Quorum Sensing Inhibiting Activity of Cefoperazone and Its Metallic Derivatives on Pseudomonas aeruginosa

The last decade has witnessed a massive increase in the rate of mortalities caused by multidrug-resistant Pseudomonas aeruginosa. Therefore, developing new strategies to control virulence factors and pathogenicity has received much attention. One of these strategies is quorum sensing inhibition (QSI) which was developed to control Pseudomonas infection. This study aims to validate the effect of one of the most used β-lactam antibiotics; cefoperazone (CFP) and its metallic-derivatives on quorum sensing (QS) and virulence factors of P. aeruginosa. Assessment of quorum sensing inhibitory activity of CFP, cefoperazone Iron complex (CFPF) and cefoperazone Cobalt complex (CFPC) was performed by using reporter strain Chromobacterium violaceum ATCC 12472. Minimal inhibitory concentration (MIC) was carried out by the microbroth dilution method. The influence of sub-MICs (1/4 and 1/2 MICs) of CFP, CFPF and CFPC on virulence factors of P. aeruginosa was evaluated. Data was confirmed on the molecular level by RT-PCR. Also, molecular docking analysis was conducted to figure out the possible mechanisms of QSI. CFP, CFPF, and CFPC inhibited violacein pigment production of C. violaceum ATCC 12472. Sub-MICs of CFP (128- 256 μg/mL), and significantly low concentrations of CFPC (0.5- 16 μg/mL) and CFPF (0.5- 64 μg/mL) reduced the production of QS related virulence factors such as pyocyanin, protease, hemolysin and eliminated biofilm assembly by P. aeruginosa standard strains PAO1 and PA14, and P. aeruginosa clinical isolates Ps1, Ps2, and Ps3, without affecting bacterial viability. In addition, CFP, CFPF, and CFPC significantly reduced the expression of lasI and rhlI genes. The molecular docking analysis elucidated that the QS inhibitory effect was possibly caused by the interaction with QS receptors. Both CFPF and CFPC interacted strongly with LasI, LasR and PqsR receptors with a much high ICM scores compared to CFP that could be the cause of elimination of natural ligand binding. Therefore, CFPC and CFPF are potent inhibitors of quorum sensing signaling and virulence factors of P. aeruginosa.


INTRODUCTION
Pseudomonas aeruginosa is an encapsulated, Gram-negative, rod-shaped opportunistic pathogen that infects plants, animals, and human (Borges et al., 2018). P. aeruginosa can colonize critical body organs such as lungs, urinary tract, and kidney with fatal pathological effects especially on immunocompromised patients (Turkina and Vikström, 2019).
Treatment of P. aeruginosa infections by antibiotics is getting ineffective due to the worldwide spread of multi-drug resistant isolates. In addition, P. aeruginosa possesses diverse virulence factors including biofilm, pyocyanin, elastase, aminopeptidase, chitinase, protease, lipase, alginate, and hydrogen cyanide (Schuster and Greenberg, 2006). All these virulence triats assist the spreading and dissemination of Pseudomonas infection in different organs.
CFP is a bactericidal third-generation cephalosporin possessing extended-spectrum activity against Gram-negative bacilli. Resistance to CFP has been developed where CFP was ineffective in the treatment of Pseudomonas infection. Thus, it is necessary to develop strategies to combat virulence factors without acquiring resistance. Masoud and coauthors have developed CFP metallic derivatives with antimicrobial activity against Gram-negative bacilli (Masoud et al., 2017). To our knowledge, QSI activity of CFP and its derivatives have not been evaluated. So, the aim of this study is to investigate the effect of CFP and its metallic complexes on QS circuits and virulence factors as a new strategy for combating Pseudomonas infection.

