Ismail Fliss, STELA Dairy Research Center, Nutraceuticals and Functional Foods Institute, Université Laval, G1K 7P4 Québec City, QC, Canada. E-mail: email@example.com
The aim of this study is to evaluate the capacity of three bacteriocin producers, namely Lactococcus lactis subsp. lactis biovar diacetylactis UL719 (nisin Z producer), L. lactis ATCC 11454 (nisin A producer) and Pediococcus acidilactici UL5 (pediocin PA-1 producer), and to grow and produce their active bacteriocins in Macfarlane broth, which mimics the nutrient composition encountered in the human large intestine.
Methods and Results
The three bacteriocin-producing strains were grown in Macfarlane broth and in De Man–Rogosa–Sharpe (MRS) broth. For each strain, the bacterial count, pH drop and production of organic acids and bacteriocins were measured for different period of time. The ability of the probiotic candidates to inhibit Listeria ivanovii HPB 28 in co-culture in Macfarlane broth was also examined. Lactococcus lactis subsp. lactis biovar diacetylactis UL719, L. lactis ATCC 11454 and Ped. acidilactici UL5 were able to grow and produce their bacteriocins in MRS broth and in Macfarlane broth. Each of the three candidates inhibited L. ivanovii HPB 28, and this inhibition activity was correlated with bacteriocin production. The role of bacteriocin production in the inhibition of L. ivanovii in Macfarlane broth was confirmed for Ped. acidilactici UL5 using a pediocin nonproducer mutant.
The data provide some evidence that these bacteria can produce bacteriocins in a complex medium with carbon source similar to those found in the colon.
Significance and Impact of the Study
This study demonstrates the capacity of lactic acid bacteria to produce their bacteriocins in a medium simulating the nutrient composition of the large intestine.
Probiotics are defined as living micro-organisms that have beneficial effects on the health of the host when administered in adequate doses. Many beneficial effects have been attributed to probiotics, including improvement in digestion and intestinal transit (Marteau et al. 2002), anti-diarrhoeal effects (Saavedra 1995), prevention of eczema (Isolauri et al. 2000) and prevention of food allergies (Matricardi 2002). However, restoring the balance of the intestinal microbiota following antibiotic treatment and inhibition of enteric pathogens remain the most studied effects. Several clinical studies have demonstrated that strains of Lactobacillus reuteri (Shornikova et al. 1997) and Bifidobacterium bifidum (Saavedra 1995) may be very effective in preventing or mitigating bacterial and viral diarrhoea. Clinical studies conducted by Anukam et al. (2006) have shown that Lactobacillus rhamnosus GR1 or Lact. reuteri RC14 can be used to improve the effectiveness of certain antibiotics in the treatment for Clostridium difficile secondary infection. While the exact mechanisms remain unknown, many hypotheses have been advanced to explain the antimicrobial activity of probiotics, notably competition for binding sites on the epithelium (Conway et al. 1987), decreased intestinal cell permeability (Stratiki et al. 2007), competition for nutrients (Sonnenburg et al. 2006), degradation of toxin receptors (Castagliuolo et al. 1996), stimulating the immune system (Kaila et al. 1992) and production of various inhibitory substances such as organic acids (Vandenbergh 1993), bio-surfactants or hydrogen peroxide (Pridmore et al. 2008), all previously reviewed (Fliss et al. 2011). Secretion of cell signalling compounds such as reutericyclin (Gänzle 2004) and bacteriocins has also been proposed as an antimicrobial mechanism. However, we have found only one scientific study proving that these substances are produced under digestive tract conditions (Corr et al. 2007).
Bacteriocins are peptide products of ribosomal synthesis (Bowdish et al. 2005) and exert bactericidal or bacteriostatic effects against bacteria phylogenetically close to the producer strain (Tagg et al. 1976), which is immune to its own bacteriocin (Cotter et al. 2005). We have shown that pediocin producer Pediococcus acidilactici UL5 is highly resistant to simulated gastrointestinal tract (GIT) conditions and that its ability to produce pediocin is not affected under these conditions (Kheadr et al. 2010). We have also shown that pediocin has minimal effect on normal intestinal flora (Le Blay et al. 2007). These findings suggest that this strain could be considered as relevant probiotic candidate. An in vivo study (Dabour et al. 2009) conducted on ICR mice showed that Ped. acidilactici UL5 is resistant to passage through the mouse GIT and that colonies recovered from faecal samples 1 day after ingestion were still be able to produce active pediocin PA-1. Administering purified pediocin PA-1 at 250 μg day−1 for three consecutive days reduced faecal listerial counts by up to two log cycles relative to a control group. Neither Ped. acidilactici UL5 nor purified pediocin PA-1 had any negative effect on feed intake or body weight. However, we did not show in this study that Ped. acidilactici UL5 was able to multiply and produce pediocin under GIT conditions. To date, only one study (Corr et al. 2007) has demonstrated the inhibition of a pathogen (Listeria monocytogenes) in direct association with bacteriocin production by a probiotic strain (Lactobacillus salivarius UCC118).
