Comparison of Activity of Commercial Protective Cultures and Thermophilic Lactic Acid Bacteria against Listeria monocytogenes: A New Perspective to Improve the Safety of Sardinian PDO Cheeses

Listeria monocytogenes contamination that occurs during and post-processing of dairy products is a serious concern for consumers, and bioprotective cultures can be applied to control the growth of the pathogen in sheep milk cheeses. However, to respect specifications provided for protected designation of origin (PDO) cheeses, only autochthonous microorganisms can be used as bioprotective cultures in these products. This in vitro study aimed to evaluate thermophilic lactic acid bacteria (LAB) isolated from sheep milk as bio-preservative agents to control L. monocytogenes growth in PDO cheese. Results were compared with those obtained with a commercial protective culture (cPC) composed of a Lactiplantibacillus plantarum bacteriocin producer designed to inhibit L. monocytogenes growth in cheese. The in vitro antilisterial activities of n.74 autochthonous LAB and a cPC were tested against 51 L. monocytogenes strains using an agar well diffusion assay. In addition, 16S rRNA sequencing of LAB isolates with antilisterial activity was conducted and strains of Lactobacillus helveticus, Lactobacillus delbrueckii subsp. indicus, Lactobacillus delbrueckii subsp. sunkii, Lactobacillus delbrueckii subsp. lactis and Enterococcus faecalis were identified. In this study, 33.6% (74/220) bacterial strains isolated from milk had characteristics compatible with thermophilic LAB, of which 17.6% (13/74) had in vitro antilisterial activity. These results demonstrate that raw sheep milk can be considered an important source of autochthonous thermophilic LAB that can be employed as protective cultures during the manufacturing of Sardinian PDO cheeses to improve their food safety. The use of bioprotective cultures should be seen as an additional procedure useful to improve cheese safety along with the correct application of good hygienic practices during manufacturing and the post-processing stages.


Introduction
Listeria monocytogenes is a Gram-positive, psychrotrophic, foodborne pathogen responsible for listeriosis. In the European Union (EU), an estimated 1600 people contract listeriosis each year, with approximately 260 fatalities [1]. This makes listeriosis one of the most serious foodborne diseases and the fifth most widespread zoonosis in the EU [2]. The severe infection is most likely to impact pregnant women and their newborns, adults aged 65 or older, and people with weakened immune systems [1]. In 2020, 1876 invasive human cases of L. monocytogenes were confirmed. The case fatality rate was as high as 13.0%, but decreased compared with 2019 and 2018. The 26% decrease in the incidence of foodborne pathogen infections detected in 2020 could be due to health interventions to prevent the transmission of SARS-CoV-2. As a result of travel restrictions, infections associated with LABs are also physiologically present in the gastro-intestinal tract of humans and animals and are an essential part of the gut microbiome [25][26][27]. These characteristics make certain LAB strains excellent candidates for use in ready-to-eat products such as cheese, and numerous studies demonstrated their effectiveness against L. monocytogenes growth [23,[28][29][30].
Commercially available protective cultures, composed of selected LABs, are considered safe and easy to use in food, but they also have some limitations. In fact, PDO regulations forbid the addition of substances or microorganisms that are not derived from the product itself [31]. Moreover, previous experiments using a mesophilic commercial protective culture (such as L. plantarum) found that the high temperatures applied during some production processing steps (45)(46)(47) • C) were more appropriate for thermophilic LABs than for mesophilic cultures [32].
Based on these data and with the aim of introducing protective cultures into the production processes of PDO cheeses, the present study aimed to isolate, identify, and select autochthonous thermophilic LABs isolated from Sardinian raw sheep milk that could potentially be used as protective cultures against L. monocytogenes in Pecorino Sardo PDO cheese. This preliminary study also compared the antimicrobial activity of one commercial protective culture to these autochthonous thermophilic LABs against L. monocytogenes and evaluated their adaptability to cheese production.

