Metabolite fingerprinting of phytoconstituents from Fritillaria cirrhosa D. Don and molecular docking analysis of bioactive peonidin with microbial drug target proteins

Fritillaria cirrhosa D. Don (Liliaceae), a valuable and critically endangered medicinal herb of northwest India, including Jammu and Kashmir, grows in temperate to alpine regions of the Himalaya. It is known as the traditional herb for cardiovascular diseases, respiratory diseases, and metabolic disorders. The plant bulbs are precious and are used to cure many other health complications. The current study analysed the phytoconstituents by liquid chromatography-mass spectrometry (LC–MS) of different crude extracts (methanolic, petroleum ether, and ethyl acetate) of F. cirrhosa. The LC–MS analysis from the bulbs of F. cirrhosa yielded 88 bioactive compounds, with the vast majority having therapeutic applications. Further, determination of minimum inhibitory concentrations (MICs) by broth microdilution method of F. cirrhosa against tested bacterial and fungal pathogens showed remarkable results with MICs ranging between 6.25–200 µg/mL and 50–400 µg/mL, respectively. Subsequently, these 88 identified phytocompounds were tested for their bioactivity through ADMET prediction by SwissADME and in silico molecular docking studies. Results revealed that Peonidin might have maximum antibacterial and antifungal activity against various microbial protein drug targets among the phytochemical compounds identified. Furthermore, the highest binding affinity complex was subjected to molecular dynamic simulation (MDS) analysis using Desmond Schrodinger v3.8. The root-mean-square deviation (RMSD) graphs obtained through the molecular dynamic simulations indicated the true bonding interactions, further validated using the root-mean-square fluctuation (RMSF) graphs which provided a better understanding of the amino acids present in the proteins responsible for the molecular motions and fluctuations. To our best knowledge, this is the first description of the phytochemical constituents of the bulbs of F.cirrhosa analyzed through LC–MS, which show pharmacological significance. The in silico molecular docking and molecular dynamics study of peonidin was also performed to confirm its broad-spectrum activities based on the binding interactions with the antibacterial and antifungal target proteins. The present study results will create a way for the invention of herbal medicines for several ailments by using F. cirrhosa plants, which may lead to the development of novel drugs.


