Microbiological Quality and Variability of Natural Microbiota in Croatian Cheese Maturing in Lambskin Sacks

Cheese in lambskin or sheepskin sacks (Sir iz mišine) is one of Croatian traditional farm produces dating back to the times when Illyrians and Thracians herded their sheep on the pastures of Dinara Mountain. They used sheepskin sacks to store and transport cheese from the mountain to the sett lements in the valley (1). Similar cheese matured in animal skin sacks is also produced in Bosnia and Herzegovina, Montenegro, Turkey and Lebanon (2).


Introduction
Cheese in lambskin or sheepskin sacks (Sir iz mišine) is one of Croatian traditional farm produces dating back to the times when Illyrians and Thracians herded their sheep on the pastures of Dinara Mountain.They used sheepskin sacks to store and transport cheese from the mountain to the sett lements in the valley (1).Similar noleptic properties and to indigenous microorganisms living in raw milk or natural starters (4).However, raw milk can also contain pathogenic bacteria that have been raising public health concern since the beginnings of dairy industry.The most common pathogenic bacteria in raw milk and milk products are Salmonella, Listeria monocytogenes, Staphylococcus aureus and Escherichia coli (5).
It is therefore necessary to characterise traditional cheese not only to protect its diversity but also to control the quality (6).Yet artisanal cheese matured in animal skin has been poorly characterised, except for some Turkish and Lebanese varieties (7)(8)(9).
It is known that the microbiological quality of dairy products can be improved by adding a starter culture such as lactic acid bacteria (LAB) that will prevent the growth of pathogenic microorganisms (5,(10)(11)(12)(13).The best way to preserve specifi c qualities of traditional cheese is to use an indigenous starter culture.Yet, lambskin cheese is traditionally produced without adding one.
The aim of our study is to characterise this unique product for the fi rst time and to determine the natural microbiota in traditional dairy products.

Cheese production and sampling
Given that there are about ten small producers of cheese in lambskin sacks of similar size in the area of Knin (Croatia), we picked at random three that produce cow's milk cheese (sampling locations I-III) and three that produce sheep's milk cheese (sampling locations IV-VI).All samplings took place from May to July 2013.
For each cheese sample about 60 L of milk were taken.Evening milk was cooled to 7-12 °C, depending on the farm, and mixed with fresh morning milk.All producers used the same commercial microbial rennet in the ratio of 1:10 000 (Milaco, Split, Croatia) at temperatures between 32 and 37 °C.Aft er 30 to 50 min, the curd was scooped into small irregular pieces or cut with a knife into 3 cm×3 cm×3 cm cubes.It was then heated to 34 to 40 °C (the temperature varied by farm) and stirred by hand or a scoop for 20 to 30 min, at which point the fi rst whey was drained.The curd was then shaped by hand, scooped and left to drain overnight through a cloth under its own weight.Aft er the rest of the whey was removed, the curd was cut into 10 cm×9 cm×5 cm pieces, salted with coarse sea salt and stuff ed in lambskin sacks.Cheese matured in lambskins for 45 days at 16 to 18 °C and relative humidity from 65 to 80 %.The average mass of mature cheese was about 20 kg.
The lambskin sacks in which the cheese matures are made by gently skinning a freshly slaughtered lamb and carefully scraping off the remaining fl esh and fat to avoid tearing.The skin is then cleaned with water, sun-dried, and fi nally smoked.Before it is fi lled with curd, the skin is submerged in whey or hot water.All openings are then sealed tight with a rope, except the neck, which is left open for stuffi ng with curd.
For microbiological analysis one of the following samples was taken per farm: raw cow's and sheep's milk, curd and cheese matured for 45 days.Milk samples were taken on the same day aft er the morning milking and curd samples immediately aft er coagulation.
In addition, two more samples of cheese were taken from each farm on days 15 and 30 of ripening for determination of LAB and staphylococci.Also, three empty lambskin sacks were taken for sampling.
Cheese samples (about 100 g per sample) were taken with a sterile knife through the neck opening of the sack.Curd (day 0) and cheese (days 15, 30 and 45 of ripening) were taken from each batch and transported to the laboratory on the same day in a mobile cooler at 4 °C.About 0.5 cm of the sample surface were scraped off and the rest was stored at -80 °C until analysis.Aft er each sampling, the skin sacks were tied up tight again and left in the ripening room at the family farm.

