Quorum-Sensing Regulation of Constitutive Plantaricin by Lactobacillus plantarum Strains under a Model System for Vegetables and Fruits

Abstract

This study aimed at investigating the regulatory system of bacteriocin synthesis by Lactobacillus plantarum strains in vegetables and fruits in a model system. Sterile and neutralized cell-free supernatant (CFS) from L. plantarum strains grown in MRS broth showed in vitro antimicrobial activities toward various indicator strains. The highest activity was that of L. plantarum C2. The antimicrobial activity was further assayed on vegetable and fruit agar plates (solid conditions) and in juices (liquid conditions). A regulatory mechanism of bacteriocin synthesis via quorum sensing was hypothesized. The synthesis of antimicrobial compounds seemed to be constitutive under solid conditions of growth on vegetable and fruit agar plates. In contrast, it depended on the size of the inoculum when L. plantarum C2 was grown in carrot juice. Only the inoculum of ca. 9.0 log CFU ml⁻¹ produced detectable activity. The genes plnA, plnEF, plnG, and plnH were found in all L. plantarum strains. The genes plnJK and plnN were detected in only three or four strains. Reverse-phase high-performance liquid chromatography purification and mass spectrometry analysis revealed the presence of a mixture of eight peptides in the most active fraction of the CFS from L. plantarum C2. Active peptides were encrypted into bacteriocin precursors, such as plantaricins PlnJ/K and PlnH and PlnG, which are involved in the ABC transport system. A real-time PCR assay showed an increase in the expression of plnJK and plnG during growth of L. plantarum C2 in carrot juice.

INTRODUCTION

Bacteria synthesize antimicrobial compounds, in particular peptides, to compete within complex communities. Depending on the mechanism of biosynthesis, two types of antimicrobial peptides are found: ribosomally synthesized peptides, which exhibit a relatively narrow spectrum of activity and mainly inhibit closely related bacteria that share the same ecological niche, and non-ribosomally synthesized peptides, some of which show broader spectra of activities and affect both bacteria and fungi (1).

Lactic acid bacteria synthesize a wide range of bacteriocins (2–4), which has been of interest because of their potential use as natural food preservatives (5, 6). Based on the biochemical and genetic properties, bacteriocins are grouped into four classes (I to IV) (7, 8). Both class I and II bacteriocins are small (3 to 10 kDa), cationic, amphiphilic, and membrane-active peptides. Class I bacteriocins, or lantibiotics, contain the unusual amino acids lanthionine and methyllanthionine. Class II bacteriocins are synthesized as precursor molecules which, in most of the cases, contain a leader peptide or the so-called double-glycine type (7–10). This leader peptide is recognized and cleaved by a dedicated ABC transporter, which results in the translocation of the mature and active bacteriocin into the surrounding medium (11). The class II bacteriocins are subdivided into three subgroups: IIa, Listeria-active peptides with the consensus sequence YGNGVXC near the N terminus; IIb, which includes bacteriocins whose antimicrobial activities depend on the complementary activity of two different peptides (7, 9); IIc, thiol-activated peptides that require reduced cysteine residues for activity. Class III bacteriocins are high-molecular-mass (>30 kDa), heat-labile proteins. Class IV bacteriocins are complex peptides that contain lipid or carbohydrate moieties, which are essential for activity.

Overall, the biosynthesis of bacteriocins is a high-energy-consuming process, and it provides advantages to the producer strain only if the cost/benefit ratio is favorable (12, 13). Consequently, the synthesis of bacteriocins by lactic acid bacteria is controlled by sophisticated molecular regulatory systems (e.g., induction and catabolic repression). For instance, several class II bacteriocins are regulated by a three-component regulatory system (7, 9, 14–16). The synthesis of bacteriocin may be an unstable phenotype (17–19), and this instability is attributed to decreased levels of the inducing factor under certain environmental conditions (e.g., temperature, cell density) (14, 16). The synthesis of bacteriocins (e.g., plantaricin C) may also be repressed by high concentrations of glucose, which acts as a mediator of catabolic repression (20).

Lactic acid bacteria have to adapt their metabolism to changing and, in particular, stressful environmental conditions. Raw fruits and vegetables are prone to environmental changes and microbial interactions. Adaptation requires that bacteria sense the multitude of extracellular signals and respond by controlling the expression of an adequate repertoire of genes. Under these conditions, the synthesis of bacteriocins has a unique role in the ecology of lactic acid bacteria and represents an efficient tool to control the spoilage and poisoning of raw vegetables and fruits (21). Lactobacillus plantarum is a highly heterogeneous and versatile species (22) that is widely found in fermented vegetables and fruits (23). Most of its regulatory systems are strictly related to detection of specific environmental signals (24), competition with other bacteria, and survival under a variety of conditions. Numerous bacteriocins have been described for L. plantarum (25–27). The bacteriocin pln locus contains five operons. plnEFI and plnJKLR encode bacteriocins and immunity proteins; plnGHSTUV encodes the ABC transport system to secrete peptides, which contains double-glycine N-terminal leaders; plnABCD encodes peptides for the signal-transducing pathway; the last operon, plnMNOP, harbors genes with unknown functions. The PlnA peptide induces the transcription of the above five operons (28). All pln operons are repressed under nonproducing conditions (29). Environmental factors such as chemical compounds from plant materials, storage conditions, and the presence of a dense microbial population (quorum) affect the accumulation of bacteriocins (30). To the best of our knowledge, the synthesis of plantaricins and the regulatory mechanism have not been investigated under conditions similar to those present in foods. Recently, the metabolic and proteomic profiles of L. plantarum during vegetable and fruit fermentations were studied, and new insights into its response under hostile environmental conditions were reported (31; P. Filannino, G. Cardinali, C. G. Rizzello, S. Buchin, M. De Angelis, M. Gobbetti, R. Di Cagno, submitted for publication).

This study aimed at investigating the antimicrobial activity of several strains of L. plantarum in a vegetables and fruits model system. The regulatory system and the environmental determinants that affected the bacteriocin phenotype were investigated. Bacteriocins were purified and identified through nano-liquid chromatography–electrospray ionization–mass spectrometry, and the expression levels of genes that are responsible for the synthesis were estimated.

