Indigenous raw milk microbiota influences the bacterial development in traditional cheese from an alpine natural park

International Journal of Food Microbiology 92 (2004) 141 – 151 www.elsevier.com/locate/ijfoodmicro

Indigenous raw milk microbiota influences the bacterial development in traditional cheese from an alpine natural park E. Poznanski a, A. Cavazza a, F. Cappa b, P.S. Cocconcelli b,c,* b

a Istituto Agrario di S. Michele. Via E. Mach 1, I-38010 San Michele all’Adige (TN), Italy Istituto di Microbiologia, Universita` Cattolica del Sacro Cuore, via Emilia Parmense 84, 29100 Piacenza, Italy c Centro Ricerche Biotecnologiche, via Milano 24, 26100 Cremona, Italy

Received 7 April 2003; received in revised form 29 July 2003; accepted 16 September 2003

Abstract Nostrano di Primiero is a 6-month ripened cheese produced from raw milk collected in the Paneveggio – Pale di San Martino Natural Park area in the Italian Dolomites. In summer, this cheese is made using milk collected from two different areas, Passo Rolle and Vanoi, in the Paneveggio Natural Park. During the experiment, the milk from the two areas was separately processed, and cheeses were made in the same cheese factory using the same technological process. The microbiota of raw milk and cheeses of the two areas was isolated and the dominant population was monitored by RAPD analysis and identified by 16S rRNA sequence. The milk of the Passo Rolle area was mainly composed of mesophilic strains, thermophilic Streptococcus thermophilus, and low amounts of enterococci were also found; the milk of the Vanoi area was dominated by mesophilic microbiota mostly Lactococcus lactis ssp. cremoris and ssp. lactis and Lactobacillus paracasei ssp. paracasei. The plating of the natural starter culture revealed the presence of a relevant community of thermophilic cocci and lower amounts of enterococci. The dynamic population analysis showed the importance of the natural starter culture in the first 2 days of cheese ripening in both cheeses. Moreover, the large biodiversity observed in the raw milks was also detected in the cheeses during ripening. The Vanoi cheese was dominated by Enterococcus faecium and Streptococcus macedonicus in the first two days and mesophilic 21 Lb. paracasei ssp. paracasei became the most represented population after 15 days of ripening. In the first few days, the Rolle cheese was characterized by being mainly composed of thermophilic S. macedonicus and S. thermophilus and secondarily by mesophilic cocci. During ripening, the microbiota composition changed, and at 15 days, mesophilic lactobacilli were the dominant population, but later, this was mainly composed of mesophilic cocci and lactobacilli. The taxonomical identification by 16S rRNA sequence confirmed a large biodiversity related to raw milk microbiota and only five strains of S. macedonicus, Lactobacillus plantarum, 21 Lb. paracasei ssp. paracasei, Lactobacillus fermentum and E. faecium were detected in both cheeses. D 2004 Elsevier B.V. All rights reserved. Keywords: Raw milk cheese; Lactic acid bacteria; Natural starter culture; Nostrano di Primiero

1. Introduction * Corresponding author. Istituto di Microbiologia, Universita` Cattolica del Sacro Cuore, via Emilia Parmense 84, 29100 Piacenza, Italy. Tel.: +39-523-599251; fax: +39-523-599246. E-mail address: [email protected] (P.S. Cocconcelli). 0168-1605/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ijfoodmicro.2003.09.006

The particular flavour and typical organoleptic properties of raw milk cheeses are associated with specific attributes of raw milk, related to the race and nutrition of dairy cows, the basic traditional cheese-

142

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151

making process and the natural microbiota (Corroler et al., 1998; Beresford et al., 2001) responsible for the fermentation process and ripening. A variety of artisanal and traditional Italian cheeses produced in the Alps are manufactured using raw milk and their fermentation is driven by adventitious microbiota (Cappa et al., 2002; Senini et al., 1997) or by bacteria added to milk as natural starter cultures (NSCs). It has been demonstrated that raw milk cheeses have more intense and typical flavour than pasteurised milk cheeses (Albenzio et al., 2001; Demarigny et al., 1997; Beuvier et al., 1997), and, besides modification connected to rennet coagulation properties and indigenous milk enzymes, the main differences concern the large number of bacteria and related enzymes associated with raw milk. Nostrano di Primiero is a 6-month ripened cheese produced from raw milk collected from a particular area of Paneveggio– Pale di San Martino Natural Park in the Italian Dolomites. During summer, it is obtained from milk produced in the alpine pastures of two different areas, Passo Rolle and Vanoi, in the Paneveggio Natural Park. The milk from the two regions is collected and processed separately, but using the identical technology, in the same cheese factory located in the lower valley. The cheese is produced using raw milk from the evening milking, skimmed overnight, mixed with the morning-collected milk. A better understanding of the role played by the natural microbiota in the production of traditional cheeses requires a deeper study of the microbial community involved in the fermentation process and increased knowledge of the population dynamics during cheese manufacture. Advanced molecular approaches have allowed the study of cheese microbiota and the identification of species and strains involved in the fermentation of raw milk cheese (Wouters et al., 2002). Recent studies have shown that artisanal cheeses have different and typical microbial population dynamics related to the production technology and geographic area of origin (Baruzzi et al., 2000; Cappa et al., 2002). In these cheeses, the microbiota is quite heterogeneous (Morea et al., 1999), its composition changes during the cheese ripening and the strains dominating the first stages are not necessarily predominating in the later periods (Mannu and Paba, 2002). In raw milk cheeses, the non starter lactic acid bacteria (NSLAB) support the cheese making process

