Production of a nisin Z/pediocin mixture by pH-controlled mixed-strain batch cultures in supplemented whey permeate

Authors


Dr C. Lacroix, Centre de recherche en sciences et technologie du lait STELA, Pavillon Paul Comtois Université Laval, Ste-Foy, Québec, Canada, G1K 7P4 (e-mail: christophe.lacroix@aln.ulaval.ca).

Abstract

The conditions for high production of nisin Z and pediocin during pH-controlled, mixed-strain batch cultures in a supplemented whey permeate medium with Lactococcus lactis subsp. lactis biovar. diacetylactis UL719, a nisin Z producer strain, and variant T5 of Pediococcus acidilactici UL5, a pediocin-producing strain resistant to high concentrations of nisin, were studied. Mixed cultures were performed at 37 °C and pH 5·5 by first inoculating with variant T5 and then with L. diacetylactis UL719 after 8 h incubation, and were compared with single-strain batch cultures. High productions of both nisin Z and pediocin were obtained after 18 or 16 h incubation during mixed cultures, with titres of 3000 and 730 AU ml−1, or 1060 and 1360 AU ml−1, respectively, corresponding to approximately 75 and 55, or 25 and 100 mg l−1 of pure nisin Z and pediocin, respectively. In pure cultures, nisin Z and pediocin productions were higher than in mixed cultures, and maximum activities were obtained after 10 h incubation, with approximately 10 000 AU ml−1 (250 mg l−1 pure nisin Z) and 2500 AU ml−1 (190 mg l−1 pure pediocin). During mixed cultures, significant pediocin degradation was observed in the culture supernatant fluid after 16 h incubation, together with a sharp drop in Ped. acidilactici UL5 cell viability. In the test conditions, the feasibility of producing a nisin/pediocin mixture by mixed-strain fermentation was demonstrated. The bacteriocin mixture produced in a supplemented whey permeate medium could be used as a natural food-grade biopreservative with a broad activity spectrum.

Bacteriocins are bacterial antimicrobial proteins that are active against bacteria closely related to the producing organism and, depending on the bacteriocin, against a wide range of Gram-positive bacteria ( Klaenhammer 1993; Jack et al. 1995 ). Bacteriocinogenic lactic acid bacteria (LAB) may play an important role in fermented food products because they have been shown to inhibit the growth of food spoilage or pathogenic micro-organisms ( Jack et al. 1995 ). Among bacteriocins produced by LAB, pediocin and nisin are the most studied, not only because they exhibit a broad spectrum of activity, but also because they are bactericidal at low concentrations and exhibit thermal and pH stability in foods ( Ray & Hoover 1993; De Vuyst & Vandamme 1994). Nisin, produced by Lactococcus lactis subsp. lactis, is the only bacteriocin currently approved by the Food and Drug Administration as GRAS (Generally Recognized As Safe) ( Federal Register 1988). Nisin is currently used in more than 40 countries for specific food applications ( Delves-Broughton 1990). Recently, much attention has been focused on pediocin because of its high level of activity against Listeria species ( Muriana 1996).

To inhibit pathogenic or spoilage micro-organisms, bacteriocinogenic strains or partially purified bacteriocins can be added to foods ( Muriana 1996). However, the effectiveness of bacteriocins in foods may be reduced by different factors ( Hanlin et al. 1993 ; Muriana 1996). First, the minimum inhibitory concentration (MIC) differs widely among bacteriocins and sensitive strains ( Muriana 1996). Secondly, the activity spectrum of bacteriocins produced by Gram-positive bacteria is usually limited and does not include Gram-negative bacteria. Harris et al. (1992a) also demonstrated that bacteriocin-resistant variants may appear and grow in the presence of a bacteriocin, and therefore limit its efficacy. Despite the large number of publications on bacteriocin resistance mechanisms, they are still not entirely understood ( Ming & Daeschel 1995; Maisnier-Patin & Richard 1996; Goulhen et al. 1998 ).

