Characterization of Lactic Acid Bacteria in Pecorino di Farindola Cheese and Manufacturing with a Lacticaseibacillus paracasei Autochthonous Culture

: This study focused on the characterization of lactic acid bacteria (LAB) in a renowned traditional Italian cheese, Pecorino di Farindola, in order to select an autochthonous culture and investigate its potential for the improvement of safety and functional properties. Two hundred and six LAB isolated throughout production and maturation from nine cheese lots of three farms were identiﬁed by 16S rRNA gene sequencing and tested for the presence of genes encoding virulence factors, vancomycin resistance (for enterococci), biogenic amines (BAs) and bacteriocin production for antimicrobial activity; and for the capacity to survive in the gastrointestinal tract (GIT) based on tolerance to low pH and bile salts and adhesion to CaCo-2 cells. A Lacticaseibacillus paracasei isolate was used in cheese making and determined a decline of spiked Listeria monocytogenes and Escherichia coli O157 faster than in the control cheese. The autochthonous bacterial groups were numerically unaffected, apart from lactobacilli that were recovered in higher numbers in cheese with the addition of the L. paracasei strain. Based on repetitive extragenic palyndrome (Rep) polymerase chain reaction (PCR) proﬁles, the added strain possibly dominated until day 88 in the cheese. Results encourage further trials with autochthonous cultures in order to efﬁciently inhibit hazardous bacteria and to enrich a functional microbiota in Pecorino di Farindola and similar cheeses.


Introduction
Traditional cheeses are a food category with economical relevance for defined geographical areas and are appreciated by consumers for their peculiar organoleptic features and genuineness [1]. Most of these products in Italy are obtained from raw milk, added or not with natural whey cultures, following the European Union (EU) guidelines in order to ensure safety [2]. Pecorino di Farindola cheese is an example of dairy product from this category, and it is included in the list of the Italian traditional cheeses held by the Italian Ministry of Agriculture [3]. It is prepared, curiously only by women, from raw and non-refrigerated ewe's milk in nine towns of a mountainous district of the Abruzzo region. Animals providing milk are fed exclusively by grazing, cereals and fava beans that are cultivated locally. The product has an ageing time from three to five months [4].
Previous studies have provided evidence of the presence of high amounts of biogenic amines (BAs) in this cheese, mainly tyramine, whose formation was primarily attributed to the amino acid decarboxylase activity of enterococci and Enterobacteriaceae. However, a correlation between the composition of cultivable microbial groups and BA content could not be established [5]. The high BA content in this cheese is favoured by the use of pig rennet that is unique for its production process [6].
In this and other similar products, the antagonistic actions exerted by naturally occurring lactic acid bacteria (LAB) have a preeminent role in preventing the growth of undesired and pathogenic microorganisms. However, variability in the composition of the indigenous microbiota can cause instability in the safety promoting activities of LAB. The use of added cultures selected among autochthonous LAB could improve safety without modifying the typical manufacturing process and sensory characteristics [7].
Therefore, in this study, the LAB microbiota of Pecorino di Farindola cheese was characterized in order to select strains devoid of hazardous traits, which are able to persist during ageing and to antagonize pathogens, to be used in cheese making trials. Conventional bacterial counts were used to provide a quantitative estimate of the cultivable bacteria that predominated, mostly impacted biochemically changes in the cheese during maturation and were ingested with the product. In addition, this allowed isolation of LAB strains for further characterization. LAB isolated at the end of cheese maturation were also assessed for the prerequisites for survival in the gastrointestinal tract (GIT). In fact, traditional cheese is recognized as a good source of LAB able to colonize GIT that could confer functional properties to the product if able to exert probiotic actions [8].
LAB isolates, including enterococci, were obtained during cheese maturation from nine lots manufactured by three producers. These were screened by polymerase chain reaction (PCR) for genes encoding the production of nisin and other lantibiotics, as a bacteriocin class with broad inhibitory spectrum and therapeutic potential [9], enterocins, enterococcal virulence factors and vancomycin resistance genes and BA production. Moreover, isolates from cheeses at the end of ripening, which are those ingested by consumers, were evaluated for tolerance to low pH and bile salts, adhesion to intestinal CaCo-2 cells and production of antibacterial substances. Among these, a Lacticaseibacillus paracasei isolate was tolerant to low pH and bile salts and was able to inhibit the growth of Listeria monocytogenes in vitro. This was used as added culture in a cheese making trial. In particular, the abilities of this autochthonous L. paracasei isolate to inhibit L. monocytogenes and Escherichia coli O157, two pathogens of major concern in this type of products [10,11], and to dominate during cheese maturation were evaluated. were used as positive control strains in specific PCR reactions, antimicrobial production tests and cheese making trial. All culture media were provided by Biolife Italiana, S.r.l. (Milan, Italy). Enterococci and rod shaped LAB were propagated in De Man, Rogosa and Sharpe (MRS) broth (20.0 g/L glucose, 10.0 g/L peptone, 8.0 g/L meat extract, 4 g/L yeast extract, 3 g/L sodium acetate, 2 g/L triammonium citrate, 2 g/L K 2 HPO 4 , 0.2 g/L MgSO 4 , 0.05 g/L MnSO 4 , 1 mL sorbitan monooleate and pH 6.5) and incubated for 24-48 h at  [12]. Samples were collected from nine batches of cheese produced in three dairy factories belonging to the producer association "Consorzio di Tutela del Pecorino di Farindola" and comprised milk, curd and cheese at days 7, 14, 21, 35, 49, 63, 77, 91, 121 and 150 of ripening during the period March to September 2018. Samples were transferred into sterile containers, in the manufacturing plant or in the ageing rooms, transported to the laboratory and stored in refrigeration conditions for a maximum of 48 h until analysis. The collected amounts were 1 L of milk, 500 g of curd and a whole piece of cheese. The a w determination was carried out with the Aqualab 4TE device (Meter Group, Pullman, WA, USA) according to the norm ISO 18787:2017 [13]. The determination of pH was carried out using a Seven Easy pH Meter (Mettler-Toledo S.P.A., Milan, Italy) on a suspension of 10 g of sample homogenized in 90 mL of deionized water for 1 min in a Stomacher 400 (Biosigma, Cona, VE, Italy).

