INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

During the past decade, Listeria monocytogenes has been incriminated in numerous food-borne outbreaks and several sporadic episodes of listeric illness (25). The emergence and persistence of L. monocytogenes on a large variety of dairy, ready-to-eat, and processed foods has led to enhanced interest in antimicrobials for its control. In addition to conventional antimicrobials (organic acids, radiation, packaging, etc.), interest in the use of bacteriocins from food-grade lactic acid bacteria (LAB) has increased. Bacteriocins were defined as ribosomally produced precursor polypeptides or proteins that, in their mature (active) form, exert an antibacterial effect against a narrow spectrum of closely related bacteria. Most of the reported bacteriocins are produced by LAB, which are naturally present in a lot of food products or are added for their technological and preserving characteristics (40).

However, in most studies, when Listeria is exposed to such antibacterial activity, emergence of resistant cells is frequently reported (35). The mechanisms underlying the bacteriocin resistance phenomenon are largely unknown. Because bacteriocin acts mainly in the cytoplasmic membrane, potential modifications of bilayer lipid content and quality have been investigated. Resistance to nisin has been correlated with both modified fatty acid and phospholipid composition (27).

Even if differences in protein expression between sensitive target cells and resistant cells are potentially numerous, the roles of proteins in bacteriocin resistance are unclear. In some target cell species, specific membrane-located bacteriocin receptors of a proteic nature have been identified (42). Modifications or the absence of such receptors could lead to resistance. Some killer toxin-resistant mutants of Saccharomyces cerevisiae expressed much smaller amounts of a protein which acts as a docking protein, facilitating toxin binding to the membrane, where it forms lethal ion channels, like bacteriocins do (38). Synthesis of new membrane proteins could interfere with bacteriocin anchorage on the receptor or in the membrane. In Lactococcus lactis subsp. lactis biovar diacetylactis, the nisin resistance gene nsr encodes a putative protein with a molecular mass of 35 kDa. A strongly hydrophobic region supports the prediction that this protein is an integral membrane protein which could decrease bacteriocin activity. Decreased bacteriocin penetration could also appear to result from membrane protein oversynthesis, as observed by Koch et al. (22) in multidrug-resistant mouse and hamster cells. Synthesis of an enzyme able to degrade the bacteriocin is also a potential efficient resistance mechanism. Jarvis (20) described a nisinase which inactivated nisin. Moreover, cell wall proteins could play a crucial role in resistance, as clearly demonstrated by Dielbandhoesing et al. (10) for two cell wall proteins in nisin resistance of yeast cells. Knowledge of the involvement of proteins in bacteriocin resistance, even if studied in gram-positive bacteria, could also highlight the role of outer membrane proteins in gram-negative resistance, which is probably essential, as demonstrated for Omp4 for the bacteriocin 28b resistance phenotype in Escherichia coli (18).

Two-dimensional electrophoresis (2DE) of proteins is currently the highest-resolution analytical technique available for the study of protein expression patterns. This technique has already been used for studying minocycline-susceptible and -resistant Mycobacterium smegmatis (44). Comparative proteome analysis of Mycobacterium tuberculosis virulent and nonvirulent vaccine strains was carried out with the help of 2DE (21). 2DE can be an important resource in identifying proteins involved in bacteriocin resistance. Thus, 2DE is a powerful tool to highlight the biochemical mechanisms governing development of cell resistance and then will help in the design of new efficient molecules or mixing of molecules with different cell targets.

In this paper, we report physiological and metabolic differences between Listeria monocytogenes variants sensitive (wild type) and resistant to the action of divercin V41, a bacteriocin produced by the LAB Carnobacterium divergens V41. Moreover, 2DE was carried out to study differential protein expression in these two characterized variants.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains, culture conditions, and bacteriocin production. Experiments were carried out with Listeria monocytogenes P (serotype 4b), a wild-type sensitive strain isolated from vacuum-packed cold-smoked salmon (Escola Superior de Biotechnologia, Porto, Portugal) and its bacteriocin-resistant variant (RV41) obtained as described below. The bacteriocin used for resistance studies was divercin V41, the bacteriocin produced by Carnobacterium divergens V41 (34).

