The prevalence of Listeria monocytogenes and Staphylococcus aureus and their virulence genes in bulk tank milk in Kosovo

Introduction: Milk is considered to be a healthy, nutritious food product. Microbiological quality is an important aspect in evaluating the quality of milk. Methodology: A total of 603 bulk tank milk samples from 221 farms distributed across ten different regions were collected for milk quality assessment. Quality was judged by total viable count, and the prevalence of two foodborne pathogens (Listeria monocytogenes and Staphylococcus aureus) by using selective media and 16S rRNA gene sequencing. The presence of virulence genes was detected by polymerase chain reaction (PCR) using specific primers. Results: Milk from only 7% (15/221) of farms were found to comply with the European Union standard. Interestingly, the microbiological quality of milk from the larger herd size farms (more than 10 cows) was better than in smaller herds. L. monocytogenes was found in 2.7% (6/221) of farms, and all the examined L. monocytogenes isolates were positive with respect to the virulence genes prfA, actA, and hlyA. S. aureus was found in 39.8% (88/221) of the farms. In total, 30.7% (27/88) of the staphylococci were positive for enterotoxin production. The enterotoxins identified were toxin B (40.7%), toxin D (33.4%), toxin C (18.5%), and toxin A (7.4%). Conclusions: The total number of bacteria in milk was very high. The presence of two foodborne pathogens in raw milk represents a great health risk to consumers. To improve the microbial quality of milk in Kosovo, important measures to improve hygiene, including better information, guidance, and control, are needed.


Introduction
Milk is a nutritious food product for humans, and it is obtained from a variety of animal sources, such as cows, goats, sheep, and buffaloes.Bovine milk contains lipids, proteins (caseins, whey), carbohydrates (lactose), amino acids, vitamins, and minerals (particularly calcium) essentials of good nutrition [1].However, due to its nutritious properties, milk is also known as an excellent growth environment for both non-pathogenic and pathogenic bacteria [1,2].Staphylococcus aureus, Salmonella spp., Listeria monocytogenes, and Escherichia coli, which are normally considered significant pathogens, are often found in milk or milk products in both industrialized and developing countries [3][4][5].L. monocytogenes can cause meningitis, septicemia, and abortion in humans and animals [6][7][8].S. aureus is the main cause of subclinical mastitis in animals, and some strains have the ability to produce toxins, especially when present in high cell numbers (> 10 5 colony-forming units (cfu)/mL).Staphylococcal food infection in humans is often accompanied by vomiting, diarrhea, and abdominal cramps [9,10].The causes of contamination are many, including loss of udder integrity of infected animals, poor hygiene, or practices leading to further spreading the pathogens in farms, milking production facilities, and during the milking process [11].Numerous studies have shown that factors such as animal health, lactation, season, climate, animal breed, feeding practice, housing, hygiene of premises, milking practice and storage, and milking facilities and equipment play important roles in the quality of bulk tank milk [12][13][14][15][16].
Kosovo is a young and small country located in the Balkan region.Milk, which is an important part of daily nutrition in the country, is produced in a quantity of approximately 257,500 tons per year [17].The majority of dairy farms (94%) in Kosovo are traditional type (also called subsistence farms), 5% of the farms are semi-commercial farms, and only 1% are commercial farms; the differentiation between the latter two is based on the number of cows present on the farm [17,18].Subsistence farmers produce milk mostly for their own consumption, while semi-commercial or commercial farmers sell their milk at small nearby dairy factories or at local markets.In European countries, the milk quality is much better compared to that in developing countries.This may be due to more rigorous inspection control practices from authorities, better managing of the farms, better hygiene, and better infrastructure in general [19,20].To our knowledge, no systematic microbial analysis has been done on bulk tank milk in Kosovo, especially with regard to the problem of foodborne pathogens.Hygiene and good management of bulk tank milk and its products are some of the important aspects in achieving high microbiological quality, and this issue needs to be addressed before Kosovo can export dairy products, especially to the European Union market, where adherence to high quality standards is required.In this study, we systematically evaluated the microbiological quality of raw cow milk in Kosovo in general and with special focus on the prevalence of two foodborne pathogens (L.monocytogenes and S. aureus).

Sample collection
Between December 2011 and June 2012, bulk tank cow milk samples were collected from 221 farms in Kosovo.These farms were chosen from ten different regions: Decan (15 farms), Ferizaj (26 farms), Gjakova (15 farms), Istog (21 farms), Lipjan (20 farms), Podujeva (15 farms), Rahovec (60 farms), Skenderaj (20 farms), Sharr (16 farms), and Theranda (13 farms).To obtain a reliable and consistent dataset, milk samples were collected every second month and three times from each farm except for the farms from Rahovec, where samples were collected only at two time points due to technical problems.This yielded 603 samples in total.The samples were kept in sterile plastic tubes and transported in an ice box (4°C-8°C) to the laboratory within 4 hours of pick-up at collection points, and the samples were immediately processed for further analyses.To collect information about herd size, milking system, cow breed, and health status, a questionnaire was prepared and completed during personal interviews with the owners.