In the present study, we sought to demonstrate that three bacteriocin producers, namely nisin Z producer Lactococcus lactis subsp. lactis biovar diacetylactis UL719 (Meghrous et al. 1997), nisin A producer L. lactis ATCC 11454 (Steen et al. 1991) and pediocin PA-1 producer Ped. acidilactici UL5 (Daba et al. 1991), are able to grow and synthesize active bacteriocin in Macfarlane broth (Macfarlane et al. 1998), which mimics the composition of medium encountered the human large intestine.
Materials and methods
Bacterial strains and culture media
Pediococcus acidilactici UL5 (producer of pediocin PA-1) and Lactococcus lactis subsp. lactis biovar diacetylactis UL719 (L. lactis UL719, producer of nisin Z) were obtained from the Dairy Research Center (STELA, Université Laval) culture collection and were isolated from cheddar and raw milk cheeses, respectively. Lactococcus lactis ATCC 11454 (producer of nisin A) was obtained from the American Type Culture Collection (Rockville, MD, USA). A nonproducing mutant of Ped. acidilactici UL5, called bac(−), was used as a control (Daba et al. 1991). Listeria ivanovii HPB 28 (L. ivanovii HPB 28) was obtained from Health Protection Branch, Health and Welfare, Ottawa, ON, Canada. The lactic acid bacteria were grown in De Man–Rogosa–Sharpe (MRS) (De Man et al. 1960) broth (Oxoid, Nepean, ON, Canada), while L. ivanovii HPB 28 was grown in tryptic soy broth supplemented with 0·6% (w/v) yeast extract (TSBYE) obtained from Difco Laboratories, Sparks, MD, USA. Fermentations were conducted using MRS broth or Macfarlane broth.
The broth described by Macfarlane et al. (1998) was used, but with a modified bile salt concentration. It contained the following constituents in grams per litre of distilled water: pectin (citrus) 2·0, oat spelt xylan 2·0, larchwood arabinogalactan 2·0, guar gum 1·0, inulin 1·0, potato starch 5·0, porcine gastric type III mucin 4·0, casein hydrolysate 3·0, peptone 5·0, tryptone 5·0, yeast extract 4·5, bile salt no. 3 0·05, KH2PO4 0·5, NaHCO3 1·5, NaCl 4·5, KCl 4·5, MgSO4·7H2O 1·25, CaCl2·2H2O 0·1, MnCl2·4H2O 0·2, FeSO4·7H2O 0·005, haemin 0·05, l-cysteine HCl 0·8, Tween 80 1·0, 0·5 ml of vitamin solution (Gibson and Wang 1994). The vitamin solution contained the following compounds in milligram per litre of distilled water: pyridoxine-HCl 20·0, 4-aminobenzoic acid 10·0, nicotinic acid 10·0, biotin 4·0, folic acid 4·0, cyanocobalamin 1·0, thiamine 8·0, riboflavin 10·0, phylloquinone 0·005, menadione 2·0 and pantothenate 20·0. It was filter-sterilized using membrane of 0·2 μm pore size and added to the autoclaved cooled medium.
Growth conditions and bacteriocin production
Bacteriocin producers were grown in MRS broth or Macfarlane. Five hundred millilitres of the two tested media was inoculated at 1% with an overnight (18 h) culture grown in MRS broth at 30°C and then incubated in an anaerobic chamber (Forma scientific anaerobic system Model 1025; Forma Scientific, Marietta, OH, USA) at 37°C for 24 h without agitation. Samples were taken every 1·5 h for the first 12 h and every 2 h during the last 12 h. For each sample, viable bacterial count, pH, residual sugar concentrations, organic acid production and bacteriocin production were determined as described below. An overnight culture of each bacteriocin producer was centrifuged at 5000 g for 5 min at 4°C and pellets were re-suspended in MRS medium for co-culture experiments. Both lactic acid bacteria (LAB) and L. ivanovii HPB 28 were inoculated at 1% and grown for 24 h at 37°C under anaerobic conditions. During fermentation, samples were withdrawn aseptically every 2 h for analysis of cell growth, organic acids, sugars, pH and bacteriocin activity.