The use of a combination of bacteriocins has been proposed to reduce the probability of selection and the development of bacteriocin-resistant variants ( Klaenhammer 1993). Hanlin et al. (1993) demonstrated an increased bactericidal activity against undesirable micro-organisms when using a mixture of pediocin and nisin compared with a single bacteriocin. The synergistic action of mixed bacteriocins could be attributed to two effects. First, different modes of bacteriocin action might explain the different levels of sensitivity of micro-organisms to these components ( Hanlin et al. 1993 ). Thus, a variant that is resistant to the first bacteriocin in a sensitive culture could eventually be eliminated by the other bacteriocin in the mixture ( Muriana 1996). Secondly, the use of a mixture of two bacteriocins might broaden their activity spectrum as a strain that is resistant to one bacteriocin could be weakened by the other bacteriocin and then become sensitive to the first bacteriocin.

Mixed cultures of a nisin producer and a non-bacteriocinogenic, nisin-resistant strain have been used for sauerkraut production ( Harris et al. 1992b ). However, there are no published data on the production of bacteriocin mixtures by mixed-strain fermentation with two bacteriocinogenic cultures. This production would require the selection of efficient bacteriocin-producing strains that are resistant to the bacteriocin produced by the other strain of the mixed culture. Another difficulty is the development of accurate methods to quantify selectively each bacteriocin in the mixture.

Lactococcus lactis subsp. lactis biovar. diacetylactis UL719 produces a nisin derivative (nisin Z) that differs from nisin A by a single amino acid at position 27 ( Mulders et al. 1991 ; Meghrous et al. 1997 ). Although nisins A and Z have similar antimicrobial activities, nisin Z exhibits improved solubility and diffusion characteristics at pH > 6·0, which are important for foods ( De Vos et al. 1993 ). Recently, Amiali et al. (1998) reported high nisin Z production in a supplemented whey permeate medium during pH-controlled batch fermentation with this strain, with a maximum total activity of 20 500 and 4500 IU ml−1 with or without aeration, respectively. Goulhen et al. (1998) recently isolated and characterized variants of Pediococcus acidilactici UL5, a pediocin producer, which were resistant to various concentrations of nisin. The purpose of this study was therefore to demonstrate the feasibility and identify the optimum conditions for the production of a mixture of nisin Z and pediocin by batch fermentation of a supplemented whey permeate medium, using L. diacetylactis UL719 and the nisin-resistant variant T5 from Ped. acidilactici UL5.

Materials and methods

Strains

The pediocin producer strain, Pediococcus acidilactici UL5, was described by Huang et al. (1994) and Daba et al. (1994) , and was previously identified as Leuconostoc mesenteroides UL5 ( Daba et al. 1993 ). Variants of Ped. acidilactici UL5 resistant to different nisin A concentrations were isolated and characterized by Goulhen et al. (1998) . Variant T5 of Ped. acidilactici UL5 was shown to resist up to 32·5 μg nisin using the agar diffusion method. Lactococcus lactis subsp. lactis biovar. diacetylactis UL719 was previously isolated from raw milk cheese and shown to produce nisin Z ( Meghrous et al. 1997 ). Listeria ivanovii ATCC 19119 was obtained from the American Type Culture Collection (Rockville, MA, USA).

Culture medium

The fermentations were carried out in whey permeate (6%, w/v) supplemented with yeast extract (2%, w/v) and Tween-80 (0·1%, v/v), based on the optimum composition tested by Daba et al. (1993) for pediocin production, and by Amiali et al. (1998) for nisin Z production. Daba et al. (1993) reported limited growth of Ped. acidilactici UL5 in this medium because the strain cannot use lactose. Therefore, glucose (0·5%, w/v) was added to the supplemented whey permeate medium (SWPM). A solution of yeast extract was prepared by adding 40 g yeast extract (Rosell Institute Inc., Montréal, QC, Canada) to 100 ml deionized water. The yeast extract solution and Tween-80 were autoclaved for 15 min at 110 °C, then added to filter-sterilized (0·22 μm, Cameo 25 N, MSI, Westboro, MA, USA) whey permeate supplemented with glucose. The MRS medium ( De Man et al. 1960 ) was obtained from Rosell Institute Inc.

