Letters in Applied Microbiology ISSN 0266-8254

ORIGINAL ARTICLE

Quorum sensing inhibition activity of garlic extract and p-coumaric acid

S.F. Bodini, S. Manfredini, M. Epp, S. Valentini and F. Santori
Department of Environmental Biotechnologies, ISRIM Scarl, Strada di Pentima, Terni, Italy

Keywords Agrobacterium tumefaciens, Chromobacterium violaceum, DMSO, garlic, LuxR-type receptors, p-coumaric acid, Pseudomonas chlororaphis, quorum sensing. Correspondence Sergio F. Bodini, ISRIM Scarl, Strada di Pentima, 4 05100 Terni, Italy. E-mail: sofobi@hotmail.com (or) s.bodini@isrim.it

Abstract Aims: The goal of this work was to investigate the inuence of DMSO, garlic extract and p-coumaric acid on bacterial quorum sensing (QS). Methods and Results: The decreases in the QS responses of QS reporter strains Escherichia coli pSB401 and pSB536, Agrobacterium tumefaciens NTL4, Chromobacterium violaceum 5999 and wt 494, Pseudomonas putida IsoF gfp and environmental Pseudomonas chlororaphis were quantied in relation to growth inhibitory effects. DMSO showed no signicant QS-specic effects on the strains tested even at close-to-lethal concentrations. Garlic extracts antagonized the activity of QS receptors LuxR, AhyR and TraR, but were toxic at higher concentrations. P-coumaric acid fully inhibited QS responses of 5999, NTL4 and P. chlororaphis, with no inuence on cell viability. Conclusions: The quorum sensing inhibition activity of garlic was extended to novel receptors, and p-coumaric acid was found to possess previously undescribed QS antagonist properties. Signicance and Impact of the Study: The results suggest that p-coumaric acid might act as QS inhibitor. Further studies are required to understand its role in the regulation of QS and investigate structurally related compounds.

2001 0169: received 27 January 2009, revised 17 July 2009 and accepted 17 July 2009
doi:10.1111/j.1472-765X.2009.02704.x

Introduction The quorum sensing is a cell densitydependent circuit common among bacteria (for review see Gonzalez and Keshavan 2006) mediated by diffusible molecules termed autoinducers (AIs). In most Gram-negative species, the AIs consist of constitutively synthesized N-acyl homoserine lactones (AHLs). The AHL-dependent QS system is regulated by two protein homologues of LuxI-AHL synthase and LuxR-AHL-response regulator, known for controlling the expression of bioluminescence in Vibrio scheri. When AHLs reach a threshold concentration, the transcriptional complex formed by the interaction of AHLs with their cognate regulator binds to DNA sequences termed lux boxes, upstream of the promoters of genes regulated by QS. This provides the population with distinctive density-dependent behaviours like antibiotic production, virulence, bioluminescence, motility, symbiosis and biolm formation. The strategy to treat bacterial diseases with antibiotics is nowadays seriously
ISRIM SCARL 49 (2009) 551555

threatened by the spread of drug-tolerant strains, which cause persistent infections, often involving the formation of highly resistant biolms. Understandably, the observations that biolm production and the activation of other pathologically signicant virulence factors were strictly interrelated with the expression of the QS alarmone raised new expectations for the discovery of anti-pathogenic drugs that are capable of interfering with the bacterial communication system, without inducing lethal effects (Rasmussen and Givskov 2006). A number of qualitative quorum sensing inhibition (QSI) screening protocols were thus based on biomonitor organisms whose QS responses were easily detectable (Steindler and Venturi 2007). In this work, DMSO, garlic extracts and p-coumaric acid were tested for their QS antagonist activity on bacterial strains possessing a LuxRI-type communication system. DMSO is a highly polar solvent, widely used in QSI assays as an aprotic alternative for water (Ni et al. 2008). Garlic is an extensively consumed vegetable, known for its antioxidant, anti-inammatory and antimicrobial
551

QSI activity of garlic and p-coumaric acid

S.F. Bodini et al.

