Results of agar well diffusion assay showed that cephalosporins displayed growth inhibitory zones reporting their antimicrobial activity against C. violaceum CV026 (Figure 1A and Supplementary Figure 1). In addition to the antimicrobial zone, there was a distinct halo zone of pigment inhibition at the interface of bacterial growth (purple color) and antimicrobial zone (Figure 1A). The halo zone of pigment inhibition was designated as the anti-QS zone. Each antibiotic was tested in eight dilutions (4 wells in each plate) with numbers 1-8 indicating highest to lowest concentrations of antibiotics (from 512 μg to 0.4 μg). Most of the antibiotics showed a significant zone of anti-QS activity (Supplementary Figure 2). The largest anti-QS zone was exhibited by CP (0.66 + 0.87 mm) (Figure 1A), CF (0.48 + 0.52 mm) (Supplementary Figure 3A), and CT (0.24 + 0.36 mm) (Supplementary Figure 3B). Other cephalosporin antibiotics showed a small yet significant zone (p < 0.05) of anti-QS activity including imipenem (0.19 + 0.07 mm), doripenem (0.26 + 0.87 mm), meropenem (0.25 + 0.13 mm) and ertapenem (0.31 + 0.09 mm) (Supplementary Figures 1D-H, 2).

Further, we determined the MICs of CP, CF, and CT to validate the anti-QS activity at sublethal concentrations against the CV026 strain. MICs of CP, CF, and CT were 128, 8, and 16 μg/mL (Supplementary Figure 5). The CV026 strain showed lower absorbance at 8, 4, and 2 μg/mL for CT as compared to the control (Supplementary Figure 5B). However, the cells were present as indicated by the visible turbidity indicating the cells survived the antibiotic treatment. CF also showed lower absorbance at 4, 2, and 1 μg/mL concentrations (Supplementary Figure 5C). Similarly, CP showed lower absorbance at 64, 32, 16, and 8 μg/mL. However there was no significant effect on the absorbance at 4, 2, 1, and 0.5 μg/mL concentrations (Supplementary Figure 5D). To quantify the anti-QS activity using pigment inhibition, we selected the 1/2th, 1/4th, 1/8th, and 1/16th concentrations of MIC values of CP, CT, and CF antibiotics. CV026 showed purple pigment production in microtiter wells with visible cell growth. Quantification of the pigment was performed by measuring the OD at 567 nm. The intensity of each well was compared with the control. CP at 32.0 μg/mL and 16.0 μg/mL significantly (p < 0.05) inhibited pigment production (Figures 1B,C), while CF inhibited pigment production at 2.0 μg/mL (Supplementary Figures 3C,D). CT showed significant pigment inhibition at 4.0 μg/mL (Supplementary Figures 3E,F). In these experiments, we have used sublethal concentrations of antibiotics throughout. In all cases, bacterial growth was devoid of pigment production, indicating that cephalosporins interfered with QS pathways in C. violaceum CVO26 with minimum effect on bacterial growth.

