We describe a comprehensive detection system for 18 kinds of classical and newly described staphylococcal superantigenic toxin genes using four sets of multiplex PCR. Superantigenic toxin genotyping of Staphylococcus aureus for 69 food poisoning isolates and 97 healthy human nasal swab isolates revealed 32 superantigenic toxin genotypes and showed that many S. aureus isolates harbored multiple toxin genes. Analysis of the relationship between toxin genotypes and toxin genes encoding profiles of mobile genetic elements suggests its possible role in determining superantigenic toxin genotypes in S. aureus as combinations of toxin gene-encoding mobile genetic elements.
Staphylococcal enterotoxins (SEs) are emetic toxins, and staphylococcal food poisoning resulting from the consumption of food contaminated with SEs is one of the most common food-borne illnesses . In addition, SEs and the SE-related toxin, toxic shock syndrome toxin-1 (TSST-1), are members of the superantigenic toxin family and have the ability to stimulate large populations of T cells having a particular Vβ element in their T-cell receptors (TCR). This stimulation subsequently leads to a massive proliferation of T cells and the uncontrolled release of proinflammatory cytokines, which cause life-threatening TSS [2–4]. SEs have been divided into five serological types (SEA though to SEE) based on their antigenicity . In recent years, new types of SEs (SEG, SEH, SEI, SEJ, SEK, SEL, SEM, SEN, SEO, SEP, SEQ, SElR and SEU) have been reported [2,5–11]. Several attempts to detect superantigenic toxin, SE and TSST-1 genes in S. aureus isolates have been made; these studies have shown that multiple superantigenic toxin genes are commonly found among S. aureus isolates [12–17]. These newly described SEs have been designated as members of the SE family based on their sequence similarity with classical SEs. The International Nomenclature Committee for Staphylococcal Superantigen Nomenclature (INCSSN) has recommended that only staphylococcal superantigens that induce emesis following oral administration in a monkey model should be designated as SE while other related toxins that either lack emetic properties in this model or have not been tested should be designated staphylococcal enterotoxin-like (SEl) superantigens . Based on this recommendation from INCSSN, the toxins SEJ, SEK, SEL, SEM, SEN, SEO, SEP, SEQ, and SEU should be renamed SElJ, SElK, SElL, SElM, SElN, SElO, SElP, SElQ, and SElU, respectively.
On the other hand, it has been known that certain SE (SEl) and TSST-1 genes are associated with mobile genetic elements such as pathogenicity islands, prophages, and plasmids [6,7,9,19–23]. These facts imply that superantigenic toxin genes are transferred horizontally between staphylococcal strains. There is a possibility that these mobile genetic elements have played an important role in the evolution of S. aureus as a pathogen. To date, there is a need for a method to comprehensively detect and identify the large family of superantigenic toxin genes; such a method would be a powerful tool for evolutionary analysis of the pathogenicity of S. aureus, as well as for diagnostic and epidemiological purposes. Here, we report fine superantigenic toxin genotyping of S. aureus isolates using a multiplex PCR system that is capable of detecting 18 kinds of staphylococcal superantigenic toxin genes. The analysis between superantigenic toxin genotypes and toxin genes encoding profiles of mobile genetic elements provides a hypothesis on possible role for determination of superantigenic toxin genotypes in S. aureus.
2Materials and methods
2.1Bacterial strains and culture media
A total of 177 S. aureus samples were used in this study. Of these, 11 strains were reference strains, including full genome sequencing strains (N315; DDBJ/GenBank/EMBL BA000018, Mu50; BA000017, MW2; BA000033) (Table 1). Sixty-nine isolates were obtained from 30 food poisoning outbreaks diagnosed by 10 local government laboratories in Japan from 1990 to 2002; S. aureus isolates were isolated from patient feces, patient vomit, or the foods involved, and collected from the laboratories. Among these 30 food poisoning outbreaks, 6 were diagnosed as SE-unidentified, meaning that all S. aureus isolates from each outbreak were negative for production of SEA to SED by commercial SET-RPLA kit (DENKA Seiken Co. Ltd., Tokyo, Japan). In this study, we included 21 isolates obtained from these SE-unidentified outbreaks. In addition to the 69 food poisoning isolates, 97 isolates were obtained from nasal swabs of healthy humans in Japan from 2000 to 2004. Bacterial cultures were grown in brain heart infusion (BHI) broth prior to purification of genomic DNA.
Total DNA of S. aureus was purified with the QIAamp DNA purification kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's instructions. The concentration of DNA solution was determined according to A260 values.
