The ethyl acetate fraction (8 g) containing QA as a major compound was chromatographed over silica gel (300 g) using H₂O:methanol as a solvent system of increasing polarity. The H₂O:methanol fractions (30:70) and (20:80) showed the presence of three compounds having similar Rf values. These fractions were mixed (3.95 g) and further subjected to column chromatography using silica gel (200 g) and butanol:acetic acid:water (B:A:W) as a solvent system of increasing polarity (El-Bassossy, 2022) (Supplementary Table S1, S2). QA was isolated in pure form in two fractions of B:A:W, 3:1:1 and 4:1:0, and its structure was confirmed using spectroscopic techniques.

The release of bacterial cell constituents into the supernatant was tested according to the reported method (Lv et al., 2011). The cells from the bacterial cultures (100 mL) were collected by centrifugation at 5,000 rpm for 10 min. They were washed three times with double-distilled water and resuspended in phosphate buffer saline (0.1 M, pH 7.0). Then, the bacterial cell suspensions (100 mL) were incubated at 37°C for 1 h in the presence of each treatment at its MIC value. The samples (2 mL) were centrifuged at 12,000 rpm for 2 min, and the supernatant (1 mL) obtained in each case was studied for absorption under a UV–visible spectrophotometer (Shimadzu) at 260 nm to determine the release of bacterial cellular contents. A correction was made for the absorption of the suspension with phosphate-buffered saline (PBS; pH 7.0) containing the same concentration of compounds after 2 min of contact with the tested bacterial strains along with the untreated cells taken as the control.

The synergistic interaction of QA and the most effective derivative with the standard antibiotic streptomycin was determined using the checkerboard method (Pourkhosravani et al., 2021). On a 96-well microtiter plate, each treatment was diluted in the well columns, and the standard antibiotic was diluted in rows at two-fold dilution of MIC ranging from MIC to 1/16 of MIC. Briefly, each treatment (50 µL) at MIC was mixed with streptomycin (50 µL) at MIC with serial two-fold dilution in Muller–Hinton broth. After that, the log-phase bacterial (100 µL) inoculum was added to all the wells and incubated at 37°C for 24 h. The MICs of each treatment and its combinations were recorded. The results were expressed in the fractional inhibitory concentration index (FICI) as follows: FICI = FIC A + FIC B FIC_A = MIC of treatment in combination/MIC of treatment alone.

FIC_B = MIC of streptomycin in combination/MIC of streptomycin alone.

FICI ≤0.5 shows synergistic interaction; 0.5 < FICI ≤1 shows an additive effect; 1 < FICI ≤4 reveals indifference or no effect; and FICI ≥4 shows antagonistic effects (Gutierrez et al., 2009).

The antibiofilm activity of the most effective treatment was studied using the crystal violet method (Sanchez et al., 2016) to check the mechanism of antibacterial activity by estimating the specific biofilm formation (SBF) through UV spectroscopy.

The nutrient broth (5 mL) was added to a 96-well microtiter plate along with 50 μL of OD₅₆₀ = 0.02 (10⁶ CFU/mL) microbial culture followed by the addition of the most effective treatment at its MIC, and 25%, 50%, and 75% of the MIC value. The plate containing streptomycin served as the standard, while the one without treatment served as the control. The plates were further incubated at 37°C for 48 h without disturbance, allowing for ring formation. After incubation, the supernatant was removed very carefully with the help of a micro-pipette, and each well was washed thoroughly with distilled water (three times) to remove planktonic or unadhered cells. The plates were then dried in air for half an hour. Thereafter, 5 mL of crystal violet dye (1% aqueous solution) was added to the plates and kept at room temperature for 15 min. After that, the solution was discarded, followed by gentle washing with distilled water (three times) to remove the excess stain. Then, 200 μL of ethanol (95%) solution was added to each well and incubated for 15 min. After that, the absorbance was recorded at 570 nm using an ELISA plate reader (Thermo Scientific, Multiscan Go). The specific biofilm formation was calculated as follows: SBF = AB −  CW / G , where SBF is the specific biofilm formation.

