Impact of Nisin and Nisin-Producing Lactococcus lactis ssp. lactis on Clostridium tyrobutyricum and Bacterial Ecosystem of Cheese Matrices

Clostridium tyrobutyricum spores survive milk pasteurization and cause late blowing of cheeses and significant economic loss. The effectiveness of nisin-producing Lactococcus lactis ssp. lactis 32 as a protective strain for control the C. tyrobutyricum growth in Cheddar cheese slurry was compared to that of encapsulated nisin-A. The encapsulated nisin was more effective, with 1.0 log10 reductions of viable spores after one week at 30 °C and 4 °C. Spores were not detected for three weeks at 4 °C in cheese slurry made with 1.3% salt, or during week 2 with 2% salt. Gas production was observed after one week at 30 °C only in the control slurry made with 1.3% salt. In slurry made with the protective strain, the reduction in C. tyrobutyricum count was 0.6 log10 in the second week at 4 °C with both salt concentration. At 4 °C, nisin production started in week 2 and reached 97 µg/g after four weeks. Metabarcoding analysis targeting the sequencing of 16S rRNA revealed that the genus Lactococcus dominated for four weeks at 4 °C. In cheese slurry made with 2% salt, the relative abundance of the genus Clostridium decreased significantly in the presence of nisin or the protective strain. The results indicated that both strategies are able to control the growth of Clostridium development in Cheddar cheese slurries.


Introduction
Clostridium (C.) tyrobutyricum is a Gram-positive, obligately anaerobic endosporeforming bacterium. In semi-hard and hard cheeses, it is capable of metabolizing lactic acid and producing butyric acid, hydrogen and carbon dioxide, which leads to major defects such as late blowing, swelling and splitting of blocks, brittle texture and foul odor [1,2].
Since the cost of bringing cheese to this stage of ripening is considerable, these defects cause significant economic losses [3,4]. Different approaches have been used to eliminate C. tyrobutricum from value-added food products, especially cheese. These include physical methods like bactofugation or microfiltration of milk. Such methods remain expensive in addition to causing significant modification of milk composition. Moreover, microfiltration should be applied to skim milk since fat globules clog filters [5,6]. Antimicrobial substances such as nitrate, lysozyme or salt have been used to prevent growth of C. tyrobutyricum [7,8]. The problem with nitrate is its conversion to nitrite by milk bacteria that express xanthine oxidase [9], possibly leading to the formation of nitrosamine, a carcinogen [7]. Also, such methods are not permitted for the Protected Designation of Origin (PDO) products [10]. The problem with lysozyme is that it comes from hen's egg,

Bacterial Cultures
Lactococcus lactis ssp. lactis CUC-H and Lactococcus lactis ssp. cremoris CUC 222 were obtained from Agriculture and Agri-Food Canada, St-Hyacinthe Research and Development Center (used as a starter). The nisin-A producing Lactococcus lactis ssp. lactis 32 strain (the only nisin-A producing strain) was obtained from Laval university culture collection. All LAB strains were activated for 16 h in De Man, Rogosa and Sharpe (MRS), Nutri Bact, Canada, broth medium at 37 • C and transferred once before use. The Clostridium tyrobutyricum ATCC 25755 was activated in Reinforced Clostridial Medium (RCM), Nutri Bact, Canada, under anaerobic conditions at 37 • C and transferred once [24].

Spore Production
Spores were prepared according to a published method [25] with the following modifications: RCM broth was inoculated with culture (1% v/v) and incubated for two weeks at 37 • C in an anaerobic chamber (Forma Scientific, Inc., Ohio, OH, USA). The spores were collected by centrifugation at 5000× g at 4 • C for 10 min and resuspended in sterilized reconstituted skim milk, Nutri Bact, Canada, (12% w/v). The spore suspension was heated at 85 • C for 5 min to kill the vegitative cells. Spores were enumerated using RCM agar and incubated under anarobic condition at 37 • C [24].