in the later phases, being often found as secondary flora during ripening (McSweeney et al., 1993; Shakeel-Ur et al., 2000; Demarigny et al., 1996). The NSLAB composing the natural milk microbiota are differentiated in mesophilic and thermophilic bacteria (Beresford et al., 2001) and play a relevant role in the cheese made from raw milk, increasing the diversity of the flavours, as has been reported (Corroler et al., 1998; Grappin and Beuvier, 1998). On the contrary, most dairy industries employ starter strains selected and routinely subcultured in milk for a faster acidification; as a consequence of this technology, the number of strains present is reduced and a certain uniformity of the products is observable (Wouters et al., 2002). Although information has been obtained recently on the role of natural starter cultures, little information is yet available on the influence of the environmental contamination of milk on raw milk cheese fermentation. The observation that cheese produced from raw milk collected during summer pastures presents better sensorial properties than cheese manufactured during winter, when cows are feed with hay and live in cattlesheds, suggests that, in addition to diet variation, environmental microbiota may also have a role in the fermentation process and the quality of cheese. Due to the peculiar situation of Nostrano di Primiero production, milk collected from two different natural park areas processed in the same cheese factory using the same natural starter cultures; this cheese could be considered as a good model for the evaluation of the role of environmental contamination of lactic acid bacteria on cheese fermentation. The purpose of the present study was to identify the role of lactic acid bacteria from natural milk culture and environmental microbiota on cheese fermentation and ripening, by comparison of cheeses manufactured with the same technology in the same cheese factory but collected from two different alpine areas, by means of molecular techniques for bacterial community development analysis.

2. Materials and methods 2.1. Sampling and colony enumeration Two areas of milk production were selected: the ‘‘Passo Rolle’’, where the dairy farms are bigger and

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151

more concentrated in a smaller area, and the ‘‘Vanoi’’, characterised by smaller alpine farms not so close together. The milk from the two areas was taken to the dairy farm, where two identical processes took place. The only difference between the two processes was the milk’s origin. Three samples of milk at the beginning of production and cheese at 24 h, 48 h, 7 days, 15 days, 1 month and 2 months were collected and immediately frozen in liquid nitrogen. The samples were stored at 80 jC for microbial analysis. Cheese samples (10 g) were homogenised in 90 ml peptone solution (0.1% w/v) for 2 min in a Stomacher 400 blender at high speed, and decimal dilutions of the homogenates were prepared with sterile peptone solution. Aliquots of the dilutions were plated onto the following media: MRS acidified with lactic acid at pH 5.5 incubated anaerobically for 3 days at 30 and 45 jC for the recovery of mesophilic and thermophilic lactobacilli, respectively; M17 agar incubated aerobically for 3 days at 30 jC and for 3 days 45 jC, for the count of mesophilic and thermophilic cocci, respectively. Microscopic observation, Gram staining and catalase test were used to confirm the enumeration of lactic cocci and rods. Kanamycin Aesculin Azide Agar Base (KAA) incubated aerobically for 24 h at 37 jC was used for enumeration of enterococci. All the media were from Oxoid (Basingstoke, UK). Anaerobiosis was achieved using Anaerocult A (Merck, Darmstadt, Germany). After the incubation period, the plates with 25 – 250 colony forming units (cfu) were selected for enumeration and isolation. The results were expressed as cfu g 1. 2.2. Cheese-making manufacture Nostrano di Primiero is a raw-milk cheese produced by mixing the milk from the evening milking, skimmed overnight at room temperature, with the morning-collected milk. For its production, a natural starter culture in milk is used to drive the primary fermentation. This milk natural culture was achieved by scalding for 15 min raw milk to 65 jC and incubating overnight at 45 jC, and was used routinely as 3%

