Vanillic acid from Actinidia deliciosa impedes virulence in Serratia marcescens by affecting S-layer, flagellin and fatty acid biosynthesis proteins

Serratia marcescens is one of the important nosocomial pathogens which rely on quorum sensing (QS) to regulate the production of biofilm and several virulence factors. Hence, blocking of QS has become a promising approach to quench the virulence of S. marcescens. For the first time, QS inhibitory (QSI) and antibiofilm potential of Actinidia deliciosa have been explored against S. marcescens clinical isolate (CI). A. deliciosa pulp extract significantly inhibited the virulence and biofilm production without any deleterious effect on the growth. Vanillic acid was identified as an active lead responsible for the QSI activity. Addition of vanillic acid to the growth medium significantly affected the QS regulated production of biofilm and virulence factors in a concentration dependent mode in S. marcescens CI, ATCC 14756 and MG1. Furthermore vanillic acid increased the survival of Caenorhabditis elegans upon S. marcescens infection. Proteomic analysis and mass spectrometric identification of differentially expressed proteins revealed the ability of vanillic acid to modulate the expression of proteins involved in S-layers, histidine, flagellin and fatty acid production. QSI potential of the vanillic acid observed in the current study paves the way for exploring it as a potential therapeutic candidate to treat S. marcescens infections.

In recent years, awareness of the consumption of fruits for health promotion has increased. Actinidia deliciosa (Kiwifruit) has been classified as an excellent source of vitamin C, dietary fibre, vitamin E and potassium, based on the US FDA's definition 13 . Kiwifruit has been known for its antioxidant, cardiovascular preventive and laxative activities 14 . Furthermore, consumption of kiwifruit selectively enhances the growth of intestinal bacteria such as Bifidobacterium and Lactobacillus 15 . Since the QSI potential of kiwifruit has never been explored and the presence of a wide range of secondary metabolites in kiwifruit has eventually led our focus to explore its QSI potential against S. marcescens.

Results and Discussion
Evaluation of QSI potential of A. deliciosa pulp extract (ADPE) against S. marcescens and identification of active principle. A number of studies have shown that bacteria, fungi and plants produce QSI compounds to protect themselves and compete with invading organisms. Due to their non-toxic and multifunctional nature, QSIs from edible sources are gaining attention in augmenting antimicrobial therapy 16 . S. marcescens strains from environmental and clinical origins were able to produce the intracellular prodigiosin pigment. Since prodigiosin production is under the direct regulation of the QS signalling mechanism 17 ; the QSI potential of ADPE was initially assessed by prodigiosin assay. ADPE exhibited concentration-dependent prodigiosin inhibitory activity (Fig. 1a). At a higher concentration (20%) ADPE greatly reduced the prodigiosin production (85%) in S. marcescens when compared to control (Fig. 1a). The prodigiosin inhibitory activity of ADPE clearly suggests the presence of QS inhibitor(s).
In addition, dose-dependent inhibition of QS-regulated production of haemolysin ( Fig. 1a) and biofilm formation (Fig. 1a) was observed with ADPE treatment without any reduction in the growth of marcescens (Fig. 1b). Light microscopic observation of biofilms formed in the presence and absence of ADPE further confirmed the potent antibiofilm activity (Fig. 1c). These results affirm the ability of ADPE to interfere with the QS mechanism of S. marcescens.
To identify the active lead(s) responsible for the QSI activity of ADPE, the ethanolic extract of ADPE (ADPEE) was extracted subsequently with hexane (ADPEE-H), chloroform (ADPEE-C), ethyl acetate (ADPEE-E) and butanol (ADPEE-B). The QSI activity of the dried solvent extracts was assessed based on its ability to reduce the QS-regulated prodigiosin production. The highest percentage of prodigiosin pigment reduction was observed in the ADPEE-C extract at 2 mg/mL (Fig. 1d). Henceforth, the effect of ADPEE-C extract on biofilm formation, production of protease and lipase production of S. marcescens was analysed. The ADPEE-C extract also showed a concentration-dependent inhibition of biofilm, protease and lipase (Fig. 1e) without any interference in the growth of S. marcescens (Fig. 1f).
The major constituents of ADPEE-C extract were identified using GC-MS analysis. The major components found in the ADPEE-C extract are vanillic acid (34.43%) and n-Hexadecanoic acid (11.85%) ( Supplementary  Fig. 1). Other major constituents present in the ADPEE-C are listed in supplementary Table 1. Vanillic acid is a well-known generally regarded as safe (GRAS) flavoring agent with antioxidant, anti-lipid peroxidative, anti-inflammatory and neuroprotective/cognitive effects.

Vanillic acid inhibits the QS regulated virulence factors and biofilm formation in S. marcescens.
Pure vanillic acid was assayed for its QSI activity against S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 virulence factors. Extracellular proteases of S. marcescens modulate the host's immune response by inducing the inflammatory response 18 . Compounds/extracts inhibiting the production of protease can be used to potentiate the host's innate immune response as well as antimicrobial therapy. In this context, the concentration-dependent decrease in the production of the secreted protease was observed in the cell free culture supernatant (CFCS) of S. marcescens CI (Fig. 2a), S. marcescens ATCC 14756 (Fig. 2b) and S. marcescens MG1 (Fig. 2c) grown in the presence of vanillic acid.