Bacterial Strains, Media, and Conditions
Chromobacterium violaceum ATCC 12472 reporter strain was used in the assay of QSI activity of CFP and its metallic complexes (McClean et al., 1997). The reporter strain was grown on Luria-Bertani (LB) media containing (tryptone, 10 g/L; yeast extract, 5 g/L; and NaCl, 10 g/L) at pH 7, and incubated at 28°C for 48 hours (Bertani, 2004). P. aeruginosa clinical isolates were isolated from urine samples and named Ps1, Ps2, and Ps3 according to the Helsinki declaration in handling, use and care of human subjects for medical research. Ethical approval was obtained from the Institutional Review Boards of Faculty of Medicine, Alexandria University, Egypt.
These clinical isolates were identified as P. aeruginosa according to laboratory biochemical standards. P. aeruginosa PAO1 and P. aeruginosa PA14 standard strains were also used for assessment of QSI effects of the tested chemical compounds as positive controls and P. aeruginosa PAO-JP2 (△lasI::Tn10, Tcr; △rhlI::Tn501-2, Hgr) is a QS double-mutant strain and has been used as a negative control . All P. aeruginosa strains were cultivated in LB media and incubated overnight at 37°C.

Cefoperazone and Metallic Derivatives
CFP complexes with FeCl 3 .6H 2 O, CoCl 2 , MnCl 2 , NiCl 2 , and CrCl 3 (Sigma Aldrich, USA) (Table 1) were prepared according to the method reported by Masoud and coauthors (Masoud et al., 2017). Each Metal chlorides (1 M) was dissolved in 40 mL ethanol and CFP (1 M) was dissolved in 40 mL doubledistilled water. Then, metal chloride ethanolic solutions and CFP solutions were mixed in a molar ratio of 1:1. Each reaction mixture was refluxed for 4 hours and left overnight. The formed precipitate was filtered and washed with 10 ml ethanol 95% w/v, and dried in a vacuum desiccator over anhydrous CaCl 2 . The formed CFP metal complexes were confirmed by UV and thermal analysis (Masoud et al., 2017). Stock solutions of cefoperazone-metallic derivatives were prepared at a concentration 5 mg/ml in DMSO (20% v/v).

QSI Assay of Cefoperazone and
Metallic Complexes Using C. violaceum ATCC 12472 C. violaceum ATCC 12472 reporter strain was used to determine the QSI activity of CFP and different complexes. The culture of C. violaceum was propagated at 28°C with 200 rpm agitation in a shaking incubator for 48 hours. The double-layer culture plate method was performed where 15 mL of LB medium (2% w/v agar) was poured into the plates (9 cm). Solidified LB plates were overlaid with 8 mL of soft LB medium (1% w/v agar) containing 100 mL of C. violaceum, and the plates were left to completely solidify. Wells were cut in the agar using a sterile cork borer (10 mm diameter). CFP and its derivative CFPC, and CFPF (100 mL) at 1/2 and 1/4 MICs were loaded in the corresponding wells and the plates were incubated at 28°C. The inhibition zone diameter of the violet color of violacein pigment around each well was measured after 24 hours (McClean et al., 1997;McLean et al., 2004). In each plate, DMSO was applied in one well as a control.

Determination of Minimal Inhibitory Concentrations
Compounds that showed inhibition of violacein pigment production were selected. The minimal inhibitory concentrations of the selected compounds; CFP, CFPC, and CFPF were determined using the microtitre plate assay method. Muller Hinton broth (100 µL) was distributed in each well, 100 µL of the tested compound was added to the first well. Two-fold serial dilution of the tested compound was performed in the subsequent wells to have serial dilutions of 512,256,128,64,32,16,8,4, 1, and 0.5 µg/mL. Each well was inoculated with 1 X10 5 CFU of P. aeruginosa cultures Ps1, Ps2, Ps3, PAO1, PA14, and PAO-JP2. Wells containing media only and wells receiving media inoculated with the test strains were included in each plate as negative and positive controls, respectively. All plates were incubated overnight at 37°C and the microbial growth in each well was visually detected. The MIC was calculated as the lowest concentration that eliminated microbial growth compared to the positive control (CLSI, 2014). Sub-MICs of CFP, CFPC and CFPF were calculated, and the tested strains were propagated in the presence of sub-MICs of the tested compounds.

