Antonie van Leeuwenhoek DOI 10.1007/s10482-015-0503-6
ORIGINAL PAPER
Isoprenyl caffeate, a major compound in manuka propolis, is a quorum-sensing inhibitor in Chromobacterium violaceum Adrian Tandhyka Gemiarto . Nathaniel Nyakaat Ninyio . Siew Wei Lee . Joko Logis . Ayesha Fatima . Eric Wei Chiang Chan . Crystale Siew Ying Lim
Received: 6 January 2015 / Accepted: 4 June 2015 Ó Springer International Publishing Switzerland 2015
Abstract The emergence of antibiotic-resistant bacterial pathogens, especially Gram-negative bacteria, has driven investigations into suppressing bacterial virulence via quorum sensing (QS) inhibition strategies instead of bactericidal and bacteriostatic approaches. Here, we investigated several bee products for potential compound(s) that exhibit significant QS inhibitory (QSI) properties at the phenotypic and molecular levels in Chromobacterium violaceum ATCC 12472 as a model organism. Manuka propolis produced the strongest violacein inhibition on C. violaceum lawn agar, while bee pollen had no detectable QSI activity and honey had bactericidal activity. Fractionated manuka propolis (pooled fraction 5 or PF5) exhibited the largest violacein inhibition
zone (24.5 ± 2.5 mm) at 1 mg dry weight per disc. In C. violaceum liquid cultures, at least 450 lg/ml of manuka propolis PF5 completely inhibited violacein production. Gene expression studies of the vioABCDE operon, involved in violacein biosynthesis, showed significant (Ctwo-fold) down-regulation of vioA, vioD and vioE in response to manuka propolis PF5. A potential QSI compound identified in manuka propolis PF5 is a hydroxycinnamic acid-derivative, isoprenyl caffeate, with a [M-H] of 247. Complete violacein inhibition in C. violaceum liquid cultures was achieved with at least 50 lg/ml of commercial isoprenyl caffeate. In silico docking experiments suggest that isoprenyl caffeate may act as an inhibitor of the violacein biosynthetic pathway by acting as a competitor for the FAD-binding pockets of VioD and
A. T. Gemiarto N. N. Ninyio S. W. Lee J. Logis C. S. Y. Lim (&) Department of Biotechnology, Faculty of Applied Sciences, UCSI University, No 1, Jalan Menara Gading, UCSI Heights, 56000 Cheras, Kuala Lumpur, Malaysia e-mail: [email protected]
A. Fatima Faculty of Pharmaceutical Sciences, UCSI University, No 1, Jalan Menara Gading, UCSI Heights, 56000 Kuala Lumpur, Malaysia
Present Address: A. T. Gemiarto Comparative Genomics Centre, James Cook University, Molecular Sciences Bldg 21, Townsville, QLD 4811, Australia
E. W. C. Chan Department of Food Science and Nutrition, Faculty of Applied Sciences, UCSI University, No 1, Jalan Menara Gading, UCSI Heights, 56000 Cheras, Kuala Lumpur, Malaysia
Present Address: N. N. Ninyio Department of Microbiology, Faculty of Science, Kaduna State University, P.M.B 2339, Tafawa Balewa Way, Kaduna, Kaduna State, Nigeria
123
Antonie van Leeuwenhoek
VioA. Further studies on these compounds are warranted toward the development of anti-pathogenic drugs as adjuvants to conventional antibiotic treatments, especially in antibiotic-resistant bacterial infections. Keywords Quorum sensing inhibition Manuka propolis Isoprenyl caffeate vio operon Antibiotic resistance Chromobacterium violaceum
Introduction Quorum sensing (QS) describes the form of communication in bacterial communities, based on diffusible small signaling molecules, in a population densitydependent manner. The emergence of antibioticresistant bacteria is one of the most pressing problems in modern medical microbiology, more so the threat posed by multidrug-resistant bacterial infections, which are resistant to nearly all known classes of antibiotics. Gram-negative bacteria are the main culprits in antibiotic-resistance acquisition, which is further complicated by a dearth of new antibiotics against them. Many efforts have been attempted to combat this resistance, including the development of novel and more potent antibiotics, but to no avail. As a result, there are increasing numbers of bacterial infections for which no adequate therapeutic options exist (Slama 2008; Souli et al. 2008). However, recent findings on the regulation of virulence in bacterial pathogens have opened the door to the development of a new class of drugs called anti-pathogenic drugs (Hentzer et al. 2003; Wagner et al. 2004; Winser and Williams 2001). These drugs are able to suppress bacterial virulence, regardless of multidrug resistance, by interfering with the bacterial QS system (Rasmussen and Givskov 2006). Some examples of phenotypes controlled by QS are virulence, surface adhesion, and extracellular polymer production (Waters and Bassler 2005). Therefore, by inhibiting the QS mechanisms in pathogens, their virulence can be suppressed, and subsequently, the infection can be controlled (Hentzer and Givskov 2003). Furthermore, because QS does not affect normal bacterial growth, it is less likely to trigger generation of drug resistance (Rasmussen and Givskov 2006).
123
Several natural extracts have demonstrated quorum-sensing inhibitor (QSI) properties against Gramnegative bacteria. Examples include halogenated furanones extracted from Delisea pulchra (Martinelli et al. 2004), extracts of Terminalia catappa (Taganna et al. 2011), epigallocatechin gallate (Taganna and Rivera 2008), and propolis (Bulman et al. 2011; Lamberte et al. 2011). Bee products, particularly propolis, have been identified in the traditional medicine of various cultures around the world. Previous studies in the United States, Denmark and Philippines have shown that propolis extracts were able to inhibit QS in Gramnegative bacteria (Bulman et al. 2011; Lamberte et al. 2011; Rasmussen et al. 2005). However, the potential QSI component(s) in propolis extracts were not explored by these studies. Furthermore, no studies have yet been done on manuka propolis, an example of a single-source propolis, and investigations of evidence of QS inhibition at the molecular level by studying the expression profiles of QS-controlled genes are lacking. This is the first study to identify a class of phytochemical in manuka propolis, isoprenyl caffeate, as a potential QSI, as well as to confirm the QSI activity of manuka propolis in C. violaceum as a model organism at the molecular level. We also show, for the first time, that the mechanism of violacein inhibition by isoprenyl caffeate may be through competitive inhibition of VioD and VioA in the C. violaceum biosynthetic pathway. The findings of this study can be further applied to develop propolis-based Gram-negative anti-pathogenic drugs to combat the phenomenon of rising antibiotic resistance.
Materials and methods Commercial bee products and extraction Four commercial sources of bee products (bee pollen, manuka propolis and two types of honeys) were used in this study. The dried bee pollen pellets and honey were obtained from Ee Feng Gu Bee Farm, Cameron Highlands, Malaysia, while the manuka propolis tincture (Royale Manuka Propolis) and manuka honey (HNZ) were from New Zealand. With the exception of the manuka propolis tincture, twenty grams of each bee product were extracted three times with 200 ml of
Antonie van Leeuwenhoek
methanol for three hours. The extracts were filtered, evaporated and freeze-dried. A stock solution of 0.5 g/ml of each extract was prepared in Luria–Bertani (LB) broth. Violacein inhibition assay for extracts of bee products
50 °C, and subsequently freeze-dried (Alpha 1–4 LD Plus, Martin Christ, Germany) at 0.1 mbar and -50 °C overnight. Phytochemical analysis of the manuka propolis pooled fractions Determination of total phenolic content
Chromobacterium violaceum ATCC strain 12472 was grown and maintained in LB broth at 26 °C. Violacein inhibition assays were carried out as previously described by Lamberte et al. (2011), with slight modifications. Lawn agar plates were prepared by spreading 100 ll of overnight C. violaceum culture (adjusted to OD720 of 0.1) on the surface of LB soft agar (0.5 % of agar–agar) using sterilized glass plating beads (4 mm diameter). Subsequently, 10 mm wells were punched through the C. violaceum lawn agar and filled with 60 ll of each extract. After 24 h incubation at 26 °C, the violacein inhibition property of the extracts was assessed by measuring the diameter of the yellowish opaque halo present (indicating bacterial growth) in the absence of the purple violacein pigmentation of the bacterial lawn (indicating QS inhibition) surrounding each well. From this assay, manuka propolis exhibited the strongest inhibition of C. violaceum violacein production. Therefore, only manuka propolis was selected for further downstream investigations. Manuka propolis fractionation The manuka propolis tincture (14 g) was fractionated with a MCI Gel CHP-20P resin (Supelco, Germany), using a water:methanol ratio at a 0–100 % step gradient. Eluents (30 ml for each fraction) were collected in 50 ml Falcon tubes. The column was flushed with ethyl acetate to remove any substances that may be retained in the column. Eluents from column chromatography were pooled into fractions based on thin layer chromatography (TLC) analysis with Silica Gel 60 F254 plates (Merck, Germany) using a chloroform:methanol:water ratio of 7:3:0.5 or a petroleum ether:ethyl acetate ratio of 6:4. Spots were detected by UV illumination (254 nm and 366 nm) and by spraying with 10 % sulphuric acid with heating. Fractions with a similar TLC profile were pooled and 8 pools were collected in total (labeled as PF1-PF8, indicating the sequence of elution from early to late). Prior to storage, the solvent was dried in a rotary evaporator (Rotavapor R-210, Buchi, Switzerland) at
The total phenolic content (TPC) of extracts was determined using the Folin–Ciocalteu assay. Samples (300 ll each) were introduced into test tubes, followed by 1.5 ml of Folin–Ciocalteu’s reagent (10 times dilution) and 1.2 ml of sodium carbonate (7.5 % w/v). The tubes were allowed to stand for 30 min before absorbance at 765 nm was measured. TPC was expressed as gallic acid equivalent (GAE) in mg/ 100 g material. The calibration equation for gallic acid was y = 0.0111x - 0.0148 (R2 = 0.9998), where y is the absorbance and x is the concentration of gallic acid in mg/l. Determination of total flavonoid content Total flavonoid content (TFC) of extracts was determined using the aluminium chloride colorimetric method. Samples (500 ll) were introduced into test tubes, followed by 1.5 ml of 75 % methanol, 0.1 ml of 1 M potassium acetate and 2.8 ml of ultra-pure water. The tubes were incubated for 30 min at room temperature before absorbance at 415 nm was measured. TFC was expressed as quercetin equivalent (QE) in mg/100 g material. The calibration equation for quercetin was y = 0.0043x (R2 = 0.9997), where y is the absorbance and x is the concentration of quercetin in mg/l. Determination of total hydroxycinnamic acid content Total hydroxycinnamic acid content (HCAC) of extracts was determined using the sodium molybdate assay. Samples (300 ll) were transferred into test tubes, followed by 2.7 ml of molybdate reagent (prepared by dissolving 16.5 g sodium molybdate, 8 g dipotassium hydrogen phosphate and 7.9 g potassium dihydrogen phosphate in one l of ultra-pure water). The tubes were vortexed briefly and incubated for 10 min at room temperature before absorbance at 370 nm was measured. HCAC was expressed as chlorogenic acid equivalent (CGAE) in mg/100 g
123
Antonie van Leeuwenhoek