143

inoculum (v/v) for the day after Nostrano di Primiero production. After mixing, the milk with the NSC in a vat, calf rennet was added. Curdling occurred at 35 jC, the curd was scalded at 48 jC and then cut into rice-like particles and kept at the same temperature until pouring into molds. The cheese was then brined at 28 jC for 2 days. Ripening followed for 4 –12 months at 18 jC. 2.3. Isolation of microorganisms and phenotypic characterization From each countable plate, at least five colonies were isolated for each morphologically different colony types observed. The cells were Gram stained and subjected to catalase test. After microscopic observation, the Gram-positive, catalase-negative colonies were picked with a sterile toothpick and transferred onto two fresh plates of MRS or M17 agar if rods or cocci, respectively. These plates were incubated anaerobically at 15 and 45 jC if rods and at 10 and 45 jC if cocci for 7 days at the lower temperatures and for 2 days at the upper one, in order to verify the ability of the isolated strains to grow at different temperatures. The grown colonies were finally subcultured in liquid MRS (rods) or M17 (cocci) and stored at 80 jC with 20% glycerol (v/v). 2.4. RAPD analysis RAPD analysis was performed on 161 isolated strains from samples collected in the Rolle and Vanoi areas and on 20 from the NSC. The isolates for the fingerprint analysis were chosen from the dominant bacterial group of each selection step. One milliliter of the overnight liquid culture was centrifuged at 10,000 rpm for 5 min, and the pellet was washed twice with sterile distilled water. The cells were than resuspended in 1 ml of distilled water and DNA of the cells was extracted using the Instagenek Matrix (Bio-Rad Laboratories, Hercules, CA, USA) following the manufacturer’s instruction. PCRs were performed in a total volume of 50 Al in PTC-100 Thermal Cycler (MJ Research, Boston, USA). The PCR mixture consisted of 20 Al of DNA, 5 U of Taq polymerase (Biotaqk DNA Polymerase, Bioline, London, UK); 2.5 mM MgCl2; 50 AM (each) dATP, dCTP, dGTP,

144

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151

dTTP; 1 AM Primer (5V-AGCAGGGTCG-3V). Primers were supplied by Primm s.r.l. (Milan, Italy). Amplification conditions were: 20 cycles of 94 jC for 1 min, 29 jC for 1 min, 72 jC for 2 min (with a slope of 1.30 min from 29 to 72 jC) followed by 45 cycles of 94 C for 30 s, 55 jC for 30 s, 72 jC for 30 s and 72 jC for 5 min (Cocconcelli et al., 1997). The PCR products were separated on 3% SeaKem LE agarose gel (Cambrex Bioscience, Rockland, ME, USA) in Tris –acetate buffer at 100 V and then stained in 0.5 Ag ml 1 ethidium bromide solution (Sigma Aldrich, Milan, Italy). Gels were photographed with Kodak DC90 camera and RAPD profiles were analysed using the Diversity Database Fingerprinting Software ver. 2 (Bio-Rad Laboratories). The DNA molecular weight 1 kb ladder plus from Life Technologies Gibco BRL (Gaithersburg, MD, USA) was used as size standard. 2.5. 16S rDNA sequencing and molecular identification The DNA extracted as described above was used to amplify the 16S rDNA region with the primers P1 (5VGCGGCGTGCCTAATACATGC-3V) and P4 (5V-ATCTACGCATTTCCACCGCTAC-3V). PCRs were performed in a total volume of 50 Al in PTC-100 Thermal Cycler (MJ Research). Taq polymerase (Biotaqk DNA Polymerase London, UK) was purchased from Bioline and deoxynucleoside triphosphates and DNA molecular weight markers from Promega (Madison, WI, USA). rDNA amplification and purification were performed as already described (Cocconcelli et al., 1997). The sequence of the 16S rDNA gene amplified fragments was determined by using ABI PRISMk Big-Dye Terminator Cycle Sequencing (Perkin Elmer, Alameda, CA) following the supplier’s instructions. The primer P1 was used to determine the partial 16S rDNA gene sequence. The sequence analysis was obtained with the Applied Biosystem 373A Automated DNA Sequencer (Perkin Elmer). 2.6. Data analysis The software packages of the European Bioinformatic Institute (www.ebi.ac.uk) was used for the analysis and comparison of DNA sequences of isolated strains. Taxonomical identification and similarity

rank (S_ab) calculation were performed, comparing the 16S rDNA sequences of isolates, with the sequences present in the small sub-unit database (SSU-Prok) of Ribosomal Database Project (Maidak et al., 2001).

3. Results 3.1. Microbial composition of milk and milk starter culture In the course of this work, a single day production of the cheeses from the two different geographical areas of the natural park was studied, and the bacterial community development was analysed up to 60 days of ripening. Table 1 shows the microbiological composition of the milk samples collected from the two areas, Vanoi and Rolle, after the mixing of evening skimmed milk with the full fat morning milk, and of the NSC. In Vanoi milk, a mesophilic microbiota, rods and cocci, was found, reaching a level greater than 105 cfu g 1; in the same sample thermophilic cocci were detected at a lower level (104 cfu g 1). In contrast, in the milk sample from the Rolle area the counts did not exceed 105 cfu g 1, being mainly composed of mesophilic and thermophilic cocci, while mesophilic lactobacilli were not detected. Enterococci were also found in lower amounts. The bacterial composition of natural starter culture in milk was determined after 12 h of fermentation at 45 jC and a cooling step at 6 jC for 72 h. Plating revealed the presence of a large community of thermophilic cocci, 2  108 cfu g 1, and, a lower amount, of enterococci (103 cfu g 1). Thermophilic lactobacilli were not found in any of the analysed samples.

