Uncovering the bactericidal potential of extract and multi-targeting phytochemicals from Mirabilis longiflora L. leaves against multidrug-resistant Pseudomonas aeruginosa and Bacillus cereus

Abstract

Mirabilis longiflora L. has been used traditionally in some parts of Bangladesh for the treatment of headaches, infectious diseases, painkillers, skin disease, herpes infection, and wound healing, but its effects on multidrug-resistant (MDR) bacteria have remained unidentified. Therefore, we aimed to determine the antibacterial activity of the methanol extract of M. longiflora L. leaves (MEMLL) against MDR Pseudomonas aeruginosa and Bacillus cereus and to recognize possible multitargeting antibacterial phytocompounds through in silico computational approaches targeting the LasR and LpxC proteins in MDR Pseudomonas aeruginosa and the FosB and PlcR proteins in MDR Bacillus cereus. PPS, FT-IR, and GC-MS were used for profiling of the phytocompounds in MEMLL. The antimicrobial activity of MEMLL was evaluated using in vitro agar-well diffusion, MIC, and MBC assays. In silico methods were applied to identify multi-targeting agents from GC-MS-annotated phytocompounds. MEMLL showed dose-dependent antibacterial activity and exposed the presence of 33 phytochemicals in GC-MS analysis. Among these, 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334) was identified as a potential antibacterial phytocompound as it exhibited multi-modal and strong binding affinity towards LasR, LpxC, FosB, and PlcR, favorable pharmacokinetics, drug-likeness, physicochemical, and toxicity properties. Finally, Molecular dynamics (MD) simulations demonstrated the structural stability of CID 14,334 within the active sites of LasR, LpxC, FosB, and PlcR. The results of this study offer scientific validation for the traditional use of M. longiflora L. in bacterial infection-related diseases. It also suggests that 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one from M. longiflora L. might be responsible for the antibacterial activity and could act as a phytopharmacological lead for the development of LasR and LpxC inhibitors against MDR Pseudomonas aeruginosa and FosB and PlcR inhibitors against MDR Bacillus cereus.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-40444-3.

Introduction

Currently, antimicrobial resistance is a major source of global morbidity and mortality¹and one of the major challenges facing global health today². Nowadays, it is becoming the leading cause of death worldwide³. Antibacterial resistance mechanisms have been developed by a large number of bacteria⁴, which poses a growing financial burden on the global health system⁵. Gram-negative (Pseudomonas aeruginosa) and gram-positive (Bacillus cereus) bacteria are of special concern because they have altered metabolic pathways for survival and persistence, making them resistant to current antibiotics^6,7. Their infectivity, virulence, and pathogenicity, enabled by Quorum Sensing (QS), make it challenging to eradicate even during antimicrobial therapy^8–10.

The QS systems of P. aeruginosa are encoded by bacterial lasR/lasI and rhlR/rhlI genes¹⁰. The lasI gene encodes the LasI protein, which is responsible for the synthesis of the quorum-sensing (QS) signaling molecule 3-oxo-dodecanoyl homoserine lactone (3OC12-HSL), also called autoinducer. Similarly, the rhlI gene produces the RhlI protein, synthesizing another QS signaling molecule, N-butyryl-l-homoserine lactone (C4-HSL). Upon release from bacteria, these signaling molecules diffuse in the environment to interact with the adjacent microbes. The receptive microbe acquires a signal that binds to its specific receptors. In P. aeruginosa, two major receptors (i) LasR, the product of lasR, and (ii) RhlR, the product of rhlR, bind with 3-oxo-dodecanoyl homoserine lactone (3OC12-HSL) and N-butyryl-l-homoserine lactone (C4-HSL), respectively. These bindings form the complexes of LasR: 3OC12-HSL and RhlR: C4-HSL, thereby activating LasR and RhlR¹¹. The activation of LasR helps the bacteria to control the regulation of genes responsible for the synthesis of elastase, hemolysin, protease, and exotoxin-A, all of which are crucial for the production of exopolysaccharide (EPS), where all the members in the community concurrently produce EPS, thereby promoting the production of biofilm. This trait can significantly enhance the potency of virulent manifolds. However, the activation of RhlR regulates the production of virulence factors such as rhamnolipid, pyocyanin, elastases, and swimming and swarming motility factors, which are essential for biofilm establishment. The activation of LasR also activates the transcription of the rhlR gene for the production of the RhlR protein. This RhlR protein successively activates the rhlI gene to produce the RhlI protein. Not only does the activation of LasR activate the lasI gene to produce the LasI protein, but it also establishes a positive feedback loop that amplifies the production of QS signaling molecules. Additionally, the two quorum-sensing systems in P. aeruginosa are arranged hierarchically, with the rhlI/rhlR system functioning under the control of the lasI/lasR system, as the expression of rhlR and rhlI relies on the LasR protein produced by the lasR gene¹². QS system studies have shown that P. aeruginosa lasR mutants exhibit significantly reduced virulence and invasiveness in various in vivo infection models. Sitagliptin, a drug used to treat type-2 diabetes, has been found to interact with the LasR receptor in P. aeruginosa and effectively suppress biofilm production¹³. These findings indicate that LasR is a key molecule in quorum sensing (QS), biofilm formation, and antibiotic resistance.

LpxC is an essential zinc-dependent metalloenzyme vital for the survival of many Gram-negative bacteria, such as Pseudomonas aeruginosa. It catalyzes the first crucial step in lipid A (endotoxin) biosynthesis foundational component of lipopolysaccharides (LPS), which play a key role in cell wall integrity, biofilm development, and antibiotic resistance in many Gram-negative bacteria, including P. aeruginosa^14,15. LpxC is highly conserved across most Gram-negative bacterial species and lacks any homologous counterpart in the human genome¹⁵. Given their interconnected roles in biofilm formation, virulence, and antibiotic resistance, LasR and LpxC emerge as promising targets for the development of new antibacterial agents. Thus, inhibiting both LasR and LpxC with a single molecule may provide a promising approach for developing a drug to combat multidrug-resistant P. aeruginosa.

FosB is the M²⁺-dependent thiol transferase fosfomycin antibiotic resistance enzyme encoded by the FosB gene. This enzyme functions by catalyzing the nucleophilic addition of either L-cysteine (L-Cys) or bacillithiol (BSH) to the carbon 1 (C1) of the fosfomycin antibiotic, which opens the epoxide ring, resulting in a modified compound that renders the antibiotic ineffective and non-bactericidal^16,17.

PlcR is part of the PlcR-PapR-OppABCDF integrated quorum-sensing system and serves as the pleiotropic regulator of genes encoding extracellular factors potentially involved in biofilm formation and pathogenicity of Bacillus cereus¹⁸. PlcR transcription is autoinduced. The functioning of PlcR requires a 48-amino acid signaling peptide, PapR, which is encoded 70 bp downstream of plcR¹⁹.

The PapR peptide is expressed as a propeptide under the control of PlcR, exported out of the cell, processed into the active peptide either during export or in the extracellular medium, and then re-imported into the cell via the oligopeptide permease OppABCDF¹⁸. In this way, PapR accumulates inside the cell and functions as an autoinducer for PlcR. The PlcR-PapR complex then binds the PlcR box, a specific target site in the promoter region with a conserved, palindromic sequence (TATGNANNNNTNCATA), which activates the transcription of a vast array of genes coding for proteins located in the cell wall or extracellular space. These proteins are responsible for biofilm formation, antibiotic resistance, and pathogenicity^18,19.