Materials and Methods
LAB counts were carried out as described by Del Matto et al. [14]. Namely, LAB were counted and isolated on MRS agar medium and M17 agar medium (19.0 Di-sodiumglycerophosphate, 11.0 g/L agar, 5 g/L lactose, 5 g/L meat digest, 5 g/L tryptone, 5 g/L soya peptone, 2.5 g/L yeast extract, 0.5 g/L ascorbic acid, 0.25 g/L MgSO 4 and pH 6.9), incubated at 37 • C for 72 h in anaerobic jars and Slanetz and Bartley (SB) medium (20.0 g/L tryptose, 13 g/L agar, 5 g/L yeast extract, 2.0 g/L glucose, 4.0 g/L K 2 HPO 4 , 0.4 g/L sodium azide, 0.2 g/L 2,3,5-Triphenyltetrazolium Chloride and pH 7.2), selective for enterococci, and incubated at 37 • C for 24 h in aerobiosis. Presumptive LAB were tested for catalase activity by transferring an individual colony onto a glass slide and adding a drop of 3% (v/v) H 2 O 2 . Gram staining was also carried out.
Tolerance to low pH was tested for strains of lactobacilli grown in MRS broth at 37 • C for 24 h. A 10 µL aliquot of culture was inoculated into 990 µL of phosphate-buffered saline (PBS) (8 g/L NaCl, 2.9 g/L Na 2 HPO 4 ·2H 2 O, 0.2 g/L KH 2 PO 4 and 0.2 g/L KCl) with pH adjusted to 2.5, 3.0 or 6.2 (control). Cell suspensions were incubated for 2 h at 37 • C. Counts were carried out on MRS agar before and after incubation, and the survival rate was expressed as percentage of Log CFU/mL at the tested pH compared to the initial count.
For the determination of tolerance to bile salts, strains were grown in MRS broth at 37 • C for 24 h and then inoculated at 2% (v/v) ratio into MRS broth containing 0.3%, 0.5% or 1% (w/v) bile salts (Ox-bile, Merck S.p.A., Serono, RM, Italy) and incubated at 37 • C. Optical density at 600 nm (OD 600 ) was measured after 24 h using a BioPhotometer D30 (Eppendorf S.r.l., Milan, Italy) and compared to a control culture grown in MRS broth without bile salts. The results were determined as percentage of OD 600 in the presence of bile salts compared to the control. All experiments were performed in triplicate.
The deferred antagonism test was used to evaluate the in vitro production of inhibitory compounds by LAB isolates against pathogenic bacteria. Briefly, 2 µL of a bacterial culture grown for 16-18 h at 37 • C were spotted onto the surface of BHI agar (BHI broth + Bacto agar 1.5% w/v) and incubated at 30 • C for 18-24 h. The plates were than overlaid with 8 mL of BHI soft agar (BHI broth + Bacto agar 0.6% w/v) seeded with 8 µL of indicator strain grown for 16 h at 37 • C in BHI broth. After incubation for 16-18 h at 37 • C, the plates were screened for the presence of growth inhibition zones in the indicator layer.

