactions with the target protein. Subsequently, the compounds with the most favorable binding scores were subjected to further evaluation, specifically concerning drug-likeness and essential pharmacokinetic characteristics. To perform these assessments, we employed the online server SwissADME, accessible at http://www.swissadme.ch, which integrates valuable insights into the pharmacokinetic characteristics of the selected compounds [19,20].

2.3. Recomputing docking positions and analyzing intermolecular interactions

To refine our selection process, the most promising compounds from Withania somnifera against PanK were subjected to a redocking procedure. This involved redocking these leading compounds, alongside reference compound (ZVW) (Pubchem ID: 71307936), into their respective binding pockets within PanK. The redocking procedure was conducted with the XP Glide module in the Schrödinger suite, following default parameters [21]; Schrödinger Release 2018-3), ensuring the accuracy and reliability of the molecular interactions and binding conformations. This step aimed to validate and further assess the affinities of the chosen compounds for binding, thus ensuring that their predicted interactions with PanK were robust and consistent with our expectations. This iterative process enhances our confidence in the potential of these compounds as inhibitors of PanK and contributes to the rigor of our drug discovery efforts.

2.4. MD simulations

MDS (Molecular dynamics simulations) were performed to assess the stability (equilibrium) of the protein-ligand complex involving the predicted potent compound and the PanK protein [22,23]. The simulations were executed using the Desmond version 4.4 plugin, an integral component of the Schrödinger-Maestro version 11.7 software suite (Schrödinger software package version 2018-3). These molecular dynamics simulations were conducted using the Desmond MD System, developed by D. E. Shaw Research in New York, NY, in 2018, was employed for this purpose. To facilitate the preparation of the simulation system, we utilized the system builder module within Desmond. All systems were engaged for the solvation by situating it in an orthorhombic box (10 Å) with the TIP4P model. To maintain neutrality, the entire system was neutralized by introducing Na + or Cl-counterions, and further minimization was carried out using default settings. Finally, using a Nose Hoover thermostat and the Martyna Tobias Klein method the entire system was simulated with temperature (300K) and a pressure (1.01325 bar) under default parameters [24]. Each protein-ligand complex system and control protein-ligand complex system were processed for 100 ns MD simulation. Default settings were considered to collect 2000 frames. In the concluding phase, we analyzed the obtained trajectories through the employment of Desmond’s Simulation Interaction Diagram (SID) module. This analysis served as a pivotal step in our extended molecular dynamics (MD) simulations, allowing for comprehensive investigation of the atomic interactions and dynamics observed during the course of our study.

2.5. Binding affinity computations

The simulated complexes' net binding free energy was computed utilizing the MM/PBSA method. This computational method was executed employing the g_mmpbsa tool for the analysis within the GROMACS software package, as described by Refs. [25,26]. To perform MM-PBSA calculations, we implemented a singular trajectory protocol technique, wherein we simulated distinct conformations of the protein-ligand complex, adhering to the methodology outlined by Kumar et al. (2020). The calculation of the free binding energy between the ligand complex and PanK was performed using coordinates extracted at 10-ps intervals from the concluding 10 ns of the molecular dynamics simulation runs.

The equation calculated ΔG_bind for each protein-ligand complex:

ΔG_bind = ⟨G_complex⟩ - ⟨G_Protein⟩ - ⟨G_Ligand⟩

here, ΔG_bind is the free binding energy, where G complex is the protein-ligand complex’s free energy. G_Protein and G_Ligand are the free energies of the protein and ligand.

2.6. In vitro antituberculosis bioassays

In this investigation, a meticulously characterized strain of M. tuberculosis, distinguished by its slow growth rate and avirulent nature, was employed for the purpose of drug susceptibility studies. The experimental protocol involved the utilization of leaf extracts derived from Withania somnifera cultivated under distinct temperature conditions, namely elevated temperature (ET) and ambient temperature (AT), Rutin and Morkotin A pure form extract as the test drug. Isoniazid (INH) served as the standard drug for comparative analysis.