Therefore, both proteins (FosB and PlcR) present themselves as appealing targets for novel antibacterial discovery. By logically inhibiting multiple factors of antibiotic resistance in bacteria, multi-target drugs are the best choice for addressing the emergence and spread of drug-resistant infectious bacteria, as no singular or simple strategy is sufficient.

Medicinal plant products have gained significant interest for their capacity to provide a broad spectrum of structurally diverse compounds with numerous reported biological activities, including antibacterial properties with multi-targeting antimicrobial functionality. This makes them a valuable resource for the discovery of novel antimicrobial substances with potential new mechanisms of action²⁰. Furthermore, the traditional use of medicinal plants in folk medicine serves as a valuable foundation for the discovery of antimicrobial agents from natural sources. These plants are known to produce bioactive compounds capable of counteracting bacterial resistance. Notably, there have been no reported cases of bacteria developing resistance to plant-derived antimicrobials. For instance, plant-derived coumarins exhibit potent antibacterial activity against Staphylococcus aureus. Similarly, berberine, isolated from Tinospora cordifolia (Willd.) Miers ex Hook. f. & Thomson, has demonstrated strong efficacy against Gram-positive bacteria, including drug-resistant strains of Staphylococcus aureus and Mycobacterium tuberculosis²¹.

M. longiflora L. is a medicinal plant in the Nyctaginaceae family. It is a shrub-like perennial plant with a tube 10.0–10.5 cm long. The plant is mainly visited and pollinated by Manduca quinquemaculata (moth) and is commonly found in India, Bangladesh, the United States, Arizona, New Mexico, and Texas²². It is known as Sondha Moni in Bengali and Sweet 9 o’clock in English. M. longiflora has been used in traditional medicine systems to treat skin infections, cuts, and wounds. The pharmacological activities of the hydroalcoholic extract of the leaves of M. longiflora L. have also been reported to attenuate ethanol-induced gastric ulcers through its antioxidant and anti-inflammatory effects²³. Although M. longiflora L. holds importance in traditional medicine and is known for its effects on gastric ulcers, its antibacterial effects and the molecular mechanisms of its compounds against multidrug-resistant bacteria have yet to be explored. Our research aimed to identify compounds that can modulate more than a single biological target in MDR P. aeruginosa and MDR B. cereus, leading to effectively blocking its robust antibiotic resistance mechanisms. Therefore, reasonably multi-target drugs would be the best choice for inhibiting LasR and LpxC from P. aeruginosa and FosB and PlcR from B. cereus, which have gained remarkable relevance in drug discovery owing to the complexity of multifactorial causes of infectious MDR bacterial diseases. Therefore, this study concentrated on examining the in vitro antibacterial properties and in silico polypharmacological drug profiles of phytocompounds from M. longiflora. Specifically, the in silico study targeted the virulence and multidrug resistance of P. aeruginosa and B. cereus by focusing on biofilm formation in both bacteria and the synthesis of lipid A in the outer membrane of P. aeruginosa. This thorough study comprised in vitro experiments, including agar-well diffusion, MIC, and MBC assays, while in silico experiments entailed molecular docking, pharmacokinetics, drug-likeness, toxicity analysis, and molecular dynamics simulation to validate the efficacy of the docked compounds. The results obtained from this study are expected to provide significant contributions in terms of understanding potential pharmaceutical applications of M. longiflora in combating MDR P. aeruginosa and B. cereus, and they are likely to contribute to the development of novel multi-target directed drug-lead compounds for tackling infections linked to multidrug resistance (Fig. 1).

Fig. 1.

Graphical workflow summarizing the antibacterial evaluation of Mirabilis longiflora L. leaves extract against Pseudomonas aeruginosa and Bacillus cereus, featuring FT-IR, GC-MS, molecular docking, ADME analysis, and molecular dynamics simulation to identify potential lead compounds. The schematic illustration was created using BioRender (https://biorender.com).

Materials and methods

Chemicals and reagents

Methanol, ethanol, H₂SO₄, Fehling’s solutions A and B, and CuSO₄ were purchased from Merck, Germany. The lead acetate, sodium hydroxide, and NaCl were purchased from BIOSOL, India. Hydrochloric acid was obtained from RCI Labscan Ltd, Thailand; Bacto agar and LB media from Liofilchem, Italy; chloroform from Pallav Chemicals Pvt. Ltd, India; and sodium nitroprusside from Loba Chemie Pvt. Ltd, India. Ampicillin discs were sourced from Bio-Rad in the USA. Anhydrous Na₂CO₃, FeCl₃, sodium citrate, and ninhydrin were acquired from Sigma-Aldrich in Germany in this study.

Plant material collection and identification

Mirabilis longiflora L. plants and leaves were collected in March 2023 from Chowgacha, Jashore, Bangladesh. Taxonomic identification of the plant was performed by Dr. Sardar Nasiruddin, a taxonomist at the National Herbarium in Dhaka, Bangladesh, where the specimen was preserved under the accession number DACB: 32,613. Permission to collect Mirabilis longiflora L. was obtained from the institutional Ethical Committee, and all collections were conducted in accordance with national biodiversity and conservation regulations. Permission to conduct all experimental research was also obtained from the institutional authority in compliance with relevant institutional, national, and international guidelines and legislation. The methodology and background of the research were reviewed and approved by the Institutional Ethical Committee of the Faculty of Biological Science, Jashore University of Science and Technology (Approval No. ERC/FBST/JUST/2021-91). The collected leaves were washed with running tap water and air-dried in an air-conditioned room at approximately 25 °C. The dried plant material was then finely pulverized into powder and stored in a tightly sealed container for subsequent experimentation.

Plant extracts preparation

Plant leaves were extracted according to the methods described previously by Rahman et al., with minor changes²⁴. A total of 100 g of powdered plant material was split into four separate 500-mL conical flasks. To each flask, 100 mL of methanol was introduced, and the flasks were then kept within a shaking incubator (JSSI300T, JSR, South Korea), undergoing agitation at 250 rpm for 72 h at 37 °C. At first, the mixture was filtered with cotton gauze and then filtered with Whatman no. 1 filter paper. Using a rotary evaporator (DLAB Scientific Inc., CA, USA), the filtrates were concentrated under vacuum at room temperature to obtain concentrated extracts of M. longiflora leaves. 7.5 g of crude methanol extract, equivalent to 7.5% by dry weight, was obtained from 100 g of the initial powdered plant material.

Bacterial strain collection

A glycerol stock containing the strains of MDR P. aeruginosa (Gene Bank Accession Number: OK355439) and MDR B. cereus is resistant to several antibiotics. This bacterial strain was isolated and identified from wastewater in a medical facility. Its antibiotic susceptibility was assessed by testing it against antibiotics, Ampicillin²⁵. The multidrug-resistant Bacillus cereus was obtained as a glycerol stock from the Department of Biotechnology and Genetic Engineering at Islamic University, Kushtia, Bangladesh. After being separated from the drainage water of a medical center, these strains were tested for antibiotic susceptibility against a variety of antibiotics and were found to be multiple antibiotic-resistant.