Divercin MIC determination. The MIC of divercin V41 was determined after growth on Elliker broth or the simulated cold smoked fish system (SCSFS) (12) at 37°C in microtiter plates containing 100 µl of 1% glucose-supplemented Muller-Hinton broth (Biokar) in each well. The total protein content of the purified bacteriocin stock solution was found to be 700 µg/ml as determined by the bicinchoninic acid assay (Pierce, Rockford, Ill.). The divercin V41 was serially diluted in Mueller-Hinton broth (1:1). Each well received 100 µl of inoculum suspension (10⁵ cells per well), and the microtiter plate was incubated overnight at 37°C. After incubation, yellow wells were positive for growth and red wells were negative. The MIC was arbitrarily defined as the amount of total bacteriocin in the highest dilution which inhibited L. monocytogenes growth.

Isolation of a divercin V41-resistant variant of L. monocytogenes. After MIC experiments, L. monocytogenes P cells were picked from the well containing 87.5 µg of divercin per ml and grown in divercin-free Elliker medium. One of the resistant variants, Listeria monocytogenes RV41, was isolated on a Palcam plate (selective agar and supplement; Merck, Nogent sur Marne, France), and divercin sensitivity and resistance were verified according to the spotting method described by Pilet et al. (34). The frequency of isolation of resistant variants was determined according to the method of Dykes and Hastings (13), in the presence of divercin V41 at five times the MIC.

Antibiogram of the two L. monocytogenes variants. To determine if the resistance to divercin confers a particular antibiotic phenotype to the resistant variant, 19 antibiotics (Table 1) acting on different cell targets were tested. The method used was based on the disk technique using Mueller-Hinton medium (3).

Serotyping and lysotyping of the two L. monocytogenes variants. Serotyping of Listeria monocytogenes strains was performed as described previously (39). Phage typing was carried out according to reference 37 with 29 well-characterized phages isolated from lysogenic strains.

Two-dimensional PAGE. 2D polyacrylamide gel electrophoresis (PAGE) experiments were performed essentially according to the method of Gormond and Phan-Thanh (16). When not indicated, chemicals and materials were from Pharmacia-Biotech (Orsay, France).

Sample preparation. Sensitive Listeria monocytogenes and RV41 cells were grown at 37°C until the middle exponential phase (OD₅₅₀ = 0.5, which corresponds to 10⁹ CFU/ml) in 10 ml of Elliker broth. Cells were harvested by centrifugation (4,000 × g, at 25°C for 30 min). Pellets were washed three times in 10 ml of physiological water (8.5 g of NaCl per liter) and recovered in 0.2 ml of Tris (2-amino-2-hydroxymethyl-1,3-propanediol) buffer (10 mM, pH 7.2) containing 5 mM Mg²⁺, 5 µl of a cocktail of protease inhibitors containing leupeptin, pepstatin, and PEFABLOC [4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride; Boehringer, Mannheim, Germany] at 2 mg/ml, 5 µl of deoxyribonuclease and ribonuclease (100 mg/ml; Boehringer), and 20 µl of a solubilization solution containing 20% CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]propane sulfonate}, 10% DTT (dithiothreitol) and 20% IPG (immobilized pH gradient) 4/7 buffer. Cells were sonicated (Vibra cell; Bioblock, Illkirch, France) three times for 1 min at a power setting of 5 and 50% pulse at 4°C. The mixture was incubated at room temperature for 30 min. Urea (7 M) was added, samples were vigorously agitated (Vortex level 5) at room temperature for 15 min, and the mixture was finally centrifuged (20,000 × g at room temperature for 20 min). Supernatants could be stored at 20°C. Proteins were quantified by the Bradford method (5).