MATERIALS AND METHODS

Microorganisms and growth conditions.

Six strains of Lactobacillus plantarum were assayed for antimicrobial activity (Table 1). Previously, all strains had been selected as potential probiotic candidates (36). Propagation was routinely carried out in MRS broth (Oxoid) at 30°C for 24 h under anaerobiosis. Bacterial strains used as indicators in this study and growth conditions are shown in Table 1.

TABLE 1.

Bacterial strains used in this study

Purpose for use and strain name	Source	Culture collection	Growth conditions	Reference
Potential bacteriocin producing strains
L. plantarum C2	Carrot	DISSPAa	MRS broth at 30°C	32
L. plantarum C5	Carrot	DISSPA	MRS broth at 30°C	32
L. plantarum POM20	Tomato	DISSPA	MRS broth at 30°C	33
L. plantarum POM42	Tomato	DISSPA	MRS broth at 30°C	33
L. plantarum 1LS9	Pineapple	DISSPA	MRS broth at 30°C	34
L. plantarum 1LS16	Pineapple	DISSPA	MRS broth at 30°C	34
Indicator strains
Escherichia coli DSM 30083	Urine	DSMZb	Luria-Bertani broth at 37°C	DSMZ
Enterobacter aerogenes DSM 30053	Sputum	DSMZ	Luria-Bertani broth at 37°C	DSMZ
Enterococcus durans DSM 20633	Dried milk	DSMZ	Brain heart infusion medium at 37°C	DSMZ
Yersinia enterocolitica DSM 4780	Glanders-like infection of face	DSMZ	Brain heart infusion medium at 37°C	DSMZ
Weissella confusa DSM 20196	Sugar cane	DSMZ	MRS broth at 30°C	DSMZ
Leuconostoc lactis DSM 20202	Milk	DSMZ	MRS broth at 30°C	DSMZ
Clostridium coccoides DSM 935	Mouse feces	DSMZ	Reinforced clostridial medium at 37°C, anaerobic	DSMZ
Lactobacillus sakei SAL1	Sourdough	DISSPA	MRS broth at 30°C	35
Bacillus megaterium F6	Fresh vegetables	DISSPA	Luria-Bertani broth at 37°C	36

^aDISSPA, Department of Soil, Plant and Food Sciences (University of Bari, Bari, Italy).

^bDSMZ, Leibniz Institute German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany).

Preliminary screening.

Twenty-four-hour-old cells of L. plantarum strains grown in MRS broth were harvested by centrifugation (10,000 × g, 10 min at 4°C), washed twice in 50 mM sterile potassium phosphate buffer (pH 7.0), resuspended in sterile distilled water to a final optical density at 620 nm (OD₆₂₀) of 2.5 (corresponding to ca. 9.0 log CFU ml⁻¹), and used to inoculate fresh MRS broth to 7.0 log CFU ml⁻¹. After incubation for 24 h at 30°C, cell-free supernatants (CFS) were recovered by centrifugation (10,000 × g, 10 min at 4°C) and used to assay antimicrobial activity in a well diffusion assay (37). Analyses were carried out on 15 ml of agar-H₂O (2%, wt/vol) overlaid with 5 ml of different soft agar media, which contained ca. 4.0 log CFU ml⁻¹ of an overnight culture of the indicator strains (Table 1). Wells (5 mm in diameter) were cut into agar plates, and 100 μl of neutralized (1 N NaOH) and sterile CFS was added. Plates were stored for 1 h at 4°C to permit the radial diffusion of CFS, and incubation was allowed at 30 or 37°C for 24 h.

Characterization of antimicrobial activity.

CFS were treated with trypsin (EC 3.4.21.4; Sigma-Aldrich Co.), as described by Atanassova et al. (38). Heat stability of CFS was determined by heating at 100°C for 5 min. After treatments, the residual activity of CFS was checked through the agar diffusion assay. To exclude the effect of hydrogen peroxide, CFS were also treated with catalase (1 mg ml⁻¹; Sigma).

Antimicrobial activity in the vegetables and fruits model system.

Pineapple, carrot, and tomato juice media were prepared according to the methods described by Di Cagno et al. (32). Vegetables and fruits were separately homogenized, centrifuged (10,000 × g, 20 min at 4°C), heated (121°C for 10 min), filtered through a Whatman apparatus (Polycarp 75 SPF; Whatman International, Maidstone, England) through a Durapore 0.22-μm-pore-size filter (Millipore Co., Bedford, MA), and stored at −20°C before use. L. plantarum strains were grown in vegetable and fruit juices that corresponded to the source from which they were previously isolated. Twenty-four-hour-old cells were harvested by centrifugation (10,000 × g, 10 min at 4°C) and washed twice in 50 mM sterile potassium phosphate buffer (pH 7.0). Harvested cells were used to assay the antimicrobial activities under solid and liquid conditions.

To determine the antimicrobial activities under solid conditions, vegetable and fruit agar plates were prepared according to the method described by Jacxsens et al. (39). Carrots and tomatoes were washed with cold tap water (for ca. 60 s), while outside spirals were removed from pineapple fruits. All raw matrices were cut with a kitchen knife and blended by using a food processor (type 4290; Braun, Kronberg, Germany). Blends were centrifuged (16,270 × g, 20 min), and supernatants were heated for 2 h at 80°C to denature enzymes and proteins and then filtered through a Whatman apparatus (Polycarp 75 SPF; Whatman International) by using Durapore 0.22-μm-pore-size filter (Millipore Co.). Finally, extracts were mixed with 1.5% (wt/vol) agar and autoclaved for 15 min at 121°C. The overlay technique described by Sathe et al. (40) was used. Briefly, L. plantarum strains were grown in the form of 2-cm streaks on vegetable and fruit agar plates and incubated at 30°C for 48 h. Further, plates were overlaid with agar medium inoculated with the indicator strains (cell density of ca. 6.0 log CFU ml⁻¹) and incubated at 30°C for 24 to 72 h. After incubation, clear zones of inhibition were measured.