Experimental Design
An experiment was designed to compare a commercial protective culture (cPC) and autochthonous thermophilic LAB with anti-listerial activity, with the following four steps:

1.
Step one: Isolation, morphological and biochemical characterisation of LABs. Bacteria with morphological and biochemical characteristics typical of LAB and able to grow at 45 • C were isolated from Sardinian raw sheep milk; 2.
Step two: In vitro assessment of cPC and LABs antilisterial activity. The antilisterial activities of cPC and LAB cultures were tested and compared using an agar well diffusion assay (AWDA) method.

3.
Step three: Investigation of the mode of action of cPC and LABs. AWDA using cPC and LAB cell-free supernatants was conducted. 4.
Step four: Evaluation of antilisterial activity and viable count enumeration of cPC and LABs after growth in "scotta" (the residual whey from ricotta cheese production). cPC and LAB antilisterial activities were tested using scotta as growth media to investigate their adaptability to the manufacturing technology of the cheese.

Step One: Isolation, Morphological and Biochemical Characterisation of LABs
To isolate potential antilisterial LAB from Sardinian raw sheep milk, four samples of milk, each containing a milk pool from the milk transport tanks, were collected from a local cheese making plant. These raw milk pool samples were collected from several primary production holdings located in different areas of Sardinia and pooled in milk transport tanks, from which the samples were taken. All the milk samples were collected in sterilised bottles, transported to the laboratory at controlled temperature (4 ± 1 • C) and used within 2 h of sampling.
Four pools of raw milk were serially diluted from 10 −1 to 10 −4 by transferring 1 mL (1:10 dilution) to 9 mL sterilised 0.85% (w v −1 ) NaCl [33]. Serial dilutions were spread into Petri dishes containing Man Rogosa and Sharpe (MRS, Biolife, Milano, Italy) agar (pH 5.5, adjusted with 0.5 N HCl) for isolation of lactobacilli, M17 (Microbiol, Cagliari, Italy) agar with lactose 0.5% (Farmitalia Carlo Erba S.p.A., Milano, Italy) for isolation of lactococci (45 • C incubation, aerobic and anaerobic conditions) and Elliker (Sigma-Aldrich, St. Louis, Missouri, USA) agar with 0.5% beef extract (Biolife, Milano, Italy) for cultivation and enumeration of Streptococcus thermophilus [34][35][36]. All experiment stages were performed in triplicate. Statistical analysis on the set of data obtained was performed by calculating the mean and standard deviation of the numbers of colonies obtained to measure the differences of each observation from the mean; in addition, the data obtained were compared using statistical analysis of variance (ANOVA, ANalysis Of VAriance). Morphologically distinct colonies were submitted to Gram stain evaluation and catalase and oxidase tests. The Gram-positive and catalase-negative isolates were selected as presumptive LABs [37], later confirmed using 16S rRNA gene sequencing. The isolates were stored at −20 • C and −80 • C in MRS, L-M17, or B-Elliker broth containing 10% glycerol for further analysis and molecular identification.

2.3.
Step Two: In Vitro Assessment of cPC and LABs Antilisterial Activity The antilisterial activities of cPC and LAB isolates from step 1 were evaluated against 51 L. monocytogenes strains. Of these 51 strains, 47/51 (92.1%) were previously isolated from Sardinian cheese-making plants (both from dairy products and the environment) [38], while the others were reference strains (ATCC19111, NCTC1088, ATCC764, ATCC19115).
The cPC "Lyofast LPAL" (Lactiplantibacillus plantarum) manufactured by Sacco System (Cadorago, Italy) and 74 autochthonous thermophilic presumptive LABs were initially screened via agar well diffusion assay (AWDA). cPC and autochthonous LABs were inoculated in their specific growth media and incubated overnight at the previously described conditions.
For the AWDA, an overnight culture of each of the L. monocytogenes strains was inoculated into Brain Heart Infusion agar (BHA, Biolife, Milano, Italy) to obtain a final concentration of 10 5 CFU/mL and poured into Petri dishes [39]. Wells of 8 mm diameter were in the agar plates, and 50-100 µL of cPC or LAB cultures was added to each well [40,41]. The plates were then incubated at 37 • C for 24 h [42]. The presence of an inhibition zone around the wells was considered as a positive for antilisterial activity.

