The media used in this study were Luria-Bertani (LB) broth, LB agar and tryptone (Lab M Limited, Lancashire, United Kingdom), Mueller Hinton broth, Mueller Hinton agar, and tryptone soya broth (Oxoid, Hampshire, UK). Dimethyl sulphoxide (DMSO), febuxostat, elastin congo red, and N-hexanoylhomoserine lactone were the products of Sigma (St. Louis, USA). Other chemicals were of pharmaceutical grade.

Pseudomonas aeruginosa PAO1 strain and Chromobacterium violaceum CV026 mutant strain were obtained from the Department of Microbiology, Faculty of Pharmacy, Mansoura University, and the Department of Microbiology, Faculty of Pharmacy, Ain Shams University, respectively.

In order to determine the minimum inhibitory concentration of febuxostat, the broth microdilution method was used [14] . Two fold serial dilutions of febuxostat in Mueller-Hinton broth were prepared and then added to the wells of 96-wells microtiter plate in 100 µl aliquots. PAO1 suspension in Mueller-Hinton broth was prepared from an overnight culture with approximate cell density of 1 × 10⁶ CFU/ml and 100 µl of the prepared suspension were added to each dilution. The minimum inhibitory concentration was calculated as the lowest concentration of the drug that showed no visible growth after 20 h of incubation of the plate at 37˚C.

To guarantee that febuxostat has no effect on the growth of PAO1, the turbidities of PAO1 in the presence and absence of 1/8 MIC of febuxostat was determined according to Nalca et al. [15] . Overnight culture of PAO1 was prepared and used to inoculate LB broth with and without febuxostat. LB broth tubes were incubated at 37˚C for overnight and the turbidities of LB cultures treated with febuxostat and control LB cultures were measured at 600 nm by using Biotek Spectrofluorimeter (Biotek, USA).

The ability of febuxostat to inhibit the quorum sensing-regulated violacein pigment was assessed using the method of Choo et al. [16] . Chromobacterium violaceum CV026 cultured in LB broth and incubated overnight. The bacterial suspension optical density was adjusted to 1 at 600 nm and aliquots of 100 μl of LB broth with the autoinducer N-hexanoyl homoserine lactone in the presence and absence of 1/8 MIC of febuxostat were delivered to the wells of a 96-well microtiter plate followed by the addition of aliquots of 100 μl of the prepared suspension. The plate was incubated for 16 h at 28˚C and was then left at 60˚C until complete dryness. Violacein pigment was eluted by addition of aliquots of 100 μl of DMSO and incubation at 30˚C with shaking. Negative control was also prepared by using DMSO alone and the absorbance at 590 nm was determined by using Biotek Spectrofluorimeter (Biotek, USA).

The potential biofilm inhibition by febuxostat was estimated according to the modified method of Stepanovic et al. [17] . Overnight culture of PAO1 in TSB was prepared and diluted to achieve an approximate cell density of 1 × 10⁶ CFU/ml. Aliquots of 0.1 ml of the bacterial suspension were delivered into the wells of 96 well sterile microtiter plate in the presence and absence of 1/8 MIC of febuxostat. The plate was incubated at 37˚C for 24 h and then the free swimming planktonic cells were aspirated. The wells were washed 3 times using sterile phosphate buffered saline (PBS, pH 7.2) and the adherent cells were fixed by methanol (99%) for 20 minutes. The wells were stained with crystal violet (1%) for a period of 20 minutes. Excess stain was washed off with distilled water and glacial acetic acid (33%) was used to elute crystal violet after air-drying of the plate. The absorbance of solubilized stain was measured at 590 nm with Biotek spectrofluorimeter (Biotek, USA).

In order to further analyze the inhibition of biofilm formation by febuxostat, biofilms were formed on sterilized cover slips placed in 50 ml centrifuge tubes containing TSB with and without 1/8 MIC of febuxostat and inoculated with P. aeruginosa PAO1 suspension adjusted to optical density of 600 nm at 1. After incubation of the tubes at 37˚C for 16 h, the cover slips were removed and washed with phosphate-buffered saline to remove any planktonic cells. The attached biofilms were fixed with methanol, stained with crystal violet (1%) and examined under the light microscope using the high power (400× magnification).

To detect the possible inhibitory activities of febuxostat on swimming and twitching motilities, the modified method of Rashid andKornberg was used [18] . Swimming agar plates (tryptone 1%, sodium chloride 0.5% and agar 0.3%) containing febuxostat (1/8 MIC) and control plates were prepared. The plates were stabbed in the center with 5 µl of diluted overnight culture of PAO1 in tryptone broth. After incubation of the plates for 24 h at 37˚C, the swimming zones were measured. For assay of twitching motility inhibition, 2 µl of the prepared culture were used to stab-inoculate LB agar plates (1%) with and without febuxostat (1/8 MIC) and incubated at 37˚C for 48 h. To measure the twitching zones, the agar was removed, stained with crystal violet after air-drying of the plates. The stain was removed; the plates were washed with water and dried.