In vitro studies

Antibacterial activity assessment by agar-well diffusion assay

The antibiotic potency of MEMLL against MDR P. aeruginosa and B. cereus was evaluated using both agar-well diffusion methods, following the procedures outlined in earlier research²¹. The frozen strains of MDR P. aeruginosa and B. cereus were thawed and plated separately on LB agar medium. The plates were incubated at 37 °C for the growth of bacterial colonies. A selected single colony was inoculated into 25 mL of LB broth and cultured at 37 °C with constant agitation at 250 rpm until the optical density (OD) at 600 nm wavelength reached 0.4, as measured with a Multiskan Sky Microplate Spectrophotometer (Multiskan Sky with Cuvette and Touch Screen, Cat # A51119700C, Thermo Fisher Scientific, USA). 50 µL of the bacterial culture was evenly dispersed on LB agar plates, and four wells were made using a sterilized cork borer. Stock solutions of MEMLL (500 µg/mL) were prepared in methanol and serially diluted to obtain concentrations of 250, 125, and 62.5 µg/mL. Each dilution was added to individual wells on LB agar plates, and a standard ampicillin disc was placed at the center as a positive control. The plates were incubated at 37 °C for 16 h, after which zones of inhibition were measured to evaluate antibacterial activity. The experiment was performed in triplicate for consistency.

MIC and MBC analysis

The minimum inhibitory concentration (MIC) of MEMLL was determined using a two-fold serial dilution method. A stock solution of MEMLL (500 µg/mL) was serially diluted with LB broth in 25 mL glass tubes to obtain final concentrations of 250, 125, and 62.5 µg/mL. Each tube, except the control (containing only bacterial culture), was inoculated with 50 µL of mid-exponential phase bacterial suspension. The tubes were incubated at 37 °C for 24 h, and bacterial growth was assessed by observing turbidity. The lowest concentration of MEMLL, which prevented visible bacterial growth, was identified as the MIC. For the determination of MBC, 50 µL of each bacterial culture from glass tubes was transferred onto LB agar plates separately for subculturing at 37 °C for 16 h. The lowest concentration of MEMLL that completely halted the growth of the bacterial colony on the agar plate was recorded as the MBC. Each trial was carried out three times to ensure accuracy.

Analytical analysis of MEMLL

Qualitative preliminary phytochemical screening

Various standard color change methods were employed to classify phytochemical classes within the MEMLL, as described previously²¹. To identify flavonoids, 25 mg of MEMLL was dissolved in 2.5 mL of methanol and slowly mixed with a 5% NaOH solution, resulting in a strong yellow color in the alkaline reagent test. A few drops of 10% HCl were added to the alkaline solution, and the resulting colorless solution confirmed the presence of flavonoids. In the FeCl3 test, a 5% ferric chloride solution was gradually added to 2.0 mL of a 10 mg/mL MEMLL solution prepared in distilled water, producing a greenish-black or reddish-black coloration that indicated tannins. For the lead acetate test, 2 mL of the same MEMLL solution was combined with 10% lead acetate solution, which led to the formation of a gray-white or creamy gelatinous precipitate, also confirming tannins. Terpenoids and steroids were identified through the Salkowski test, where 10 mg of MEMLL was mixed with 8 mL of chloroform, filtered, and the clear filtrate was split into two separate test tubes for further analysis. Concentrated H₂SO₄ was gently added to the tube edges, yielding a brown-layer at the upper interface/Bluish-brown layer at the interface/formation of a yellow-coloured lower layer, confirming the presence of a terpenoid. In the second tube, concentrated H₂SO₄ was introduced to MEMLL, agitation ensued, and the appearance of a blackish layer at the bottom/dark-black color on the bottom surface/Red color appears in the lower layer signified the presence of the steroid. The presence of saponins in the respective extracts was determined by foaming and frothing tests. In this case, 0.5 mg/mL of MEMLL was combined with 3 mL of distilled water, agitated for a few minutes, and then left to stand. The foam persisted even after standing for ten minutes, indicating the presence of saponins. Fehling’s test entailed mixing 25 mg of MEMLL with equal volumes of Fehling’s solutions A and B, followed by 5–10 min boiling in a boiling water bath. The emergence of a sudden yellowish color/reddish orange precipitate indicated the presence of reducing sugar. In the ninhydrin test, 0.2 g of ninhydrin was dissolved in 10 mL of ethanol, and a few drops of ninhydrin were poured into 1 mL solution (0.5 mg/mL) of MEMLL in distilled water. The sample was heated using a water bath, and the appearance of a violet-blue or purple color confirmed the presence of proteins and free amino acids. In the sodium nitroprusside test, sodium nitroprusside was dissolved in distilled water, then 1 mL of a 0.5 mg/mL MEMLL solution in ethanol was added. After thorough mixing, 5% sodium hydroxide was added drop by drop. The development of a dark brown color indicated the presence of ketones. In the Keller-Kiliani test, 25 mg of MEMLL was added to 4 mL of glacial acetic acid, followed by adding 1 mL of 5% ferric chloride solution along with 1 mL of dilute HCl (1 N); the formation of a brown ring at the interface indicates the presence of cardiac glycosides. In Mayer’s test, 50 mg of MEMLL was dissolved in 5 mL of aqueous 2% HCl in a boiling water bath. Subsequently, the mixture was filtered to obtain a clear portion. In a test tube, 1 mL of Mayer’s reagent was combined with 1 mL of this acid-treated clear solution. The appearance of creamy-colored precipitation/Creamy-white precipitation indicates the presence of alkaloids.

FT-IR spectroscopic analysis

The Fourier-transform infrared spectroscopy (FT-IR) analysis of MEMLL was carried out using a previously established protocol²⁶. Briefly, the plant extract was transformed into a KBr pellet, which was subsequently positioned within the FT-IR sample compartment. The absorption spectrum was captured within the wavenumber spectrum ranging from 4500 to 400 cm^− 1, using a resolution of 4 cm^− 1.

Gas chromatography-mass spectrometry (GC-MS) analysis

GC-MS analysis was performed as previously described²¹. Phytocompounds in MEMLL were identified using a Shimadzu triple-quadrupole GC-MS-TQ8040 system. Helium was used as the carrier gas, and separation occurred on an Rtx-5MS capillary column measuring 30 m in length, 0.25 mm in internal diameter, and with a 0.25 μm film thickness. The oven temperature program started at 50 °C, held for 1 min, then increased to 200 °C for 2 min, and finally raised to 300 °C for 7 min. Throughout the 40-minute investigation, the sample injector temperature was consistently maintained at 250 °C, and one microliter of the sample was injected into the GC column’s injection port at a steady flow rate of 1 mL/min in the spitless mode. The detector was set up with the following parameters: mass spectrometry using Q3 scan mode, scanning a mass-to-charge ratio (m/z) range from 50 to 600, a scan speed of 2000, interface temperature at 250 °C, ion source temperature at 230 °C, a scan time of 0.3 s per scan, and an ionization energy of 70 eV. Identification of metabolites in the phytocompounds was based on comparing retention times and spectral patterns, which were then confirmed by matching with entries in the National Institute of Standards and Technology (NIST) database.

In silico study of antibacterial activity of MEMLL

Protein structure retrieval and preparation

The tertiary structures of the LasR (PDB:2UV0)²⁷, LpxC (PDB:2VES)²¹, FosB (PDB:8G7F)²⁸, and PlcR (PDB:3U3W)²⁹proteins in PDB format were collected from the RCSB protein data bank (https://www.rcsb.org/). The molecular crystal structures of LasR, LpxC, FosB, and PlcR were resolved with a resolution of 1.40 Å, 1.90 Å, 2.40 Å, and 2.04 Å, respectively. Initially, the structure of these proteins was prepared using the default parameter of the protein preparation wizard in the Maestro module of the Schrödinger suite version 2020-3³⁰. In this preparation method, the assigned binding order, creating disulfide bonds, hydrogens, and missing side chains, was added, along with water, and hetero atoms were removed from the respective protein structures. Lastly, the structure of proteins was optimized by applying the OPLS3e force field.