We performed molecular docking of the CviR receptor with CP, CF, and CT using Autodock Vina (Figures 1D,E, Supplementary Figures 3G,H, and Table 1). To find interacting amino acids in the binding pockets of the CviR receptor, we selected amino acids within 5Å distance of each ligand. The natural ligand of the CviR receptor, C10-HSL, showed the highest predicted binding affinity with −6.9 kcal mol^–1. The stabilization of C10-HSL in the binding pocket of the receptor is based on hydrogen bond interactions of the carbonyl group with Trp84, the amide carbonyl group with both Tyr80 and Ser155, and the secondary amine with Asp97 (Figure 1D and Table 1). The last carbon on the alkane tail of C10-HSL formed an alkyl interaction with Met72 and a π-alkyl interaction with Tyr88. CT displayed the most favorable interactions with the CviR receptor, with a predicted affinity of −6.5 kcal mol^–1. CT formed four hydrogen bonds; one between Met72 and the carbonyl on the beta-lactam ring, one between Ile69 and the carboxylic acid, and two more hydrogen bonds with Tyr88 and Asn92 from the amine connected to the aromatic ring (Supplementary Figure 3G, Table 1). A π-sulfur interaction was formed between Met72 and the aromatic ring, an alkyl interaction between Ala94 and the non-aromatic ring thioether, as well as a π-alkyl interaction with the aromatic ring Ala94. In addition to this, an unfavorable interaction was formed between Ala94 and the proton of the amine connected to the aromatic ring, which may be the reason for the slightly lower predicted energy than the natural ligand. CP showed the second highest predicted binding affinity of -6.3 kcal mol^–1 with the CviR receptor. CP formed π-alkyl interactions with Met72 and Ala94 and an amide-π stacking interaction with Gln70 (Figure 1E and Table 1). Two hydrogen bonds were formed between (1) Ile69 and the proton of the primary amine connected to the aromatic ring and (2) Met72 and the nitrogen in the methoxy-amino group. Alkyl interactions were formed with Met72 and Val75 residues. In addition, an unfavorable donor-donor interaction was detected between the proton of the carboxylic acid and Ser89. Among the three inhibitors, CF showed the lowest affinity for the CviR receptor (−6.1 kcal mol^–1). CF formed a hydrogen bond with Ser89 from the primary amine connected to the aromatic ring, and another with Met72 from a carboxylic acid group. Near the carboxylic acid group, four alkyl interactions were formed with the neighboring carbon atoms (Supplementary Figure 3H, Table 1).

The antibiotic treatment showed significant inhibition of pyocyanin production as compared to control (Figure 2 and Supplementary Figure 4). PAO1 showed significant pyocyanin pigment production OD_{690 nm} (1.67 + 0.19) and with CT (1.02 + 0.18), CF (0.77 + 0.27), CP (0.48 + 0.17). There is a clear reduction of pyocyanin production by CP, CF, and CT supplementation. These results collectively suggested that CP, CF, and CT interfered with molecular pathways associated with pyocyanin production.

Motility phenotypes were analyzed for PAO1 control and QS mutant strains. We found that the QS mutant strains showed no significant difference in swimming motility, which suggests that the flagella-dependent swimming motility is independent of QS pathways (Figures 3A_i, 4A). Interestingly all three antibiotics CP, CF, and CT showed significant swimming motility inhibition (p < 0.001) of PAO1 as well as the isogenic QS mutant strains (Figure 4A). The diameters of the zones are mentioned as table insert in Figure 4A. Further, results showed that the ΔlasR and ΔrhlR mutant strains showed a significant difference (p < 0.05) in swarming motility as compared to the control indicating that the motility is dependent on las and rhl QS pathways (Figures 3B_i, 4B). The pqsR and pqsA mutant strains showed no significant difference in the swarming motility zone as compared to the control PAO1. Interestingly all three antibiotics CP, CF, and CT, showed significant swarming motility inhibition (p < 0.001) for wild-type PAO1 and mutant strains (Figures 3B_i, 4B). Effect on twitching motility was also analyzed on PAO1 and its isogenic mutant strains, and it was found that the ΔlasR/ΔrhlR mutant strains showed significant difference (p < 0.05) in the twitching motility (Figures 3C_i, 4C), indicating that the twitching motility is dependent on las and rhl QS pathways. Interestingly only two antibiotics CP and CT, showed significant twitching motility inhibition (p < 0.001), but surprisingly, CF showed no motility inhibition (Figures 3C_i, 4C). Our results showed that CP and CF have a strong inhibitory effect on motility phenotypes.