AB is the OD at 570 nm of the attached and stained bacteria.

CW is the OD at 570 nm of the stained control wells containing only the bacteria-free medium.

G is the OD at 630 nm of cell growth in broth (Niu and Gilbert, 2004).

The SBF values were classified into three categories.

Strong biofilm: SBF> 2.00.

Intermediate biofilm: SBF between 1 and 2.

Weak biofilm SBF <1.00.

All the results were expressed in terms of mean ± standard deviation with three replications. Two-way ANOVA followed by Tukey’s B test was carried out for the investigation of antibacterial activity data, and p < 0.05 was accepted as significant.

QA was found to be the major compound present in the ethyl acetate fraction of the methanol extract of citrus fruits dropped in April, as revealed by GC-MS analysis (Supplementary Figure S1, S2). QA has also been reported to be present in the dropped immature fruits of different citrus species during the early stages of development (Albertini et al., 2006; Marrubini et al., 2015). Various chlorogenic acids (esters of caffeic acid and quinic acid) have been isolated from the ethyl acetate extract of Zanthoxylum bungeanum leaves (Chang et al., 2018). It was isolated from the ethyl acetate fraction by column chromatography as a white solid with a melting point in the range of 160–162°C. It was derivatized into QA₁, QA₂, and QA3. The structures of QA and its derivatives were confirmed using spectroscopic techniques (Supplementary Figure S3–S14) and by reference to the published data in the literature (Wang et al., 2013).

Among QA and its derivatives, the compound QA₁ showed maximum antibacterial potential against all the tested bacteria with maximum diameter inhibition zones at 1,500 μg/mL of 33.10 mm, 38.86 mm, 32.13 mm, and 35.46 mm against E. coli, Bacillus sp., Y. enterocolitica, and S. aureus, respectively (Figure 2). This was followed by compound QA₂, which showed 29.30 mm and 22.50 mm zones of inhibition against Bacillus sp. and S. aureus, respectively, and compound QA₃, which showed zones of inhibition of 27.20 mm and 24.61 mm against E. coli and Y. enterocolitica, respectively, at the same concentration, as shown in Supplementary Tables S3–S6.

The compound QA₁ also exhibited the lowest MIC values against all the tested bacteria in the range of 80–120 μg/mL (Figure 3). QA was found to be the least effective against all the bacterial strains, with maximum values of MIC (300–450 μg/mL) against the tested bacteria.

Based on their MIC values, the antibacterial activity of QA and its derivatives against the tested bacteria followed the following order:

Streptomycin > QA₁ > QA₂ > QA₃> QA (Bacillus sp. and S. aureus).

Streptomycin > QA₁ > QA₃> QA₂ > QA (E. coli and Y. enterocolitica).

Depending upon the MIC values, the antibacterial potential of QA₁ (most effective treatment) against the tested bacteria decreased as follows.

Bacillus sp. > S. aureus > Y. enterocolitica > E. coli.

All the treatments were found to be less effective than the standard streptomycin, which showed an MIC in the range of 5–15 μg/mL against all the tested bacterial strains. All the results were significantly different from each other by (p < 0.05), as indicated by two-way ANOVA followed by Tukey’s-B test.

QA has been found effective against several Gram-positive and Gram-negative bacteria (Ercan and Dogru, 2022). It has been reported in the literature that QA inhibited the growth of S. aureus ATCC 6538 by reducing the bacterial DNA content, having an MIC of 2.5 mg/mL (Bai et al., 2018). Various QA derivatives isolated from Z. bungeanum leaves via ethyl acetate extraction exhibited consistent and moderate bactericidal activity with an MIC and minimum bactericidal concentration (MBC) of 5 mg/mL and 10 mg/mL, respectively (Chang et al., 2018). Various acyl quinic acids detected in Geigeria alata extracts exhibited antibacterial potential against S. aureus in which 3,4,5-tricaffeoylquinic acid possessed maximum antibacterial activity (MIC/MBC = 2.5 mg/mL) (Zheleva-Dimitrova et al., 2017).