Preparation of ligand

Medicinal plants contain a wide variety of phytochemicals that can be used for the development of new drugs. Thirty-three (33) compounds were obtained in MEMLL during GC-MS analysis. These phytochemicals were retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) in 3D SDF format and prepared using the LigPrep wizard available in the Maestro Schrödinger Suite v11.4. The chemical structures were ionized at the general possible states of 7.0 ± 2.00 pH, using Epik to create tautomers. The chemical structures were further optimized using the OPLS3e force field³¹.

Molecular docking studies

Molecular docking analysis is utilized to identify the binding interactions between the target proteins and specific ligands³². Plant phytochemicals metabolically annotated by GC-MS analysis have been docked with the targeted proteins by applying the Glide v-11.4 and Maestro v-12.5.139 packages from the Schrödinger Suite. The OPLS3e force field was employed in a standard precision mode of docking. The receptor grids were generated using native inhibitors that are complexed with the desired proteins. The thirty-three (33) compounds, along with the control drugs, were subjected to molecular docking with the target proteins (LasR, LpxC, FosB, and PlcR). For LasR, the central organizes of the grid box were set at X = 23.63 Å, Y = 16.67 Å, Z = 79.374 Å with a 26.179 ų grid range, for LpxC, the grid box was set at X = 57.069 Å, Y = 37.55 Å, and Z = 15.97 Å with a 23.64 ų grid range. Whereas the FosB receptor grid box was created at X = – 6.367 Å, Y = 19.502 Å, and Z = – 7.98 Å with a 30 ų grid range. Finally, for the PlcR receptor, the grid box was generated at X = 11.90 Å, Y = 5.67 Å, and Z = 31.358 Å with a 26.179 ų grid range. The interactions between the target proteins (LasR, LpxC, FosB, and PlcR) and ligands were evaluated by calculating binding energies. The different chemicals and ligand-binding residues were figured out by the Maestro viewer from the docking pose file.

Pharmacokinetics (PK) and toxicity analysis

The process of computational drug design and development involves an initial assessment of diverse factors, including physicochemical properties, lipophilicity, water solubility, pharmacokinetics (GI absorption and BBB permeability), adherence to RO5 for drug-likeness, synthetic accessibility, and toxicity for optimizing a molecular candidate to become an effective drug. Thus, a publicly available online tool called the Swiss-ADME server (www.swissadme.ch) was employed to analyze the mentioned features except for the toxicity of the phytocompounds that have multi-targeting potential with a molecular docking score higher than that of the molecular docking score of the control drug ampicillin. In modern drug design, toxicology prediction helps to determine how a specific compound affects humans, animals, plants, or the environment. Therefore, ProTox-II, a freely accessible online server, was used to assess the toxicity of the selected phytocompounds³³.

Molecular dynamics (MD) simulation

MD simulations were used to investigate how the protein-ligand complexes maintained their structural integrity in a particular physiological environment. To evaluate the physical motions and behavior of the compounds selected in the macromolecular environment, 100 ns MD simulations were conducted to examine the stability of protein-ligand complexes. The physical motions of atoms in protein molecules were examined using Schrödinger’s Desmond module (Release 2020-3) operating in a Linux environment³⁰. Simple point-charge (SPC) water molecules were added to each complex’s cubic box, which had dimensions of 10 × 10 × 10 ų, to maintain a constant system volume. Na⁺ and Cl^- ions were randomized into the system to maintain a salt content of 0.15 M. The OPLS3e force field enabled the stabilization and relaxation of the system³¹. During the simulation, the temperature was kept at 300.0 K, and the pressure was kept at 1.01325 bar in the NPT (constant pressure-constant temperature) ensemble. Through the computation of parameters including RMSD, RMSF, Rg, and SASA, the stability and dynamic properties of the complexes were assessed.

Statistical analysis

Antibacterial activity results are presented as the average value followed by the standard deviation (STDEV) from three independent replicates. These experiments were carried out with varying concentrations of MEMLL.

Results

Antibacterial activity of MEMLL

Our results demonstrated that MEMLL inhibited the growth of MDR P. aeruginosa and B. cereus in a concentration-dependent manner, with inhibition zones ranging from 9.666 ± 1.154 to 16.666 ± 1.154 mm for P. aeruginosa (Fig. 2A) and 9.666 ± 0.577 to 15 ± 1 mm for MDR B. cereus (Fig. 2E) when tested at concentrations ranging from 62.5 to 500 µg/mL. Seemingly, the MEMLL exhibited higher antimicrobial activity for MDR P. aeruginosa (Fig. 2B) compared to the antimicrobial activity against MDR B. cereus (Fig. 2F). Interestingly, there was no antibacterial action against MDR P. aeruginosa and MDR B. cereus. shown by the control drug, ampicillin, due to its known resistance, showed no antibacterial activity. The MIC of MEMLL was determined to be 250 µg/mL for both the MDR P. aeruginosa and MDR B. cereus (Fig. 2C,G), whereas the minimum concentration at which they were bactericidal (MBC) was identified as 125 µg/mL for MDR P. aeruginosa (Fig. 2D) and 250 µg/mL for MDR B. cereus (Fig. 2H). It is worth noting that a greater amount of MEMLL was needed for the complete eradication of MDR B. cereus than for the complete eradication of MDR P. aeruginosa, as determined by the MBC assay (Fig.  2D,H).

Fig. 2.

Antibacterial activity of MEMLL against MDR P. aeruginosa and B. cereus. The zones of inhibition by various concentrations (62.5 to 500 µg/mL) of MEMLL against MDR P. aeruginosa (A) and B. cereus (E). MEMLL at various concentrations demonstrated zones of inhibition in millimeters (mm) against MDR P. aeruginosa (B) and B. cereus (F). The MIC in µg/mL against MDR P. aeruginosa (C) and B. cereus (G). The MBC in µg/mL against MDR P. aeruginosa (D) and B. cereus (H).

Preliminary phytochemical screening in MEMLL

The presence of the classes of phytochemicals in MEMLL was determined through color change methods, which serve as a first step in speculating the various types of phytochemicals in MEMLL listed in Supplementary Table S1. In the alkaline reagent test, extracts exhibited a strong yellow color that faded to colorless when 10% HCl was added, confirming the presence of flavonoids. Tannins were identified through the development of a reddish-black coloration upon treatment with 5% FeCl₃ solution and the appearance of a gray-white precipitate with 10% lead acetate solution. Salkowski’s test indicated the presence of terpenoids as a brown-layer at the upper interface and steroids as a blackish layer at the bottom. The MEMLL showed the presence of saponins by the foaming stability on top of the test tube that appeared even after shaking for ten minutes. Fehling’s test showed the presence of reducing sugars by forming a reddish-orange precipitate. The ninhydrin test produced a violet-blue color instead of an intense yellow color, confirming the presence of proteins and free amino acids. The sodium nitroprusside test exhibited a dark-brown color, signifying the presence of a ketone. In the Keller-Killiani test, MEMLL did not show the brown ring at the interface upon successive treatment of MEMLL with glacial acetic acid, 5% ferric chloride solution, and dilute HCl (1 N), indicating the absence of cardiac glycosides. MEMLL did not show a creamy-colored precipitation/Creamy-white precipitation in Mayer’s test, suggesting the absence of alkaloids in the MEMLL.