We used Environmental Scanning Electron Microscopic (ESEM) to image biofilms produced by PAO1 (Figure 5). ESEM imaging and live cell counts were performed on every alternate day (days 1, 3, 5, and 7) on the urinary catheter surface. On day 1, PAO1 control biofilms showed the presence of a uniform monolayer of cells attached to the urinary catheter surface. Cells were uniformly distributed within the extracellular polysaccharide matrix (Figure 5A_i). The live cell count of the day 1 biofilm was high (6.39 ± 1.89 log cfu) (Figure 5A). In antibiotic-treated groups, CP showed low cell counts with significantly less polysaccharide matrix distributed throughout the catheter surface (Figure 5A_ii). CP treatment significantly reduced (p < 0.001) the log cfu count as compared to the control (3.83 ± 0.32 log cfu). CF and CT both showed a significant amount of exopolysaccharide distributed on the catheter surface (Figures 5A_iii,A_iv). Both antibiotics showed reduced live cell counts of 4.70 ± 0.26 and 5.50 ± 0.43 log cfu (Figure 5A). Three-day-old-biofilms of PAO1 of the control group showed the formation of a thick layer of bacterial cells embedded in an exopolysaccharide matrix (Figure 5B_i). Biofilms appeared to be complex in structure with the presence of long polymer threads, mucilaginous matrix, and high cellular growth. Live cell counts were higher on day 3 in control groups (8.28 ± 1.13 log cfu) (Figure 5B). The CP-treated catheter surface showed low cell counts attached to the urinary catheter surface at day 3 (Figure 5B_ii). The bacterial live cell count on day 3 in CP treated urinary catheters was lower (5.01 + 0.56 log cfu) as compared to the control group (Figure 5B). CF and CT thin biofilm formation on the catheter surface (Figures 5B_iii,B_iv). The bacterial cells appeared to be longer in size as compared to the control. Log cfu count of the treated groups were significantly reduced (p < 0.05) as compared to the PAO1 control (CT 6.0 ± 0.56 log cfu and CF 6.6 ± 0.73 log cfu) (Figure 5B). Five-day-old-biofilms of the PAO1 control group showed thick biofilm (Figure 5C_i). On day 5, the biofilms started to construct mushroom-shaped structures. At this stage, biofilms might be completely resistant to antimicrobial agents due to the protection provided by the exopolysaccharide layer. These clusters of live bacterial cells help in the formation of the three-dimensional architecture of bacterial biofilm. As expected, the biofilm showed a high live cell count in five-day-old biofilms (10.05 ± 0.86 log cfu) (Figure 5C). In CP treated groups, the biofilm appeared to be defective (Figure 5C_ii), and the log cfu count of the live cell was significantly lower, indicating the potent antibiofilm activity of CP (Figure 5C). In CT and CF, treated biofilms were condensed and covered with thick polysaccharide matrix on the catheter surface (Figures 5C_iii,C_iv). However, there was no significant difference in the log cell count of live cells in CF treated biofilms as compared to the control group. CT treated biofilms showed a significant difference in the log cfu count of bacterial cells (7.01 ± 0.19 log cfu) (Figure 5C). Moreover, on day-7 control biofilms appeared highly complex three dimensional with numerous mushroom-shaped secondary structure on the biofilm surface (Figure 5D_i). In the untreated control group, biofilms showed a larger assembly of multi-mushroom shaped structures, and the log cfu count of live cells in 7-day-old-biofilm was found to be very high (12.60 ± 0.75 log cfu). The seven-day-old CP treated biofilms were still defective in biofilm architecture. The mushroom-shaped structures were absent, and individual cells can be visualized on the catheter surface. In treated biofilms, bacterial cells appeared as long thread shaped and produced significantly less polysaccharide with a low log cfu count of 7.81 ± 0.65 (Figure 5D). CT and CF biofilms were covered with mushroom-shaped structures, and few individual cells were visible, indicating the thick three-dimensional biofilm formation. Live cell count of the catheter surface of these antibiotic-treated groups was found to be non-significant as compare to control (Figure 5D).