Lipolytic enzymes are utilised by most of the bacterial pathogens to establish the infection by degrading the phospholipid bilayer of host cells and to manipulate the host cell signalling pathways 19 . Production of lipase was significantly reduced in all the three S. marcescens used in the present study. Briefly, 94 and 62% of inhibition of lipase production was noted in S. marcescens CI (Fig. 2a) and S. marcescens MG1 (Fig. 2b) respectively, at 250 µg/ mL of vanillic acid. Whereas at 500 µg/mL of vanillic acid 65% of lipase inhibition was observed in S. marcescens ATCC 14756 (Fig. 2c). Haemolysin is a well-characterised virulence factor of S. marcescens. Vanillic acid reduced the haemolytic activity of S. marcescens CI and S. marcescens ATCC in a concentration-dependent mode. At 250 and 500 µg/mL of vanillic acid 68.75 and 56.52% of haemolysin production was observed in S. marcescens CI (Fig. 2a) and S. marcescens ATCC (Fig. 2b), respectively. These results show that vanillic acid has a potential to interfere with the QS mechanism of S. marcescens.
Biofilm formation has been considered as a major event in the establishment of chronic infections and confers extreme resistance to a broad spectrum of antibiotics. Vanillic acid has been reported for its ability to inhibit QS-dependent violacein pigment production in C. violaceum 20 and biofilm formation in Aeromonas hydrophila 21 . Since QS is involved in the biofilm formation of S. marcescens 22 , the anti-QS activity of vanillic acid is expected to have a significant negative effect on biofilm formation as well as swarming motility. In the present study, vanillic acid exhibited dose-dependent antibiofilm activity against S. marcescens CI (Fig. 2a), S. marcescens ATCC (Fig. 2b) and S. marcescens MG1 (Fig. 2c). Briefly, 63.6 and 57.64% of biofilm inhibition was noted at 250 µg/mL of vanillic acid in S. marcescens CI (Fig. 2a) and S. marcescens MG1 (Fig. 2c), respectively. At the same time, 50.84% (Fig. 2a) of biofilm inhibition was observed in S. marcescens ATCC grown in the presence of 500 µg/ mL of vanillic acid.
In addition, biofilms of S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 formed in the absence and presence of vanillic acid were observed under light microscope and the results revealed the reduction in biofilm formation and surface coverage (Fig. 2d). CLSM analysis also further confirmed the antibiofilm potential of vanillic acid (Fig. 2e). Swarming motility confers adaptive resistance to antibiotics and allows pathogenic bacteria Quorum sensing inhibitory potential of ADPE against S. marcescens. The ADPE inhibited the QS regulated virulence factors such as biofilm formation, prodigiosin, protease, hemolysin and lipase (a) in a dose dependent manner without inhibiting the growth of S. marcescens (b). Light microscopic visualization of S. marcescens biofilm formed in the absence and presence of ADPE (c). Effect of solvent extracts of ADPEE against prodigiosin production and the growth (d) of S. marcescens. ADPEE-C exhibited the concentration dependent inhibition of biofilm formation, protease and lipase (e) without affecting the growth (f). Growth of S. marcescens was measured at 600 nm after 18 h incubation at 30 °C. Error bars represent standard deviations from the mean (n = 6 [biological triplicates in experimental duplicates]). Statistical significance was analyzed using one way ANOVA-Duncan's post-hoc test. a, b, c and d indicate the significant difference p < 0.05, p < 0.01, p < 0.005 and p < 0.001, respectively.

Figure 2.
Quorum sensing inhibitory potential of active lead vanillic acid. Effect of pure vanillic acid on the QS regulated biofilm formation and prodigiosin, protease, hemolysin and lipase production of S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 (a). Light (b) and CLSM [Three dimensional micrographs] (c) analyses corroborate the antibiofilm activity of vanillic acid. Effect of vanillic acid on the swarming motility of S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 (d). Error bars represent standard deviations from the mean (n = 6 [biological triplicates in experimental duplicates]). Statistical significance was analyzed using one way ANOVA-Duncan's post-hoc test. a, b, c and d indicate the significant difference p < 0.05, p < 0.01, p < 0.005 and p < 0.001, respectively. to move across the moist, solid and viscous medium. Presence of vanillic acid in the growth medium was found to inhibit the swarming motility of all the three S. marcescens strains used in the present study (Fig. 2f).
As an ideal QSI is expected to have no/least interference with the basal growth of bacteria 23 , the effect of vanillic acid on the growth of S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 was assessed by measuring the cell density at 600 nm. Results of growth measurement clearly suggest that vanillic acid does not have antibacterial activity at the tested concentrations. In addition, growth curve analysis was also carried out and the results revealed the non-antibacterial/bacteriostatic nature of vanillic acid at 125, 250 and 500 µg/mL (Fig. 3a,b and c).