The Effect of Sub-Inhibitory Concentrations on Bacterial Growth
P. aeruginosa strains Ps1, Ps2, Ps3, PAO1, PA14 and PAO-JP2 were propagated in the presence of 1/2 MICs of CFP, CFPC and CFPF. Control untreated strains were also cultivated under the same conditions. In brief, LB broth media (25 mL) supplemented with 1/2 MIC of the tested compounds CFP, CFPC and CFPF was inoculated with overnight culture of Ps1, Ps2, Ps3, PAO1, PA14 and PAO-JP2 to achieve 0.05 OD 600 nm at zero time. From each mixture, 1 mL was collected at different time intervals and the OD 600 nm was estimated. In addition, viable counts of treated and untreated P. aeruginosa strains Ps1, Ps2, Ps3, PAO1, PA14 and PAO-JP2 were estimated after 18 hours using the pour plate method. The collected samples were diluted 1:10 and 1 mL of each dilution was inoculated in LB agar (15 mL) and distributed in 9 cm plate. The plates were solidified and incubated at 37°C for 18 hours. The count of bacterial cells of each sample was calculated as the viable P. aeruginosa colonies X dilution factor and represented as CFU/mL (Standards Australia, 1995).
2.6 Effect of Sub-MICs of the Tested Compounds on Pseudomonas Virulence Factors CFP, CFPC and CFPF at 1/2 and 1/4 MICs did not exhibit any significant effect on microbial growth. So, they were assessed for their effect on different virulence factors of P. aeruginosa. The tested P. aeruginosa strains Ps1, Ps2, Ps3, PAO1, PA14, and PAO-JP2 (negative control) were propagated in the presence of 1/2 and 1/4 MICs of CFP, CFPC, and CFPF (Musthafa et al., 2012). The untreated strains were also grown under the same conditions. Assay of different virulence factors was performed in the presence and absence of the tested compounds (El-Mowafy et al., 2017).

Pyocyanin Assay
Pyocyanin quantification was performed by using King's A broth media. Overnight culture of P. aeruginosa strains was inoculated into 5 mL of King's A media. Both untreated and treated cultures with sub-MICs (1/2 and 1/4 MICs) of CFP, CFPC, and CFPF were incubated at 37°C for 48 hours with shaking. Pyocyanin was extracted from the supernatant with 3 mL chloroform. The mixture was centrifuged at 3000 rpm for 10 min. Chloroform fractions were transferred to a new tube, then 1 mL of 0.2 M HCL was added, and the mixture was re-centrifuged for 5 min at 3000 rpm. The absorbance of the aqueous layer was measured at OD 520 nm. Pyocyanin concentration was calculated from the equation: pyocyanin concentration (µg/mL)= OD 520 nm × 17.072 (He et al., 2014;Saurav et al., 2016).

Total Protease Production
The tested P. aeruginosa strains were inoculated in LB broth media supplemented with 1/2 and 1/4 MICs of CFP, CFPC, and CFPF and grown overnight at 37°C. The untreated strains were also grown under the same conditions. Cells were centrifuged and proteolytic activity of the treated and untreated P. aeruginosa isolates was measured (Rossignol et al., 2008). Proteolytic activity was determined using skimmed milk assay technique (Skindersoe et al., 2008;El-Mowafy et al., 2017). The assay relies on determining the change in the turbidity of skimmed milk accompanying protease activity. Skimmed milk was freshly prepared by dissolving 1.25 gm of skimmed milk in 100 mL sterile distilled water at 60°C. Then, 0.5 mL of culture supernatant was added to 1 mL of the prepared skimmed milk, and the mixture was incubated at 37°C for 1 hour. The degree of clearance of skimmed milk was determined by measuring the turbidity at OD 600 nm. The decrease in the OD 600 nm was indicated by the clearance of the skimmed milk with elevated proteolytic activity (El-Mowafy et al., 2014).

Determination of Hemolysin Activity
The tested P. aeruginosa strains were propagated in the presence of 1/2 and 1/4 MICs of CFP, CFPC, and CFPF overnight at 37°C (Musthafa et al., 2012). The untreated strains were also grown under the same conditions. Cells were centrifuged and the hemolytic activity of treated and untreated P. aeruginosa isolates was measured (Rossignol et al., 2008). RBCs (obtained from sheep) were washed three times with sterile physiological saline and re-suspended in tris-buffer (pH 7.4, 0.025 M) to a final concentration of 2% v/v. For hemolysin assay, 700 mL of erythrocytes suspension was mixed with 500 mL of treated and untreated cell-free supernatants and incubated for 2 hours at 37°C. The suspension was centrifuged at 3000 rpm for 10 min at 4°C, and cell lysis was assessed by determining absorbance at OD 540 nm.