DNA Extraction
DNA was extracted from 300 µL fresh bacterial culture with a Maxwell ® Systems-Automated Sample Processing (Promega Italia S.r.l., Milan, Italy) using the Maxwell ® 16 Tissue DNA Purification Kit (Promega Italia), according to the manufacturer's instructions.

PCR Reactions
Amplification reactions were performed according to previously reported methods using the PCR 2x Master Mix (Promega Italia) in a GeneAmp PCR System 9700 (Thermo Fisher Scientific). PCR products were analyzed by capillary electrophoresis using the QIAxcel ® Advanced System (Qiagen, London, UK). For the identification of bacteria, the MicroSEQ ® 500 16S rDNA Bacterial Identification System (Thermo Fisher Scientific), amplifying a 500 bp fragment from the 5 end of the 16S ribosomal RNA gene, was used according to the manufacturer's instructions.
L. casei and L. paracasei were differentiated by PCR, according to Ward and Timmins [20], after 16S ribosomal RNA gene sequencing.
Repetitive Extragenic Palyndrome (Rep) PCR was carried out with the GTG 5 primer as described by Versalovic et al. [23]. Image analysis and clustering were performed with the BioNumerics V7.5 software (Applied Maths N.V., Sint Martens Latem, Belgium), using the Dice coefficient for pairwise comparison and the Unweighted Pair Group Method using Arithmetic Averages (UPGMA) for clustering. Some isolates were analyzed in duplicate in the same experiment in order to establish a cut off for considering isolates distinct. Consequently, the cut off was set at 90% similarity.

Sequencing
Purification of PCR fragments for sequencing was performed with the Cytiva Illustra™ ExoProStar™ 1-Step kit (Merck, Paris, France). The sequencing master mix was prepared with the BigDye ® Terminator v3.1 Cycle Sequencing Kit (Thermo Fisher Scientific), and the final purification step was performed with Sephadex ® G-50 (VWR International S.r.l., Milan, Italy). Sequence analysis was performed with ABI-Prism 3500 Genetic Analyzer (Thermo Fisher Scientific). Bacterial isolates were identified relative to the species level according to the highest score obtained by nucleotide BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 13 July 2021) alignment.

Test of LAB Adhesion to CaCo-2 Cells
The Human Colon Epithelial Cancer cell line CaCo-2 (ATCC, Rockville, MD, USA) was used as surrogate system in order to evaluate strain capacity to adhere to intestinal cells. CaCo-2 cells were seeded in 24-well tissue culture plates at the density of 25,000 cells/cm 2 and grown for 15 days to simulate mature human enterocytes [24]. LAB isolates were grown for 16-18 h at 37 • C and centrifuged from 1 mL of culture at 10,000× g for 10 min at 4 • C. The pellet was resuspended in 1 mL of sterile PBS. The concentration of bacterial cells in these suspensions was determined by plate count. Then, 0.5 mL of LAB suspension was added to CaCo-2 cell monolayers that were previously washed five times with sterile PBS. After 1 h of incubation at 37 • C, the wells were gently washed several times with PBS to remove unbound bacteria. Adhered bacteria were then detached from CaCo-2 cells using trypsin-ethylenediaminetetraacetic acid (EDTA) (0.05% w/v trypsin, 0.02% w/v EDTA) solution and counted on plate. The adhesion capacity was expressed as percentage of CFU/mL of bacteria detached from cells on the CFU/mL in the inoculum. The experiments were carried out in triplicate.