Fermentation conditions

Batch cultures were performed in a 1·5 litre BioFlo III bioreactor (New-Brunswick, Edison, NJ, USA) with a culture volume of 1 litre and agitation set at 70 rev min−1 by a wide-blade impeller. The SWPM was adjusted to pH 6 with 1 n HCl. During fermentation, pH was controlled at 5·5 by the addition of 5 n NH4OH. This pH set point was chosen to represent a compromise between the pH optima of 5·0 for pediocin production ( Daba et al. 1993 ) and 6·0 for nisin Z production ( Amiali et al. 1998 ). After 18 h incubation, the pH control was stopped to activate the maturation enzymes, which are involved in the cleavage of signal peptides at acidic pH ( Ray & Hoover 1993).

Temperature has a strong effect on cell growth and bacteriocin production ( Ray & Hoover 1993). The maximum pediocin production in MRS medium with Ped. acidilactici UL5 was observed at 37–40 °C ( Daba et al. 1993 ). Ray & Hoover (1993) reported optimal growth temperature and maximum pediocin production at 37 °C in TGE broth. On the other hand, L. diacetylactis UL719 showed similar growth profiles at 30 or 37 °C in SWPM, but nisin Z production was fourfold higher at 37 than at 30 °C after 6 h incubation (data not shown). Therefore, a temperature of 37 °C was chosen for batch fermentations with the mixed culture.

Single strain cultures with variant T5 of Ped. acidilactici UL5 and L. diacetylactis UL 719 were inoculated at 1% (v/v) with an overnight culture prepared in SWPM incubated at 30 °C. For mixed-strain cultures, variant T5 from Ped. acidilactici UL5 was first inoculated (1%, v/v) and then incubated for 8–9 h to an O.D.650 of approximately 0·5. The culture of L. diacetylactis UL719 was then inoculated at 1% (v/v) and incubated for a total of 24 h.

Two, four and three replications were performed, respectively, for batch fermentations with L. diacetylactis UL719, nisin-resistant variant T5 of Ped. acidilactici UL5 and mixed cultures with both strains. Reported data are means, and standard deviations were calculated from replicated fermentations.

Analytical methods

Cell enumeration and growth rate evaluation.

Cell growth was estimated by the pour plate method using a spiral plater (Model D, Spiral System, Cincinnati, OH, USA) and 15 cm diameter Petri dishes (Fisher Scientific, Whitby, ON, Canada). Citrate- and lactose-positive L. diacetylactis UL719 were enumerated selectively by plating on KMK medium ( Kempler & McKay 1980) after incubation at 30 °C for 48 h. The lactose- and citrate-negative cells of variant T5 of Ped. acidilactici UL5 did not grow on KMK medium under these conditions. They were enumerated on MRS agar after 48 h incubation at 47 °C; these conditions prevented the growth of L. diacetylactis UL719. The O.D.650 was also monitored hourly (LKB diode array spectrophotometer, Bromma, UK). The specific growth rate was determined from the equation of a least-square fitted line through Ln (cell counts) vs incubation time from exponential growth phase of each strain.

Determination of sugars and organic acids.

Glucose, lactose and lactic acid concentrations were determined in duplicate by HPLC analysis (Waters, Milford, CT, USA) using an Ion 300 column and an Ion polymeric guard column (Interaction Chemicals Inc., Mountainview, CA, USA). Glucose was measured by a refractive index detector and lactic acid by u.v. at 210 nm. The mobile phase was 0·0049 n H2SO4. Bacterial cells were removed by centrifugation at 10 000 g for 10 min at 4 °C. The supernatant fluids were 20-fold diluted and sterilized through a 0·45 μm pore size filter before analysis.

Bacteriocin assays.

Soluble and cell-bound nisin Z and pediocin activities in culture samples were measured in duplicate by a critical dilution microtitre assay. Bacterial cells were removed by centrifugation at 10 000 g for 10 min at 4 °C. The supernatant fluids were sterilized through a 0·22 μm pore size filter (Cameo 25 N, MSI, Westboro, MA, USA) and stored at −20 °C for the determination of soluble bacteriocin activity. To determine the cell-bound activity, the cellular pellet was resuspended in 0·02 n HCl, boiled for 10 min to release cell-bound bacteriocin and kept on ice ( White & Hurst 1968). The sample was centrifuged at 10 000 g for 10 min at 4 °C and the supernatant fluid was stored at −20 °C. Total activity was obtained by adding soluble and cell-bound activities.