Minimum inhibitory concentration by broth dilution method. Minimum inhibitory concentration
in the present study was performed through Muller Hinton broth dilution method. In the present study, 10% ADC was supplemented in broth followed by performing a two-fold serial concentration from 0.00 to 0.20 µg/ mL of antibiotics, including ciprofloxacin and amphoterecin-B in different combinations with fungal and bacterial cultures. Moreover, each inoculated plate was kept in an incubator at 37 °C for 24 to 48 h. To monitor the growth of cells and observe growth inhibition, the minimum inhibitory concentration of the drug was noted. MIC was considered as the lowest concentration of an antimicrobial at which no balanced growth was observed. In addition, colony-forming units were, measured on day 28th of incubation at 37 °C for better use.
In terms of drug interaction, the microbial cultures were grown in various concentrations of ciprofloxacin and amphotericin B in multiple combinations. The antimicrobials were added to fabricate 8 × MIC for both the antimicrobials followed by two fold dilutions leading to 1/64 × MIC concentration ranges of each drug both alone and in combination were assured 31 . Ligand and receptor preparations. The molecular docking analysis was carried out for all the selected phytocompounds from the LC-MS data fractions. Chemical structures of phytocompounds retrieved from PubChem database in SDF format and converted into protein data bank (PDB) format using PyMol Version 3.0. Software. The selected targets for this study were based on previous studies, and all the 3D crystal structures were fetched from (PDB) in PDB format. The structural information of the Peonidin identified from the Fritillaria cirrhosa was also retrieved from the PubChem database in SDF format and converted into PDB format using an open babel server. Peonidin is one of the compounds identified from LC-MS analysis. It exhibits apoptotic effects in the breast cancer cells 32 .
Peonidin is an anthocyanide that can be easily eliminated from the human body. Anthocyanins (ACNs), including peonidin, are natural bioactive compounds with many pharmacological effects: antioxidant, antiinflammatory, prevention of age-related chronic diseases: cardiovascular disease (CVD), cancers, neurodegenerative, and eye-related diseases. ACNs also have antiviral properties. Recent in vitro studies have shown that they can inhibit the replication of viruses such as herpes simplex, parainfluenza virus, syncytial virus, HIV, rotavirus, and adenovirus 33 . The broad spectrum of pharmacological properties supported by preclinical and clinical evidence associated with low toxicity make their pharmacotherapeutic use very attractive. Peonidin was selected as an appropriate ligand in the current study based on its pharmacological properties. Compounds with antimicrobial activities disable bacteria by targeting key bacterial metabolism components, including dihydropterote synthase (DHPS), penicillin binding protein (PBP), elongation factor Tu, 1,3 beta glycan, ABC transporter and beta-tubulin . These six key proteins play a critical role in the life cycle of fungi and bacteria. Penicillin-binding proteins (PBPs) are membrane-associated proteins that play a vital role in forming cell wall 34,35 . As previously indicated, antibiotics target the dihydropteroate synthase enzyme 36 . This enzyme participates in the folate synthesis pathway, directly connected to nucleic acid synthesis 37 . Inhibiting this enzyme with a strong antibiotic will have a more profound and irreversible effect on the bacterium's protein production. EF-Tu's primary role is to transport amino-acylated tRNAs to the ribosome. Since the 1970s, antibiotics have targeted EF-Tu as a therapeutic 38 . ABC transporters (ATP-binding cassette) are found in all three domains of life and mediate transmembrane transport of a wide range of substrates, including medicines, proteins, carbohydrates, amino www.nature.com/scientificreports/ acids, ions and carbohydrates. Glucans have long been known to have immunomodulatory effect 39 . Tubulin are the protein subunits of microtubules which helps in cell division of prokaryotic and eukaryotic organisms 40 .
In view of this, the molecular docking study was carried out to examine the binding interactions of peonidin with these microbial drug targets. All the targeted receptors were fetched from the protein data bank (PDB) in the PDB format and subjected for purification and refinement. All the receptors were prepared for molecular docking analysis using Biovia Discovery Studio and eliminate undesirable ligands, molecules of water, and other impurities from all of the proteins involved in the study.
Purification and refinement of proteins and ligands. Biovia Discovery studio was used to eliminate undesirable ligands, molecules of water, and other impurities from all of the proteins involved in the study. Polar hydrogens were added to the protein during the preparation process for improved interactions, followed by Kollman charges. The 2D structure and 3D structure of peonidin were retrieved using the PubChem database (PubChem CID: 5281708). The 3D structures of dihydropteroate synthase (DHPS), penicillin-binding protein (PBP), elongation factor Tu, 1,3 beta glycan, ABC transporter and beta-tubulin were retrieved from the Protein Data Bank (Fig. 1).

Molecular docking studies.
The docking calculation and algorithms to reproduce the pocket binding residues was verified employing co-crystal ligands. In the current study, we have used the Lamarckian Genetic Algorithm in Autodock Vina 4.2 to initiate the docking analysis with the standard parameters. Furthermore, all the transformation were clustered by considering the 4.0A RMSD and the most favourable binding poses was selected depending on the lowest free energy and inhibition constant. The molecular docking studies were performed on all selected phytocompounds from the LC-MS data sections. Chemical structures of phytocompounds were obtained in SDF format from the PubChem database and were translated to PDB format with PyMol Version 3.0. The targets for this investigation were chosen based on the evidence from past research, and all 3D crystal structures were collected in .pdb format from the Protein Data Bank. Based on the isolation modes and species-specific variants, the crystal structures of proteins such as dihydropterote synthase, penicillin-binding protein, and elongation factor Tu (Ef-Tu), 1,3 beta glycan, ABC transporter and beta-tubulin were retrieved.
In-silico drug-likeness and toxicity predictions. Drug-likeness of compounds was predicted based on an already established concept by Lipinski et al. 41 . The structures of all some pharmacologically important compounds were converted to their canonical simplified molecular-input line-entry system (SMILE). They submitted to the SwissADME and PreADMET tool to estimate in silico pharmacokinetics, such as the number of hydrogen donors, hydrogen acceptors, rotatable bonds, and total polar surface area of a compound. These compound's organ toxicities and toxicological endpoints were predicted using PreADMET and OSIRIS Property 42 . www.nature.com/scientificreports/ The selection of compounds as drug candidates was determined by a drug score parameter. The higher the drug score value, the higher the compound's chance of being considered a drug candidate 43 .
Molecular dynamic simulation. The Desmond Schrodinger package was employed for the molecular dynamic simulation with the highest binding affinity docked complex to analyse the docked calculation and interpretation. The analysis indulged simulation run for 100 ns to check the stability of structures. Desmond's inbuilt parameters were also used for protein preparation and bond orders. In the case of bond orders, the addition of hydrogen binds, and filling of missing residues are also optimized during the initial process. OPLS force field was used to initiate the builder system for the standard simulation grid box. And the conformation were retrieved in the form of simulation trajectories which could be analyzed using Root mean square deviation and root mean square fluctuation graphs through hydrogen bond analysis and radius of gyration plots.