IEF. The 1D separation was carried out on immobilized pH gradients (4/7 Immobilin dry strips of 18 cm) as described by Görg et al. (15) with Multiphor II apparatus. The following voltage gradient was applied: from 0 to 50 V in 0.02 h; 50 V for 1 h; from 50 to 150 V in 0.02 h; 150 V for 1 h; from 150 to 300 V in 0.02 h; 300 V for 2 h; from 300 to 3,500 V in 5 h; and 3,500 V for 11 h. Protein samples (100 µg) were loaded into cups at the anode end. After isoelectric focusing (IEF), strips were equilibrated in a solution containing 50 mM Tris (pH 6.2), 6 M urea, 30% glycerol (vol/vol), 2% (wt/vol) sodium dodecyl sulfate (SDS), and 0.3% (wt/vol) DTT followed by a second bath with the same solution, but with 4.5% (wt/vol) iodoacetamide in place of DTT.

SDS-PAGE. The 2D separation was performed, essentially according to the method of Laemmli (24), in an IsoDalt apparatus (Hoeffer, San Francisco, Calif.) with a 14% acrylamide separating gel, but without a stacking gel and at constant voltage (below 180 V). Large plate gels (200 by 250 by 1 mm) were used to improve resolution. Ten gels were run simultaneously in one tank to improve reproducibility. A low-molecular-weight electrophoresis calibration kit (Amersham-Pharmacia-Biotech, Buckinghamshire, England) was used for protein molecular mass (daltons) reference standards (phosphorylase b, 94,000; bovine serum albumin, 67,000; ovalbumin, 43,000; carbonic anhydrase, 30,000; soybean trypsin inhibitor, 20,100; and -lactalbumin, 14,400).

Analysis of protein spots on 2D gels. Images were scanned (CanoScan FB620 P, Canon, France). Protein gel analysis was performed with Melanie II 2D PAGE software (release 2.2; Bio-Rad, Ivry sur Seine, France). Reference points (landmarks) were marked on images to align and match gels. After gel alignment and matching, pairs (spots present in several gels) could be highlighted. The reverse function evidenced the differences between each gel. Ten gels per sample were analyzed and compared. Spots present in at least nine gels were considered to be consistent spots and were taken into account.

Protein identification. Coomassie blue-stained proteins which had been separated by 2D gel electrophoresis were excised and washed five times with 30 µl each of 40% n-propanol, followed by five washes with 30 µl each of 0.2 M ammonium carbonate containing 50% acetonitrile. The gel pieces were completely dried under reduced pressure (Speed-Vac; Savant). To the dried gel pieces, 0.5 µg of trypsin in 10 µl of 100 mM ammonium carbonate (digestion buffer) was added to allow reswelling of the gel pieces. About 15 µl of digestion buffer was added to completely immerse the gel pieces. Digestion was carried out at 37°C for 2 h. The supernatant was collected, and the gel pieces were extracted with 15 µl of 0.1% formic acid, followed by 15 µl of acetonitrile. Extraction was repeated twice, and all supernatants were pooled and dried in a Speed Vac. To be able to remove supernatants from the gel pieces, two 500-µl Eppendorf tubes arranged concentrically were used. The tube containing gel pieces was pierced with a hypodermic needle to generate a hole large enough to allow the liquid to be centrifuged into the lower tube, but to retain the gel pieces. Also in order to avoid possible cross-contamination, all buffers and wash solutions were pipetted with clean Hamilton syringes, which were reserved solely for the handling of the same solutions, making sure that the needle never touched the gel pieces.