material. The calibration equation for chlorogenic acid was y = 0.0062x (R2 = 0.9975), where y is the absorbance and x is the concentration of chlorogenic acid in mg/l. Violacein inhibition assay for pooled fractions of manuka propolis Disc diffusion assays were carried out using paper discs (diameter 6 mm) loaded with various amounts of manuka propolis pooled fractions (dissolved in methanol, 0.1–1 mg per disc). All impregnated discs were air-dried before placing on C. violaceum lawn agar, with kanamycin (40 lg per disc) as the bactericidal control and methanol as the solvent control. After overnight incubation at 26 °C, the plates were examined for yellowish opaque halo inhibition zones in the absence of purple pigmentation around the discs. From this assay, manuka propolis pooled fraction 5 (PF5) showed the strongest violacein inhibition. Therefore, out of the eight pooled fractions, only PF5 was further investigated. Manuka propolis PF5 dose–response violacein inhibition assay Briefly, an overnight (18 h) culture of C. violaceum ATCC 12472 at OD720 of 0.1 was divided into 5 ml aliquots in 11 mm petri dishes, then added with 0, 50, 100, 150, 200, 250, 300, 350, 400, 450 and 500 lg/ml of manuka propolis PF5, respectively. The plates were incubated overnight (18 h) at 26 °C, with constant shaking at 50 rpm. After incubation, a 1 ml aliquot of each manuka propolis PF5-treated culture was used for the determination of OD at 720 nm (for cell density) and 577 nm (for violacein production). All treatments were carried out in biological triplicates. Mean OD720/ OD577 ratios from the manuka propolis PF5-treated cultures were compared to the untreated culture using Student’s unpaired T test (p = 0.05) to assess the degree of violacein inhibition. Differential gene expression assay of the vioABCDE operon Total RNA was immediately isolated from the remainder of all the treated and untreated cultures above using the Geneaid Total RNA Mini Kit (Geneaid, Taiwan) and treated with Qiagen RNase-
123
Free DNase-I (Qiagen, Germany) according to the respective manufacturers’ protocols. Total RNA quality and purity were monitored via agarose gel electrophoresis and the A260/A280 ratio. RNA was used immediately for cDNA first strand synthesis using RevertAid reverse transcriptase and random hexamers (all from Fermentas, Lithuania), according to the manufacturer’s protocols. Real-time PCR was performed for the vioABCDE operon genes, using rpoB as the reference gene, with the respective gene-specific primers (Table 1) using the SensiFAST SYBR No-ROX Kit (Bioline, UK). Cycling conditions were as follows: Activation at 95 °C for 10 min, followed by 40 cycles of denaturation at 94 °C for 15 s, annealing at the respective primers’ temperatures as in Table 1 for 20 s and elongation at 72 °C for 30 s. This was followed by a melting reaction from 68 to 95 °C. All reactions were run in triplicates in an ABI StepOne (Applied Biosystems, USA) real-time thermal cycler. The Pfaffl mathematical method of relative quantification was applied to determine differential expression according to the formula: Fold change = EtarDCt/ErefDCt (Pfaffl 2001). Therefore, cDNA reverse-transcribed from untreated control C. violaceum RNA was used to generate five-fold serial dilutions. Standard curves with amplification efficiencies (E) between 1.8 and 2.2 (where 2.0 is perfect efficiency) and correlation coefficient (R2) values of at least 0.99 were constructed for all genes. Differential expression of each condition was compared between each treatment condition and the untreated control, with rpoB as the reference gene for normalization. Data were presented as fold change (FC), where |FC| = ±1 indicates no change in expression and |FC| [ ±2 indicates significant differential expression. Identification of isoprenyl caffeate as a major compound(s) in manuka propolis PF5 High performance liquid chromatography (HPLC) Manuka propolis PF5 was dissolved in 100 % methanol and analysed using reverse-phase HPLC (Agilent 1200 series, Agilent Technology, USA) with a C18 column (4.6 mm 9 150 mm 9 5 lm). A 20-min linear gradient from 10 to 100 % methanol was used to elute the sample at 0.5 ml/min. Mobile phases were acidified with 0.1 % trifluoroacetic acid
Antonie van Leeuwenhoek Table 1 Primers for the rpoB housekeeping gene and the vioABCDE operon
Gene
GenBank Gene ID
Primer sequences
Ta (°C)
rpoB
2548952
50 -GCCCACACTTCCATCTCACCGAAAC-30
60
50 -TCCAAGACCCAGATGACCCTGTTCG-30 vioA
2547942
50 -TCCTCTCCTTCGGCCACGCA-30
60
50 -GCGCGCTTCAGCCTGGGTTA-30 vioB
2548008
50 -TGGAACAGGAAGTGCGGATG-30 0
5 -CCGGCAACAACCATTTCTCC-3 vioC vioD
58
0
2548007
50 -GATCATATCGCCCTGCAAGC-30
58
2548006
50 -CCGCTACTACTTCGAGCACA-30 50 -GTACTCGGACACGATGAGCAC-30
62
50 -CACCTTGGCGACGTATTCGG-30 vioE
2548075
50 -AGCGCCTATGTGTCGTACTG-30
55
50 -TTGAACAGGCCGTCTATCCG-30
to prevent ionisation of analytes. Elutions were monitored at 5 different wavelengths (210, 254, 268, 280 and 365 nm). Liquid chromatography–mass spectrometry (LC–MS) PF5 was dissolved in 100 % methanol and analysed using a Flexar SQ300 mass spectrometer (Perkin Elmer, USA). Sample components were separated with a C18 column (2.1 mm 9 150 mm 9 3.5 lm). A 20-min linear gradient from 10 to 100 % methanol acidified with 1 % trifluoroacetic acid was used to elute the sample at 0.2 ml/min. Liquid chromatography–mass spectrometry mass spectra were acquired in negative ion mode. Mass up to 3000 m/z was measured. Docking studies Prior to purchasing isoprenyl caffeate as a pure compound for QSI dose–response assays, docking experiments between isoprenyl caffeate, a major compound in manuka propolis PF5, and the Vio proteins A, C, D and E were performed using the CDOCKER protocol on Discovery Studio 2.5.5 (Accelrys Inc.) (Wu et al. 2003). As the three-dimensional (3-D) structure of VioB is currently still unavailable and no homologous protein structures are available in any database, therefore VioB was not included in the docking experiments. The 3-D structures of VioD (3C4A.pdb) and VioE (2ZF4.pdb) of C. violaceum were downloaded from Protein Data Bank (Berman et al. 2000). The structures for VioA and VioC were modelled using the homology modelling server of Swiss-PDB (Schwede et al. 2003). The apo structures
were selected because the binding site is identified in these structures. The structure for isoprenyl caffeate (CID_528179) was downloaded from PubChem. Control docking was done using ligands bound in the PDB files. Binding energy of the bound structures was calculated using the Calculate Binding Energy protocol of Discovery Studio. The best bound structures according to the energy profile were visualized using Pymol (Schro¨dinger) and the two-dimensional image of the binding site was elucidated using LIGPLOT (Wallace et al. 1995). Isoprenyl caffeate dose–response violacein inhibition assay To confirm the QSI effect of isoprenyl caffeate on C. violaceum, disk diffusion assays of C. violaceum with 25-100 lg per disc of commercial isoprenyl caffeate [caffeic acid 1,1-dimethylallyl ester (Fluka, Germany)] dissolved in methanol, were carried out (as in Violacein inhibition assay for pooled fractions of manuka propolis). Dose–response assays with 25–100 lg/ml of commercial isoprenyl caffeate were also carried out on C. violaceum liquid cultures (as in Manuka propolis PF5 dose–response violacein inhibition assay).
Results Manuka propolis exhibits violacein inhibition In the C. violaceum well-diffusion assay with the four bee products (±0.5 g/ml for each sample), yellowish
123
Antonie van Leeuwenhoek
opaque halos were observed around the wells containing manuka propolis and manuka honey, indicating violacein inhibition activity, with inhibition zone diameters of 27 ± 2 and 25 ± 3 mm, respectively (Fig. 1a). Bactericidal effects, indicated by the presence of a clear halo around a well, were not observed for manuka propolis, but slight bactericidal effects were observed for manuka honey. No discoloration was observed around the wells containing Malaysian honey and Malaysian bee pollen, suggesting that these bee products did not inhibit violacein production, nor had any bactericidal effects. Therefore, only manuka propolis was selected for further investigations. Hydroxycinnamic acid as the major phytochemical associated with violacein inhibition The phytochemical screening of TPC, TFC, and HCAC confirmed that all eight manuka propolis fractions (PF1-PF8) are rich in phenolic compounds (Table 2). PF6 had the highest flavonoid content, while manuka propolis PF5 had the highest level of hydroxycinnamic acids. PF1-PF3 were not included in violacein inhibition investigations due to insufficient yield. Of PF4-PF8, the strength of violacein inhibition was observed to correlate with hydroxycinnamic acid content, where manuka propolis PF5 exhibited the strongest violacein inhibition activity (24.5 ± 2.5 mm at 1 mg per disc) (Fig. 1b). Manuka propolis PF5 significantly inhibits violacein production without inhibiting bacterial growth Dose–response assays of C. violaceum liquid cultures with 0-500 lg/ml of manuka propolis PF5 saw an inverse correlation between manuka propolis PF5 concentrations and intensity of the purple colour, where a concentration of at least 300 lg/ml of manuka propolis PF5 was found to significantly decrease (p \ 0.05) the intensity of the purple colour compared to the untreated control, while treatment with at least 450 lg/ml of manuka propolis PF5 resulted in turbid yellowish suspensions with no hint of purple, indicating bacterial growth in the absence of violacein production. Normalised colony forming unit (cfu) counts from viable cell count assays using Plate Count Agar (PCA) showed that manuka propolis PF5, even at
123
Antonie van Leeuwenhoek b Fig. 1 Manuka propolis (MP) shows the strongest violacein
inhibition followed by manuka honey (MH) with slight bactericidal effects, while Malaysian honey (MS H) and Malaysian bee pollen (MS BP) do not show signs of violacein inhibition (a). Manuka propolis PF5 is the pooled fraction with the strongest inhibition of violacein production, where 1 mg of manuka propolis PF5 produced a 24.5 ± 2.5 mm violacein inhibition zone with no antibacterial activity (b). 25 lg of isoprenyl caffeate was the minimum inhibitory concentration that exhibited 6.22 ± 0.10 mm of violacein inhibition with no antibacterial activity, while 100 lg of isoprenyl caffeate gave the strongest inhibition of violacein production (9.25 ± 0.12 mm inhibition zone produced) without antibacterial activity. IC denotes isoprenyl caffeate (c)
500 lg/ml, did not have significant antibacterial activity (p [ 0.05). Manuka propolis PF5 influences vioABCDE operon expression In the CviIR system of C. violaceum, CviI synthesizes the autoinducer C10-homoserine lactone (C10-HSL) which binds to and activates CviR, a cytoplasmic DNA-binding transcription factor, resulting in activation of the vioA promoter. The genes involved in violacein biosynthesis from L-tryptophan are organised into a single operon: the vioABCDE operon (August et al. 2000; Brazilian National Project Consortium 2003; Stauff and Bassler 2011). The vioABCDE operon that controls the violacein biosynthetic pathway (Fig. 2) contains five genes (each with
their own ribosome-binding sites) which encode for the enzymes tryptophan 2-monooxygenase (vioA), polyketide synthase (vioB), monooxygenase (vioC), hydroxylase (vioD) and an enzyme with no characterized homologues but is involved in catalysing the formation of intermediates at various points of the pathway towards the subsequent synthesis of violacein (vioE) (Ryan et al. 2008). In the present study, we found that treatment of C. violaceum with 300-450 lg/ml of manuka propolis PF5 consistently down-regulated vioD and vioE by at least two-fold compared to the untreated control (Fig. 3). In addition, vioA was also significantly down-regulated by nine-fold with 450 lg/ml of manuka propolis PF5 treatment, compared to the untreated control (Fig. 3). Treatment with 500 lg/ml of manuka propolis PF5, a higher concentration than the 450 lg/ml required to completely inhibit violacein production, significantly up-regulated vioB, vioD and vioA, with no significant changes in vioE expression, compared to the untreated control. We hypothesize that this change in the expression profile of the vio genes, especially the increase in expression of vioB and vioD, may be due to switching from the canonical pathway of violacein synthesis (which is now completely inhibited) that uses VioA to convert L-tryptophan to indole-3-pyruvic acid, to the alternative pathway (Fig. 2) which uses VioD to convert L-tryptophan to 5-hydroxytryptophan in the absence of VioA.