In our study, QA₁ carrying a lactone moiety has been found to be more effective than its parent compound, QA. Similar results have also been reported in the literature where the antibacterial potential of QA was less than that of its derivative, QA₁. QA showed MIC values at 100 μg/mL, 1,024 μg/mL, 1,024 μg/mL, and 1,021 μg/mL; however, QA₁ showed comparatively lower MIC values at 100 μg/mL, 512 μg/mL, 512 μg/mL, and 512 μg/mL against S. aureus, Staphylococcus epidermidis, Pseudomonas aeruginosa, and Mycobacterium tuberculosis, respectively (Rezende et al., 2014). It has been reported that QA alters the fluidity of cell membranes by disturbing oxidative phosphorylation along with altering levels of fatty acids and glycerophospholipids. After crossing the cell membrane, it inhibits protein synthesis by altering ribosome function and synthesis of aminoacyl-tRNAs (Bai et al., 2019).

All the compounds caused a significant increase in cell constituent release in all the tested bacteria compared to the control, as indicated by an increase in OD values. Among the four compounds, the maximum cell constituent release was caused by QA₁, followed by QA₂, QA₃, and QA. Among the four tested bacteria, QA₁ caused the maximum cell constituent release in Bacillus sp., having the highest OD value, 0.56 ± 0.80, followed by E. coli (OD; 0.419 ± 0.23), S. aureus (OD; 0.380 ± 0.30), and Y. enterocolitica (OD; 0.375 ± 0.96), as shown in Figure 4 . On the other hand, QA₂ caused the maximum release of cell constituents in E coli (OD; 0.389 ± 0.64), followed by Bacillus sp. (OD; 0.358 ± 0.51), S. aureus (OD; 0.295 ± 0.45), and Y. enterocolitica (OD; 0.261 ± 0.55). QA₃ showed the maximum release of cell constituents in Bacillus sp (OD; 0.256 ± 0.72), followed by S. aureus (OD; 0.210 ± 0.84), E. coli (OD; 0.180 ± 0.72), and Y. enterocolitica (OD; 0.150 ± 0.50). QA showed the lowest bacterial constituent release in Bacillus sp (OD; 0.125 ± 0.22), followed by S. aureus (OD; 0.106 ± 0.23), E. coli (OD; 0.098 ± 0.46), and Y. enterocolitica (OD; 0.050 ± 0.65) (Figure 4). These differences in cell constituent releases in four tested bacteria may be due to differences in the chemical structures of the tested compounds. Loss of bacterial cell constituents means that the treatment resulted in irreversible damage to the cytoplasmic membrane. It has been reported in the literature that various organic acids strongly affected bacterial cell membrane integrity, as indicated by an increase in OD₂₆₀ values and bacterial liquid conductivity (Liu et al., 2020).

Our results were also supported by the fact that QA had a damaging effect on the cell membrane as observed by transmission electron microscopy, which led to hyperpolarization and decreased membrane fluidity by interacting with the phenylalanine residues of the membrane protein as observed by fluorescence quenching technique (Bai et al., 2018).

Synergistic interaction between plant-derived natural products and synthetic antibiotics may allow the use of a lower doses of antibiotics to minimize their side effects (Nidhi et al., 2020). The synergistic interactions of the parent compound (QA) and the most effective derivative (QA₁) against a standard antibiotic, streptomycin, were studied through the checkerboard method.

The results were expressed in terms of FICI, as shown in Table 1. Among all the treatments, the compound QA₁ carrying a lactone moiety showed synergistic interaction with streptomycin against all the tested bacteria, E. coli, Y. enterocolitica, Bacillus spp. and S. aureus with FICI values of 0.31, 0.37, 0.37, and 0.29 respectively. QA showed indifferent behavior with streptomycin against E. coli, Y. enterocolitica, and S. aureus with FICI values at 1.12, 1.00, and 1.12, respectively, while it showed additive behavior with streptomycin against Bacillus sp. with an FICI value of 0.62, as shown in Table 1. QA₁ showed synergistic behavior and a reduced MIC of the standard antibiotic (streptomycin) from 8.70 to 0.54 μg/mL against E. coli, from 8.0 to 1.0 μg/mL against Y. enterocolitica, from 15.0 to 1.80 μg/mL against Bacillus sp., and from 5.0 to 0.62 μg/mL against S. aureus.