FT-IR analysis showed the presence of polysaccharides, nucleic acid, proteins and fatty acids in extracted EPS of S. marcescens control and vanillic acid treated (Fig. 3c). The presence of peaks at 1,200-900 cm −1 indicates the C-OH stretching and C-O-C, C-O ring vibrations of carbohydrates 24,25 peaks at 1,250-1,220 cm −1 corresponds to the phosphodiester, free phosphate, and monoester phosphate functional groups of phosphodiester, DNA/RNA backbone, phospholipids and phospho sugars 26 . Peaks at 1,540 and 1,650 cm −1 indicated the N-H bending, C-N stretching of Amide II and C = O stretching, C-N bending of Amide I proteins, respectively 25 (Fig. 3c). In addition, a peak at 3,000-2,800 cm −1 corresponds to C-H vibrations of the functional groups of fatty acids 24,25 . The polysaccharide region (1,200-900 cm −1 ) of vanillic acid treated EPS was found to be increased slightly. However reduction in the absorbance of a peak at 1120 cm −1 -1140 cm −1 which corresponds to the C-O-C stretching vibrations of carbohydrates was observed in vanillic acid treated EPS. Whereas the decrease in the IR absorbance at 1,250-1,220 cm −1 (nucleic acid) 27,28 , 1,540 and 1,650 cm −1 (Amide II and Amide I proteins) 26 , 3,000-2,800 cm −1 (fatty acid) 25,29 was observed in vanillic acid treated EPS. Hence alterations in the biofilm architecture upon vanillic acid treatment could be attributed to the reduction of nucleic acid, amide I and II proteins and fatty acid content of EPS. When compared to the control, vanillic acid treated EPS showed a decrease in the absorbance of a broad peak at 3,800-3,100 cm −1 corresponding to -OH group (responsible for the hydration of EPS). Comparison of FT-IR spectra of control and treated EPS revealed a noticeable decrease in the IR absorbance of the nucleic acid, protein and fatty acid content and not in the carbohydrate content of the EPS (Fig. 3c). In addition to carbohydrate, other macromolecules such as nucleic acid, protein and fatty acid are also the main constituents required for the formation of EPS matrix. The decrease in the nucleic acid, protein and fatty acid content of EPS upon vanillic acid treatment could be one of the underlying mechanisms of its antibiofilm activity. Furthermore, the EPS inhibitory potential of vanillic acid is expected to decrease the survival of S. marcescens under in vivo conditions. In vivo anti-QS and antibiofilm activity of vanillic acid. C. elegans has been used as one of the simple and successful in vivo preclinical models to study the bacterial pathogenesis and screening of bioactive compounds with antibacterial, anti-QS and antibiofilm activities etc 30 . Anti-QS and antibiofilm agents have been already reported to rescue/increase the survival of C. elegans during pathogenic infection 31,32 . Hence, in the present study the in vivo anti-QS and antibiofilm activity of vanillic acid against S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 are evaluates using C. elegans. Worm liquid-killing assay was used to evaluate the ability of the vanillic acid to rescue the C. elegans from S. marcescens infection. No significant difference in the survival of worms fed with E. coli OP50 + vehicle control and E. coli OP50 + 500 µg/mL of vanillic acid confirms the non-toxic nature of vanillic acid (Fig. 4a). In addition, results of killing assay revealed that S. marcescens CI (Fig. 4a) is more pathogenic to C. elegans than S. marcescens ATCC (Fig. 4b) and S. marcescens MG1 (Fig. 4c). Briefly, the complete killing (100%) was observed at 78 h in S. marcescens CI infected C. elegans (Fig. 4a). Whereas complete killing of C. elegans was observed at 90 h and 102 h upon S. marcescens ATCC (Fig. 4b) and S. marcescens MG1 (Fig. 4c), respectively. Interestingly, vanillic acid increases the survival of C. elegans during S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 infection by 51 (Fig. 4a) Effect of vanillic acid treatment on the cellular proteome of S. marcescens. The results of the virulence assays prompted us to study the underlying molecular mechanism of QS inhibition of vanillic acid in S. marcescens. To get an overview, cellular proteins (each 30 µg) from the exponential phase cultures of S. marcescens CI grown in the presence and absence of vanillic acid (250 μg/mL) were separated in SDS-PAGE. CBB stained SDS-PAGE gel showed interesting protein profile variations in the cellular proteome of S. marcescens cultured in the presence of vanillic acid when compared to the control. Then, the cellular proteome of control and vanillic acid treated S. marcescens CI were analysed using 2DGE in biological triplicates ( Supplementary Fig. 2). The protein spots present in the control and treated gels were detected and matched using Image Master Platinum 7 software (GE Healthcare). Based on the densitometric analysis, among the detected spots (579), 27 spots and 21 spots were found to be down-regulated and upregulated by more than 1.5 fold, respectively (Fig. 5). Based on statistical significance (ANOVA 0.05), 19 down-regulated and 14 upregulated protein spots were selected for protein identification. Protein spots with more than 2 fold differential expression were identified using nano-LC MS/MS analysis. Furthermore, 1.9 to 1.5 fold differentially expressed protein spots were identified using MALDI-TOF/ TOF analysis. The list of differentially expressed proteins and their functions are presented in Table 1. Gene ontology analysis of differentially expressed proteins revealed their involvement in the biosynthesis of flagella, amino acids, prodigiosin, outer membrane, carbohydrates and lipids of S. marcescens (Fig. 6).