The micro-method assay consisted of serial twofold dilutions of cell-free extracts in 125 μl MRS broth in a 96-well microplate (Becton Dickinson Labware, Lincoln Park, NJ, USA). Each well was inoculated with 25 μl of the indicator strain. Nisin Z activity was estimated selectively with an overnight culture diluted 1/100 in MRS broth of the indicator strain, Ped. acidilactici UL5, which is nisin Z-sensitive and immune to pediocin. Pediocin activity in pure culture of variant T5, and total pediocin and nisin Z activities in mixed culture samples, were measured with an overnight culture of indicator strain L. ivanovii ATCC 19119, a nisin- and pediocin-sensitive strain, diluted 1/1000 in MRS broth. Microplates were incubated for 22 h at 37 °C, and 12–14 h at 30 °C, for L. ivanovii ATCC 19119 and Ped. acidilactici UL5, respectively. The bacteriocin titre was defined as the reciprocal of the highest dilution causing inhibition of the indicator strain, and expressed in arbitrary units per millilitre (AU ml−1).

The specific contributions of nisin Z and pediocin to the total activity titre determined in samples from mixed cultures were estimated indirectly. For nisin Z, the titre determined with indicator strain Ped. acidilactici UL5 was used to calculate an equivalent titre with reference to the indicator strain, Listeria ivanovii ATCC 19119, using the following relationship established using samples from a pure culture with L. diacetylactis UL719 and NisaplinTM (the commercial preparation containing 2·5% nisin A; Aplin and Barrett Ltd, Dorset, UK); 1 AU nisin Z determined with L. ivanovii ATCC 19119 corresponded to 32 AU with Ped. acidilactici UL5 as indicator organism. Then, the activity of pediocin in the mixed samples was estimated by subtracting the calculated activity of nisin Z from the total activity determined with indicator strain L. ivanovii ATCC 19119.

The relationship between AU of nisin Z measured with Ped. acidilactici UL5 and international units (IU) was determined using HPLC-purified nisin Z and AmbicinTM (pure nisin A; Aplin and Barrett Ltd); 1 AU corresponded to 1 IU (40 IU = 1 μg pure nisin Z) ( Bouksaïm et al. 1998 ). The relationship between AU of pediocin measured with L. ivanovii ATCC 19119 and weight of pure bacteriocin (μg) was determined using HPLC-purified pediocin ( Daba et al. 1994 ); 13 AU corresponded to approximately 1 μg pure pediocin.

The bacteriocin titres were highly reproducible for the repeated fermentations with pure and mixed cultures, and differed by no more than one well, corresponding to a twofold dilution.

Results

Batch cultures with L. diacetylactis UL719

Maximal population during pH-controlled batch fermentation of SWPM by L. diacetylactis UL719 (1·2 × 1010 cfu ml−1) was obtained after 8–10 h incubation, and cell counts decreased to 6·0 × 109 cfu ml−1 after 24 h of incubation ( Fig. 1). Lactose consumption and lactic acid production stopped after 10 h incubation. About 54% of the initial lactose (40·7 ± 0·7 g l−1) was consumed after 24 h fermentation, with a corresponding production of lactic acid of 22·5 ± 2·7 g l−1 at the end of incubation.

Figure 1.

Mean nisin Z soluble (□, ▪) and cell-bound (bsl00016, bsl00016) activities, and viable cell counts (○, •) during pure cultures (open symbols, bsl00016) of Lactococcus diacetylactis UL719 (two repetitions) and mixed cultures (solid symbols, bsl00017) of variant T5 of Pediococcus acidilactici UL5 and L. diacetylactis UL719 (three repetitions). Nisin Z activity was determined with the microtitre assay and indicator strain, Ped. acidilactici UL5. To facilitate comparisons, time 0 h for the mixed culture corresponds to the inoculation time of L. diacetylactis UL719, which was carried out after 8 h incubation of variant T5