Table 1 Microbiological composition of the milk samples collected in the Vanoi and Rolle area and of the milk natural starter culture (cfu g 1) Mesophilic Mesophilic Thermophilic Enterococci bacilli cocci cocci Milk from 1.7  105 Vanoi Milk from < 100 Rolle Natural starter < 100 culture

4.4  105

1.2  104

< 100

8.6  104

9.5  104

1.5  103

< 100

2.0  108

1.0  103

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151

145

3.2. Bacterial community development during fermentation and ripening The evolution throughout the ripening process of the viable counts in the cheeses manufactured with milk from the two areas, Vanoi and Rolle, is reported in Figs. 1 and 2, respectively. In the milk added with the natural starter cultures from Vanoi, thermophilic cocci dominated the microbial population, as expected after the addition of the milk starter culture. Mesophilic rods and cocci were also present, although at lower levels, and thermophilic lactobacilli and enterococci constituted a low portion of the microbial community of milk. After 48 h of curd fermentation, thermophilic cocci were again one the most represented populations (2.2  107 cfu g 1), together with mesophilic cocci (1.6  107 cfu g 1) and, to a lesser extent, lactobacilli. After 15 days of ripening, the mesophilic rods dominated the microbial community (9.3  108 cfu g 1), which was composed also of mesophilic and thermophilic cocci (2.4  108 and 1.8  108 cfu g 1, respectively) and by the enterococci, which reached their maximum level in cheese (close to 107 cfu g 1). Mesophilic and thermophilic lactobacilli and mesophilic cocci predominated the bacterial community after 30 days, and 1 month later, a similar microbial composition of cheese was found. A different microbial community development was observed in the cheese produced using milk from the

Fig. 2. Microbial evolution during fermentation of curd and cheese obtained from milk from the Rolle area ( – 5 – , mesophilic lactobacilli; – n – , thermophilic lactobacilli; : : :o: : :, mesophilic cocci; : : : : : :, thermophilic cocci; — —, enterococci).

.

y

Rolle area, as reported in Fig. 2. The count of thermophilic cocci, present in the milk and added with the NSC, increased reaching 1  108 cfu g 1 after 48 h of fermentation. Moreover, these bacteria dominated the microbiota during the first week of cheese ripening. The other bacterial groups analysed during this work, which were already present in raw milk, increased in number during the first seven days, reaching values varying from 107 to 108 cfu g 1. Compared to Vanoi cheese, the thermophilic lactic acid bacteria and enterococci were in higher amounts, while bacteria enumerated at 30 jC were detected at similar levels. During the ripening, the different bacterial groups remained almost constant in number. After 15 days, as well as after 30 days of ripening, mesophilic rods and cocci were the predominant populations. In the Rolle cheese, the evolution of the enterococcal population was significant, since after 7 days and until the end of the ripening period, their number exceeded 107 cfu g 1 in the microbial community. 3.3. Analysis of population by means of RAPD strain typing

Fig. 1. Microbial evolution during fermentation of curd and cheese obtained from milk from the Vanoi area ( – 5 – , mesophilic lactobacilli; – n – , thermophilic lactobacilli; : : : o: : : , mesophilic cocci; : : : : : : , thermophilic cocci; — —, enterococci).

.

y

Since the aim of this study was, besides the identification of the bacterial community of Nostrano di Primiero, the recognition of the role of adventitious milk microbiota, a RAPD protocol, which had already proven to be effective in strain discrimination (Ba-

146

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151

ruzzi et al., 2000), was applied to type the dominant populations of NSC, milk and cheese ripened for 2, 15 and 30 days. The RAPD analysis revealed that the bacterial community of NSC was composed of three dominant strains (Fig. 3, lanes 3 – 6), since three different fingerprint patterns were found analysing the thermophilic cocci isolated from the milk culture. The population of Vanoi milk was dominated by mesophilic cocci, and two dominating populations were identified by RAPD fingerprints, as shown in Fig. 3 (lanes 1 and 2). Lactic cocci were also the dominant population in cheese after 2 days, where a more complex bacterial community was found, since four different RAPD profiles were detected at this stage (Fig. 4, lanes 1 –4). Lactobacilli became the most represented population in Vanoi cheese after 15 days and six different strains were detected (Fig. 5, lanes 3– 8). This bacterial group remained dominant in the bacterial community after 30 days of ripening, when five strains were identified (Fig. 5, lanes 1, 2, 9– 11). The milk produced in the Rolle area presented an heterogeneous coccal population, composed of eight strains, (Fig. 3, lanes 8 – 15), and these bacterial groups remained dominant after 2 days of fermentation, with a restricted number of biotypes identified (Fig. 4, lanes 6 – 8). At 15 days of cheese ripening,

lactobacilli were the principal population typed by RAPD analysis (Fig. 5, lanes 13 –22). Cocci dominated again after 30 days, when nine different strains were detected (Fig. 3, lanes 10– 16). The application of RAPD analysis on the dominant bacterial population during the fermentation and ripening processes of Nostrano di Primiero allowed the identification of 11 and 18 dominant biotypes in the Vanoi and Rolle dairy samples, respectively. 3.4. Taxonomical identification of dominant strains in Nostrano di Primiero The strains distinguished using RAPD fingerprint analysis were taxonomically identified by sequencing the 5V region of 16S rDNA. The composition of the studied microbiota is reported in Table 2, where percentages refer to the frequency of isolation of a single bacterial population among the analysed bacterial community. Differences in species composition between the bacterial community development of the two studied cheeses during ripening were detected, as reported in Table 2. The NSC was dominated by a microbiota composed of Streptococcus thermophilus, while Streptococcus suis was found in lower percentages. A lower cell number of Enterococcus faecium (103 cfu g 1) was detected in this starter culture.

