نوع مقاله : پژوهشی- انگلیسی
نویسندگان
سازمان پژوهش های علمی و صنعتی ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Introduction: In the present study, the inhibitory effect of nisin- producing Lactococcus lactis during co- culture and pure standard nisin were assessed against selected foodborne pathogenes in growth medium and Iranian UF Feta cheese. In comparison L lactis, not only proves flavor but also plays a better role in microbial quality of Iranian UF Feta cheese as a model of fermented dairy products.
Materials and methods: L. lactis subsp. lactis as nisin producer strain, Listeria monocytogenes, Escherichia coli and Staphylococcus aureus as pathogenic strains were inoculated in Ultra- Filtered Feta cheese. Growth curve of bacterial strains were studied by colony count method in growth medium and UF Feta cheese separately and during co- culture with L. lactis. Nisin production was determined by agar diffusion assay method against susceptible test strain and confirmed byRP- HPLC analysis method.
Results: Counts of L. monocytogenes decreased in cheese sample containing L. lactis and standard nisin, to 103 CFU/g after 7 days and it reached to undetectable level within 2 weeks. S. aureus counts remained at its initial number, 105 CFU/g, after 7 days then decreased to 104 CFU/g on day 14 and it was not detectable on day 28. E. coli numbers increased in both treatments after 7 days and then decreased to 104 CFU/g after 28 days. Despite the increasing number of E. coli in growth medium containing nisin, due to the synergistic effect of nisin and other metabolites produced by Lactococcus lactis and starter cultures, the number of E. coli decreased with slow rate.
Discussion and conclusion: The results showed, L. monocytogenes was inhibited by L. lactis before entering the logarithmic phase during co- culture. S. aureus was also inhibited during co- culture, but it showed less sensitivity in comparison with L. monocytogenes. However, the number of E. coli remained steady in co- culture with L. lactis. Also, we found that, in all cheese samples, E. coli decreased slowly after 28 days which may be due to the synergistic inhibitory effects of nisin and other metabolites produced by L. lactis and starter culture strains. These conditions are compatible to UF Feta cheese making processes. The usage of L. lactis is more effective in terms of pathogenic inhibitory in comparison withfree nisin. Using L. lactis as an adjunct starter culture can assist microbial quality improvement and prevent important pathogens, which may survive during food processing, because of the production of beneficial metabolites.
کلیدواژهها [English]
Introduction
Food protection against spoilage and pathogenic bacteria has always been a high priority for the public health. In addition, the unhealthy effects of processed food and the disadvantages of chemical food preservatives, encourage researchers to innovate new strategies for food safety (1). They also struggle to cope with the increasing demands of consumers for high quality nutritious food with improved organoleptic characteristics. Biopreservation, the use of microorganisms or their products for preserving food, is a new approach to food safety. Recent studies demonstrate that certain Lactic acid bacteria (LAB) are stabile to use as biopreservative. They can not only compete well for nutrients with pathogens, but also produce antimicrobial active metabolites. Moreover, they are harmless to humans (2 and 3).
LAB can improve quality of fermented products by inhibiting the pathogens, providing hygienic safety, extending shelf life and enhancing sensory properties (4). Food protection effect of LAB on pathogenic and spoilage bacteria is through the production of substances such as, organic acids, hydrogen peroxide, diacetyl, acetoin and also bacteriocins, which can act as bio- preservative (4 and 5) . Use of LAB especially bacteriocinogenic strains in the form adjunct culture is an appropriate substitute for preservatives in food fermentation processes to prevent and combat undesirable bacteria (2, 6, 7 and 8).