Soluble and cell-bound nisin Z production was determined by the microtitre assay using Ped. acidilactici UL5 as indicator strain. Soluble activity was maximum after 10 h incubation (9216 AU ml−1) and decreased progressively to 1192 AU ml−1 after 24 h incubation ( Fig. 1). Nisin Z cell-bound activity was first detected after 4 h fermentation and increased continuously to 2560 AU ml−1 after 24 h. The cell-bound nisin Z activity as a percentage of total activity increased with incubation time, corresponding to 4, 13, 24 and 68% of the total activity after 6, 12, 16 and 24 h incubation, respectively.

Batch cultures with variant T5 of Ped. acidilactici UL5

A maximum cell count of 6 × 108 cfu ml−1 of variant T5 was obtained after 12 h incubation during pH-controlled batch fermentation of SWPM ( Fig. 2). All the glucose in the medium was consumed after 18 h incubation. Apparently, glucose added at 0·5% in the SWPM was not a limiting factor for cell growth, as shown by the high residual glucose concentration during the early stationary phase (1·6 ± 0·5 g l−1 after 12 h incubation). Variant T5 did not use the lactose in SWPM and lactose concentration did not change significantly with time. Lactic acid production after 24 h was about 3·5 ± 0·3 g l−1.

Figure 2.

Mean pediocin-soluble (□, ▪) and cell-bound (bsl00016, bsl00017) activities and cell counts (○, •) during pure culture (open symbols, bsl00016) of variant T5 of Pediococcus acidilactici UL5 (four repetitions) and mixed cultures (solid symbols, bsl00017 with variant T5 and Lactococcus diacetylactis UL719 (three repetitions). Pediocin activity was determined with the microtitre assay and indicator strain, Listeria ivanovii ATCC 19119. Correction for the contribution of nisin Z to total titre determined with L. ivanovii ATCC 19119 was performed to estimate pediocin activity during mixed cultures after 8 h incubation, as described in Materials and Methods

Soluble and cell-bound pediocin activities were determined with L. ivanovii ATCC 19119 as indicator strain. Soluble pediocin activity was maximum during the early stationary growth phase (2250 AU ml−1 after 16–18 h incubation) and decreased to 1700 AU ml−1 after 24 h incubation ( Fig. 2). Cell-bound pediocin activity increased with fermentation time up to 14 h (176 AU ml−1) and remained relatively stable thereafter, representing only 4–9% of total pediocin activity ( Fig. 2).

Mixed cultures with L. diacetylactis UL719 and variant T5 of Ped. acidilactici UL5

The pH-controlled batch cultures with two bacteriocinogenic strains were carried out in two steps. First, variant T5 from Ped. acidilactici UL5 was inoculated (1%, v/v) and incubated until the O.D.650 reached 0·5, i.e. after approximately 8 h. Then, L. diacetylactis UL719 was inoculated at 1% (v/v). This procedure was adopted to promote the growth of variant T5, which was much less competitive than L. diacetylactis UL719 when both cultures where inoculated simultaneously, and consequently, to increase pediocin production in mixed cultures. Therefore, the first 8 h corresponded to a pure culture of variant T5, as is clearly shown in Fig. 2. The viable cell counts of variant T5 remained relatively stable between 8 and 16 h incubation and sharply decreased as a function of time thereafter ( Fig. 2). The growth of L. diacetylactis UL 719 inoculated at 1% after 8 h of incubation in mixed cultures was reduced compared with pure cultures; 8 h after inoculation, viable cell counts of L. diacetylactis UL719 were approximately fourfold lower in mixed cultures than in the pure cultures ( Fig. 1).

After 8 h incubation of variant T5, the residual glucose and lactic acid concentrations were 3·2±0·4 g l−1 and 1·9±0·2 g l−1, respectively ( Fig. 3). After inoculation with L. diacetylactis UL 719, the residual glucose was totally consumed after 14 h incubation, while lactose concentration did not change during the 8–14 h incubation period ( Fig. 3). At the end of the 24 h fermentation, 8·6 ± 0·4 g l−1 of lactose and 4·5 ± 0·7 g l−1 of glucose were consumed, and 12·9 ± 1·3 g l−1 of lactic acid was produced ( Fig. 3).