Fig. 3. RAPD profiles of the dominating strains isolated from Vanoi milk (lanes 1 and 2), NSC (lanes 3 – 6) and from Rolle milk (lanes 8 – 15). Lane 1: Lactoc. lactis ssp. lactis NdP142; lane 2: Lactoc. lactis ssp. cremoris NdP147; lane 3: E. faecium NdP162; lane 4: S. suis NdP168; lane 5: S. thermophilus NdP169; lane 6: S. thermophilus NdP173; lane 7: molecular weight marker; lane 8: S. macedonicus NdP111; lane 9: S. thermophilus NdP113; lane 10: E. faecalis NdP115; lane 11: E. faecalis NdP122; lane 12: Lactoc. garviae NdP128; lane 13: Lactoc. garviae NdP154; lane 14: Lactoc. garviae NdP159; lane 15: S. macedonicus NdP160.

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151

147

Fig. 4. RAPD profiles of the dominating strains of cocci isolated in cheese from the two areas: lanes 1 – 4: cocci from Vanoi; lanes 6 – 17: cocci from Rolle. Lane 1: S. macedonicus NdP1; lane 2: E. faecalis NdP5; lane 3: E. faecium NdP10; lane 4: P. pentosaceus NdP11; lane 5: molecular weight marker; lane 6: S. macedonicus NdP15; lane 7: E. faecalis NdP19; lane 8: S. thermophilus NdP20; lane 9: E. faecium NdP25; lane 10 E. faecium NdP26; lane 11: E. faecalis NdP28; lane 12: Lactoc. lactis ssp. lactis NdP31; lane 13: Lactoc. lactis ssp. lactis NdP33; lane 14: Lactoc. lactis ssp. cremoris NdP38; lane 15 E. faecium NdP40; lane 16: E. faecalis NdP45; lane 17: E. faecalis NdP30.

A different microbiota composition was revealed in the milk collected in the two areas: the bacterial strains isolated from Vanoi milk were Lactococcus lactis ssp. cremoris, Lactococcus lactis ssp. lactis, Lactobacillus paracasei ssp. paracasei and, to a lesser extent, Streptococcus macedonicus. On the contrary, in Rolle milk, the community was mostly constituted of thermophilic bacteria with a predominance of S. thermophilus strains and a minority presence of S. macedonicus, and Enterococcus faecalis. Lower amounts of bacteria belonging to the Lactoc. lactis and Lactococcus garviae species were isolated from this sample, too. After 2 days of fermentation the dominant microbiota in Vanoi cheese was mainly constituted of S. macedonicus and E. faecium while in the Rolle cheese the thermophilic cocci S. mace-

donicus and S. thermophilus were detected. Differences were observed also during the ripening: in Vanoi cheese at 15 days the predominant species was Lb. paracasei, and one strain of 21 Lb. plantarum and one of Lactobacillus zeae were also isolated. At 30 days the dominant bacteria were identified as Lb. paracasei, Lactobacillus fermentum and 21 Lb. plantarum. In contrast, in 15-days Rolle cheese a complex bacterial community of mesophilic rods, belonging to the species 21 Lb. paracasei, 21 Lb. fermentum, Lactobacillus pentosus, 21 Lb. plantarum, Lactobacillus rhamnosus, was found. Pediococcus pentosaceus was also isolated at this stage. At 30 days, enterococci, E. faecium and E. faecalis, became the dominant group while Lactoc. lactis ssp. lactis and Lactoc. lactis ssp. cremoris were isolated in lower numbers.

Fig. 5. RAPD profiles of the dominating strains of rods isolated from the two areas: lanes 1 – 11: rods from Vanoi; lanes 13 – 22: rods from Rolle. Lane 1: 21 Lb. plantarum NdP74; lane 2: 21 Lb. paracasei ssp. paracasei NdP76; lane 3: 21 Lb. paracasei ssp. paracasei NdP52; lane 4: 21 Lb. paracasei ssp. paracasei NdP57; lane 5 21 Lb. zeae NdP58; lane 6: 21 Lb. plantarum NdP59; lane 7: 21 Lb. paracasei ssp. paracasei NdP65; lane 8: 21 Lb. paracasei ssp. paracasei NdP66; lane 9: 21 Lb. paracasei NdP78; lane 10: 21 Lb. fermentum NdP79; lane 11 21 Lb. paracasei NdP81; lane 12: molecular weight marker ; lane 13: 21 Lb. fermentum NdP87; lane 14: 21 Lb. paracasei NdP88; lane 15: 21 Lb. paracasei NdP90; lane 16: 21 Lb. paracasei ssp. paracasei NdP92; lane 17: 21 Lb. rhamnosus NdP93; lane 18: 21 Lb. paracasei ssp. paracasei NdP94; lane 19: 21 Lb. fermentum NdP95; lane 20: 21 Lb. pentosus NdP96; lane 21: 21 Lb. plantarum NdP98; lane 22: 21 Lb. paracasei ssp. paracasei NdP101.