Table 2 Phytochemical screening of manuka propolis pooled fractions Pool
TPC (mg GAE/g)a
PF1
204 ± 5.69
TFC (mg QE/g)a
HCAC (mg CGAE/g)a
16 ± 1.17
37 ± 1.87
Violacein inhibition zone (mm) at 100 lg per discb NA
PF2
775 ± 8.52
83 ± 0.82
102 ± 1.06
NA
PF3 PF4
458 ± 4.49 347 ± 1.55
52 ± 1.75 121 ± 1.05
94 ± 2.03 158 ± 1.03
NA 17.3 ± 1.2
PF5
366 ± 5.15
112 ± 0.59
324 ± 3.12
20.5 ± 2.2
PF6
375 ± 3.14
220 ± 0.84
142 ± 0.86
6.5 ± 0.2
PF7
373 ± 1.03
106 ± 1.61
114 ± 1.02
PF8
99 ± 4.00
116 ± 1.78
34 ± 0.33
10.5 ± 0.8 ND
Values for PF5 are in italics for emphasis TPC total phenolic content; TFC total flavonoid content; HCAC hydroxycinnamic acid content; NA not available; ND not detected a
Values of TPC, TFC and HCAC are mean ± SD (n = 3). Results are expressed in terms of per gram of fraction, where GAE is the gallic acid equivalent, QE is the quercetin equivalent, and CGAE is the chlorogenic acid equivalent b
Values of violacein inhibition zone are means of three biological replicate ± SD (n = 3)
123
Antonie van Leeuwenhoek
acid derivative with [M-H] = 247 (Fig. 4), which was identified as isoprenyl caffeate (IUPAC: 3-methylbut-2enyl (E)-3-(3,4-dihydroxyphenyl)prop-2-enoate) (Falca˜o et al. 2010).
Potential Vio protein-isoprenyl caffeate interactions
Fig. 2 Violacein biosynthetic pathway. Either VioA or VioD generates indole-3-pyruvic acid (IPA) imine from L-tryptophan, where VioB then converts the IPA imine into a dimer, which is transformed into prodeoxyviolacein only in the presence of VioE, after which VioC and VioD hydroxylate prodeoxyviolacein to form violacein (adapted from August et al. 2000; Ryan et al. 2008)
Manuka propolis PF5 consists mainly of hydroxycinnamic acids and flavonoids, with isoprenyl caffeate as the major component The highest HPLC peak was located at approximately 17.2 min, where a UV-spectrum analysis of this peak revealed that the major component belongs to the hydroxycinnamic acid group, which has a unique UVspectrum profile, with a shoulder at 290-300 nm and a peak at 320–330 nm (Gouveia and Castilho 2012). Liquid chromatography–mass spectrometry (LC–MS) analysis confirmed the presence of a hydroxycinnamic
123
Docking studies of isoprenyl caffeate-Vio protein complexes painted a possible explanation for the different responses of the vioA, D and E genes to treatment with different concentrations of manuka propolis PF5. The predicted binding energies between isoprenyl caffeate and the VioA, VioC, VioD and VioE enzymes are -108.76, -123.57, -148.29 and -133.40 kcal/mol, respectively. From the results, the most probable target of isoprenyl caffeate is VioD, with the strongest binding energy of -148.29 kcal/mol (Fig. 5). These results are in agreement with our experimental evidence, as inhibition of VioD would prevent the formation of the purple pigment (August et al. 2000). Analysis of the VioD binding site indicates that isoprenyl caffeate may competitively inhibit the binding of flavin adenine dinucleotide (FAD), which is required for the activation of this enzyme. The essential binding residues in the site are Ala8, Gly9, Glu32, Lys33, Asn34, Arg97, Arg105 and Asp269. Glu32, Arg97 and Arg105 contribute to hydrogen bonding of both ligands to the enzyme. The binding energy of isoprenyl caffeate (-148.29 kcal/mol) was lower compared to FAD (-800.235 kcal/mol). However, blocking of the same residues is enough to hinder the activity of the enzyme. Since VioA is the rate-limiting enzyme in the violacein biosynthesis pathway, it is important to show the interactions of isoprenyl caffeate, if any, with the enzyme. Because the three-dimensional structure of VioA is not available, docking was carried out with a modelled structure obtained using the ModBase modelling server. The template used by the server was L-glutamate oxidase from Streptomyces sp (2E1M.pdb). The only returned structure was the binding pocket of the oxidase (chain A) that had a sequence identity of 44 %. The binding pocket is the same as FAD (Fig. 6).The predicted binding energies of isoprenyl caffeate and FAD to the modelled part are -108 and -417 kcal/mol, respectively. The binding pocket residues include
Antonie van Leeuwenhoek
Fig. 3 Increasing concentrations of manuka propolis PF5 (0–500 lg/ml) inhibits violacein production of C. violaceum ATCC 12472. From left to right 0 (untreated control), 50, 100, 150, 200, 250, 300, 350, 400 and 450 lg/ml of manuka propolis PF5-treated cultures, with the 500 lg/ml treated culture (phenotypically similar to the 450 lg/ml treated culture) directly behind at the second row (a). Consolidated real-time PCR differential expression profile of the vioABCDE operon genes. Differential expression of each gene with 300, 350, 400,
450 and 500 lg/ml of manuka propolis PF5, relative to the untreated control, is presented as fold change, where no change in expression is expressed as 1. All fold changes have been normalized to rpoB as the reference gene. Data are means of fold changes with standard deviations from three independent experiments amplified in triplicates. An asterisk indicates significant differential expression ([± two-fold change compared to control) (b)
Gly65, Ala66, Gly67, Ile68, Glu88, Ala89, Arg97 and Gly96. Isoprenyl caffeate forms hydrogen bonds with Ala66, Glu88 and Arg97, while FAD binds to Ala69, Met123, Arg124, Met354 and Arg97. Thus, Arg97 is the common binding residue for the hydrogen bonding of isoprenyl caffeate and FAD to the binding pocket. Despite the limitations of the modelled VioA structure, combining the binding energy and docking interaction with our experimental results, it is proposed that VioA could also be a target for isoprenyl caffeate.
Isoprenyl caffeate is a quorum-sensing inhibitor A complete violacein inhibition zone of 6.22 ± 0.12 mm was achieved with at least 25 lg of isoprenyl caffeate (Fig. 1c), with no significant antibacterial effects as determined by plate count assay (data not shown). In liquid cultures, inhibition of C. violaceum violacein production was achieved with at least 50 lg/ml of commercial isoprenyl caffeate (data not shown). Thus, it is likely that isoprenyl caffeate, with violacein-inhibition properties and yet
123
Antonie van Leeuwenhoek Fig. 4 Representative ionogram of manuka propolis PF5 detected by Flexar SQ300. Isoprenyl caffeate (M-H = 247, Rt = 14.44 min) was identified as a major compound of manuka propolis PF5
Fig. 5 vioD docking studies. Crystal structure of FAD bound to vioD (a). Isoprenyl caffeate bound to the modelled pocket of VioD (b). LIGPLOT of FAD bound to VioD showing the binding site residues (c). LIGPLOT of VioD structure showing the important residues in the binding pocket (d). All common binding residues are encircled
123
Antonie van Leeuwenhoek Fig. 6 VioA docking studies. Crystal structure of L-glutamate oxidase from Streptomyces sp. with FAD in the binding pocket (a). Isoprenyl caffeate bound to the modelled pocket of VioA (b). LIGPLOT of the L-glutamate oxidase structure showing the important residues in the binding pocket (c). LIGPLOT of VioA structure showing the important residues in the binding pocket (d). The common binding residues are encircled
no significant antibacterial properties towards C. violaceum, is indeed a quorum-sensing inhibitor.