Figure 3.

Mean (three repetitions) of glucose (▪), lactose (•) and lactic acid (○) concentrations in mixed cultures of SWPM with variant T5 of Pediococcus acidilactici UL5 and Lactococcus diacetylactis UL719. The pH and temperature were, respectively, controlled at 5·5 and 37 °C

Total production of nisin Z and pediocin in mixed cultures was measured by the microtitre assay using indicator strain L. ivanovii ATCC 19119. Nisin Z titre was also estimated with indicator strain Ped. acidilactici UL5, sensitive to nisin and immune to pediocin. The specific contribution of nisin Z and pediocin to the total titres was calculated as described in Materials and Methods, using the correspondence for nisin Z titres determined with Ped. acidilactici UL5 and L. ivanovii ATCC 19119. During the first 8 h of incubation, pediocin production was similar in mixed and pure cultures ( Fig. 2). Soluble pediocin activity remained approximately constant between 8 and 16 h, averaging 1360 AU ml−1, and was lower than the maximum of 2515 AU ml−1 obtained after 16 h incubation of pure cultures. Cell-bound pediocin activity was similar in pure and mixed cultures. It accounted for approximately 11 and 22% of total pediocin activity after 10 and 18 h incubation for mixed cultures, compared with only 4% for pure cultures ( Fig. 2).

The nisin Z production profile was very similar in pure and mixed cultures after inoculating with L. diacetylactis UL 719, even though nisin Z titres were approximately threefold lower in mixed cultures ( Fig. 1). A maximum total nisin Z titre of 2976 AU ml−1 was observed 10 h after the inoculation of mixed cultures with L. diacetylactis UL719, corresponding to 18 h of fermentation ( Fig. 1). This value was lower than the maximum nisin Z titre of 10 176 AU ml−1 obtained after 10 h incubation of pure cultures. Cell-bound nisin Z activity increased progressively during mixed culture, and accounted for about 20% of the total nisin Z activity after 20–24 h ( Fig. 1).

Discussion

The production of a mixture of two broad-spectrum bacteriocins, nisin Z and pediocin, by pH-controlled batch fermentation of SWPM required the elucidation of strain compatibilities and their specific demands for growth and bacteriocin production. Pure cultures of L. diacetylactis UL719 and variant T5 of Ped. acidilactici UL5 were first studied using conditions from previous experiments that gave high cell growth and bacteriocin production by both strains, i.e. a pH of 5·5, a temperature of 37 °C, and a whey-based medium supplemented with 2% yeast extract, 0·5% glucose and 0·1% Tween-80. The maximum cell count for L. diacetylactis UL719 obtained after 8 h incubation under these conditions (1·2 × 1010 cfu ml−1) was about two-thirds that reported by Amiali et al. (1998) at pH 5·5 and 30 °C in a whey-based medium containing 1% yeast extract, 0·1% Tween-80 and no glucose. Moreover, Amiali et al. (1998) obtained a maximum total nisin Z titre of 4100 IU ml−1 after 8 h incubation, which remained constant until the end of the 24 h fermentation. In this study, a higher maximum for total nisin Z activity of 9216 AU ml−1 (1 IU = 1 AU) was obtained after 10 h incubation, but this titre decreased to 5461 AU ml−1 at the end of the culture, suggesting that nisin Z degradation occurred at 37 but not at 30 °C. This effect could also be partially attributed to readsorption of soluble nisin Z to the cell wall of the producer strain. The cell-bound nisin Z titre increased from 0 to 2560 AU ml−1 between 8 and 24 h of incubation ( Fig. 1); this titre was approximately fivefold higher than that reported by Amiali et al. (1998) for the same period of incubation. As shown in Fig. 1, soluble nisin Z production parallels that of biomass and thus, shows primary metabolite kinetics ( De Vuyst & Vandamme 1994; Amiali et al. 1998 ). The effect of temperature in the suboptimal range from 30 to 37 °C on cell growth and nisin Z production may be strain-specific. Matsusaki et al. (1996) reported that nisin Z production by L. lactis IO-1 at 37 °C, which is the optimal temperature for growth and lactic acid production, was only 75% of the activity obtained at 30 °C.