Microbiological analyses
Classical microbiological methods were used to isolate and identify pathogenic bacteria (aft er 0, 2 and 45 days of storage) and for the isolation of staphylococci and LAB, the samples stored for 15 and 30 days were also used.At the same time lamb sacks were analysed for indigenous microbial populations.
A mass of 10 g of each sample was homogenised in 90 mL of sterile 0.5 % NaCl solution for 3 min using a Stomacher Lab Blender (Seward Limited, Worthing, UK) and serially diluted before plating (pour plate method for total aerobic mesophilic bacteria and spread plate method for other bacteria) on selective media.Total aerobic mesophilic bacteria were counted aft er incubation on nutrient agar (Merck, Darmstadt, Germany) at 37 °C for 48 h, Enterobacteriaceae aft er incubation on Violet Red Bile Glucose (VRBG) agar (Biolife Italiana Srl, Milan, Italy) at 37 °C for 48 h, E. coli aft er incubation on RAPID'E. coli 2 agar (Biolife) at 37 °C for 48 h, Staphylococcus aureus aft er incubation on Baird-Parker (BP) agar (Biolife) at 37 °C for 24 h, mould and yeasts aft er incubation on potato dextrose agar (Biolife) at 25 °C for 96 h, and lactic acid bacteria after incubation on de Man, Rogosa and Sharpe (MRS) agar (Biolife) at 30 °C for 48-72 h.Salmonella sp. was grown in Rappaport-Vassiliadis (RV) Salmonella enrichment broth (Merck) according to ISO 6579:2002 method ( 14), followed by subculturing on xylose lysine deoxycholate (XLD) agar (Biolife) at 37 °C for 24-48 h.Listeria monocytogenes was detected using a two-step selective enrichment procedure in Fraser broth according to ISO 11290-1:1996 method ( 15), followed by subculturing on PALCAM agar (Merck) at 37 °C for 24 h.The microbial growth was determined using traditional plate counting and the results are expressed as the logarithm of colony forming units per gram of cheese (log CFU/g).

Identifi cation of microbial population
To identify the indigenous microbial population, standard biochemical API ® 50 CH (for LAB) and API ® staph (for staphylococci) identifi cation strips (bioMérieux, Vienna, Austria) were used.As many as 250 colonies were analysed in both cow's and sheep's milk, curd and cheese samples (aft er 0, 2, 15, 30 and 45 days of storage) as well as 80 colonies obtained from the lambskin sack samples.All colonies were collected aft er incubation on different agar media (MRS and BP).API identifi cation of LAB and staphylococcal isolates was confi rmed using a Microfl ex LT™ matrix-assisted laser desorption-ionisation time-of-fl ight mass spectrometer (MALDI-TOF MS; Bruker Daltonik, Bremen, Germany) according to Babić et al. (11).For identifi cation, the peaks from the generated mass spectra were compared with the reference spectra of the integrated database using the MALDI Biotyper Software package (Bruker Daltonik).

Determination of enterotoxins
To screen staphylococcal isolates from cow's and sheep's milk for potential ability to produce enterotoxins and to detect the presence of staphylococcal enterotoxins in dairy samples, the TECRA Staphylococcal Enterotoxins (SET) Visual Immunoassay (VIA) kit (TECRA Company, Kassel, Germany) was used.The kit can detect staphylococcal enterotoxins A (SEA), B (SEB), C (SEC), D (SED) and E (SEE).

Statistical analysis
All measurements/counts were carried out in triplicate.The results are expressed as mean value±standard deviation (S.D.).