Quantification of Biofilm Formation Using Microtiter Plate Assay
Flat bottomed polystyrene microtiter plates were used to evaluate slime production and biofilm formation. Overnight cultures of the tested P. aeruginosa strains were diluted with sterile LB broth to 0.5 McFarland. Treated and untreated cultures (100 mL) were transferred to each well and the plates were incubated for 24 hours at 37°C for mature biofilm formation. The content of each well was aspirated using Pasteur pipette and each well was rinsed three times with 200 ml of physiological saline. The plates were shaken to remove all non-adherent cells and the remaining attached bacteria were fixed with 150 mL of absolute methanol for 15 min, then, the plates were emptied and left to dry. Sessile cells bound to the wells were stained with 150 mL of crystal violet (1% v/v) for 10 min. Excess stain was rinsed off, and the plates were washed with water. The plates were air-dried and the dye bound to the wells was eluted with 150 mL of 33% (v/v) glacial acetic acid (Adonizio et al., 2008). The absorbance was measured at OD 490 nm using a microtiter plate reader (Diaket, ELISA plate reader).

Expression of QS Genes
The effect of CFP, CFPC, and CFPF at sub-MICs on the expression of QS genes lasI and rhlI in P. aeruginosa PAO1 was assessed by RT-PCR. The untreated and treated PAO1 with 1/2 MIC of CFP, CFPC, and CFPF were propagated, and cells were collected at the middle of exponential phase. Total RNA was extracted by Triazole reagent (Sigma Chemicals, USA). Chloroform (100 µL) was added to each sample, the tubes were incubated for 2-3 min at room temperature and centrifuged at 12.000 rpm for 15 min at 4°C. The aqueous phase was collected in RNase free Eppendorf tubes and chloroform step was repeated for complete purification of RNA. Isopropanol (300 µL) was added to each tube, the tubes were mixed for 1 min and centrifuged at 12.000 rpm for 15 min at 4°C. The supernatant from the previous step was removed leaving RNA pellet. The pellet was washed twice with 1 ml of ethanol 75% w/v. The sample was gently mixed, and centrifuged at 10.000 rpm for 5 min at 4°C. The supernatant was discarded, and RNA pellet was air dried for 10-15 min. The concentration and the purity for each RNA sample were determined using NanoDrop (ND-1000 Spectrophotometer, NanoDrop Technologies, Wilmington, Delaware, USA).
Complementary DNA (cDNA) was synthesized using Quanti-Tect Reverse Transcription kit (Qiagen, Germany) according to the manufacturer's instructions. RT-PCR reaction mixture was composed of cDNA, 4 µL of the 5X FIREPol Eva Green, qPCR Mix (Solis Bio-Dyne, Estonia), 2 nM of each primer ( Table 2), and RNAase free water to final volume 20 µL. RT-PCR was performed as follows: one cycle at 95°C for 15 min, followed by 40 cycles each cycle programmed as denaturation at 95°C for 15 s, annealing for 30 s and extension at 72°C for 30 s. RT-PCR was performed using a Rotor-Gene Q thermocycler (Qiagen, Germany). A negative control containing RNAase/ DNAase-free water instead of cDNA was included in each run.
The expression of the target genes in both treated and untreated samples was analyzed and normalized against rpoD expression. The level of gene expression in untreated and treated samples was calculated relative to the untreated PAO1.

Molecular Docking
CFP, CFPC, and CFPF were docked into the active site of P. aeruginosa ligand-binding domain at PDB ID: 1RO5 (Gould et al., 2004), PDB ID: 2UV0 (Bottomley et al., 2007) and PDB ID: 4JVD (Ilangovan et al., 2013) to evaluate their binding affinities, and to determine their inhibition activities and binding modes at the active site of LasI, LasR, and PqsR, respectively. The crystal structures of LasI, LasR, and PqsR were picked up from the protein data bank. All bound water ligands were removed from the protein. All components were constructed on ChemBioDraw using ChemBioOffice ultra v.14 software "ChemOffice, scientific personal productivity tools -PerkinElmer Informatics". The energy was minimized by using MM2, Jop Type. Docking studies were performed using the Molsoft program, internal coordinate mechanics (ICM) 3.4-8C was applied as reported by Abagyan et al. (1994) by converting PDB file 1RO5, 2UV0, and 4JVD into an ICM object. ICM aims to find the global minimum energy that describes the interaction between the ligand and the receptor. The modes of the interaction of the autoinducer molecule 3-oxo-C12-HSL within 2UV0 and 2-nonyl-4hydroxyquinoline (NHQ) within 4JVD were used as standard docked models.