Discussion To be potential anti-pathogenic compounds, ideally QSI compounds should not interfere with normal bacteria growth, otherwise the bacteria has the tendency to generate resistance against these QSI compounds, rendering them as being no different from antibiotics (Rasmussen and Givskov 2006). In this study, out of the four types of bee products investigated, the two Malaysian bee products (honey and bee pollen) did not exhibit any violacein inhibitory properties. Manuka honey contained both violacein-
inhibition and antibacterial activities, as was also found by Wang et al. (2012), whereas manuka propolis only inhibited violacein production without antibacterial effects. The violacein inhibition property of manuka propolis was also stronger than that of manuka honey, thus manuka propolis was selected for further investigation in this study. Interestingly in the present study, manuka honey was identified to have a low content of phenolic compounds, which are known to exhibit QSI activity. It is possible that the violacein inhibition and bactericidal effect of manuka honey may be attributed to its high sugar content, as seen in a previous study by Lee et al. (2011), which proposed that the QSI activity of diluted honey against E. coli O157:H7 was through a mechanism of action where a high sugar concentration
123
Antonie van Leeuwenhoek
may prevent the E. coli acyl homoserine lactones (AHL) from re-entering the cells; therefore blocking AHL-receptor complex formation. Manuka propolis fraction 5 (PF5) exhibited the strongest violacein inhibition, followed by PF4, where both fractions were observed to contain high amounts of phenolics, flavonoids and, particularly, hydroxycinnamic acid. Previous studies investigating the QSI properties of various natural sources suggest that phenolic compounds, such as phenolic acids and flavonoids, might be the active component responsible (Annapoorani et al. 2012; Lee et al. 2011; Singh et al. 2009). Other PFs, with low hydroxycinnamic acid content, showed weak or no violacein inhibition properties. Thus, the active QSI compound(s) in manuka propolis is suggested to belong to the hydroxycinnamic acid group, particularly isoprenyl caffeate. This suggestion is supported by the work of Brackman et al. (2008; 2011), which found cinnamaldehyde (a derivative of the hydroxycinnamic acid group) and cinnamaldehyde derivatives to possess QSI effects on Vibrio spp. as a QS model organism. It would seem that this family of compounds are active QSI compounds, regardless of the aldehyde or carboxylic acid forms (Brackman et al. 2008; 2011). The minimum concentration of manuka propolis PF5 required for phenotypically observable QSI effects in C. violaceum is 300 lg/ml. This concentration is much less than the 28 mg/ml of Filipino propolis ethanolic extract required for violacein inhibition in C. violaceum ATCC 12472 (Lamberte et al. 2011) and the 10 mg/ml of Spanish propolis tincture required for violacein inhibition in C. violaceum CV026 (Bulman et al. 2011), a strain inherently unable to produce the signaling molecule N-hexanoyl homoserine lactone (C6-HSL). In these previous studies, both types of propolis were not subjected to further fractionation and purification. In this study, manuka propolis, a single-source propolis, was fractionated and subjected to further purification, where we identified a major compound— isoprenyl caffeate—as the likely QSI compound present in manuka propolis. Complete violacein inhibition in C. violaceum liquid cultures was achieved with a minimum concentration of 50 lg/ml of commercial isoprenyl caffeate, thus confirming it as a QSI compound. The molecular observations of the manuka propolis PF5 dose–response gene expression experiments correlate with our in silico molecular docking experiments for isoprenyl caffeate, the postulated QSI compound in
123
manuka propolis PF5. Manuka propolis PF5 treatment of 300 lg/ml began down-regulating vioD and vioE, with isoprenyl caffeate showing the highest binding energy with VioD, followed by VioE. This suggests that the way which isoprenyl caffeate can inhibit the violacein biosynthetic pathway is by competitively inhibiting the binding of FAD to the VioD enzyme, thus deactivating the enzyme and preventing the hydroxylation of prodeoxyviolacein to the purple pigment violacein. In turn, less VioE is required. Subsequently, the higher concentration of manuka propolis PF5 required to significantly down-regulate vioA (450 lg/ml) compared to vioD and vioE is also evident in the docking results, where a comparatively low binding energy was obtained for the binding of isoprenyl caffeate to the VioA enzyme. As VioA is the rate-limiting enzyme that is responsible for the introduction of oxygen into the first intermediate product of the pathway, indole 3-pyruvic acid, deactivation of VioA through the competitive inhibition of FAD-binding by isoprenyl caffeate may have caused the first substrate to be incorrectly formed, resulting in the inhibition of the violacein biosynthetic pathway.
Conclusions This is the first study to elucidate isoprenyl caffeate, a major compound in manuka propolis, as a potential QSI compound, where a commercial isoprenyl caffeate pure compound was confirmed to exhibit QSI properties on C. violaceum solid and liquid cultures without significant antibacterial effects. By correlating the phenotypic observation of violacein inhibition to the molecular responses of the vioABCDE operon in C. violaceum, it was observed that violacein inhibition was achieved in tandem with the down-regulation of vioA, vioD and vioE. Docking studies suggest that isoprenyl caffeate acts as a competitor of the FADbinding pockets of VioA and VioD. Acknowledgments This work was done with funding from UCSI University’s Faculty of Applied Sciences and the MAKNA Cancer Research Award 2012. We thank Prof. Tom Coenye for his constructive comments of the early findings of this study and helpful insights into the QSI activity of cinnamaldehyde and cinnamaldehyde derivatives. We also thank the Drug Design & Development Research Group (DDDRG) of the University of Malaya for use of the docking software.
Antonie van Leeuwenhoek Conflict of interest The authors declare that they have no conflict of interest.
References Annapoorani A, Parameswari R, Pandian SK, Ravi AV (2012) Methods to determine antipathogenic potential of phenolic and flavonoid compounds against urinary pathogen Serratia marcescens. J Microbiol Methods 91:208–211 August PR, Grossman TH, Minor C, Draper MP, MacNeil IA, Pemberton JM, Call KM, Holt D, Osburne MS (2000) Sequence analysis and functional characterization of the violacein biosynthetic pathway from Chromobacterium violaceum. J Mol Microbiol Biotechnol 2:513–519 Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, Bourne PE (2000) The protein data bank. Nucleic Acids Res 28:235–242 Brackman G, Defoirdt T, Miyamoto C, Bossier P, Van CS, Nelis H, Coenye T (2008) Cinnamaldehyde and cinnamaldehyde derivatives reduce virulence in Vibrio spp. by decreasing the DNA-binding activity of the quorum sensing response regulator LuxR. BMC Microbiol 8:149 Brackman G, Celen S, Hillaert U, Van Calenbergh S, Cos P, Maes L, Nelis HJ, Coenye T (2011) Structure-activity relationship of cinnamaldehyde analogs as inhibitors of AI2 based quorum sensing and their effect on virulence of Vibrio spp. PLoS One 6:e16084 Brazilian National Project Consortium (2003) The complete genome sequence of Chromobacterium violaceum reveals remarkable and exploitable bacterial adaptability. Proc Natl Acad Sci USA 100:11660–11665 Bulman Z, Lee P, Hudson AO, Savka MA (2011) A novel property of propolis (bee glue): anti-pathogenic activity by inhibition of N-acyl-homoserine lactone mediated signaling in bacteria. J Ethnopharmacol 138:788–797 Falca˜o SI, Vilas-Boas M, Estevinho LM, Barros C, Domingues MR, Cardoso SM (2010) Phenolic characterization of Northeast Portuguese propolis: usual and unusual compounds. Anal Bioanal Chem 396:887–897 Gouveia SC, Castilho PC (2012) Validation of a HPLC-DAD– ESI/MSn method for caffeoylquinic acids separation, quantification and identification in medicinal Helichrysum species from Macaronesia. Food Res Int 45:362–368 Hentzer M, Givskov M (2003) Pharmacological inhibition of quorum sensing for the treatment of chronic bacterial infections. J Clin Invest 112:1300–1307 Hentzer M, Wu H, Andersen JB, Riedel K, Rasmussen TB, Bagge N, Kumar N, Schembri MA, Song Z, Kristoffersen P, Manefield M, Costerton JW, Molin S, Eberl L, Steinberg P, Kjelleberg S, Høiby N, Givskov M (2003) Attenuation of Pseudomonas aeruginosa virulence by quorum sensing inhibitors. EMBO J 22:3803–3815 Lamberte LE, Cabrera EC, Rivera WL (2011) Activity of the ethanolic extract of propolis (EEP) as a potential inhibitor of quorum sensing-mediated pigment production in Chromobacterium violaceum and virulence factor production in Pseudomonas aeruginosa. Philipp Agric Sci 94:14–22
Lee JH, Park JH, Kim JA, Neupane GP, Cho MH, Lee CS, Lee J (2011) Low concentrations of honey reduce biofilm formation, quorum sensing, and virulence in Escherichia coli O157:H7. Biofouling 27:1095–1104 Martinelli D, Grossmann G, Se´quin U, Brandl H, Bachofen M (2004) Effects of natural and chemically synthesized furanones on quorum sensing in Chromobacterium violaceum. BMC Microbiol 4:25 Pfaffl MW (2001) A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 29:e45 Rasmussen TB, Givskov M (2006) Quorum-sensing inhibitors as anti-pathogenic drugs. Int J Med Microbiol 296:149–161 Rasmussen TB, Bjarnsholt T, Skindersoe ME, Hentzer M, Kristoffersen P, Ko¨te M, Nielsen J, Eberl L, Givskov M (2005) Screening for quorum-sensing inhibitors (QSI) by use of a novel genetic system, the QSI Selector. J Bacteriol 187:1799–1814 Ryan KS, Balibar CJ, Turo KE, Walsh CT, Drennan CL (2008) The violacein biosynthetic enzyme VioE shares a fold with lipoprotein transporter proteins. J Biol Chem 283:6467–6475 Schwede T, Kopp J, Guex N, Peitsch M (2003) SWISSMODEL: an automated protein homology-modeling server. Nucleic Acids Res 31:3381–3385 Singh BN, Singh BR, Singh RL, Prakash D, Dhakarey R, Upadhyay G, Singh HB (2009) Oxidative DNA damage protective activity, antioxidant and anti-quorum sensing potentials of Moringa oleifera. Food Chem Toxicol 47:1109–1116 Slama TG (2008) Gram-negative antibiotic resistance: there is a price to pay. Crit Care 12(Suppl 10):1–7 Souli M, Galani I, Giamarellou H (2008) Emergence of extensively drug-resistant and pandrug-resistant Gram-negative bacilli in Europe. Euro Surveill 13:19045 Stauff DL, Bassler BL (2011) Quorum sensing in Chromobacterium violaceum: DNA recognition and gene regulation by the CviR receptor. J Bacteriol 193:3871–3878 Taganna JC, Rivera WL (2008) Epigallocatechin gallate from Camellia sinensis L. (Kuntze) is a potential quorum sensing inhibitor in Chromobacterium violaceum. Sci Diliman 20:24–30 Taganna JC, Quanico JP, Perono RM, Amor EC, Rivera WL (2011) Tannin-rich fraction from Terminalia catappa inhibits quorum sensing (QS) in Chromobacterium violaceum and the QS-controlled biofilm maturation and LasA staphylolytic activity in Pseudomonas aeruginosa. J Ethnopharmacol 134:865–871 Wagner VE, Gillis RJ, Iglewski BH (2004) Transcriptome analysis of quorum-sensing regulation and virulence factor expression in Pseudomonas aeruginosa. Vaccine 22(Suppl 1):S15–S20 Wallace AC, Laskowski RA, Thornton JM (1995) LIGPLOT: a program to generate schematic diagrams of protein-ligand interactions. Protein Eng 8:127–134 Wang R, Starkey M, Hazan R, Rahme LG (2012) Honey’s ability to counter bacterial infections arises from both bactericidal compounds and QS inhibition. Front Microbiol 2012:144
123
Antonie van Leeuwenhoek Waters CM, Bassler BL (2005) Quorum sensing: cell-to-cell communication in bacteria. Annu Rev Cell Dev Biol 21:319–346 Winser K, Williams P (2001) Quorum sensing and the regulation of virulence gene expression in pathogenic bacteria. Int J Med Microbiol 291:131–143
123
