Abstract
The aim of this study was to characterize enterotoxigenic Staphylococcus aureus recovered from raw cow milk from two geographical regions of Poland using pulsed-field gel electrophoresis (PFGE) and multilocus sequence typing (MLST). Among 610 samples tested, 229 (37.5%) were positive for S. aureus and 30 (13.1%) of them possessed at least one gene encoding enterotoxins. The sec marker was the most commonly identified (12; 40.0% isolates), followed by the sed (9; 30.0%), sea (6; 20.0%), and seb (1; 3.3%) genes. Some S. aureus possessed a combination of the sea and sec or sea and seb toxin markers. Only two (6.7%) of the enterotoxin gene-positive isolates were not able to produce enterotoxins in vitro. Genotypic analysis with the PFGE method of a total of 50 toxigenic S. aureus isolates from the present and previous studies identified 16 clonal groups. Furthermore, MLST revealed the presence of 15 sequence types with the most common being ST45 and ST1. The results of this study indicate that raw cow milk may be a source of S. aureus with classical enterotoxin genes, which may pose a potential threat for the consumers' safety.
Introduction
S
To date, 23 SE and SE-like toxins (SEls) have been described, including five defined as “classical” (SEA, SEB, SEC, SED, and SEE), with a particular significance in foodborne infections (Kadariya et al., 2014; Umeda et al., 2017). Each of the classical enterotoxins has been related to outbreaks connected with the consumption of raw milk (Jørgensen et al., 2005b).
Pulsed-field gel electrophoresis (PFGE) is considered a highly discriminatory technique used to characterize genetic diversity of different bacterial pathogens, including S. aureus (Tang, 2009). Despite improvements and standardization of the PFGE protocols, comparison of data from various laboratories may be difficult (Murchan et al., 2003).
Multilocus sequence typing (MLST) technique is based on the sequence polymorphism of seven housekeeping genes present in S. aureus (Enright et al., 2000). A high reproducibility of MLST makes it useful for global comparisons of bacterial populations (Melles et al., 2007). Application of both PFGE and MLST methods gives the most accurate and discriminatory results in genotyping of S. aureus (Xie et al., 2011). Molecular analyses of S. aureus isolates recovered from raw milk may provide information about the clonal relationship of strains and may be useful in tracing the sources of food contamination in SFP investigations (Jørgensen et al., 2005a). The aim of this study was to characterize enterotoxigenic S. aureus recovered from raw cow milk using PFGE and MLST.
Materials and Methods
A total of 610 raw cow milk samples were collected from a northeastern region of Poland. Coagulase-positive staphylococci (CPS) were isolated on Baird–Parker agar with rabbit plasma fibrinogen supplement (bioMérieux, Marcy l'Étoile, France) incubated at 37°C for 24–48 h. One suspected colony from each sample tested by PCR enabled to identify the specific S. aureus 23S rRNA gene fragment (Korpysa-Dzirba and Osek, 2014). The isolates confirmed as S. aureus were taken for further analysis.
Detection of SE genes and toxins
Staphylococcal enterotoxin sea, seb, sec, sed, and see genes were identified according to the described procedure (Korpysa-Dzirba and Osek, 2011). Detection of S. aureus enterotoxins was performed after a 24-h incubation at 37°C in brain-heart infusion broth (bioMérieux) using the Vidas SET 2 test (bioMérieux) based on the ELFA (enzyme-linked fluorescent assay) technique.
Pulsed-field gel electrophoresis
PFGE analysis was used for all SE-positive S. aureus identified in the previous (Korpysa-Dzirba and Osek, 2014) and present studies, that is, 30 and 20 strains, respectively. These 50 S. aureus were tested using the method adopted from the European Union Reference Laboratory for Coagulase-Positive Staphylococci (EURL CPS), which was modified by decreasing the enzymatic digestion temperature to 25°C (Marault et al., 2011).