2.7. Bacterial culture

The H37Ra strain of M. tuberculosis designated for infection was cultivated to mid-logarithmic phase in 7H9 medium (Middlebrook) supplemented with 10 % OADC (Ovalbumin, Dextrose, and Catalase), 0.05 % Tween-80, and 0.2 % glycerol. The preparation of the 7H9 medium involves a meticulous process of crafting a nutrient-rich broth, which is then accurately distributed into individual centrifuge tubes. Stringent measures are implemented to ensure the stability of the media and prevent contamination by storing it at 4 °C. The subsequent thawing of the H37Ra stock, preserved at −80 °C, ensures the viability of the bacterial strain. Under aseptic conditions, 500 μl of the H37Ra stock is precisely introduced into each centrifuge tube containing the prepared 7H9 medium. The incubation phase follows at 37 °C, optimizing conditions for H37Ra growth, and careful monitoring continues until an optical density (OD) of 0.6 is achieved. Cultures are cryopreserved in 20 % glycerol and stored at −80 °C until utilized for infection.

2.8. Microplate alamar blue assay (MABA)

The calculation of MIC (Minimum Inhibitory Concentrations) against M. tuberculosis-H37Ra was conducted through the Alamar Blue assay. This assay utilizes the inherent reducing capability of viable cells to convert resazurin into the fluorescent compound resorufin. To mitigate evaporation within inner wells, 200 μl of deionized water was added to the outer perimeter of 96-well flat-bottom plates.

Various concentrations starting from 2400 μg/ml of the test compounds, including rutin, Morkotin A, and a plant leaf extract, along with isoniazid (INH) as a positive control, were serially diluted in dimethyl sulfoxide (DMSO) and dispensed into a 96-well plate. Control wells contained drugs and cells only. Subsequently, 50 μl of the bacterial culture was added to all wells, excluding control wells. After sealing the plates with parafilm, they were incubated for six days at 37 °C.

After the 6-day incubation period, Alamar Blue reagents were introduced to all wells at a 1:10 ratio (20 μl of Alamar Blue dye in 200 μl of media) and re-incubated for an additional 24 h at 37 °C. The color conversion in each well was then recorded. The absence of a color change (blue) indicated no growth, while a color shift to pink signified bacterial growth. The MIC was represented as the minimal (lowest concentration) of a compound that resulted in 100 % inhibition of bacterial growth, denoted by the absence of a color change from blue to pink.

3. Results

3.1. SBVS analysis

To predict best interaction mode between ligands and target protein to form a stable complex using scoring, we perform SBVS (Structure-Based Virtual Screening) [27]. The principal aim of this study is to pinpoint natural compounds with the potential to inhibit the M. tuberculosis-Pank protein effectively and, thereby, potentially used for the treatment of M. tuberculosis infection. To achieve this, 83 phytochemicals of Withania somnifera subjected to computational screening for the M. tuberculosis PanK protein resulted in docking scores spanning from −13.28 to −1.88 kcal/mol (Table S2, Supplementary File). According to the docking scores obtained, the following four top compounds, Morkotin A (CID- 10190763), 2,3-Dihydrowithanolide E (CID- 131751517), Withaoxylactone (CID- 101687981), and Rutin (CID- 5280805) (Fig. 1), were recognized as specific inhibitors due to their substantial docking scores: −13.2 kcal/mol, −10.4 kcal/mol, −9.01 kcal/mol, and −8.8 kcal/mol, respectively, were identified as specific inhibitors and, consequently, selected for further redocking and an analysis of intermolecular interactions. Notably, the chosen phytochemicals exhibited acceptable docking energy score in compassion to reference inhibitor (ZVW) i.e. −8.5 kcal/mol of M. tuberculosis-PanK protein.

Fig. 1.