Vanillic acid treatment down regulated the expression of surface layer protein (slaA) in S. marcescens by 4.68 fold. Protein surface layer (S-layers) present in Gram-positive and Gram-negative bacteria has been known to involve in cell stabilization (mechanical, thermal and osmotic) 33 , compartmentalisation, protection from host immunological defences (phagocytosis) 34 , trapping of ions and immobilization of proteins 35 . SlaA of S. marcescens is partially similar to the Caulobacter crescentus paracrystalline S-layer protein and is recognized and transported into extracellular medium by the Lip system 36 . Role of slaA in biofilm formation of S. marcescens has not been identified yet. However, S-layer has been shown to have a role in adherence and maintaining cell surface hydrophobicity (CSH) of certain Gram-positive and Gram-negative bacteria 37 and is one of the determinants required for the biofilm formation. Hence, down-regulation of slaA by vanillic acid could modulate the CSH in S. marcescens which is expected to affect the biofilm formation. Furthermore, vanillic acid also down-regulated (−2.11 fold) the expression of flagellin and upregulated (+2.57 fold) the expression of the long polar fimbrial chaperone (LpfB). In S. marcescens, flagellin is synthesized by the flhDC operon and polymerized into flagella, whereas LpfB is involved in the organization of cell wall and pili. Disruption of flhDC operon in S. marcescens CH−1 has been shown to be defective in swarming motility 38 . Flagella and swarming motility required for the initial attachment of biofilm formation 39 . Hence, down-regulation of flagellin by vanillic acid could be responsible for the reduction in biofilm formation. In addition, flagellar M-ring protein was also down-regulated upon vanillic acid treatment. Aminohydrolases are involved in nucleotide and amino acid metabolism 40 . Proteomic analysis revealed a 2 fold down-regulation of the expression of amidohydrolase of S. marcescens upon vanillic acid treatment.
Scientific REPORts | 7: 16328 | DOI:10.1038/s41598-017-16507-x Expression of probable phospholipid-binding protein (MlaC) was found to be upregulated upon vanillic acid treatment. MlaC is involved in bacterial intermembrane phospholipid trafficking 41 and mlaC mutant of Escherichia coli has been shown to have increased phospholipids content in the outer membrane 42 . Hence upregulation of MlaC by vanillic acid could alter the phospholipid content of S. marcescens outer membrane. Urocanate hydratase (HutU) is involved in histidine metabolism found to down-regulated upon vanillic acid treatment by 2.70 fold when compared to control. Biosynthesis of purines and pyrimidines depends on the availability of histidine and DNA synthesis is linked to e-DNA release, which is also required for the biofilm formation 43 . In addition, HutU has been reported to have biofilm formation of Acinetobacter oleivorans 44 and Acinetobacter baumannii 43 . It is likely that down-regulation of HutU by vanillic acid could also be responsible for the reduction of biofilm formation in S. marcescens.
UniProtKB shows that uncharacterized protein A0A0U6CCF3 is 90% identical in terms of sequence similarity to the biofilm development protein (bsmB [A0A0N1UW53]) of S. marcescens subsp. marcescens Db11. In S. marcescens, bsmB has already been shown to regulate by AHL mediated QS and plays a crucial role in biofilm development 22 . Recently phytol has been reported to inhibit the biofilm formation in S. marcescens by targeting the expression of bsmB 45 . Hence, we hypothesized that down-regulation of uncharacterized protein A0A0U6CCF3 in S. marcescens CI by vanillic acid could be responsible for the biofilm inhibition.
Enzymes such as beta-ketoacyl synthase (PigJ) and 4-hydroxy-2, 2′-bipyrrole-5-methanol synthase (PigH) involved in prodigiosin biosynthesis of S. marcescens 15 were found to be down-regulated upon vanillic acid treatment. These results further validate the prodigiosin inhibitory potential of vanillic acid. Furthermore, 3-deoxymannooctulosonate-8-phosphatase (KdsC) which is known to involve in the biosynthesis of lipopolysaccharide and bacterial outer membrane synthesis 46 was found to be down-regulated upon vanillic acid treatment. Since components of outer membrane and LPS are immunogenic and required for biofilm formation in Gram-negative bacterial pathogens 46,47 , down regulation of KdsC by vanillic acid could also be responsible for the reduction in the biofilm of S. marcescens.
