Staphylococcal Enterotoxin Genes in Coagulase-Negative Staphylococci—Stability, Expression, and Genomic Context

In the current study, we screened a collection of coagulase-negative staphylococci (CoNS) isolates for orthologues of staphylococcal enterotoxins (SEs) involved in S. aureus-related staphylococcal food poisoning (SFP). The amplicons corresponding to SEs were detected in S. chromogenes, S. epidermidis, S. haemolyticus, S. borealis, S. pasteuri, S. saprophyticus, S. vitulinus, S. warneri, and S. xylosus. All amplicons were sequenced and identified as parts of known S. aureus or S. epidermidis SE genes. Quantitative real-time PCR allowed determining the relative copy number of each SE amplicon. A significant portion of the amplicons of the sea, seb, sec, and seh genes occurred at low copy numbers. Only the amplicons of the sec gene identified in three isolates of S. epidermidis displayed relative copy numbers comparable to sec in the reference enterotoxigenic S. aureus and S. epidermidis strains. Consecutive passages in microbiological media of selected CoNS isolates carrying low copy numbers of sea, seb, sec, and seh genes resulted in a decrease of gene copy number. S. epidermidis isolates harbored a high copy number of sec, which remained stable over the passages. We demonstrated that enterotoxin genes may occur at highly variable copy numbers in CoNS. However, we could identify enterotoxin genes only in whole-genome sequences of CoNS carrying them in a stable form at high copy numbers. Only those enterotoxins were expressed at the protein level. Our results indicate that PCR-based detection of enterotoxin genes in CoNS should always require an additional control, like analysis of their presence in the bacterial genome. We also demonstrate S. epidermidis as a CoNS species harboring SE genes in a stable form at a specific chromosome site and expressing them as a protein.


Introduction
Staphylococcal enterotoxins (SEs) produced by S. aureus are recognized causes of staphylococcal food poisoning (SFP) [1]. Coagulase-negative staphylococci (CoNS) for a long time were considered only as commensal bacteria, unable to cause infection, and their presence in clinical material was considered as contamination with bacteria colonizing the skin. Nevertheless, numerous studies imply that coagulase-negative staphylococci may also be enterotoxigenic and potentially cause food poisoning. The first descriptions of possible CoNS food poisoning appeared as early as in the 1950s and 1970s [2,3]. The first CoNS strains to produce TSST-1 and SEA enterotoxins were identified in the second half of the 1980s and came from cases of toxic shock in humans [4]. Since then, enterotoxigenic coagulase-negative staphylococci have been identified repeatedly in samples derived from human clinical material, as well as from cases of mastitis in cattle and other ruminants [5][6][7][8].
The number of reports on the enterotoxigenicity of CoNS has been increasing. The method most commonly used for determining the presence of enterotoxin genes is PCR, allowing the detection of small copy numbers of a given gene [9][10][11][12]. However, there are insufficient data about the structure of genetic elements containing enterotoxin genes in the genomes of CoNS and the expression of these genes. Some researchers have observed that the intensity of PCR amplicons corresponding to SEs genes is significantly lower for CoNS than S. aureus, despite using comparable amounts of genomic DNA. In addition, a gradual or complete loss of PCR signal, as a result of subsequent passages of bacteria, has been reported [9,13]. Such phenomenon has not been noted for enterotoxigenic S. aureus strains. Scientists have hypothesized that staphylococcal enterotoxin genes in some CoNS isolates are unstable and the cell population of such isolates is heterogenic in terms of toxin gene content. As a result of cell division, only some cells may carry genetic elements encoding enterotoxin genes [9].
There are scarce data on the possibility of the acquisition of elements bearing enterotoxin genes by staphylococci. In enterotoxigenic S. aureus, the genes encoding staphylococcal enterotoxins are located on mobile genetic elements such as plasmids, prophages, S. aureus pathogenicity islands (SaPIs), genomic islands (νSa), or in the proximity of staphylococcal chromosome cassette elements (SCC) [1]. Such locations potentially enable horizontal gene transfer (HGT) between bacteria. To date, HGT between S. aureus strains [14][15][16], staphylococcal species (SaPI transfer from S. aureus to S. xylosus), and between genera of bacteria (SaPI transfer from S. aureus to L. monocytogenes) [17] has been described in vitro. However, the only known genetic element carrying enterotoxin genes among CoNS is the composite S. epidermidis pathogenicity island, SePI. First described in a human S. epidermidis FRI909 isolate, SePI carries the orthologs of sec and sel of S. aureus encoding SEC and SEL, respectively [18]. However, efforts to mobilize composite SePI from the S. epidermidis FRI909 strain remain ineffective [18].
Numerous PCR-based studies aiming to screen for SE genes report a variable, often frequent occurrence of these genes in CoNS. In turn, studies searching for SE genes based on genome sequence analysis failed to detect SE genes in CoNS, except in some S. epidermidis and coagulase-variable S. hyicus and S. agnetis isolates [19][20][21][22]. We assumed that the SE genes detected in some CoNS strains may occur in their cells in stable and transient forms, the last likely representing a step in SE transfer to CoNS. Therefore, our goal was to assess the origin, stability, and abundance of the DNA carrying SE genes in CoNS isolates obtained from food products and farm animals. We also determined the genomic context of these genes and the expression of CoNS enterotoxins.
Amplicons obtained from CoNS isolates corresponding in size to sea, seb, sec, and seh products of the reference S. aureus strains were sequenced (Genomed S.A., Warsaw, Poland).
Then, using the BLAST algorithm, their sequences were compared with the sequences available in the GenBank. We noted that the nucleotide sequences of the sea, seb, and seh amplicons obtained from the DNA of CoNS were homologous to the corresponding genes of known S. aureus enterotoxins. The sequence identity was 100% for sea and seb and 99.72-100% for seh. The sec amplicons found in S. epidermidis and S. pasteuri isolates had sequences 100% consistent with the sec sequence in S. epidermidis FRI909 and S. epidermidis 4S [13,18] (Table 3). Observation of electrophoresed SE amplicons in CoNS showed that the intensity of a significant portion of them (n = 26) was lower than the intensity of the corresponding amplicons obtained from the reference S. aureus or S. epidermidis strains. Only in the case of three enterotoxigenic S. epidermidis isolates was the intensity of sec amplicons comparable to that of the corresponding reference S. aureus and S. epidermidis strains. To explore the reasons for this observation, we determined the relative copy numbers of the SE genes in all tested CoNS isolates.