The 2D structural representations of the chosen prospective compounds, (a) Morkotin A (b) Rutin, (c) Withaoxylactone, (d) 2, 3-Dihydrowithanolide E.

3.2. Molecular interaction analysis via Re-docking

After completing the SBVS, it becomes imperative to carry out a Re-docking analysis to validate the affinity of the compounds identified during the virtual screening process with the active binding site residues. This is crucial because SBVS is a faster method but is somewhat less precise in its predictions [28]. Therefore, we opted for a stringent XP docking approach in the redocking process to ensure that the chosen ligands effectively bind with the active site residues of M. tuberculosis PanK. In this context, the re-docked complex exhibited significantly reduced binding energies when compared to the reference inhibitors (ZVW).

Intermolecular interaction analysis was employed to ascertain and identify the characterization of molecular contacts establishment among the docked ligands (2, 3-Dihydrowithanolide E, Rutin, Withaoxylactone and Morkotin A) and target protein i.e., M. tuberculosis-PanK. Herein, each docked complex (2, 3-Dihydrowithanolide E, Rutin, Withaoxylactone and Morkotin A) (Fig. 2 and Table 1) was observed for hydrophobic bonds, hydrogen bond formation (H-bond), and pi-pi interactions of amino acid residues of M. tuberculosis-PanK (target protein). In details, the docked PanK- Morkotin A complex, three hydrogen bonds (MET-144, TYR-182 and ARG-238 residues) were noted to have formed, along with one pi-pi stacking (TYR-182 residue) and one pi-cation interaction (LYS-147). Additionally hydrophobic (VAL-99, ALA-100, LEU-132, MET-144, TYR-153, TYR-177, TYR-182, LEU-203, TYR-235, PHE-239, MET-242, ALA-246, PHE-247, PHE-254, TYR-257 ILE-272, and ILE-276 residues), polar (HIE-145, HIS-253 and ASN-277 residues), negative (GLU-42 and ASP-129 residues), positive (LYS-147, HIP-179 and ARG-238 residues), and glycine (GLY-148 residue) interactions were providing equilibrium (Fig. 2a and b) to the docked complex. The PanK-Rutin (protein-ligand complex) was noted for formation of three hydrogen bonds (HIP-179, ARG-238 and THR-245 residue), in addition to complementary interactions with significant residues like hydrophobic (VAL-99, ALA-100, LEU-132, TYR-153, TYR-177, TYR-182, LEU-203, TYR-235, PHE-239, MET-242, ALA-246, PHE-247, PHE-254, TYR-257, ILE-272 and ILE-276 residues), pi-pi stacking (TYR-182), polar (THR-245 and ASN-277 residues), negative (GLU-42 and ASP-129 residues), positive (LYS-147, HIP-179 and ARG-238 residues), pi-cation (LYS-147 residue) and glycine (GLY-148) interactions (Fig. 2c and d).

Fig. 2.

3D and 2D interaction diagram of compounds; (a–b). Morkotin A, (c–d). Rutin, (e–f). Withaoxylactone (g–h) 2, 3-Dihydrowithanolide E. The docked complexes exhibit interactions with their respective ligands in the M. tuberculosis-PanK protein’s binding pocket within a 4 Å radius.

Table 1.

Selected phytochemical inhibitors targeting the PanK protein and their molecular interactions.