Wu G, Robertson DH, Brooks CL 3rd, Vieth M (2003) Detailed analysis of grid-based molecular docking: a case study of CDOCKER-A CHARMm-based MD docking algorithm. J Comput Chem 24:1549–1562
ORIGINAL PAPER
Isoprenyl caffeate, a major compound in manuka propolis, is a quorum-sensing inhibitor in Chromobacterium violaceum Adrian Tandhyka Gemiarto . Nathaniel Nyakaat Ninyio . Siew Wei Lee . Joko Logis . Ayesha Fatima . Eric Wei Chiang Chan . Crystale Siew Ying Lim
Received: 6 January 2015 / Accepted: 4 June 2015 Ó Springer International Publishing Switzerland 2015
Abstract The emergence of antibiotic-resistant bacterial pathogens, especially Gram-negative bacteria, has driven investigations into suppressing bacterial virulence via quorum sensing (QS) inhibition strategies instead of bactericidal and bacteriostatic approaches. Here, we investigated several bee products for potential compound(s) that exhibit significant QS inhibitory (QSI) properties at the phenotypic and molecular levels in Chromobacterium violaceum ATCC 12472 as a model organism. Manuka propolis produced the strongest violacein inhibition on C. violaceum lawn agar, while bee pollen had no detectable QSI activity and honey had bactericidal activity. Fractionated manuka propolis (pooled fraction 5 or PF5) exhibited the largest violacein inhibition
zone (24.5 ± 2.5 mm) at 1 mg dry weight per disc. In C. violaceum liquid cultures, at least 450 lg/ml of manuka propolis PF5 completely inhibited violacein production. Gene expression studies of the vioABCDE operon, involved in violacein biosynthesis, showed significant (Ctwo-fold) down-regulation of vioA, vioD and vioE in response to manuka propolis PF5. A potential QSI compound identified in manuka propolis PF5 is a hydroxycinnamic acid-derivative, isoprenyl caffeate, with a [M-H] of 247. Complete violacein inhibition in C. violaceum liquid cultures was achieved with at least 50 lg/ml of commercial isoprenyl caffeate. In silico docking experiments suggest that isoprenyl caffeate may act as an inhibitor of the violacein biosynthetic pathway by acting as a competitor for the FAD-binding pockets of VioD and
A. T. Gemiarto N. N. Ninyio S. W. Lee J. Logis C. S. Y. Lim (&) Department of Biotechnology, Faculty of Applied Sciences, UCSI University, No 1, Jalan Menara Gading, UCSI Heights, 56000 Cheras, Kuala Lumpur, Malaysia e-mail: [email protected]
A. Fatima Faculty of Pharmaceutical Sciences, UCSI University, No 1, Jalan Menara Gading, UCSI Heights, 56000 Kuala Lumpur, Malaysia
Present Address: A. T. Gemiarto Comparative Genomics Centre, James Cook University, Molecular Sciences Bldg 21, Townsville, QLD 4811, Australia
E. W. C. Chan Department of Food Science and Nutrition, Faculty of Applied Sciences, UCSI University, No 1, Jalan Menara Gading, UCSI Heights, 56000 Cheras, Kuala Lumpur, Malaysia
Present Address: N. N. Ninyio Department of Microbiology, Faculty of Science, Kaduna State University, P.M.B 2339, Tafawa Balewa Way, Kaduna, Kaduna State, Nigeria
123
Antonie van Leeuwenhoek
VioA. Further studies on these compounds are warranted toward the development of anti-pathogenic drugs as adjuvants to conventional antibiotic treatments, especially in antibiotic-resistant bacterial infections. Keywords Quorum sensing inhibition Manuka propolis Isoprenyl caffeate vio operon Antibiotic resistance Chromobacterium violaceum
Introduction Quorum sensing (QS) describes the form of communication in bacterial communities, based on diffusible small signaling molecules, in a population densitydependent manner. The emergence of antibioticresistant bacteria is one of the most pressing problems in modern medical microbiology, more so the threat posed by multidrug-resistant bacterial infections, which are resistant to nearly all known classes of antibiotics. Gram-negative bacteria are the main culprits in antibiotic-resistance acquisition, which is further complicated by a dearth of new antibiotics against them. Many efforts have been attempted to combat this resistance, including the development of novel and more potent antibiotics, but to no avail. As a result, there are increasing numbers of bacterial infections for which no adequate therapeutic options exist (Slama 2008; Souli et al. 2008). However, recent findings on the regulation of virulence in bacterial pathogens have opened the door to the development of a new class of drugs called anti-pathogenic drugs (Hentzer et al. 2003; Wagner et al. 2004; Winser and Williams 2001). These drugs are able to suppress bacterial virulence, regardless of multidrug resistance, by interfering with the bacterial QS system (Rasmussen and Givskov 2006). Some examples of phenotypes controlled by QS are virulence, surface adhesion, and extracellular polymer production (Waters and Bassler 2005). Therefore, by inhibiting the QS mechanisms in pathogens, their virulence can be suppressed, and subsequently, the infection can be controlled (Hentzer and Givskov 2003). Furthermore, because QS does not affect normal bacterial growth, it is less likely to trigger generation of drug resistance (Rasmussen and Givskov 2006).
123
Several natural extracts have demonstrated quorum-sensing inhibitor (QSI) properties against Gramnegative bacteria. Examples include halogenated furanones extracted from Delisea pulchra (Martinelli et al. 2004), extracts of Terminalia catappa (Taganna et al. 2011), epigallocatechin gallate (Taganna and Rivera 2008), and propolis (Bulman et al. 2011; Lamberte et al. 2011). Bee products, particularly propolis, have been identified in the traditional medicine of various cultures around the world. Previous studies in the United States, Denmark and Philippines have shown that propolis extracts were able to inhibit QS in Gramnegative bacteria (Bulman et al. 2011; Lamberte et al. 2011; Rasmussen et al. 2005). However, the potential QSI component(s) in propolis extracts were not explored by these studies. Furthermore, no studies have yet been done on manuka propolis, an example of a single-source propolis, and investigations of evidence of QS inhibition at the molecular level by studying the expression profiles of QS-controlled genes are lacking. This is the first study to identify a class of phytochemical in manuka propolis, isoprenyl caffeate, as a potential QSI, as well as to confirm the QSI activity of manuka propolis in C. violaceum as a model organism at the molecular level. We also show, for the first time, that the mechanism of violacein inhibition by isoprenyl caffeate may be through competitive inhibition of VioD and VioA in the C. violaceum biosynthetic pathway. The findings of this study can be further applied to develop propolis-based Gram-negative anti-pathogenic drugs to combat the phenomenon of rising antibiotic resistance.
Materials and methods Commercial bee products and extraction Four commercial sources of bee products (bee pollen, manuka propolis and two types of honeys) were used in this study. The dried bee pollen pellets and honey were obtained from Ee Feng Gu Bee Farm, Cameron Highlands, Malaysia, while the manuka propolis tincture (Royale Manuka Propolis) and manuka honey (HNZ) were from New Zealand. With the exception of the manuka propolis tincture, twenty grams of each bee product were extracted three times with 200 ml of
Antonie van Leeuwenhoek
methanol for three hours. The extracts were filtered, evaporated and freeze-dried. A stock solution of 0.5 g/ml of each extract was prepared in Luria–Bertani (LB) broth. Violacein inhibition assay for extracts of bee products
50 °C, and subsequently freeze-dried (Alpha 1–4 LD Plus, Martin Christ, Germany) at 0.1 mbar and -50 °C overnight. Phytochemical analysis of the manuka propolis pooled fractions Determination of total phenolic content
Chromobacterium violaceum ATCC strain 12472 was grown and maintained in LB broth at 26 °C. Violacein inhibition assays were carried out as previously described by Lamberte et al. (2011), with slight modifications. Lawn agar plates were prepared by spreading 100 ll of overnight C. violaceum culture (adjusted to OD720 of 0.1) on the surface of LB soft agar (0.5 % of agar–agar) using sterilized glass plating beads (4 mm diameter). Subsequently, 10 mm wells were punched through the C. violaceum lawn agar and filled with 60 ll of each extract. After 24 h incubation at 26 °C, the violacein inhibition property of the extracts was assessed by measuring the diameter of the yellowish opaque halo present (indicating bacterial growth) in the absence of the purple violacein pigmentation of the bacterial lawn (indicating QS inhibition) surrounding each well. From this assay, manuka propolis exhibited the strongest inhibition of C. violaceum violacein production. Therefore, only manuka propolis was selected for further downstream investigations. Manuka propolis fractionation The manuka propolis tincture (14 g) was fractionated with a MCI Gel CHP-20P resin (Supelco, Germany), using a water:methanol ratio at a 0–100 % step gradient. Eluents (30 ml for each fraction) were collected in 50 ml Falcon tubes. The column was flushed with ethyl acetate to remove any substances that may be retained in the column. Eluents from column chromatography were pooled into fractions based on thin layer chromatography (TLC) analysis with Silica Gel 60 F254 plates (Merck, Germany) using a chloroform:methanol:water ratio of 7:3:0.5 or a petroleum ether:ethyl acetate ratio of 6:4. Spots were detected by UV illumination (254 nm and 366 nm) and by spraying with 10 % sulphuric acid with heating. Fractions with a similar TLC profile were pooled and 8 pools were collected in total (labeled as PF1-PF8, indicating the sequence of elution from early to late). Prior to storage, the solvent was dried in a rotary evaporator (Rotavapor R-210, Buchi, Switzerland) at
The total phenolic content (TPC) of extracts was determined using the Folin–Ciocalteu assay. Samples (300 ll each) were introduced into test tubes, followed by 1.5 ml of Folin–Ciocalteu’s reagent (10 times dilution) and 1.2 ml of sodium carbonate (7.5 % w/v). The tubes were allowed to stand for 30 min before absorbance at 765 nm was measured. TPC was expressed as gallic acid equivalent (GAE) in mg/ 100 g material. The calibration equation for gallic acid was y = 0.0111x - 0.0148 (R2 = 0.9998), where y is the absorbance and x is the concentration of gallic acid in mg/l. Determination of total flavonoid content Total flavonoid content (TFC) of extracts was determined using the aluminium chloride colorimetric method. Samples (500 ll) were introduced into test tubes, followed by 1.5 ml of 75 % methanol, 0.1 ml of 1 M potassium acetate and 2.8 ml of ultra-pure water. The tubes were incubated for 30 min at room temperature before absorbance at 415 nm was measured. TFC was expressed as quercetin equivalent (QE) in mg/100 g material. The calibration equation for quercetin was y = 0.0043x (R2 = 0.9997), where y is the absorbance and x is the concentration of quercetin in mg/l. Determination of total hydroxycinnamic acid content Total hydroxycinnamic acid content (HCAC) of extracts was determined using the sodium molybdate assay. Samples (300 ll) were transferred into test tubes, followed by 2.7 ml of molybdate reagent (prepared by dissolving 16.5 g sodium molybdate, 8 g dipotassium hydrogen phosphate and 7.9 g potassium dihydrogen phosphate in one l of ultra-pure water). The tubes were vortexed briefly and incubated for 10 min at room temperature before absorbance at 370 nm was measured. HCAC was expressed as chlorogenic acid equivalent (CGAE) in mg/100 g
123
Antonie van Leeuwenhoek