First, the antimicrobial efficacy of aminoglycosides was tested against P. aeruginosa PAO1. Kanamycin sulfate showed poor antibacterial activity against P. aeruginosa PAO1 (MIC: 64 μg/mL); however, in the presence of CP, the MIC of kanamycin has significantly decreased to 16 μg/mL (Figure 6A). Streptomycin sulfate showed potent efficacy against PAO1, and the MIC was found to be 8 μg/mL (Figure 6B). In combination with CP, the MIC of streptomycin was significantly reduced to 2 μg/mL. Similarly, the MIC of neomycin was found to be 16 μg/mL, and in combination with CP, the MIC was decreased to 8 μg/mL (Figure 7C). Gentamicin and tobramycin were found to be highly effective antibiotics against PAO1 with the MIC values of 2 μg/mL (Figures 6D,E). However, in combination with CP, the MIC values were decreased to 1 μg/mL. Our results showed that the CP at a sub-inhibitory concentration in combination with major aminoglycosides led to the increased antimicrobial potency of aminoglycosides. This combination needs to be tested in animal models to validate its effectiveness against P. aeruginosa and could be established as a clinically useful combinational therapy to treat P. aeruginosa infections.

We docked 3-oxo-C12HSL to its binding pocket in the LasR receptor. The predicted binding affinity was −8.0 kcal mol^–1. 3-oxo-C12HSL formed three hydrogen bonds, one with Ser129 and the amide carbonyl, and two more with Asp73, Thr75, and the proton of the secondary amine (Figure 7A and Table 1). A hydrophobic interaction connected Tyr56 and the ring carbon neighboring the secondary amine, supported by two alkyl interactions between the carbon at the end of the hydrophobic tail with Leu40 and Ala50. Next, the three selected antibiotics were docked to the same ligand-binding pocket of LasR. CP formed two hydrogen bonds, one between the carboxyl group and Asp65 and another between Ala50 and the proton of the secondary amine connected to the beta-lactam ring (Figure 7B). A π- sulfur interaction was formed between Phe167 and the sulfur atom of the aromatic ring, an electrostatic interaction was formed between Asp65 and the charged nitrogen atom, a carbon-hydrogen bond was formed between Asn49 and the beta-lactam carbonyl, and an alkyl interaction was formed between Ile52 and the thioether group. CP showed a comparative docking score of −5.6 kcal mol^–1. CF formed three hydrogen bonds; one between Gly54 and a carboxyl group, and two between the primary amine with Tyr56 and Asn55. Lys16 formed a carbon-hydrogen bond with the carboxyl group and an alkyl interaction with a nearby methyl group (Figure 7C and Table 1). A π - alkyl interaction was formed between Ala58 and the aromatic ring, and an unfavorable electrostatic interaction was formed between Glu62 and the carboxyl nearest to the beta-lactam ring. CF showed a comparatively low docking score of −5.9 kcal mol^–1. CT formed two hydrogen bonds; one between Asn49 and the primary amine and another between Lys16 and the non-aromatic ring thioether (Figure 7D and Table 1). A π - sulfur interaction was formed between Phe167 and the aromatic ring thioether. Carbon hydrogen bonds were formed between Ala58 and the methyl group connected to the triazine ring; and between Asn49 and the beta-lactam carbonyl. Three alkyl interactions were formed; one from Ile52 to the non-aromatic thioether ring, two from the triazine ring to Ala58 and Arg61, respectively. CT showed a comparatively higher docking score of −6.6 kcal mol^–1.