S. No.	Drug	Docking Score (kcal/mol)	H-Bond	Pi-Pi stacking	Hydrophobic	Polar	Negative	Positive	Glycine	Pi-cation
1.	Morkotin A	−13.2	MET-144, TYR-182, ARG-238	TYR-182	MET-144, ILE-276, LEU-203, TYR-182, LEU-132, TYR-153, TYR-177, ALA-100, VAL-99, MET-242, PHE-239, ALA-246, PHE-247, TYR-257, TYR-235, ILE-272, PHE-254	HIE-145, HIS-253, ASN-277	GLU-42, ASP-129	LYS-147, HIP-179, ARG-238	GLY-148	LYS-147
2.	Rutin	−10.4	HIP-179, ARG-238, THR-245	TYR-182	ILE-276, TYR-182, TYR-177, LEU-132, LEU-203, TYR-153, VAL-99, ALA-100, TYR-235, PHE-239, MET-242, ALA-246, PHE-247, TYR-257, PHE-254, ILE-272	THR-245, ASN-277	GLU-42, ASP-129	LYS-147, HIP-179, ARG-238	GLY-148	LYS-147
3.	Withaoxylactone	−9.0	LYS-147, TYR-182, HIP-179, ARG-238, ASN-277	–	ILE-276, VAL-99, ALA-100, MET-242, PHE-239, PHE-247, TYR-235, TYR-182, PHE-254, TYR-257, ILE-272	ASN-277, SER-104, THR-127	ASP-129, GLU-201	LYS-147, HIP-179, ARG-238, LYS- 103	–	–
4.	2,3-Dihydrowithanolide E	−8.8	TYR-235, TYR-257, HIP-179	–	ILE-276, TYR-235, ILE-272, LEU-132, TYR-153, LEU-203, VAL-99, TYR-177, TYR-182, PHE-254, TYR-257, PHE-239, MET-242, PHE-247	ASN-277	ASP-129	LYS-147, HIP-179	GLY-148	–
5	ZVW	−8.5	TYR-177	–	VAL-99, ALA-100, LEU-132, TYR-153, TYR-177, TYR-182, LEU-203, TYR-235, PHE-239, MET-242, PHE-247, PHE-254, TYR-257, ILE-272 and ILE-276	ASN-277	ASP-129	LYS-147, HIP-179	GLY-148	–

In a similar vein, the docked complex involving PanK and Withaoxylactone established five hydrogen bonds with specific residues (LYS-147, HIP-179, TYR-182, ARG-238, and ASN-277). Furthermore, this docked complex exhibited interactions with critical residues, encompassing hydrophobic moieties (VAL-99, ALA-100, TYR-182, TYR-235, PHE-239, MET-242, PHE-247, PHE-254, TYR-257, ILE-272, and ILE-276), polar groups (SER-104, THR-127, and ASN-277), as well as negatively charged counterparts (ASP-129 and GLU-201), and positive (LYS- 103, LYS-147, HIP-179 and ARG-238 residues) (Fig. 2e and f). The PanK-2, 3 Dihydrowithanolide E docked complex also showed three hydrogen bonds (HIP-179, TYR-235 and TYR-257 residues), hydrophobic (VAL-99, LEU-132, TYR-153, TYR-177, TYR-182, LEU-203, TYR-235, PHE-239, MET-242, PHE-247 PHE-254, TYR-257, ILE-272 and ILE-276 residues), polar (ASN-277 residue), negative (ASP-129 residue), positive LYS-147 and HIP-179 residues) and glycine (GLY-148 residues) interactions (Fig. 2g and h). Phytochemical compounds of Withania somnifera exhibited pi–pi interactions with the PanK protein. Additionally, reference inhibitor, ZVW (PubChem ID: 71307936) (Fig. S1, Supplementary File), showed only one hydrogen bond (TYR-177 residue), hydrophobic interaction (VAL-99, ALA-100, LEU-132, TYR-153, TYR-177, TYR-182, LEU-203, TYR-235, PHE-239, MET-242, PHE-247, PHE-254, TYR-257, ILE-272 and ILE-276 residues), polar (ASN-277 residue), negative (ASP-129 residue), positive (LYS-147, HIP-179 and ARG-238 residues) and glycine (GLY-148 residue). 2, 3-Dihydrowithanolide E, Rutin, Withaoxylactone, and Morkotin A are finalized phytochemicals that are putative inhibitors of PanK compared to the reference compound (ZVW), according to an examination of the interaction profiles of all docked poses.