To determine the antimicrobial activities under liquid conditions, 24-h-old cells of L. plantarum strains cultivated on vegetable and fruit juices (ca. 7.0 to 9.0 log CFU ml⁻¹) were used to inoculate fresh vegetable and fruit juices at 30°C for 24 h. CFS were recovered, and antimicrobial activity was assayed in a well diffusion assay. Freeze-dried concentrated (10×) CFS were also assayed. Bacillus megaterium F6 was the indicator strain used for these assays.

The induction of the antimicrobial activity of L. plantarum C2 was determined according to the method of Rojo-Bezares et al. (41). Twenty microliters of a 24-h-old culture of L. plantarum C2 (inoculum of ca. 7.0 and 8.0 log CFU ml⁻¹) grown on carrot juice was mixed with 50 μl of CFS of a 24-h-old culture of the same strain (inoculum of ca. 9.0 log CFU ml⁻¹) and 930 μl of fresh carrot juice. After incubation at 30°C for 24 h, the mixture was centrifuged and the supernatant was assayed for antibacterial activity by using the agar well diffusion method.

Purification and identification of antimicrobial compounds.

CFS were fractionated by ultrafiltration (Ultrafree-MC centrifugal filter units; Millipore), using four different membranes with cutoffs of 50, 30, 10, and 5 kDa. Aliquots of 400 μl were centrifuged at 10,000 × g for 60 min. After ultrafiltration, fractions were used for the agar diffusion assay. The active fractions were further analyzed and partially purified by reversed-phase high-performance liquid chromatography (RP-HPLC), using an XTerra MS C₁₈ 5-μm, 4.6- by 250-mm column (Waters, Brussels, Belgium) and an Äkta purifier HPLC, equipped with a UV-900 detector (GE Healthcare Bio-Sciences Corp., Piscataway, NJ) at 214 nm. Aliquots (100 μl) of the CFS were added to 0.05% (vol/vol) trifluoroacetic acid (TFA) and centrifuged at 10,000 × g for 10 min. The supernatant was filtered through a 0.22-μm-pore-size filter and loaded onto the XTerra MS C₁₈ column. The gradient elution was at a flow rate of 0.8 ml min⁻¹, at room temperature, using a mobile phase composed of water and acetonitrile (CH₃CN) and containing 0.05% TFA. The CH₃CN concentration was increased linearly from 5 to 46% between 16 and 62 min and from 46 to 100% between 62 and 72 min. Thirty fractions (2 ml) were recovered by using a FRAC 920 automatic fraction collector (GE Healthcare). The antimicrobial activity after each step of purification was assayed by using the agar well diffusion method as described above. Fraction 3, with the highest antimicrobial activity, was freeze-dried to remove solvents, redissolved in sterile water, and subjected to mass spectra analysis by nano-LC–electrospray ionization-mass spectrometry (nano-LC-ESI-MS). The Finningan LCQ Deca XP Max ion trap mass spectrometer (ThermoElectron, San Jose, CA), equipped with a nano-ESI interface, was used. According to the manufacturer's instruction, MS spectra were automatically collected by the Xcalibur software (ThermoElectron) in positive ion mode. Total ion current (m/z range, 50 to 2,000) and selected ion monitoring spectra were recorded and processed with the software BioWorks 3.2 (ThermoElectron). The concentration of peptides in CFS and purified fractions was determined by the o-phthaldialdehyde (OPA) method (42).

DNA isolation, PCR conditions, and sequencing of amplification products.

Total DNA was obtained according to the method of De Los Reyes-Gavilán et al. (43). Table 2 shows the set of specific primers used to amplify genes that are involved in the plantaricin locus. Fifty microliters of each PCR mixture contained the following: 200 μM each deoxynucleoside triphosphates, 1 μM both primers, 2 mM MgCl₂, 2 U of Taq DNA polymerase (Promega Corporation, Madison, WI) in the supplied 1× buffer, and ca. 50 ng of template DNA. The T100 thermal cycler (Bio-Rad, Hercules, CA) was used under the following conditions: denaturation at 94°C for 2 min, followed by 30 cycles of denaturation at 94°C for 15 s, annealing at 60°C for 30 s, and polymerization at 72°C for 2 min, plus a final polymerization step at 72°C for 4 min. PCR products were separated through electrophoresis on a 1.5% (wt/vol) agarose gel (Gibco BRL, France) and stained with a fluorescent nucleic acid gel stain (Biotium Inc., Hayward, CA). Reactions of DNA sequencing were performed by Eurofins MWG GmbH (Ebersberg, Germany). Sequence comparison was carried out using the Basic BLAST database. Translation of the nucleotide sequence was performed using the Omiga software (Oxford Molecular, Madison, WI) or ExPASy translation routine on the ExPASy Molecular Biology Server of the Swiss Institute of Bioinformatics (http://ca.expasy.org/). Similarity searches were carried out with the advanced BLAST algorithm available at the National Center for Biotechnology Information site (htpp://www.ncbi.nlm.nih.gov/).

TABLE 2.