Proteomic analysis of vanillic acid treated S. marcescens revealed the upregulation of multiple stress resistance protein (bhsA), which is predicted to be involved in stress response mechanism. BhsA of S. marcescens is 52.1% similar to putative outer membrane protein ycfR (bhsA) of E. coli in terms of the sequence. In E. coli, deletion ycfR (bhsA) has been shown to increase the biofilm, aggregation and CSH 48 . Upregulation of bhsA upon vanillic acid treatment could affect the biofilm formation in S. marcescens. In addition, vanillic acid treatment was found to upregulate the expression of 60 kDa chaperonin (groL) protein. GroL prevents the misfolding and helps in refolding/proper folding of proteins synthesized under stress condition 49 . Beta-ketoacyl-acyl carrier protein synthases are involved in fatty acid biosynthesis in most of the bacteria. Hence they have been identified as one of the promising targets for broad-spectrum antibacterial agents 50 . Down-regulation of beta-ketoacyl-acyl carrier protein synthase I, acetyl-coenzyme A carboxylase carboxyl transferase subunit alpha and beta-ketoacyl-acyl-carrier-protein synthase I, 6-deoxyerythronolide B synthase suggested the ability of vanillic acid to interfere the fatty acid biosynthesis of S. marcescens. Furthermore, the expression of 3-oxoacyl-[ acyl-carrier-protein] reductase (fabG) of S. marcescens was down regulated by vanillic acid. FabG has already been reported to regulate the virulence of Pseudomonas syringae pv. tabaci through AHL and fatty acid biosynthesis 51 . Whereas, in P. aeruginosa fabG determines the acyl chain length of 3-oxo-C 12 -HSL 52 . Role of fabG in QS and biofilm formation of S. marcescens has not been established yet. To get a preliminary information global sequence alignment analysis was carried out using EMBOSS Needle and the results revealed that the fabG of S. marcescens shares 64.8 and 78.1% identity and similarity, respectively to fabG of P. syringae pv. tabaci (A0A0W0PM71) and 63.2 and 75.3% identity and similarity, respectively to fabG (O54438) of P. aeruginosa. Hence it is hypothesized that down-regulation of fabG by vanillic acid could have a negative effect on QS-regulated virulence factors production in S. marcescens. In addition to inhibition of fatty acid biosynthesis pathway, amino acid (arginine, proline, cysteine and methionine) and butanoate metabolism have been predicted to inhibit the biofilm formation 53 . In the present study also down-regulation (1.5 fold) of cysteine synthase was noted upon vanillic acid treatment. Furthermore, down-regulation in the expression of aspartate ammonia lyase (aspA) was observed upon vanillic acid treatment. AspA has been reported to involve in the conversion of L-aspartate into fumarate and ammonia 53 . AspA mediated ammonia generation has already been reported to increase the intracellular polyamines synthesis, which in turn alters the membrane permeability, increases resistance to antibiotics and oxidative List of differentially expressed proteins of S. marcescens upon vanillic acid treatment identified using nano-LC-MS/MS analysis  54 . In contrast, the expression of butyryl-CoA dehydrogenase of S. marcescens was upregulated upon vanillic acid treatment. These results revealed the ability of vanillic acid to modulate the expression of proteins involved in the fatty acid and amino acid biosynthesis in S. marcescens CI. Mutants of Serratia plymuthica RVH1 lacking splI (AHL synthase), has been shown to modulate the mixed-acid and butanediol fermentation. Briefly, mutation in splI gene resulted in enhanced and decreased acid and butanediol production respectively, in S. plymuthica 55 . Interestingly in the present study vanillic acid treatment resulted in the upregulation of 2, 3-butanediol dehydrogenase (budC) of S. marcescens, which is under the control of QS. For further confirmation, Methyl Red Voges Proskauer (MR-VP) test was carried out in the present investigation and the results revealed that, vanillic acid treatment enhances the acid production and decreases the butanediol production. Though, 2, 3-butanediol dehydrogenase (budC) is upregulated upon vanillic acid treatment, the production of butanediol is reduced when compared to control ( Fig. 7a and b). Butanediol production in S. plymuthica RVH1 and S. marcescens MG1 is regulated by AHL mediated QS and the growth of S. marcescens was not altered in the presence of vanillic acid. Hence reduction of butanediol production in S. marcescens upon vanillic acid treatment suggests its ability to modulate the AHL synthesis.
In addition, vanillic acid treatment down regulated the expression of signal recognition particle receptor (FtsY) protein by 1.5 fold. FtsY is actively involved in the translocation/insertion of nascent membrane proteins into the cytoplasmic membrane 56 . Furthermore, molecular chaperone OsmY was found to be down regulated upon vanillic acid treatment in S. marcescens. EMBOSS Needle global sequence alignment analysis revealed that OsmY is 91.7% identical and 95.6% similar in terms of sequence with transport-associated protein (A8G9G9) of Serratia proteamaculans and which is shown to interact with the lipid A 1-diphosphate synthase (lpxT) and outer membrane protein assembly factor (BamA) involved in LPS biosynthesis and assembly of outer membrane proteins, respectively [interaction can be found in STRING database] 57 . Down-regulation of FtsY and OsmY suggests the possible involvement of vanillic acid to modulate the transport of membrane proteins. Vanillic acid treatment was found to upregulate the expression of aldehyde dehydrogenase B (+4.74 fold) in S. marcescens. Recently, aromatic aldehyde dehydrogenases from Sphingobium sp. strain SYK-6 have been reported to catalyze the conversion of vanillin to syringaldehyde 58 . Hence, upregulation of aldehyde dehydrogenase B could be the possible mechanism responsible for the survival of S. marcescens in the presence of vanillic acid. Oxidoreductase has been reported to regulate the biofilm formation and virulence in Salmonella enterica serovar Typhimurium 59 and Burkholderia cenocepacia 60 . In the present study, vanillic acid treatment was found to down-regulate the expression of putative oxidoreductase by 2.7 fold. In addition, vanillic acid treatment was found to modulate the expression of ferrous iron transporter B (+5.77 fold), delta-aminolevulinic acid dehydratase (+2.40 fold) and several uncharacterized proteins ( Table 1).