Standard plate counting (SPC)
Total viable counts (TVCs) of microorganisms were determined following the guidelines of International Standards of Organization (ISO 4833:2003) using plate count agar (PCA) [21].Briefly, collected samples were serially diluted and 100 µL from higher dilutions (10 -8 to 10 -2 ) was plated on PCA plates (Oxoid, Basingstoke, United Kingdom) and incubated aerobically for 72 hours at 30°C before total colonyforming units were determined.

Isolation and enumeration of S. aureus
Isolation of S. aureus from bulk tank milk was performed according to the international standard procedure (ISO 6888-1:1999) using Baird-Parker agar supplemented with egg yolk tellurite emulsion (Oxoid, Basingstoke, United Kingdom) [22].The plates were incubated under aerobic conditions at 37°C for 24 to 48 hours.From each milk sample, five randomly selected egg yolk reaction-positive colonies (an indication of S. aureus) and five egg yolk reaction-negative isolates (for other Staphylococcus spp.) were selected for further characterization.All isolates were tested for Gram staining and catalase activity, hemolytic properties, and their ability to coagulate rabbit plasma (tube coagulase test) and to produce clumping factor (Staphylase test; Oxoid, Basingstoke, United Kingdom).The strains that showed similarity were grouped, and only one isolate from each group was frozen in stab agar and further analyzed.Colonies were re-grown on brain-heart infusion (BHI) broth and stored at -80°C with 13% glycerol.

DNA isolation, PCR, and DNA sequencing
Bacterial genomic DNA was extracted by using a modified purification protocol described previously [23].PCR was done on genomic DNA to amplify 16S rRNA genes, using appropriate primers described previously (Table 1) [24,25].The thermal profile used for PCR was at 90°C for 1 minute, and 35 cycles at 95°C for 15 seconds, 54°C for 30 seconds, and 70°C for 90 seconds.The PCR product was purified by using a QIA quick PCR purification kit (Qiagen, Hilden, Germany), and sequenced by an ABI Prism 377 DNA sequencing system (Applied Biosystems, Darmstadt, Germany).

Determination of staphylococcal toxin production
Detection of staphylococcal enterotoxins A, B, C, and D was performed using the reverse passive latex agglutination kit for enterotoxin (SET-RPLA) (Oxoid, Basingstoke, United Kingdom).Before testing, each colony was plated in Baird-Parker agar and incubated for 24 hours at 37°C under aerobic conditions.One colony per each sample where S. aureus was found at more than 10 3 cfu/mL (88 isolates) was used in toxin detection.The pure cultures were regrown on tryptic soya broth (Oxoid, Basingstoke, United Kingdom) following manual instructions of the SET-RPLA kit without quantification (Oxoid, Basingstoke, United Kingdom).

Isolation and enumeration of L. monocytogenes
Isolation of L. monocytogenes from raw milk was performed according to the EN ISO 11290-2 standard procedure of the International Organization for Standardization [26].Raw milk samples were plated in triplicate directly onto Oxford agar and PALCAM agar plates (Sigma-Aldrich, St. Louis, United States).Plates were incubated at 37°C and scored for presumptive Listeria spp.Colony hydrolysis, black colonies (indication for Listeria spp.) were observed at 24 and 48 hours.Presumptive Listeria isolates were then grown in BHI broth (Oxoid, Basingstoke, United Kingdom) for further testing, such as oxidase with 1% tetramethyl-p-phenylenediamine dihydrocholride, catalase with 3% hydrogen peroxide, and Gram staining using a three-step staining kit (Oxoid, Basingstoke, United Kingdom).Hemolytic activity was determined by stabbing blood agar (Columbia with 5% sheep blood; Thermo Fisher Scientific, Lenexa, Kansas, United States), which was incubated at 37°C for 48 hours.Single colonies were regrown in BHI broth (Oxoid, Basingstoke, United Kingdom) and stored at -80°C with 13% glycerol.

Virulence gene detection
The following virulence genes were detected: prfA encoding a positive regulatory factor A; hyl encoding a pore-forming toxin (listeriolysin O, LLO); and actA, encoding an actin assembly-inducing protein.Primers are listed in Table 1.Confirmatory identification of L. monocytogenes was performed by 16S rRNA gene sequencing.

Statistical analysis
Statistical analysis of TVC and the prevalence of the two foodborne pathogens (S. aureus and L. monocytogenes) were carried out with R software, version 3.1.0.To evaluate the significant differences and mean values, two-way analysis of variance (ANOVA) was applied, followed by Tukey's test.Statistical significance was defined when a p value was below 0.005.