3.3. MDS assessment

MDS is a technique used to observe the movement of individual atoms within a protein molecule over a period of time. This simulation is performed under the influence of a force field, which controls the physical laws governing the interactions between atoms and these interactions are based on the principles of classical mechanics. The intermolecular interaction stability of 2, 3-Dihydrowithanolide E, Rutin, Withaoxylactone and Morkotin A with PanK protein was examined by simulating each docked complex to a 100 ns simulation, where they were exposed to a solvent environment, maintained at specific temperature and pressure conditions. This allowed us to assess the behaviour and interactions of these complexes over the course of the simulation. We derived the molecular dynamics properties for each complex from the 100 ns trajectory by utilizing the SID feature within the Maestro-Desmond interface. The extracted information encompassed key metrics such as RMSD, RMSF, and protein-ligand contact maps. RMSD served to assess structural variations, RMSF illuminated residue and atomic mobility, and these factors collectively contributed to dynamic stability evaluation. Furthermore, interaction maps were scrutinized to ascertain the intermolecular bonding within the docked configurations.

We initiated the RMSD calculations for both protein and ligand in the docked complexes involving potential phytochemicals and PanK by using the initial pose as a fixed reference frame. Notably, all of the docked PanK-ligands displayed rather extensive residency in the PanK’s selective pocket as a function of the 100-ns MDS interval, with exception of observed deviations in ligand conformations were within an acceptable range when compared to the reference compounds. In the process of 100 ns MD simulation, the Cα atom of PanK in PanK-ligand complexes remained thermally stable and exhibited the RMSD values typically ranged from 1.5 Å to 3.5 Å on average with 2, 3-Dihydrowithanolide E, Withaoxylactone, and Morkotin A, compared to reference compounds (ZVW) (Fig. 3 and Fig. S2, Supplementary file). However Rutin-PanK complex showed acceptable deviation till 80ns but at 100 ns it get separated and attain instability. The RMSF values (Figs. S3 and S4, Supplementary file) also supported these observations. The phytochemical compounds (2, 3-dihydro withanolide E, Rutin, Withaoxylactone, and Morkotin A) were observed to exhibit superior stability as comparison to the reference inhibitors (ZVW).

Fig. 3.

RMSD (Root mean square deviation) values were acquired for the alpha carbon atoms of M. tuberculosis PanK protein (blue curves) and phytochemicals (red curves) from the docked complexes., viz. (a) Morkotin A (b) Rutin, (c) Withaoxylactone, (d) 2, 3-Dihydrowithanolide E. , were extracted from 100ns MD simulation time. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Various dynamic parameters were assessed for each ligand to understand the conformational stability within protein-ligand complexes (Fig. S5, Supplementary file). Ligand radius of gyration (rGyr) fell within 4.0–4.9 Å for Withaoxylactone and 2,3-dihydro withanolide E, while it ranged from 4.8 to 6.8 Å for Morkotin A and Rutin. Notably, Morkotin A exhibited a higher number of intra-molecular hydrogen bonds (intraHB). Solvent-accessible surface area (SASA) spanned 100–410 Å, with molecular surface area (MolSA) and polar surface area (PSA) ranging from 150 to 600 Å. Furthermore, we examined protein-ligand interactions in the 100 ns MD simulations of all docked complexes, focusing on key elements like hydrogen bonds, hydrophobic contacts, ionic interactions, and water bridges (Fig. 4, Fig. 5). In the PanK-Morkotin A complex, hydrogen bonds involving Met-144, Tyr-182, Tyr-235, Arg-238, and Tyr-257 represented 30–80 % of the interactions, while hydrophobic interactions with Tyr-182, Phe-254, and Ile-272 constituted 60 % of the simulation. Water bridges formation were exhibited by residues Asp179 and His-179, where His-179 showed for more than 100 % of water bridge formation during the 100 ns of simulation time. Additionally, Lys-103 residues formed an ionic bond for 10 % simulation time (Fig. 4, Fig. 5 a). During a 100 ns MD simulation, the PanK-rutin docked complex showed hydrogen bonding with several residues (Lys-147, His-179, Tyr-182, Ala-246, and Ser-252) for 20 %–70 % of the time. Additionally, hydrophobic interactions (involving Tyr-182, Phe-254, and Ile-272) were observed for 30 %, 40 %, and 90 % of the time, respectively. The simulation also revealed that other residues (Asp-129, His-179, Leu-180, Tyr-235, Asp-249, Tyr-257, and Asn-277) were involved in the interaction through a water bridge for 50 % and 75 % of the time. (Fig. 4, Fig. 5 b). In the protein contact mapping analysis of the Withaoxylactone-PanK complex, it was observed that hydrogen bonding (Arg-108, Lys-147, His-179, Tyr-182 and Asp-277) were exhibited for 20–50 % of the 100 ns simulation time, whereas more than 30 % of hydrophobic interaction was recorded for Leu-180, Tyr-182, Phe-239, Phe-254, and Ile-276 residues throughout the simulation. Notably, Lys-103 residue showed more than 50 % of the hydrogen bond along with the water bridge interaction (Fig. 4, Fig. 5 c). In comparison, the docking of PanK-2, 3-Dihydrowithanolide E resulted in relatively few residues engaging in interactions. Specifically, Tyr-177 and His-179 formed hydrogen bonds for over 50 % of the time, and Tyr-235 and Tyr-257 engaged in interactions for over 60 % of the simulation. Hydrophobic interactions were observed with Leu-132 and Ile-272 for 20 % of the time and with Phe-254 for over 40 % of the simulation. Additionally, Arg-238 participated in water-bridge formation for 75 % of the time during the 100 ns simulation.