Molecular docking was also performed independently with the PqsR receptor. The natural ligand formed no hydrogen bonds but formed ten alkyl interactions between its two rings with Leu208, Ala168, Ile149, Pro238, and Ala102, as well as five alkyl interactions between the last carbon in the hydrophobic tail and Leu189, Val170, Trp234, Val211, Ile236 of PqsR receptor (Figure 7E and Table 1). The natural ligand showed a docking score of −7.4 kcal mol^–1. CP formed two hydrogen bonds from its primary amine to Arg209 and Leu208. A carbon-hydrogen bond was formed between the methoxyl group and Thr265 (Figure 7F and Table 1). Two π-σ interactions were formed from the aromatic ring to Leu208 and Ile236. An alkyl interaction and a π-alkyl interaction were formed from the thioether to Ile263 and Tyr258, respectively. The docking score of CP was found to be −6.7 kcal mol^–1. Similarly, CF formed a hydrogen bond between its primary amine and Pro210 as well as a carbon-hydrogen bond from Leu207 to the six-membered ring (Figure 7G and Table 1). An unfavorable interaction was also formed between Glu259 and the carboxyl nearest the beta-lactam ring. CF showed a docking score of −6.2 kcal mol^–1. CT formed four hydrogen bonds; one between its primary amine and Glu151, one between Leu207 and the carboxyl group, one between Ile236 and the proton on the triazine ring, and lastly, one between Leu197 and a carboxyl on the triazine ring (Figure 7H and Table 1). The docking score of CT was −6.7 kcal mol^–1. These conformations would obstruct the entry into the binding pocket, and therefore may interrupt ligand-receptor interactions.

We have used C. elegans as an animal model to investigate the protective effect of CP, CF, and CT against P. aeruginosa PAO1. Advantage of its transparent nature, internal organs can be visualized directly. In healthy worms, the intestine, uterus, proximal and distal gonads (Figures 8A_i,A_ii) were found to be intact, and no tissue damage was observed. The diseased/dead worm has blunt-body curvature with no/reduced body movement; these were distinct phenotypic characteristics of the tissue damage or diseased state of the C. elegans (Dimov and Maduro, 2019). PAO1 infected C. elegans showed potent damage of intestinal tissue after 72 h leading to 100% mortality in infected worms. Percentage survival of C. elegans at 12, 24, and 28 h was 88.0 + 3.60%, 65.0 + 4.30% and 36.6 + 8.30% and at 72 h, it has achieved 100% mortality (Figure 8F). Interestingly, the sub-inhibitory concentration of antibiotic treatment showed a protective effect in C. elegans infected groups. CP treatment showed a significant increase in the percentage survival of C. elegans. At 12, 24, 48, and 72h the percentage survival was 95.6 + 2.51%, 92.33 + 2.50%, 83.33 + 4.16%, and 65.33 + 5.03% (Figure 8F). Similarly, CF treated groups also showed a significant increase in worm survival. At 12, 24, 48, and 72 h the % survival was found to be 93.6 + 1.52%, 85.33 + 5.03%, 77.0 + 5.56% and 58.0 + 4.30% (Figure 8F). CT showed a similar protective effect, and treated groups showed reduced mortality. At 12, 24, 48, and 72 h, the percentage survival was found to be 92.0 + 2.60%, 84.33 + 5.56%, and 48.66 + 8.08% (Figure 8F). All three antibiotics showed a significant reduction in the C. elegans mortality, indicating the protective effect of a sub-inhibitory concentration of cephalosporin antibiotics. We also tested the effect of quorum sensing mutant strains on C. elegans survival. Percentage survival of DeltalasR mutant at 12, 24, 48 and 72 h was 92.66 + 1.52%, 82.0 + 5.29%, 59.0 + 9.53% and 31.66 + 11.23% (Figure 8G). Similarly, ΔrhlR mutant strain showed 89.0 + 3.60%, 77.6 + 10.20%, 55.0 + 10.0% and 27.33 + 7.63% (Figure 8G). Similarly, ΔpqsR and ΔpqsA mutant strains were also showed 86.33 + 5.13%, 71.33 + 6.11%, 47.66 + 11.23%, 26.66 + 10.40% and 84.33 + 4.04%, 75.66 + 7.76%, 57.0 + 12.53%, and 36.66 + 5.77% survival rates of C. elegans (Figure 8G). We have also used an uninfected group with E.coli OP50 to check the C. elegans mortality, and no mortality was observed until 72 h. The result of our experiment showed that the sub-MIC concentration of antibiotics protected C. elegans from P. aeruginosa infection and led to an increase in the survival of the worm. Finally, we have compared the percentage mortality of all the groups with the control PAO1 group, and all the groups have shown a significant reduction in the percentage mortality at 72 h (Figure 8H). Additionally, mortality rates were higher in the QS mutant strains as compared to the antibiotic-treated groups (Figure 8G). This indicates that cephalosporins might have additional alternative mechanisms apart from their anti-QS activity to protect C. elegans from P. aeruginosa infection.