material. The calibration equation for chlorogenic acid was y = 0.0062x (R2 = 0.9975), where y is the absorbance and x is the concentration of chlorogenic acid in mg/l. Violacein inhibition assay for pooled fractions of manuka propolis Disc diffusion assays were carried out using paper discs (diameter 6 mm) loaded with various amounts of manuka propolis pooled fractions (dissolved in methanol, 0.1–1 mg per disc). All impregnated discs were air-dried before placing on C. violaceum lawn agar, with kanamycin (40 lg per disc) as the bactericidal control and methanol as the solvent control. After overnight incubation at 26 °C, the plates were examined for yellowish opaque halo inhibition zones in the absence of purple pigmentation around the discs. From this assay, manuka propolis pooled fraction 5 (PF5) showed the strongest violacein inhibition. Therefore, out of the eight pooled fractions, only PF5 was further investigated. Manuka propolis PF5 dose–response violacein inhibition assay Briefly, an overnight (18 h) culture of C. violaceum ATCC 12472 at OD720 of 0.1 was divided into 5 ml aliquots in 11 mm petri dishes, then added with 0, 50, 100, 150, 200, 250, 300, 350, 400, 450 and 500 lg/ml of manuka propolis PF5, respectively. The plates were incubated overnight (18 h) at 26 °C, with constant shaking at 50 rpm. After incubation, a 1 ml aliquot of each manuka propolis PF5-treated culture was used for the determination of OD at 720 nm (for cell density) and 577 nm (for violacein production). All treatments were carried out in biological triplicates. Mean OD720/ OD577 ratios from the manuka propolis PF5-treated cultures were compared to the untreated culture using Student’s unpaired T test (p = 0.05) to assess the degree of violacein inhibition. Differential gene expression assay of the vioABCDE operon Total RNA was immediately isolated from the remainder of all the treated and untreated cultures above using the Geneaid Total RNA Mini Kit (Geneaid, Taiwan) and treated with Qiagen RNase-
123
Free DNase-I (Qiagen, Germany) according to the respective manufacturers’ protocols. Total RNA quality and purity were monitored via agarose gel electrophoresis and the A260/A280 ratio. RNA was used immediately for cDNA first strand synthesis using RevertAid reverse transcriptase and random hexamers (all from Fermentas, Lithuania), according to the manufacturer’s protocols. Real-time PCR was performed for the vioABCDE operon genes, using rpoB as the reference gene, with the respective gene-specific primers (Table 1) using the SensiFAST SYBR No-ROX Kit (Bioline, UK). Cycling conditions were as follows: Activation at 95 °C for 10 min, followed by 40 cycles of denaturation at 94 °C for 15 s, annealing at the respective primers’ temperatures as in Table 1 for 20 s and elongation at 72 °C for 30 s. This was followed by a melting reaction from 68 to 95 °C. All reactions were run in triplicates in an ABI StepOne (Applied Biosystems, USA) real-time thermal cycler. The Pfaffl mathematical method of relative quantification was applied to determine differential expression according to the formula: Fold change = EtarDCt/ErefDCt (Pfaffl 2001). Therefore, cDNA reverse-transcribed from untreated control C. violaceum RNA was used to generate five-fold serial dilutions. Standard curves with amplification efficiencies (E) between 1.8 and 2.2 (where 2.0 is perfect efficiency) and correlation coefficient (R2) values of at least 0.99 were constructed for all genes. Differential expression of each condition was compared between each treatment condition and the untreated control, with rpoB as the reference gene for normalization. Data were presented as fold change (FC), where |FC| = ±1 indicates no change in expression and |FC| [ ±2 indicates significant differential expression. Identification of isoprenyl caffeate as a major compound(s) in manuka propolis PF5 High performance liquid chromatography (HPLC) Manuka propolis PF5 was dissolved in 100 % methanol and analysed using reverse-phase HPLC (Agilent 1200 series, Agilent Technology, USA) with a C18 column (4.6 mm 9 150 mm 9 5 lm). A 20-min linear gradient from 10 to 100 % methanol was used to elute the sample at 0.5 ml/min. Mobile phases were acidified with 0.1 % trifluoroacetic acid
Antonie van Leeuwenhoek Table 1 Primers for the rpoB housekeeping gene and the vioABCDE operon
Gene
GenBank Gene ID
Primer sequences
Ta (°C)
rpoB
2548952
50 -GCCCACACTTCCATCTCACCGAAAC-30
60
50 -TCCAAGACCCAGATGACCCTGTTCG-30 vioA
2547942
50 -TCCTCTCCTTCGGCCACGCA-30
60
50 -GCGCGCTTCAGCCTGGGTTA-30 vioB
2548008
50 -TGGAACAGGAAGTGCGGATG-30 0
5 -CCGGCAACAACCATTTCTCC-3 vioC vioD
58
0
2548007
50 -GATCATATCGCCCTGCAAGC-30
58
2548006
50 -CCGCTACTACTTCGAGCACA-30 50 -GTACTCGGACACGATGAGCAC-30
62
50 -CACCTTGGCGACGTATTCGG-30 vioE
2548075
50 -AGCGCCTATGTGTCGTACTG-30
55
50 -TTGAACAGGCCGTCTATCCG-30
to prevent ionisation of analytes. Elutions were monitored at 5 different wavelengths (210, 254, 268, 280 and 365 nm). Liquid chromatography–mass spectrometry (LC–MS) PF5 was dissolved in 100 % methanol and analysed using a Flexar SQ300 mass spectrometer (Perkin Elmer, USA). Sample components were separated with a C18 column (2.1 mm 9 150 mm 9 3.5 lm). A 20-min linear gradient from 10 to 100 % methanol acidified with 1 % trifluoroacetic acid was used to elute the sample at 0.2 ml/min. Liquid chromatography–mass spectrometry mass spectra were acquired in negative ion mode. Mass up to 3000 m/z was measured. Docking studies Prior to purchasing isoprenyl caffeate as a pure compound for QSI dose–response assays, docking experiments between isoprenyl caffeate, a major compound in manuka propolis PF5, and the Vio proteins A, C, D and E were performed using the CDOCKER protocol on Discovery Studio 2.5.5 (Accelrys Inc.) (Wu et al. 2003). As the three-dimensional (3-D) structure of VioB is currently still unavailable and no homologous protein structures are available in any database, therefore VioB was not included in the docking experiments. The 3-D structures of VioD (3C4A.pdb) and VioE (2ZF4.pdb) of C. violaceum were downloaded from Protein Data Bank (Berman et al. 2000). The structures for VioA and VioC were modelled using the homology modelling server of Swiss-PDB (Schwede et al. 2003). The apo structures
were selected because the binding site is identified in these structures. The structure for isoprenyl caffeate (CID_528179) was downloaded from PubChem. Control docking was done using ligands bound in the PDB files. Binding energy of the bound structures was calculated using the Calculate Binding Energy protocol of Discovery Studio. The best bound structures according to the energy profile were visualized using Pymol (Schro¨dinger) and the two-dimensional image of the binding site was elucidated using LIGPLOT (Wallace et al. 1995). Isoprenyl caffeate dose–response violacein inhibition assay To confirm the QSI effect of isoprenyl caffeate on C. violaceum, disk diffusion assays of C. violaceum with 25-100 lg per disc of commercial isoprenyl caffeate [caffeic acid 1,1-dimethylallyl ester (Fluka, Germany)] dissolved in methanol, were carried out (as in Violacein inhibition assay for pooled fractions of manuka propolis). Dose–response assays with 25–100 lg/ml of commercial isoprenyl caffeate were also carried out on C. violaceum liquid cultures (as in Manuka propolis PF5 dose–response violacein inhibition assay).
Results Manuka propolis exhibits violacein inhibition In the C. violaceum well-diffusion assay with the four bee products (±0.5 g/ml for each sample), yellowish
123
Antonie van Leeuwenhoek
opaque halos were observed around the wells containing manuka propolis and manuka honey, indicating violacein inhibition activity, with inhibition zone diameters of 27 ± 2 and 25 ± 3 mm, respectively (Fig. 1a). Bactericidal effects, indicated by the presence of a clear halo around a well, were not observed for manuka propolis, but slight bactericidal effects were observed for manuka honey. No discoloration was observed around the wells containing Malaysian honey and Malaysian bee pollen, suggesting that these bee products did not inhibit violacein production, nor had any bactericidal effects. Therefore, only manuka propolis was selected for further investigations. Hydroxycinnamic acid as the major phytochemical associated with violacein inhibition The phytochemical screening of TPC, TFC, and HCAC confirmed that all eight manuka propolis fractions (PF1-PF8) are rich in phenolic compounds (Table 2). PF6 had the highest flavonoid content, while manuka propolis PF5 had the highest level of hydroxycinnamic acids. PF1-PF3 were not included in violacein inhibition investigations due to insufficient yield. Of PF4-PF8, the strength of violacein inhibition was observed to correlate with hydroxycinnamic acid content, where manuka propolis PF5 exhibited the strongest violacein inhibition activity (24.5 ± 2.5 mm at 1 mg per disc) (Fig. 1b). Manuka propolis PF5 significantly inhibits violacein production without inhibiting bacterial growth Dose–response assays of C. violaceum liquid cultures with 0-500 lg/ml of manuka propolis PF5 saw an inverse correlation between manuka propolis PF5 concentrations and intensity of the purple colour, where a concentration of at least 300 lg/ml of manuka propolis PF5 was found to significantly decrease (p \ 0.05) the intensity of the purple colour compared to the untreated control, while treatment with at least 450 lg/ml of manuka propolis PF5 resulted in turbid yellowish suspensions with no hint of purple, indicating bacterial growth in the absence of violacein production. Normalised colony forming unit (cfu) counts from viable cell count assays using Plate Count Agar (PCA) showed that manuka propolis PF5, even at
123
Antonie van Leeuwenhoek b Fig. 1 Manuka propolis (MP) shows the strongest violacein
inhibition followed by manuka honey (MH) with slight bactericidal effects, while Malaysian honey (MS H) and Malaysian bee pollen (MS BP) do not show signs of violacein inhibition (a). Manuka propolis PF5 is the pooled fraction with the strongest inhibition of violacein production, where 1 mg of manuka propolis PF5 produced a 24.5 ± 2.5 mm violacein inhibition zone with no antibacterial activity (b). 25 lg of isoprenyl caffeate was the minimum inhibitory concentration that exhibited 6.22 ± 0.10 mm of violacein inhibition with no antibacterial activity, while 100 lg of isoprenyl caffeate gave the strongest inhibition of violacein production (9.25 ± 0.12 mm inhibition zone produced) without antibacterial activity. IC denotes isoprenyl caffeate (c)
500 lg/ml, did not have significant antibacterial activity (p [ 0.05). Manuka propolis PF5 influences vioABCDE operon expression In the CviIR system of C. violaceum, CviI synthesizes the autoinducer C10-homoserine lactone (C10-HSL) which binds to and activates CviR, a cytoplasmic DNA-binding transcription factor, resulting in activation of the vioA promoter. The genes involved in violacein biosynthesis from L-tryptophan are organised into a single operon: the vioABCDE operon (August et al. 2000; Brazilian National Project Consortium 2003; Stauff and Bassler 2011). The vioABCDE operon that controls the violacein biosynthetic pathway (Fig. 2) contains five genes (each with
their own ribosome-binding sites) which encode for the enzymes tryptophan 2-monooxygenase (vioA), polyketide synthase (vioB), monooxygenase (vioC), hydroxylase (vioD) and an enzyme with no characterized homologues but is involved in catalysing the formation of intermediates at various points of the pathway towards the subsequent synthesis of violacein (vioE) (Ryan et al. 2008). In the present study, we found that treatment of C. violaceum with 300-450 lg/ml of manuka propolis PF5 consistently down-regulated vioD and vioE by at least two-fold compared to the untreated control (Fig. 3). In addition, vioA was also significantly down-regulated by nine-fold with 450 lg/ml of manuka propolis PF5 treatment, compared to the untreated control (Fig. 3). Treatment with 500 lg/ml of manuka propolis PF5, a higher concentration than the 450 lg/ml required to completely inhibit violacein production, significantly up-regulated vioB, vioD and vioA, with no significant changes in vioE expression, compared to the untreated control. We hypothesize that this change in the expression profile of the vio genes, especially the increase in expression of vioB and vioD, may be due to switching from the canonical pathway of violacein synthesis (which is now completely inhibited) that uses VioA to convert L-tryptophan to indole-3-pyruvic acid, to the alternative pathway (Fig. 2) which uses VioD to convert L-tryptophan to 5-hydroxytryptophan in the absence of VioA.