FT-IR spectroscopic functional group analysis in MEMLL

FTIR spectroscopy aids in the identification of functional groups of phytocompounds with the help of characteristic peak positions in FT-IR spectra. The vibrational mode of a particular bond usually provides a characteristic peak in FT-IR spectra, which helps in the assignment of a functional group of phytochemicals. As shown in Fig. 3, the FT-IR spectra results displayed distinct peaks, confirming the presence of various functional groups (Supplementary Table S2) in MEMLL. Particularly, MEMLL displayed peaks at 3650–3000 for primary and secondary amines or amides (N–H stretch), 2924 and 2851 for aldehyde (C–H), 1733 for ester or ketone (C = O), 1632 for aromatic (C=C), 1383 for amines (C–N), 1166 for alcohols/esters (C–O), 1074 for sulfoxide (S=O), 1050 and 828 for alkyl halide (C-X) and 800–430 for alkyl chloride (C–Cl). In this investigation, a broad peak was identified at approximately 3650 –3000 cm^− 1 signifies the existence of secondary N-H bonds associated with amines or amides. the medium peak was found at 1632 cm^− 1, indicating the presence of an aromatic C=C bond. This is obvious, as most of the phytochemicals contain a lot of aromatic compounds, and the amount of this (aromatic C =C) bond should be much higher than any other bond in the sample. Additionally, evidence of the carbonyl group was substantiated by a peak at 1733 cm^− 1, presented as a broad shoulder, indicative of diverse molecular environments. This observation corresponds to the prevalence of carbonyl compounds in phytochemicals, wherein the abundance of this bond surpasses that of other bonds in the sample. Moreover, the FT-IR spectral analysis indicated the existence of a sulfonamide (S=O) bond, suggesting a high likelihood of sulfonyl compounds in the natural products. The FTIR analysis also identified various components, such as hydrocarbons, alkyl halides, aldehydes, and alcoholic compounds.

Fig. 3.

Fourier transform-infrared spectra (FT-IR) of MEMLL. The peaks at 3650 –3000 cm^− 1 denote primary and secondary amines or amides (N–H stretch), 2924 and 2851 cm^− 1 indicates aldehyde (C–H), 1733 cm^− 1 represents ester or ketone (C=O), 1632 cm^− 1 designates aromatic (C=C), 1383 cm^− 1 defines amines (C–N), 1166 cm^− 1 explains alcohols/esters (C–O), 1074 cm^− 1 explicates sulfoxide (S=O), 1050 and 828 cm^− 1 elucidates alkyl halide (C-X) and 800 –430 cm^− 1 reveals alkyl chloride (C–Cl).

GC-MS analysis of MEMLL

MEMLL showed 33 peaks in the GC-MS chromatogram, and each peak denoted a unique, respective compound (Fig. 4). By comparing average peak areas to total areas with retention time (RT), we calculated their relative percentage values, as represented in Supplementary Table S3. Peaks with retention time (RT) confirm the presence of particular phytochemicals that have distinct identities based on their structure and chemical formula. Chemical compounds were ranked in order of abundance according to their relative concentration in MEMLL, which was determined by calculating the overall peak area (%). The most abundant phytochemicals were 13-Docosenamide, (Z)- (51.16%), phytol (8.51%), 8,11,14-Docosatrienoic acid, methyl ester (6.36%), hexadecanoic acid, methyl ester (5.81%), and methyl 9-cis,11-trans-octadecadienoate (3.01%).

Fig. 4.

GC-MS analysis of MEMLL reveals distinct peaks that correspond to different metabolites found in the extract.

Binding affinity analysis through molecular Docking

Molecular docking is an essential method to determine the fundamental binding mode of a ligand to a macromolecule with a 3D structure in computer-aided drug design and structural molecular biology³⁴. Binding affinity studies were conducted between phytochemicals obtained from GC-MS with LasR and LpxC receptors for P. aeruginosa, and FosB and PlcR receptors for B. cereus. Among 33 phytocompounds, 10 compounds exhibited multi-modal binding affinity with all the target proteins (Supplementary Table S4). Among the compounds, 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334) demonstrated multi-modal activity and a significantly higher negative binding affinity for all targets compared to the control drug, ampicillin (CID 6249). This compound displayed molecular docking interaction-based binding energy score − 7.427 kcal/mol for LasR, – 6.991 kcal/mol for LpxC, – 5.77 kcal/mol for FosB, and − 5.216 kcal/mol for PlcR, whereas ampicillin exhibited docking score − 7.169 kcal/mol for LasR, – 6.491 kcal/mol for LpxC, – 4.805 kcal/mol for FosB, and − 3.341 kcal/mol for PlcR. Since this compound encompasses multi-modal and better binding poses towards all targets than the control drug, ampicillin was selected for further evaluation through a rational drug design process.

Interpretations of 3D, and 2D structure of protein-ligands interaction

The molecular docking interactions of 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334) and control drug ampicillin (CID 6249) with the target proteins (LasR, LpxC, FosB, and PlcR) have been figured by the Maestro module of the Schrödinger suite, depicted in Figs. 5A-D and 6A-D. Different kinds of noncovalent bonds, like hydrophobic, polar, hydrogen bonds, and other bonds were contribute to the binding of ligands to the active pockets of target proteins to make an interaction between protein and ligands. Non-covalent bond formation in drug discovery plays playing crucial role in that ligands don’t alter the structure of the protein, rather they bind with the target protein efficiently. Hydrogen bond formation is an important feature for drug binding and proper metabolization³⁵. As shown in Supplementary Table S5, the selected compound CID 14,334 with the LasR protein formed two hydrogen bonds, four hydrophobic bonds, and several other bonds, while CID 6249 (ampicillin) with the LasR receptor formed no hydrogen bonds, eight hydrophobic bonds, and other bonds. When CID 14,334 interacted with the LpxC receptor demonstrated one hydrogen, six hydrophobic, and other bonds, besides the control drug (CID 6249), which formed five hydrogen, eight hydrophobic, and other bonds. Likewise, CID 14,334 with the FosB receptor exhibited the presence of one hydrogen, three hydrophobic, and other bonds, whereas ampicillin interacted with FosB by forming two hydrogen, three hydrophobic, and other bonds. In the case of PlcR, ampicillin displayed two hydrogen, three hydrophobic, and other bonds, but CID 14,334 showed the presence of three hydrogen, two hydrophobic, and other bonds, which represents a better interaction for the compound 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334).

Fig. 5.

The interaction of 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334) and control drug ampicillin (CID 6249) with the target proteins LasR and LpxC was depicted in both 3D (left) and 2D (right) formats. 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334) bound to the active pocket of LasR (A) and LpxC (C). Control drug ampicillin (CID 6249) interacted with the active pocket of LasR (B) and LpxC (D).

Fig. 6.

The interaction of 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334) and control drug ampicillin (CID 6249) with the target proteins FosB and PlcR was depicted in both 3D (left) and 2D (right) formats. 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334) bound to the active pocket of FosB (A) and PlcR (C). Control drug ampicillin (CID 6249) interacted with the active pocket of FosB (B) and PlcR (D).