Primers used for detection of genes that belong to the pln locus

Target gene	Codified bacteriocin	Target microorganism(s) or function	Forward primer (5′–3′)	Reverse primer (5′–3′)	Annealing temp (°C)	Amplicon size (bp)	Reference(s)
plnA	Plantaricin A	Pediococcus pentosaceus, Lactobacillus plantarum, Lactobacillus casei, Lactobacillus sakei, Lactobacillus viridescens, Carnobacterium piscicola	GTACAGTACTAATGGGAG	CTTACGCCATCTATACG	53.0	450	27, 74
plnH	Plantaricin H	Accessory factor for PlnG	TCTTACACGATCAAGGCAAC	TGTGCCATTACTTACCTGTTTC	53.4	526	27
plnG	Plantaricin G	ABC transporter, plantaricin A system	TGCGGTTATCAGTATGTCAAAG	CCTCGAAACAATTTCCCCC	52.8	453	25, 27
plnEF	Plantaricins E/F	P. pentosaceus, L. plantarum, L. casei, L. sakei, L. viridescens, C. piscicola, P. acidilactici, Lactobacillus curvatus, Micrococcus luteus, Listeria innocua, Enterococcus casseliflavus	GGCATAGTTAAAATTCCCCCC	CAGGTTGCCGCAAAAAAAG	53.2	428	70, 75
plnJ	Plantaricins J/K (peptide J)	P. pentosaceus, L. plantarum, L. casei, L. sakei, L. viridescens, C. piscicola, P. acidilactici, L. curvatus, M. luteus, L. innocua, E. casseliflavus	TAACGACGGATTGCTCTG	AATCAAGGAATTATCACATTAGTC	51.0	475	70, 75
plnK	Plantaricins J/K (peptide K)	P. pentosaceus, L. plantarum, L. casei, L. sakei, L. viridescens, C. piscicola, P. acidilactici, L, curvatus, M. luteus, L. innocua, E. casseliflavus	AATCGCAGTGACTTCCAGAAC	AGAGCAATCCGTCGTTAATAAATG	53.7	469	70, 75
plnS	Plantaricin S	L. plantarum, Lactobacillus delbrueckii, L. fernentum, Lactococcus lactis subsp. cremoris, L. lactis subsp. lactis, Leuconostoc mesenteroides subsp. dextranicum, L. mesenteroides subsp. mesenteroides, L. paramesenteroides, P. pentosaceus, Streptococcus thermophilus, Clostridium tyrobutyricum, E. faecalis, L. curvatus, L. helveticus, L. sakei	GCCTTACCAGCGTAATGCCC	CTGGTGATGCAATCGTTAGTTT	60.0	320	55, 67
plnW	Plantaricin W	L. lactis, Oenococcus oenos, L. mesenteroides, P. acidilactici, P. pentosaceus, E. faecalis, L. innocua, L. monocytogenes, Propionibacterium freudenreichii, Staphylococcus aureus	TCACACGAAATATTCCA	GGCAAGCGTAAGAAATAAATGAG	55.0	165	68
plnN	Plantaricin N	Function unknown	ATTGCCGGGTTAGGTATCG	CCTAAACCATGCCATGCAC	51.9	146	27
plnNC8βα	Plantaricin NC8	L. plantarum	GGTCTGCGTATAAGCATCGC	AAATTGAACATATGGGTGCTTTAAATTCC	60.0	207	26

RNA isolation and transcript analysis by quantitative real-time PCR.

Cells of L. plantarum C2 inoculated on carrot juice to ca. 7 and 9 log CFU ml⁻¹ and incubated at 30°C for 24 h were harvested (after different times: 0, 2, 4, 8, 12, 15, 19, and 24 h) and centrifuged at 9,000 × g for 10 min at 4°C. Cell pellets were frozen at −80°C until use for RNA extraction. Total RNA isolation from the same amount of cells (ca. 1 × 10⁹) was carried out by using the RNeasy plant minikit, as recommended by the manufacturer (Qiagen, Hilden, Germany). Quality control of RNA was checked through agarose gel electrophoresis. The RNA concentration was measured in a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Rockland, DE). cDNA synthesis (500 ng of RNA resuspended into a volume of 20 μl), from starting RNA samples, was performed using the QuantiTect reverse transcription kit (Qiagen) and random hexamers for priming according to the protocol supplied for RT-PCR amplification. The effective elimination of genomic DNA contamination was carried out by using the gDNA Wipeout buffer (7×; Qiagen). All reactions were set up in a Rotor Gene 6000 instrument (Corbett Life Science, New South Wales, Australia) equipped with a 36-well reaction rotor and with the Rotor-Gene SYBR green PCR kit (Qiagen), according to the protocol supplied. The reaction mixture (25 μl) contained 12.5 μl of SYBR green PCR master mix (Qiagen), 1 to 4 μl of the cDNA sample, and appropriate (1 μM) PCR primers (Table 3). The assays were carried out in triplicate. PCR required an initial denaturation at 95°C for 5 min, followed by a 40-cycle amplification consisting of denaturation at 95°C for 5 s and annealing and extension at 60°C for 10 s. After the last cycle of each amplification, a melt curve analysis, with a temperature range of 60°C to 95°C ramping at 1°C/5 s, was performed to determine the product specificity. RT-PCR amplification products were further loaded on agarose 2% gels to check the correspondence to amplicon size. Data were normalized to levels of the 16S rRNA housekeeping gene and analyzed using a comparative cycle threshold method (ΔΔC_T). The levels of expression of genes were compared using the relative quantification method (44). Real-time data are presented as the relative change compared to L. plantarum C2 incubated for 2 h (control). Error bars show the standard deviations (SD) of the ΔΔC_T value (44, 45).

TABLE 3.

RT-PCR primer sequences used to determine plantaricin gene expression

Target gene	Codified bacteriocin	Forward primer (5′–3′)	Reverse primer (5′–3′)	Amplicon size (bp)	Reference
plnA	Plantaricin A	AAAATTCAAATTAAAGGTATGAAGCAA	CCCCATCTGCAAAGAATACG	108	66
plnB	Plantaricin B	CCAGAGCGGAGAATTCAGAG	GTACAACTAACCCGCGCAAT	92	66
plnC	Plantaricin C	AGCAGATGAAATTCGGCAG	ATAATCCAACGGTGCAATCC	108	25
plnEF	Plantaricins E/F	GTTTTAATCGGGGCGGTTAT	ATACCACGAATGCCTGCAAC	85	66
plnJ	Plantaricins J/K (peptide J)	TAAGTTGAACGGGGTTGTTG	TAACGACGGATTGCTCTGC	102	66
plnK	Plantaricins J/K (peptide K)	TTCTGGTAACCGTCGGAGTC	ATCCCTTGAACCACCAAGC	97	66
plnN	Plantaricin N	GCCGGGTTAGGTATCGAAAT	TCCCAGCAATGTAAGGCTCT	102	66
plnG	Plantaricin G	TGCGGTTATCAGTATGTCAAAG	CCTCGAAACAATTTCCCCC	453	41

Statistical analyses.

Data (at least three replicates) were subjected to one-way analysis of variance, and pairwise comparison of treatment means was achieved by using Tukey's procedure at a P level of <0.05, utilizing the statistical software Statistica 7.0 for Windows.

RESULTS

Preliminary screening.

After 24 h of cultivation in MRS broth, the sterile and neutralized CFS from six Lactobacillus plantarum strains were assayed for in vitro antimicrobial activities (in well diffusion assays). Table 4 shows the inhibition activities toward various indicator strains. The latter were chosen since their occurrence in raw vegetables and fruits is likely the outcome of environmental contamination. Escherichia coli is commonly present in freshwater environments (46), and Enterobacter aerogenes, Enterococcus durans, and Yersinia enterolitica are often involved in food spoilage, including plant materials (47, 48). Although Clostridium coccoides is considered among the most predominant bacterial groups in human gut microbiota (49), it may be involved in canned vegetable spoilage (50). Bacillus megaterium is a common soil bacterium and a plant endophyte (51). Weissella, Leuconostoc, and Lactobacillus are some of the genera of lactic acid bacteria that most commonly colonize vegetables and fruits (23).

TABLE 4.

Inhibitory spectrum of CFS from six Lactobacillus plantarum strains grown in MRS for 24 h at 30°C and evaluated in the well diffusion assay

Indicator microorganism	Inhibitory activity of CFS from straina
	L. plantarum C2	L. plantarum C5	L. plantarum POM20	L. plantarum POM42	L. plantarum 1LS9	L. plantarum 1LS16
Escherichia coli DSM 30083	++	+++	++	+	++	++++
Enterobacter aerogenes DSM 30053	++++	+++	+++	++	++	++
Enterococcus durans DSM 20633	++	−	−	−	+	−
Yersinia enterocolitica DSM 4780	++	−	−	−	−	+
Weissella confusa DSM 20196	−	−	−	−	−	−
Leuconostoc lactis DSM 20202	+++	+	+	++	+	+
Clostridium coccoides DSM 935	−	−	−	−	−	−
Lactobacillus sakei SAL1	−	−	−	−	−	−
Bacillus megaterium F6	+++	+++	+	++	+	++

^aInhibitory activity was scored as follows: −, no inhibition; +, halo of inhibition diameter of 0.5 to 4 mm; ++, halo of inhibition diameter of 4 to 6.0 mm; +++, halo of inhibition diameter of 6.0 to 7.5 mm; ++++, halo of inhibition diameter of >7.5 mm.

None of the L. plantarum strains showed antimicrobial activity toward Weissella confusa DSM 20196, Clostridium coccoides DSM 935, or Lactobacillus sakei SAL1. Strains showed an almost common spectrum of activity. The only exceptions were Enterococcus durans DSM20633, which was inhibited by L. plantarum C2 and 1LS9, and Yersinia enterocolitica DSM 4780, which was affected by L. plantarum C2 and 1LS16. The antimicrobial activity of the CFS from L. plantarum C2 was the most intense.

The antimicrobial activities of CFS were assayed at various intervals of time during incubation in MRS broth at 30°C (Table 5). The indicator strain B. megaterium F6 was used for this assay. Only the CFS of L. plantarum C2 showed a halo of inhibition (ca. 2.5 mm) already after 8 h of incubation (cell density of ca. 9.4 log CFU ml⁻¹). The other L. plantarum strains did not show halos of inhibition before 12 h (ca. 1.5 to 2.4 mm; cell densities were 9.5 to 9.7 log CFU ml⁻¹). The antimicrobial activities of CFS of all strains were almost completely lost after digestion with trypsin. The activity was unaffected by heating at 100°C for 5 min, nor by treatment with catalase (data not shown).

TABLE 5.

Antimicrobial activities^a of CFS assayed at various times during 24 h of incubation in MRS broth at 30°C

Strain	Time (h)	Cell density (log CFU ml−1)	Halo of inhibition/cell density (mm log CFU−1 ml−1)
Lactobacillus plantarum C2	6	7.99 ± 0.18	ND
	8	9.40 ± 0.19	0.27 ± 0.02
	12	9.69 ± 0.15	0.35 ± 0.01
	24	10.06 ± 0.21	0.45 ± 0.02
L. plantarum C5	6	8.01 ± 0.18	ND
	8	9.29 ± 0.12	ND
	12	9.51 ± 0.20	0.16 ± 0.03
	24	9.99 ± 0.18	0.47 ± 0.02
L. plantarum POM20	6	8.21 ± 0.21	ND
	8	9.33 ± 0.15	ND
	12	9.72 ± 0.14	0.23 ± 0.02
	24	10.05 ± 0.21	0.33 ± 0.01
L. plantarum POM42	6	8.16 ± 0.12	ND
	8	9.61 ± 0.15	ND
	12	9.69 ± 0.18	0.22 ± 0.01
	24	9.99 ± 0.20	0.40 ± 0.02
L. plantarum 1LS9	6	8.48 ± 0.21	ND
	8	9.40 ± 0.12	ND
	12	9.71 ± 0.15	0.17 ± 0.02
	24	10.10 ± 0.19	0.50 ± 0.03
L. plantarum 1LS16	6	8.12 ± 0.12	ND
	8	9.56 ± 0.18	ND
	12	9.72 ± 0.16	0.25 ± 0.03
	24	10.02 ± 0.19	0.57 ± 0.02

^aThe indicator strain Bacillus megaterium F6 was used. ND, not determined.

Antimicrobial activity in the vegetables and fruits model system.

After precultivation in the indicated vegetable and fruit juices, the antimicrobial activity of L. plantarum strains was assayed on vegetable (carrot and tomato) and fruit (pineapple) agar plates (solid conditions), which mimicked food-like environments. All strains showed halos of inhibition toward B. megaterium F6 that ranged from ca. 4.5 ± 0.2 mm (strain C2) to 1.8 ± 0.1 mm (strains 1LS9 and 1LS16). Further, the antimicrobial activity was investigated under liquid conditions. The strains were inoculated at various levels (ca. 7.0 to 9.0 log CFU ml⁻¹) on vegetable and fruit juices, and the growth at 30°C was allowed for 24 h. Apart from the size of the inoculum, cell densities of ca. 9.2 ± 0.2 to 9.8 ± 0.1 log CFU ml⁻¹ were found after 24 h. Then, CFS were collected and used for the well diffusion assay. None of the strains showed antimicrobial activity toward B. megaterium F6 when the inoculum on vegetable and fruit juices was below ca. 9.0 log CFU ml⁻¹. The same was found when concentrated CFS were used. Although inoculated with ca. 7.0 and 8.0 log CFU ml⁻¹, the CFS of L. plantarum C2 acquired antimicrobial activity when the CFS from the cultures started with ca. 9.0 log CFU ml⁻¹ were added to carrot juice at the beginning of the incubation.

Based on the above results, the active compounds and the expression levels of genes responsible for the antimicrobial activities were further characterized with L. plantarum C2.

Purification of antimicrobial compounds.

CFS of L. plantarum C2, which was grown on carrot juice (inoculum of ca. 9.0 log CFU ml⁻¹), had a peptide concentration of 13.64 ± 0.22 mg ml⁻¹. Preliminarily, CFS was fractionated by ultrafiltration. Inhibitory activity toward the indicator B. megaterium F6 was found in all four fractions (cutoffs of 50, 30, 10, and 5 kDa). Consequently, further characterization was carried out only on the fraction having a molecular mass lower than 5 kDa. This fraction was subjected to partial purification through RP-HPLC, and 30 fractions were separated (the concentration of peptides varied from 0.50 ± 0.02 to 6.72 ± 0.04 mg ml⁻¹). Although weak antimicrobial activities were found in fractions 2 and 4 (0.58 ± 0.01 and 5.2 ± 0.12 mg of peptides ml⁻¹, respectively), the highest inhibition was found for fraction number 3 (5.75 ± 0.03 mg of peptides ml⁻¹). As shown toward the indicator B. megaterium F6, the MIC of fraction 3 was 1.02 mg ml⁻¹. This fraction was subjected to a further step of RP-HPLC purification and nano-LC-ESI-MS analysis. Eight different peptides, containing 7 to 32 amino acid residues, were identified (Table 6). Three peptides were fragments of the PlnJ bacteriocin precursor peptide and two peptides were fragments of the PlnK bacteriocin precursor peptide. Sequences corresponding to fragments encrypted into PlnH and to a fragment of PlnG were also found. The elaboration of the mass spectra in SIM mode revealed the presence of peptides that corresponded to the most intense peaks (f27-53 of PlnJ and f27-53 of PlnK) at the 500-ppm level.

TABLE 6.

Sequences of peptides contained in the purified antimicrobial fraction of the CFS of Lactobacillus plantarum C2

Sample no.	Sequencea	Score	Molecular mass (Da)	Deltab	Fragment	NCBI accession no./source protein
1	MNKMIKDLDVVDAFAPISNNK	35	2,364.77	−0.18	f1–21	AFJ79560/PlnJ, L. plantarum
2	GAWKNFWSSLRKGFYDGKAGRAIRR	42	2,929.34	−0.21	f27–53	AFJ79560/PlnJ, L. plantarum
3	GAWKNFWSSLR	32	1,352.53	0.31	f27–37	AFJ79560/PlnJ, L. plantarum
4	RSRKNGIGYAIGYAFGAVERAVLGGSRDYNK	40	3,503.92	0.42	f1–32	AFJ79559/PlnK, L. plantarum
5	GGSRDYNK	45	896.93	−0.32	f50–57	AFJ79559/PlnK, L. plantarum
6	TAQQVVSNTK	42	1,075.19	0.22	f176–185	ADE34579/PlnH, L. plantarum
7	TIKATKT	38	761.909	0.11	f319–324	ADE34579/PlnH, L. plantarum
8	ENGNTLSG	35	791.11	0.17	f615–622	ADE34578/PlnG, L. plantarum

^aStandard single-letter amino acid codes were used.

^bDelta indicates the difference between the calculated mass and the expected mass.

Detection of genes responsible for plantaricin synthesis.

Ten genes, which are present in the pln locus and are responsible for the synthesis of plantaricins, were searched in the L. plantarum strains (Table 2). PCR amplifications gave 7 DNA fragments with the expected amplicon size. The similarity searches of the gene sequences were carried out with BLAST analyses (http://www.ncbi.nlm.nih.gov/blast) (Table 7). Table 8 summarizes the presence/absence of plantaricin genes in the strains of this study. All strains harbored genes encoding plnA, plnEF, plnG, and plnH. The two-peptide plnJK genes were found in three strains (C2, POM20, and POM42), while strain C5 was positive for plnK but not for plnJ. The gene encoding the putative prebacteriocin (plnN) was only found in L. plantarum C2, POM20, and POM42. None of the strains harbored plnNC8, plnS, or plnW genes.

TABLE 7.

Similarity searches^a of the purified PCR products for plantaricin genes in the Lactobacillus plantarum strains used in this study

Genes	Length (bp)	Accession no.b	% similarity	Homologous protein/strain
plnA	450	ABD15221.1	100	Plantaricin A precursor peptide, induction factor/L. plantarum J51
plnEF	428	2RLW_A	100	Chain A, three-dimensional structure of two peptides that constitute two-peptide bacteriocin plantaricins E/F/L. plantarum C11
plnJ	475	ZP_07076967.1	100	Bacteriocin peptide PlnJ/L. plantarum subsp. plantarum ATCC 14917
plnK	469	YP_004888428.1	99	Putative bacteriocin precursor peptide PlnK/L. plantarum WCFS1
plnG	453	ADE34578.1	100	PlnG/L. plantarum PCS20
plnH	526	ADE34579.1	100	PlnH/L. plantarum PCS20
plnN	146	ZP_07076969.1	100	Bacteriocin precursor peptide PlnN/L. plantarum C11

^aSimilarity searches of the gene sequences were carried out with BLAST analyses (http://www.ncbi.nlm.nih.gov/blast).