Pearce Test
To select the best ratio between Lactococcus strains the Pearce test was done. The Lactococcus strains were grown in MRS medium for 16 h at 30 • C. Cultures were centrifuged at 5000× g for 5 min and the pellets were washed twice with 0.85% saline solution. All strains were standardized to an optical density of 0.2 at 600 nm using 0.15% peptone water, Nutri Bact, Canada. Sterilized reconstituted skim milk (40 mL, 12% w/v) was inoculated with standardized strains suspension at different ratios and incubated on a Cheddar cheese temperature cycle [26,27]. Samples were taken aseptically at 0 h, 2 h and 5 h for the determination of pH, bacterial count and bacteriocin production.

Nisin Encapsulation
Nisin in a commercial form (Nisaplin ® , 2.5% nisin, Cayman Chemicals, Ann Arbor, MI, USA) was purified on a Sep-Pack C18 column at 4 • C at a flow rate of 3 mL/min and concentrated with Speed-Vac for 6 h at 50 • C Its concentration was measured by Highperformance liquid chromatography (HPLC) [24]. It was then microencapsulated in 1% sodium alginate (Sigma Aldrich, Oakville, ON, Canada) with 0.5% non-gelatinized resistant starch (Manitoba Starch products Inc, Carberry, MB, Canada) as described previously [26] using an Inotech IE-20 encapsulator (Inotech Biosystems International, Inc., Rockville, MD, USA) with a 300 µm nozzle and syringe pump set at a flow rate of 2.91 mL/min to form droplets stabilized on contact with sterile 5% calcium chloride solution at pH 6.00. The encapsulator vibration frequency was set at 2000 Hz. The resulting beads were stirred at 300 rpm for 4 h at room temperature. The encapsulation efficiency was determined as described previously. Microcapsules were added to milk to obtain a nisin concentration of 50 µg/mL [24].

Cheese Slurry Preparation and Experimental Treatments
The cheese slurry was prepared as described previously [26]. Pasteurized milk was acidified with glucono-δ-lactone (Sigma, St-Louis, MO, USA). The unsalted fresh curd was lyophilized, then reduced to powder in a Quadro ComIL disc grinder (Waterloo, ON, Canada) and irradiated at 5 kGy with Cobalt 60 for 2 h. The pH was adjusted with concentrated lactic acid. The experimental formulations are shown in Table 1. Each slurry was divided into three groups of 100 g as follows: Group 1: Cheddar cheese starter (8 log 10 cfu/g) + C. tyrobutyricum (4 log 10 cfu/g) Group 2: Cheddar cheese starter (8 log 10 cfu/g) + C. tyrobutyricum (4 log 10 cfu/g) + encapsulated nisin Group 3: Cheddar cheese starter (8 log 10 cfu/g) + C. tyrobutyricum (4 log 10 cfu/g) + L. lactis ssp. lactis 32 (nisin A producer) The slurry was incubated for four weeks at 4 • C or for two weeks at 30 • C to accelerate the reactions to mimic the condition after one year. Chemical composition, bacterial counts and free nisin concentration were measured weekly. Weighed samples (about 2.5 g) were Foods 2021, 10, 898 4 of 18 blended with 22.5 mL of peptone water in a stomacher (Seward model 400, Norfolk, UK) and diluted serially. For C. tyrobutyricum counts, a portion of each dilution was heated to 85 • C for 10 min before plating (1 mL) using RCM medium at 37 • C under anaerobic condition. All cheese analyses were carried out in triplicate. Fat and moisture contents were measured using the CEM (SMART 6 Moisture/Solids Analyzer).

Nisin Determination by HPLC
Nisin extraction from cheese was performed as described in Jarvis, (2016) with some modifications. Two grams of cheese slurry were added to 8 mL of 0.02 N HCl then homogenized for 1 min at high speed before the slurry was boiled at 98 • C for 5 min. The extract was cooled immediately to 20 • C followed by centrifugation at 10,000× g, 4 • C for 30 min. The fat layer was removed, and the supernatant was filtered through a 0.2 µm filter before being analyzed by a reverse-phase HPLC (HPLC Hewlett Packard Series 1100; Agilent technologies Canada Inc., Saint-Laurent, QC, Canada) [24].