Experimental Production of Pecorino di Farindola Cheese
Two batches of Pecorino di Farindola cheese were prepared from 300 L of raw milk previously tested for the absence of L. monocytogenes and E. coli O157. Milk was contaminated with L. monocytogenes ATCC 7644 and EURL 12M0B098, E. coli O157 VTEC ISS, ATCC 25952 and EURL C07 at final concentrations of about 3 Log CFU/mL. The preparation of the inoculum was carried out in accordance with the European Union Reference Laboratory for Listeria monocytogenes (EURL) guidelines for L. monocytogenes [25] by revitalizing frozen cultures at 37 • C for 18 h in BHI and adapting the strains at the temperature of milk processing (35 • C) in new BHI broth cultures for 24 h.
In one batch (E cheese), the isolate L. paracasei L11121 was added at a final concentration of about 6 Log CFU/mL by using 1.5 L of culture grown in MRS broth for 24 h at 37 • C in aerobic conditions. Bacterial cultures were added after the transfer of the raw milk in the coagulation vat. Cheese manufacturing was carried out according to the procedure adopted by the farms associated with the "Consorzio di Tutela del Pecorino di Farindola". Briefly, milk was heated at 35 • C, and pig rennet was added at 2.5 mL/L final concentration. Casein precipitation was obtained in 40 min. The curd was mechanically broken by using a lyre and left to rest for about 10 min. Temperature was then raised to 37 • C for 5 min. The curd was transferred into polyethylene molds and, after whey purging and manual pressing, the cheese-wheels were surface salted with coarse salt and left to season on wooden shelves in climatic chambers (Piardi ™ CC2000, Castenedolo, Italy) at 18 • C and 65% relative humidity (RH) for 148 days. Twenty-six cheese wheels of about 2 kg were produced. The spiked pathogens, LAB on MRS medium, LAB on M17 medium, enterococci and coagulase positive staphylococci (CPS), were monitored by plate count at different maturation times.

Statistical Analyses
Two-way analysis of variance (ANOVA) was applied to compare count data of each microbial group during maturation between cheese lots and between E cheese and control cheese (C). The data series were considered significantly different for p < 0.05. The data of isolate tolerance to low pH, bile salts and cell adhesion capacity were analyzed by correlation clustering. All statistical analyses were carried out using PAST (Paleontological Statistic Software) V 4.06b, (https://www.nhm.uio.no/english/research/infrastructure/ past/, accessed on 12 August 2021) [26].

LAB Dynamics in Pecorino di Farindola Cheese
MRS and M17 media were used to improve the recovery of LAB. In particular, M17 medium is commonly used to count and isolate lactic cocci. However, this medium also sustains the growth of lactobacilli so that lactic cocci are not detected when lactobacilli predominate in the sample [27]. Most of the bacteria enumerated on M17 medium were of coccoid shape until day 14. Afterwards prevalently rod shaped LAB grew on this medium. On MRS medium, rod shaped bacteria were isolated at all times, but coccoid cells were also retrieved and were identified as enterococci by 16S rRNA gene sequencing. In particular, enterococci could be isolated on M17, MRS media and on the more specific SB medium.
The evolution of LAB counted on different culture media in the Pecorino di Farindola cheese lots analyzed in this study is shown in Figure 1. Although there were differences in the evolution among manufacturers and, for Dairy 1, among lots, enterococci reached high numbers during maturation in all cheeses and were still high in the products ready for consumption, being in the range of 5-7 Log CFU/mL. LAB showed similar evolution trends among lots and dairies, except for Lot 3 of Dairy 1, where LAB grown on M17 medium declined drastically after day 91.

Identification of Dominant LAB in Pecorino di Farindola Cheese
Predominant LAB involved in the maturation of Pecorino di Farindola cheese were isolated from each growth medium at different maturation times, obtaining 206 isolates. Although the exact load of each isolate in the sample of origin could not be determined, its order of magnitude can be estimated from the count on the respective isolation medium. Isolate identification Although there were differences in the evolution among manufacturers and, for Dairy 1, among lots, enterococci reached high numbers during maturation in all cheeses and were still high in the products ready for consumption, being in the range of 5-7 Log CFU/mL. LAB showed similar evolution trends among lots and dairies, except for Lot 3 of Dairy 1, where LAB grown on M17 medium declined drastically after day 91.