Fig. 4.

Protein-ligand interactions mapping for M. tuberculosis PanK protein with the potential phytochemicals, i.e., (a). Morkotin A (b) Rutin, (c) Withaoxylactone, (d). 2, 3 Dihydrowithanolide E, extracted from 100 ns simulations.

Fig. 5.

Formulate a visual depiction that outlines the interactions between PanK protein and the docked compounds in a schematic manner i.e., (a). Morkotin A (b) Rutin, (c) Withaoxylactone, (d). 2, 3-Dihydrowithanolide E, extracted from 100 ns simulations.

In contrast, the reference complex ZVW exhibited substantial interactions with PanK, including hydrophobic interactions with His-179 and Tyr-182 for more than 100 % of the time, indicating strong and consistent interactions (Fig. S6, Supplementary file). A water bridge formed at Tyr-235 for 66 % of the simulation duration, and Tyr-177 was involved in a hydrogen bond for 90 % of the simulation time (Fig. S7, Supplementary file). Overall, the MD simulation analysis suggests that the selected compounds show significant stabilization within the active pockets of PanK, making them promising candidates for potent inhibitors against M. tuberculosis-PanK.

3.4. ADME analysis

ADMET profiling is crucial in early drug discovery to assess drug-likeness and pharmacokinetics. We analyzed the top four compounds along with a reference compound (ZVW) (Table S4, Supplementary file). None of these compounds cross the brain/blood barrier. 2, 3-Dihydrowithanolide E meets Lipinski’s rule of five, while Withaoxylactone has one violation, and Morkotin A and Rutin have three violations. Natural substances may not always adhere to drug-likeness rules due to active transport. Overall, these compounds show promising medicinal properties.

3.5. Binding energy calculations

The net binding free energy analysis revealed strong binding affinity of all four docked ligands (rutin, withaoxylactone, 2, 3-dihydro withanolide E, and Morkotin A) for the PanK protein, outperforming the reference complex. Notably, rutin exhibited the highest affinity with a binding free energy of −220.09 kcal/mol. These results indicate favorable interactions within PanK’s selective pocket over a 100 ns interval.