QA₁, when used in combination with streptomycin, resulted in an 8–16× reduction in the MIC values of streptomycin. Our results are supported by the literature where sesquiterpene lactones were isolated from Centratherum punctatum Cass. showed limited antibacterial activity individually against drug-resistant E. coli and K. pneumoniae. However, when combined with ampicillin, they exhibited a synergistic effect, significantly enhancing antibacterial efficacy. The MIC values of the standard antibiotic, ampicillin, when used with different lactones, showed a reduction from 1,250 μg/mL to 78–625 μg/mL for E. coli and from 2,500 μg/mL to 78–1,250 μg/mL for K. pneumoniae. The experiment illustrated a pronounced synergistic interaction, as indicated by the FICI index values ranging from 0.185 to 1.00 for E. coli and 0.28 to 0.75 for K. pneumoniae (Chukwujekwu et al., 2018). The bactericidal activity of caffeic acid was improved when quinic acid was used in synergy due to a reduction in log survival ratio (3.71 ± 0.23 CFUs/mL vs. 5.45 ± 0.39 CFUs/mL when used alone) at low pH, thereby suggesting the synergistic potential of quinic acid (Kabir et al., 2014). It has also been reported that QA exhibited a synergistic antibiofilm effect with levofloxacin at concentrations below its MIC (Lu et al., 2021).

Biofilm represents a multicellular entity of the bacterial colonies present in the self-produced matrix that protects bacteria under adverse environmental conditions. It is considered one of the major factors responsible for prolonged bacterial infections as it protects them from antibiotic treatment (Jain and Parihar, 2018). QA₁ and its combination with the standard antibiotic, streptomycin, were tested for their antibiofilm activity against all the test bacteria as it revealed maximum antibacterial potential against these bacteria along with a synergistic interaction with streptomycin. The results were further interpreted in terms of specific biofilm formation values.

QA₁, when used alone, exhibited intermediate biofilm formation with SBF values of 1.33, 1.25, and 1.87 against E. coli, Bacillus sp. and S. aureus, respectively, and strong biofilm formation with an SBF value of 2.55 against Y. enterocolitica. However, when QA₁ @ MIC/4 was used in combination with streptomycin (MIC/8), the pair showed weak biofilm formation with SBF values of 0.98 and 0.80 against Y. enterocolitica and Bacillus sp., respectively. It also showed weak biofilm formation (SBF = 0.90) @ MIC/4 against E. coli when used in combination with streptomycin (MIC/16) and intermediate biofilm formation (SBF = 1.20) @ MIC/6 against S. aureus when used in combination with streptomycin (MIC/8) (Table 2).

In one study, QA showed inhibition of extracellular polymeric material (EPS) secretion in the biofilm of P. aeruginosa at some sub-MICs (Lu et al., 2021). Moreover, QA, when treated @ 0.3125–1.25 mg/mL, significantly reduced biofilm formation of S. aureus from 55% to 70% as compared to the control. It significantly reduced biofilm mass by decreasing viability and also reduced the metabolic activity of biofilm cells (Bai et al., 2019). Various chlorogenic acids isolated from an ethyl acetate extract of Z. bungeanum leaves showed decreased metabolic activity of the cells in the biofilm of S. aureus with 31.1%–65.4% inhibition (Chang et al., 2018). Similar results were reported in which a natural product (α-amylase) showed antibiofilm activity against P. aeruginosa and S. aureus at 200 μg/mL (Lahiri et al., 2021).