Role of slaA in biofilm formation and osmotic stress resistance in S. marcescens has not been studied yet. However, surface layer proteins of Lactobacillus acidophilus ATCC 4356 have been shown to protect the cells form NaCl induced osmotic stress 61 and are responsible for adherence and cell surface hydrophobicity (CSH) 37 . To this end, CSH assay was carried out and the results revealed the concentration dependent reduction in CSH of S. marcescens upon vanillic acid treatment (Fig. 7c). Reduction in the CSH hinted the possible role of slaA in the CSH of S. marcescens. Furthermore, in the present study, we have also evaluated the effect of vanillic acid treatment on the osmotic stress resistance of S. marcescens and the results revealed that vanillic acid treated cells are unable to grow normally in the presence of 0.5 and 1 M NaCl when compared to control. These results suggest the ability of vanillic acid to modulate the osmotic stress resistance in S. marcescens by down regulating the expression of slaA protein.
Further analyses are required to unveil the role of slaA in biofilm formation and virulence of S. marcescens.

Summary
In conclusion, vanillic acid has been identified as an active principle responsible for QSI and antibiofilm potential of kiwifruit (A. deliciosa) against S. marcescens. Virulence assays revealed the concentration dependent inhibitory activity over QS regulated biofilm, protease, prodigiosin and lipase production and swarming motility of S. marcescens CI and S. marcescens ATCC. In addition, vanillic acid also inhibited the protease and lipase production and biofilm formation in non-pigmented S. marcescens MG1. In addition, light and CLSM microscopic analyses further confirmed the antibiofilm activity of vanillic acid against S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 used in this study. Furthermore, FTIR analysis of EPS extracted from control and vanillic acid treated S. marcescens CI showed the reduction in the nucleic acid, amide I and II proteins and fatty acid content. In vivo assays confirms the ability of vanillic acid to rescue C. elegans form the S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 infection by inhibiting the QS regulated virulence and biofilm formation. Mass spectrometric identification of differentially expressed proteins and gene ontology analyses revealed the ability of vanillic acid modulates to the proteins involved in S-layers, flagellin, amino acid and fatty acid production in S. marcescens CI. The present study suggests that vanillic acid with its non-toxic nature and QSI potential can serve as a choice to potentiate the treatment strategy to overcome the QS-and biofilm-mediated infections caused by S. marcescens.

Materials and Methods
Ethical statement. Sheep blood was collected from the Karaikudi municipality modern slaughter house, Karaikudi and used to evaluate the effect of vanillic acid treatment on the hemolysin production in S. marcescens. Normally sheep blood is discarded in the butchery and hence no specific ethical permission was needed.  material and preparation of A. deliciosa pulp extract (ADPE). A. deliciosa (Kiwifruit Zespri International Limited, New Zealand) was collected from the local market and the pulp was ground and centrifuged to collect the A. deliciosa pulp extract (ADPE). The ADPE was filter sterilized using 0.2 µm on the membrane filter (Millipore Corp., USA).

Bacterial strain and growth conditions. S. marcescens
Clinical isolate used in this study is a clinical strain isolated from a urine sample, identified by 16S rRNA gene sequencing (GenBank Accession Number: FJ584421). In addition, S. marcescens ATCC 14756 and S. marcescens MG1 was also used to evaluate the anti-QS and antibiofilm activity of vanillic acid.
To determine the effect of ADPE on QS regulated extracellular virulence factors, LB medium (1% tryptone, 1% sodium chloride and 0.5% yeast extract) supplemented with and without (10-20% v/v, in increments of 5%) filter sterilized ADPE was inoculated with 1% (v/v) of the overnight culture of S. marcescens (OD 600nm = 0.5) and incubated at 30 °C for 18 h at 120 rpm. After incubation, the cell density of control and treated S. marcescens CI was measured at 600 nm followed by centrifugation at 12,000 rpm for 15 min at 4 °C to collect the cell free culture supernatant (CFCS). CFCSs were stored at −20 °C for protease, lipase and hemolysin assay. The cell pellet was used for prodigiosin assay. Experiments are performed in biological triplicates in experimental duplicates.

Measurement of prodigiosin.
