The molecular structure-optimized coordinates were performed utilizing Gaussian 16 software (Frisch et al., 2016.). The study utilized the def2TZVP basis set and M06-2X functional, both known for their high-quality theoretical approaches, and used solvent model density (SMD) for solvent modeling (Zhao and Truhlar, 2008; Hill, 2013). The software programs Gaussian View 6 and Origin 8 were utilized to construct the graphical representations. The frontier molecular orbitals (FMOs) (Bulat et al., 2004) were plotted in order to estimate the band gap energy. Additionally, the wavefunction input files for the title compounds were obtained via Multiwfn 3.8 wavefunction analyzer software (Lu and Chen, 2012a). The analyses of electrostatic potential (Murray and Politzer, 2011) and Fukui functions (Fukui, 1982) were also completed using Multiwfn 3.8. The VMD 1.9.1 application was utilized to generate all isosurface maps (Humphrey, et al., 1996) and Multiwfn output files.

Molecular docking was used for the predictive combination mode of levofloxacin and andrographolide with the protein AlgW. According to Li et al. (2021), the structure of AlgW is from the Protein Data Bank (https://www.rcsb.org/), and its PDB code is 7CO2. Afterward, levofloxacin and andrographolide were downloaded from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) in. sdf format, and CIDs of levofloxacin and andrographolide are 149096 and 5318517, respectively. Levofloxacin was taken as a reference antibiotic in this study. All molecular docking operations were completed using Schrödinger 2022 software. First, the low-energy 3D structure of ligands was generated using the LigPrep panel (Mahto et al., 2014). Next, the protein preparation wizard module was used to take AlgW from its original state to the proper state for calculations, such as assigning bond orders, adding hydrogens, and H-bond assignment (Rolta et al., 2021). The protein binding sites were found by the sitemap panel. Finally, the Glide Ligand Docking Panel was used to run docking jobs using prepared proteins (Ashfaq et al., 2023).

Molecular dynamic (MD) simulation monitored the motions of macromolecules over time (Hildebrand et al., 2019). The protein–ligand complex was simulated in the CHARMM36 force field and implemented using GROMACS 2020.6 software (Abraham, et al., 2015).

In order to generate the topology of compounds and force field parameters, the CGenFF server was used. The protein was put into a cube box of water of the simple point charge (spc216) model, and the distance between the protein peripherals and the box edge was 1 nm. Na⁺ and Cl⁻ions were added to neutralize the charge on the complex and maintain its electrical neutrality. Then, the energy of the MD system was minimized by steepest descent minimization. The system temperature was stabled by the NVT (number of particles, volume, and temperature) ensemble, and the system pressure was stabled by the NPT (number of particles, pressure, and temperature) ensemble. Finally, position restraints were released, and production MD was run to collect data at 100 ns (Hou et al., 2023).

The overnight PAO1 was diluted to a final optical density (OD₆₀₀) of 0.05 in LB broth. The cell suspensions were taken every hour, and the cell density was determined by measuring the absorbance value at 600 nm using a microplate reader to monitor the growth of bacteria.

As previously mentioned, biofilm formation was quantified using crystal violet staining (Scoffield et al., 2017). In brief, overnight PAO1 was 100 times diluted in brand-new M63 minimal medium that has been enhanced with arginine and magnesium sulfate and then incubated for 12–16 h at 37 °C. A volume of 1 mL of diluted cell suspension was transferred into a 24-well plate, and different concentrations of andrographolide and levofloxacin were added, followed by incubation for 48 h. Only untreated cultures that contained 0.625% DMSO were used as a negative control. After incubation, wells were washed with sterilized phosphate-buffered saline (PBS) three times to remove the M63 minimal medium. The adhered cells were stained with 0.05% crystal violet for 30 min at room temperature, followed by washing the wells three times with PBS. Ethanol (95%) was used to elute the biofilm, and the absorbance of the eluted solution was determined at 570 nm using a microplate reader. The assay for crystal violet was measured three times. The difference between the three mean values was calculated using the one-sample t-test; p < 0.01 was considered statistically significant.

In order to understand the molecular chemical reactivity, molecular stability, electronic characteristics, and optical properties of molecules, the frontier molecular orbitals [highest occupied molecular orbital (HOMO)–least unoccupied molecular orbital (LUMO)] (Bredas, 2014) are often employed. When the ΔE value is large, the HOMO–LUMO energy gap is less, which suggests that the chemical reactivity is lower (Gopalakrishnan et al., 2014). Band gap energies are computed, and the HOMO–LUMO is given an estimate using a def2TZVP functional based solvation methodology for several polar solvents (DMSO and water).

In order to acquire a greater comprehension of the chemical properties of the title’s molecules, the electronic parameters were investigated. These data were obtained by conceptual DFT analysis from the Multiwfn 3.8 program. Based on the data presented in Table 1, it can be observed that both the HOMO and LUMO, together with their adjacent orbitals, possess negative energies, which suggests that the molecule is in a stable state (Zhou and Parr, 1990). Figure 1 illustrates the energy gap in the gas phase. The title’s compounds have energy gaps, with values of 749.69 and 622.01 kJ/mol, respectively. A positive reflection of its molecular electrical transport capabilities was observed in the levofloxacin compound’s ΔE at lower values. In addition, the electron cloud primarily associated with levofloxacin is located on the carbon–carbon bond of the ring in the LUMO that corresponds to π→π*. Additional information and details can be found in Supplementary Table S1.