properties, whose QSI activities on LuxR- and LasR-based biosensors were demonstrated by Rasmussen et al. (2005). P-coumaric acid is a natural compound primarily produced by plants both as a phenylpropanoid intermediate of the lignin pathway and in response to wounds and nutritional stresses, which was recently recognized as a precursor of the cumaroylAHL signal molecule (Schaefer et al. 2008). Our study aimed at (i) providing quantitative data on the impact of DMSO on QS assays, (ii) investigating the effects of garlic on QS receptors other than LuxR and LasR and (iii) looking for putative QSI activity of p-coumarate. The experimental data were processed to quantify the QSI responses in relation to the potential interference of antibiotic effects. Our experiments revealed that (i) DMSO affected QS signals but compromised cell growth at concentrations higher than 310%, (ii) garlic extracts induced signicant concentration-dependent QS antagonist effects against E. coli mutants and Agrobacterium and (iii) millimolar amounts of p-coumarate inhibited QS of both environmental P. chlororaphis and CviR- and TraR-based QS reporter strains, without affecting growth rate. Materials and methods Bacterial strains, growth conditions and reagents Escherichia coli pSB401 and pSB536, Agrobacterium tumefaciens NTL4 and Pseudomonas putida IsoF gfp were generously provided by the International Center of Genetic Engineering and Biotechnology, Trieste, Italy. Chromobacterium violaceum 5999 and wt 494 were pur` chased from the Coleccion Espanola de Cultivos Tipo, Valencia, Spain. Pseudomonas chlororaphis (PSECL) was isolated from a diesel-contaminated soil and characterized by BIOLOG metabolic ngerprinting (Biolog, Hayward, CA, USA). All strains were cultured aerobically with shaking at 120 rev min)1: pSB401, pSB536 and 5999 at 30C in LB broth with tetracycline, ampicillin and kanamycin, 5, 100 and 25 lg ml)1, respectively, NTL4 at 28C in AB minimal broth with 30 lg ml)1 of gentamicin, IsoF gfp at 30C in LB broth Gen40 and PSECL at 30C in AB minimal broth with 05% glucose. Garlic extracts The extraction of garlic was performed with toluene as described by Rasmussen et al. (2005). The toluene homogenate was ltered and extracted with an equivalent volume of water. The aqueous phase was concentrated in a Speedvac (Savant Instruments Inc., Holbrook, NY, USA). The concentration of garlic extracts intended for QSI assays was expressed in garlic equivalents that dened
552

the ratio of the fresh weight of extracted crude garlic to the nal assay volume, so as to associate the detected activities to the original amount of garlic cloves. QSI assays All tests were performed on freshly QS-induced cultures in the absence and presence of the putative QS mimic. pSB401 and pSB536 After 4-h incubation, bioluminescence of 50 ll of cell culture was measured with a Microtox Model 500 Analyzer (Strategic Diagnostic Inc., Newark, NJ, USA). NTL4 After 3-h incubation, two ml of cell culture was added to 100 ll of methylumbelliferyl-b-d-galactoside in DMSO (5 mg ml)1) and, after 60 min at 37C, uorescence at 445 nm following excitation at 362 nm was measured with a PerkinElmer LS50B (PerkinElmer Inc., Wellesley, MA, USA). 5999 and wt 494 After 24-h incubation, 100 ll of cell culture was centrifuged (3600 g, 10 min) to precipitate insoluble violacein. One millilitre of DMSO was added to the pellet. The solution was centrifuged (3600 g, 10 min), and the absorbance at 585 nm of the supernatant was measured with a spectrophotometer Uvikon 992 (Kontron Instruments, Watford, UK). IsoF gfp After 24-h incubation, green uorescence at 515 nm following excitation at 475 nm was measured with a PerkinElmer LS50B. PSECL After overnight incubation, 3 ml of cell culture was centrifuged (3600 g, 10 min). The supernatant was concentrated in a Speedvac, and the residue was derivatized for 30 min at 37C with 20 ll of N-methyl-N-trimethylsilyltriuoroacetamide in 20 ll of pyridine (Roessner et al. 2000). One microlitre of sample was injected in a gas chromatographymass spectrometry (GC-MS) system consisting of a 1077 Universal Capillary Injector operated in splitless mode, a Star 3400 and a Saturn II ion-trap (Varian Inc., Walnut Creek, CA, USA). The carrier gas was helium set at 2 ml min)1. GC was performed on a 30-m long, 025-mm internal diameter CP-Sil 8 CB Low Bleed MS column with 02-lm lm thickness. Injection temperature was 250C. The oven temperature was held constant at 100C for 2 min, ramped at 4C min to 280C and held constant for 13 min. Mass spectra were
ISRIM SCARL 49 (2009) 551555

S.F. Bodini et al.