Determination of the Relative Copy Numbers of SE Genes in CoNS Isolates
In all CoNS isolates in which the presence of amplicons corresponding to the enterotoxin genes was detected by PCR, we assessed the relative number of copies of SE genes in genomic DNA.
We showed that in the DNA of 26 enterotoxigenic CoNS, the relative copy number of the SE genes was multiple orders of magnitude lower than in the DNA of the appropriate reference S. aureus or S. epidermidis strains (Table 4). In S. saprophyticus 5-8 isolate, the relative copy number of the sea gene was 622,702 times lower than in the reference S. aureus strain. In S. xylosus WS08.1 isolate in which the seb gene was found, the relative copy number of this gene was 503,324 times lower than in the reference S. aureus strain. In S. pasteuri 65.1 and 34.1.3 isolates, the relative copy number of the sec gene was 3974 and 5980 times lower than in S. aureus FRI913, and 4568 and 6873 times lower than in S. epidermidis 4S, respectively. In S. pasteuri 34.30 isolate with the confirmed sec amplicon, the sensitivity of qPCR turned out to be insufficient to evaluate the copy number of the gene. The relative copy number of the seh gene was determined in 21 CoNS isolates. In 10 of them, the determination of the number of seh copies was below the qPCR sensitivity. In the remaining 11 CoNS, the relative number of the seh gene copies was 1050 to 3,461,012 times lower than in S. aureus FRI137.
In three S. epidermidis isolates of food origin, i.e., S. epidermidis 29, 34, and 65.2, sec amplicons were detected. Sequences of these amplicons were 100% identical to the sequence of the sec gene of S. epidermidis 4S and S. epidermidis FRI909. The relative copy numbers of the sec gene in S. epidermidis 29, 34, and 65.2 isolates were from 1.5 times lower to 1.8 times higher than the sec copy number in the reference S. aureus strain FRI913 and from 1.7 times lower to 1.5-fold higher than the sec copy number of S. epidermidis 4S strain. The relative copy numbers of the sec gene were comparable to S. aureus and S. epidermidis reference strains only in those enterotoxigenic CoNS.

Determination of Stability of Staphylococcal Enterotoxin Genes
We found that in S. xylosus WS08.1 isolate, in which the seb gene amplicon was identified, the relative number of enterotoxin gene copies decreased to values not measurable by qPCR after one passage. In S. saprophyticus 5-8 isolate, positive for the sea gene amplicon, the relative number of copies of this enterotoxin gene varied in the range of 0.0000062828 to 0.0000003002 throughout five consecutive passages. In S. pasteuri 65.1 isolate, in which the sec gene amplicon was found, the relative number of copies of this enterotoxin gene varied from 0.0000005755 to values not detectable by qPCR throughout five consecutive passages. In the isolate of S. epidermidis 275lp, in which the amplicon of the seh gene was found, the relative copy number of this gene varied in the range of 0.0000563722 to 0.0000003596 throughout 5 successive passages (Table 5). Of the 3 S. epidermidis isolates in which the relative sec copy numbers were comparable to those found in the reference S. aureus and S. epidermidis strains, S. epidermidis 29 isolate was selected for subsequent passages on microbial medium. As a result of five consecutive passages, the relative copy number of the sec gene ranged from 0.69 to 1.08.