Electrophoresis was performed in the CHEF-DR II (Bio-Rad, Hercules, CA) under the following conditions: initial pulse of 1 s, final pulse of 45 s, and 6 V/cm at 14°C for 19 h. After staining with ethidium bromide and destaining with distilled water, the gel was visualized with a UV transilluminator Gel Doc XR+ (Bio-Rad). The band patterns were compared using Dice coefficient with a 1% band position tolerance with BioNumerics v.6.8 software (Applied Maths, Sint-Martens-Latem, Belgium). A PFGE dendrogram was created based on the underweighted-pair group method with arithmetic averages (UPGMA) with cutoff set on 80% similarity (Murchan et al., 2003).
Multilocus sequence typing
MLST was used according to the procedure described previously (Enright et al., 2000). Seven housekeeping genes were targeted: carbamate kinase (arcC), shikimate dehydrogenase (aroE), glycerol kinase (glp), guanylate kinase (gmk), phosphate acetyltransferase (pta), triosephosphate isomerase (tpi), and acetyl coenzyme A acetyltransferase (yqiL) (Holden et al., 2004). The primers for their amplification were as recommended in MLST database (
Amplifications were carried out in both forward and reverse directions using the same sets of primers as for PCRs following DNA sequencing (Hennekinne et al., 2003). Allelic profiles were assigned by comparing the obtained sequences with those existing in the S. aureus MLST database. The sequence type (ST) was determined by comparing the combination of seven MLST alleles using BioNumerics v.7.6 (Applied Maths).
Results
Detection of enterotoxigenic S. aureus and toxin production
Two hundred twenty-nine of 610 (37.5%) milk samples collected in the present study were positive for coagulase-positive S. aureus. Among them, 30 (13.1%) isolates carried at least one gene encoding classical SEs. The vast majority (28; 93.5%) of the toxigenic strains had one toxin gene, while the remaining two S. aureus were positive for two SE markers. The most frequently detected were the sec gene (12; 40.0% strains) followed by the sed (9; 30.0%), sea (6; 20.0%), and seb (1; 3.3%) markers. Single strains (3.3%) harbored a combination of the sea with sec and sea with sed toxin genes. The results obtained by the ELFA technique showed that all but two (93.3%) S. aureus strains were able to produce staphylococcal enterotoxins in vitro. One of them harbored the sec and the second one the sed genes, respectively.
PFGE and MLST genotyping
PFGE revealed a high diversity among 50 S. aureus analyzed. A total of 16 clusters, containing from 1 to 14 strains, were identified. The minimal similarity between the different patterns was 47.3%. The largest cluster (14 strains) had 13 isolates with the sec enterotoxin gene, including one gene-positive strain that was not able to produce toxin in vitro (negative result in the Vidas SET2 test) (Fig. 1). The second numerous cluster (10 strains) covered nine isolates with the sea gene. Both these groups contained S. aureus from the northeast as well as southeast regions of Poland.

PFGE dendrogram with characteristics of the Staphylococcus aureus isolates. PFGE, pulsed-field gel electrophoresis; ST, sequence type.
Further molecular analysis performed by MLST revealed that all isolates were classified to 15 STs. The most numerous STs were ST45 and ST1 with 17 and 9 strains, respectively. The molecular methods used in this study revealed groups of sec-positive ST45 (13 strains) and sea-positive ST1 (7 strains) S. aureus originating from both geographical regions combined in two PFGE clusters. The results of PFGE and MLST analysis are shown in Figure 1.
Discussion
As shown here, 12.6% of the isolated S. aureus had the classical enterotoxin genes, mainly sec and sed. However, the previous study from Poland with a lower number of raw cow milk samples indicated that 68.8% of CPS were enterotoxigenic (Bystroń et al., 2005). In Portugal and the Czech Republic, S. aureus with classical SE genes was also recovered from raw cow milk at levels of 20% and 37.1%, respectively (Zouharova and Rysanek, 2008; Pereira et al., 2009). Recently, a similar prevalence of S. aureus with SEA to SEE genes in cow milk (9.4%) to the reported in the present study has been described in Jordan (Obaidat et al., 2018).