3.6. In vitro analysis

Natural medicines are often touted for their perceived lack of side effects, cost-effectiveness, and ready availability, making them potentially advantageous for the well-being of humanity over extended periods [29]. The rich presence of diverse phytochemicals with anti-tubercular properties suggests that exploring natural products could prove fruitful in uncovering novel leads for anti-TB interventions [30]. The anti-TB activity of rutin, Morkotin A, and leaf extracts of Withania somnifera was observed against M. tuberculosis H37Ra. M. tuberculosis, characterized as a slow-growing intracellular pathogen with a complex cell envelope containing mycolic acids and various unique lipids, presents a formidable challenge in developing effective treatments [31,32]. The leaf extracts of Withania somnifera, specifically elevated (ET) and ambient (AT), demonstrated notable anti-tubercular activity against M. tuberculosis H37Ra (Fig. 6.). The determination of the minimum inhibitory concentration (MIC) using the MABA assay revealed values of 97.5 μg/ml and 195 μg/ml for AT and ET extracts, respectively. The MIC of control i.e INH was 0.125 μg/ml. These findings underscore the substantial anti-TB activity exhibited by W. somnifera in this study, suggesting its potential as a promising anti-TB agent.

Fig. 6.

Microplate assay showing anti-tubercular activity of Rutin, Morkotin A leaf extracts under ambient (AT) and elevated (ET) temperatures. Blue wells indicate bacterial inhibition, while pink wells indicate growth. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

4. Conclusion

The application of in silico drug discovery methodologies in this research has demonstrated their value as efficient tools for the early stages of drug development. By employing computational techniques, this study identified potential drug candidates, elucidated their structural interactions with target proteins, optimized their properties, and predicted safety profiles, significantly reducing time and cost while enhancing the likelihood of success. Focusing on Pantothenate Kinase (PanK), a key enzyme in the CoA biosynthetic pathway and a molecular target for anti-TB therapeutics, the study explored the therapeutic potential of phytochemicals from Withania somnifera (Ashwagandha). The rich phytochemical diversity of this plant, known for its medicinal value, holds significant promise for anti-tubercular drug discovery.

From a pool of 83 compounds, four—Morkotin A, Rutin, Withaoxylactone, and 2,3-Dihydrowithanolide E—emerged as promising PanK inhibitors, with superior binding interactions compared to reference inhibitors. Molecular docking, ADME analysis, and extensive molecular dynamic simulations confirmed their stability, drug-like properties, and favorable pharmacokinetic attributes. Supporting these findings, previous studies reported anti-TB activity in rutin and leaf extracts of Withania somnifera against M. tuberculosis H37Ra. The AT and ET treated leaf extracts of Withania somnifera demonstrated notable anti-tubercular activity, with minimum inhibitory concentrations of 97.5 μg/ml and 195 μg/ml, respectively. These results reinforce the potential of natural medicines as cost-effective, accessible, and low-side-effect alternatives for long-term human health benefits.

This study highlights the power of in silico methods in accelerating drug discovery and underscores the untapped potential of natural products in developing anti-TB therapeutics. The results suggest further studies are needed to confirm the safety and efficacy of these promising candidates. Future advancements in computational tools and collaborative efforts between computational and experimental researchers will likely enhance the development of novel, effective, and safe treatments for tuberculosis, benefiting both patients and the broader field of medicine.

Author contributions

A.S.: conceptualization the study, developed the research framework, and were responsible for drafting and finalizing the manuscript. U.M: conceptualization the study, developed the research framework, and were responsible for drafting and editing the manuscript. P.Y. provided critical support in data analysis and interpretation. While P.Y. was involved in aspects of the laboratory workflow, he was also involved in final drafting of the manuscript

Use of generative AI in scientific writing

The authors declare that they have not used generative AI and AI- assisted technologies in the writing process of this paper.

Funding sources

This research received no funding.

Conflict of 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.

Appendix A. Supplementary data

The following is the Supplementary data to this article: