Action of divergicin M35, a class IIa bacteriocin, on liposomes and Listeria

Authors


Ismaïl Fliss, Dairy Research Center STELA, Pavillon Paul Comtois, Université Laval, Québec, Qc, Canada, G1K 7P4. E-mail: ismail.fliss@aln.ulaval.ca or

Djamel Drider, Laboratoire de Microbiologie Alimentaire et Industrielle, ENITIAA, rue de la Géraudière, BP 82225, 44322 Nantes Cedex 3, France. E-mail: drider@enitiaa-nantes.fr

Abstract

Aims:  The mode of action of divergicin M35, a class IIa bacteriocin, was studied against Listeria monocytogenes with sensitive (DivS) and resistant (DivM) phenotypes, as well as on synthetic phospholipid liposomes.

Methods and Results:  Divergicin-induced release of 1,6-diphenyl-1,3,5-hexatriene (DPH) from zwitterionic (DMPC) and anionic (DMPC/DMPG, 4:1) liposomes, divergicin binding to liposomes, intracellular ATP concentration, cation efflux, cell affinity for hydrocarbons and cell lysis were measured and cell damage was visualized by fluorescence imaging and transmission electron microscopy. Divergicin M35 at 5 μg ml−1 induced DPH efflux from anionic and zwitterionic liposomes at rates of about 2·58% and 1·61% per minute, respectively. DPH efflux rate from anionic liposomes was reduced by about 1·83% and 2·1% per minute in the presence of Li+ and Ca2+, respectively. Binding affinity of divergicin M35 to anionic and zwitterionic liposomes was about 86% and 63%, respectively. Intracellular ATP decreased in the sensitive and the resistant strains by 96·7% and 72·8%, respectively after 20 min of exposure to 5 μg ml−1 divergicin M35. Lysis of the sensitive strain reached 57% in 18 h at a concentration of 5 μg ml−1 when compared with the lysis of the divergicin-resistant strain (38·8%). The K+ and Na+ efflux from the divergicin-sensitive strain reached 87% and 80% of the total ion content within 5 min of exposure. This strain also showed higher affinity for hydrocarbons.

Conclusions:  The cell death of listerial strains upon addition of divergicin M35 could result from ATP depletion, K+ and Na+ efflux, and bacteriolysis. This triple biological effect was attenuated in the DivM strain.

Significance and Impact of the Study:  This study contributed to the understanding of the mode of action of divergicin M35, a pediocin-like bacteriocin.

Introduction

Bacteriocins are small peptides with antimicrobial activity that are produced by bacteria and directed against bacteria (Tagg et al. 1976; Klaenhammer 1993; Jack et al. 1995; Ennahar et al. 2000) other than the producer species. Bacteriocins produced by lactic acid bacteria are the most common bacteriocins and have been studied thoroughly (Montville and Chen 1998; Héchard and Sahl 2002). Based on their biochemical and genetic properties, three classes of bacteriocins are recognized. Class I contains small, post-translationally modified peptides called lantibiotics, which contain unusual amino acids such as lanthionine. Class II includes unmodified bacteriocins, which are divided into subclasses IIa (pediocin-like bacteriocins), IIb (bacteriocins made up of two polypeptide chains), and IIc (other single-chain bacteriocins). Class III bacteriocins are thermosensitive peptides. A fourth class, which contains complex molecules, was included in Klaenhammer's classification (1993), but is currently the subject of discussion and not formally recognized (Cotter et al. 2005). The pediocin-like bacteriocins are small (<10 kDa), listericidal, heat-stable, nonlanthionine-containing peptides of 37–48 amino acids. They consist of two distinct domains, including the highly conserved hydrophilic N-terminal β-sheet sequence YGNGV(X)C(X)4C(X)V(X)4A, where X denotes any amino acid. Class IIa bacteriocins such as sakacin G, plantaricin 423, pediocin PA-1/AcH, divercin V41, enterocin A, and divergicin M35 contain an additional C-terminal disulfide bridge, which plays an important role in stabilizing the three-dimensional structure of the C-terminal domain (Fimland et al. 2005). These C-terminal-stabilized bacteriocins display higher antimicrobial potency than bacteriocins containing only one disulfide bridge. This difference has not yet been explained, nor is it clear how these class II bacteriocins inhibit or inactivate the food-borne pathogen Listeria monocytogenes. In this study, we have examined the potency of divergicin M35 against three strains of Listeria and its behaviour with liposomes.

Materials and methods

Bacterial strains and culture conditions

Carnobacterium divergens M35 (Tahiri et al. 2004) was grown at 30°C under anaerobic conditions in de Man, Rogosa, and Sharpe (MRS) broth (de Man et al. 1960; Rosell Institute Inc., Montreal, PQ, Canada) containing 0·1% Tween 80 (v/v). Listeria monocytogenes LSD 530 referred to as divergicin-sensitive (DivS) strain was obtained from the Canadian Food Inspection Agency (Laboratory Services Division, Ottawa, ON, Canada), while the divergicin M35-resistant variant of L. monocytogenes LSD 530 referred to as DivM strain was recently isolated (Naghmouchi et al. 2006). Listeria innocua and L. monocytogenes were grown in tryptic soy broth (TSEYB; Difco Laboratories, Sparks, MD, USA) supplemented with yeast extract (0·6% w/v) and incubated aerobically without shaking for 16 h at 30°C. All strains were kept frozen at −80°C and subcultured at least three times at 24-h intervals prior to tests.

Divergicin M35 purification

Divergicin M35 produced by C. divergens M35 was purified to homogeneity by the method of Tahiri et al. (2004).

Preparation of 1,6-diphenyl-1,3,5-hexatriene-loaded liposomes

Dimyristoyl phosphatidylglycerol (DMPG), dimyristoyl phosphatidylcholine (DMPC), dipalmitoyl phosphatidylglycerol (DPPG), and dipalmitoyl phosphatidylcholine (DPPC) were obtained from Avanti Polar Lipids (Alabaster, Ala). DMPC/DMPG 4 : 1 and DPPC/DPPG 4 : 1 liposomes (relatively anionic at pH 6·0), as well as DMPC and DPPC liposomes (zwitterionic at pH 6·0) were prepared using the protocol of Fiorini et al. (1987). The fluorescent probe DPH (1,6-diphenyl 1,3,5-hexatriene; Invitrogen Inc., Burlington, Ontario, Canada) was incorporated into the lipid suspension before drying the mixture. The resuspended liposomes had a lipid/probe molar ratio of 0·1%.

Measurement of DPH release from liposomes

Divergicin-induced release of DPH from liposomes was determined by monitoring the increase in fluorescence of the suspension following addition of 2·5 and 5 μg ml−1 of purified divergicin M35 in medium buffered at pH 6·0 as per the protocol of Kaiser and Montville (1996). Emission at 350 nm following excitation at 490 nm was measured with a spectrofluorometer (Fluo-star Galaxy®; BMG Lab technologies, Durham, NC, USA). Total DPH release was brought about by adding Triton X-100 (0·2%, v/v). DPH efflux at time t was expressed as per cent efflux = [(FtF0)/(FF0)] × 100, where Ft was the fluorescence at time t, F0 was initial fluorescence, and F was the fluorescence after Triton X-100 addition. DPH efflux rates (per cent per minute) were calculated from the slope of a tangent to the efflux curve at 100 s after divergicin addition. To determine the effect of cations on divergicin-induced DPH release, divalent cations (Mg2+, Ca2+ and Mn2+) and monovalent cations (K+, Na+ and Li+) were added at final concentrations of 10 and 20 mmol, respectively.

Divergicin M35 binding to liposomes

Bacteriocin binding to liposomes was determined by adding 0·1 ml of divergicin M35 (0·15 mg ml−1) to 0·9 ml of a liposome suspension (1 mg ml−1) prepared without DPH (Pantev et al. 2003). The suspension was stirred at 220 rev min−1 at room temperature for 15 min and then centrifuged (15 000 g, 30 min). The supernatant was tested by enzyme-linked immunosorbent assay (ELISA) using antidivergicin M35 polyclonal antibodies. Percentage of bound bacteriocin was calculated as 100 – (CN/C0 × 100), where C0 and CN are initial and final concentrations of unbound divergicin, respectively.

Fluorescence imaging microscopy

DPH-loaded liposomes were observed in the absence or presence of divergicin M35 (at 5 μg ml−1) through the 60X objective of a fluorescence microscope (model BX51; Olympus, Lake Success, NY, USA) equipped with a 100 W mercury lamp. High-resolution images were captured by a cooled CCD camera (CoolSNAP-fx, 12-bit digitization; Roper Scientific, Inc., Trenton, NJ, USA). Excitation and emission filters used with the dichroic mirror were 400 ± 10 and 530 ± 20 nm, respectively. The CCD acquisition time used was 2 s. Images were processed for contrast and brightness and analysed with Image PRO-PLUS 4·5 software (Media Cybernetics, Silver Spring, MD, USA).

Effect of divergicin M35 on cell lysis

Cell lysis was measured using sterile flat-bottom 96-well microtitre plates (Falcon, Becton-Dickinson and Company, Franklin Lakes, NJ, USA). Cells of DivS, DivM and L. innocua obtained by centrifugation (10 000 g, 4°C, 10 min) from 10 ml of 6-h cultures were washed and resuspended in 5 ml of sodium phosphate buffer (20 mmol l−1, pH 6·5). Bacterial suspension (125 μl) was placed in microtitre plate wells and 125 μl of sodium phosphate buffer containing divergicin M35 at 0, 0·625, 1·25, 2·5, or 5 μg ml−1 was added. Plates were incubated at 30°C for up to 18 h. Absorbance at 650 nm was measured using a Thermomax microplate reader (Molecular Devices, Opti-Ressources, Charny, QC, Canada). Percentage of cell lysis was calculated as 100 – (At/A0 × 100), where A0 and At were absorbance measured at 0 and 3 or 18 h of incubation, respectively.

Measurement of divergicin-induced ATP leakage from bacterial cells

The DivS and DivM strains grown to an A650 of 0·5 were harvested by centrifugation (10 000 g, 4°C, 10 min), washed twice with 50 mmol l−1 of morpholine ethanesulfonic acid (MES) buffer (pH 6·5, Sigma-Aldrich, Oakville, ON, Canada) and kept on ice until use. Cells were energized prior to ATP measurement by suspending in MES buffer (half of the original culture volume) containing 0·2% dextrose and incubated for 30 min at 30°C. Total and extracellular ATP concentrations were measured for up to 20 min of exposure to 0, 1·25, 2·5, or 5 μg ml−1 of purified divergicin M35 using the bioluminescence method of Guihard et al. (1993) and an ATP bioluminescence assay kit (Sigma-Aldrich, Oakville, ON, Canada). A standard curve prepared based on the bioluminescence response of known pure ATP concentrations was utilized. ATP levels were expressed as nmol mg−1 of cell dry mass.

Determination of divergicin-induced ion efflux from bacterial cells

The DivS and DivM strains were grown in TSEYB broth at 30°C to the exponential growth phase. Cells were centrifuged (10 000 g, 4°C, 10 min), washed twice with MES buffer, resuspended in MES buffer containing glucose (0·2%, w/v) plus divergicin M35 at 2·5 or 5 μg ml−1 and incubated at 30°C for up to 30 min. Filtrates obtained using 0·45-μm Millipore HA filters were analysed for K+ and Na+ content using an atomic absorption spectrophotometer (Perkin-Elmer 3300, Ueberlingen, Germany). To determine the total Na+ and K+ content, cells were treated with 2 μg ml−1 nisin A (Aplin and Barrett Ltd., Beaminster, UK) for 1 h as described by Abee et al. (1994). Ion efflux was expressed in nmol mg−1 of cell dry mass vs incubation time.

Transmission electron micrographs of divergicin-treated bacterial cells

Listeria monocytogenes was grown to mid-log growth phase at 30°C in TSEYB with or without divergicin M35 (20 μg ml−1), harvested by centrifugation (5000 g, 10 min, 4°C), microencapsulated in agarose (3%, w/v), and fixed for 2 h at 4°C in 0·05% glutaldehyde (Marivac Ltd., Halifax, NS, Canada) and 2·5% paraformaldehyde in 0·1 mol l−1 sodium cacodylate buffer at pH 7·2 (JBS-CHEM, Dorval, PQ, Canada). Samples were washed four times (5 min each) with cacodylate buffer, postfixed for 2 h at 4°C in cacodylate buffer with 0·1% (w/v) osmium tetroxide (Sigma), washed four times, dehydrated in a graded ethanol series, embedded in Quetol (Marivac, Halifax, Canada), and polymerized at 60°C for 48 h (Abad et al. 1987). Ultra-thin sections (0·1 μm) were cut with an ultramicrotome (Reichert-Jung, Vienna, Austria) and collected on Formvar-coated nickel grids (JBEM, Dorval, Canada). Grids were washed in cacodylate buffer, dried, stained with uranyl acetate and lead citrate, and examined with a 1200 EX microscope (JEOL, Peabody, MA) at 80 kV.

Determination of cell affinity for hydrocarbons

The DivS and DivM strains were grown to the logarithmic growth phase, harvested by centrifugation (5000 g, 10 min), washed twice, and resuspended at a concentration of 108 CFU ml−1 in 0·15 mol l−1 NaCl solution. Hexadecane or decane (0·4 ml) was vortexed for 60 s and the suspension was allowed to stand for 20 min. Absorbance by the aqueous phase at 400 nm was measured using a spectrophotometer (Molecular Devices Corp., Sunnyvale, CA, USA). The percentage of cells extracted into each hydrocarbon liquid was determined by the following formula: 100 × [1 – (A/A0)], where A0 and A are the absorbances by the aqueous phase prior to and after mixing.

Results

Effect of divergicin M35 on synthetic liposomes

Exposure of synthetic liposomes to divergicin M35 resulted in a gradual release of entrapped DPH. DPH leakage rates measured after 100 s in the presence of divergicin M35 at concentrations of 2·5 and 5 μg ml−1 were about 1·38% and 2·58%, respectively, per minute from DMPC/DMPG 4 : 1 liposomes and about 0·8% and 1·61%, respectively, per minute from DMPC liposomes (Fig. 1a). The proportion of loaded DPH released in the presence of divergicin at 5 μg ml−1 was 13·4% for DPPC/DPPG 4 : 1 and 10·.3 % for DPPC liposomes after 980 s (Fig. 1b) compared with 17·6% from DMPC/DMPG 4 : 1 and 9·2% from DMPC liposomes after 980 s (Fig. 1c). The addition of divalent cations Mg2+, Mn2+, and Ca2+ reduced divergicin-induced DPH efflux from DMPC/DMPG 4 : 1 liposomes by 57%, 48%, and 82%, respectively, while reductions owing to the addition of monovalent cations K+, Na+, and Li+ were 52%, 47% and 71%, respectively (Fig. 1d). Fluorescence micrographs of DPH-loaded liposomes in the absence or presence of divergicin M35 at a concentration of 5 μg ml−1 indicated that divergicin M35 disrupted liposomes (Fig. 2a,b).

Figure 1.

 (a) Rate* of 1,6-diphenyl-1,3,5-hexatriene (DPH) efflux from liposomes of DMPC/DMPG 4 : 1 (white) and DMPC (grey) in the presence of divergicin M35 at concentrations of 2·5 and 5 μg ml−1. (b) DPH efflux† from liposomes of DPPC/DPPG 4 : 1 (circle) and DPPC (diamond) over time in the presence of divergicin M35 at a concentration of 5 μg ml−1. (c) DPH efflux‡ from DMPC/DMPG (square) and DMPC (diamond) liposomes over time in the presence of divergicin M35 at a concentration of 5 μg ml−1. (d) Rate* of DPH efflux from DMPC/DMPG (black) and DMPC (white) liposomes in the presence of divergicin M35 at a concentration of 5 μg ml−1 upon addition of mono- and divalent cations. Notes:*Per cent per minute calculated from the slope of the tangent to the curve at 100 s. †Per cent of total release induced by Triton X-100 (0·2%, v/v). Fitted curve is based on a logarithmic and polynomial function. ‡Per cent of total release induced by Triton X-100 (0·2%, v/v). Fitted curve is based on a logarithmic function.

Figure 2.

 Fluorescence images of 1,6-diphenyl-1,3,5-hexatriene (DPH)loaded liposomes (DMPC) in the presence (a) or absence (b) of divergicin M35 at a concentration of 5 μg ml−1.

Binding of divergicin M35 to liposomes

ELISA measurements indicated that the relatively anionic DMPC/DMPG, 4 : 1 and DPPC/DPPG 4 : 1 liposomes were bound to 86% and 67% of the added divergicin M35, respectively, while the zwitterionic DMPC and DPPC liposomes were bound to 63% and 65% of divergicin M35.

Effect of divergicin M35 on cell lysis

The effect of divergicin M35 at concentrations of 0·625, 1·25, 2·5, and 5 μg ml−1 on listerial cell lysis is shown in Fig. 3. Three hours of contact with cells at all divergicin M35 concentrations had a little effect on the lysis of the DivS and DivM strains but a noticeable effect on the lysis of L. innocua (13·97% ± 1·8 to 51·2% ± 0·18) at the higher concentrations (Fig. 3a). Eighteen hours of contact produced 33·9 ± 2·2 to 57 ± 1·41% lysis of the DivS strain, 30·9 ± 0·43 to 38·8 ± 1·2% lysis of the DivM strain and 37·1 ± 1·6 to 82·6 ± 0·32% lysis of L. innocua, all in a concentration-dependent manner (Fig. 3b). It should be noted that incubation in the phosphate buffer without bacteriocin produced measurable time-dependent lysis.

Figure 3.

 Lysis of the divergicin-sensitive (DivS) strain (square), divergicin-resistant (DivM) strain (triangle), and Listeria innocua (diamond) in contact with divergicin M35 at concentrations of 0·625, 1·25, 2·5, and 5 μg ml−1 for 3 h (a) and 18 h (b) at 30°C. Error bars represent the standard deviation calculated from duplicate experiments.

Measurement of intracellular and extracellular ATP concentration

ATP levels in energized cells of the DivS and DivM strains increased to 6·65 and 3·98 nmol mg−1, respectively after 20 min of incubation in the absence of divergicin M35. In the presence of the bacteriocin, intracellular ATP was depleted in a time- and concentration-dependent manner (Fig. 4a,b). After 20 min at 5 μg ml−1, ATP levels in DivS and DivM strains were 3·3% and 27·7%, respectively, of the initial levels. No extracellular ATP was detected at any of the divergicin concentrations tested (results not shown).

Figure 4.

 Intracellular ATP levels of the divergicin-sensitive (DivS) strain (left) and the divergicin-resistant (DivM) strain (right) in contact with 0 (diamond), 1·25 (square), 2·5 (triangle), and 5 μg ml−1 (circle) concentrations of divergicin M35.Values presented are the means of two independent bioluminescence measurements.

Divergicin-induced ion efflux from Listeria

Total K+ and Na+ contents of freshly prepared cells of the DivS and DivM strains after treatment with nisin A at a concentration of 2 μg ml−1 for 1 h or divergicin M35 at a concentration of 2·5 or 5 μg ml−1 for 20 min at 30°C have led to Na+ and K+ efflux. After treatment with 2 μg ml−1 of nisin A, the total amount of K+ released from the DivS and DivM strains was quite similar [344·61 ± 5·45 (nmol mg−1 of cell dry mass) in the case of DivS and 348·50 ± 2·74 in the case of DivM]. A notable decline was detected in total K+ release from both strains: the amounts of K+ released were 337·8 ± 7·23 and 180 ± 2·3, respectively, when the DivS and DivM cells were treated with divergicin M35 at 5 μg ml−1 (Fig. 5a). The total Na+ released was 610·99 ± 6·17 for DivS and 646·37 ± 2·33 for DivM when the cells were treated with nisin. This amount slightly changed when divergicin M35 (at 5 μg ml−1) was used (595 ± 3·18 for DivS and 621·5 ± 3·10 for DivM) (Fig. 5b). For the DivM strain, the K+ and Na+ efflux reached 180·4 ± 2·3 and 621·5 ± 3·1 nmol mg−1, respectively. After 20 min of divergicin treatment (at 5 μg ml−1), the K+ efflux from DivS and DivM strains were about 92% and 52%, respectively, of the total contents. Remarkably, no extracellular calcium (Ca2+), manganese (Mn2+), or iron (Fe2+) cations were found in the media (data not shown).

Figure 5.

 (a) Time and concentration-dependent K+ efflux from divergicin-sensitive (DivS) (open symbols) and divergicin-resistant (DivM) strains (filled symbols) cells treated with 2·5 (triangle) and 5 (diamond) μg ml−1 of divergicin M35. (b) Time and concentration-dependent Na+ efflux from DivS (open symbols) and DivM strains (filled symbols) cells treated with 2·5 (square) and 5 (circle) μg ml−1 of divergicin M35. Values are the mean of three independent measurements.

Transmission electron microscopy

The transmission electron microscopy (TEM) images of L. monocytogenes not treated with divergicin M35 revealed regular rod-shaped cells with a continuous uniform cell wall (Fig. 6a,b) whereas a treatment with divergicin M35 at a concentration of 20 μg ml−1 caused cell wall damage (Fig. 6c,d). Furthermore, the damaged cells showed dispersed intracellular material likely attributable to the destruction of a major part of the bacterial cell wall. The outer layers of the DivM strain were rounder and rougher than the outer layers of the DivS strain (Fig. 6a,b).

Figure 6.

 Electron micrographs of thin sections of the divergicin-sensitive (DivS) strain (a) and the divergicin-resistant (DivM) strain (b) grown to logarithmic growth phase. Panels c and d show the strain 530 with divergicin M35 at a concentration of 20 μg ml−1. CW, cell wall; CM, cytoplasmic membrane; IM, intracellular material; LCW, lysed cell wall. Magnification = 60 000X. Scale bar = 200 nm.

Cell affinity for hydrocarbons

The DivS strain displayed greater affinity for both hexadecane and decane than the DivM strain (Fig. 7).

Figure 7.

 Affinity of the divergicin-sensitive DivS and divergicin-resistant (DivM) strains for hexadecane (striated) and decane (grey). Values presented are the means of two independent measurements.

Discussion

The mode of action of class IIa bacteriocins from gram-positive bacteria has been extensively studied (Cotter et al. 2005; Drider et al. 2006). These bacteriocins are believed to induce permeabilization of the target cell membrane, probably by forming ion-selective pores, which results in dissipation of the proton motive force and depletion of intracellular ATP (Montville and Chen 1998). In direct comparison, divergicin M35 caused depletion of intracellular ATP, particularly in the DivS strain. This class IIa bacteriocin also induced DPH efflux from synthetic liposomes, in greater amounts from DMPC/DMPG 4 : 1 liposomes than DMPC liposomes. The DMPG has been identified as a major constituent of listerial membranes, whereas DMPC has been shown to be present in greater proportions than DMPG in membranes of listerial strains of bacteriocin-resistant phenotype (Verheul et al. 1997; Mazzotta and Montville 1999). This induced leakage of DPH suggests that the class II bacteriocin does not require any specific protein receptor. Nevertheless, the presence of a specific target for bacteriocins on the bacterial cell surface has been suggested by Fimland et al. (1998). The authors observed that a pentadecapeptide fragment starting from the central hinge region and including the C-terminus of pediocin PA-1/AcH inhibited pediocin PA-1/AcH activity. They hypothesized that this pentadecapeptide fragment may interact with a docking molecule involved in pediocin PA-1/AcH recognition/binding. It has also been reported that leucocin A activity requires a chiral receptor at the surface of the target cell (Yan et al. 2000).

The higher affinity of DivS for alkane liquids leads us to speculate that the cell surfaces of the DivS and DivM strains may contain different amounts of OCOO and HSOinline image groups, as reported for Lactobacillus spp. cell surfaces (Pelletier et al. 1997; Monica et al. 2002). The presence of such anionic groups could explain the variations in binding of cationic bacteriocins such as divergicin M35.

Our findings indicate that divergicin M35 produced nearly complete Na+ efflux for both DivS and DivM strains (based on comparison with nisin A). However, the incomplete K+ efflux for DivM suggests that DivS and DivM differ significantly in their sensitivity to divergicin M35. The release of K+ ions from sensitive cells has been observed also for other bacteriocins such as Lactococcin G (Moll et al. 1996) and Piscicocin CS526 (Suzuki et al. 2005). The Lactococcin G does act as a potassium-conducting pore (Moll et al. 1996). Divergicin M35 and Lactococcin G are not only specific for K+ ions, but also for monovalent Na+ ions. Remarkably, Lactococcin G also causes influx of Na+ into the cells (Moll et al. 1996). Intracellular Na+ is known to be toxic. Finally, another pediocin-like bacteriocin named Piscicocin CS526, was shown to induce the efflux of K+ ions from the target cells, causing dissipation of the transmembrane potential (Δφ) of the cell membrane (Suzuki et al. 2005). In direct comparison, Chen and Montville (1995) reported that K+ efflux from L. monocytogenes Scott A induced by pediocin PA-1/AcH at 10 AU ml−1 reached a maximum of 540 nmol mg−1, a value which is 1·6 and 3 times higher than that obtained with nisin A for the DivS and DivM strains, respectively. In the present study, the effect of divergicin M35 on the DivS strain reached its maximum after 10 min. The absence of divalent cations in the medium suggests that the induced ion permeability was limited for both divergicin M35 and nisin A. This is not surprising as the production of visible damage to the DivS strain by divergicin M35 required hours of exposure. Such damage was localized in the membrane, which normally leads to cell death following evacuation of cytoplasmic material.

The concentration of 5 μg ml−1 divergicin M35 was required to bring about cell lysis. However, as we recently reported (Naghmouchi et al. 2006), the minimum inhibitory concentration (MIC) against L. monocytogenes LSD 530 was less than 0·128 μg ml−1. Remarkably, at this concentration (0·128 μg ml−1), there is no ATP or Na/K leakage, which may indicate that the pore formation is unlikely to be significant at these concentrations. This phenomenon seems to be analogous to the situation with that of nisin, which at high concentrations can bring about cell lysis in the absence of a receptor but at low concentrations the lipid II receptor is required. Nisin appear to form channels in bacterial membranes using lipid II, the prenyl chain-linked donor of the peptidoglycan building block, both as a receptor and as an intrinsic component of the pore (Breukink et al. 2003). The length of the chain of lipid II plays an important role in maintaining pore stability. The interaction with lipid II is required for pore formation, and the pores are stable for seconds (Wiedemann et al. 2004). The data presented herein would seem to indicate that divergicin M35 activity at 5 μg ml−1 is dependent on pore formation but that another (receptor-mediated) mechanism is responsible for inhibitory activity at 0·128 μg ml−1. For understanding, the mode of action of class IIa bacteriocins at cellular and molecular levels, see a recent review by Drider et al. (2006).

Being a positively charged peptide, divergicin M35 may act through electrostatic interaction with negatively charged phospholipid head groups. This seems plausible as binding to the relatively anionic DMPC/DMPG 4 : 1 liposomes was 1.3 times greater than that observed for the zwitterionic liposomes, suggesting that binding of divergicin M35 to liposomes was charge-dependent. Overall, it should be pointed out that a significant difference was observed in bacteriocin binding between the anionic liposomes DMPC/DMPG and DPPC/DPPG (86% and 67%, respectively), which corresponded to DPH-efflux of 17·6% and 13·4%, respectively. On the other hand, there was no significant difference in bacteriocin binding between the anionic liposome DPPC/DPPG and the zwitterionic (DMPC and DPPC) liposomes (67% and 63–65%, respectively) but the differences in DPH-efflux were significant (13·4% and 9·2–10·3%, respectively). This could be explained by several means such as differences in the electrostatic forces between the antimicrobial peptide agent-lipids and integration and orientation of the peptide within bilayers membranes. When DPPC liposomes were used in fluorescence studies, a synthetic peptide from VP3 capsid protein of hepatitis A virus was detected in hydrophobic environment. The addition of a 5% of a charged lipid, DPPG, to the preparation changed the preference of the peptide towards a polar surrounding (Sospedra et al. 2000). Further, the integration of gramicidin A, a 15-residue hydrophobic peptide, in membrane bilayers composed of different anionic and zwitterionic lipids revealed that in DMPC membranes the orientation of lipids was comparable with that in the absence of peptide, while in DMPS, DMPG, and DMPE the degree of orientation of peptide and lipid chain was less than in that of DMPC (Kota et al. 2004).

Furthermore, cations such as Ca2+ and Li+ resulted in reduced divergicin binding to synthetic liposomes. In summary, we may conclude that listericidal potency of divergicin M35 may result from induced efflux of K+ and Na+ ions, ATP depletion, and lysis.

Acknowledgements

This work was carried out under the auspices of the Canadian Research Network on lactic acid bacteria and funded by The National Science and Engineering Research Council of Canada, Agriculture and Agri-Food Canada, Novalait Inc., Dairy Farmers of Canada and by the Province of Quebec Fond pour les Chercheurs et l'Avancement de la Recherche. Karim Naghmouchi is a recipient of a postgraduate scholarship from the Tunisian government. The authors would like to thank Gülhan U. Yüksel (University of Idaho) for critical reading of this manuscript.

Ancillary