L. lactis is one of the most important members of LAB. L. lactis strains produce nisin, which has a broad spectrum inhibitory effect on Gram- positive bacteria (9). Several studies have shown that nisin could inhibit various Gram- positive bacteria such as L. monocytogenes in food (9 and 10). Among dairy products, cheeses especially soft ones such as UF Feta cheese are good media to transfer pathogens to human. Pathogenic bacteria such as L. monocytogenes, E. coli and S. aureus can survive during manufacturing soft cheeses such as Feta (11). A number of studies have reported the effects of nisin and nisin producing strain on the pathogenic bacteria in different kinds of cheese such as, Cheddar, Camembert and Gouda (6, 9 and 12). El- Gazzar et al. assessed antagonistic effect between L. monocytogenes and Lactococci during fermentation of products from ultrafiltered skim milk (14). Based on their results, more inactivation was observed in permeate than skim milk and retentate. However, no published reports are available that evaluate the effects of nisin and L. lactis on pathogenic bacteria on UF Feta cheese. In manufacturing Feta cheese, removal of water from milk by ultrafilteration leads to significant increase in concentration of microbial contaminants (13 and 14). The main purpose of this study was to investigate one industrial applicable method by adding one of safe known lactic acid bacteria (L. lactis) in starter culture of ultra- filtered condensed milk to produced fermented dairy products. Therefore, control and decrease of any bacterial pollution even Gram negative bacteria such as E. coli caused by inappropriate pasteurization or post contamination, in UF cheese by adding safe bio- preservative such as nisin or nisinogen strain (L. lactis), is great importance.
Materials and methods
Bacterial strains and media
All bacterial strains were obtained from the Persian Type Culture Collection (PTCC), Iran. L. lactis subsp. lactis PTCC 1336 was used as a nisin producer strain. L. monocytogenes PTCC 1301, E. coli PTCC 1399 and S. aureus PTCC 1431 were used as pathogenic strain bacteria. All stock cultures were maintained at- 80°C in skim milk and 20% glycerol. Working cultures of L. lactis were kept on de Man, Rogosa and Sharp (MRS) medium (HiMedia Laboratory, India) and pathogenic bacteria were maintained on BHI agar (Oxoid, Hampshire, England) at 4°C. Micrococcus luteus PTCC 1169 was used as an indicator strain for nisin bioassay and was maintained like pathogenic bacteria.
Growth curve study of L. lactis
Growth curve and pH profile of L. lactis were plotted. L. lactis growth was measured by colony count in MRS and M17 broth. The culture (24 h) of L. lactis was inoculated into the growth media to give a final concentration of 10 7 CFU/ml. They were incubated in rotary shaker incubator at 30°C, 100 rpm and the samples were taken every 4 hours. Colony count was carried out by pour plate method on MRS and M17 agar and incubation at 30°C for 48 h. Nisin production in the two media was also assessed and compared.
Determination of nisin activity produced by L. lactis
Sample preparation
Samples, from culture of L. lactis, were taken every 4 hours during 48 h. Then, the samples were adjusted to pH 2 by concentrated HCl (Merck, Germany), containing 0.1% v/v of Tween 20 (Sigma Chemical Co., St. Louis, Mo.), and were heated at 90°C for 5 min. Subsequently, they were centrifuged at 12,000 × g for 10 min. The collected supernatants were then filtered through a 0.22 mm membrane filter (Millipore® Corp., Bedford, MA), and were used for nisin activity quantification by agar well diffusion bioassay method (10 and 15).
Nisin activity quantification
Nisin determination was carried out by agar diffusion method (15). BHI medium (Merck), consisting of 0.75% agar (Bacto agar, Difco) and 1% volume per volume (v/v) Tween 20 (Sigma Chemical Co., St. Louis, Mo.) was prepared. The medium was then inoculated with 1% v/v of a 24- h culture of Micrococcus luteus with an optical cell density of 1.7 at 600 nm to get approximately 108 CFU/ml of the medium. Four wells were bored on each plate using a sterilized borer and 50 μl of each supernatant sample was placed into each well in triplicate, and the fourth well was filled with blank (50 μl of 0. 02 N HCl). Before incubation at 37°C for 24 h, all plates were incubated overnight at 4°C for diffusion of nisin. The plates were examined for diameter of inhibition zones using a digital caliper (AACO, China).
In order to plot the standard curve of nisin, different concentrations of nisin were prepared. For this purpose, at first, a stock solution of nisin (1000 IU/ml) was prepared by dissolving 0.025 g of commercial nisin 106 IU/g (Sigma Chemical Co., St. Louis, Mo.) in 25 ml of sterile 0.02 N HCl. Then using 0.02 N HCl, various concentrations of standard nisin solutions (500, 400, 300, 200, 100, 50, 25, 10 and 5 IU/ml) were prepared. The standard curve was constructed by plotting diameters of inhibition zones versus the logarithm10 of nisin concentrations (16).