Table 2 Phytochemical screening of manuka propolis pooled fractions Pool
TPC (mg GAE/g)a
PF1
204 ± 5.69
TFC (mg QE/g)a
HCAC (mg CGAE/g)a
16 ± 1.17
37 ± 1.87
Violacein inhibition zone (mm) at 100 lg per discb NA
PF2
775 ± 8.52
83 ± 0.82
102 ± 1.06
NA
PF3 PF4
458 ± 4.49 347 ± 1.55
52 ± 1.75 121 ± 1.05
94 ± 2.03 158 ± 1.03
NA 17.3 ± 1.2
PF5
366 ± 5.15
112 ± 0.59
324 ± 3.12
20.5 ± 2.2
PF6
375 ± 3.14
220 ± 0.84
142 ± 0.86
6.5 ± 0.2
PF7
373 ± 1.03
106 ± 1.61
114 ± 1.02
PF8
99 ± 4.00
116 ± 1.78
34 ± 0.33
10.5 ± 0.8 ND
Values for PF5 are in italics for emphasis TPC total phenolic content; TFC total flavonoid content; HCAC hydroxycinnamic acid content; NA not available; ND not detected a
Values of TPC, TFC and HCAC are mean ± SD (n = 3). Results are expressed in terms of per gram of fraction, where GAE is the gallic acid equivalent, QE is the quercetin equivalent, and CGAE is the chlorogenic acid equivalent b
Values of violacein inhibition zone are means of three biological replicate ± SD (n = 3)
123
Antonie van Leeuwenhoek
acid derivative with [M-H] = 247 (Fig. 4), which was identified as isoprenyl caffeate (IUPAC: 3-methylbut-2enyl (E)-3-(3,4-dihydroxyphenyl)prop-2-enoate) (Falca˜o et al. 2010).
Potential Vio protein-isoprenyl caffeate interactions
Fig. 2 Violacein biosynthetic pathway. Either VioA or VioD generates indole-3-pyruvic acid (IPA) imine from L-tryptophan, where VioB then converts the IPA imine into a dimer, which is transformed into prodeoxyviolacein only in the presence of VioE, after which VioC and VioD hydroxylate prodeoxyviolacein to form violacein (adapted from August et al. 2000; Ryan et al. 2008)
Manuka propolis PF5 consists mainly of hydroxycinnamic acids and flavonoids, with isoprenyl caffeate as the major component The highest HPLC peak was located at approximately 17.2 min, where a UV-spectrum analysis of this peak revealed that the major component belongs to the hydroxycinnamic acid group, which has a unique UVspectrum profile, with a shoulder at 290-300 nm and a peak at 320–330 nm (Gouveia and Castilho 2012). Liquid chromatography–mass spectrometry (LC–MS) analysis confirmed the presence of a hydroxycinnamic
123
Docking studies of isoprenyl caffeate-Vio protein complexes painted a possible explanation for the different responses of the vioA, D and E genes to treatment with different concentrations of manuka propolis PF5. The predicted binding energies between isoprenyl caffeate and the VioA, VioC, VioD and VioE enzymes are -108.76, -123.57, -148.29 and -133.40 kcal/mol, respectively. From the results, the most probable target of isoprenyl caffeate is VioD, with the strongest binding energy of -148.29 kcal/mol (Fig. 5). These results are in agreement with our experimental evidence, as inhibition of VioD would prevent the formation of the purple pigment (August et al. 2000). Analysis of the VioD binding site indicates that isoprenyl caffeate may competitively inhibit the binding of flavin adenine dinucleotide (FAD), which is required for the activation of this enzyme. The essential binding residues in the site are Ala8, Gly9, Glu32, Lys33, Asn34, Arg97, Arg105 and Asp269. Glu32, Arg97 and Arg105 contribute to hydrogen bonding of both ligands to the enzyme. The binding energy of isoprenyl caffeate (-148.29 kcal/mol) was lower compared to FAD (-800.235 kcal/mol). However, blocking of the same residues is enough to hinder the activity of the enzyme. Since VioA is the rate-limiting enzyme in the violacein biosynthesis pathway, it is important to show the interactions of isoprenyl caffeate, if any, with the enzyme. Because the three-dimensional structure of VioA is not available, docking was carried out with a modelled structure obtained using the ModBase modelling server. The template used by the server was L-glutamate oxidase from Streptomyces sp (2E1M.pdb). The only returned structure was the binding pocket of the oxidase (chain A) that had a sequence identity of 44 %. The binding pocket is the same as FAD (Fig. 6).The predicted binding energies of isoprenyl caffeate and FAD to the modelled part are -108 and -417 kcal/mol, respectively. The binding pocket residues include
Antonie van Leeuwenhoek
Fig. 3 Increasing concentrations of manuka propolis PF5 (0–500 lg/ml) inhibits violacein production of C. violaceum ATCC 12472. From left to right 0 (untreated control), 50, 100, 150, 200, 250, 300, 350, 400 and 450 lg/ml of manuka propolis PF5-treated cultures, with the 500 lg/ml treated culture (phenotypically similar to the 450 lg/ml treated culture) directly behind at the second row (a). Consolidated real-time PCR differential expression profile of the vioABCDE operon genes. Differential expression of each gene with 300, 350, 400,
450 and 500 lg/ml of manuka propolis PF5, relative to the untreated control, is presented as fold change, where no change in expression is expressed as 1. All fold changes have been normalized to rpoB as the reference gene. Data are means of fold changes with standard deviations from three independent experiments amplified in triplicates. An asterisk indicates significant differential expression ([± two-fold change compared to control) (b)
Gly65, Ala66, Gly67, Ile68, Glu88, Ala89, Arg97 and Gly96. Isoprenyl caffeate forms hydrogen bonds with Ala66, Glu88 and Arg97, while FAD binds to Ala69, Met123, Arg124, Met354 and Arg97. Thus, Arg97 is the common binding residue for the hydrogen bonding of isoprenyl caffeate and FAD to the binding pocket. Despite the limitations of the modelled VioA structure, combining the binding energy and docking interaction with our experimental results, it is proposed that VioA could also be a target for isoprenyl caffeate.
Isoprenyl caffeate is a quorum-sensing inhibitor A complete violacein inhibition zone of 6.22 ± 0.12 mm was achieved with at least 25 lg of isoprenyl caffeate (Fig. 1c), with no significant antibacterial effects as determined by plate count assay (data not shown). In liquid cultures, inhibition of C. violaceum violacein production was achieved with at least 50 lg/ml of commercial isoprenyl caffeate (data not shown). Thus, it is likely that isoprenyl caffeate, with violacein-inhibition properties and yet
123
Antonie van Leeuwenhoek Fig. 4 Representative ionogram of manuka propolis PF5 detected by Flexar SQ300. Isoprenyl caffeate (M-H = 247, Rt = 14.44 min) was identified as a major compound of manuka propolis PF5
Fig. 5 vioD docking studies. Crystal structure of FAD bound to vioD (a). Isoprenyl caffeate bound to the modelled pocket of VioD (b). LIGPLOT of FAD bound to VioD showing the binding site residues (c). LIGPLOT of VioD structure showing the important residues in the binding pocket (d). All common binding residues are encircled
123
Antonie van Leeuwenhoek Fig. 6 VioA docking studies. Crystal structure of L-glutamate oxidase from Streptomyces sp. with FAD in the binding pocket (a). Isoprenyl caffeate bound to the modelled pocket of VioA (b). LIGPLOT of the L-glutamate oxidase structure showing the important residues in the binding pocket (c). LIGPLOT of VioA structure showing the important residues in the binding pocket (d). The common binding residues are encircled
no significant antibacterial properties towards C. violaceum, is indeed a quorum-sensing inhibitor.