Sample Preparation
One mL of the cheese suspension was centrifuged at 10,000× g for 20 min, and the fat layer was removed before the cell pellet was suspended in 500 mL of a TE 2X solution (20 mM Tris HCl pH 8.0, 2 mM EDTA). 1.25 µL of Propidium monoazide (PMA) solution (Biotium Inc., Hayward, CA, USA) was added to bacterial suspension. The tubes were shaken in the dark for 5 min to allow for maximal PMA contact with DNA, then all tubes were exposed to a PMA lamp apparatus under ice (halogen 500W, Ingenia Biosystems, Barcelona, Spain) for 15 min. Finally, all tubes were recentrifuged at 10,000× g for 15 min to recover cell pellets for DNA extraction and stored at −20 • C [28] as described below.

DNA Extraction from Cheese Samples
Total DNA was extracted from cheese samples according to Desfossés-Foucault et al.

16S rRNA Sequencing and Analysis
The diversity of the cheese microbiome was assessed by sequencing the bacterial 16S rRNA gene in the V3-V4 region using the amplification primers 341F (5' CCTACGGGNGG CWGCAG-3') and 805R (5'-GACTACHVGGGTATCTAATCC-3') adapted to incorporate the transposon-based Illumina Nextera adapters (Illumina, San Diego, CA, USA) and a sample barcode sequence allowing multiplexed sequencing. High-throughput sequencing was performed at the Institute for Integrative Systems Biology at Université Laval, QC, Canada on a MiSeq platform using 2× 300 bp paired-end sequencing (Illumina, San Diego, CA, USA). The 16S rRNA gene was profiled using demultiplexed raw data processed with Mothur software (v1.35.1) as described previously [27]. Sequences were aligned using the bacterial reference database SILVA with the align.seqs command. The sequences were clustered into operational taxonomic units (OTUs) using the OPTI parameter. Representative OTU sequences were assigned to taxa based on the Green genes reference database [28].

Statistical Analysis
A split-plot design was applied to determine the effect of time on bacterial growth and the effect of NaCl concentration, pH and time on Clostridium and starter strain counts. All experiments were performed three times. Effects were declared significant at p ≤ 0.05. Statistical analyses were carried out using the SAS general linear models' procedure (version 9.1.3, Cary, NC, USA).

Selection of Mixed Culture Ratio and Production of Encapsulated Nisin
All three Lactococcus lactis strains grew well in skim milk under Cheddar cheese temperature cycle conditions. Growth, final pH and bacteriocin production varied between strains (Tables 2-4). Strain L. lactis ssp. lactis CUC-H was the strongest acidifier after 5 h of incubation. Production of nisin-A by L. lactis ssp. lactis 32 started after 5 h. The most compatible combination was L. lactis spp. lactis CUC-H, L. lactis spp. cremoris 222 and L. lactis spp. lactis 32 at a ratio of 1:1:1.5 for a total inoculum volume of 2%. This provides optimal growth and maximal acid production.   Table 4. Nisin-A production by the nisin-producing strain of Lactococcus lactis in skim milk; pure culture or combination under the Cheddar cheese temperature cycle (n = 3).