Identification of Dominant LAB in Pecorino di Farindola Cheese
Predominant LAB involved in the maturation of Pecorino di Farindola cheese were isolated from each growth medium at different maturation times, obtaining 206 isolates. Although the exact load of each isolate in the sample of origin could not be determined, its order of magnitude can be estimated from the count on the respective isolation medium. Isolate identification based on 16S rRNA gene sequences is jointly reported in Table 1 with  In milk, curd and cheese in early ripening, the dominant microorganisms present in all three dairies, although not detected in all lots, were Lactococcus lactis subsp. lactis and Leuconostoc mesenteroides. In particular, L. lactic subsp. lactis with the nisin structural gene nisA was predominant for five lots from two dairies. The heterofermentative species Levilactobacillus brevis was the only microorganism isolated from all cheese lots and generally retrieved until advanced maturation, while other LAB species predominated according to the dairy. The L. brevis isolates did not possess amino acid decarboxylase encoding genes except for an isolate from Dairy 1 in which an ornithine decarboxylase gene, odc, was identified. Lactiplantibacillus plantarum was also isolated from all three dairies but not from all cheese lots. L. plantarum isolates from Dairy 2 were obtained from late ageing cheese, thus indicating their capacity to persist during maturation. The presence of Lacticaseibacillus paracasei was demonstrated only in two dairies. Other species found to predominate in particular maturation phases, i.e., Loigolactobacillus coryniformis, L. pseudomesenteroides, L. garviae, Streptococcus infantarius and S. lutetiensis, appeared to be specific relative to the manufacturing plant.
Enterococci exhibited high diversity in the endowment of virulence, bacteriocin and tdc genes. The tyrosine decarboxylase gene tdc, suggested to be a species-specific character in E. faecalis and E. faecium [28], was detected in all the E. faecalis isolates but not in E. faecium, except for one isolate belonging to this species. All isolates were negative for genes encoding vancomycin resistance. Different species or biotypes of enterococci co-dominated or alternated as dominant at different maturation times. In two cases E. faecalis isolates with virulence characters were still present in the product ready for consumption. Enterococci with bacteriocin encoding genes and devoid of virulence genes were also isolated from all three dairies in late ageing.

Evaluation of the Prerequisites for Survival in GIT
LAB isolated at days 121 and 150, i.e., those that are ingested in high numbers with the product, including the enterococcal isolates devoid of virulence characters, were tested for their ability to survive in acidic conditions (pH 2.5 and 3.0) in the presence of bile salts (0.3%, 0.5% and 1%) and for their ability to adhere to CaCo-2 cells. These cells are generally used as a model of the intestinal epithelial cells [29]. L. casei ATCC 393, a strain isolated from cheese and already tested for tolerance to acidity, bile salts and adhesion properties [30], was used as a reference to interpret the characteristics in the isolates examined in this study by comparison with previous work. The results are reported in Table 2. Table 2. LAB tolerance to pH 2.5 and 3.0, to bile salts 0.3, 0.5 and 1% (w/v) and adhesion capacity to CaCo-2 cells. Isolate labels "L" or "E" stand for lactobacilli and enterococci, respectively; the first following number indicates the dairy, the second the lot and the three last numbers indicate the day of isolation (121 or 150). Letters at the end distinguish isolates obtained from the same sample. At pH 3.0, all isolates tested exhibited good tolerance, while at pH 2.5 only ten were able to survive. Most isolates showed significantly higher tolerance than the reference strain to all tested bile salt concentrations. In the cell adhesion test, six enterococci, seven L. brevis isolates and L. plantarum L22150 showed better adhesion capacity than the reference strain. All the L. paracasei isolates showed a low cell adhesion capacity.

Isolate
Patterns of tolerance to GIT conditions and adhesion capacity were not species-specific, as shown by correlation clustering of average data series for the isolates listed in Table 2 ( Figure 2).  Table 2 for LAB isolates at 121 and 150 days.

LAB Activity against Pathogenic Bacteria
The same LAB isolated at days 121 and 150 were also tested for the production of antimicrobial compounds active against L. monocytogenes ATCC  Table 2 for LAB isolates at 121 and 150 days.

LAB Activity against Pathogenic Bacteria
The same LAB isolated at days 121 and 150 were also tested for the production of antimicrobial compounds active against L. monocytogenes ATCC 7644, Salmonella Typhimurium ATCC 14,028 and Staphylococcus aureus ATCC 33,862 by the deferred antagonism test. Seven enterococcal isolates, coinciding with those in which enterocin genes were detected, and L. paracasei L11121 showed activity against L. monocytogenes ATCC 7644. In particular, the enterococcal isolates produced inhibition halos with diameters, extending from the spot margin, between 4 and 10 mm, while L. paracasei L11121 produced an inhibition halo of 4 mm. The latter autochthonous strain was selected as culture to be used in a cheese making trial because of the recognized safety status of the species [31].