In the future, it might be interesting to investigate the effect of these antibiotics at the molecular level against P. aeruginosa. This may lead to the identification of novel molecular pathways in P. aeruginosa.

Chromobacterium violaceum CV026, P. aeruginosa wild type strain PAO1 and its defective QS mutant strains including ΔlasR,ΔrhlR,ΔpqsA, and ΔpqsR were kindly provided by Dr. Paul Williams (Professor of Molecular Microbiology, Faculty of Medicine and Health Science, University of Nottingham, United Kingdom). C. violaceum CV026 was cultured using Luria Broth with kanamycin at 50 μg/mL (30°C) under stationary conditions and used for the detection of acyl-homoserine lactone and screening of anti-QS activity of β-lactam antibiotics. All bacterial strains were maintained as 25% glycerol stocks and stored at −80°C. Fresh subculture was performed from frozen stocks and used for each experiment.

The following antibiotics cefepime hydrochloride (#PHR1763-1G:LRAB8503, sigma), ceftriaxone disodium salt hemi-heptahydrate (#C5793-1G: 027M4799, sigma), ceftazidime (#CDS020667-50MG:B02487097), oxacillin sodium salt (Cat#28221-5G:097M4853V), imipenem (#PHR1796-200MG:LRAB7177, sigma), doripenem hydrate (#SML1220-50MG:025M4717V, sigma), meropenem trihydrate (#USP-1392454: 50K434), ertapenem (#CDS022172-25MG:B02467020) were procured in dry powder and stored in 4^oC under desiccated conditions. Master stocks of 1024 μg/mL were made in sterile distilled water and stored in -80^oC for future use. Every time fresh stock of antibiotics was used for experiments. All the other chemicals are of ACS analytical grade.

The MIC of antibiotics was determined, according to National Committee for Clinical Laboratory Standards (NCCLS) (2002) guidelines against the standard P. aeruginosa wild type strain, PAO1. Round bottom, 96 microwell plate was used for MIC determination. Briefly, the stock solution was serially diluted (2-fold) in cation adjusted Muller Hinton broth, and 2–8 × 10⁵ cfu was used as inoculum in the total reaction of 100 μL per well in cation Muller Hinton broth (n = 3 for each antibiotic dilution. After 15 h, resulting in no visible growth was selected as the MIC. Sterile media with bacteria and sterile media with antibiotics are served as positive controls and negative controls. The final plate was also read at OD_{600 nm} to quantify the effect of the antibiotic on PAO1. MIC, 1/2th, and 1/4th MIC were tested on the growth curve profile of PAO1. Respective concentrations showing no significant effect on the growth profiles were selected for anti-virulence factors and anti-biofilm assays against P. aeruginosa.

Pseudomonas aeruginosa PAO1 wild type and QS mutant strains (ΔlasR,ΔrhlR,ΔpqsA, andΔpqsR) were assessed for motility competence in the presence of a sub-inhibitory concentration of CT, CF, and CP.

Pyocyanin was estimated using the method of Huerta et al. [34] with slight modification. Briefly, PAO1 was grown under different concentrations (10% to 100%) of Luria broth, and different concentrations were tested for pyocyanin production (n = 3 for each antibiotic). At the different concentrations of antibiotics (sub-MIC) of CT CF and CP, the production of pyocyanin production was estimated and compared with control. Chloroform was added to cell-free culture supernatant at different concentrations, and OD at 690 nm of the chloroform layer was measured.