148

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151

Table 2 Microbial composition (%) of the dominant microbiota in natural starter culture, in milk, cheese after 2, 15 and 30 days from the Vanoi and Rolle areas NSCa Vanoi

Rolle

Milk Cheese (days)

Milk Cheese (days)

2 15 30 E. faecalis NdP115 E. faecalis NdP19 E. faecalis NdP30 E. faecalis NdP5 E. faecium NdP10 E. faecium NdP25 14 Lb. fermentum NdP79 14 Lb. paracasei NdP88 14 Lb. paracasei NdP101 14 Lb. paracasei NdP52c 14 Lb. paracasei NdP57c 14 Lb. paracasei NdP65c 14 Lb. paracasei NdP66c 14 Lb. paracasei NdP92c 14 Lb. paracasei NdP94c 14 Lb. pentosus NdP96 14 Lb. plantarum NdP59 14 Lb. rhamnosus NdP93 14 Lb. zeae NdP58 Lc. garviae NdP154 Lc. garviae NdP159 Lc. lactis NdP123d Lc. lactis NdP147d Lc. lactis NdP38d Lc. lactis NdP142e Lc. lactis NdP31e Ped. pentosaceus NdP11 Ped. pentosaceus NdP110 Str. macedonicus NdP1 Str. macedonicus NdP153 Str. suis NdP168 Str. thermophilus NdP114 Str. thermophilus NdP169 Str. thermophilus NdP173 Str. thermophilus NdP20

2 15 30 5 5 3

11

17 17

10 43

22 22

+b 27

16 11 16

6 6 9 18 27 12 27

15 10

5 5 16 16 5 5

35 10 23 16 9 3 45

13 25 4

9

5 5 5

42

6 6

44

18 6 23 59 37

45

a

NSC: natural starter culture. b E. faecium NdP25 was detected in NSC at 103 CFU g onto KAA plates. c Lb. paracasei spp. paracasei. d Lc. lactis spp. cremoris. e Lc. lactis ssp. lactis.

3.5. Population dynamics The comparison of data deriving from RAPD typing and molecular identification allowed the study

of population dynamics and community development during the fermentation of Nostrano di Primiero. As previously observed (Bottazzi, 1967), the NSC microbiota resulting mainly comprised of thermophilic cocci. In particular, two strains of S. thermophilus represented 82% of isolates and one strain of S. suis the remaining 18%. Moreover, one strain of E. faecium was detected at a low percentage in NSC and this then became a dominant strain (22%) after 30 days of ripening in Rolle cheese. Differently the two dominant strains of S. thermophilus were no longer isolated in either analysed cheese at 48 h or in later samples. The molecular analysis showed a significant numbers in the microbiota of Vanoi milk of mesophilic cocci, identified as Lactoc. lactis ssp. lactis and ssp. cremoris, which were not isolated in other cheese samples. Otherwise, the S. macedonicus NdP1, isolated from milk, was found also in 2 days fermented cheeses from both areas. Two strains of mesophilic Lb. paracasei ssp. paracasei, NdP65 and NdP66, detected in the Vanoi milk in significant amount were members of the dominant population also in the cheeses at 15 and 30 days. The two strains Lb. paracasei ssp. paracasei NdP57 and Lb. plantarum NdP59 were isolated at 15 and 30 days. A completely different situation was observed in the case of the milk and cheese from the Rolle area. Eleven independent strains belonging to six different species were detected in raw milk, four of them occurring also in the subsequent phases of cheese ripening. In milk and at the beginning of cheese ripening S. thermophilus and S. macedonicus dominated the microbial fermentation. An E. faecalis NdP19 and Lactoc. lactis ssp. lactis NdP31 were identified in milk and detected also during cheese ripening. The two strains of Lactoc. garviae, detected in milk, disappeared during the first days of cheese fermentation. The microbiota of 15-day-old cheese was mainly composed of five strains of Lb. paracasei and by one of Lb. fermentum, which were not detected in previous analysed samples. Five strains, E. faecium NdP10, Lb. fermentum NdP79, Lb. paracasei NdP66, Lactob. plantarum NdP59 and S. macedonicus NdP1, were isolated from both cheese samples, indicating the presence of cross contamination during cheese manufacture or deriving as adventitious microbiota from the production environment.