Cheese Making Trial with the Addition of L. paracasei L11121 Isolate
The count values for all the microbial groups assessed in the cheese making trial with the use of L. paracasei L11121 are reported in Table 3.
According to two-way ANOVA analysis, the number of LAB grown on MRS and M17 media was significantly higher in E cheese (with addition of L. paracasei L11121) compared to C cheese (control) for p < 0.001 and p < 0.05, respectively. The decrease in E. coli O157 and L. monocytogenes in E cheese was more rapid, becoming evident from day 17 for E. coli O157 and from day 28 for L. monocytogenes. The count values of L. monocytogenes in E and C cheeses were significantly different (p < 0.01). In E cheese, this pathogen remained at levels below the admitted limit of 100 CFU/g [10] since day 28.

Rep PCR for L. paracasei L11121 Tracing during Cheese Maturation
The persistence of L. paracasei L11121 in cheese was determined by isolating LAB at each time interval and identifying L. paracasei isolates by species-specific PCR. Ten L. paracasei isolates were examined for each cheese at each time of analysis for a total of 260 isolates that were submitted to genotyping by Rep PCR. In the dendrogram reported in Figure 3, one representative profile for each group of isolates separated at 90% similarity from each sample is included. Thus, all the profiles representative of each cheese batch and time of isolation are shown.
In most cases, profiles from the same sample, always obtained in the same analysis session, that shared more than 90% similarity were considered duplicates of identical isolates. For this reason, with a few exceptions, only one isolate per sample is shown in the dendrogram (Figure 3). It can be observed that L. paracasei L11121 shared highly similar profiles with isolates derived from milk and curd of E cheese. Profiles that clustered together with those of L. paracasei L11121, milk and curd isolates from E cheese but showed lower similarity with the latter, possibly as a consequence of the incomplete reproducibility known to occur among different Rep PCR runs [32], were retrieved until day 88 from E cheese. This observation suggests that the added culture was able to persist until the minimum maturation period of Pecorino di Farindola cheese. specific PCR. Ten L. paracasei isolates were examined for each cheese at each time of analysis for a total of 260 isolates that were submitted to genotyping by Rep PCR. In the dendrogram reported in Figure 3, one representative profile for each group of isolates separated at 90% similarity from each sample is included. Thus, all the profiles representative of each cheese batch and time of isolation are shown. The isolate labels indicate the cheese batch of origin (E or C); the sample type, i.e. milk (M), curd (C) or cheese (CH); isolate identity, defined with a letter; and the day of isolation. The isolate labels indicate the cheese batch of origin (E or C); the sample type, i.e. milk (M), curd (C) or cheese (CH); isolate identity, defined with a letter; and the day of isolation.