Bacterial counts were determined by plating tenfold serial dilutions of culture on selective media using the drop plate method (Herigstad et al. 2001). Pediococcus was counted using Pediococcus Selective Medium (PSM + A) (Simpson et al. 2006) after incubation at 37°C for 2 days in an anaerobic chamber. Lactococcus was enumerated on MRS agar after aerobic incubation at 37°C for 24 h. Listeria was counted using PALCAM agar (van Netten et al. 1989) with 48 h of aerobic incubation at 37°C.
Determination of bacteriocin activity
Preparation of bacteriocin samples
One millilitre of culture was centrifuged at 10 000 g for 10 min. Supernatant was filtered through cellulose acetate membrane with 0·45-μm pore size (VWR, Mississauga, ON, Canada) and then heated for 10 min at 100°C. The samples were then stored at −80°C for subsequent use.
Agar diffusion assay
Bacteriocin activity was prescreened visually using the agar well-diffusion method (Wolf and Gibbons 1996). Briefly, MRS broth or TSBYE containing 0·75% (w/v) agar was cooled to 47°C and seeded with 1% (v/v) of an overnight culture of indicator strain (≈1 × 109 CFU ml−1), namely Ped. acidilactici UL5 for nisin and L. ivanovii HPB 28 for pediocin. The seeded agar (25 ml) was then poured into a sterile Petri dish at room temperature. Wells 7 mm in diameter were cut in the solidified agar using the wide end of a sterile 5-ml glass pipette and were filled with 80 μl of culture supernatant neutralized to pH 7 with 1 N NaOH solution. The plates were incubated at 30°C for 18 h, after which the inhibition zone diameters were determined.
Bacteriocin activity was determined using a previously described microtitration method (Daba et al. 1994). Briefly, twofold serial dilutions of sample (125 μl) with the pH adjusted to 7·0 with 5 mol l−1 NaOH were made in the wells of a flat-bottomed microtest™ polystyrene microtitration plate (96-well microtest; Becton Dickinson Labware, Franklin Lakes, NJ, USA). Each well was inoculated with 50 μl of 1000-fold diluted overnight culture to reach a final concentration of approximately 106 CFUml−1. Pediococcus acidilactici UL5 was used as the indicator strain for nisin, and L. ivanovii HPB 28 was used for pediocin. Microtitration plates were incubated at 30°C for 18 h, and absorbance at 650 nm was read using a Thermo-max molecular device spectrophotometer (OPTI-Resources, Quebec, QC, Canada). Bacteriocin activity was defined as the highest bacteriocin dilution showing complete inhibition of the indicator strain (absorbance equal to that in noninoculated medium), calculated as 2n (1000/125) where n is the number of wells showing inhibition of the indicator strain and expressed in arbitrary units per millilitre (AU ml−1).
Determination of organic acids, sugars and pH
Fermentation broth was centrifuged (10 000 g for 10 min at 4°C), and supernatant stored at −80°C until analysis. Samples were diluted 1 : 5 in the mobile phase (H2SO4, 0·0065 N), filtered on hydrophilic polyvinylidene difluoride membrane of 0·2-μm pore size (Pall, Ann Arbor, MI, USA) and sealed immediately. Glucose and lactate concentrations were determined by HPLC (Waters, Milford, MA, USA) equipped with an Ion 300 column (Transgenomic, San Jose, CA, USA), differential refractometer (Model R410; Waters) and Agilent ChemStation Version 6.2 using the method described by Doyon et al. (1991). The analysis was performed at a flow rate of 0·4 ml min−1 at 37°C with an injection volume of 100 μl. Each analysis was performed in duplicate. The mean metabolite concentrations were expressed in mmol l−1. Sample pH was measured and adjusted with a 1 N NaOH solution to 7·0 using a pH meter (VWR symphony model SP70P) for bacteriocin determination.
Growth of the three bacteriocin producers in MRS broth and Macfarlane is shown in Fig. 1. Growth of Lactococcus lactis UL719 was similar in MRS and Macfarlane broths during the first 16 h, with a maximum specific growth rate (μmax) of 0·54 and 0·55 h−1, respectively (Fig. 1a), and increases of approximately 1·6 log CFU ml−1 after 3 h. After 16 h, a significant decline was observed in MRS broth but not in Macfarlane broth. For Pediococcus acidilactici UL5, growth was maximal in MRS medium, with μmax of 0·48 h−1 and an increase of 2 log cycles after 6 h and stable numbers thereafter (Fig. 1b). In Macfarlane broth, growth was slightly slower, with μmax of 0·43 h−1 and an increase of 1·2 log CFU ml−1 during the first 10 h. Growth of L. lactis ATCC 11454 was maximal in Macfarlane broth, with μmax of 0·45 h−1 and an increase of 1·7 log CFU ml−1 in 6 h, compared to 0·41 h−1 in MRS broth (Fig. 1c).