^bAccession number corresponding to the homologous sequence.

TABLE 8.

Presence or absence of PCR amplification products for plantaricin genes in the Lactobacillus plantarum strains used in this study

Gene	Lactobacillus plantarum C2	L. plantarum C5	L. plantarum POM20	L. plantarum POM42	L. plantarum 1LS9	L. plantarum 1LS16
plnA	+	+	+	+	+	+
plnEF	+	+	+	+	+	+
plnJ	+	−	+	+	−	−
plnK	+	+	+	+	−	−
plnG	+	+	+	+	+	+
plnH	+	+	+	+	+	+
plnN	+	−	+	+	−	−
plnNC8	−	−	−	−	−	−
plnS	−	−	−	−	−	−
plnW	−	−	−	−	−	−

^a+, gene product present; −, gene product absent.

Levels of expression of plantaricin genes in Lactobacillus plantarum C2 during growth on carrot juice (liquid conditions).

The level of relative expression (RE) of plantaricin genes in L. plantarum C2 during growth (inoculum of ca. 9.0 log CFU ml⁻¹) on carrot juice was determined to establish those which were responsible for antimicrobial activity (Fig. 1C). Efficiencies for all the primer pairs were close to 100%, making them appropriate for analysis through the comparative critical threshold method (ΔΔC_T) (52). For relative quantification, the value of Δ C_T for each sample was determined by calculating the difference between the value of C_T of plantaricin target genes and the value of C_T of the 16S rRNA housekeeping gene. Then, the value of ΔΔC_T for each sample was determined by subtracting the value of ΔC_T of the calibrator (reference sample) from the ΔC_T value for the sample. Since the PCR efficiencies of the target and housekeeping genes were comparable, the normalized level of target gene expression was calculated by using the formula: 2^−ΔΔCT. As previously determined by Desroche et al. (53), a gene was considered overexpressed when its RE level was higher than 2. During growth on carrot juice, the expressed genes were plnJK and plnG. Transcripts of plnA, plnEF, plnH, and plnN were not found. Genes plnJK and plnG were variously overexpressed already after 8 h of growth. A further increase of the level of RE was found up to 19 h, which was followed by a marked decrease. Although all three genes were overexpressed, plnJ showed the highest level, followed by plnK and plnG. When the size of inoculum of L. plantarum C2 on carrot juice (liquid conditions) was ca. 7.0 or 8.0 log CFU ml⁻¹, the expression of the above genes was not found (data not shown). The antimicrobial activity of the CFS from L. plantarum C2, which were collected at various intervals of time during growth on carrot juice, was assayed, and the halos of inhibition are shown in Fig. 1A and B.

FIG 1.

Antimicrobial activity and plantaricin gene expression of Lactobacillus plantarum C2. (A) Well diffusion assay of CFS of L. plantarum C2, which was inoculated into carrot juice at ca. 9.0 log CFU ml⁻¹ and incubated at 30°C for 24 h. Numbers represent results for CFS that were collected at different times: 0, 2, 4, 8, 12, 15, 19, and 24 h. Bacillus megaterium F6 was the indicator microorganism. (B) Halo of inhibition/cell density ratio (mm log CFU⁻¹ ml⁻¹). (C) Relative expression levels of plnG (black bars), plnJ (white bars), and plnK (gray bars) genes. The calibrator condition used was the culture of L. plantarum C2 after 2 h of incubation.

DISCUSSION

Several studies have described the synthesis of bacteriocins by Lactobacillus plantarum, including investigations of the mechanism of regulation and contributions to the knowledge on the ecology adaptation of this species (29, 54, 55). To the best of our knowledge, the synthesis of plantaricins by L. plantarum has only been studied during cultivation in MRS broth (54–57). The physical and chemical characteristics of laboratory media are rather far from those of the natural food ecosystems and poorly suitable for selection of bacteriocin-producing strains for food fermentations. Raw vegetables and fruits possess intrinsic chemical and physical parameters that make them particularly complex and harsh matrices, with numerous factors influencing microbial growth and metabolite production (23). First, this study was aimed at highlighting the regulatory system for synthesisis of bacteriocins in L. plantarum in a vegetables and fruits model system.

Sterile and neutralized CFS from six L. plantarum strains that were grown in MRS broth showed in vitro antimicrobial activity. Almost similar inhibitory spectra were found against the full range of indicator microorganisms. All the strains in this study were previously shown to exert antimicrobial activity toward human pathogenic bacteria (36). The antimicrobial activity of L. plantarum C2 was the most intense. Recently, this strain was used as a functional starter to ferment an Echinacea suspension, which exhibited a marked antimicrobial activity toward Gram-positive and -negative bacteria (58). The antimicrobial activity of L. plantarum strains was assayed on vegetable and fruit agar plates (solid conditions) and juices (liquid conditions), which mimicked food-like conditions. Synthesis of antimicrobial compounds was found on agar plates (solid conditions), and it depended on the size of inoculum when strain C2 was incubated in carrot juice (liquid conditions). Only an inoculum of ca. 9.0 log CFU ml⁻¹ was efficient for detection of the activity. Lower sizes of inocula (ca. 7.0 and 8.0 log CFU ml⁻¹) resulted in the loss of the phenotype. These results suggest an autoinduction mechanism for synthesis of bacteriocins in L. plantarum C2. The synthesis of antimicrobial compounds on solid but not in liquid media has been reported for several lactic acid bacteria (59). L. plantarum WCFS1 and NC8 showed constitutive synthesis of bacteriocins on solid laboratory medium, while cultures in broth depended on the size of the inoculum. Highly diluted broth cultures of WCFS1 (below 5.0 log CFU ml⁻¹) did not show the synthesis of bacteriocins (59). Among lactic acid bacteria, most of the class II bacteriocins were clearly shown to be regulated by quorum sensing (QS) (17, 19, 60). Restoration of the class II bacteriocin phenotype is only achieved by plating the culture, after addition of a purified induction factor (e.g., autoinducing peptide [AIP]) or the CFS, which contained this factor. The CFS of L. plantarum C2 coming from an inocula of ca. 7 and ca. 8.0 log CFU ml⁻¹ acquired antimicrobial activity after the addition of CFS from the same strain, which was cultivated starting from an inoculum of ca. 9.0 log CFU ml⁻¹.