The ability of febuxostat to inhibit the proteolytic activity of PAO1, the skim milk agar method was used [19] . Overnight cultures of P. aeruginosa PAO1 in LB broth in the presence and absence of 1/8 MIC of febuxostat were centrifuged at 10,000 rpm for 15 min and the supernatants were separated. Aliquots of 100 µl of the supernatants were added to the wells made in 5% skim milk agar plates. After overnight incubation of the plates at 37˚C, the clear zones around the wells were measured in order to detect the protease inhibitory activity of febuxostat.

The possible elastase inhibition by febuxostat was assessed by the elastin congo red assay [20] . Supernatants of untreated and febuxostat treated P. aeruginosa PAO1 cultures were added in aliquots of 500 µl to elastin congo red reagent tubes. The tubes were incubated for 6 h at 37˚C with shaking and any insoluble elastin congo red was removed by centrifugation and the absorbance was measured at 495 nm using Biotek Spectrofluorimeter (Biotek, USA).

The reduction in production of the blue-green pyocyanin pigment by febuxostat treated PAO1 was estimated according to Das and Manefield [21] . PAO1 was cultured in LB broth and incubated overnight. The bacterial suspension was diluted to achieve an approximate optical density of 0.4 at 600 nm. LB broth tubes each contains 1 ml with 1/8 MIC of febuxostat and control LB broth tubes were prepared and inoculated with 10 μl of the prepared suspension. After incubation at 37˚C for 48 h, the tubes were centrifuged at 10,000 rpm for 10 minutes and the supernatants were separated. The pyoyanin was estimated by measuring the absorbance at 691 nm by Biotek Spectrofluorimeter (Biotek, USA).

Hemolytic activity of PAO1 in the presence and absence of febuxostat was determined according to Rossignol et al. [22] . The prepared cell free supernatant (500 µl) was mixed with 700 µl of fresh 2% saline suspension of erythrocytes in saline. After incubation at 37˚C for 2 h, the hemoglobin released from lysed erythrocytes was separated by centrifugation at 2500 g for 5 minutes at 4˚C and measured at 540 nm. The hemoglobin released was compared with positive control prepared by addition of 0.1% SDS to erythrocyte suspension to cause complete hemolysis and negative control prepared by incubation of erythrocytes in LB broth at the same conditions. To calculate percentage hemolysis, the following formula was used:

% hemolysis = [X − B/T − B] × 100, (X: the treated and untreated samples, B: the negative control, T: the positive control). The hemolysis of febuxostat treated cultures was expressed as % compared to hemolysis of untreated culture.

RNA of febuxostat-treated and untreated PAO1 cultures was extracted using GeneJET RNA Purification Kit (Thermoscientific, USA) according to manufacturer instructions. Overnight cultures of PAO1 in the presence and absene of 1/8 MIC of febuxostat were prepared and centrifuged at 12,000 g for 2 minutes to collect the pellets. The Pellets were then resuspended in 100 µl of Tris-EDTA buffer containing lysozyme. After incubation for 5 minutes at 25˚C, the lysis buffer containing β-mercaptoethanol was added and mixed thoroughly. After purification, RNA was eluted by addition of 100 µl of nuclease-free water and stored at −70˚C until use.

To determine the effect of febuxostat on quorum sensing genes, the relative expression levels of QS genes were analyzed in PAO1 strain treated and untreated with febuxostat by qRT-PCR. RopD gene was used to normalize the relative expression level of each gene beause the the expression levels of this gene undergo no change in febuxostat treated and untreated PAO1 cultures. Table 1 shows the primers used in this study. The protocol described in SensiFAST™ SYBR^® Hi-ROX One-Step Kit (Bioline, UK) was used for this analysis employing StepOne Real-Time PCR system (Applied Biosystem, USA). Agarose gel electrophoresis and a melting curve analysis of products were used to confirm the specific PCR amplification according to the manufacturer recommendation. The relative gene expression was calculated by using the comparative threshold cycle (∆∆C_t) method [23] .

Docking study was used to determine the molecular interaction of febuxostat with the LasR, RhlR receptors. The crystal structure of P. aeruginosa LasR ligand

*F = Forward, **R = Reverse.

binding domain and the rhlr receptor model (ID: P54292.1) were obtained from Protein Data Bank (PDB ID: 2UV0) [24] and the protein model portal [25] respectively. Febuxostat and the co-crystalized natural ligand, 3-oxo-dodecanoylhomoserine lactone were docked into the receptor active site using Molegro Virtual Docker (MVD Version 6.0). Febuxostat was drawn into Marvin Sketch V5.11.5 [26] and the most energetically favored conformer was saved as (*mol²) file format for docking. The optimal geometry of the ligand was determined during the docking process. The E monomer was selected for analysis, water was discarded, cavities were determined and the search area was set to be 9 Å from the active site center. Docking process was performed by using MolDock optimizer algorithm with 10 runs/ligand, 150 population size, 4000 max iteration and 5 poses for each ligand. MolDock docking engine [27] using docking template and the optimized ligands was executed. At last, the top returned poses were selected for analysis. For docking febuxostat and the autoinducer, butanoyl homoswerine lactone (C4-HSL), into the active site of the rhlr receptor model, the same procedure was used.