Evaluation of molecular features (physicochemical, solubility, pharmacokinetics, drug-likeness) and toxicity of properties of the selected compounds

Pharmacokinetics (PK) is the study of the dynamic motions of foreign substances as they travel through the body, including the kinetics of absorption, distribution, metabolism, and excretion (ADME), which are significantly influenced by their physicochemical properties³⁶. As shown in Supplementary Table S6, the selected hit phytocompound (6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one) exhibited physicochemical properties, including a molecular weight of ≤ 500 g/mol, HBA ≤ 10, HBD ≤ 5, TPSA ≤ 140 Ų, and RB ≤ 11, all within the desirable range, suggesting potential for high oral bioavailability. The optimal rotatable bond (RB) range is 0 to 11, and our hit phytocompound possesses an RB of zero (0), aligning well with favorable absorption. A clogP value of 1 > to ≤ 5 typically signifies strong absorption and solubility, and our hit phytocompound encounters this standard with a clogP value of 1.49. The Log S (ESOL) associated with water solubility, which ideally remains low to enhance drugs’ solubility, falls to an acceptable range of – 4.0 to 0.5 mol/L, suggesting that they can cross the membrane’s bilayer. Our hit phytocompound exhibited a logS value of -1.69 mol/L. The bioavailability of a drug is significantly influenced by human intestinal absorption (HIA); thus, HIA calculations were conducted. Our hit phytocompound exhibited high GI absorption, while the control drug ampicillin displayed low GI absorption. 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one satisfied no violation of Lipinski’s rule of five for drug-likeness and exhibited ease of synthesis within an acceptable range of synthetic accessibility score in medicinal chemistry. Uridine diphosphate glucuronosyltransferase (UGT) is an important enzyme for drug metabolism and the clearance of drugs from the renal system. The lead phytochemicals showed a 0.7 UGT-catalyzed value, which denoted 70% of clearance. However, the control drug ampicillin displayed zero (0) UGT catalyzed value, suggesting 0.0% of clearance. In terms of toxicity, computational assessments of the hit phytocompound showed that it was non-carcinogenic, non-mutagenic, non-immunogenic, and non-cytotoxic. Following these thorough assessments, 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one was selected as the best-hit phytocompound for further evaluation.

Examination of the protein-ligand complexes’ structural stability of the best-hit compound through molecular dynamics simulation

Molecular dynamics (MD) simulation is a powerful technique to observe the conformational changes where ligands and proteins are both permitted to run over a specific period in an artificial environment³⁷. An acceptable value of deviation between protein-ligand complexes is 1–3 Å. A 100 ns simulation was conducted to investigate the flexibility, structural dynamics, and binding interactions of the complex, addressing parameters such as root-mean-square deviation (RMSD), root mean square fluctuation (RMSF), the radius of gyration (Rg), solvent-accessible surface area (SASA), and protein-ligand contact analysis.

RMSD analysis

The root mean square deviation (RMSD) is utilized to calculate the difference between a protein’s backbone from its original structural conformation to its final conformation. The stability of the proteins (LasR, LpxC, FosB, and PlcR) with the best-hit phytocompound (CID 14334) and control drug ampicillin (CID 6249) complex structures was assessed by measuring the RMSD of the Cα atoms during the 100 ns simulation. For the LasR receptor, the apoprotein (Cα-RMSD), the LasR-CID 14,334 complex showed average RMSD values of 1.59, 1.76 Å, respectively, while the LasR-CID 6249 complex depicted an RMSD value of 1.88 Å, indicating the structural stability of the hit-phytochemical, as demonstrated in Fig. 7A. In the context of LpxC, the apoprotein (Cα-RMSD), the LpxC-CID 14,334 complex, and the LpxC-CID 6249 complex displayed mean RMSD values of 1.30 Å, 1.48 Å, and 1.41 Å, respectively. Notably, none of these compounds showed significant variations during the study, remaining in an equilibrium state throughout the simulation time (Fig. 7B). In the case of FosB, the mean RMSD values of apoprotein (FosB), FosB-CID 14,334 were 4.03 Å and 4.85 Å, respectively, while FosB-CID 6249 (control) was calculated at 4.55 Å value, which demonstrates the structural stability with minor fluctuation of FosB-CID 14,334 compared with the native protein (Fig. 7C). Similarly, for PlcR-CID 14,334 and CID 6249 (ampicillin) complexes as depicted in Fig. 7D, the range was 1.04–5.46 Å and 1.29–5.82 Å, whereas the apo protein showed a range of 1.07–9.16 Å, indicating better stability of the selected phytocompound compared to the PlcR-CID 6249. The average value should be included, and the conclusion should be written. In the case of the best-hit phytocompound, the deviation indicated a level of equilibration comparable to that of the apoprotein and control drug ampicillin. This led to the compound’s stability within the protein’s binding site for a significant portion of the simulation time, exceeding the stability observed in the apoprotein and protein-control drug complex.

Fig. 7.

The extracted RMSD and RMSF values from the Cα atoms of the target protein, selected ligand docked complex during a 100 ns MD simulation. (A–B) (A–D) The RMSD of the selected hit phytocompound or control drug ampicillin in complex with the target proteins (LasR, LpxC, FosB, and PlcR). (E–H) The RMSF of the selected hit phytocompound or control drug ampicillin in complex with the target proteins (LasR, LpxC, FosB, and PlcR). In both cases of RMSD and RMSF, the selected hit phytocompound or control drug ampicillin in complex with the target proteins is represented by a red and green color, respectively.

RMSF analysis

The Root Mean Square Fluctuation (RMSF) analysis can help to characterize and determine the residual changes within the protein chain during the interaction of compounds with the specific amino acid residues. The RMSF value of CID 14,334 and CID 6249 in complex with the selected receptors, LasR, LpxC, FosB, and PlcR, was analyzed to observe the alterations in the structural flexibility of the protein during attachment to specific amino acid residue positions. The RMSF value of CID 14,334 was calculated for LasR, LpxC, FosB, and PlcR compared to the apoprotein and control drug (ampicillin), illustrated in Fig. 7E-H. The CID 14,334, CID 6249 (ampicillin), and apoprotein in complex with LasR exhibited an average fluctuation of 0.87 Å, 0.73 Å, and 0.77 Å, respectively (Fig. 7E). In the LasR receptor, the SER14 (2.09 Å), ASP43 (2.33 Å), and GLU168 (3.01 Å), residual positions have large peaks when forming complex with the hit compound CID 14,334 and the control ligand CID 6249 during simulation time. The average RMSF values for the CID 14,334, and apoprotein were 0.74 Å, and 0.77 Å, respectively for the LpxC protein while the control drug (ampicillin) showed a 0.77 Å of average fluctuation presenting the lower RMSF values with minor fluctuation of hit phytochemical compared to the control compound CID 6249, as depicted in (Fig. 7F). In the case of FosB, the average fluctuation for the CID 14,334 and apoprotein were 1.45 Å and 1.76 Å, whereas CID 6249 (ampicillin) exhibited an average fluctuation of 1.40 Å, respectively. The hit compound CID 14,334 and the control ligand CID 6249 showed higher pick at ARG35 (1.86 Å), CYS 43 (1.76 Å), GLN62 (2.42 Å), LYS74 (2.30 Å), VAL98 (2.47 Å), and GLN122 (2.76 Å) residues when forming a complex with FosB receptor (Fig. 7G). In the PlcR receptors, an average RMSF value of 1.68 Å and 3.38 Å was observed for the CID 14,334 and apoprotein, in contrast to the control (ampicillin), which showed an average fluctuation of 1.63 Å, respectively. The higher peaks at the GLY18 (11.25 Å), SER32 (8.84 Å), GLN57 (8.12 Å), ILE149 (4.61 Å), and ALA187 (4.33 Å) residual fluctuations have been shown when the hit compound CID 14,334 and the control ligand CID 6249 formed a complex with the PlcR receptor (Fig. 7H). These findings suggested that the selected hit phytocompound complexes with the receptors fluctuated in an optimal range with their amino acid positions.