Chemical and Microbiological Composition of Cheese Slurry
The slurries had similar chemical compositions, which corresponded to the normal composition of Cheddar cheese (Table 5), except for the NaCl concentration. At 2.0% NaCl, no significant difference was found between pH 5.0 and 5.3 on Lactococcus counts. Therefore, only the data of pH 5.3 will be presented. At 2% salt concentration, a slight reduction was seen in the control group at 4 • C after four weeks. On the other hand, the presence of encapsulated nisin was associated with an approximately 1 log 10 reduction until the second week. The total bacterial counts increased again during the third week (p ≤ 0.05), likely due to the gradual release of nisin from capsules. In the presence of the nisin-A producer (strain 32), counts appeared stable over time. At 1.3% NaCl, the same trend overall was observed (Figure 1).
At 30 • C, the total bacterial counts in control fell by about 0.3 log 10 , a 1.0 log 10 reduction (p ≤ 0.05) was noted in the encapsulated nisin-A group at 2% salt, while it was 0.6 log 10 at 1.3% salt, and a slight reduction appeared in the presence of the protective strain, perhaps due to better bacterial growth at this temperature ( Figure 2) at 2% salt. The gradual in situ production of nisin during storage was confirmed by quantitative HPLC. These results show that these NaCl concentrations and pH do not likely interfere with starter culture growth.   At 30 °C, the total bacterial counts in control fell by about 0.3 log10, a 1.0 log10 reduction (p ≤ 0.05) was noted in the encapsulated nisin-A group at 2% salt, while it was 0.6 log10 at 1.3% salt, and a slight reduction appeared in the presence of the protective strain, perhaps due to better bacterial growth at this temperature ( Figure 2) at 2% salt. The gradual in situ production of nisin during storage was confirmed by quantitative HPLC. These results show that these NaCl concentrations and pH do not likely interfere with starter culture growth.  At 30 °C, the total bacterial counts in control fell by about 0.3 log10, a 1.0 log10 reduction (p ≤ 0.05) was noted in the encapsulated nisin-A group at 2% salt, while it was 0.6 log10 at 1.3% salt, and a slight reduction appeared in the presence of the protective strain, perhaps due to better bacterial growth at this temperature ( Figure 2) at 2% salt. The gradual in situ production of nisin during storage was confirmed by quantitative HPLC. These results show that these NaCl concentrations and pH do not likely interfere with starter culture growth. In the case of C. tyrobutyricum counts, significant reductions were noted after the first week in the presence of encapsulated nisin and in the second week in the protective strain treated group at 1.3% and 2% NaCl at 4 °C (Figure 3). In the encapsulated nisintreated group, no spores were detected after the third week (p ≤ 0.05) in RCM medium. At 2% NaCl under 4 °C, none were detected after the second week, suggesting synergism between NaCl and nisin-A. We have also observed a 0.5 log10 reduction in C. tyrobutyricum  In the case of C. tyrobutyricum counts, significant reductions were noted after the first week in the presence of encapsulated nisin and in the second week in the protective strain treated group at 1.3% and 2% NaCl at 4 • C (Figure 3). In the encapsulated nisintreated group, no spores were detected after the third week (p ≤ 0.05) in RCM medium. At 2% NaCl under 4 • C, none were detected after the second week, suggesting synergism between NaCl and nisin-A. We have also observed a 0.5 log 10 reduction in C. tyrobutyricum counts in the protective starter treated group compared to control (p ≤ 0.05) during four weeks of incubation under two salt concentrations at 4 • C. This inhibition activity was associated with the gradual in situ production of nisin during storage, as confirmed by nisin quantification using HPLC. Gas production was observed in the control group at 1.3% NaCl at 30 • C (Figure 4). The C. tyrobutyricum count decreased by 0.74 log after one week under both salt concentrations, while the count was maintained constantly at 30 • C after first week (Table 6). In the case of C. tyrobutyricum counts, significant reductions were noted after the first week in the presence of encapsulated nisin and in the second week in the protective strain treated group at 1.3% and 2% NaCl at 4 °C (Figure 3). In the encapsulated nisintreated group, no spores were detected after the third week (p ≤ 0.05) in RCM medium. At 2% NaCl under 4 °C, none were detected after the second week, suggesting synergism between NaCl and nisin-A. We have also observed a 0.5 log10 reduction in C. tyrobutyricum counts in the protective starter treated group compared to control (p ≤ 0.05) during four weeks of incubation under two salt concentrations at 4 °C. This inhibition activity was associated with the gradual in situ production of nisin during storage, as confirmed by nisin quantification using HPLC. Gas production was observed in the control group at 1.3% NaCl at 30 °C (Figure 4). The C. tyrobutyricum count decreased by 0.74 log after one week under both salt concentrations, while the count was maintained constantly at 30 °C after first week ( Table 6).       . Gas swelling of Cheddar cheese slurry (made with 1.3% NaCl) by Clostridium tyrobutyricum after one week at 30 • C, left: control; right: containing encapsulated nisin-A. Table 6. Clostridium tyrobutyricum counts in cheese slurry at 30 • C (n = 3).