For extraction of prodigiosin, cell pellets were resuspended in 1 ml acidified ethanol (4% (v/v) of 1 M Hydrochloric acid (HCl) in ethanol) and vortexed vigorously. Acidified ethanol containing solubilized prodigiosin and cell debris was subjected to centrifugation at 10,000 rpm for 10 min and the optical density of the supernatant was measured at 534 nm. The percentage of prodigiosin inhibition was calculated using the formula 17 : Total protease assay. The casein-degrading proteolytic activity was assessed by incubating each 100 µL reaction mix containing 2 mg/mL of azocaesin in 0.05 M Tris-hydrochloride, 0.5 mM CaCl 2 (pH 7.5) with 100 µl of control and ADPE treated S. marcescens CFCS at 37 °C for 15 min. A 500 µl of 10% trichloroacetic acid was added to each reaction tube to terminate the reaction, followed by incubation at −20 °C for 10 min and the tubes were centrifuged at 12000 rpm for 20 min. Finally, the absorbance of the supernatant was measured at 400 nm using Multi-Mode Microplate Reader (SpectraMax M3, US) 12 .
Lipase assay. A 100 µl of S. marcescens culture supernatant (control and ADPE treated) was incubated with 900 µl of reaction mixture containing 1 volume 0.3% (w/v) p-nitro phenyl palmitate in isopropanol and 9 volumes of 50 mM Na 2 PO 4 buffer (pH 8.0) containing 0.2% (w/v) sodium deoxycholate and 0.1% (w/v) gummi arabicum for 1 h at room temperature in dark. An equal volume of 1 M Na 2 CO 3 was added to each tube to terminate the lipolytic activity and then centrifuged at 12,000 rpm for 10 min at room temperature. The absorbance of the supernatant was measured at 410 nm 62 .
Hemolysin assay. Hemolysin production was measured in accordance with the method described earlier 18 .
Equal volume of 2% sheep red blood cells (RBCs) in phosphate buffered saline (pH 7.4) was mixed with control and ADPE treated S. marcescens CFCS at 37 °C for 2 h. After incubation, the reaction mixture was centrifuged at 8,000 rpm for 5 min and the absorbance of the supernatant was measured at 530 nm 63 .
Biofilm inhibition assay. Anti-biofilm potential of ADPE extract was determined using the method adopted by our group previously 64 . The biofilm of S. marcescens was allowed to grow in LB medium in the presence (10-20% v/v, in increments of 5%) and absence of ADPE in 24-well polystyrene plate at 30 °C for 24 h. After incubation, the wells were washed with distilled water to remove the planktonic cells. The sessile cells were stained with 0.4% crystal violet stain (CV) (w/v) for 5 minutes followed by washing off the unstained dye using distilled water. Finally, 1 mL of absolute ethanol was added to each well to dissolve the CV present in the biofilm cells. The absorbance (OD 570nm ) was determined and percentage of biofilm inhibition was calculated using the formula: Swarming motility assay. The effect of vanillic acid on the swarming motility was determined by placing 2 µL of overnight cultures of S. marcescens in the centre of swarm agar plates (peptone (5 g/L), glycerol (1% (v/v)) and agar (0.75%)). The plates were incubated at 25 °C for 20 h. The activity was determined by the decrease in the swarmed area radius 5 .
Extraction of EPS and FTIR analysis. EPS  Microscopic observation. For microscopic observation, C. elegans were exposed to S. marcescens CI, S. marcescens ATCC and S. marcescens MG1, with and without vanillic acid for 24hrs. Then control and treated C. elegans were washed with M9 buffer thoroughly and placed on a microscopic slide with 1 mM Sodium azide and visualized under bright field microscope 68 .

Colony forming unit assay (CFU).
To access the intestinal colonization of S. marcescens CI, S. marcescens ATCC and S. marcescens MG1 in C. elegans CFU was performed. Briefly synchronised L4 stage C. elegans were exposed to S. marcescens CI, S. marcescens ATCC and S. marcescens MG1, with and without vanillin 250 µg/mL for 24hrs. After 24hrs, 20 numbers of exposed worms were washed thoroughly with M9 buffer containing 1 mM sodium azide to reduce the discharge of bacteria from the worm intestine and finally crushed/ground with 400 mg of 1.0 mm of silicon carbide and vortexed for 20 minutes. The final suspension was serially diluted and plated onto specific medium 68 .
Extraction of cellular proteins. For extraction of intracellular proteins, S. marcescens was grown in the absence and presence of 250 µg/mL of vanillic acid for 18 h at 30 °C with 120 rpm in conical flasks in biological triplicates and cells were harvested by centrifugation at 8000 rpm for 12 min at 4 °C and the resulting pellets were washed twice with 50 mM tris-HCl (pH 8.0). Cell pellets were resuspended in 50 mM tris-HCl (pH 8.0) supplemented with 1% of protease inhibitor cocktail and sonicated using Ultra-sonicator (Sonics VCX 750, USA) with the parameters of 35 KHz, 750 W and 35% amplitude for 5 min with pulse time of 10 sec (for both on and off cycles) followed by centrifugation at 13,000 rpm for 30 min at 4 °C to remove the cell debris. Supernatants containing cellular proteins were purified using phenol extraction and precipitated by acetone. Precipitated proteins were washed thrice with ice-cold acetone to wash off the remaining phenol residues and air dried. Finally, the air-dried protein pellets were dissolved in urea thiourea sample buffer ( [3][4][5][6][7][8][9][10]) and centrifuged at 13,000 rpm for 20 min at 4 °C and the concentration of proteins present in the samples were quantified by Bradford assay kit (BioRad). Each 450 µg of cellular proteins form control and vanillic acid treated S. marcescens were used for isoelectric focusing (IEF) in biological triplicates 69 .