Radius of gyration analysis

The radius of gyration (Rg) provides valuable insights into the compactness and characterizes how atoms are distributed around the axis in a protein-ligand complex system. Thus, the stability of CID 14,334 or CID 6249 in complex with the LasR, LpxC, FosB, and PlcR receptors was calculated based on their Rg values to a 100 ns simulation run time, as shown in Fig. 8(A–D). The lead compound (CID 14334) in complex with LasR, LpxC, FosB, and PlcR receptors showed lower Rg values with minor fluctuations, while CID 6249 (ampicillin) exhibited a higher average Rg value with increased fluctuations, which demonstrates better stability, significantly improved compactness of the hit-phytocompound CID 14,334 than the control compound ampicillin in complex with LasR, LpxC, FosB, and PlcR receptors.

Fig. 8.

The MD simulation of the selected hit phytocompound or control drug ampicillin and target protein docked complexes over a 100 ns duration regarding the trajectories of the radius of gyration (Rg) and solvent accessible surface area (SASA). A–D The Rg value of the selected hit phytocompound or control drug ampicillin in complex with the target proteins (LasR, LpxC, FosB, and PlcR). E–H The SASA value of the selected hit phytocompound or control drug ampicillin in complex with the target proteins (LasR, LpxC, FosB, and PlcR). In both cases of Rg and SASA, the selected hit phytocompound or control drug ampicillin in complex with the target proteins is represented by a red and green color, respectively.

SASA analysis

Protein surfaces typically contain amino acid residues that provide functional sites and intermingle with other drug-like compounds to provide insight into the solvent-like behavior (hydrophilic or hydrophobic) of molecules and proteins. In this study, the hit-phytocompound CID 14,334 in complex with LasR, LpxC, FosB, and PlcR proteins demonstrated average SASA values of 9.27 Ų, 43.03 Ų, 183.88 Ų, and 61.52 Ų respectively whereas, the control compound CID 6249 (ampicillin) showed an average SASA values of 37.62 Ų, 162.86 Ų, 421.61 Ų, and 70.62 Ų respectively and indicating extensive surface area exposed to the solvent in complex with LasR, LpxC, FosB receptors of hit-phytocompound CID 14,334 displayed compared to the control compound CID 6249 (ampicillin), as represented in Fig. 8E–H. This suggests that there was effective exposure of amino acid residues to the identified compound in the complex systems.

Protein-ligand contact analysis

Protein interactions with the selected ligands and their intermolecular interactions were assessed for a 100-ns simulation time using the simulation interactions diagram. Hydrogen bonds, ionic bonds, water bridge bonds, and non-covalent (hydrophobic) bonds are the crucial bonds for the ligand binding with the respective protein’s active pocket³⁵. Therefore, the interactions between the target proteins and the best-hit phytocompound (CID 14334) or control drug ampicillin (CID 6249) in complex with the LasR, LpxC, FosB, and PlcR were analyzed and visually represented in Fig. 9A–H. The stacked bar charts for the complex of best-hit phytocompound (CID 14334) and target protein LasR are shown in Fig. 9A, with an interaction fraction value (IVF) of a maximum of 1.118 and 0.5 at the residues of ARG61 and TYR64, which make contact by using hydrogen-water bridges and hydrogen-hydrophobic-water bridges, respectively. In the complex of LpxC and CID 14,334 (Fig. 9B), the IFV 0.6 at PHE 193 and 1.2 at ASP 196 was formed by hydrogen and water bridge bonds, indicating 60%and 120%interactions with LpxC receptors. Whereas in the FosB and CID 14,334 complex, multiple interactions such as hydrogen-water bridges at the position of PHE 10 (0.125), GLH 16 (0.135), and LYS 113 (0.140), and a hydrogen-hydrophobic bond at TRP 46 (0.200) and hydrophobic-water bridge bond at ALA 48 (0.175) were formed (Fig. 9C). Similarly, in the complex of PlcR and CID 14,334 (Fig. 9D), multiple interactions, hydrogen-water bridges at the position of GLN 118 (0.4), hydrophobic-water bridges at the position of TYR 156 (0.57), and hydrogen-ionic-water bridges at the position of ASN 159 (0.4) were deliberated. On the other hand, CID 6249 (ampicillin) in complex with LasR, LpxC, FosB, and PlcR proteins established connections through hydrogen-water bridge, hydrogen-hydrophobic-water bridge, and hydrogen-ionic-water bridge interactions (Fig. 9E–H. These results suggest that there are a variety of interactions at play significant roles in stabilizing the binding among the target proteins and the designated compounds.

Fig. 9.

The bar graph illustrates the interactions among the target proteins and designated ligands over a 100 ns simulation. Here, showcasing the interactions of CID 14,334 or the control drug ampicillin with LasR and LpxC in Pseudomonas aeruginosa, as well as FosB and PlcR in Bacillus cereus. These interactions are denoted as (A) CID 14,334-LasR, (B) CID 14,334-LpxC, (C) CID 14,334-FosB, (D) CID 14,334-PlcR, (E) CID 6249-LasR, (F) CID 6249-LpxC, (G) CID 6249-FosB, and (H) CID 6249-PlcR.

Discussion

P. aeruginosa and B. cereus are emerging opportunistic human pathogens capable of causing life-threatening acute and chronic diseases in immunocompromised individuals^38,39. Current treatment strategies often require a combination of antibiotics, but the side effects and rise of drug resistance highlight the need for natural antibacterial drugs⁴⁰. Within this challenge in biomedical science, there is a growing interest in exploring phytocompounds as prospective antibacterial agents.

While earlier research has exposed the antimicrobial potential of extracts against various pathogens, there was a lack of information concerning their efficacy against MDR bacteria, particularly MDR P. aeruginosa and MDR B. cereus. Therefore, our study focused on examining the lethal effect of M. longiflora L. leaf extract and its phytocompounds on MDR P. aeruginosa and MDR B. cereus. Since LasR, in connection with LpxC, the bacterial outer membrane-forming enzyme in P. aeruginosa, as well as FosB and PlcR in B. cereus, play a pivotal role in biofilm formation and subsequent antibiotic resistance, we evaluated the in vitro antibacterial activity of MEMLL against MDR P. aeruginosa and MDR B. cereus. Subsequently, we assessed the in silico antibacterial effect of the phytocompounds in MEMLL with a precise emphasis on targeting LasR, LpxC, FosB, and PlcR.

Since we observed concentration-dependent in vitro antibacterial activity of MEMLL against MDR P. aeruginosa and MDR B. cereus, we metabolically annotated the MEMLL using GC-MS and identified 33 phytochemical metabolites. These metabolites were subjected to molecular docking analysis, which revealed 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334) as the hit phytocompound. This compound exhibited a higher negative binding affinity with all target proteins (LasR, LpxC, FosB, and PlcR) compared to the control drug ampicillin. Consistent with our findings, several previous studies have conducted molecular docking analyses of 82 phytochemicals from Cassia occidentalis L. and 51 phytochemicals from Christella dentata (Forssk.) Brownsey & Jermy, targeting LasR and LpxC. These studies found that CID 102,861, CID 136,654, and CID 7150 in Cassia occidentalis L., and CID 536,446 and CID 7734 in Christella dentata (Forssk.) Brownsey & Jermy have higher negative binding affinities towards all target proteins (LasR and LpxC) compared to the native inhibitor ligands C12-HSL for LasR, BB-78,485 for LpxC, and the control drug ampicillin^21,26.