Statistical Analysis
Each experiment was performed in triplicate. Mean and standard deviations of three independent measurements were calculated by Excel data package. Statistical analysis was performed using the GraphPad Instate software package (version 3.1) using T paired test for comparing the treated and untreated cultures. The results were assigned as significant when p ≤0.05, and highly significant when p ≤0.01 or p ≤0.001.  Figure 1). However, CFP derivatization with MnCl 2 , NiCl 2, and CrCl 3 did not exhibit QSI activity (Supplementary Figure 1). Therefore, CFP and metalliccomplexes CFPF, and CFPC were selected to study their effects on virulence factors of P. aeruginosa clinical isolates and standard strains.

Effect of Sub-Inhibitory Concentrations of the Tested Compounds on the Growth of P. aeruginosa Isolates
Viable count of P. aeruginosa was estimated after treating each isolate with 1/2 MIC of CFP, CFPC, and CFPF. Cultivation of P. aeruginosa with sub-MICs did not affect bacterial growth when compared to the control untreated cultures (Supplementary Table S1). For instance, the viable count of Ps1 isolate was 152, 155 and 160 × 10 7 CFU/mL when treated with 1/2 MIC of CFP, CFPC, and CFPF, respectively and the bacterial count of untreated Ps1 was 162× 10 7 CFU/mL. The counts of Ps2 treated with 1/2 MIC of CFP, CFPC, and CFPF were 140, 145 and 138 ×10 6 CFU/mL, respectively while the count of untreated culture was 146× 10 6 CFU/mL. Also, the viable count of Ps3 was 132, 129, and 134 ×10 7 CFU/mL upon addition of 1/2 MIC of CFP, CFPC, and CFPF, respectively while the count of untreated culture was 136× 10 7 CFU/mL. Furthermore, the viable count of PAO1 treated with 1/2 MIC of CFP, CFPC, and CFPF were 166, 160, and 164 ×10 7 CFU/mL, respectively and the viable count of untreated PAO1 was 168× 10 7 CFU/mL. Similarly, PA14 count was 175, 168, and 166 ×10 7 CFU/mL when cultivated in 1/2 MIC of CFP, CFPC, and CFPF, respectively and the viable count of the untreated culture was 177× 10 7 CFU/mL. The viable count of PAO-JP2 culture was 181, 184, 180 ×10 7 CFU/mL when cultivated with 1/2 MIC of CFP, CFPC, and CFPF, respectively and the viable count of the untreated culture was 187× 10 7 CFU/mL. Additionally, the OD 600 nm of treated and untreated cultures was measured at different time intervals. There was no effect on the bacterial growth over time in cultures treated with 1/2 MIC of CFP, CFPC, and CFPF compared to untreated cultures (Supplementary Figure S2).

Effect of Sub-MICs of the Tested Compounds on Virulence Factors
The influences of sub-MICs; 1/2 and 1/4 of CFP, CFPC, and CFPF on the virulence factors; pyocyanin, protease, hemolysin, and biofilm of P. aeruginosa clinical isolates Ps1, Ps2 and Ps3, and standard strains PAO1, PA14, and PAO-JP2 were investigated and compared with untreated cultures at the same conditions.  (Figure 2).    (Figure 3).  Figure 4).

Expression of QS-Regulated Genes
To get insights into the molecular mechanism of CFP, CFPC and CFPF in reducing QS of P. aeruginosa. RT-PCR was used to assess relative expression levels of lasI, and rhlI genes. CFP, CFPC and CFPF was found to significantly (p < 0.001) reduce the transcription of lasI by 77.25%, 94% and 88%, respectively in PAO1 compared to the untreated samples ( Figure 5A). Also, significant inhibition in rhlI gene expression in P. aeruginosa PAO1 was elucidated when subjected to CFP, CFPF and CFPC at sub-MIC. The relative expression data showed a significant reduction in rhlII gene expression by 44% in CFP treated cells (p < 0.01), 41% reduction in CFPF treated culture (p < 0.01) and 80% reduction in CFPC treated cells (p < 0.001) when compared to the untreated cells ( Figure 5B).