Treatment
Log cfu/g at 2% NaCl

In Situ Nisin Production
Nisin production by L. lactis ssp lactis strain 32 (protective strain) in cheese slurry was not affected by pH (Figures 5-7). Production was gradual, reaching 97.54 ± 1 µg/g at 2% NaCl and 80.92 ± 1 µg/g at 1.3% NaCl after four weeks at 4 • C ( Figure 5). A higher salt concentration could increase nisin production by putting the bacterial cells under more stress [28]. Somewhat more nisin was found at 30 • C, which could be due to the suitable temperature for bacterial growth to produce bacteriocins (Figure 7). These results showed that L. lactis ssp lactis strain 32 was able to survive and produce nisin in cheese slurry during storage at both temperatures. We have also shown that encapsulation improves the stability of nisin and controls its release in cheese during storage. At 4 • C and in the presence of 2% salt, the concentration of released nisin increased gradually within 2 weeks then decreased to about 30 µg/g at 4 • C due to the reduction of nisin amount inside capsules. The rise and the fall were both slower with 1.3% salt (Figure 6). At 30 • C, the rise appears to have occurred within 1 week, followed by the same tendency to drop (Figure 8).
during storage at both temperatures. We have also shown that encapsulation improves the stability of nisin and controls its release in cheese during storage. At 4 °C and in the presence of 2% salt, the concentration of released nisin increased gradually within 2 weeks then decreased to about 30 µg/g at 4 °C due to the reduction of nisin amount inside capsules. The rise and the fall were both slower with 1.3% salt (Figure 6). At 30 °C, the rise appears to have occurred within 1 week, followed by the same tendency to drop ( Figure  8). Figure 5. Nisin production in Cheddar cheese slurry containing bacteriocin-producing Lactococcus lactis ssp. lactis 32 (protective starter), during 4 weeks at 4 °C (n = 3), Different letters indicate significant difference (p < 0.05). Figure 5. Nisin production in Cheddar cheese slurry containing bacteriocin-producing Lactococcus lactis ssp. lactis 32 (protective starter), during 4 weeks at 4 • C (n = 3), Different letters indicate significant difference (p < 0.05).  . Nisin production in Cheddar cheese slurry containing bacteriocin-producing Lactococcus lactis ssp. lactis 32 (protective starter), during two weeks at 30 °C (n = 3), Different letters indicate significant difference (p < 0.05).   . Nisin production in Cheddar cheese slurry containing bacteriocin-producing Lactococcus lactis ssp. lactis 32 (protective starter), during two weeks at 30 °C (n = 3), Different letters indicate significant difference (p < 0.05). Figure 7. Nisin production in Cheddar cheese slurry containing bacteriocin-producing Lactococcus lactis ssp. lactis 32 (protective starter), during two weeks at 30 • C (n = 3), Different letters indicate significant difference (p < 0.05). Figure 7. Nisin production in Cheddar cheese slurry containing bacteriocin-producing Lactococcus lactis ssp. lactis 32 (protective starter), during two weeks at 30 °C (n = 3), Different letters indicate significant difference (p < 0.05). Figure 8. Nisin released in Cheddar cheese slurry containing the encapsulated bacteriocin, during two weeks at 30 °C (n = 3), Different letters indicate significant difference (p < 0.05). Figure 8. Nisin released in Cheddar cheese slurry containing the encapsulated bacteriocin, during two weeks at 30 • C (n = 3), Different letters indicate significant difference (p < 0.05).

Metagenomics Analysis
The impact of the nisin-A-producer strain and encapsulated nisin-A on the composition of Cheddar cheese slurry microbiota during storage was expressed in terms of the relative abundance of various families of bacteria. Figures 9 and 10 show the changes in the diversity of genera over time. In the protective strain treated group (pH 5.3 at 1.3% salt) incubated at 4 • C, the genus Lactococcus was dominant at time zero, then increased significantly to 94% of the total bacteria after four weeks of incubation, followed by the encapsulated nisin treated group then the control group. The greatest drop in relative abundance of Clostridium at this temperature was observed in the encapsulated nisin treated group, followed by the bacteriocin-producing strain treated group compare to the control group.
After two weeks at 30 • C and 1.3% NaCl, the relative abundance of Clostridium spp. in the control group had increased significantly compared to the other groups. The proportion of genus Lactococcus had increased slightly and was still dominant, followed by the genus Pediococcus in the control and protective strain treated groups compared to the encapsulated nisin treated group. In this case, the genus Lactococcus decreased significantly with the increasing abundance of genus Pediococcus during the second week at 30 • C and 1.3% NaCl.
During four weeks at 4 • C and 2% salt, the relative abundance of Clostridium spp. increased in the slurry adjusted to pH 5.3 and 2% salt. Meanwhile, Lactococcus declined significantly during week four at 2% salt in all groups, while the genus Pediococcus reached 85%, 64% and 55%, respectively, in the control, encapsulated nisin and nisin producer strain treated groups. At 30 • C and 2% salt, in the control group the Clostridium abundance increased during the two weeks of incubation compare to the encapsulated nisin and the nisin producing strain treated groups (p < 0.05) due to the presence of nisin-A.