Bacteriocin production was visualized using the agar well-diffusion assay and quantized using a microtitration method. As shown by diffusion (Fig. 2), bacteriocin production was greater in MRS medium regardless of bacterial strain, with inhibition zone diameters of about 2 cm. Some bacteriocin production was observed in Macfarlane broth for all three strains, with L. lactis UL719 being the most active. The greatest production (8192 AU ml−1) was obtained from Ped. acidilactici UL5 and L. lactis ATCC 11454 after 7·5 and 10·5 h, respectively, in MRS broth, while L. lactis UL719 produced 4096 AU ml−1 within 6 h in this medium (Fig. 3a). In Macfarlane broth (Fig. 3b), L. lactis UL719 was the best producer, reaching 256 AU ml−1 after 4·5 h. In comparison, Ped. acidilactici UL5 and L. lactis ATCC 11454 did not reach 128 AU ml−1 until 7·5 and 4·5 h of growth, respectively.
Organic acids production
The lactic acid production and glucose consumption during the fermentation are shown in Table 1. The three bacteriocin-producing strains produced large amounts of lactic acid in MRS broth. None of these strains produced any organic acid in Macfarlane broth. The pH of Macfarlane broth only decreased from 6·30 to 5·90 during the fermentation, while it decreased from 6·00 to 4·40, 3·80 and 4·60 in MRS for L. lactis UL719, Ped. acidilactici UL5 and L. lactis ATCC 11454, respectively (data not shown).
Table 1. Glucose consumption and organic acid production by Lactococcus lactis subsp lactis biovar diacetylactis UL719, Pediococcus acidilactici UL5 and L. lactis ATCC 11454 in De Man–Rogosa–Sharpe (MRS) broth and Macfarlane broth, as determined by HPLC. Each value is the mean of two independent repetitions ± standard errors
Glucose (mmol l−1)
Lactate (mmol l−1)
L. lactis UL719
95·3 ± 4·4
67·4 ± 0·7
2·6 ± 0·0
76·3 ± 0·3
0·4 ± 0·0
0·1 ± 0·0
6·3 ± 0·4
8·3 ± 0·4
Ped. acidilactici UL5
99·5 ± 0·1
51·5 ± 1·0
2·7 ± 0·1
109·3 ± 2·7
0·4 ± 0·0
0·1 ± 0·0
6·9 ± 1·0
8·0 ± 0·0
L. lactis ATCC 11454
93·1 ± 3·9
69·4 ± 0·5
1·7 ± 0·9
74·6 ± 1·6
0·3 ± 0·0
0·2 ± 0·1
6·6 ± 0·8
8·0 ± 0·8
Effect of bacteriocin producers on Listeria ivanovii HPB 28 in co-culture
Figure 4 shows the growth of L. ivanovii HPB 28 in co-culture with bacteriocin producers in Macfarlane broth. In the presence of Ped. acidilactici UL5, L. ivanovii HPB 28 counts remained stable for 4 h and then dropped by more than 2 log cycles (9·3 106 CFU ml−1 to 3·8 104 CFU ml−1) after 6 h (Fig. 5a). A positive correlation was observed between the decrease in L. ivanovii HPB 28 numbers, growth of Ped. acidilactici UL5 and pediocin PA-1 production. These results were confirmed using the bac(−) mutant of Ped. acidilactici UL5, which was unable to inhibit L. ivanovii HPB 28, as shown in Fig. 4. The growth curve was similar to that obtained for pure culture of L. ivanovii HPB 28. Similar results were observed in the presence of nisin Z producer L. lactis UL719, namely a >2 log cycles drop in listerial counts after only 4 h and a nearly 5 log cycles drop after 24 h (Fig. 4). As indicated in Fig. 5(c), the decrease in L. ivanovii HPB 28 counts coincides with the onset of nisin Z production. L. lactis UL719 was able to reach 5·5 108 CFU ml−1 in Macfarlane broth within 4 h and to produce nisin Z up to a concentration of 512 AU ml−1. Listeria ivanovii HPB 28 counts decreased by nearly 1 log cycle after 6 h and by nearly 3 log cycles after 24 h of co-culture with L. lactis ATCC 11454 in Macfarlane broth (Fig. 4). Lactococcus lactis ATCC 11454 reached 3·2 108 CFUml−1 after 4 h and produced nisin A at a concentration of 192 AU ml−1 within 6 h, which likely caused the observed decrease in L. ivanovii HPB 28 counts (Fig. 5d).