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151

4. Discussion Nostrano di Primiero cheese is produced using milk collected from alpine pastures located in the natural park of Paneveggio –Pale di San Martino, an area characterised by a low anthropic impact. Previous studies have shown that in this type of raw milk cheese, besides dairy technology, a relevant role on microbial development is played by environmental contamination of milk (Grappin and Beuvier, 1998), which influence the final sensorial properties of traditional products. Nostrano di Primiero is produced in dairy farms using milk collected from two areas, Rolle and Vanoi, processed by using identical methods and technologies at the same time. In addition, during production, the same NSC is used, ripening time and conditions. Since the only difference between the two processes is the milk’s origin, this cheese could be considered a valuable model for the study of the role of milk adventitious microbiota. In order to follow the community development and the population dynamics, we chose to study a singleday production and to analyse the same cheeses throughout the ripening process. This approach allowed to monitor the growth and persistence of single bacterial strains, the variation of the dominance of microbial communities and to compare the population dynamics of the cheeses from the two geographical areas. A first difference was observed analysing the bacteria counts and composition of microbial community of the two overnight skimmed milks, the Vanoi milk exceeded 105 cfu g 1 of lactic acid bacteria, while lower levels were found in Rolle. Moreover, the analysis of isolated strains resulting by means of RAPD typing and 16S rRNA identification showed a large and different biodiversity in the Vanoi and Rolle samples. The characterization of the dominant microbiota at different ripening stages of cheese manufactured in the same way with two milk bulks originating from distinct geographic areas showed different population dynamics, linked to the microbial composition of the raw milk. The milk originating from Vanoi was richer in mesophilic microbiota (>105 cfu g 1), particularly in lactobacilli and lactococci, which grew to significant levels (6  108 cfu g 1) in the cheese after 7 days

149

of ripening, and dominated the microbial community until the end of the observation period. Otherwise, thermophilic cocci and enterococci were higher in number in the milk coming from the Rolle area, and their high level was detected in 2-day ripened cheese and after 30 days. Also, in this cheese, the 15-day bacterial community was dominated by mesophilic lactobacilli, present in the milk at a very low level ( < 100 cfu g 1) and increased up to 108 cfu g 1 after 15 days. Since in the cheese making of Nostrano di Primiero, a natural starter culture is used to promote the acidification process, its microbiota was studied. Two strains of S. thermophilus were found to compose more than 80% of the NSC bacterial community, as observed in a previous study on milk natural starter cultures (Bottazzi, 1967). In addition, in the same sample, S. suis NdP168 and E. faecium NdP25 were found at lower levels. After 2 days of cheese fermentation, none of these strains were detected in the analysed community of both cheeses, while in the 30-day ripened Rolle cheese, the E. faecium NdP25 strains appeared as part of the dominant microbial community. In both milk samples, S. macedonicus was detected, and the strain NdP1 was found be the dominant strain in the microbial community of both cheeses after 2 days of fermentation. The different composition of microbiota resulting from the analysis of the raw milk from the Vanoi and Rolle area underlines the influence of the environment on the milk contamination. This difference continues also in the Vanoi and Rolle cheeses during their ripening; the identified species correspond to the NSLAB most frequently isolated from cheeses during ripening (McSweeney et al., 1993; Gobetti et al., 1999; Albenzio et al., 2001), thus pointing out the importance of microbiota coming from the environment in the production of traditional cheeses. Moreover, the two cheeses obtained from the relative milks were different concerning their organoleptic characteristics and the texture (data not shown), and this observation could be related to the diverse microbial evolution and population composition observed. In fact, the cheese made with the milk from the Vanoi area was rich in Lb. paracasei and other mesophilic rods that were detected both in the raw milk and in the cheese at 15 and 30 days; on the other hand, the Rolle cheese was dominated by different NSLAB, in par-

150

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151

ticular, enterococci already present in the milk. The NSLAB generally constituted of mesophilic lactobacilli can be responsible for the rate of maturation and can contribute to the formation of the typical flavour of this kind of product (Albenzio et al., 2001), but in Nostrano di Primiero cheese, there were also detected typical thermophilic species such as S. macedonicus. In the NSLAB isolated during the time of observation study, only five strains were detected in both cheeses. These bacteria, representing 17% of the whole dominant microbial community, were probably present in the dairy equipment, in the dairy environment or in NSC at low levels. The residual 73% of the isolated NSALB strains were typical for either one of the two products. The study of Nostrano di Primiero by means of molecular techniques revealed that the adventitious NSLAB present in the raw milk or contaminating the cheese during manufacture, due to their tolerance to the hostile environment of cheese during ripening, play a fundamental role in the community development of this raw milk cheese. In addition, Nostrano di Primiero has been demonstrated to be a good model for comprehending the importance of the indigenous microbiota of raw milk, and the role played in traditional cheeses ripening with NSC and the adventitious microbiota originating in the milk or environment.