In our subsequent identification of the best-hit phytochemicals obtained from the molecular docking analysis, we found that 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334) exhibited favorable physicochemical, lipophilicity, water solubility, pharmacokinetics, drug-likeness, medicinal chemistry, and toxicity properties. Therefore, we ensured it was a best-hit phytochemical and a potential candidate for an antibacterial agent. This led to its selection for additional MD simulation, projecting it to be an attractive lead compound for antibacterial treatment against MDR P. aeruginosa and MDR B. cereus. Similarly, it has been reported that methyl dihydrojasmonate (CID 102861), methyl benzoate (CID 7150), and 4a-methyl-4,4a,5,6,7,8-hexahydro-2(3 H)-quinazolinone (CID 136654) in Cassia occidentalis L., as well as bicyclo[4.3.0]nonane, 2,2,6,7-tetramethyl-7-hydroxy- (CID 536446), and 1,4-diethylbenzene (CID 7734) in Christella dentata (Forssk.) Brownsey & Jermy were identified as the best-hit compounds against LasR and LpxC in MDR P. aeruginosa, followed by MD simulation to identify the lead compound^21,26.

In all trajectories of the MD simulation, CID 14,334 exhibited stable binding in the active pockets of LasR, LpxC, FosB, and PlcR receptors over a 100 ns simulation period, indicating its potential as a multi-targeting antibacterial lead phytochemical candidate against MDR P. aeruginosa and B. cereus. Following this, research has not only identified lead phytochemicals (CID 102861, CID 7150, CID 136654, CID 536446, and CID 7734) against LasR and LpxC in P. aeruginosa^21,26but also revealed lead phytochemicals (CID 70825, CID 25247358, CID 54685836, and CID 1983) against the main protease (Mpro) of SARS-CoV-2 for combating COVID-19⁴¹ and lead chemicals (CID 95842900, CID 137030374, CID 124958150, and CID 110126793) against phosphorylated RET (pRET) tyrosine kinase for the treatment of various cancers⁴².

Our identified lead phytochemical, 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334), falls within the category of carbonyl compounds or ketones. Following our results, it has been reported that styryl 3-chloro-4-nitrophenyl ketones display antibacterial activity against both Gram-positive (Bacillus subtilis, Staphylococcus aureus, and Micrococcus luteus) and Gram-negative (Escherichia coli and Pseudomonas aeruginosa) bacteria⁴³. Our findings also align with previous research demonstrating that 4a-Methyl-4,4a,5,6,7,8-hexahydro-2(3 H)-naphthalenone (CID 136654), also classified as a carbonyl compound or ketone, found in the ethyl acetate extract of Cassia occidentalis L. leaves, inhibited LpxC, a key enzyme that catalyzes the synthesis of lipopolysaccharides (LPSs) in the bacterial cell wall and cell membrane, and prevented the LasR receptor from binding to the quorum-sensing signaling molecule acyl homoserine lactone, responsible for cell-to-cell signaling in Gram-negative bacteria in Pseudomonas aeruginosa²¹. In accordance with our research, 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one has also been identified in the ethanol extract of the algae Sargassum horneri and has demonstrated anti-inflammatory potential, both for the extract and the compound itself, against LPS-induced inflammation⁴⁴. This indicates its expanded pharmacological potential in addition to its antibacterial activity. Our hit phytocompound exhibited a logS value of -1.69 mol/L associated with an acceptable range of water solubility (-4.0 to 0.5 mol/L), suggesting its crossing of the membrane’s bilayer. In agreement with our results, it has been reported that carbonyl compounds or ketones (Mannich ketones), due to their water solubility, can impair the cell membrane of bacteria and have reported antibacterial activity against Gram-negative bacteria (Pseudomonas aeruginosa, Escherichia coli) and Gram-positive bacteria (Staphylococcus aureus, Staphylococcus saprophyticus, Micrococcus luteus, and Bacillus subtilis). Moreover, it has been reported that 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one in the aerial parts of Origanum ehrenbergii Boiss. has antioxidant activity⁴⁵.

We evaluated the in vitro antibacterial activity of the extract to support the in silico antibacterial activity. Similarly, the ethyl acetate extract of Ruellia prostrata Poir., containing 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one, along with other phytocompounds, exhibited dose-dependent in vitro antibacterial activity against both Gram-positive bacteria (Bacillus cereus ATCC 14579, Bacillus infantis ATCC 15697, and Exiguobacterium sp. ATCC BAA1283/AT1b) and Gram-negative bacteria (Escherichia coli IFO 3007, Pseudomonas aeruginosa ATCC 27853, and Acinetobacter lwoffii ATCC 15309)⁴⁶. Considering all, 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one has emerged as the most promising lead phytochemical for the development of antibacterial therapy targeting MDR P. aeruginosa, B. cereus, and related infectious diseases.

Despite the encouraging in vitro and in silico results, this study has some limitations. Phytochemical characterization was performed using GC-MS, which primarily identifies volatile and semi-volatile compounds; therefore, certain non-volatile phytochemicals that could be better resolved using LC-MS may not have been detected. In addition, the antibacterial potential of the selected phytochemical has not yet been validated in in vivo models. Future studies incorporating LC-MS-based profiling and in vivo investigations are required to further substantiate the therapeutic potential of the identified lead compound.

Conclusion

In this investigation, the leaves of M. longiflora showed antibacterial activity against MDR P. aeruginosa and B. cereus, as demonstrated by the inhibitory zones in the agar well diffusion assay. The lead compound, 6-Hydroxy-4,4,7a-trimethyl-5,6,7,7a-tetrahydrobenzofuran-2(4 H)-one (CID 14334), was identified through in silico studies. This compound inhibited LasR of P. aeruginosa, FosB and PlcR of B. cereus, which are pivotal signaling receptors responsible for biofilm formation, virulence, and multidrug resistance in both P. aeruginosa and B. cereus. Additionally, it inhibited LpxC of P. aeruginosa, a key enzyme essential for the biosynthesis of lipid A, a vital component of the bacterial outer membrane in Gram-negative bacteria. These results suggest the potential for developing new natural plant-based inhibitors to combat antibiotic-resistant infections. Further, in vivo experiments are required to confirm our findings.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

MEMLL

Methanolic Extract of Mirabilis longiflora leaves

MDR

Multidrug resistance

FT-IR

Fourier-transform infrared spectroscopy

PPS

Preliminary phytochemical screening

GC-MS

Gas chromatography-mass spectrometry

MIC

Minimum Inhibitory Concentration

MBC

Minimum Bactericidal Concentration

OPLS3e

Extending Force Field Coverage for Drug-Like Small Molecules

TIP3P

Transferable intermolecular potential with 3 points

ADME

Absorption, Distribution, Metabolism, and Excretion

SID

Simulation Interaction Diagram

RMSD

Root mean square deviation

RMSF

Root mean square fluctuation

Rg

Radius of gyration

SASA

Solvent accessible surface area

Author contributions

Shahina Akhter: investigation, methodology, formal analysis, writing-original draft. Md. Enamul Kabir Talukder: investigation, data curation, methodology, formal analysis, writing-original draft, writing-review & editing. Md. Tarikul Islam: investigation, methodology, formal analysis, writing-original draft, writing-review & editing. Md. Baha Uddin: investigation, methodology, formal analysis. Nafis Fuad Shahir: investigation, methodology, formal analysis. Nazia Islam Rafi: investigation, methodology, formal analysis. Sadia Israt: investigation, methodology, formal analysis. Rahat Alam: resources, software, writing-review & editing. Mohamad Abu Hena Mostofa Jamal: resources, writing-review & editing. Md. Mashiar Rahman: conceptualization, resources, data curation, validation, supervision, funding acquisition, writing-review & editing.

Funding

This research received no external funding.

Data availability

Available upon reasonable request from the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.