The principal goal of this study was to evaluate the capacity of lactic acid bacteria to grow and produce bacteriocins in media of composition similar to that encountered in the human large intestine, namely Macfarlane broth, which may represent the soluble matter present in the proximal portions of the large intestine. Three bacteriocin producers were studied, namely Lactococcus lactis UL719, L. lactis ATCC 11454 and Pediococcus acidilactici UL5, which produce, respectively, nisin Z, nisin A and pediocin PA-1.
Different patterns of growth and bacteriocin synthesis were obtained for the bacteriocin producers. Lactococcus lactis UL719 grew identically in MRS broth and Macfarlane broth, while Ped. acidilactici UL5 grew better in MRS broth and L. lactis ATCC 11454 grew better in Macfarlane broth. We also showed that the three strains produced their respective bacteriocins at different levels in this medium. This could be explained by the lack of simple carbohydrates in Macfarlane broth. Guerra and Pastrana (Guerra and Pastrana 2002) showed that nisin and pediocin production depend on proper carbon/nitrogen (C/N) ratios in the medium. The composition of Macfarlane broth is not optimal for bacteriocin production. Acidification of the medium was also very low, organic acid production requires readily available hexose carbohydrates such as lactose or glucose (Mattey 1992), and Macfarlane broth does not contain any, and because the medium is buffered by high protein concentration.
We showed that Listeria ivanovii HPB 28 was able to grow alone in Macfarlane broth. However, in co-culture with Ped. acidilactici UL5, its numbers dropped as pediocin PA-1 production peaked at 32 AU ml−1. In contrast, L. ivanovii HPB 28 was not inhibited in co-culture with the bac(−) Ped. acidilactici UL5 mutant. As both strains were able to grow in Macfarlane broth, we can conclude that the inhibition of L. ivanovii HPB 28 was due to pediocin PA-1 production. Similarly, inhibition of L. ivanovii HPB 28 in co-culture with L. lactis UL719 was correlated with nisin Z production. Even though no bac(−) mutant was used, this inhibition was clearly due to nisin Z production. Similar results were obtained for co-culture with L. lactis ATCC 11454.
This study is the first to show the involvement of bacteriocin production in the inhibitory activity of lactic acid bacteria in a medium of a nutrient composition resembling the contents of the large intestine. The bacteria seem to remain stable and active under these conditions. This is in agreement with one of our previous studies in which we showed that even though bacteriocin was partially degraded in the small intestine, the gavage of balb/C mice with high concentrations of pure pediocin PA-1 resulted in a significant lower faecal load of Listeria as well as significantly lower translocation to the liver and spleen (Dabour et al. 2009). Moreover, Corr et al. (2007) used Lactobacillus salivarius UCC118 (a probiotic strain of human origin and producer of bacteriocin ABP-118) to protect mice infected with Listeria monocytogenes, proving that bacteriocin production inhibits this pathogen in vivo.
Ingesting bacteriocin-producing strains is preferable to ingesting purified bacteriocin, because the latter is likely to be degraded in the small intestine by proteases. We have previously demonstrated this in the case of pediocin using the TNO dynamic model of the stomach and small intestine (Kheadr et al. 2010). To overcome this problem, bacteriocin concentrations higher than the minimal inhibitory concentration (MIC) values or protection by gastrointestinal resistant encapsulation materials is necessary. We believe that the use of the producer strain itself offers several advantages over the use of purified bacteriocins. The anti-pathogenic effect is prolonged, particularly when the strain is able to colonize and grow in the colonic environment. Moreover, the inhibitory activity may involve other mechanisms of action such as competition and exclusion.
We have shown that L. ivanovii HPB 28 is inhibited in co-culture with any of our bacteriocin producers in Macfarlane broth and that this was not due to organic acid production; indeed, this medium does not contain the simple carbohydrates necessary for organic acid production. We also observed a correlation between the increase in bacteriocin production and the decrease in Listeria counts. We can therefore conclude that our bacteriocin producers were able to inhibit L. ivanovii HPB 28 growth in Macfarlane colonic fermentation medium by means of their bacteriocin production.
This work was supported by the National Science and Engineering Research Council of Canada (NSERC), the Fonds de recherche du Québec – Nature et technologies (FQRNT).