Acknowledgements We thank Caseificio Sociale di Primiero and Natural Park of Paneveggio –Pale di San Martino for the kind collaboration and Dr. Maria Giulia Parisi for technical assistance. This work was supported by Caseificio Sociale di Primiero and by project MiPAF ‘‘Valorizzazione e salvaguardia della microflora caratteristica delle produzioni casearie italiane’’ article no 22.

References Albenzio, M., Corbo, M.R., Rehman, S.U., Fox, P.F., De Angelis, M., Corsetti, A., Sevi, A., Gobetti, M., 2001. Microbiological and biochemical characteristics of Canestrato Pugliese cheese made from raw milk, pasteurized milk, or by heating the curd

in hot whey. International Journal of Food Microbiology 67, 35 – 48. Baruzzi, F., Morea, M., Maturante, A., Cocconcelli, P.S., 2000. Changes in the Lactobacillus community during Ricotta forte cheese natural fermentation. Journal of Applied Microbiology 89, 807 – 814. Beresford, T.P., Fitzsimons, N.A., Brennan, N.L., Cogan, T.M., 2001. Recent advances in cheese microbiology. International Dairy Journal 11, 259 – 274. Beuvier, E., Berthaud, K., Cegarra, S., Dasen, A., Pochet, S., Buchin, S., Duboz, G., 1997. Ripening and quality of Swiss-type cheese made from raw, pasturized or microfiltered milk. International Dairy Journal 7, 311 – 323. Bottazzi, V., 1967. Microbiology of natural starter culture in milk. Scienza e Tecnica Lattiero-Casearia 18, 14 – 23. Cappa, F., Cattivelli, D., Cocconcelli, P.S., 2002. Study of microbial community in Fromadzo cheese produced in alpine pastures. Scienza e Tecnica Lattiero-Casearia 53, 127 – 138. Cocconcelli, P.S., Parisi, M.G., Senini, L., Bottazzi, V., 1997. Use of RAPD and 16S rDNA sequencing for the study of Lactobacillus population dynamics in natural whey culture. Letters in Applied Microbiology 25, 8 – 12. Corroler, D., Mangin, I., Desmasures, N., Gueguen, M., 1998. An ecological study of lactococci isolated from raw milk in the Camembert cheese registered designation of origin area. Applied and Environmental Microbiology 64, 4729 – 4735. Demarigny, Y., Beuvier, E., Dasen, A., Duboz, G., 1996. Influence of raw milk microflora on the characteristics of Swiss-type cheeses: I. Evolution of microflora during ripening and characterization of facultatively heterofermentative lactobacilliLe Lait 76, 371 – 387. Demarigny, Y., Beuvier, E., Buchin, S., Pochet, S., Grappin, R., 1997. Influence of raw milk microflora on the characteristics of Swiss-type cheeses: II. Biochemical and sensory characteristics. Le Lait 77, 151 – 167. Gobetti, M., Folkertsma, B., Fox, P.F., Corsetti, A., Smacchi, E., De Angelis, M., Rossi, J., Kilcawley, K., Cortini, M., 1999. Microbiology and biochemistry of Fossa (pit) cheese. International Dairy Journal 9, 763 – 773. Grappin, R., Beuvier, E., 1998. Possibile implications of milk pasteurization on the manufacture and sensory quality of ripened cheese. International Dairy Journal 7, 751 – 761. Maidak, B.L., Cole, J.R., Lilburn, T.G., Parker Jr., C.T., Saxman, P.R., Farris, R.J., Garrity, G.M., Olsen, G.J., Schmidt, T.M., Tiedie, J.M. 2001. The RDP-II (Ribosomal Database Project). Nucleic Acids Research 29, 173 – 174. Mannu, L., Paba, A., 2002. Genetic diversity of lactococci and enterococci isolated from home-made Pecorino Sardo ewes’ milk cheese. Journal of Applied Microbiology 92, 55 – 62. McSweeney, P.L.H., Fox, P.F., Lucey, J.A., Jordan, K.N., Cogan, T.M., 1993. Contribution of the indigenous microflora to the maturation of cheddar cheese. International Dairy Journal 3, 613 – 634. Morea, M., Baruzzi, F., Cocconcelli, P.S., 1999. Molecular and physiological characterization of dominant populations in traditional Mozzarella cheese processing. Journal of Applied Microbiology 87, 574 – 582.

E. Poznanski et al. / International Journal of Food Microbiology 92 (2004) 141–151 Senini, L., Cappa, F., Cocconcelli, P.S., 1997. Use of rRNA-targeted oligonucleotide probes for the characterization of the microflora from fermentation of Fontina cheese. Food Microbiology 14, 469 – 476. Shakeel-Ur, R., Banks, J.M., McSweeney, P.L.H., Fox, P.F., 2000. Effect of ripening temperature on the growth and significance

151

of non-starter lactic acid bacteria in Cheddar cheese made from raw or pasteurised milk. International Dairy Journal 10, 45 – 53. Wouters, J.T.M., Ayad, E.H.E., Hugenholtz, J., Smit, G., 2002. Microbes from raw milk for fermented dairy products. International Dairy Journal 12, 91 – 109.