Discussion
A great interest has developed around the use of lactic acid bacteria (LAB) as protective cultures for preserving food or to prolong their shelf life. The bio-protective activity of LAB is related to their ability to produce substances having biological activities at very low concentrations, including organic acids (acetic, lactic and propionic acid), hydrogen peroxide, diacetyl, acetaldehyde and antimicrobial peptides such as bacteriocins [29,30].
The present study aimed to assess the potential of bioprotective lactic acid cultures as well as their metabolites with antimicrobial activity (nisin) for the control of Clostridium tyrobutyricum in Cheddar-type cheese slurries.
The strains L. lactis spp. lactis CUC-H, 32 and L. lactis spp. cremoris 222 were selected to be used as a protective starter. To select the suitable ratio between the selected strains, the Pearce test was done. The results showed that the most compatible ratio was L. lactis spp. lactis CUC-H, L. lactis spp. cremoris 222 and lactis 32 at a ratio of 1:1:1.5 for a total inoculum volume of 2%. A combination of a nisin-producing l. lactis spp. cremoris JSI02 and l. lactis spp. lactis C2 at a total inoculation volume of 6% (4% C2 and 2% JSI02) reportedly reached a pH ≤ 5.4 in 5 h [31]. In addition, the combinations of l. lactis spp. cremoris and l. lactis spp. lactis and l. lactis spp. lactis biovar diacetylactis at a ratio 1.5:1.5:1 for a 2% total inoculum reportedly grew and produced nisin Z [21]. In all these cases, bacteriocin was produced during fermentation of milk. It remains to be shown whether bacteriocin production continues during cheese storage too.
The chemical composition of Cheddar cheese slurries and the effect of pH and salt on bacterial growth and bioactive compound production in milk has been studied [17]. Under the two NaCl concentrations, there is no significant effect of pHs on Lactococcus counts. The C. tyrobutyricum was not detected in the group treated with encapsulated nisin in the presence of 2% salt, from the second week of storage at 4 • C. The addition of the protective ferment inhibited the growth of C. tyrobutyricuim from the second week. Similar results have been reported for four clinical isolates of Clostridium difficile 2.5 log 10 cfu/mL reductions from 10 6 cfu/mL in the presence of 3.2 µg/mL nisin [32]. Clostridium beijerinckii at 2.2 × 10 5 spores/mL reportedly failed to germinate in pasteurized ovine milk inoculated with a bacteriocin-producing strain of L. lactis ssp. lactis, thus preventing late blowing even after 30 days of incubation at 23 • C [33].
The nisin production from L. lactis ssp. lactis 32 was followed up during the Cheddar slurries incubation. The production of nisin was increased gradually for all cheese groups. This is in agreement with studies of nisin Z produced by L. lactis biovar diacetylactis during ripening of Cheddar cheese, in which activity declined from over 300 units per g to less than 30 units in six months at 7 • C [21]. On the other side, the capsules maintain stability and control the nisin release in cheese during storage. The nisin released from capsules decreased after the second week. Similar results have been obtained for liposomeencapsulated nisin Z during a second month of storage at 7 • C [21]. Moreover, the nisin producing L. lactis ssp. lactis INIA 415 strain delayed the late blowing defect (LBD) by seven days during ripening at 14 • C in the cheese produced from milk contaminated with 3.1 log spores/ mL of C. tyrobutyricum CECT 4011 spores [34]. The same inhibitory effect was observed against C. tyrobutyricum CECT 4011 using the L. lactis IPLA 729 strain that produces nisin-Z in a combination with starter culture during the manufacturing of Vidiago cheese, a semi-hard farmhouse variety, produced in Asturias, Spain. The counts of C. tyrobutyricum were reduced by 3 logs 10 at nisin concentration of 1600 AU / ml compared to an increase in the control group of 3.2 logs 10 during the 30 days of storage at 12 • C [35].
The microbial ecosystem of Cheddar cheese slurries was tested, at low salt concentration (1.3% NaCl). The relative abundance of Clostridium at 4 • C dropped significantly in the encapsulated nisin treated group, followed by the nisin-producing strain treated group compare to the control group. The effect of lowered salt concentration on bacterial growth in ripening cheese has been studied. In soft cheese surface-ripened for 27 days at 12 • C, Pseudomonas fragi, there was an increase even at the concentration of 1.8% NaCl, but it was significantly lower than the growth with a concentration of 1.3% [13]. In postaging, Listeria monocytogenes contamination of Cheddar cheese, the L. monocytogenes count decreased by 0.14 to 1.48 log 10 at 0.7 and 1.8% salt concentration, respectively, during 90 days at 4 • C [36]. Artisanal cheeses collected from 11 different regions in Brazil revealed the dependence of microbial population dynamics on production conditions and subsequent processing steps [37]. Although LAB were always dominant, thermophile bacteria such as Streptococcus were found dominant in cheeses from the North, Cerrado and Serro regions. In cheeses from the Northeast, mesophiles (Leuconostoc) were more likely to dominate. Cheeses from the North also contained high counts of Lactobacillus fermentum. Salers cheese produced from raw milk by traditional methods on farms showed growing LAB counts on the first day of production, based on single-strand conformation polymorphism analysis, and was dominated by Lactococcus lactis, Streptococcus thermophilus, Lactobacillus pentosus and Lactobacillus plantarum after three months of ripening [38].
The metagenomic profile of Gouda cheese containing PMA showed that viable counts of starter culture strains (5 L. lactis cremoris, 2 L. lactis lactis and 1 Leuconostoc mesenteroides) increased until the brining step, at which point the counts dropped by 2-3 orders of magnitude, based on amplified fragment length polymorphism quantified by lineage-specific qPCR because of the differential inactivation rates. During ripening, the relative abundance of two citrate-positive L. lactis ssp. lactis, a L. lactis ssp. cremoris and a Ln. mesenteroides ssp. cremoris increased, while other strains from the same genus disappeared due to the effects of salt and low pH [39]. Furthermore, the Lactococcus lactis species showed the highest mean DNA coverage during ripening of semi-hard Maasdam-type cheeses after 12 days in a warm room or 37 days in a cold room. Finally, a study of bacterial dominance and richness in Cotija cheese using high-throughput 16S rDNA sequencing showed that the cheeses were dominated by three genera: Lactobacillus, Leuconostoc and Weissella, followed by more than 500 non-dominant genera grouped into 31 phyla of bacteria and archaea [39].