Mixed Culture of Bacterial Isolates and Incubation of Bacterial Culture with Genomic DNA
The strains of S. epidermidis 65.2 and S. pasteuri 65.1 were isolated from a sample of a single food product. In both isolates, the amplicons corresponding to the sec gene were found by PCR. S. epidermidis 65.2 isolate carried sec at a high copy number (0.62) while S. pasteuri 65.1 at a low copy number (0.0001387956). S. epidermidis 65.2 maintained the relative copy number of the sec gene at a constant level after five consecutive passages in the microbial medium. In turn in the S. pasteuri 65.1 isolate, complete loss of the amplicon corresponding to the sec gene was observed after six passages. We assessed the possibility of restoring the sec gene in S. pasteuri 65.1 by its co-culture with the enterotoxigenic S. epidermidis 65.2 isolate. Prior to this experiment, one S. pasteuri 65.1 colony was subjected to six passages on an agar plate, which resulted in a complete loss of the PCR signal for the sec gene. After a 2-day co-culture of the isolates, we observed the sec amplicon in 0-1 of 20 colonies of tested S. pasteuri 65.1 isolate. After 5 and 8 days of culture, the sec amplicon was noted in 2-4 and 3-4 of 20 examined colonies of S. pasteuri 65.1, respectively.
In all S. pasteuri 65.1 colonies derived from co-culture with S. epidermidis 65.2, we determined the relative copy number of the sec gene by qPCR (Table 6). It was found that the relative copy number of the sec gene in all studied colonies was 263,790-5,187,877 times lower than in the reference S. aureus strain FRI913, from 298,063 to 5,861,914 times lower than in the S. epidermidis 65.2 isolate and from 40 to 779 times lower than that determined in natural S. pasteuri 65.1 isolate when isolated from food. Two selected S. pasteuri colonies denoted as II7 and II8, which acquired the sec amplicon after the 5-day co-culture with S. epidermidis 65.2, were subjected to 5 consecutive passages in the microbial medium. After each passage, we assessed the signal for the sec gene by PCR. The presence of the sec signal ceased to be PCR-detectable after the third and fourth passage for bacteria derived from colonies II7 and II8, respectively.
To evaluate the DNA containing enterotoxin genes persistence possibility in the culture of CoNS, S. pasteuri 65.1 isolate was incubated with the genomic DNA extracted from S. epidermidis 65.2. Prior to the experiment, one S. pasteuri 65.1 colony was subjected to six consecutive passages in the microbial medium until the complete disappearance of the signal for the sec gene. After two days of incubation of S. pasteuri 65.1 with a solution of S. epidermidis 65.2 genomic DNA, the sec gene was detected in 5 out of 8 examined S. pasteuri colonies. After 5 and 8 days of incubation, we confirmed the presence of the sec gene in 7 and 2 out of 8 examined S. pasteuri colonies, respectively.
For all S. pasteuri 65.1 colonies that displayed the presence of the sec gene after incubation with S. epidermidis 65.2 DNA, the relative copy number of the gene was determined by qPCR. In all colonies studied, the relative copy number of the sec gene was 12,864 to 355,068 times lower than in the reference strain of S. aureus FRI913, from 14,535 to 401,200 times lower than in S. epidermidis 65.2 and from 2 to 53 times lower than determined in the natural isolate of S. pasteuri 65.1 after isolation from food (Table 7).

Genome Sequencing and Analysis of Selected Bacterial Isolates
The genomic sequence of the enterotoxigenic S. pasteuri 65.1 isolate carrying a lowcopy-number, unstable sec gene was determined. The isolate derived from the same food product as S. epidermidis 65.2, carrying a high-copy-number, stable sec gene. The number of contigs in the genomic sequence of S. pasteuri 65.1 was 146 and the sequence length was 2,618,483 bp (Table 8). Using the Genome Comparator module of the BIGSdb platform [23], we screened the genomic sequence of S. pasteuri 65.1 for the sec gene previously detected in this isolate by PCR. We did not identify the sec gene in the genomic sequence of S. pasteuri 65.1 isolate. The gene was previously documented in S. epidermidis 65.2 genome [21].