Minimum inhibitory concentration concentrations of aminoglycosides, namely, gentamicin sulfate (Sigma, Cat# G1914), neomycin sulfate (Sigma, Cat# PHR1491), streptomycin sulfate (Sigma, Cat# S6501), tobramycin sulfate (Sigma, Cat# T1783), kanamycin sulfate (Sigma, Cat# 60615) were determined using standard 96 well plate assay as mentioned previous section in detail. To determine CP’s combinatorial effect with an aminoglycoside, we added CP at a sub-inhibitory concentration (0.5 μg/mL) in each well with different dilutions of aminoglycosides. The MIC was determined by visual detection of bacterial growth and optical density measurement of each well. OD_{600 nm} values were plotted as a heat map using Graph pad prism 8.0.

Molecular docking studies were performed using Autodock Vina (Version 1.5.6) following standard protocols. The crystal structures of CviR (PDB ID: 3QP8), LasR (PDB ID: 3IX3), and PqsR (PDB ID: 4JVI) were downloaded from the protein databank, at http://www.pdb.org. Ligands, ions, water molecules, and small molecules were removed, and the file was saved in pdbqt format using autodock tools. Ligand structure files in PDB format were downloaded from www.rcsb.org. Structures that were not available in PDB format were converted from sdf using an online converter¹. The binding site was located by highlighting key amino acids on the receptor protein in autodock tools. Then, a grid box was placed to encase the binding site area while minimizing the total volume. A config.txt file was created to specify the receptor, ligand, and the x, y, and z dimensional coordinates of the docking site. The coordinates for protein receptors were provided as in Supplementary Table.1). For extensive simulation, the ‘exhaustiveness’ option was set as 10.0. Autodock Vina was run, and the best docking conformations were visualized using Pymol (Zalman 3D) software and BIOVIA Discovery studio. Docking conformations of each ligand were analyzed by examining their total energy score. Docking scores of cephalosporin antibiotics were compared with the natural ligand of each receptor.

Nematode infection model (Musthafa et al., 2012b; Husain et al., 2013) with modification was used to access the anti-virulence effect of cephalosporins. C. elegans N2 strain was procured for CGC. The worms were cultivated in an NGM Petri plate with a bacterial lawn of OP50 cells. L2 and L3 stage nematode (20–40 worms) were transferred to 10 NGM plates to grow enough worms for the experiment. After the 4-5 days growth, the worms were collected in liquid NGM media, and the concentration of worms was adjusted to approx. 100 worms/mL (n = 100 for each group). P. aeruginosa PAO1 was grown at 40% LB media at 37°C stationary conditions with and without sub-inhibitory concentrations of CP, CF, and CT. The young adult nematodes were infected with PAO1 (control and antibiotic-treated) and incubated at 25°C for 15 h. After the incubation, worms were centrifuged to remove free bacteria and transferred to fresh liquid NGM media with and without the sub-inhibitory concentration of cephalosporins. QS mutant strain was also collected and processed similarly to evaluate its effect on nematode killing. Worms were confirmed dead if they did not show any moment under the microscope. Every 15 h, worms were immediately observed under a light microscope for the mortality estimation. The number of dead worms was counted, and the percentage mortality was calculated as follows: Survival (%) = Total no of worms-Number of dead worms/Total no of worms X100.

All experiments were performed in triplicate and repeated on different days. The effect of antibiotic treatment on pigment production, pyocyanin, motility, live cell count in biofilms, and percentage survival/mortality of C. elegans in different groups was evaluated using a two-way ANOVA test. The p values were calculated, and p < 0.05 was considered significant. Results were analyzed using Graph Prism 8.0 software. Values were expressed as mean ± standard deviation (SD).