Conclusions
To study the effectiveness of biological antimicrobial formulas in cheese, three groups of Cheddar cheese slurry were designed: a control group challenged with Clostridium tyrobutyricum spores without any protection; a similar group containing a bacteriocin-producing lactic ferment and another similar group containing encapsulated nisin. All groups included slurry made at pH 5.0 or pH 5.3 and with 1.3% or 2% NaCl and stored at 4 • C or 30 • C. Clostridium tyrobutyricum was not detected in the 2% NaCl + encapsulated nisin group after two weeks at 4 • C. In the presence of the protective strain and 2% NaCl, Clostridium viable spore counts fell by 1 log after one month at 4 • C. Also, the bacteriocin was produced during storage at 4 • C. Using classical enumeration techniques, it was shown that Lactococcus was the dominant genus during storage in all treatment cases. However, 16S rRNA sequencing showed dynamics among bacteria of different genera. After four weeks at 4 • C, Pediococcus become dominant in the presence of encapsulated nisin or the protective strain and 2% salt. In the encapsulated nisin group, the genus Clostridium decreased during storage at 4 • C, but its 16S rRNA could be detected. It was not countable using classical techniques. Finally, the classical and the molecular quantitative methods both showed that the bacteriocinproducing strain and encapsulated nisin are both able to control the growth of Clostridium in Cheddar cheese slurries.