Results and Discussion
Earlier studies on the composition, hygiene and quality of the cheese produced from raw milk in diff erent countries have shown signifi cant variations in hygiene standards, but all warn of health risks for the consumer (8,12,16).In this sense, our study expands and supports the knowledge gathered so far with the fi ndings on Croatian cheese in lambskin sacks, which is produced from raw cow's or sheep's milk on family farms in poorly controlled conditions, mostly during the summer.To establish the safety of cheese in lambskin sacks, our fi ndings were compared with the Croatian cheese safety standards (17).
The microbiological analysis showed that two-thirds of the samples did not meet the safety criteria due to the high S. aureus and Enterobacteriaceae counts (Tables 1 and  2).However, colonies of Salmonella and L. monocytogenes were not found in any of the samples.Colonies of E. coli were not found in cow's milk but were present in sheep's milk samples (Table 2).
Milk can be contaminated by these microorganisms because of the poor hygiene, or animal udders and vessels that are not cleaned before milking.The high micro- organism count can also be the consequence of poor handling and storage conditions.Overnight storage of milk at 7.2 °C allows microorganisms to grow much faster than at lower temperatures (4,12).
The microbiological quality of lambskin sack samples was satisfactory, as no pathogens were found (results not shown).
As the limit of S. aureus of 2 log CFU/g (17) was exceeded in two-thirds of milk, curd and cheese samples, all samples were tested for staphylococcal enterotoxins (SE), which resulted in detecting staphylococcal enterotoxin C (SEC) in two sheep's milk cheese samples (Table 2).These results indicate that this phase of sheep's milk cheese production is also at risk and that there is room for improvement in the production hygiene and/or selection of raw materials.The increase in S. aureus count in the curd during the fi rst days of ripening (Table 2) suggests the possible risk at this stage of production and the need to take measures to improve the hygienic conditions.Various authors have confi rmed rapid increase in S. aureus at the beginning of ripening (18,19).The conditions under which cheese in lambskin sacks is processed are poor, which can be one of the major sources of contamination of the curd with S. aureus besides raw milk.SEC in the two sheep's milk cheese samples may also originate from animals suffering from mastitis.Our results are in agreement with literature sources which claim that only a small percentage of coagulase-positive staphylococci (CPS) is potentially enterotoxic and explain why SE are also rarely detected in the samples where CPS exceeds the values of 10 5 CFU/g (14,(20)(21)(22).
In contrast, only three cow's milk and curd samples and none of the cheese samples had S. aureus counts above the safety limit.One of the reasons for the bett er microbiological quality of cow's milk could be the higher lactic acid bacteria (LAB) counts that were enumerated in this milk compared to the sheep's milk (Tables 1 and 2), which confi rms the fi ndings by Durmaz (23), Kells and Gilmour (24), Lamprell et al. (25) and Schoder et al. (26).Many studies (5,(10)(11)(12)(13) have shown that lactic acid bacteria added to traditional fermented food as starter cultures can inhibit the growth of pathogenic and spoilage microorganisms, mainly due to the production of lactic acid, but they can also produce many other antimicrobial metabolites such as bacteriocins, organic acids, hydrogen peroxide, etc.
However, many traditionally fermented dairy products in Croatia and elsewhere are produced without the addition of starter cultures.To protect their specifi c qualities such as fl avour and aroma, it is essential to understand the indigenous microbial interactions during fermentation and to isolate and characterise autochthonous starter cultures that can improve their safety (10,13,27).
Using biochemical analysis nine indigenous bacterial species were identifi ed in milk, curd and cheese samples: Lactobacillus plantarum, Lactococcus lactis, Leuconostoc mesenteroides, Lactobacillus paracasei, Lactobacillus curvatus, Lactobacillus brevis, Staphylococcus xylosus, S. aureus and S. epidermidis, with 99-99.9% accuracy (Table 3).The prevalent strains of LAB were Lactococcus lactis and Lactobacillus paracasei in dairy samples from cow's milk and Leuconos-  4).These fi ndings give an important insight into the fermentation and microbial ecology of cheese in lambskin sacks and suggest that the isolated LAB and staphylococcal strains (S. xylosus) may play an important role in defi ning its recognisable and desired aroma.The exact role of these bacteria in the taste formation remains to be elucidated.

Conclusion
Our study was the fi rst to isolate and identify the indigenous microbial population in Croatian traditional cheese in lambskin sacks obtained from either raw cow's or sheep's milk.It has shown that sheep's milk, curd and cheese in lambskin sacks contain bacteria that are generally considered indicators of poor hygiene, but indigenous LAB were also identifi ed that may serve as starter cultures and improve cheese quality and safety.We consider these fi ndings as a fi rst step in future research to identify indigenous LAB that can be used as starter cultures (through their technological and functional characterisation) and which hygiene improvements to use, without aff ecting the unique organoleptic properties of cheese in lambskin sacks.This will set the grounds for obtaining the quality scheme Protected Designation of Origin, which would make this product desirable for potential markets in Croatia and all over the world.

Table 2 .
Results of microbiological analyses and detection of staphylococcal enterotoxins in the samples of raw sheep's milk, curd and ripened cheese in lambskin sacks *not satisfactory criterion (≤2 log CFU/g), **not satisfactory criterion (≤5 log CFU/g), LD=limit of detection Table1.Results of microbiological analyses and detection of staphylococcal enterotoxins in the samples of raw cow's milk, curd and ripened cheese in lambskin sacks

Table 3 .
Results of biochemical identifi cation of natural microbial populations of raw cow's and sheep's milk, curd and ripened cheese in lambskin sacks Lactobacillus plantarum in those from sheep's milk.In lambskin sacks the prevalent lactobacilli were Leuconostoc mesenteroides and Lactobacillus plantarum (Table

Table 4 .
Results of biochemical identifi cation of microbial population in the lambskin sacks