For mixed cultures, we used a nisin-resistant variant, T5, of Ped. acidilactici UL5, which was isolated on a nisingradient and shown to resist high concentrations of nisin A ( Goulhen et al. 1998 ). The effect of culture medium and incubation time during batch culture with Ped. acidilactici UL5, at 30 °C and without pH control, was studied by Daba et al. (1993) . Maximum biomass and pediocin production was observed after 9 h incubation in MRS broth supplemented with 2% yeast extract, giving 6·9 × 109 cfu ml−1 and 8192 AU ml−1, respectively. High biomass and pediocin production was also obtained in whey permeate supplemented with 2% yeast extract and 0·1% Tween-80, giving 7·6 × 108 cfu ml−1 and 2048 AU ml−1, respectively ( Daba et al. 1993 ). These values are very similar to those obtained in the present study of pH-controlled batch culture at 37 °C and variant T5 of Ped. acidilactici UL5; maximum biomass production and soluble pediocin titre were obtained after 10–12 h incubation and remained constant thereafter at 6 × 108 cfu ml−1 and 1900 AU ml−1, respectively ( Fig. 2). External control of pH did not appear to affect pediocin production in SWPM, which supports earlier reported data using MRS medium ( Daba et al. 1993 ). Finally, the maximum soluble titre (1900 AU ml−1) obtained between 10 and 24 h incubation in this study was very similar to the 2048 AU ml−1 reported by Daba et al. (1993) after 24 h incubation in MRS broth at the same pH of 5·5. Therefore, the conditions of batch fermentations selected in this study resulted in high pediocin production (estimated at 150 mg l−1) in SWPM.

Production of a mixture containing two broad-spectrum bacteriocins has never been reported, possibly because of the incompatibility or sensitivity of strains producing different bacteriocins, and to other problems related to data collection and interpretation. Although variant T5 was resistant to high concentrations of nisin, its competitiveness in mixed culture with L. diacetylactis UL719 was limited. These data might be explained in part by the low growth rate of variant T5 compared with L. diacetylactis UL719 during pH-controlled batch cultures (0·42 ± 0·03 h−1 and 1·23 ± 0·09 h−1, respectively), and the competition for the carbon source (glucose) in mixed cultures. Pediococcus acidilactici UL5 does not utilize lactose ( Daba et al. 1993 ) and therefore relies exclusively on glucose added to SWPM. Lactococcus diacetylactis UL719 consumed both glucose and lactose in the SWPM, leading to rapid utilization of residual glucose when this culture was inoculated after 8 h of mixed fermentation ( Fig. 3). Exhaustion of glucose was observed after 14 and 18 h incubation in mixed and pure cultures of variant T5, respectively. In order to promote cell growth and pediocin production by variant T5 in mixed culture, L. diacetylactis UL719 was inoculated 8 h after the inoculation of variant T5; this corresponded to the end of the exponential growth phase of T5 in pure culture.

In the test conditions, the feasibility of producing a nisin/ pediocin mixture was demonstrated by mixed-strain fermentations. High bacteriocin titres for both bacteriocins were obtained by a two-step fermentation with successive inoculations with variant T5 followed by, after 8 h incubation, L. diacetylactis UL719. However, the maximum total nisin Z production in mixed culture was significanty lower than in pure cultures; these maxima were 2976 and 10 176 AU ml−1, respectively, obtained 10 h after inoculation with L. diacetylactis UL719. This difference might be explained by the corresponding lower cell counts for mixed (4·2 × 109 cfu ml−1) rather than pure cultures (1·2 × 1010 cfu ml−1). In order to avoid pediocin degradation, which occurred during long incubation periods, and to produce high nisin Z titres, the fermented broth should be optimally harvested after approximately 16 h incubation ( Figs 1 and 2). Increasing the inoculation rate of L. diacetylactis UL719 might lead to a reduction in incubation time which would then enable the maximum nisin Z titre to be achieved and the fermented medium to be harvested after only 16 h fermentation ( Fig. 2). Finally, bacteriocin production did not appear to be stimulated in mixed culture as indicated by the similar mean specific pediocin (4·5 and 6·3 × 10−6 AU cfu−1 ml−1) or nisin (8·4 and 7·4 × 10−7 AU cfu−1 ml−1) production for pure and mixed cultures at incubation times corresponding to maximum titres, and considering the low accuracy of the serial twofold dilution activity test.