Two-dimensional gel electrophoresis (2DGE) and image analysis. Protein samples (each 400 µg)
from control and vanillic acid treated S. marcescens were diluted with urea thiourea rehydration buffer (7 M urea, 2 M thiourea and 2% [CHAPS]) supplemented with 12.5 mg/mL destreak reagent and 0.5% IPG buffer [pH 3-10] to a final volume of 350 µl. Immobiline DryStrip gel strips (18 cm, non-linear, pH 3-10) were rehydrated with thiourea rehydration buffer containing cellular proteins for 12 h at 20 °C and the rehydrated strips were iso-electrically focused in IPGphor 3 system using standard parameters. After IEF, IPG strips were subjected to reduction and alkylation with DTT and iodoacetamide (IAA), respectively. Then the strips were placed on 25 cm × 22 cm × 1 mm 10-15% gradient sodium dodecyl sulphate-polyacrylamide gels (SDS-PAGE), sealed with 0.3% agarose solution containing a trace amount of tracking dye Bromophenol blue and electrophoresis was carried out at 100 V for 1 h and 150 V until the dye front had reached the bottom of the SDS-PAGE gel in Ettan DALT six apparatus (GE Healthcare). After electrophoresis, the gels incubated in fixative solution (40% methanol, 10% glacial acetic acid (GAA) and 50% MilliQ H 2 O) for 3 h. Protein spots were visualized using mass spectrometry compatible colloidal coomassie brilliant blue (CBB) G-250 staining for 12 h and destained with MilliQ H 2 O until the background becomes clear. Then, image acquisition was done with Image Scanner III (GE Healthcare) and analysed by ImageMaster 2D Platinum 7.0 software (GE Healthcare) 69 .
Protein spots excision and in-gel trypsin digestion. The protein spots with 1.5 fold differential regulation (up/down) were excised from the gels and the excised gel plugs were destained with 50% acetonitrile (ACN) containing 25 mM ammonium bicarbonate (NH 4

Identification of differentially expressed proteins by nano LC-MS/MS and MALDI-TOF/TOF analyses.
Protein spots with ≥2 fold differential regulation were analysed using nano reverse phase-LC (Thermo Scientific, USA) coupled with an Orbitrap Elite Mass spectrometer (Thermo Scientific, USA) using standard parameters 71 . The MS data acquisition was done in positive ion mode (m/z 350-4000 Da) using Xcalibur software (Version 2.2.SP1.48, Thermo Scientific, USA). Protein identification was done with Proteome Discoverer software v.1.4 (Thermo Scientific) using the following parameters: 2 missed cleavages, 10 ppm and 0.5 Da as precursor mass tolerance and fragment mass tolerance, respectively. In addition, cystine carbamidomethylation as fixed modification, methionine oxidation, N-terminal acetylation and phosphorylation (S, T, Y) as variable modification. Protein spots with 1.5 to 1.9 fold differential regulation were analysed using MALDI-TOF/TOF analysis. Prior to analysis, internal calibration was carried out using TOF-Mix ™ (LaserBio Labs, France). For MALDI-TOF/ TOF analysis, 1 µL of purified peptides were mixed with 1 µL of the alpha-cyano-4-hydroxy cinnamic acid matrix (10 mg/mL) on MALDI target plate. Mass spectra were acquired in positive reflectron mode and the monoisotopic peak list (m/z range of 700-4000 Da) was used for protein identification using MS-Fit (http://prospector. ucsf.edu) online software using standard parameters 69 .
MR-VP test. S. marcescens was allowed to grow in MR-VP broth supplemented with and without 125 and 250 µg/mL of vanillic acid for 24 h. After incubation, to detect the acid production, methyl red indicator dye was added and the changes in the colour was visually observed and photographed. To detect the butanediol formation Barritt's reagent A and B were added to the culture, mixed well and formation of red coloration was visually observed and photographed.
Osmotic sensitivity assay. To analyse the effect of vanillic acid treatment on the osmotic sensitivity, S. marcescens was allowed to grow in LB medium supplemented with 0.5 and 1 M NaCl in the presence and absence of 125 and 250 µg/mL of vanillic acid at 30 °C and cell density was measured periodically for 24 h with 2 h interval 61 .

Statistical analysis.
All the experiments were performed independently in triplicate to confirm reproducibility. Data were analyzed by one-way analysis of variance (ANOVA) followed by Duncan's post-hoc test with a significant P value of <0.05 using the SPSS (Chicago, IL, USA) statistical software package.