Molecular docking is a useful approach for discovering new drugs. It is based on searching for drug targets like enzymes or receptors and also bioactive components in natural products (Gupta and Singh, 2012; Sarkar et al., 2022). DNA gyrase is one of the attractive targets for antibacterial drugs; it is involved in bacterial DNA replication and transcription. This enzyme belongs to the topoisomerases, which control topological transitions of DNA and catalyze negative supercoiling of circular DNA by ATP hydrolysis (Collin et al., 2011). The enzyme has also been used as a target for antibacterial drugs as it is essential for all bacteria and is absent in eukaryotes (Heddle and Maxwell, 2002; Kumar et al., 2014). Docking analysis was carried out to better understand the mode of action of bioactive components against the bacterial enzymes transpeptidase and β-lactamase. Transpeptidases catalyze the nucleophilic carbonyl substitution process involved in cross-linking of peptidoglycan in the bacterial cell wall. β-lactamase inactivates betalactam antibiotics by hydrolyzing the peptide bond of the characteristic four-membered beta-lactam ring, rendering the antibiotic ineffective. By understanding the interactions between target enzymes and the bioactive components through docking studies, researchers can design more potent inhibitors to overcome antibiotic resistance and enhance the efficacy of natural antibiotics in clinical use (Elfaky et al., 2020; Shidiki and Vyas, 2022).

Hence, molecular docking studies were performed to predict the interaction of QA and its derivatives with the binding sites of DNA gyrase, transpeptidase, and β-lactamase. The crystal structures of the three enzymes (PDB codes 1KZN, 1MWT, and 6BU3, respectively) were chosen as the protein model for the current study (Jayashree et al., 2010; Bansal et al., 2014; Elfaky et al., 2020).

The computational studies revealed that the cognate ligand (clorobiocin), which is an aminocoumarin antibiotic that inhibits DNA gyrase, bound to the same pocket as in the experimentally-crystallized complex (PDB Code: 1KZN), with a good docking score (−9.1 kcal/mol, Figure 5). This confirmed the robustness of the adopted docking protocol for DNA gyrase. This pocket is an ATP-binding site, and clorobiocin is a known competitive ATP inhibitor. Furthermore, streptomycin showed −2.82 kcal/mol of binding energy against the target enzyme, transpeptidase (Figure 6), and ampicillin showed a binding score of −7.38 kcal/mol with β-lactamase (Figure 7).

Two of the four ligands analyzed (compounds QA₁ and QA₂) docked to the same ATP-binding site as clorobiocin, with docking scores of −5.8 kcal/mol and −5.6 kcal/mol, respectively (Figure 8). This suggested that these two ligands competed for the ATP-binding site and could act as potential gyrase inhibitors. Specifically, QA₁ interacted with VAL167 and VAL43 amino acid residues through hydrogen bonding and formed hydrophobic interactions with ILE78 and ALA47 amino acid residues that had the highest docking score (−5.8 kcal/mol) among all the ligands (Figure 8).

In contrast, the compound QA₂ formed a single hydrogen bond with GLY77 and revealed no apparent hydrophobic contacts (Figure 6). Moreover, the remaining two ligands (compounds QA and QA₃) bound away from the ATP-binding site, with docking scores of −4.2 kcal/mol and −5.4 kcal/mol, respectively. QA₃ bound near the alpha-helical region of the protein and formed hydrogen bonds with ASN46, ALA96, VAL120, and SER121. QA bound near the beta sheets of the protein and formed hydrogen bonds with the HIS55, LYS75, and THR163 amino acid residues (Figure 8). None of these three ligands formed hydrophobic interactions with protein residues (Table 3).

QA₁ also showed a binding score of −4.70 kcal/mol against transpeptidases, which was more than that of streptomycin (2.82 kcal/mol). QA₁ formed three hydrogen bonds with ASN393, GLN396, and GLY282 amino acid residues and one hydrophobic interaction with LYS281 amino acid residues of transpeptidase (Figure 9). QA₁ also showed a binding score of −6.30 kcal/mol against β-lactamase, which was less than that of ampicillin (−7.38 kcal/mol). QA₁ formed two hydrogen bonds with GLU96 and LYS98 and one hydrophobic interaction with the VAL95 amino acid residues of β-lactamase (Figure 10). However, the remaining compounds showed comparatively less binding energy in the range of −3.65 to −4.57 kcal/mol against transpeptidase and −4.83 to −5.57 kcal/mol against β-lactamase. Among all the three targets tested, QA₁ showed the maximum binding energy against β-lactamase followed by DNA gyrase and transpeptidase and, hence, justified its mode of action.