Results
Preliminary phytochemical screening. The phytochemical study of various extracts from the bulbs of F.
cirrhosa revealed a variety of phytochemicals. The essential constituents present in petroleum ether, ethyl acetate and methanolic extracts as represented in Table 2.
Chemical composition. The separation of individual components presents in a solvent-based on their mass/charge ratio is the working principle behind Liquid Chromatography-Mass Spectroscopy (LC-MS). The chemical composition of the extracts from the F. cirrhosa, analyzed by LC-MS, are represented in (Table 3; Fig. 2

Minimum inhibitory concentration.
In terms of novel antimicrobial drugs, we have investigated the antimicrobial compounds extract from F. cirrhosa that showed significant inhibitory activity shown in Table 4 and observed to be non-toxic. The present study elaborates the detailed antimicrobial activity and in silico study for better validation.  Table 4.

Molecular docking analysis.
The LC-MS analysis revealed that F. cirrhosa bulb extracts contained 88 bioactive compounds (Table 3). These phytocompounds were analyzed for activities against bacterial and fungal target proteins. The docking studies were carried out for phytoligands using the AutoDock Vina programme to elucidate the binding affinities to the target proteins. In general, among the 88 phytocompounds. All the six microbial drug targets related protein structures were docked against best-screened ligand peonidin obtained after ADME screening based on different parameters like solubility, toxicity, absorption, molecular weight and excretion. Docking was performed using AutoDock Vina with Chimera Plugin. Subsequently, the best-docked complex was screened based on their binding energy. Bonding between ligand and target protein were visualized using Biovia discovery studio. 2D-interaction diagrams showing different bonds formed between ligand and target protein were also visualized using Biovia discovery studio. Peonidin (2-(4-hydroxy-3-methoxyphenyl) chromenylium-3,5,7-triol) exhibited the best binding conformations with the lowest binding energy values with bacterial (− 8.2 kcal/mol), fungal (− 8.2 kcal/mol) target proteins (Tables 5 and 6). Our findings are supported by a previous study that showed that the lower the binding energy score, the better the protein-ligand binding stability was identified 44 . The peonidin formed the most excellent ligand-protein complexes with six tested microbial target proteins compared with other compounds. According to the binding energies, the docking results of peonidin with target proteins were ranked. The docking results of peonidin with the proteins, such as dihydropterote synthase, penicillin-binding protein, elongation factor Tu, 1, 3 beta glycan, ABC transporter and betatubulin confirmed that the ligand has a higher affinity for penicillin-binding protein, which is a key regulator in cell wall synthesis and maintenance (Fig. 6). In comparison to ciprofloxacin, the extracted peonidin natural molecule comparatively showed same docking efficiency with PBP within the binding pocket. The docked structure was imaged to illustrate the ligand interactions with significant amino acids TYR A:561, ILE A:371, ASN A:397, PHE A:450, and GLN A:452 through Vander Waal forces as well as hydrogen bonding (Fig. 6). The binding affinity of the selected compound against antibacterial and antifungal proteins are represented in (Tables 5, 6

Molecular dynamic simulation.
The best-hit compound was considered based on the binding affinity and transformation for molecular dynamic studies to analyze the stability of the compound in contrast to the protein of our interest. For simulation studies, the Desmond Schrodinger suite program was used for receptor complexes followed by the RMSD of protein exhibited a stable trajectory during the dynamic run for 100 nano seconds studies. Simulation of 100 nano seconds indicates significant conversions and stable conformations when comparing protein-ligand RMSD, as shown in Fig. 7. Regarding dynamic studies, RMSD exposed a fluctuated trajectory till 30 nano seconds, and there are fewer fluctuations between ligand and protein conformations (Fig. 7). Overall, both the receptor and ligand RMSD was observed in a stable format and can be seen in (Figs. 7,8,9,10). In addition, all the amino acid interactions of our complexes were also analyzed accordingly, as well as the unique hydrogen bonds noted like TYR 561, ILE 371, ASN 397, and so on.