Production of SEs by CoNS Isolates
We checked the ability to produce SEs in selected CoNS isolates in which sea, seb, sec, and seh amplicons were detected by PCR. SEA production was determined in the S. saprophyticus 5-8 isolate, SEB in S. xylosus WS08.1, SEC and SEL in the S. pasteuri 65.1, S. epidermidis 65.2, S. epidermidis 29, and S. epidermidis 34 isolates. We assessed the SEH production in the S. saprophyticus 51.2 and S. epidermidis 54.1 isolates. Appropriate S. aureus reference strains were used as positive controls and the S. epidermidis 4S strain was used as a control for SEC and SEL. Negative controls consisted of reference S. aureus strains that did not produce specific enterotoxins ( Table 9). Cultures of the above-mentioned isolates were carried out in BHI broth. At 6 and 24 h, the number of bacteria was determined, and the concentration of appropriate enterotoxins was measured using ELISA in post-culture liquids. The number of bacteria ranged from 4.3 log CFU/mL for S. saprophyticus 5-8 to 9.4 log CFU/mL for S. aureus CCM5757 and S. aureus FRI913 after 6 h of cultivation, and from 8.7 log CFU/mL for S. aureus FRI913 and S. xylosus WS08.1 to 10.0 log CFU/mL for S. aureus FRI137 after 24 h of cultivation (Table 10). The only CoNS isolates able to produce SEs were the S. epidermidis 29, 34, and 65.2 isolates, all of which secreted SEC and SEL. S. epidermidis 29 produced SEC at 476-1180 ng/mL and SEL at 58-223 ng/mL, S. epidermidis 34 produced SEC at 337-933 ng/mL and SEL at 18-76 ng/mL, while S. epidermidis 65.2 produced SEC at 235-1023 ng/mL and SEL at 17-20 ng/mL after 6 and 24 h of culture, respectively. In the post-culture liquids obtained from the S. saprophyticus 5-8, S. xylosus WS08.1, S. pasteuri 65.1, S. saprophyticus 51.4, and S. epidermidis 54.1 isolates, no SEs were detected (Table 10).