Step Three: Investigation of the Mode of Action of cPC and LABs
In order to distinguish between antibacterial activity related to bacteriocins or bacteriocin-like substances, an AWDA protocol was conducted. Overnight cultures of cPC and LABs identified as effective against L. monocytogenes in the previous experiment were centrifuged at 14,000× g for 5 min [43]. The cell-free supernatants (CFSs) were adjusted to pH 6.5 with 1 M NaOH in order to rule out acid inhibition [44,45] and subsequently treated with 1 mg mL −1 of catalase at 25 • C for 30 min to eliminate the possible inhibitory action of hydrogen peroxide. Plates containing semi-solid BHA previously inoculated with 10 5 CFU/mL of L. monocyotogenes [39] were prepared and AWDA was performed, inoculating wells with CFS filtered using a 0.22 µm filter (Millex SLGP033RS, Millipore, Bedford, MA, USA) [46]. Plates were examined after incubation at 37 • C for 24 h [47] and inhibition zone diameters were measured.
To confirm the protein nature of the antilisterial compound produced by cPC as described by the producer, its CFS was treated with the enzyme proteinase K. Proteinase K is a highly active protease, stable over a wide range of pHs and capable of withstanding high temperatures. To evaluate the proteolytic susceptibility of bacteriocins produced by cPC (L. plantarum), its CFS was treated at 37 • C for 2 h with proteinase K at a final concentration of 1 mg/mL. After adjusting the pH to 6.5, the samples were filtered through a 0.22 µm sterile filter (Millipore) and the antimicrobial activity was determined by the AWDA method as previously described [46].
The best-performing LAB strains, meaning those that produced the largest inhibition zone, were used for other experiments described below.

2.5.
Step Four: Evaluation of Antilisterial Activity and Viable Count Enumeration of cPC and LABs after Growth in "Scotta" The seven autochthonous thermophilic presumptive LABs that exhibited the best in vitro activity against L. monocytogenes were selected and tested, along with the cPC, for their ability to grow in scotta. Following the same procedure used to prepare the scotta-innesto (the typical starter culture) [48], a modified scotta-innesto (MSI) was prepared: cPC and LABs were suspended in scotta and incubated overnight at 37 • C for cPC and 45 • C for LABs. For cPC and LAB enumeration, 1 mL of the appropriate dilution was transferred into Petri dishes and 10-12 mL of their selective growth agar media was added. Finally, plates were incubated in anaerobic conditions for 72 h at 30 • C and 45 • C, respectively [49,50]. Data obtained were statistically analysed by calculating the mean and standard deviation and compared using statistical analysis of variance (ANOVA, ANalysis Of VAriance).
AWDA was performed as previously described with minor modifications. Briefly, BHA spiked with L. monocytogenes at a final concentration of 10 5 CFU/mL was poured into Petri dishes. Then, wells were cut in the agar and filled with the overnight cultures of scotta inoculated with cPC and LABs. These plates were incubated for 24 h at 37 • C and inhibition zones were measured [51]. All experiments were performed in triplicate.

Molecular Identification of Autochthonous Thermophilic Presumptive Lactic Acid Bacteria
Seven autochthonous thermophilic presumptive LABs that demonstrated antilisterial activity after AWDA were sent to B.M.R. Genomics (Padova, Italy) for 16S rRNA gene sequencing to identify isolates at the species level according to the protocol of Clarridge [52]. B.M.R. Genomics performed 16S gene amplification and sequencing, after which a search for the homology of the sequences was made using the BLAST algorithm [53,54] available from the NCBI server URL: http://blast.ncbi.nlm.nih.gov/Blast.cgi (accessed on 1 December 2022).

Step One: Isolation, Morphological and Biochemical Characterisation of LABs
Autochthonous thermophilic bacterial isolates were obtained from Sardinian raw sheep milk following cultivation on MRS pH 5.5, L-M17, and B-Elliker media. Tables 1 and 2 list the viable counts obtained. Table 1. Viable colony counts (log 10 CFU/mL, mean ± standard deviation) of thermophilic LABs from Sardinian raw sheep milk samples plated on MRS pH 5.5, L-M17, and B-Elliker in aerobic conditions at 45 • C.