The nucleotide sequences of all PCR primers used in this study and their respective amplified products are listed in Table 2. The primer sets used to detect selj, selk, sell, selm, seln, selo, selp, selq and selr genes were designed according to published nucleotide sequences [6,7,9,20–23]. These primer sets were designed to anneal to unique regions and generate amplicons that would allow identification of each se gene based on the molecular weight of its PCR product (Table 2). The primer sets used to detect tst-1 and sea to see were those described by Becker et al. . The primer sets used to detect seg, seh and sei were described by Omoe et al. . As an internal positive control, we used primers that are specific to S. aureus to amplify fem A and fem B genes [16,24]. To construct a multiplex PCR system, four sets (Set 1; sea, seb, sec, sed, see, fem B: Set 2; seg, seh, sei, selj, selp, fem A: Set 3; selk, selm, selo, tst-1, fem A: Set 4; sell, seln, selq, selr, fem B) of 10× primer master mixes (containing 2 μM each primer) were prepared.
Table 2. Nucleotide sequences and predicted size of PCR products for the staphylococcal superantigen-specific oligonucleotide primers
To evaluate the specificity of the newly designed primer sets for detecting selj, selk, sell, selm, seln, selo, selp, selq and selr genes, uniplex PCR using each primer pair was performed. The amplification was performed in an automated thermalcycler with a hot bonnet (Takara PCR Thermal Cycler MP). The reaction mixture (50 μl) for uniplex PCR contained 0.4 μM of each primer, 2 mM MgCl2, 200 μM each of dGTP, dATP, dTTP and dCTP (Takara Syuzo Co., Kyoto, Japan), 0.5U of TaKaRa EX Taq DNA polymerase (Takara), and 5 μl of 10× buffer (Takara). Thermal cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 60 s were repeated 30 times. The DNA fragments obtained from uniplex PCR were subcloned to pGEM-easy vector (Promega, Madison, WI) and subjected to nucleotide sequencing analysis using an ABI3100-avant automatic DNA sequencer (Applied Biosystems, Foster City, CA).
Multiplex PCR of each primer set was performed with QIAGEN Multiplex PCR Kit (QIAGEN) according to manufacturer's instructions. Each reaction mix (50 μl) consisted of 25 μl of 2× QIAGEN Multiplex PCR Master Mix (containing QIAGEN HotStartTaq DNA polymerase, QIAGEN multiplex PCR buffer, and dNTP mix), 0.2 μM of each primer, and 10–100 ng of template DNA. DNA amplification was carried out with the following thermal cycling profile: an initial denaturation of DNA and QIAGEN HotStartTaq DNA polymerase activation at 95°C for 15 min was followed by 35 cycles of amplification (95°C for 30 s, 57°C for 90 s, and 72°C for 90 s), ending with a final extension at 72°C for 10 min. PCR products were resolved by electrophoresis in 3% NuSieve 3:1 agarose gel (Cambrex Bio Science Rockland, Inc., Rockland, ME) in 0.5× TBE (Tris-boric acid-EDTA) buffer, stained by 0.5 μg/ml of EtBr, and visualized on a transilluminator.
3Results and discussion
3.1Development of multiplex PCR system for detection of se and tst-1 genes
First of all, we tried amplifying target DNA of newly designed PCR primers for selj, selk, sell, selm, seln, selo, selp, selq and selr genes. Uniplex PCR using each primer set with total DNA of reference S. aureus strains was performed: S. aureus 196E total DNA for selj and selr; S. aureus MW2 for selk and selq; and S. aureus N315 for sell, selm, seln, selo and selp. The sizes of PCR products obtained by these uniplex PCRs were identical to those predicted from the design of the primers (data not shown). Then, these PCR products were subcloned to pGEM-easy vector and subjected to nucleotide sequencing analysis. The DNA sequences of these clones of se genes were almost exactly identical to the published DNA sequences of the respective se genes. These results showed that the newly designed PCR primer sets could amplify respective se genes with specificity.
The combinations of primer sets and reaction conditions for the multiplex PCR were optimized to ensure that all PCR products of target genes were satisfactorily amplified. We ultimately constructed four optimized multiple primer sets, as described in Section 2. Fig. 1 shows the results of multiplex PCR when total DNAs of reference S. aureus strains were used as a templates. As a positive control, a mixture of total DNA of S. aureus 196E, S6, FRI-326, FRI-569 and N315 was used. Reliable amplification of PCR products was observed in all multiplex PCR reactions using the four primer sets. The sizes of the PCR products obtained from the positive control and the reference strains corresponded to their predicted sizes (Table 2). Furthermore, the toxin gene genotypes of all reference strains determined by multiplex PCR were exactly identical to the toxin gene genotypes determined by full genome sequencing (N315, Mu50 and MW2) or southern blot analysis (196E, S6, FRI-361, FRI-326, FRI-569, 834 and Saga1) (Table 1). When Milli-Q water was used as a negative control instead of template genomic DNA, no PCR products were observed in any of the four sets of multiplex PCR.