QSI activity of garlic and p-coumaric acid

acquired from m z 50-650 with scan time of 1 s and evaluated using the Varian WS program. Retention time of phenazine 1-trimethylsilyl carboxylic acid (TMS-PCA) was 36125 min. Peak integration of the abundance of the distinctive ion at m z 281 allowed to calculate PCA concentration. Cell growth Cell growth was monitored by plating E. coli strains, Iso gfp and PSECL on LB agar, NTL4 on Czapek yeast agar and C. violaceum strains on LB soft agar. Results Quantitative QSI bioassays in liquid media containing increasing concentrations of DMSO, garlic and p-coumaric acid were performed on ve different QS bioreporters characterized by inability to produce their own AI signal, but still able to respond to exogenous active signal molecules (McClean et al. 1997; Winson et al. 1998; Luo et al. 2001; Venturi 2006) and a reference wild-type strain (Table 1). The effects of p-coumarate on QS were also tested on an environmental P. chlororaphis isolate by monitoring, via GC-MS, the QS-regulated synthesis of the yellow antibiotic PCA. The bacterial survival was monitored by cell culture, because OD measurements were found to be strongly affected by the presence of dead cells. At high concentrations, DMSO was toxic against all strains tested with no signicant QSI activity (Table 2), whereas garlic showed strain- and concentration-dependent QSI activity (Fig. 1). When tested on Chromobacterium strains and P. putida, garlic induced remarkable decreases in both QS response and growth within narrow concentration ranges that prevented the unequivocal recognition of anti-QS properties. In contrast, a signicant QSI activity of garlic on LuxR-, AhyR- and TraR-based reporter strains could be identied in the presence of 225 g ml)1 of garlic equivalents. However, when the concentration was doubled, the antimicrobial effect prevailed, and the extracts seriously affected bacterial survival. QSI assays performed in the presence of millimolar amounts of p-coumaric acid showed that E. coli
Table 1 Strains used in this study
Strain pSB401 pSB536 NTL4 5999 wt 494 IsoF gfp PSECL Host

Table 2 Minimum DMSO concentration affecting quorum sensing assays

Strain pSB401 pSB536 NTL4 5999 494 IsoF gfp DMSO (%) 5 10 10 3 15 10

strains were left unaffected and P. putida suffered toxic effects (Fig. 2). Instead, intriguingly, p-coumarate antagonized the QS activity of CviR-, PhzR- and TraR-based reporter strains in a concentration-dependent manner, as evidenced by the fact that QS-dependent productions of violacein by C. violaceum, PCA by P. chlororaphis and b-galactosidase by A. tumefaciens were strongly inhibited, without comparable growth decreases. This was also visually corroborated by the absence, respectively, of purple pigmentation, yellow colour and uorescence. Discussion Aiming at tackling the problem of antibiotic resistance, the investigation into substances characterized by antipathogenic rather than antibiotic properties motivated, in recent years, widespread research efforts that allowed to recognize the presence of QS mimics in a large variety of plant parts, extracts and exudates (for review see Janssens et al. 2008). Three main ndings emerged from the present study: (i) the use of DMSO as a solvent in QS-related studies shall be limited to concentrations lower than 310%, presumably as a result of the occurrence of prelethal alterations; (ii) the QSI activity of garlic extracts was extended from LuxR to different LuxR-type receptors like AhyR and TraR and (iii) p-coumaric acid showed putative antagonist activity against Chromobacterium, Agrobacterium and Pseudomonas QS. The latter result was fully unexpected because our original aim was to test the ability of p-coumarate in accelerating the QS response. Our

Based on quorum sensing system LuxI R (V. scheri) AhyI R (A. hydrophyla) TraI R (A. tumefaciens) CviI R (C. violaceum) CviI R (C. violaceum) PpuI R (P. putida) PhzI R (P. chlororaphis)

Autoinducer added (lmol l)1) C6-HSL (1) C4-HSL (10) 3-oxo-C6-HSL (5) C6-HSL (05) Self-generated 3-oxo-C12-HSL (1) Self-generated

Reporter system Bioluminescence Bioluminescence b-galactosidase Violacein Violacein Fluorescence PCA

E. coli E. coli A. tumefaciens C. violaceum C. violaceum P. putida P. chlororaphis

ISRIM SCARL 49 (2009) 551555

553

QSI activity of garlic and p-coumaric acid

S.F. Bodini et al.

100 50 0 401 % of control 100 50 0 5999 100 50 0 494

a A b 0 a A b B

A c C 5

100 50 0 536 100 B c c C 5 25 50 0 Ntl4 100 bB bB 50

a A c 0 a A c 0 a A 2

B d 2 A c 5 C 5 A B c 7 d C 7
Figure 1 Effect of garlic extracts on quorum sensing (QS) expression (open bars) and growth (closed bars) of tested strains. Data were reported as mean percentages (SE, 25 tests) of the corresponding controls. QS values were measured as relative luminescence units (RLU) ml)1 for E. coli pSB401 and pSB536, absorbance for C. violaceum 5999 and wt 494 and relative uorescence units (RFU) ml)1 for A. tumefaciens NTL4 and P. putida IsoF gfp. Growth was measured as colony-forming units (CFU) ml)1. Bars marked by different letters are signicantly different based on Tukeys test (P < 005).