Raw Sheep Milk Samples
Culture Media MRS pH 5.5 4 L-M17 5 B-Elliker 6 Viable colony counts in aerobic conditions in MRS pH 5.5 were significantly lower (p < 0.01) than in L-M17 and B-Elliker, while no significant difference was observed between L-M17 and B-Elliker (p > 0.05).
For bacteria grown in anaerobiosis, the situation was similar to that of bacteria grown in aerobic conditions. The bacterial count in MRS pH 5.5 was significantly lower (p < 0.01) than in L-M17 and B-Elliker.  Comparing the two growth conditions used (aerobic and anaerobic), the bacterial counts in L-M17 and B-Elliker were significantly higher (p < 0.01) in the samples grown in anaerobiosis. No significant differences were found for samples grown in MRS pH 5.5.
From sheep milk, we isolated 220 strains, and 102/220 strains demonstrated a catalasenegative reaction. The Gram stain demonstrated that 28/102 were Gram-negative and 74/102 were Gram-positive bacteria; 92/102 were classified as oxidase-negative. In addition, 49/102 isolates were rod-shaped and the remaining 53/102 were coccus-shaped. The resulting 74/220 (33.6%) Gram-positive and catalase-negative strains were evaluated for their inhibitory activity against L. monocytogenes using the agar well diffusion assay method.

Step Two: In Vitro Assessment of cPC and LAB Antilisterial Activity
After overnight growth in their specific culture media, the cPC demonstrated in vitro antilisterial activity against all 51 L. monocytogenes strains tested and produced an inhibition halo with a diameter > 2.1 cm. We found that 13/74 (17.6%) autochthonous presumptive LABs demonstrated antimicrobial activity against L. monocytogenes and, of these, 7/13 (53.8%) produced an inhibition halo with a diameter ranging from 1.1 cm to 2.0 cm, while the remaining 6 LABs had an inhibition halo with a diameter between 0.1 and 1 cm. Inhibition halos produced by cPC and LABs were significantly different (p < 0.01). Therefore, the seven LABs that demonstrated the highest efficacy against L. monocytogenes were used in the next phases.

Step Three: Investigation of cPC and LAB Modes of Action
Using CFS, cPC produced an inhibition halo with a diameter > 2.1 cm in AWDA, while the same experiment demonstrated negative results for all LAB strains. With regard to the evaluation of proteinase K's effect on the cPC bactericidal activity, no anti-listeria activity after treatment with proteinase K was found. This result confirmed the manufacturer indications that report that the cPC L. plantarum strain is a bacteriocin producer.
These results confirmed that cPC antilisterial activity is related to bacteriocin production, as declared by the manufacturer. Among the bacterial strains isolated from raw sheep milk that were positive on the AWDA, 7/13 (53.8%) demonstrated the largest inhibition zone and therefore were chosen to perform further investigation.

3.4.
Step Four: Evaluation of Antilisterial Activity and Viable Count Enumeration of cPC and LABs after Growth in "Scotta" cPC and LAB viable counts were evaluated after overnight growth in scotta in order to investigate their adaptability to dairy matrices. The same samples grown in scotta were tested for their inhibitory activity against L. monocytogenes via AWDA and demonstrated significant differences (p < 0.01). Inhibition zone diameters were evaluated and compared with data obtained from the traditional AWDA, in which cPC and LABs grown in traditional media instead of scotta were used. The halo of inhibition was significantly higher (p < 0.01) when cPC grew in the traditional growth medium than in scotta. After growth in traditional media and in scotta, inhibition halos produced by LABs demonstrated no significant differences (p > 0.05). Results are listed in Table 3. Table 3. The first column lists the analysed samples. The second column shows the viable count (log 10 CFU/mL ± standard deviation) of cPC and the 7 best-performing LABs after overnight growth in scotta. The third column shows results of AWDA: antilisterial activities in vitro (diameter of the inhibition halo) of cPC and LAB grown in traditional media and in scotta are compared.  Table 3 shows that the bacterial strains L1 MRS 1b and L2 MRS 9b grown in "scotta" demonstrated a more effective in vitro antilisterial activity compared to the same strains grown in traditional media, with a significantly larger inhibition halo diameter measured (p < 0.01). On the other hand, cPC demonstrated a significantly larger inhibition halo after growth in MRS pH 5.5 (p < 0.01). No significant differences were detected for the other strains (p > 0.05).