RP- HPLC analysis of nisin
In order to confirm nisin production, HPLC analysis was carried out. The cells of L. lactis were removed by centrifugation (10,000 × g, 30 min, 4°C) and nisin was precipitated by 30% ammonium sulfate (Merck, Germany) and kept overnight at 4°C while being stirred gently. Precipitated proteins were collected by centrifugation (12,000 × g, 30 min, 4°C), resuspended in 2 ml sterile solution of 0. 02 N HCl (Merck, Germany) and dialyzed for 24 h in Spectra/Por® no. 7 dialysis tubing (Spectrum laboratories Inc., USA, molecular weight cut off, 2000 Daltons). This fraction was concentrated by freeze- drying, resuspended in 1.5 ml of 0.02 N HCl sterile solutions, and filtered through a 0.22 μm membrane filter (Millipore® Corp. Bedford, MA).
The above preparation and a 2% standard nisin (Sigma Chemical Co., St. Louis) solution were further purified by gel filtration chromatography on Sephadex G- 25 as described by Pirad et al. (17). After the gel filtration chromatography, fractions containing nisin (determined by bioassay method) were polled and dialyzed against 0. 02 N HCl with two changes of the solution and were further concentrated by freeze- drying. The lyophilized samples were injected into a Knauer HPLC unit (Model K- 1001, Knauer, Germany) equipped with an analytical reversed phase (RP) C18 column (EurosilBioselect, Knauer, Germany) for retention time measurement (18). Samples were eluted with the mobile phase consisting of 0.1% (v/v) trifluoroacetic acid (TFA) in a mixture of water (eluent A) and acetonitrile (eluent B) (Both reagents were HPLC grade, obtained from Merck, Germany). Samples were initially eluted with 100% A for 5 min, then with a linear gradient 0 – 50% B over 45 min, followed by a linear gradient to 100% B over 5 min maintained at 100% B for 7 min. The flow rate was maintained at 1 ml/min, absorbance was monitored at 215 nm using K- 2501 UV detector (Knaure, Germany) and the column was kept at a constant temperature of 35°C. Data analysis was performed using a chromatography software package (ChromGate, version 3.1, Knaure, Germany).
Determination of inhibitory effect of L. lactis
The effect of L. lactis on growth of each test strain was evaluated separately in MRS medium (HiMedia Laboratory, India). For this purpose, overnight cultures of each pathogenic bacterium and L. lactis were inoculated into MRS broth simultaneously to give the concentration of 105- 106 CFU/ml and 106- 107 CFU/ml, respectively. They were incubated in rotary shaker incubator at 30°C, 100 rpm and the numbers of bacteria were enumerated every 4 hours. For pathogenic bacteria the colony count was determined using pour plate method on appropriate selective medium after 24 h incubation at 37°C. Listeria selective agar (Difco, USA) was used for L. monocytogenese andBaird Parker agar (Oxoid, Hampshire, England) and EMB (Difco, USA) agar were used for S. aureus and E. coli, respectively.
Minimum inhibitory concentration (MIC) of nisin
The minimum inhibitory concentration (MIC) of nisin was assessed by broth dilution method (19). For this purpose, a stock solution containing 10,000 IU/ml of nisin was prepared by dissolving 0.1 g nisin in 10 ml 0.02 N HCl. Then it was diluted serially, from 5000 to 125 IU/ml, using BHI broth.
Tubes containing different dilutions of nisin were inoculated with an overnight culture of each pathogenic bacterium (106 CFU/ml) separately and incubated at 37°C, 30°C and 8°C. The growth was inspected visually and spectrophotometrically at 600 nm after 24, 48, 72 h and 1 week of incubation. The blank was the medium alone incubated under the same conditions. The MIC was defined as the lowest concentration of nisin that inhibited the growth of bacterium after the incubation period. The MIC determination repeated independently for 3 times, always with Micrococcus luteus (a sensitive strain to nisin) as positive control.