25

0 aA

125 aA b A

a A b C 2 5 b C 7

15

0 25 5 0 IsoF 1) Garlic equivalents (g ml

100 50

aA

aA

a A bB bB bB bB 100 50 0 PSECL 00 a A 100 50 0 Ntl4 aA A a b 00

a A B c 06 A b 30 100 50 b 61 aA aA B 61 A b 91 c 91 A
Figure 2 Effect of p-coumaric acid on quorum sensing (QS) expression (open bars) and growth (closed bars) of tested strains. Data were reported as mean percentages (SE, 25 tests) of the corresponding controls. QS values were measured as RLU ml)1 for E. coli pSB401 and pSB536, absorbance for C. violaceum 5999 and wt 494, RFU ml)1 for A. tumefaciens NTL4 and P. putida IsoF gfp and ion counts (m z = 281) for P. chlororaphis PSECL. Growth was measured as CFU ml)1. Bars marked by different letters are signicantly different based on Tukeys test (P < 005).

0 401 00 % of control 100 50 0 5999 00 aA a A

06

30 A b 30

61 A b 61 aA

91 122 A b 91 100 50 aA

100 50

C B d d 0 0 0 IsoF 00 61 91 122 494 00 91 122 152 536 00 30 61 91 p-coumaric acid concentration (mmol l1)

expectations were motivated by several publications disclosing a stimulatory role played by p-coumarate in molecular pathways directly and indirectly associated with the QS system, like (i) being a precursor of p-coumaroylAHL, a signal molecule used by the photosynthetic bacterium Rhodopseudomonas palustris (Schaefer et al. 2008) and (ii) stimulating the vir regulon, a two-component machinery carrying the virulence proteins associated with the QS system of A. tumefaciens (Peng et al. 1998). In contrast, our investigations undertaken on a nonvirulent vir-defective A. tumefaciens mutant and a CviI-mutant of C. violaceum lacking the AHL synthase showed that millimolar concentrations of p-coumaric acid caused full inhibition of the QS system. Intriguingly, we also observed that p-coumaric acid repressed the antibiotic production by environmental P. chlororaphis, a competitive bacterium extensively used as a biocontrol agent against fungal plant pathogens (Whistler and Pierson 2003). These results suggest that depending upon strain and concentration
554

p-coumaric acid may exert stimulatory or inhibitory effects in the regulation of QS. Similar strain-dependent behaviours were observed among QS autoinducers. Exemplarily, the natural agonist of the PpuR receptor system, 3-oxo-C12-AHL, also acts as an antagonist of the QS-regulated activity in Chromobacterium (McClean et al. 1997). Questions still remain as to the possible causes of these previously unknown anti-QS properties of p-coumaric acid. An important observation is that earlier recognized QS antagonists were typically found active at much lower concentrations than observed for p-coumaric acid. The AI-2 antagonism showed by pyrogallol and analogues (Ni et al. 2008) was detected in the micromolar range and, accordingly, 50 lmol l)1 l-canavanine from Medicago sativa was sufcient to affect QS expression on Sinorhizobium meliloti (Gonzalez and Keshavan 2006), and 15 lmol l)1 curcumin from Curcuma longa effectively reduced the pathogenicity of P. aeruginosa PAO1 QS (Rudrappa and Bais 2008). There arose the
ISRIM SCARL 49 (2009) 551555

S.F. Bodini et al.