Maximum nisin production during pH-controlled batch cultures, obtained under the presumed optimal conditions, has been reported to be in the range 2000–3800 IU ml−1 ( De Vuyst & Vandamme 1994; Matsusaki et al. 1996 ). Recently, Amiali et al. (1998) reported a very high nisin Z production (20 500 IU ml−1 corresponding to approximately 500 mg l−1 of pure nisin) during pH-controlled, aerated batch fermentation of SWPM with L. diacetylactis UL719, with high cell-bound activity accounting for 80% of the total titre. Total nisin Z and pediocin production for mixed cultures was about 3000 AU ml−1 and 730 AU ml−1 (corresponding to 75 and 55 mg l−1 pure nisin and pediocin, respectively) after 18 h incubation, and 1063 and 1359 AU ml−1 (25 and 100 mg l−1, respectively) after 16 h incubation. Therefore, the ratio of nisin to pediocin in fermented broth was very dependent on incubation time and increased from 0·25 to 1·34 during the 16–18 h incubation period, although total bacteriocin production remained stable at approximately 125 mg l−1.

The estimation of both nisin Z and pediocin titres in broth samples from mixed cultures is a complex issue; it requires either the selection of an indicator strain that is sensitive to one of the bacteriocins in the mixture, or the specific inactivation of either bacteriocin. Both nisin Z and pediocin are small, heat-stable peptides which are relatively stable over a large pH range. Our first approach consisted of testing different enzymes at a concentration of 1 mg ml−1 for specific inactivation of one bacteriocin. Trypsin was shown to inactivate pediocin but not nisin Z after 1 h incubation at 37 °C. However, nisin Z titre could be directly estimated in culture samples by the microtitre assay using indicator strain Ped. acidilactici UL5, which is nisin-sensitive but immune to pediocin. However, conditions for specifically inactivating nisin Z by enzymes were not found. In addition, a strain that was sensitive to pediocin but resistant to nisin was not found. Listeria ivanovii ATCC 19119 and Ped. acidilactici UL5 were then selected for determining total bacteriocin and nisin Z titres, respectively, and the pediocin titre was estimated indirectly, as described in the Materials and Methods. However, because of the low sensitivity of Listeria strains to nisin compared with pediocin, the nisin Z contribution to total titres, as estimated using L. ivanovii ATCC 19119, was always low, and represented a maximum of 10% of the total titre in mixed culture. Therefore, when considering the low accuracy of the twofold dilution micro-method, the nisin Z contribution to total bacteriocin titres, as determined with L. ivanovii ATCC 19119, was within the experimental error of the method.

In conclusion, this study demonstrated the feasibility of, and identified conditions for, the production of a bacteriocin mixture in mixed-strain fermentation of SWPM by successive inoculations with variant T5 of Ped. acidilactici UL5, a nisin-resistant and pediocin-producing strain, and L. diacetylactis UL719, a nisin-producer strain. In the test conditions, high production of both pediocin and nisin Z was obtained after 16–18 h fermentation. Additional research might be required to optimize the inoculation rate of L. diacetylactis UL719 to increase the production of nisin Z while avoiding pediocin degradation, which occurs during long incubation periods exceeding 16 h. The bacteriocin mixture produced in SWPM could be used as a natural food biopreservative with a broad spectrum of activity resulting from the synergistic action of nisin Z and pediocin ( Hanlin et al. 1993 ).

Acknowledgements

This work was supported by the Conseil des Recherches en Pêches et en Agro-alimentaire du Québec (systemic program) and the Fond pour les Chercheurs et l’Avancement de la Recherche from the Government of Quebec. The authors wish to thank Aplin and Barrett Ltd for kindly supplying NisaplinTM and AmbicinTM.

Ancillary