Enumeration of microorganisms
In total, 603 raw milk samples were collected during the period of December 2011 to June 2012.The samples were derived from 221 farms distributed in 10 different regions.In most cases, the samples were collected 3 times from each farm with 2-3 months between the time points.For the farms from the region of Rahovec, the samples were collected only twice due to technical problems with transport.The first sampling was carried out during December and January, the second sampling during February and March, and the third one during May and June.Most of the farms  included in this study were commercial farms (185/221) that had more than 5 cows per farm (Supplementary Table 1).The TVCs varied considerably between the farms, ranging from 10 4 to 10 7 cfu/mL of milk.Only 6.8% of the farms (15 of 221) complied with the accepted values of the EU (< 105 cfu/mL), but 43% of the farms fulfilled the less strict regulation of Ministry of Agriculture, Forestry and Rural Development of Kosovo (MAFRED) (< 5 × 10 5 cfu/mL milk) (Figure 1).Within the individual farms, only small differences (10 2 to 10 3 cfu/mL) in TVCs were found between samples collected at different time points.The high differences in TVCs were observed between different farms (Supplementary materials Table 1).
Based on the results obtained from the coagulase test, milk from 88 farms was apparently contaminated with S. aureus, while milk from 44 farms was    2).

Discussion
Raw milk represents an important part of nutrition in Kosovo, but little is known about the microbiological quality of raw bulk milk, especially with regard to contamination and the prevalence of foodborne pathogens.To investigate this aspect, milk samples from 221 farms across the country were collected and analyzed.Results showed that only 6.8% (with total cell numbers < 10 5 cfu/mL) of the samples complied with the EU regulations of the microbial quality of milk [32].The MAFRED regulations are less strict, with a TVC of 5 x 10 5 cfu/mL as the upper limit [33]; however, only 43% of the samples (95/221) were acceptable within this national standard [33].
From each farm, samples were collected every second month at three time points.This sampling regime allowed us to assess the general microbial quality of a farm over time and partly during changing seasons (winter to summer).No significant variations in the microbial quality of milk was observed, but a significant variation in microbial quality between farms, irrespective of geographical location, was recorded.This indicates that contamination was most likely not related to the regions or season but rather to the milk production facilities at the individual farms (Supplementary Table 1).We found 43% of the farms (95/221) satisfied the MAFRED standard with regard to TVC limit (< 5 × 10 5 cfu/mL).Interestingly, in a previous study in 2007 from Kosovo, a large number of samples (from 364 farms) of bulk tank milk were collected, and 80 of these farms (20.6%) were found with TVCs below 5 × 10 5 cfu/mL [35].Thus, these studies indicate that some improvement in milk quality occurred in recent years.Another study from 2008 on the microbial quality of raw milk in Macedonia, a neighboring country of Kosovo, revealed that 41.45% of raw milk had TVCs below 4 × 10 5 cfu/mL, which is the upper acceptance level of TVC in milk in Macedonia [33].Using this limit (4 × 10 5 cfu/mL), only about 34% of farms in Kosovo have acceptable TVCs, suggesting that the current raw milk quality in Kosovo is inferior to milk in Macedonia in 2008.In addition, in different countries in Europe, the microbiological quality of raw milk is much better than in developing countries such as Kosovo [20,36].It is well documented that several factors can influence the microbial quality of raw milk, including hygienic conditions of the farms and feed and herd size [12][13][14][15][16].The farms in the regions Ferizaj, Gjakova, and Rahovec were shown to have highest level of TVCs compared to farms in other regions (Supplementary Table 1).This means that measurements such as more and stricter inspection and better management practices need to be implemented in these regions.Sraiiri et al. [37] showed that monitoring of herd management practices may allow a better indirect control of raw milk management than any of other factors, such as increasing the number of sampled cows or increasing the frequency of microbiological analyses of the raw milk.Similarly, another report showed that frequent microbiological analyses of raw milk and sharing the results with farmers can increase their hygiene awareness, which will eventually lead to better farming practices, thereby improving the microbial quality of milk [38].Another factor that affects the quality of raw milk is herd size [12,15].Our results showed that there were significant differences between the farms with large herd size (> 10 cows) and the farms with small hard size (< 5 cows) (Figure 2), which is in line with the results from a previous study where the quality of milk was shown to be better in large herds than in small ones [36].
The high number of TVCs found in raw bulk milk might indicate the presence of a great diversity of microorganisms, some of which can be foodborne pathogens such us Salmonella, Listeria, Staphylococcus, or Escherichia coli.This study conducted an analysis of the presence of two different foodborne pathogens, L. monocytogenes and S. aureus.
It was found that S. aureus was present in milk samples in 39.8% of the farms examined, and in 36 farms (16.3%), the counts of S. aureus were more than 10 5 cfu/mL.The presence of S. aureus in milk can originate from fecal contamination of the udder or milk equipment [39].In recent years, it has been reported that the prevalence of S. aureus in raw bulk milk in EU varied between 12% and 75% [29].In Balkan countries, few studies have been reported.The prevalence of this bacterium was 18% and 32% in farms in Albania and Bosnia and Herzegovina, respectively [4,30].Hence, our results indicate that prevalence of S. aureus in Kosovo is higher compared to Albania but not to Bosnia and Herzegovina.
One of the causative agents of foodborne intoxication is Staphylococcus spp.toxin production, with S. aureus being the most prominent species [28].Based on the food safety standards, Staphylococcus spp., and particularly S. aureus at high numbers (> 10 5 cfu/mL), can produce toxins that severely affect consumers [40].At higher cell numbers, S. aureus can trigger toxin production (A, B, C, D, and E). S. aureus in our study was present at high cell levels (> 10 5 cfu/mL) in 30.7% of bulk tank milk samples tested.Several countries have reported that outbreaks of staphylococcal food poisoning are related to SEA, SEB, SEC, and SED enterotoxin [29].In Bosnia and Herzegovina, it has been reported that 43% of the isolated S. aureus from milk carry one or more genes encoding staphylococcal toxins with highest prevalence of SEB, SEC, and SED toxins, a result that is comparable to that in our study [4].
L. monocytogenes is reported to be present in poorly preserved silage and/or in poor hygiene conditions in farms that are the common source of infection in dairy cows [41,42].Different species of Listeria exist, but only L. monocytogenes and L. ivanovii have been reported to be the major pathogens in humans and animals, respectively.Cleaning procedure is not sufficient to eliminate L. monocytogenes [43].Based on our results, the presence of L. monocytogenes in bulk tank milk was low (2.7%).Previous reports have shown that the prevalence of L. monocytogenes varied between 2.5% and 6% in bulk tank milk in Europe [43,44], while in one study from New York, L. monocytogenes was identified in 12.6% of dairy herds [45].Consumption of such unpasteurized and contaminated raw bulk milk or processed milk would pose a potential health risk for consumers due to the high cell concentrations (up to 10 3 cfu/mL).It is well documented that production of cheese (particularly soft cheeses and fresh cheeses) or other milk-based products from unpasteurized milk infected with L. monocytogenes might increase the risk of infection [46].In Kosovo, some of the traditional cheeses are made from unpasteurized milk.Notably, regional differences in the prevalence of S. aureus and L. monocytogenes were not observed (Supplementary Table 1).However, differences were observed on the farm level rather than regional level, which might indicate differences in individual milking or hygiene practices.No obvious correlation between total counts and contamination with S. aureus and/or L. monocytogenes was found (Supplementary Table 1).This indicates that raw bulk milk is a good reservoir of these two foodborne pathogens, regardless the presence or absence of other bacteria in the same milk.However, S. aureus and L. monocytogenes were mostly present in raw milk collected from farms with a herd size smaller than eight cows.This indicates that these two foodborne pathogens are present more frequently in farms with small herds than in farms with large herds.