Origin, Stability, Abundance, and Genomic Context of CoNS Enterotoxin Genes
Enterotoxins produced by staphylococci are the cause of staphylococcal food poisoning. Undoubtedly, Staphylococcus aureus is the main producer of these agents [1]. However, over the years, there have been numerous reports confirming the enterotoxigenicity of CoNS isolated from food and animal sources [24,25]. Researchers confirmed the presence of genes encoding enterotoxins orthologous to S. aureus SEC and SEL in the genome of the S. epidermidis strain FRI909 [18]. The number of reports on the enterotoxigenic CoNS is significant [9,10,13,26,27], but certain research limitations prevent a reliable assessment of this problem. A number of studies conducted so far, showing the presence of amplicons corresponding to enterotoxin genes using PCR, have not confirmed the production of enterotoxins by CoNS at the protein level. Furthermore, sequencing of the obtained SE amplicons and determination of the whole-genome sequence of enterotoxigenic CoNS are rarely performed.
We characterized CoNS isolates from food, cow's milk, and farm animals, in which we demonstrated the presence of amplicons corresponding to the SE genes by PCR. Sequence analysis of the sea, seb, sec, and seh amplicons in CoNS isolates showed that they correspond to the sequences of known enterotoxins of S. aureus or S. epidermidis. The signal strength of the majority of these amplicons after electrophoretic separation was weaker than for the SE amplicons of S. aureus. The exception was the sec gene amplicons in three isolates of S. epidermidis, which harbor the sec gene.
We have shown that in a significant part of enterotoxigenic CoNS isolates, the relative number of SE gene copies was much lower than in S. aureus strains. The only isolates with relative copy numbers of SE genes comparable to the reference strains of S. aureus and S. epidermidis 4S strain described earlier [13] were three isolates of S. epidermidis in which we detected the sec gene by PCR. We quantified the SE signals referring them to the tufA gene known to occur in a single copy in CoNS cells, allowing to indirectly compare the number of SE copies to the number of CoNS cells. Based on our research it can be interpreted that in CoNS isolates carrying the SE genes at low copy numbers a given SE gene occurs in one bacterial cell per thousands or millions non-toxigenic cells.
Our results also indicate that CoNS isolates in which the relative number of enterotoxin genes is significantly lower than in S. aureus strains may lose the PCR signal of the SE genes. The phenomenon of the loss of genetic elements containing SE genes by CoNS during culture in microbial media may suggest that SE genes occur in CoNS in an unstable form, and bacteria can lose them during cell division. Alternatively, this observation may indicate a continuous transmission of elements containing SE genes from staphylococci that harbor stable enterotoxin genes to the CoNS, which may occur in the environment. Isolation of the CoNS from the environment and further cultivation under laboratory conditions would break the horizontal gene-transfer process. This scenario seemed probable when we isolated the S. epidermidis 65.2 strain displaying a stable sec gene in a high copy number, together with the S. pasteuri 65.1 strain with the unstable sec gene present in a low copy number from a single food product.
To determine whether the sec gene is integrated into the genome of S. pasteuri 65.1, whole-genome sequencing of the natural S. pasteuri 65.1 isolate was performed. Our analysis did not show the presence of the sec gene or the elements of the composite SePI in the whole genomic sequence of the S. pasteuri 65.1 isolate. According to current knowledge, the sec gene is part of certain SaPIs and the S. epidermidis composite SePI. Researchers have convincingly described the possibility of mobilization and horizontal SaPI transfer involving helper phages between S. aureus strains and between S. aureus and other bacterial species in vitro [14][15][16][17]. Attempts to mobilize the helper phage Φ909 in S. epidermidis FRI909 strain and transfer the SePI element by Madhusoodanan et al. [18] did not have any effect. This may suggest that the composite SePI element has no potential for transmission between S. epidermidis and other microorganisms.
To determine the possibility of horizontal transfer of genetic elements containing enterotoxin genes between CoNS species, we performed a joint culture of sec-positive staphylococci isolated from the same food product, i.e., S. epidermidis 65.2 and S. pasteuri 65.1. The unstable sec gene in the natural isolate of S. pasteuri 65.1, derived from the food product, was eliminated in the process of subsequent passages in the microbiological medium. During the co-culture of S. epidermidis 65.2 and S. pasteuri 65.1, we observed the appearance of amplicons corresponding to the sec gene in some of the S. pasteuri colonies. The relative copy number of the sec gene acquired by S. pasteuri 65.1 was at a level corresponding to the low copy number of enterotoxin genes in the CoNS that possess unstable enterotoxin genes. Further passages of selected S. pasteuri 65.1 colonies that acquired sec resulted in a gradual disappearance of the sec amplicon up to its complete loss, as previously seen in the natural isolate of S. pasteuri 65.1 isolated from the food product. The results obtained indicate that elements containing staphylococcal enterotoxin genes can be transferred in the food environment between CoNS species.
Given that the presence of the sec gene or any of the SePI elements has not been identified in the complete genome of the natural S. pasteuri 65.1 isolate, we assumed that the observed acquisition of sec by S. pasteuri may not occur due to the transfer of the SePI element containing the sec gene to S. pasteuri cells, but only as a result of the persistence of extracellular DNA of S. epidermidis 65.2 in the culture of the S. pasteuri isolate. We evaluated this possibility by incubating the S. pasteuri 65.1 culture with genomic DNA from S. epidermidis 65.2. The results of this experiment indicate the probability of persistence of, most likely extracellular, DNA on the surface of CoNS cells. This could mean that the observed in vitro transfer of the sec gene between S. epidermidis and S. pasteuri was not the result of the mobilization of the genetic element and its horizontal transfer. The results obtained indicate that the generated amplicons corresponding to the sequences of SE genes from CoNS isolates may be associated with the long-term persistence of DNA containing enterotoxin genes in cultures of these bacteria.
The genomes of enterotoxigenic S. epidermidis 29, 34, and 65.2 strains isolated from food were already sequenced and analyzed for their SE gene contents and their genomic context [21]. These strains produced stable signals for PCR-identified SE genes with high copy numbers comparable to reference enterotoxigenic S. aureus and S. epidermidis strains. We identified genes encoding S. aureus SEC and SEL orthologs in the genomic sequences of the 29, 34, and 65.2 S. epidermidis strains. Genes encoding these orthologs were located in a composite genomic-island SePI that had the same genomic integration site in each S. epidermidis strain [21].

Food Safety Hazards and Clinical Relevance of CoNS Enterotoxins
To assess food safety hazards related to enterotoxigenic CoNS, we determined the ability to produce staphylococcal enterotoxins by selected isolates of these bacteria. Based on the results obtained, we can assume that the only CoNS isolates capable of SE production are those in which enterotoxin genes are present in a stable form in high copy numbers. Isolates of CoNS in which enterotoxin genes in an unstable form and low copy numbers are present are not capable of producing enterotoxins, even under optimal laboratory conditions. The enterotoxigenicity of staphylococci contributed to a number of human and animal diseases, including organ and systemic disorders [28,29]. However, to date, all these cases were only associated with SE-producing S. aureus. There are scarce data on the correlation between the enterotoxigenicity of CoNS and its clinical importance. Currently available WGS-based research indicates that other virulence factors, such as genes encoding adhesion factors, biofilm production, hemolysins, and exoenzymes, may affect the pathogenicity of CoNS [30]. We recently found that SE produced by S. epidermidis modifies the intestinal immune response and contributes to gut damage in mice [31].