Discussion
In this study, variability was observed in the identity of dominant bacteria in Pecorino di Farindola cheese among producers and among lots from the same producer during the whole maturation process. Nevertheless, constant microbiological features could be defined, i.e., the predominance of Lactococcus lactis subsp. lactis and Leuconostoc mesenteroides in curd and in cheese in early ripening and the predominance of Levilactobacillus brevis and enterococci throughout ripening.
An unusual observation was that the species Lacticaseibacillus paracasei, found to be the most persistent in many other traditional cheeses [14], was not isolated from all manufacturers. L.
brevis, present in all cheese lots, was found to persist in the product ready for consumption. This species has not been characterized much for its role in cheese flavor formation and safety improvement, although its presence has also been documented in other cheeses from raw ewe and goat milk [33]. In addition, strains of L. brevis with probiotic properties have been described [34,35]. These aspects should be addressed in further studies on L. brevis isolates from Pecorino di Farindola cheese, together with the safety concern of BA formation and horizontal gene transfer (HGT) of the involved genetic determinants reported for this species [36]. The fact that no amino acid decarboxylase genes, except an odc gene found in one isolate, were detected in this study indicates that this LAB species is not frequently responsible for BA formation in Pecorino di Farindola cheese.
Enterococci were found to be abundant in this cheese, as in other cheeses made from raw milk [33,37]. A high diversity was observed for these bacteria on the basis of the presence of genes encoding beneficial or undesired traits. It was surprising that the tdc gene was not detected in most of the E. faecium isolates examined in this study since it was suggested that the presence of this gene is a species-specific trait [29]. This finding is in favor of the safety status of the E. faecium isolates from Pecorino di Farindola cheese in which the presence of virulence characters was not frequent and bacteriocin genes were detected. Therefore, it would be interesting to investigate the ability of these E. faecium isolates to outcompete enterococci with risk characters, such as the examined E. faecalis isolates.
Since individual strains of enterococci can present risk characters, the Qualified Presumption of Safety (QPS) status cannot be conferred to the whole genus or species [37]. On the other hand, the flavor formation ability and probiotic prerequisites demonstrated also in this study, as well as the production of broad spectrum bacteriocins [37,38], represent desirable properties for autochthonous microbiota components of traditional cheeses. Therefore, standardized procedures should be fixed in order to individually evaluate enterococci for safety-related aspects to allow their use as autochthonous cultures. These procedures must include the demonstration that these bacteria are not prone to acquiring unwanted genetic traits by HGT in the cheese environment as it was reported for some strains [39,40].
Analogous issues regard other dominant bacterial species found in this study, e.g., L. garviae, Streptococcus infantarius and S. lutetiensis, that can cause infections in humans [41,42] but include also strains that carry beneficial characters and are devoid of virulence traits [43,44].
The assessment of the capacity of isolates obtained at the end of maturation to withstand the harsh conditions of GIT indicated that enterococci, L. brevis and L. plantarum isolates from Pecorino di Farindola cheese are potentially able to survive in GIT upon consumption of the product. This observation encourages future investigations of the probiotic characteristics of the isolates to enrich the product with autochthonous cultures able to enhance its functional properties.
The cheese making trial was carried out with an autochthonous added culture, a L. paracasei isolate, since this species has a proven suitability for use as adjunct culture in cheese [45] and a recognized safety status [31]. The added culture possibly dominated in the experimental cheese until day 88, which is the minimum maturation time for Pecorino di Farindola cheese. Indeed, Rep PCR profiles clearly similar to that of L. paracasei L11121 were obtained for L. paracasei isolates only from E cheese until advanced maturation. These should be considered different from the added strain because they show less than the 90% profile similarity fixed in this study for replicate profiles of the same strain. However, this value regarded genotypic profiles obtained in the same experiment, while low reproducibility among different experiments was reported for this typing method [32]. Therefore, the fact that L. paracasei L11121 was able to persist until the minimum maturation period of Pecorino di Farindola cheese cannot be ruled out.
The favorable effects observed from the use of the L. paracasei culture, i.e., the enrichment in lactobacilli and faster pathogen decline, can be considered relevant functional and safety improvements. Indeed, L. monocytogenes was below the tolerance limits in E cheese already at day 28 and was absent at the minimum ripening time of 90 days. In a previous investigation on the same cheese with no added cultures, the pathogen was undetectable after the minimum ripening period only in one out of three batches examined [46]. A variability in the presence of an anti-listerial microbiota probably determined the different behavior of the pathogen in different cheese batches, confirming the necessity of using added cultures to ensure its absence before product commercialization.
Results suggest further developments of mixed cultures able to enhance safety improvements by inhibiting enterococci with risk characters and CPS. CPS in this cheese appeared to be of particular concern since they reached numbers that could cause staphylococcal toxin accumulation during early ripening. These bacteria exceeded the critical limit [10,47] until day 88 also in E cheese, thus requiring the selection of autochthonous LAB strains able to antagonize them.
In particular, lactococci with the nisin structural gene, found to naturally dominate in some lots of Pecorino di Farindola cheese during early ripening, should be further investigated for the capacity to inhibit detrimental microorganisms. Indeed, it was demonstrated that nisin is produced in cheese and that its production can continue in GIT with consequent health promoting effects [48].

Conclusions
In this study, a "cultivable" core LAB microbiota of Pecorino di Farindola cheese was defined based on the presence of some species in all the cheese lots examined. It was placed in evidence that this cheese can be considered a source of microorganisms that could exert beneficial effects, with the potential of being established in GIT and to express antimicrobial activity. On the other hand, enterococci with virulence characters and CPS were found to be present in the product at high levels and were not affected by the addition of an autochthonous L. paracasei culture. Therefore, the use of native cultures in association should be further investigated to ensure safety and increase health promoting features of Pecorino di Farindola and other similar cheese types.