3.2Superantigenic toxin gene genotyping of S. aureus isolates from food poisoning outbreaks and healthy human nasal swabs using multiplex PCR
A total of 166 S. aureus isolates were subjected to superantigenic toxin gene genotyping analysis. A total of 32 superantigenic toxin genotypes were observed among the 166 isolates (Table 3). All of the 166 isolates tested harbored the fem A and fem B genes. Of the 69 isolates that originated in food poisoning, all isolates were diagnosed as positive for se genes. Thirteen SE-genotypes were observed in food poisoning-related isolates. Forty (58%) isolates were associated with the sea gene, and the majority of these isolates possessed other se genes. Among the 97 healthy human isolates, only 77 isolates (79.4%) were diagnosed as se-positive. A total of 25 genotypes were observed in healthy human isolates. In contrast to the trend in food poisoning-related isolates, there were only 8 (8.3%) sea-associated isolates.
Table 3. S. aureus superantigenic toxin genotypes and relationship with mobile genetic elements
S. aureus superantigenic toxin genotypes
Suspected genomic islands and plasmids
SFPa (n= 69)
Human nasal swab (n= 97)
Total (n= 166)
aIsolates from Staphylococcal food poisoning outbreaks.
S. aureus harboring classical superantigenic toxin genes
S. aureus harboring classical and new superantigenic toxin genes
sea, sec, sell
?Sa3mu, Type II νSa3
sea, seg, tst-1
?Sa3mu +seg, tst-1
sea, seb, selk, selq
?Sa3mu, νSa1(SaPI3) Or ?Sa3mw +seb
sea, sed, selj, selr
sea, seh, selk, selq
?Sa3mw +seh, or ?Sa3mu +seh, selk, selq
sea, seb, seh, selk, selq
?Sa3mu, νSa1(SaPI3) +seh or ?Sa3mw +seb, seh
sea, seg, sei, seln, tst-1
?Sa3mu +seg, sei, seln, tst-1
sea, seg, sei, selm, seln, selo
?Sa3mu, Type I νSaβ
seb, selk, selq
seb, selk, selq, selp
? Sa3n, νSa1 (SaPI3)
seb, seg, sei, selm, seln, selo
Type I νSaβ+seb
sec, seg, sei, sell, selm, seln, selo
Type II νSa3, Type I νSaβ
sec, seg, sei, sell, selm, seln, selo, tst-1
Type I νSa4 or νSa2, Type I νSaβ
sed, seg, sei, selj, selm, seln, selo, selp, selr
?Sa3n, Type I νSaβ, pIB485
tst-1, seg, seh, seln
tst-1, seg, sei, seln
tst-1, seg, sei, selk, selm, seln, selo
νSa1 (SaPI1), Type I νSaβ
S. aureus harboring new superantigenic toxin genes
seg, sei, selm, seln
seg, sei, selm, seln, selo
Type I νSaβ
seg, sei, selm, seln, selo, selp
?Sa3n, Type I νSaβ
seg, sei, selj, selm, seln, selo, selr
Type I νSaβ, pF5
seg, sei, selk, selm, seln, selo, selq
Type I νSaβ+selk, selq
seh, selk, selq
S. aureus harboring no superantigenic toxin gene
Twenty-one isolates from 6 SE-unidentified food poisoning outbreaks possessed newly identified se genes (seg, sei, selj, selm, seln, selo, selp, or selr). The superantigenic toxin genotypes of isolates within each outbreak were the same (2 outbreaks: seg, sei, selm, seln, selo; 2 outbreaks: seg, sei, selm, seln, selo, selp; 1 outbreak: seg, sei, selj, selm, seln, selo, selr; 1 outbreak: selj, selr). However, it is difficult to conclude that these newly identified SEs were responsible for these food poisoning outbreaks. The emetic activity of the newly described SEs has not been proved, except in the cases of SEG and SEI . To confirm the relationship between these newly identified SEs and food poisoning, it is important to demonstrate the emetic activity of these newly described SEs using an experimental primate model. Recent studies have shown that specific non-primate animal models, such as the ferret and the house musk shrew, respond to SEs and exhibit emetic reactions [25,26]. However, as recommended by INCSSN, the primate model is still the gold standard for estimating the emetic activity of SEs. Moreover, detection of superantigenic toxin genes in S. aureus isolates does not imply expression of these genes by these isolates. Omoe et al.  have shown that seg- and sei-harboring S. aureus isolates produce very low levels of SEG and SEI in vitro, although transcription of mRNA of SEG and SEI in these isolates was proven by reverse-transcriptase PCR analysis. Demonstration of toxin production at levels that are sufficient to cause diseases by strains harboring these se genes is also needed. It has been reported that the production of specific SEs may depend on the host environment and may play a role in the adaptation of S. aureus to the host . SE production should be assessed in vitro, in vivo and in food, using an immunological detection method such as ELISA to confirm the relationship between newly identified SEs and diseases.