Conclusions
We demonstrated that enterotoxin genes can occur at highly variable copy numbers in CoNS. However, we could identify enterotoxin genes only in the whole-genome sequences of CoNS that carried them in high numbers of copies. Only those enterotoxins were expressed at the protein level. The enterotoxin genes appearing in CoNS isolates at low copy numbers may be a result of a transient transfer of DNA containing enterotoxin genes from co-occurring enterotoxigenic staphylococci. Our results indicate that the detection of enterotoxin genes based on PCR in CoNS should always require an additional control, such as analysis of the presence of the SE gene in the bacterial genome or detection of the SE protein. Our survey also revealed S. epidermidis to be a CoNS species harboring SE genes in a stable form in the bacterial chromosome and expressing them as a protein.

Bacterial Strains' Isolation and Staphylococcal Species' Identification
In this research, 141 food products (92 meat products, 34 samples of raw cow's milk, and 15 samples of rennet cheese) and 50 swine nasal swabs collected in a slaughterhouse were screened for the presence of enterotoxigenic coagulase-negative staphylococci. Assuming that enterotoxigenic CoNS may enter the food chain from animal sources, food samples represented products obtained from raw material not processed thermally during production. Similarly, milk was non-processed, and swine swabs were taken from animals just after stunning and bleeding.
A food sample (5 ± 1 g) or a swab was placed in 20 mL of Giolitti-Cantoni broth (Merck, Darmstadt, Germany). After 24-48 h of incubation at 37 • C, positive samples (blackening of the medium or a gray-black precipitate) were streaked onto Baird-Parker agar (Merck, Darmstadt, Germany) and incubated for another 48 h. Next, a loopful of biomass from the plate was transferred into 5 mL of brain-heart-infusion broth (BTL, Warsaw, Poland) and incubated for 18-24 h at 37 • C. From these cultures, DNA was isolated to identify the presence of sea-sed or seh genes. In case of a positive result, separate bacterial colonies were screened for the presence of amplicons of the corresponding staphylococcal enterotoxins. To exclude contamination with enterotoxigenic S. aureus, the CoNS DNA was inspected for coagulase (clf ) and nuclease (nuc) genes. Species identification of enterotoxigenic CoNS isolates was performed by analyzing fragments of the dnaJ or tufA gene sequences and their comparison with the GenBank database resources.

PCR Amplification
The tested isolates were screened for the presence of staphylococcal enterotoxin sea, seb, sec, sed, and seh genes using PCR. Each reaction mixture (25 µL) consisted of 2.5 µL 10× DreamTaq Buffer (Thermo Fischer Scientific Inc., Waltham, MA, USA), 0.5 µL of 10 mM dNTPs mix (Thermo Fischer Scientific Inc., Waltham, MA, USA), 0.1 µL of each 10 µM primer, 0.5 U of DreamTaq polymerase (Thermo Fischer Scientific Inc., Waltham, MA, USA), and 1 µL of genomic DNA. Reference S. aureus strains FRI913, CCM5757, FRI1151m, and FRI137 were used as PCR controls (Table 11). All amplicons corresponding to S. aureus enterotoxin genes were sequenced (Genomed S.A., Warsaw, Poland). The presence of clf and nuc genes was determined by duplex PCR. The composition of the reaction mixture was the same as described above. All primers used in this study and all PCR protocols are listed in Supplementary Table S1.

Determination of Relative Gene Copy Numbers in the Tested Staphylococcal Isolates
In all isolates of enterotoxigenic CoNS, the relative copy number of enterotoxin genes was determined by quantitative PCR (qPCR) using a CFX Connect™ thermal cycler (Bio-Rad Laboratories Inc., Hercules, CA, USA). The tufA gene was selected as the reference gene for signal normalization for enterotoxin genes. This gene is present in the genomes of CoNS species in one copy per genome. The primers used for SE genes' amplification were the same as those used in PCR, and the primers used for amplification of the housekeeping gene were as follows: tuf2-F 5 -GACTACGCTGAAGCTGGTGAC-3 and tuf3-R 5 -AGTACGGAAATAGAATTGTGG-3 . The reaction mixture (20 µL) consisted of 10 µL of KAPA SYBR ® FAST qPCR Master Mix (2X) (Sigma-Aldrich Co., Saint Louis, MO, USA), 0.8 µL of each 1.25 µM primer, and 100 ng of genomic DNA. The qPCR amplification conditions included: initial denaturation at 95 • C for 10 min, followed by 40 cycles of amplification (95 • C for 15 s, 55 • C for 20 s, and 72 • C for 40 s). PCR specificity was monitored by analyzing the melting curves of PCR products in a temperature range from 90 • C to 56 • C. For each primer pair, PCR efficiencies were determined by running serial five-fold dilutions of the template. Bio-Rad CFX Manager 3.1 software (Bio-Rad Laboratories Inc., Hercules, CA, USA) was used for the calculations. The Ct values of a given enterotoxin gene were normalized to the Ct of the housekeeping tufA gene, taking into calculation the amplification efficiency of the gene, using the normalized expression method (∆∆Ct) [32]. Then, the values of the relative copy number of each enterotoxin gene were calculated by referring them to the sample showing the lowest normalized expression value, which was assigned the value of 1. Therefore, the values of the relative copy numbers of a given enterotoxin gene showed the copy numbers of that gene relative to the arbitrarily chosen lowest value. Each qPCR run consisted of two replicates for each sample.