Molecular Identification of Autochthonous Thermophilic Presumptive Lactic Acid Bacteria
The results obtained by 16S rRNA gene sequencing confirm that the isolates were lactic acid bacteria. More specifically, the selected LABs with antilisterial activity in ADWA were identified as Lactobacillus delbrueckii subsp. sunkii, Lactobacillus delbrueckii subsp. lactis, Lactobacillus delbrueckii subsp. indicus, Lactobacillus helveticus, and Enterococcus faecalis (Table 4).

Discussion
Currently, much emphasis is placed on the safety of food production methods by producers, competent authorities, and consumers. Therefore, it is essential to search for strategies that guarantee food safety while, at the same time, maintaining the characteristics of specific products. To address this topic, this study focused on investigating the use of bioprotective cultures to inhibit the growth of L. monocytogenes in Sardinian PDO cheeses. Many studies have demonstrated the efficacy of commercially produced protective cultures and their potential for use in various food applications. The use of bioprotective cultures is a promising strategy for enhancing the safety of many kinds of food such as meat products, in which a significant reduction (p < 0.01) >1 Log 10 CFU/g of L. monocytogenes has been demonstrated [55], as well as dairy products [56,57].
The main objective of this work was to select autochthonous LABs that could potentially be used as bioprotective cultures to improve the safety of traditional PDO cheeses, such as Pecorino Sardo DOP or other Sardinian DOP dairy products. LABs are called "autochthonous" to indicate their environmental and indigenous origin [58,59]. Raw milk is characterised by a dynamic and diverse microbial community. Bacteria in milk have been widely reported, but the majority of bioprotective strains have been isolated from human or bovine milk [60]. For this reason, the present study investigated Sardinian sheep milk, commonly used for cheese production in Sardinia, as a source of potential biopreservative lactic acid bacteria. During pasteurisation, which is an important step to eliminate milk-borne pathogens, the number of microorganisms, including LABs, is reduced 20-fold [60,61]. However, some of the bacteria that do not survive pasteurisation could have beneficial effects in dairy products, in particular with regard to the protection of consumer health, such as those stemming from their ability to produce antimicrobial compounds active against pathogens [62] and maintain and reinforce the human immune system [63]. In response to this problem, developments in scientific research have proposed some strategies, such as the applications of protective cultures [64,65]. The addition of autochthonous LABs during cheesemaking can represent a good compromise between the safeguarding of traditional products, the adoption of solutions that allow high levels of food safety, and the possibility to maintain traditional manufacturing technologies [66]. In agreement with the results of other investigations [67], this study demonstrated a prevalence of Gram-positive bacteria in raw sheep milk, almost equally divided between rod-shaped bacilli (50.7%) and cocci (49.3%). Of these bacteria, 33.6% were identified as presumptive LABs due to their morphological and biochemical features, and 13/74 (17.6%) demonstrated antilisterial activity in vitro. This result highlights that Sardinian sheep milk is a good source of bioprotective microorganisms that could substitute for commercial bioprotective cultures. The data reported in this work can be compared with other similar studies in which indigenous LABs were selected and used as bioprotective cultures in foodstuffs. Our results show that LABs isolated from Sardinian raw milk present antilisterial capability that is superior to both the results of Ortolani et al. (2010), who found that 14.9% LAB strains isolated from raw milk and soft cheese exhibited antilisterial activity [28], and Reuben et al. (2020), who observed that 6.5% of LABs isolates from goat and cow raw milk were inhibitory to L. monocytogenes [68]. Considering cheese as a source of bioprotective bacteria, further results on the percentages of indigenous LABs with antilisterial activities can be found in the literature [30,69,70].
Traditional food production, such as the production of PDO cheeses, is part of Sardinian regional culture and is very important in terms of the economic activity of the region. The production specification of Pecorino Sardo PDO provides that "the whole sheep milk is inoculated only with bacteria from the origin area" [71]. The use of native microorganisms with inhibitory action against L. monocytogenes allows improvements in food safety in compliance with PDO procedures, specifications, and guidelines. Use of autochthonous thermophilic LABs, e.g., lactobacilli, to inhibit foodborne pathogens potentially present in dairy products makes them suitable microorganisms for use in cheesemaking [72]. In this study, thermophilic autochthonous LABs were tested for two reasons: Firstly, thermophilic LABs are more tolerant of the temperatures encountered during cheese production than mesophilic bacteria. Secondly, it could be useful to introduce protective cultures during the cheesemaking process, as is commonly performed for starter cultures. In fact, bio-protective cultures are traditionally sprayed onto the cheese surface [64], but this step is time-consuming and requires specialised personnel. The use of microorganisms with bio-preservative potential are attractive to dairy industries, but there is the need for an easier mode of application for the food safety operators. To evaluate this, this study investigated the possibility of inoculating the protective microorganisms in the milk during the cheesemaking process rather than spraying post-production. Scotta was identified as a medium compatible with dairy processes. "Scotta" is a residual milk by-product of ricotta cheese production and is currently used in Sardinia to produce "Scotta-innesto", a whey starter obtained by incubating starter bacteria in scotta at 42-45 • C for 18-20 h [48].
The use of scotta during sheep milk cheese production is a well-established procedure for preparing the scotta-innesto used for different kinds of Sardinian PDO cheese [73]. With this study, we evaluated whether scotta could be an appropriate medium to inoculate bioprotective cultures during PDO cheese production. In this work, a modified scottainnesto (MSI) was created in which the starter culture was replaced with cPC and LABs. The results of AWDA show that in vitro antilisterial activity was maintained in the MSI. Therefore, there is potential in inoculating the milk with bioprotective cultures during the cheese-making process.