QSI activity of garlic and p-coumaric acid

need to consider if other relevant facts interfered with our experimental study. From a chemical point of view, specic control tests performed in the absence of the bacterial species allowed us to reject the hypothesis that p-coumarate nonspecically affected the absorbance or uorescence. Biologically, because p-coumarate is known for inuencing a wide variety of metabolic activities as a result of strong anti-oxidant and radical-scavenging properties (Choi et al. 2007), the existence of additional actions of p-coumaric acid, such as enzyme inhibitions and interference with regulatory mechanisms, cannot be excluded, and further studies are needed to directly examine whether the decrease in the QS signal could be simply referable to an inhibitory effect of p-coumarate on the secondary metabolism. However, the demonstrated fact that cinnamaldehyde also induces QS modulation (Niu et al. 2006) corroborates the possibility that the class of cinnamate derivatives encompasses one or more QS antagonists, other than p-coumaric acid, and more powerful at lower concentrations. Indeed, the use of natural low weight compounds to attenuate bacterial pathogenicity is a novel and attractive approach particularly if, at the dosages used, these QS inhibitors are nontoxic and produce no adverse consequences on host-indigenous bacterial ora. Acknowledgements This work was supported by the Italian private foundation Fondazione Cassa di Risparmio di Terni e Narni. We thank Mike Givskov, Vittorio Venturi, Stephen Farrand and Anna Rita Cicalini for providing strains and Centro Servizi Bibliotecari, Universita di Perugia, sezione Terni for providing references. References
Choi, S.W., Lee, S.K., Kim, E.O., Oh, J.H., Yoon, K.S., Parris, N., Hicks, K.B. and Moreau, R.A. (2007) Antioxidant and antimelanogenic activities of polyamine conjugates from corn bran and related hydroxycinnamic acids. J Agric Food Chem 55, 39203925. Gonzalez, J.E. and Keshavan, N.D. (2006) Messing with bacterial quorum sensing. Microbiol Mol Biol Rev 70, 859875. Janssens, J.C., de Keersmaecker, S.C., de Vos, D.E. and Vanderleyden, J. (2008) Small molecules for interference with cell-cell-communication systems in Gram-negative bacteria. Curr Med Chem 15, 21442156. Luo, Z.Q., Clemente, T.E. and Farrand, S.K. (2001) Construction of a derivative of Agrobacterium tumefaciens C58 that

does not mutate to tetracycline resistance. Mol Plant Microbe Interact 14, 98103. McClean, K.H., Winson, M.K., Fish, L., Taylor, A., Chhabra, S.R., Camara, M., Daykin, M., Lamb, J.H. et al. (1997) Quorum sensing and Chromobacterium violaceum: exploitation of violacein production and inhibition for the detection of N-acylhomoserine lactones. Microbiology 143, 37033711. Ni, N., Choudhary, G., Li, M. and Wang, B. (2008) Pyrogallol and its analogs can antagonize bacterial quorum sensing in Vibrio harveyi. Bioorg Med Chem Lett 18, 15671572. Niu, C., Afre, S. and Gilbert, E.S. (2006) Subinhibitory concentrations of cinnamaldehyde interfere with quorum sensing. Lett Appl Microbiol 43, 489494. Peng, W.T., Lee, Y.W. and Nester, E.W. (1998) The phenolic recognition proles of the Agrobacterium tumefaciens VirA protein are broadened by a high level of the sugar binding protein ChvE. J Bacteriol 180, 56325638. Rasmussen, T.B. and Givskov, M. (2006) Quorum-sensing inhibitors as anti-pathogenic drugs. Int J Med Microbiol 296, 149161. Rasmussen, T.B., Bjarnsholt, T., Skindersoe, M.E., Hentzer, M., Kristoffersen, P., Kote, M., Nielsen, J., Eberl, L. et al. (2005) Screening for quorum-sensing inhibitors (QSI) by use of a novel genetic system, the QSI selector. J Bacteriol 187, 17991814. Roessner, U., Wagner, C., Kopka, J., Thretewey, R.N. and Willmitzer, L. (2000) Simultaneous analysis of metabolites in potato tuber by gas chromatographymass spectrometry. Plant J 23, 131142. Rudrappa, T. and Bais, H.P. (2008) Curcumin, a known phenolic from Curcuma longa, attenuates the virulence of Pseudomonas aeruginosa PAO1 in whole plant and animal pathogenicity models. J Agric Food Chem 56, 19551962. Schaefer, A.L., Greenberg, E.P., Oliver, C.M., Oda, Y., Huang, J.J., Bittan-Banin, G., Peres, C.M., Schmidt, S. et al. (2008) A new class of homoserine lactone quorum-sensing signals. Nature 454, 595599. Steindler, L. and Venturi, V. (2007) Detection of quorum sensing N-acyl homoserine lactone signal molecules by bacterial biosensors. FEMS Microbiol Lett 266, 19. Venturi, V. (2006) Regulation of quorum sensing in Pseudomonas. FEMS Microbiol Rev 30, 274291. Whistler, C.A. and Pierson, L.S. III (2003) Repression of phenazine antibiotic production in Pseudomonas aureofaciens strain 30-84 by RpeA. J Bacteriol 185, 37183725. Winson, M.K., Swift, S., Fish, L., Throup, J.P., Jrgensen, F., Chhabra, S.R., Bycroft, B.W., Williams, P. et al. (1998) Construction and analysis of luxCDABE-based plasmid sensors for investigating N-acyl homoserine lactone-mediated quorum sensing. FEMS Microbiol Lett 163, 185192.

ISRIM SCARL 49 (2009) 551555

555