Cheese production
UF cheese was manufactured from retentate with 1% starter culture and 1% salt (provided by Bel- Sahar, roozaneh dairy company, Qazvin, Iran). Overnight culture of each pathogenic bacterium was inoculated to retentate to give concentration of 105 to 106 CFU/ml separately. For treatment containing L. lactis, the retentate was inoculated to reach 107 to 108 CFU/ml of L. lactis. In treatment with nisin, 300 IU/ml nisin was added to retentate after the inoculation of pathogenic bacteria. For each bacterium the positive control cheese was prepared by adding each pathogen without the addition of nisin or L. lactis. Subsequently, the retentate was poured in cups and after reaching pH 4.7 to 4.8, 0.2 ml rennet (provided by Bel- Sahar, roozaneh dairy company) per 100 ml retentate was added. The cheese cups were maintained at 8°C and analyzed during a period of 30 days (20).
Microbiological analysis
The samples of cheese were taken on days 1, 7, 14 and 28. A sample (2.5 g), taken from different parts of cheese, was pooled and homogenized with 22.5 ml of sterile sodium citrate (Merck, Germany) solution (9). The cheese samples were then serially diluted in sterile 2% sodium citrate. Bacteriacounts were determined on duplicate plates of their selective media, as described earlier.
Nisin extraction and bioactivity in cheese
For extraction of nisin from cheese, acid extraction method was performed. Cheese samples (2.5 g) were mixed and homogenized with 10 ml of sterile 0. 02 N HCl and the pH was adjusted to 2 with 10 N HCl. Then samples were kept in boiled water for 5 min. After cooling and temperature adjustment to 20°C, the volume was adjusted to 12.5 ml and the samples were centrifuged at 10000 × g for 30 min at 4°C. Subsequently, the supernatant was transferred to sterile containers and maintained at 4°C for 30 min in order to solidify the fat phase. The supernatant was filtered through a 0.22 mm filter (Millipore® Corp., Bedford, MA). The pH of filtrates was adjusted to 5.5 with sterile 4 N NaOH (Merck, Germany), after which, they were kept at 4°C for 1 h before further experiments (6). Nisin activity was determined by agar diffusion method which was described earlier.
To make sure that the results are reproducible, each experiment was conducted at least two times in triplicates.
Results
Growth curve of L. lactis and nisin production
MRS and M17 are used as suitable media for lactic acid bacteria (10). In this study the growth ofL. lactis and maximum nisin production were investigated in both media. The lag and log phase of L. lactis in MRS and M17 were approximately the same but maximum growth rate in MRS was slightly better (Fig. 1. a). However, L. lactis produce more nisin in MRS (Fig. 1. b).
The maximum nisin production in MRS was 477 IU/ml, whereas in M17 it was 245 IU/ml. Moreover, pathogenic bacteria were able to grow better in MRS medium than that of M17 (data not shown). Therefore, MRS was chosen for the rest of the experiments.
Fig. 1- Growth curve of L. lactis base on colony count in MRS (dotted line) and M17 (solid line) (a), Inhibition zones of supernatant from L. lactis growth in two media: MRS and M17 (b). B: Blank; M. luteus PTCC 1169 was used as indicator strain
Fig. 2- Growth curve of L. lactis (dotted line) and curve of nisin production (solid line) (a); pH profile during L. lactis growth (b)
The number of L. lactis cellsincreased following a short lag phase after 4 h in MRS broth (Fig. 2. a). The highest absorbance was also observed between the 12th and 16th h of the growth. No nisin activity was observed at the first 4 h of L. lactis growth; however, nisin production started to escalate sharply and reached its highest level, 477 IU/ml, after 12 h parallel with the logarithmic phase of L. lactis growth. Subsequently, the growth stayed steady during the next 18 h. Nisin quantity then decreased gradually may be by increases of protease activity after 30 h. The pH value began to decline in parallel with bacterial growth and gradually reached 4.6 ± 0.2 after 48 h due to acid production from carbon sources of medium. Minimum pH value was observed after 20 h (Fig. 2b).
RP- HPLC analysis
For reconfirmation of nisin production by L. lactis, bioassay was followed by RF- HPLC. Samples of fermentation broth and standard nisin solution were eluted with the mobile phase consisting of Tri- fluoroacetic acid (TFA) in a mixture of water and acetonitrile. RP- HPLC analysis indicated that the purified peptide, produced by L. lactis and standard nisin has identical retention times (from 46.0 to 47.0 minutes) and these fractions had the same nisin activity (data not shown) (Fig. 3).