Conclusions
L. monocytogenes is one of the most important microorganisms of concern in the dairy sector. Therefore, identifying strategies to improve cheese safety is of paramount importance. Given the key role of the dairy sector in Sardinia, maintaining and improving the safety standards of Sardinian PDO cheeses becomes even more important, given the huge economic losses that could arise in the case of L. monocytogenes contamination. In addition, consumers are increasingly aware of food safety issues, but at the same time, require products that are free from chemical additives. Results of this study identify strategies that could meet the needs of dairy farms and the demands of the consumer and allow compliance with the criteria laid down in community legislation.
Autochthonous thermophilic LABs identified in the present study are excellent candidates for bio-protective cultures concerning Sardinian PDO cheeses. The results obtained in this study show that Sardinian sheep milk is an excellent source of lactic acid bacteria with antilisterial activity and that these LAB strains show the potential to improve food safety in compliance with PDO specifications. In accordance with previous studies, the MRS culture medium appeared the most appropriate for isolation and growth of bioprotective lactic acid bacteria [74,75]. Evaluation of in vitro cPC and LAB inhibitory activity against L. monocytogenes found the effectiveness of bioprotective lactic acid bacteria against this relevant pathogen for the dairy industry. Autochthonous Lactobacillus species could be considered for new potential bioprotective cultures; moreover, their excellent adaptability to dairy matrices such as scotta and their optimum growth temperature (45 • C) makes them particularly suitable to be used for inoculation during the cheesemaking processes. These preliminary results will be useful to carry out challenge tests conducted by inoculating the two types of bioprotective cultures (commercial and autochthonous) in milk during cheese production in order to assess whether the antilisterial activity demonstrated in vitro is confirmed in the cheese. The use of bioprotective cultures, in combination with good hygiene in food processing, may become a promising strategy to control pathogen contamination in cheeses and improve food safety. Despite the encouraging results obtained from this preliminary study, it is essential for producers and consumers to take further precautions to prevent the contamination of cheeses, especially during post-process handling.