Conclusions
The present results indicate that the hygienic quality of bulk tank milk in Kosovo is poor and needs to be improved.Potentially enterotoxigenic S. aureus and virulent L. monocytogenes are occasionally found in Kosovo farms.There is a need to increase the hygiene in dairy farms, and some measures should be directed to improve hygienic conditions during the milking process and milk transport.Possible reasons for the presence of high total viable count contamination include lack of adequate cleaning and improper milking disinfection equipment and suboptimal milk storage and transport.One of the important issues in milk production in Kosovo dealing with poor microbiological quality of milk, which can be overcome by improving the sanitation in the farms.This can be accomplished by providing more information to farmers about the relationship between hygiene and milk quality and also through relevant and practical training given by veterinary and health authorities.Increased herd size and more organized and professional farming practices will probably improve the microbiological quality of raw milk as well.

Figure 1 .
Figure 1.Total viable counts in raw milk from 221 farms in Kosovo.

Figure 2 .
Figure 2. Mean value of total viable counts in raw milk from different group of farms in three different periods.

Figure 3 .
Figure 3. Mean value of total viable counts in raw milk from different regions.

Table 1 .
DNA primers used in PCR for detection of virulence genes in Listeria monocytogenes and 16s rRNA gene sequencing.
F: forward primer; R: reverse primer

Table 2 .
Presence of Staphylococcus spp.and Listeria spp. in raw bulk milk.

Table 3 .
Virulence genes in L. monocytogenes identified by PCR.