Binding Affinity Analysis for LasR and Ligands
The ICM score and hydrogen bonds between compounds and the surrounding amino acids were used to predict the binding modes, their binding affinities, and the orientation of these compounds at the active site of LasI, LasR, and PqsR. The scoring functions of the compounds were calculated from minimized ligand-receptor complexes. CFP, CFPC and CFPF were docked at the binding site of LasI. CFP showed high ICM score of -123.78 and formed seventeen hydrogen bonds with Arg30, Arg104, Ile107, Thr144, Lys150, Arg172, and Glu171. Both derivatives; CFPC and CFPF gave higher ICM scores of -132.85 and -191.69, respectively (Table 4). CFPC binds with the receptor by ten hydrogen bonds with Lys167, Arg172, Ile170, and Glu171. While CFPF form nine hydrogen bonds with Arg30, Arg104, Thr144, Thr145, Phe105, and Thr121 ( Figure 6, Supplementary Table S2).
Molecular docking of CFP, CFPC and CFPF with LasI, LasR, and PqsR of P. aeruginosa are indicated in the Supplementary Tables S2-S4.
Additionally, P. aeruginosa has a pronounced ability for biofilm assembly. Pseudomonas biofilms are resistant to most of the conventional antibiotic regimens. Inhibition of biofilm formation makes bacterium more susceptible to immune system and phagocytosis by neutrophils (Hentzer et al., 2002;Hentzer and Givskov, 2003). CFP, CFPC, and CFPF at 1/2 and 1/4 MICs showed a significant reduction in biofilm formation (Figure 4). Quorum sensing regulates the production of several extracellular virulence factors and promotes biofilm maturation, meaning that it has a key role in the pathogenesis of P. aeruginosa (Wagner et al., 2003). Azithromycin (Swatton et al., 2016), imipenem (El-Mowafy et al., 2017), piperacillin/tazobactam (Aleanizy et al., 2021) inhibit biofilm formation via interruption with the QS cascade. In addition, QSI activity of garlic renders P. aeruginosa biofilms sensitive to tobramycin and phagocytosis (Bjarnsholt et al., 2005).
On the molecular level, we tested the activity of CFP, CFPC, and CFPF on the expression of QS genes lasI, and rhlI. Data revealed that CFP, CFPC, and CFPF at 1/2 MIC exhibited significant reduction in the relative expression of lasI by 77.25%, 94% and 88%, respectively ( Figure 5). The decrease in the expression level of lasI was consistent with reduction in all virulence factors controlled by lasI/R. The interaction of LasR with 3-oxo-C12-HSL, induces las system and triggers the production of virulence factors such as elastase, alkaline protease, hydrogen-cyanide synthase, exotoxin A, and secretion apparatus (Pessi et al., 2001). Previous researches elucidated that inhibition of lasI/R QS circuit is associated with reduction of P. aeruginosa virulence traits including elastase, total protease, hemolysin, and biofilm (Tateda et al., 2001;Gupta et al., 2011).
Also, the relative expression of rhlI was significantly reduced by CFP, CFPC, and CFPF ( Figure 5). The lasI/R and rhlI/rhlR circuits are interconnected and suppression of lasI is associated with subsequent inhibition of rhlI gene expression . The rhl system is also necessary for optimal production of lasB elastase, lasA protease, pyocyanin (Latifi et al., 1996). Furthermore, the rhl system is involved in the regulation of hcnABC, and alkaline protease Pessi et al., 2001).
Hence, the tested compounds CFP, CFPC, and CFPF when docked into the active sites of LasI, LasR and PqsR receptors, they bound with the important amino acids with high ICM scores (Figures 6-8). This indicates that the tested compounds are potential inhibitors of QS and related virulence factors. Data is also supported by phenotypic and molecular studies.
In conclusion, this study explores a new character of CFP to comprise significant inhibition of QS and elimination of associated virulence factors. CFPC and CFPF at 2-8 folds lower concentration than CFP significantly eliminated QS and virulence factors of P. aeruginosa including (1) pyocyanin production (2) hemolysin (3) protease and (4) biofilm formation without affecting microbial growth. The interpretation of QSI activity of CFP, CFPC and CFPF was assessed by the molecular model analysis. It may be useful to investigate in vivo activity of CFPC and CFPF in the treatment of P. aeruginosa infections.