It has been reported that QA isolated from Lonicerae Japonicae Flos regulated core targets (rhlA, rhlR, and rhlB) in a quorum sensing system, as validated by molecular docking studies (Lu et al., 2021).

In all the docking figures, purple cartoons represent the 3D structures of the docked protein, where the ligands have been shown in different colors. Square boxes depict the residues around each ligand. Amino acid residue in the green interacted with the ligand through hydrogen bonds (hydrogen bonding distance in Å), while amino acid residues in pink were involved in hydrophobic interactions with ligands.

Current toxicity studies of drug molecules using computational simulation are cost-effective, less time-consuming, and do not involve animals. The ProTox-II webserver includes both chemical and molecular target knowledge. This platform provides information about six different toxicity classes. Class 1 represents highly toxic substances (LD₅₀ ≤ 5), Class 2 also includes “fatal if swallowed” and comparatively less toxic substances (5 < LD₅₀ ≤ 50), Class 3 marks “toxic if swallowed” substances (50 < LD₅₀ ≤ 300), Class 4 is “harmful if swallowed” substances (300 < LD₅₀ ≤ 2000), Class 5 is “may be harmful if swallowed” substances (2000 < LD₅₀ ≤ 5,000), and Class 6 includes substances with a non-toxic nature (LD₅₀ > 5,000) (Banerjee et al., 2018).

In the present work, the compounds (I–VI) belonged to toxicity classes 3–6, as observed from ProTox-II analysis. The compounds QA, QA₁, and QA₂ belonged to class 6 toxicity and, hence, are safe. They may be used as future pharma drugs. The compound QA₃ belonged to class 4, having intermediate toxicity (Table 4).

The term physiochemical refers to “drug-likeness” that determines the capacity of a molecule to befall as an oral drug regarding its bioavailability. It is an important step in the drug development stage after discovery and involves detailed studies on the drug formulation, stability, pharmacokinetics, metabolism, and toxicity (Ursu et al., 2011). This concept is usually employed to filter chemical libraries to exclude molecules with properties of incompatible behavior and to take forward molecules found in primary screening with an acceptable pharmacokinetics profile (Ursu et al., 2011).

The distinction between drug-like and non-drug-like properties of molecules is based on their certain physiochemical properties, which are generally evaluated by Lipinski’s Rule of Five (Borul and More, 2022). Lipinski’s Rule of Five comprises ranges of several physiochemical properties that make a molecule orally available (Ntie-Kang et al., 2019).

The physiochemical properties of all the compounds were found to be in the required range, except for a few violations, as shown in Table 5. The permeability of the drug molecule through the cell membrane is affected by various factors such as molecular weight and topological polar surface area (TPSA). In this study, the molecular weight of all the molecules fell within the required range (100–600 Da), and TPSAs of all the compounds except QA₂ were found to be in the required range (0–140), indicating their in vivo permeability potential (Table 5). Another parameter is log P, which measures lipophilicity and is defined as the logarithm of the partition coefficient between the aqueous and organic phases. A value of log p less than 5 is favorable for a good rate of absorption (Shukla et al., 2014). In this study, log p values of all four compounds were in the desirable range (less than 5; Table 5). In addition, the capability of the drug molecule to cross the cell’s bilayer membrane is determined by the number of hydrogen bond donors (preferably ≤5) and acceptors (preferably ≤10). In this study, hydrogen bond donors and acceptors for all four compounds were in the acceptable range.

Apart from the Lipinski parameters, the log S value determines the drug solubility, and the lowest value is preferred with the optimal range of −4 mol/L to 0.5 log mol/L. Herein, all four compounds showed the log S values in the desirable range from −1.8 mol/L to −0.3 log mol/L. Log D represents the distribution coefficient and influences the lipophilicity of ionizable compounds as it impacts the permeability property of drugs. Herein, all the compounds showed log D values between 1 and 3 as the standard limit except QA and QA₃, which showed values of −1.48 and −0.48, respectively.