Determination of Stability of Enterotoxin Genes in CoNS Isolates
In SE gene-positive CoNS isolates, the stability of these genes was determined. Bacteria were streaked onto BHI agar (BTL, Warsaw, Poland) and incubated overnight at 37 • C. A single, well-separated bacterial colony was transferred to a fresh medium every 24 h. Two to five subsequent passages were carried out. During each passage, a portion of the selected bacterial colony was transferred into BHI broth for genomic DNA isolation. DNA obtained from subsequent bacterial passages was tested by qPCR to determine the relative copy numbers of SE genes, as described above.

Mixed Culture of Bacterial Isolates and Incubation of Bacterial Culture with Genomic DNA
To investigate the possibility of transfer of genetic elements carrying enterotoxin genes between CoNS isolates, the enterotoxigenic S. epidermidis 65.2 strain containing a stable sec gene of high relative copy number was cultured together with the S. pasteuri 65.1 strain, in which the unstable, low-copy sec gene was lost as a result of consecutive passages in a microbial medium. The co-culture was conducted as described before by McCarthy et al. [33] with minor modifications. Serial dilutions of the samples were plated on MSA medium (BTL, Poland) and incubated at 37 • C for 24-48 h. The medium allowed the distinction between S. epidermidis and S. pasteuri based on the colony morphology (yellow versus white colonies). The culture was continued for eight days with sampling after two, five, and eight days. At each time point, 20 well-separated colonies of S. pasteuri 65.1 were collected, transferred into 5 mL of BHI broth, and incubated overnight at 37 • C with agitation at 230 rpm. For each sample, the DNA was extracted to determine the relative copy number of the sec gene. Two selected colonies of S. pasteuri 65.1, positive for the sec gene after co-culture with S. epidermidis 65.2, were subjected to five consecutive passages on an agar medium to assess the stability of the gene.
To evaluate the persistence of extracellular DNA containing enterotoxin genes in CoNS cell culture, the S. pasteuri 65.1 isolate, in which the unstable sec gene was lost, was incubated with genomic DNA of S. epidermidis 65.2. To do so, 6 mL of BHI broth was inoculated with a single colony of S. pasteuri 65.1 and incubated overnight at 37 • C with agitation at 80 rpm. Next, 60 µL of diluted bacterial culture (OD 600 = 0.03) was transferred into 6 mL of fresh BHI medium containing 30 µg of S. epidermidis 65.2 strain genomic DNA (sec +). The amount of genomic DNA added was determined based on the extraction of extracellular DNA from S. epidermidis 65.2 post-culture liquid. S. pasteuri 65.1 was cultured for eight days under the conditions described above. Bacteria were diluted daily by transferring 60 µL of the culture to 6 mL of fresh medium. After two, five, and eight days, samples of 100 µL were taken, and their serial dilutions were plated on BHI agar (overnight incubation at 37 • C). After that, eight colonies were selected for qPCR.