Discussion To be potential anti-pathogenic compounds, ideally QSI compounds should not interfere with normal bacteria growth, otherwise the bacteria has the tendency to generate resistance against these QSI compounds, rendering them as being no different from antibiotics (Rasmussen and Givskov 2006). In this study, out of the four types of bee products investigated, the two Malaysian bee products (honey and bee pollen) did not exhibit any violacein inhibitory properties. Manuka honey contained both violacein-
inhibition and antibacterial activities, as was also found by Wang et al. (2012), whereas manuka propolis only inhibited violacein production without antibacterial effects. The violacein inhibition property of manuka propolis was also stronger than that of manuka honey, thus manuka propolis was selected for further investigation in this study. Interestingly in the present study, manuka honey was identified to have a low content of phenolic compounds, which are known to exhibit QSI activity. It is possible that the violacein inhibition and bactericidal effect of manuka honey may be attributed to its high sugar content, as seen in a previous study by Lee et al. (2011), which proposed that the QSI activity of diluted honey against E. coli O157:H7 was through a mechanism of action where a high sugar concentration
123
Antonie van Leeuwenhoek
may prevent the E. coli acyl homoserine lactones (AHL) from re-entering the cells; therefore blocking AHL-receptor complex formation. Manuka propolis fraction 5 (PF5) exhibited the strongest violacein inhibition, followed by PF4, where both fractions were observed to contain high amounts of phenolics, flavonoids and, particularly, hydroxycinnamic acid. Previous studies investigating the QSI properties of various natural sources suggest that phenolic compounds, such as phenolic acids and flavonoids, might be the active component responsible (Annapoorani et al. 2012; Lee et al. 2011; Singh et al. 2009). Other PFs, with low hydroxycinnamic acid content, showed weak or no violacein inhibition properties. Thus, the active QSI compound(s) in manuka propolis is suggested to belong to the hydroxycinnamic acid group, particularly isoprenyl caffeate. This suggestion is supported by the work of Brackman et al. (2008; 2011), which found cinnamaldehyde (a derivative of the hydroxycinnamic acid group) and cinnamaldehyde derivatives to possess QSI effects on Vibrio spp. as a QS model organism. It would seem that this family of compounds are active QSI compounds, regardless of the aldehyde or carboxylic acid forms (Brackman et al. 2008; 2011). The minimum concentration of manuka propolis PF5 required for phenotypically observable QSI effects in C. violaceum is 300 lg/ml. This concentration is much less than the 28 mg/ml of Filipino propolis ethanolic extract required for violacein inhibition in C. violaceum ATCC 12472 (Lamberte et al. 2011) and the 10 mg/ml of Spanish propolis tincture required for violacein inhibition in C. violaceum CV026 (Bulman et al. 2011), a strain inherently unable to produce the signaling molecule N-hexanoyl homoserine lactone (C6-HSL). In these previous studies, both types of propolis were not subjected to further fractionation and purification. In this study, manuka propolis, a single-source propolis, was fractionated and subjected to further purification, where we identified a major compound— isoprenyl caffeate—as the likely QSI compound present in manuka propolis. Complete violacein inhibition in C. violaceum liquid cultures was achieved with a minimum concentration of 50 lg/ml of commercial isoprenyl caffeate, thus confirming it as a QSI compound. The molecular observations of the manuka propolis PF5 dose–response gene expression experiments correlate with our in silico molecular docking experiments for isoprenyl caffeate, the postulated QSI compound in
123
manuka propolis PF5. Manuka propolis PF5 treatment of 300 lg/ml began down-regulating vioD and vioE, with isoprenyl caffeate showing the highest binding energy with VioD, followed by VioE. This suggests that the way which isoprenyl caffeate can inhibit the violacein biosynthetic pathway is by competitively inhibiting the binding of FAD to the VioD enzyme, thus deactivating the enzyme and preventing the hydroxylation of prodeoxyviolacein to the purple pigment violacein. In turn, less VioE is required. Subsequently, the higher concentration of manuka propolis PF5 required to significantly down-regulate vioA (450 lg/ml) compared to vioD and vioE is also evident in the docking results, where a comparatively low binding energy was obtained for the binding of isoprenyl caffeate to the VioA enzyme. As VioA is the rate-limiting enzyme that is responsible for the introduction of oxygen into the first intermediate product of the pathway, indole 3-pyruvic acid, deactivation of VioA through the competitive inhibition of FAD-binding by isoprenyl caffeate may have caused the first substrate to be incorrectly formed, resulting in the inhibition of the violacein biosynthetic pathway.
Conclusions This is the first study to elucidate isoprenyl caffeate, a major compound in manuka propolis, as a potential QSI compound, where a commercial isoprenyl caffeate pure compound was confirmed to exhibit QSI properties on C. violaceum solid and liquid cultures without significant antibacterial effects. By correlating the phenotypic observation of violacein inhibition to the molecular responses of the vioABCDE operon in C. violaceum, it was observed that violacein inhibition was achieved in tandem with the down-regulation of vioA, vioD and vioE. Docking studies suggest that isoprenyl caffeate acts as a competitor of the FADbinding pockets of VioA and VioD. Acknowledgments This work was done with funding from UCSI University’s Faculty of Applied Sciences and the MAKNA Cancer Research Award 2012. We thank Prof. Tom Coenye for his constructive comments of the early findings of this study and helpful insights into the QSI activity of cinnamaldehyde and cinnamaldehyde derivatives. We also thank the Drug Design & Development Research Group (DDDRG) of the University of Malaya for use of the docking software.
Antonie van Leeuwenhoek Conflict of interest The authors declare that they have no conflict of interest.
References Annapoorani A, Parameswari R, Pandian SK, Ravi AV (2012) Methods to determine antipathogenic potential of phenolic and flavonoid compounds against urinary pathogen Serratia marcescens. J Microbiol Methods 91:208–211 August PR, Grossman TH, Minor C, Draper MP, MacNeil IA, Pemberton JM, Call KM, Holt D, Osburne MS (2000) Sequence analysis and functional characterization of the violacein biosynthetic pathway from Chromobacterium violaceum. J Mol Microbiol Biotechnol 2:513–519 Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, Bourne PE (2000) The protein data bank. Nucleic Acids Res 28:235–242 Brackman G, Defoirdt T, Miyamoto C, Bossier P, Van CS, Nelis H, Coenye T (2008) Cinnamaldehyde and cinnamaldehyde derivatives reduce virulence in Vibrio spp. by decreasing the DNA-binding activity of the quorum sensing response regulator LuxR. BMC Microbiol 8:149 Brackman G, Celen S, Hillaert U, Van Calenbergh S, Cos P, Maes L, Nelis HJ, Coenye T (2011) Structure-activity relationship of cinnamaldehyde analogs as inhibitors of AI2 based quorum sensing and their effect on virulence of Vibrio spp. PLoS One 6:e16084 Brazilian National Project Consortium (2003) The complete genome sequence of Chromobacterium violaceum reveals remarkable and exploitable bacterial adaptability. Proc Natl Acad Sci USA 100:11660–11665 Bulman Z, Lee P, Hudson AO, Savka MA (2011) A novel property of propolis (bee glue): anti-pathogenic activity by inhibition of N-acyl-homoserine lactone mediated signaling in bacteria. J Ethnopharmacol 138:788–797 Falca˜o SI, Vilas-Boas M, Estevinho LM, Barros C, Domingues MR, Cardoso SM (2010) Phenolic characterization of Northeast Portuguese propolis: usual and unusual compounds. Anal Bioanal Chem 396:887–897 Gouveia SC, Castilho PC (2012) Validation of a HPLC-DAD– ESI/MSn method for caffeoylquinic acids separation, quantification and identification in medicinal Helichrysum species from Macaronesia. Food Res Int 45:362–368 Hentzer M, Givskov M (2003) Pharmacological inhibition of quorum sensing for the treatment of chronic bacterial infections. J Clin Invest 112:1300–1307 Hentzer M, Wu H, Andersen JB, Riedel K, Rasmussen TB, Bagge N, Kumar N, Schembri MA, Song Z, Kristoffersen P, Manefield M, Costerton JW, Molin S, Eberl L, Steinberg P, Kjelleberg S, Høiby N, Givskov M (2003) Attenuation of Pseudomonas aeruginosa virulence by quorum sensing inhibitors. EMBO J 22:3803–3815 Lamberte LE, Cabrera EC, Rivera WL (2011) Activity of the ethanolic extract of propolis (EEP) as a potential inhibitor of quorum sensing-mediated pigment production in Chromobacterium violaceum and virulence factor production in Pseudomonas aeruginosa. Philipp Agric Sci 94:14–22
Lee JH, Park JH, Kim JA, Neupane GP, Cho MH, Lee CS, Lee J (2011) Low concentrations of honey reduce biofilm formation, quorum sensing, and virulence in Escherichia coli O157:H7. Biofouling 27:1095–1104 Martinelli D, Grossmann G, Se´quin U, Brandl H, Bachofen M (2004) Effects of natural and chemically synthesized furanones on quorum sensing in Chromobacterium violaceum. BMC Microbiol 4:25 Pfaffl MW (2001) A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 29:e45 Rasmussen TB, Givskov M (2006) Quorum-sensing inhibitors as anti-pathogenic drugs. Int J Med Microbiol 296:149–161 Rasmussen TB, Bjarnsholt T, Skindersoe ME, Hentzer M, Kristoffersen P, Ko¨te M, Nielsen J, Eberl L, Givskov M (2005) Screening for quorum-sensing inhibitors (QSI) by use of a novel genetic system, the QSI Selector. J Bacteriol 187:1799–1814 Ryan KS, Balibar CJ, Turo KE, Walsh CT, Drennan CL (2008) The violacein biosynthetic enzyme VioE shares a fold with lipoprotein transporter proteins. J Biol Chem 283:6467–6475 Schwede T, Kopp J, Guex N, Peitsch M (2003) SWISSMODEL: an automated protein homology-modeling server. Nucleic Acids Res 31:3381–3385 Singh BN, Singh BR, Singh RL, Prakash D, Dhakarey R, Upadhyay G, Singh HB (2009) Oxidative DNA damage protective activity, antioxidant and anti-quorum sensing potentials of Moringa oleifera. Food Chem Toxicol 47:1109–1116 Slama TG (2008) Gram-negative antibiotic resistance: there is a price to pay. Crit Care 12(Suppl 10):1–7 Souli M, Galani I, Giamarellou H (2008) Emergence of extensively drug-resistant and pandrug-resistant Gram-negative bacilli in Europe. Euro Surveill 13:19045 Stauff DL, Bassler BL (2011) Quorum sensing in Chromobacterium violaceum: DNA recognition and gene regulation by the CviR receptor. J Bacteriol 193:3871–3878 Taganna JC, Rivera WL (2008) Epigallocatechin gallate from Camellia sinensis L. (Kuntze) is a potential quorum sensing inhibitor in Chromobacterium violaceum. Sci Diliman 20:24–30 Taganna JC, Quanico JP, Perono RM, Amor EC, Rivera WL (2011) Tannin-rich fraction from Terminalia catappa inhibits quorum sensing (QS) in Chromobacterium violaceum and the QS-controlled biofilm maturation and LasA staphylolytic activity in Pseudomonas aeruginosa. J Ethnopharmacol 134:865–871 Wagner VE, Gillis RJ, Iglewski BH (2004) Transcriptome analysis of quorum-sensing regulation and virulence factor expression in Pseudomonas aeruginosa. Vaccine 22(Suppl 1):S15–S20 Wallace AC, Laskowski RA, Thornton JM (1995) LIGPLOT: a program to generate schematic diagrams of protein-ligand interactions. Protein Eng 8:127–134 Wang R, Starkey M, Hazan R, Rahme LG (2012) Honey’s ability to counter bacterial infections arises from both bactericidal compounds and QS inhibition. Front Microbiol 2012:144
123
Antonie van Leeuwenhoek Waters CM, Bassler BL (2005) Quorum sensing: cell-to-cell communication in bacteria. Annu Rev Cell Dev Biol 21:319–346 Winser K, Williams P (2001) Quorum sensing and the regulation of virulence gene expression in pathogenic bacteria. Int J Med Microbiol 291:131–143
123
Wu G, Robertson DH, Brooks CL 3rd, Vieth M (2003) Detailed analysis of grid-based molecular docking: a case study of CDOCKER-A CHARMm-based MD docking algorithm. J Comput Chem 24:1549–1562