Discussion
F. cirrhosa is a versatile medicinal plant with numerous biological properties in its nature. In the present study, the investigation of petroleum ether, ethyl acetate, and methanol extracts from bulbs of F. cirrhosa revealed the presence of various phytoconstituents, including flavonoids and carbohydrates saponins, phenols, alkaloids, steroids, and terpenoids. These bioactive phytoconstituents could be responsible for the therapeutic ability of various extracts of F. cirrhosa. The analysis was carried out by liquid chromatography-mass spectrometry (LC-MS), one of the most widely used techniques for separating phyto-constituents. The LC-MS investigation of F. cirrhosa bulb extracts revealed the presence of 88 phytochemical compounds, which could contribute to  www.nature.com/scientificreports/ alkaloids like sinpeinine A (17), imperialine-3-β-glucoside (18), imperialine (7) and 3β-acetylimperialine have been reported from Bulbus F. cirrhosa 14 . From the above evidence, it can be elucidated that F. cirrhosa consists of the enormous potential of pharmacological constituents therapeutic phytocompounds responsible for various pharmacological actions. This is the first report on LC-MS investigation of F. cirrhosa bulb extracts to the best of our knowledge. Furthermore, the present study investigated the in vitro antimicrobial activity against various bacterial and fungal strains. The study focused on the gram-positive bacteria and gram-negative bacteria, which shows the significant efficiency. Among various fungal and bacterial strains such as E. coli, K. pneumonia, M. luteus, S. pneumonia, H. influenzae, N. mucosa and C. albicans, C. glabrata, C. Parapsilosis were observed to be more susceptible to the plant extracts with the better minimum inhibitory concentration values (Table 4) compare to co-crystals MICs (0.625, 1.25, 0.039, 0.625, 1.25, 3.12, 1.25, 2 0.5, and 2.5) at a combinational concentration. In addition, it is also mentioned to that, the therapeutic herbs have different mechanisms that reflects identical  www.nature.com/scientificreports/ antimicrobial mode of therapeutics in contrast to antimicrobial agents 10,51 . According to molecular docking, peonidin with 6 different proteins (dihydropterote synthase, penicillin binding protein, elongation factor Tu, 1,3 beta glycan, ABC transporter and beta-tubulin) exposed that the peonidin has a better interaction and binding score with penicillin binding protein as compared to other phytocompounds. The docking study results also showed that various energy sources are consistent and contribute to the overall strength of the binding interactions of peonidin for each target protein. After docking, computational molecular dynamic simulation studies revealed that various amino acids were observed to have interacted with peonidin, which is requisite for the synthesis of purines and pyrimidine nucleotides. Most of the interacted amino acids were bound with    www.nature.com/scientificreports/  www.nature.com/scientificreports/ the hydrogen bonding and Vander Wall forces, whereas, the docking outcome with dihydropteroate synthase exposed the interrelated binding affinity with targeted receptors. The trajectory of molecular dynamic simulation was completed for 100 nano seconds and the RMSD and RMSF graph values were analyzed. The constant increase in the RMSD values with respect to time indicates that the protein serves from its native conformation regularly. The higher RMSF was analyzed in the study where strong hydrogen bonding was noted with mostly polar amino acids having maximum peak.

Conclusion
The present investigation focused on identifying various bioactive compounds from the bulb extracts of F. cirrhosa for the first time by LC-MS analysis. These compounds are responsible for the different therapeutic and pharmacological properties. We have also provided evidence of bulb extracts of F. cirrhosa for its antimicrobial activity. Furthermore, the present study also explained the role of the peonidin compound against antimicrobial proteins such as dihydropteroate synthase (DHPS), penicillin-binding protein (PBP), elongation factor-Tu (Eu-Tu), 1,3 β-glycan, ABC transporter, and beta-tubulin. Among all the bacterial and fungal species, peonidin showed the best docking score against PBP and was then subjected to dynamic simulation studies to investigate  www.nature.com/scientificreports/ the stability and RMSD and RMSF values concerning protein-ligand transformations. Using the peonidin compound may enable us to develop an effective drug against pathogenic bacteria and fungal diseases. In addition, ADMET (drug-likeness) studies showed the highest drug-likeness properties of the studied compound, which suggests that the peonidin compound can act as a promising microbail drug candidate. Further investigations to determine its bioactivity, toxicity profile, and clinical studies are necessary for broad-spectrum drug discovery.