The oral bioavailability potential of drug molecules is reflected by the number of rotatable bonds with an acceptable range of ≤10 (Leeson and Springthorpe, 2007; Pollastri, 2010). In this study, compound QA showed one rotatable bond (nRot), QA₁ and QA₃ exhibited no rotatable bonds, and QA₂ showed 11 rotatable bonds (Table 5), thus reflecting all the molecules have a strong potential of oral bioavailability except QA₂ with comparatively less oral bioavailability potential.

Too many rigid bonds (nRig) (0–30) can lead to increased molecular complexity, ultimately affecting a substance’s oral bioavailability. Herein, all the compounds showed 6–14 nRig bonds. Enzymes involved in drug metabolism can interact with charged sites and further affect the normal distribution of drug molecules in the target body; therefore, the formal charge (f Char) limit is −4 to +4, and all the compounds showed zero formal charge (f Char), which is within permissible limits. More heteroatoms (nHet) indicate increased reactivity and more potential for specific types of interactions with a permissible limit of 1–15, and all the compounds showed nHet values within the range (Table 5).

The safety limit profiles of all the physiochemical properties of compounds were studied using in silico approaches represented by radial plots of each compound showing the upper limit (yellow region), lower limit (red area), and compound properties (blue line). Ideally, the blue line should lie outside the red region and inside the yellow region (Figure 11).

All the compounds were further evaluated by in silico ADMET analysis to predict their pharmacokinetic profile (Van De Waterbeemd and Gifford, 2003). The natural product score (NP score) typically ranges from −5 to +5, with positive scores indicating higher natural product-likeness and negative scores indicating lower natural product-likeness (Ekins et al., 2002). In the present results, all the compounds showed values in the limited range, showing their natural product nature (Table 6). The adsorption parameters include the Caco-2 permeability property with an optimal value of > −5.15, which depicts the absorption of orally administered drugs. In the present study, all the compounds showed permeability values in the required range, except compounds QA (−6.11) and QA₃ (−5.46); however, compound QA₁ showed maximum value (−4.79), revealing maximum permeability. Furthermore, human intestinal absorption (0–1) was in the range of 0.0013–0.937. The distribution parameter includes the blood–brain barrier (BBB), which is the ability of a drug to penetrate the blood–brain barrier. It is a crucial factor in the distribution phase of ADMET analysis, specifically concerning the ability of a drug to penetrate the central nervous system (CNS) and reach a target site in the brain. Herein, all the compounds showed probability in the range of 0.164–0.733, and QA₁ showed the highest (0.733) probability, suggesting higher distribution capacity.

Cytochrome P450 enzymes (CYP enzymes) play a crucial role in drug metabolism in humans. Ideally, antibiotics should be substrates of CYP enzymes and should be metabolized by these enzymes. The antibiotic should be efficiently processed and eliminated from the body, allowing for predictable pharmacokinetics. Antibiotics that do not strongly inhibit CYP enzymes are preferred. Non-inhibitory antibiotics are less likely to interfere with the metabolism of other drugs, reducing the risk of harmful drug interactions (Anzenbacher and Anzenbacherova, 2001). All the compounds showed varied probabilities of substrate rate in different isoforms of the CYP450 enzyme; however, QA₁ showed the highest substrate probability, suggesting its good metabolic rate as compared to other compounds. In this study, none of the compounds showed an inhibitory effect on the cytochrome P450 family of proteins, as all compounds showed a probability close to 0 except QA₃, which showed a comparatively high probability (0.87) of being a CYP450 2D6 inhibitor, as shown in Table 6.

The excretion parameter clearance (CL) was studied; its optimum range (5–15) describes a moderate elimination of drugs from the body. In the present study, all the compounds showed values in the limited range. The toxicity parameters include oral toxicity, skin sensitization, eye corrosion, eye irritation, and respiratory toxicity, and the most effective substance, QA₁, showed the probabilities as 0.077, 0.071, 0.003, 0.04, and 0.014, respectively. Some compounds showed a probability close to 1, indicating a likelihood of being toxic (Table 6).