There are many quantum chemical parameters in Koopmans’s theory (Luo et al., 2006). The following parameters are shown in Table 1: ionization energies (I), electron affinities (A), electronegativity (χ), chemical hardness (η), chemical softness (ζ), chemical potential (μ), and the global electrophilicity index (ω). According to Table 1, the chemical hardness (η) value (Parr and Pearson, 1983), the gas phase exhibits the largest degree of andrographolide, suggesting a greater level of chemical stability and reactivity. Despite this, when dissolved in water, we achieved the highest chemical softness measurement (2.809), indicating that levofloxacin exhibits greater reactivity in this specific solvent. Furthermore, the title compounds were evaluated for solvent effects, revealing that polar solvents (water and DMSO) could decrease the chemical hardness (η) value, which is much lower than in the gas phase.

Furthermore, when comparing the energy gap, we found that the levofloxacin compound had the lowest value at 6.306 eV in a water solvent. The molecules are harder when the HOMO–LUMO energy gap is larger (Chattaraj and Maiti, 2003). Conversely, it exhibits a negative correlation with chemical reactivity and softness. Theoretical parameter data have shown that the andrographolide is more reactive in polar solvents, but it may not exhibit higher biological activity than the levofloxacin compound.

The higher the absolute value of the Glide GScore, the stronger the potential affinity between the molecule and protein. The Glide GScores of levofloxacin and andrographolide with the protein AlgW were −8.030 and −3.261, respectively. The Glide GScore showed that the affinity between levofloxacin and protein binding is greater than that of andrographolide. Figure 4; Supplementary Table S4 show andrographolide (Labels a1 and a2) and levofloxacin (Labels b1 and b2) binding conformations of AlgW. It showed that around the active pocket, small molecules mainly interact with amino acid residues of proteins at the distance of 4 Å through hydrogen bonding and van der Waals forces.

The 3D docking diagram of andrographolide and levofloxacin with AlgW showed that AlgW could be attached to the active pocket of the protein. Although the absolute value of the Glide GScore between andrographolide and AlgW is smaller than that between levofloxacin and AlgW, andrographolide is a potential inhibitor of AlgW due to the Glide GScore of −3.261.

The molecular dynamics evaluation parameters for andrographolide–AlgW (red line) and levofloxacin–AlgW (black line) complexes are shown in Figure 5. RMSD can quantify the stability of protein–ligand complex structures over a certain period of time (Rudnev et al., 2022). In this study, the RMSD of andrographolide–AlgW and levofloxacin–AlgW were calculated within 100 ns. The RMSD analysis showed that the RMSD value of levofloxacin-AlgW was balanced after 20 ns. However, the RMSD value of andrographolide-AlgW began to stabilize after 10 ns. It was in equilibrium between 10 ns and 90 ns. After 90 ns, there was a slight increase in the RMSD value. The gyration radius reflected the volume and shape of the complex. The larger the turning radius of the same system, the more expansive the system becomes (Yanao et al., 2007). It showed that levofloxacin–AlgW was more expansive than andrographolide–AlgW. This study analyzed the variation in the number of hydrogen bonds between ligands and proteins over time. In molecular dynamics, the hydrogen bond number of andrographolide–AlgW was not better than that of levofloxacin–AlgW. RMSF was used to calculate the root mean square fluctuations of atomic positions in trajectories (Thangavel and Albratty, 2023). The results indicated that there was little fluctuation in the atomic positions of amino acids around the docking site of the AlgW protein. The molecular dynamics results indicated that in a 100 ns simulation, the binding of andrographolide AlgW and levofloxacin AlgW complexes was stable and persistent.

In order to explore the effect of P. aeruginosa growth with different concentrations of andrographolide, the bacteria were grown in Luria–Bertani broth (LB) agar medium for 24 h, and the results are summed up in Figure 6. The results showed that there is no inhibitory effect on bacterial growth compared with the wild type with 0.1 μg/mL, 10 μg/mL, or 100 μg/mL of andrographolide. In addition, it is reasonable to explain that andrographolide does not directly sterilize the bacterial like antibiotics, and it gave us some clues that andrographolide may have effects on components like biofilm, virulence, quorum sensing (QS) systems, or secretion systems.

Although there is no evidence to prove the activity of andrographolide against P. aeruginosa growth assay, previous studies have shown that andrographolide and 14-deoxy-11,12-didehydroandrographolide may influence the formation of biofilm in P. aeruginosa (Majumdar et al., 2020). To determine whether andrographolide will inhibit the biofilm of P. aeruginosa, we compare the wild-type P. aeruginosa (PAO1) with different concentrations of andrographolide and positive control levofloxacin, which can inhibit the formation of biofilm, by staining adherent cells with crystal violet to investigate biofilm formation (Banerjee et al., 2017a). Figure 7 shows that andrographolide with different concentrations has some effect on biofilm formation, although it is less effective than levofloxacin at the same concentration.