In an attempt to explain why bacteriocins are synthesized on solid and not in liquid media, it should be useful to consider the environments in contrast to those from which lactic acid bacteria are isolated. In nature, most lactic acid bacteria thrive on surfaces within biofilms, where they are under conditions substantially different from those that characterize the planktonic status (61). Within biofilms or colonies, cells are in close contact with their neighbors, thus enabling communication and making group decisions via QS mechanisms (62). Contrary to what happens during growth under liquid conditions, Chao and Levin (63) stated that killing sensitive strains around a bacteriocin-producing colony might markedly increase the concentration of nutrients. Adsorption of bacteria to solid surfaces such as those of fruits and vegetables may cause changes in the expression of relevant genes (e.g., bacteriocin regulation) or it may just enable bacterial communication through limited diffusion of AIP (60). In both cases, an increase in the level of AIP and the activation of the autoinduction QS mechanism might occur. Most bacteriocins that are regulated by QS mechanisms are synthesized under those culture conditions, which better mimic the natural ecological niche (56, 64). The mechanism for the synthesis of bacteriocins in L. plantarum C2 seems to have evolved for competing on solid supports, where the cost/benefit ratio to produce antimicrobial compounds appears to be more favorable than in liquid media (59).

Genes belonging to the pln locus, which are involved in the synthesis of bacteriocins, pheromones, and other peptides, are present in L. plantarum. The identities of the PCR products were confirmed through sequencing. The genes plnA, plnEF, plnG, and plnH were found in all the L. plantarum strains of this study. Although highly conserved, the genes plnJK and plnN were variously detected. L. plantarum C2, POM20, and POM42, which were respectively isolated from carrots and tomatoes, seemed to possess the complex pln locus organized into five operons. These results are in agreement with other studies that have reported these genes as the most prevalent (65, 66). Nevertheless, further analyses on pln loci might not rule out the complete organization of the pln locus into five operons, even in those strains that do not show the PCR product of all the genes. No strains carrying plantaricin S, plantaricin W, or PlnNC8 were found (65, 66). The presence of plantaricin genes encoding plantaricin S and plantaricin W are relatively rare among bacteriocinogenic L. plantarum strains (67, 68). The type of regulatory system and the presence of the various bacteriocin-related peptide genes allow us to classify L. plantarum strains into seven plantaritypes (69). The presence of the plnA gene instead of plnNC8 suggests that they belong to plantaritype group 1 and group 2, which share the common features of the plnABCD regulatory system.

RP-HPLC purification and mass spectrometry analysis of the CFS of L. plantarum C2 revealed the presence of a mixture of eight peptides in the most active fraction. Peptides corresponded to fragments of plantaricins, which were previously reported in the NCBI database. Two peptides (GAWKNFWSSLRKGFYDGKAGRAIRR and RRSRKNGIGYAIGYAFGAVERAVLGGSRDYNK) corresponded to active fragments from PlnJ and PlnK. It had been already elucidated that the release of these sequences through the ABC transporter system is necessary to reveal the antimicrobial activity (70). The two peptides together are considered a unique antibacterial unit (PlnJ/K) (71) that belongs to class II bacteriocins, and it is more active in combination when the composing peptides are present individually (70). Class II includes small, heat-stable peptides that are synthesized as precursor molecules, containing, in most cases, a leader peptide of the double-glycine type (26). This leader peptide is recognized and cleaved by a dedicated ABC transporter, resulting in the translocation of a mature and active bacteriocin into the medium (26). All two-peptide bacteriocins whose mode of action was studied (e.g., plantaricins E/F, plantaricins J/K, lactococcin G, thermophilin, lactacin F, and lactocin 705) rendered the membranes of sensitive bacteria permeable to small molecules (71). Plantaricins J/K permeabilize membranes for monovalent ions, including H⁺, but not for divalent ions, such as phosphate and Mg²⁺ (71). Some other fragments of PlnJ and PlnK were identified. None contained the GxxxG motif (71). Another three peptides were identified. They are encrypted in plantaricins PlnH and PlnG, which are expressed via the conserved part of the bacteriocin loci and are involved in the ABC transport system to secrete and process bacteriocin precursors. Domains, located in the N-terminal proteolytic region of PlnG, are responsible for recognition and cleavage of the double-glycine leader peptides from precursors of various bacteriocins, such as the plantaricins J/K (17, 26, 57). Similar peptides were also found in the antimicrobial extract of Echinacea sp., which was also fermented with L. plantarum C2 (58). The MIC of the purified fraction, which contained the mixture of the identified peptides, was comparable to the MIC values previously reported for partially purified antimicrobial peptides produced by lactic acid bacteria (58, 72). To the best of our knowledge, the antimicrobial activity of plantaricins J/K toward the Gram-negative E. coli was not found previously. A real-time PCR assay confirmed what was found through mass spectrometry analysis. The expression of plnJK genes and the ABC transporter gene of plantaricin G increased during growth of L. plantarum C2 in carrot juice (liquid conditions). Bioactivity of these gene products was confirmed through a well diffusion assay. Nevertheless, it was not possible to establish a proportional relationship between genetic expression and bioactivity at various intervals of time during growth in carrot juice, since an accumulation of the antimicrobial compounds may have an effect. The transcription process for each bacteriocin gene, including posttranscriptional and/or posttranslational modifications, may explain this relationship (73).

The use of vegetable and fruit agar plate and juice models, which mimic food-like conditions, may be considered a valuable tool to screen the synthesis of antimicrobial compounds by lactic acid bacteria and to evaluate the potential behavior and effects in food applications. The combined use of proteomic and transcriptomic approaches allowed our study of the regulatory system of plantaricins J/K by L. plantarum C2, adding another wedge in the complex mosaic of gene functionality, which is related to the selective colonization of different environmental niches.