3.3The relationship between superantigenic toxin genotypes and toxin gene-encoding mobile genetic elements
In the present study, we have shown that there are many superantigenic toxin genotypes in S. aureus isolated from food poisoning outbreaks or healthy human nasal swabs. It has been known that almost all superantigenic toxin genes are associated with mobile genetic elements such as genomic islands (pathogenicity islands, prophages and staphylococcal cassette chromosomes) and plasmids. Thus, we analyzed the relationship between superantigenic toxin genotypes obtained in this study and known superantigenic toxin gene-encoding mobile genetic elements. Table 3 summarizes the relationship between superantigenic toxin genotypes and known mobile genetic elements. One half (16/32) of superantigenic toxin genotypes observed in this study could be considered as combinations of known superantigenic toxin gene-encoding profiles of genomic islands or plasmids. For example, genotype sea, sec, sell could be a combination of ?Sa3mu (sea) and Type II νSa3 (sec, sell), and genotype sed, seg, sei, selj, selm, seln, selp, selr could be a combination of pIB485 (sed, selj, selr), Type I νSaβ (seg, sei, selm, seln, selo) and ?Sa3n (selp). Of the remaining 16 genotypes, 7 could be considered as combinations of known mobile genetic elements plus particular se genes. For example, genotype sea, seh, selk, selq could be a combination of ?Sa3mw (sea, selk, selq) plus seh or ?Sa3mu (sea) plus seh, selk, selq. The remaining 9 genotypes, such as “seb, seh”, “seh, selk, selq”, “tst-1, seg, seh, seln”, did not follow the rule of known superantigenic toxin gene profiles of mobile genetic elements. In these genotypes, we observed several gene combinations that could be considered incomplete Type I νSaβ, such as “seg, sei, selm, seln”, “seg, sei, seln”, “selm, selo”, and “seln”. Becker et al.  reported the prevalence of Type I νSaβ-related SE genes, and showed that a substantial number of isolates were found to harbor only one or two of the selm, seln, and selo genes. These results suggest the possibility of the existence of SE-encoding variants within the genomic island Type I νSaβ, or the existence of new types of mobile genetic elements encoding seg, sei, selm, seln, or selo genes. There is also the possibility of existence of many new types of toxin-gene-encoding mobile genetic elements. As shown above, it seems that the se genotype of S. aureus may be determined by mobile genetic elements it harbors. To prove this hypothesis, an effort to explore new types of mobile genetic element is needed, as well as detailed characterization of SE-encoding genomic islands. Previously, Baba et al.  mentioned that genomic island allotyping would be a useful approach to S. aureus genotyping and that this process would enable the prediction of the pathogenic capability of an S. aureus clinical strain. Our multiplex PCR system for detecting superantigenic toxin genes will be useful in determining genomic island allotypes. Recently, Sergeev et al.  reported a PCR-based microarray assay system for simultaneous detection of SE genes. This microarray system would be a powerful method for detecting several types of se genes simultaneously. However, equipment for microarrays is expensive, and microarrays are not widely used in common laboratories at present. By contrast, our PCR-based superantigenic toxin gene detection system could be performed easily in commonly equipped clinical laboratories.
In conclusion, the newly developed multiplex PCR system for comprehensive detection and identification of staphylococcal superantigenic toxin genes described here is a potentially powerful tool for diagnosis and epidemiological study of S. aureus. The data presented here suggest the system's potential role in determining superantigenic toxin genotypes as combinations of toxin gene-encoding mobile genetic elements, such as genomic islands and plasmids. Further exploration and characterization of new types of mobile genetic elements are needed.
This work was partly supported by grants-in-aids for scientific research from the Japan Society for the Promotion of Science (Grants 15580272 and 16380205).
We thank Dr. Keichi Hiramatsu (at Juntendo University) for kindly providing the S. aureus strains used in this work.