After discovery of new SE types, a number of potentially enterotoxigenic strains have increased. A total of 71.1% of S. aureus isolated from raw milk by Rall et al. (2008) had SE genes encoding both toxin groups. Similar results were obtained in Italy where 71.6% of S. aureus recovered from raw milk and milk products had both classical and new SE genes (39.5%), only classical (24%) or new toxin markers (7.4%), respectively (Morandi et al., 2009). As shown in the present study, among classical enterotoxin genes, mainly sec followed by see and sed was identified. Studies from Brazil, Norway, France, and Japan also showed that the sec gene was the most commonly found in S. aureus isolated from raw milk (da Silva et al., 2005; Jørgensen et al., 2005b; Katsuda et al., 2005; Loncarevic et al., 2005; Villard et al., 2005; Normanno et al., 2007). On the contrary, the sed-positive isolates predominated among S. aureus recovered from raw milk in Italy (Morandi et al., 2007).
In the present study, the ability of enterotoxin production was shown in the majority of the isolates (93.3%) as tested by the Vidas SET2 method. However, some authors using the reversed passive latex agglutination test observed a lower correlation between the presence of genes and toxin expression, which might be due to, for example, the production of toxins below the test detection limit (Morandi et al., 2007). It may suggest that the majority of enterotoxigenic S. aureus in favorable environmental conditions can produce classical staphylococcal enterotoxins.
PFGE analysis showed 44.8% similarity between tested S. aureus, which is lower than that described in Sweden (50.8% isolates from cheese) and France (57.5% strains from food poisoning outbreaks) (Kérouanton et al., 2007; Rosengren et al., 2010). A high diversity of isolates in the present study (16 clonal groups) may be due to different geographical origins of the samples or different sources of milk contamination. Similar results were published in the Czech Republic where 19 clonal groups were distinguished among 57 staphylococci isolated from milk products. Alibayov et al. (2014) showed that S. aureus with the sec genes clustered in a separate clonal group with 76% homology.
The sec-positive S. aureus analyzed in the present study were gathered in the largest clonal group, with a similarity of 75.7–100%, and assigned to ST45 as tested by MLST. This ST was also found among sec-positive S. aureus recovered from food poisoning in Spain (Argudin et al., 2012).
Strains of ST1, the second numerous group, had the sea gene and in one case the sec gene. These results are comparable with data from China and Germany where all S. aureus from food poisoning outbreaks were positive for the sea marker and classified to ST1 (Yan et al., 2012; Fetsch et al., 2014). ST1 and ST12 detected previously in Poland were among strains isolated from food poisoning outbreaks associated with the consumption of ice-cream in Germany (Fetsch et al., 2014). However, S. aureus ST12 from these two locations had different SE genes. In the present study, the sec-positive isolates with ST12, together with one strain of ST1460, were classified into one clonal group as tested by PFGE.
Conclusions
The presented results are unique, showing molecular characteristics of enterotoxigenic S. aureus recovered from raw cow milk. Their high genetic diversity has been found, indicating that contamination by these bacteria could originate from several sources. The genetic typing results demonstrated the existence of S. aureus groups sharing the same SE genes and STs gathered in two largest PFGE clusters. STs (ST45 and ST1) were found in other countries during food outbreak analyses, which may suggest that such S. aureus are spread in different geographical areas. Furthermore, this study provides information on the genetic structure of enterotoxigenic S. aureus from raw cow milk and may help to track the spreading of such strains during epidemiological investigations.
Footnotes
Acknowledgments
This work was supported financially by the National Science Centre in Poland (Decision No. DEC-2011/01/N/NZ7/04310) and by KNOW (Leading National Research Centre) Scientific Consortium “Healthy Animal - Safe Food” decision of Ministry of Science and Higher Education No. 05-1/KNOW2/2015.
Disclosure Statement
No competing financial interests exist.