In unpaired t tests, Graph Pad Prism 5 was used to investigate the significance of the inhibitory activities of febuxostat against swarming, swimming and twitching motilities. In paired t tests, Graph Pad Prism 5 was used to detect the significance of the effects of febuxostat against biofilm formation and eradication, protease, violacein, pyocyanin, elastase and hemolysin.

P values < 0.05 were considered statistically significant.

Febuxostat showed growth inhibitory effect against Pseudomonas aeruginosa PAO1 at a concentration of 8 mg/ml. Then, 1/8 MIC (1 mg/ml) was used to assess the activity of febuxostat against quorum sensing and virulence factors of PAO1.

Potential inbition of QS and virulence by febuxostat may be due to its effect on growth of PAO1. To avoid such possibility, PAO1 overnight cultures in LB broth with and without 1/8 MIC of febuxostat were prepared and the turbidities of the bacterial suspensions were measured at 600 nm. The difference in turbidities as a measure of growth rates in febuxostat treated and untreated PAO1 cultures was not statistically significant.

Febuxostat exerted a significant biofilm inhibitory activity. Biofilm formation by PAO1 was reduced by 64.4% (Figure 1(a)). In order to further investigate the Antibiofilm effect of febuxostat, biofilms formed on sterile glass cover slips in the absence and presence of febuxostat were stained with crystal violet and examined under the light microscope. The febuxostat treated sample showed much fewer scattered cells.

As the production of the intracellular violacein pigment in the reporter strain C. violaceum CV026 is well reported to be regulated by the quorum sensing machinery, the anti-quorum sensing activity of febuxostat was evaluated by measuring the reduction in violacein pigment. At first the effect of sub-MIC of febuxostat on growth of C. violaceum CV026 was investigated and no significant difference in growth was found between the treated and untreated samples. Febuxostat at 1/8 MIC diminished violacein production in C. violaceum CV026 by 71.61% (Figure 1(b)).

The supernatant of the febuxostat treated and untreated cultures were added to the wells in skim milk agar and the clear zones around them were measured. Febuxostat could inhibit protease by 55.18% (Figure 1(c)).

The blue green pyocyanin pigment is a redox metabolite in P. aeruginosa that has pulmonary cytotoxic effect [28] . In the presence of 1/8 MIC of febuxostat, pyocyanin production was lowered by 63.63% (Figure 1(d)).

The ability of elastase enzyme produced by febuxostat treated and untreated PAO1 culture supernatant to decompose elastin congo red was assayed. Elastolytic activity was reduced with febuxostat by 60.96% (Figure 1(e)).

Febuxostat at 1/8 MIC reduced the ability to PAO1 to swim, twitch and swarm by 70.12% (Figure 1(f)), 83.72% (Figure 1(g)), and 85.17% (Figure 2(a)), respectively.

The hemolytic activity of untreated PAO1 and febuxostat treated culture supernatants was assayed quantitatively and the hemoglobin release was measured spectrophotometrically. Febuxostat could inhibit hemolytic activity of PAO1 by 73.94% as compared to the untreated control.

To investigate the effect of febuxostat on the expression levels of QS genes of PAO1, qRT-PCR was performed (Figure 2(b)). The expression of QS genes was assessed in febuxostat treated and untreated PAO1 using 2^−∆∆Ct method. Significant decrease in the expression levels of LasI, LasR, RhlI, RhlR, PqsA and PqsR was found as compared to control PAO1 culture. The relative expression of virulence factors regulating genes were decreased from 100% in untreated PAO1 to 39.54% for LasI, 43.27% for LasR, 45.19% for RhlI, 69.87% for RhlR, 61.52% for PqsA and 28.24% for PqsR.

The molecular docking study included the quorum sensing proteins LasR and rhlr as targets for P. aeruginosa virulence inhibition. The natural ligand and febuxostat were docked into LasR active site. The binding interaction with the receptor is shown in (Figure 3(a)). The MolDock scores for febuxostatand the natural ligand along with the interacting residues are shown in Table 2. The inhibitor lacks the hydrophobic side chain that induces the correct formation of the hydrophobic core of LasR. The natural ligand is a LasR inducer that can stabilize the conformational change through its long acyl hydrophobic chain. Febuxostat has no acyl hydrophobic chain thus act as LasR inhibitor. The high docking score indicates a promising LasR antagonist activity for febuxostat.

Febuxostat was docked into rhlr receptor model active site. The binding interaction with the rhlr receptor is shown in Figure 3(b). It showed similar interaction to the auoinducer C4-HSL as shown in Table 3. The important hydrophobic acyl group of C4-HSL is responsible for conformational change to act as inducer that is absent in febuxostat which make it a possible inhibitor with a promising high antagonist activity as indicated from high docking score.