Fig. 3- RP- HPLC analysis of standard nisin and nisin purified from L. lactis. Black curve: Standard nisin; Gray curve: nisin purified from the culture medium
Inhibitory effects of L. lactis on pathogenic bacteria
L. monocytogenes was inhibited completely by L. lactis during co- culturewithin logarithmic phase in comparison with pure culture of L. monocytogenes (Fig. 4. a). In the case of S. aureus, its number started to increase in monoculture and decreased in its co- culture with L. lactis from 4.6 logs CFU/ml to 2 logs CFU/ml after 48 h (Fig. 4. b). However, it did not decrease completely before 48 h as it happened for L. monocytogenes.
In the case of E. coli, a slight inhibitory effect was observed but no reduction in its count occurred and the growth remained steady (Fig. 4. c). Number of E. coli in monoculture reached its peak after 20 h, whereas in co- culture, it maintained the same level and just showed a small change. Therefore, L. lactis had an inhibitory effect on E. coli, but it did not decrease the number of E. coli during 48 h.
Fig. 4- Growth curve of bacteria during co- culture in MRS. dotted lines (- - - - ) represent monoculture and solid lines represent co- culture in all graphs; L. monocytogenes growth curve (a), S. aureus growth curve (b) and E. coli growth curve (c)
Minimum inhibitory concentration (MIC) of nisin
MIC of nisin was compared at two different pH (6.8 and 4.7) from 24 h to 1 week. It was found that MIC of selected microorganisms was completely depended on pH and temperature. At 30°C and 37°C in culture medium (pH 6.8 to 7.2), the MIC of L. monocytogenes and S. aureus was 2000 IU/mland 3000 IU/ml, respectively. Their MIC decreased significantly at 8°C to less than 250 IU/ml. Isoelectric pH in which rennet is added to cheese is 4.7 to 4.8. MIC of L. monocytogenes and S. aureus at this pH decreased to less than 500 IU/ml at 30°C and 37°C. In the case of E. coli, no inhibitory activity was observed after 72 h at any concentration between 125 and 5000 IU/ml(data not shown). M. luteus as positive control did not grow at any of these concentrations.
Microbiological analysis of cheese
Microbiological studies of cheese illustrated that, the number of L. monocytogenes decreased from 105 CFU/g to 103 CFU/g after 7 days and subsequently no L. monocytogenes was observed from day 14 in samples containing L. Lactis. After day 7, the viable count of L. monocytogenes in controlling cheese decreased by 2 log units. On day 28, no L. monocytogenes was found in controlling cheese. The trend of reduction in cheese with nisin was the same as controlling sample, although this drop was slightly less in samples containing nisin (Table 1).
The number of S. aureus remained at 105 CFU/g after 7 days in the 2 treatments (cheese with nisin and L. Lactis) and then reduced to undetectable level after 28 days. In controlling cheese, a slight reduction was observed; however, it was not inhibited completely during the study (Table 2).
Colony count of E. coli in cheeses increased in both treatments even in controlling cheese. Number of E. coli increased to 106 CFU/g in control and cheese with L. lactis, and to 107 CFU/g in cheese containing standard nisin. It declined gradually in both treatments and even in controlling cheese then after (Table 3).