Growth Curve Determination and Sandwich ELISA of Selected Bacterial Isolates
A single bacterial colony was inoculated in 5 mL of BHI broth and incubated overnight at 37 • C with agitation at 230 rpm. Then, the bacterial suspension was diluted to an OD 600 of 0.03 in 100 mL of BHI; the culture was carried out at 37 • C with agitation at 230 rpm for 24 h. Samples for the cell count and ELISA were collected at 6 and 24 h of culture. Bacterial cells were quantified by plating serial dilutions of the culture on BHI agar. At each time point, 2 mL of bacterial culture sample was centrifuged at 12,000× g for 5 min, and 1.5 mL of supernatant was collected and stored at −20 • C until analyzed. To neutralize the A protein, 100 µL of each sample was incubated with the final concentration of 20% rabbit serum (IBSS Biomed S.A., Kraków, Poland) overnight at 4 • C.
Native SEA (Sigma-Aldrich Co., Saint Louis, MO, USA), recombinant rSEC [34], and rSEH [35] were used to determine the standard curve for the ELISA. Recombinant rSEL was obtained as described by Omoe et al. [36]. The coding region of the mature sel gene from S. aureus FRI913 was amplified by PCR and the amplicons were digested with NcoI and XhoI (Thermo Fischer Scientific Inc., Waltham, MA, USA). The digested amplicon was cloned into a pET-22b plasmid vector (Merck, Darmstadt, Germany), introduced into NovaBlue Escherichia coli cells (Merck, Darmstadt, Germany), and next transformed into E. coli Rosettagami DE3 cells (Merck, Darmstadt, Germany). The SEL expression was performed using the IPTG induction (Merck, Darmstadt, Germany). Purification of rSEL was performed on His-Select cobalt affinity gel (Merck, Darmstadt, Germany), with on-column refolding. The purity of rSEL preparation was checked with SDS-PAGE. Polyclonal rabbit anti-SEA (OriGene Technologies, Inc., Rockville, MD, USA), anti-SEC (OriGene Technologies, Inc., Rockville, MD, USA), and anti-SEH (Abcam, Cambridge, UK) antibodies were used to detect SEA, SEC, and SEH enterotoxins. Anti-SEL sera were obtained by immunizing rabbits with 30 µg of rSEL using the protocol described by Lis et al. [35]. Immunization was repeated four times at three-week intervals. Fourteen days after the last immunization, the animals were bled and the serum was prepared. The antibody titer was monitored by indirect ELISA. Monospecific polyclonal anti-SEL antibodies were obtained by affinity purification of the sera with rSEL coupled to CNBr-activated Sepharose 4B (Sigma-Aldrich Co., Saint Louis, MO, USA). The specificity of the antisera was tested by Western blotting, using culture supernatants of enterotoxigenic, sel-negative S. aureus strains FRI1151m and CCM5757, and the sel-positive strains FRI913 and FRI137.
Antibodies conjugated with biotin N-hydroxy succinimide ester (Sigma-Aldrich Co., Saint Louis, MO, USA) were used as secondary antibodies. These conjugates were obtained by incubating samples of anti-SEA, anti-SEC, and anti-SEL sera with a solution of biotin hydroxysuccinylimide ester [35]. A solution of horseradish peroxidase streptavidin conjugate was used to detect biotinylated anti-SEA, anti-SEC, and anti-SEL antibodies. As a secondary serum for the detection of SEH enterotoxin, an anti-SEH horseradish peroxidase conjugate was used (Abcam, Cambridge, UK). ELISA assay was performed as described earlier by Schubert et al. [34]. For the detection of the SEB enterotoxin, the RIDASCREEN ® SET Total kit (R-Biopharm AG, Germany) was used to qualitatively detect staphylococcal enterotoxins (A-E) according to the manufacturer's instructions. The enterotoxin concentration was measured with SEA, rSEC, rSEH, and rSEL as standards, using a four-parameter logistic curve. Plotting standard curves and calculating enterotoxin concentrations was performed using GraphPad Prism 8 software (GraphPad Software, San Diego, CA, USA). All experiments were carried out in duplicate with two biological replicates.

Genome Sequencing and Analysis of Bacterial Isolates
Enterotoxigenic S. pasteuri 65.1 was subjected to NGS sequencing. Genomic DNA was extracted with the MasterPure™ kit (Lucigen, Middleton, WI, USA) according to the manufacturer's instructions and measured using a DeNovix spectrophotometer (DeNovix Inc., Wilmington, DE, USA). De novo whole-genome sequencing was performed using MiSeq (Illumina, San Diego, CA, USA) in the PE250 mode (Genomed S.A., Warsaw, Poland). DNA libraries were prepared using the NEBNext ® DNA Library Prep Master Mix Set for Illumina ® (New England Biolabs, Ipswich, MA, USA). The obtained readings were cut with Cutadapt 1.16 [37]. De novo assembly was performed using the SPAdes algorithm 3.11.1 [38]. The obtained contigs were ordered using Mauve by aligning them to the sequence of the complete genome of the S. pasteuri SP1 (accession number CP004014) isolate available in the GenBank database. The whole genome sequence of the isolate was deposited in the BIGSdb database [23] and analyzed for the presence of genes of 23 known staphylococcal enterotoxins obtained from the GenBank database using the Genome Comparator module of the BIGSdb platform. The genome sequence of S. pasteuri 65.1 has been deposited in the DDBJ/ENA/GenBank under the accession number JAJTJN000000000. The version described in this paper is version JAJTJN010000000.