Preparation of crude flagella. One liter of culture was harvested by centrifugation at 5,000 × g, washed twice in a 0.01 M sodium phosphate-0.15 M sodium chloride (pH 7.6) (PBS) buffer, and resuspended at a ratio of 5 ml of PBS/liter of broth culture. Samples, in bottles containing glass beads, were shaken vigorously for 30 min at 20°C. The suspension was centrifuged at 5,000 × g for 15 min, and the supernatant was retained. The pellet was washed twice with PBS by vigorous pipetting to remove sheared flagella trapped within the cell mass. The supernatant and cell washing were pooled and centrifuged at 14,000 × g for 40 min to clear the remaining bacteria, and the resulting supernatant was centrifuged at 200,000 × g for 90 min to harvest crude flagella (33). One milligram of crude flagellar protein was recovered in each experiment and solubilized in 50 µl of PBS.

Sample preparation for scanning electron microscopy. Five milliliters of the cultures was filtered over membranes (0.2-µm-pore-diameter GTTP 01300; Millipore Polycarbonate, Molsheim, France). Membranes were loaded on the surface of a 0.1 M sodium cacodylate (pH 7.2) buffer containing 2.5% glutaraldehyde. Cells or flagella were fixed for at least 48 h at room temperature. Membranes were dehydrated in successive baths of ethanol (10, 25, 50, 75, 95, and 100%) for 10 to 20 min. Samples were desiccated by introducing them into a pressurized enclosure where ethanol was replaced by liquid CO₂ (10°C). Samples were desiccated by heating until the critical point was reached (31°C and 73.8 bars) without deterioration and then were metalized with gold and observed.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Physiological comparison of Listeria monocytogenes sensitive and RV41 variants. Sensitive and resistant variants present the same phenotypic characters, the same serotype (4b), and the same lysotype (1444, 1317, 3274, 2671, and 340) (data not shown). Thus, no drastic surface modifications could be postulated on these bases. The MICs for the sensitive and RV41 strains were determined after growth on Elliker and SCSFS broths. In Elliker broth, the MIC for the resistant variant was at least 10⁴ times higher than that for the sensitive strain (higher than 104 and 0.01 µg/ml¹, respectively). This result was in accordance with that obtained by Métivier (28) with L. monocytogenes Scott A. In the SCSFS medium, the MIC for RV41 was 10⁶ times higher than that for the sensitive strain (higher than 104 and 5.7 10⁵ µg/ml¹, respectively). For the latter strain, the MIC was 175 times higher on Elliker broth than on SCSFS. These results could be explained by two facts. (i) SCSFS is a less nutritive medium in which L. monocytogenes had difficulty in growing, and thus it was more sensitive to bacteriocin activity. (ii) Elliker broth contains many more molecules which are able to interfere with the bacteriocin and artificially decrease the number of molecules free to interact with the target cells. These observations underlined the impact of environmental conditions on bacterial growth and bacterial sensitivity to antibacterial agents.

5

4

5

1

1

View larger version (9K): [in this window] [in a new window]   FIG. 1.   Growth curves of the L. monocytogenes wild type (solid symbols) and the resistant variant (open symbols) on Elliker broth at 37°C.

Comparison of the protein patterns of L. monocytogenes sensitive and RV41 variants. No significant difference was observed between the two 1DE patterns of the total proteins from the two variants of L. monocytogenes P (data not shown).

View larger version (74K): [in this window] [in a new window]   FIG. 2.   Silver-stained 2D gels of 100 µg of total proteins of L. monocytogenes strains sensitive (a) and resistant (b) to the divercin V41. Arrows indicate noncommon proteins. Mr, molecular mass.

View larger version (72K): [in this window] [in a new window]   FIG. 3.   Flagellin expression and interaction with divercin V41 of sensitive and RV41 samples cultivated at 37°C. Lanes: 1, flagellin extract of the sensitive variant; 2, flagellin extract of sensitive strain plus divercin V41 (3.5 µg); 3, molecular mass standards (kilodaltons); 4, divercin V41 (3.5 µg); 5, flagellin extract of RV41 (200 µg) plus divercin V41 (3.5 µg); 6, flagellin extract (200 µg) of RV41. Mr, molecular mass.