Table 1- Inhibition of L. monocytogenes in cheese containing L. lactis and nisin (mean CFU/g ± SD)
|
Cheese samples |
1st Day |
7th Day |
14th Day |
28th Day |
|
Control cheese |
8 ± 0.05Î105 |
8 ± 1.1Î103 |
1.5 ± 0.35Î103 |
ND* |
|
Cheese with L. lactis |
8± 0.08Î105 |
1.6 ± 0.14Î103 |
ND |
ND |
|
Cheese with nisin |
8± 0.12Î105 |
7.6 ± 0.35Î103 |
4.1 ± 1.26Î103 |
ND |
*ND: Not Detectable, ≤10 CFU per sample
Table 2- Inhibition of S. aureus in cheese containing L. lactis and nisin (mean CFU/g ± SD)
|
Cheese samples |
1st Day |
7th Day |
14th Day |
28th Day |
|
Control cheese |
7 ± 0.35Î105 |
7 ± 0.27 Î105 |
1.5 ± 0.17Î105 |
3.5 ± 0.35Î102 |
|
Cheese with L. lactis |
7± 0.13Î105 |
4.5 ± 0.28Î105 |
8.2 ± 0.56Î104 |
ND* |
|
Cheese with nisin |
7± 0.21Î105 |
6.5 ± 0.35Î105 |
9 ± 0.28Î104 |
ND |
*ND: Not Detectable, ≤10 CFU per sample
Table 3- Inhibition of E. coli in cheese containing L. lactis and nisin (mean CFU/g ± SD)
|
Cheese samples |
1st Day |
7th Day |
14th Day |
28th Day |
|
Control cheese |
6 ± 0.13Î105 |
7.6 ± 0.28Î106 |
5.5 ± 0.14Î106 |
4 ± 0.7Î104 |
|
Cheese with L. lactis |
6 ± 0.11Î105 |
6.2 ± 0.49Î106 |
3.5 ± 0.35Î106 |
2 ± 0.7Î104 |
|
Cheese with nisin |
6 ± 0.22Î105 |
2.1 ± 0.35Î107 |
6.6 ± 0.35Î106 |
4.2 ± 0.4Î104 |
Determination of nisin activity in cheese
In cheese containing L. lactis, no substantial amount of nisin was detected. In cheese containing nisin (second treatment) there was a dramatic reduction in the amount of nisin during the first week after cheese making and it reached 24 IU/gand 10 IU/gat day 7 and 14, respectively. Then it further decreased to 5 IU/g at day 28 and remained steady thereafter.
Discussion and conclusion
L. monocytogenes, E. coli and S. aureus are bacteria of major concern. They can easily be transmitted to human through food (7). They can survive during processing and storage of fermented milk and cheese (7 and 11). In this study the inhibitory effect of L. lactis on these foodborne pathogens was investigated in Iranian UF Feta cheese. L. lactis plays the most important role in manufacturing fermented dairy products such as sour milk, cream, butter and different varieties of cheese and has been traditionally used as starter (4). Growth curve and nisin production of L. lactis in M17 and MRS media were distinguished. Both media are almost equal for L. lactis growth. However, nisin production in MRS broth was better than M17. In our strategies, the selected media must support not only growth of L. lactis and nisin production, but also growth of other pathogenic bacteria. Thus, MRS was chosen as an appropriate medium for L. lactis and its associated culture with pathogenic bacteria.
Nisin production of L. lactis was studied by bioassay and HPLC analysis. HPLC results confirmed that the nisin produced by L. lactis PTCC 1336 has the same retention time as standard nisin. The pH profile of L. lactis fermentation in MRS broth showed that production of acid starts immediately after lag phase. L. lactis preserves fermented food by hostile effects on other organisms through the conversion of the sugars to organic acids and consequently, lowering the pH, nutrients competition and also producing bacteriocins and other compounds (4).
As the results show, L. monocytogenes was inhibited by L. lactis before entering the logarithmic phase during co- culture. S. aureus was also inhibited during co- culture, but it showed less sensitivity in comparison with L. monocytogenes. However, the number of E. coli remained steady in co- culture with L. lactis. According to a study by Benkerroum and Sandin , L. monocytogenes was extremely sensitive to nisin at low pH and it was totally inhibited before 24 h at pH lower than 5 (21). Almore et al. reported the correlation between pH reduction and inhibition of S. aureus during co- culture with L. lactis in micro- filtrate milk (22). They reported that low pH had a notable effect on the reduction of this pathogen in parallel with other factors such as bacteriocin production. In addition, various studies indicated that LAB would play remarkable role in suppression of Gram- negative bacteria through the production of short chain fatty acids such as lactic acid, and other metabolites (23). Thus, nisin production, pH reduction and other produced metabolites can inhibit these bacteria during co- culture.
Based on MIC study, nisin revealed no inhibitory effect on E. coli. This is due to the cell membrane structure of Gram- negative bacteria (23 and 24). L. monocytogenes and S. aureus were more sensitive to nisin at pH 4.7 at 8°C. These conditions are compatible to UF Feta cheese making processes. Since the MIC of these bacteria at 8°C (the storing temperature of cheese) was less than 125 IU/ml, 300 IU/ml(7.5 mg/l) was chosen for cheese bio- preservation. The maximum level of nisin usage in the US determined by FDA is 250 mg/l (25).
In the second part of the study, the inhibitory effect of L. lactis on pathogenic bacteriain UF Feta cheese was also assessed. According to our results on MIC at 8°C and production of nisin in co- culture, two cheese treatments were designed by adding L. lactis and 300 IU/g nisin. L. monocytogenes was inhibited before day 14 of storage in samples containing L. lactis. This inhibition was confirmed by the results of Maisnier- patin et al. (12) study on inhibition of L. monocytogenes in Camembert cheese containing nisinogenic starters. According to their results, a dramatic reduction happened from 6 to 24 h and the reduction continued till the end of the second week. However, Listeria was able to restart its growth after 6 weeks. In our research in UF Feta cheese, L. lactis was able to decrease the number of L. monocytogenes to undetectable level after 2 weeks. This occurred sooner than cheese containing nisin and even controlling cheese. Higher number of L. monocytogenes in cheese containing nisin might be due to a slight inhibition of starter culture by nisin (12).
In the case of S. aureus, it was inhibited in treatments containing L. lactis and nisin before 30th day of cheese storage. Hamama et al. (7) studied the stability of S. aureus in Jben, a Moroccan cheese, in the present of nisin- producing L. lactis. They found that, reduction in number of S. aureus depended on inoculum size. In lower inoculum (103 CFU/ml) it reached undetectable level earlier than higher inoculum (105 CFU/ml). The results did not show complete inhibition of S. aureus in higher inoculum size (105 CFU/ml) during their studies. In our studies, it dropped to undetectable level from 105 CFU/ml after 28 days.
The number of E. coli in all cheese samples Containing nisin increased during the first 2 weeks of storage which may be due to resistance of Gram- negative bacteria to nisin (24). In cheese containing nisin this was greater due to slight inhibition of starter cultures by nisin and E. coli continued growth one fold higher than other treatments. This is in agreement with the results of Ramsaran et al. (11) on Feta cheese. In contrast, it was observed that, in all cheese samples E. coli decreased gradually after 28 days which may be due to the inhibitory effect of metabolites of starter culture and L. lactis (6) . Reduction in the number of pathogenic bacteria was observed in all controlling cheese. This suggests the inhibitory effect of common starter cultures. Benkerroum and Sandine observed a reduction in Listeria count to control samples because of simultaneous impact of various inhibitory substances produced during the growth in cottage cheese (21).
In cheese containing L. lactis, no significant amount of nisin was detected by biological assay after one week. Study of Maisnier- Patin et al. (12) showed the maximum amount of nisin produced by nisin- producing culture on Camembert cheese was after 9 h and its concentration decreased gradually between 9 and 24 h, after which it declined dramatically during ripening. This confirms the results of our study, in which nisin is produced at logarithmic phase of growth. The biological activity of nisin in UF Feta cheese containing 300 IU/g nisin progressively declined to 1.7% during 28 days of storage. This inactivation might be induced by enzymatic system produced by starter culture.
Although starters can assist in safety of cheese, they cannot totally guarantee microbial quality of cheese. Ultrafiltration of milk brings about concentration of macromolecules and even retention and concentration of bacteria and their spores. If a pathogenic bacterium such as Listeria exists in ultra- filtered milk, it may have more time for propagation before the reduction of pH. Thus, addition of adjunct culture (L. lactis) and antibacterial substances such as nisin can enhance the microbial quality. This research showed that the usage of L. lactis is more effective in terms of pathogenic inhibitory in comparison withfree nisin. Using L. lactis as an adjunct starter culture can assist microbial quality improvement and prevent important pathogens, which may survive during food processing, because of the production of beneficial metabolites.
Acknowledgment
This research was supported by the Iranian Research Organization for Science and Technology (IROST). Authors gratefully acknowledge the generous cooperation of Bel- Sahar Dairy Company in providing retentate. We also appreciate Dr. Farzaneh Aziz Mohseni, head of Persian Type Culture Collection (PTCC), for providing us with the bacterial strains.
)