Biological Properties of Staphylococcal Enterotoxin-Like Toxin Type R

Abstract

We investigated the biological properties of a novel staphylococcal enterotoxin-like putative toxin, staphylococcal enterotoxin-like toxin type R (SElR). Major histocompatibility complex class II molecules were required for T-cell stimulation by SElR. SElR stimulated T cells bearing receptors Vβ 3, 11, 12, 13.2, and 14. These results suggested that SElR acts as a superantigen.

Staphylococcus aureus is an important human and animal pathogen. This organism causes a broad range of diseases in humans including septicemia, toxic shock syndrome (TSS), and food poisoning, as well as bovine mastitis. Staphylococcal enterotoxins (SEs) and TSS toxin 1 are members of the superantigenic toxin family and have the ability to stimulate large populations of T cells that have a particular Vβ element of the T-cell receptor (TCR). This stimulation subsequently leads to a massive proliferation of T cells and the uncontrolled release of proinflammatory cytokines, which cause life-threatening TSS (3, 12). Five major serological types, SEA through SEE, have been characterized based on their antigenicity (2). These classical SEs are emetic toxins and causative agents in staphylococcal food poisoning. However, in recent years, many new types of SE or SE-like putative toxins have been identified (8, 9, 10, 12, 16, 17, 18, 19, 21). These new SEs and SE-like toxins were designated as members of the SE family based on their sequence similarity with classical SEs. At present, the relationship between these new SEs and human diseases including food poisoning and TSS is not fully understood.

SER is the newest SE-related putative toxin, described by Omoe et al. (16), who showed that the SER amino acid sequence had the highest homology to the SEG sequence and found that SER showed significant T-cell stimulation activity. Here we report the superantigenicity of SER and the SER production levels of S. aureus strains harboring ser gene-carrying plasmids. Based on these findings and the recommendations of the International Nomenclature Committee for Staphylococcal Superantigen Nomenclature, we rename this SE-like putative toxin as staphylococcal enterotoxin-like toxin type R (SElR).

Analysis of the requirement of MHC class II molecules for activation of T cells by SElR.

We investigated whether the presence of major histocompatibility complex (MHC) class II molecules on accessory cells (ACs) is required for SElR to activate T cells. Recombinant SElR was expressed and purified according to the work of Omoe et al. (16). Human peripheral blood mononuclear cells (PBMCs) were obtained from healthy donors by Ficoll-Conray density gradient centrifugation. T cells were obtained by the S-2-aminoethylisothiouronium-treated sheep red blood cell rosette method. They were further enriched by removal of CD16-, CD14-, CD19-, and HLA-DR-positive cells with the use of monoclonal antibodies (MAbs) to those antigens and anti-mouse immunoglobulin-coated magnetic beads (Dynabeads; Dynal, Oslo, Norway) (7). L cells transfected with the DR4 gene (8124 L cells) and control L cells (8400) were prepared as described previously (6, 13). These L cells were treated with mitomycin C, irradiated with an MBR-1404R X-ray generator (Hitachi, Tokyo, Japan) to block proliferation, and used as ACs for T-cell activation by SElR. To measure interleukin-2 (IL-2) production from stimulated T cells, IL-2 activity in the culture supernatants was determined with IL-2-dependent CTLL-2 cells. Data are presented as units of IL-2 per milliliter (23). Purified T cells from human PBMCs were stimulated in vitro with 5 or 10 μg of SElR/ml in the presence or absence of L cells transfected with DR genes (8124) or of control L cells (8400). The T cells were then examined for IL-2 production. The effect of the antibody to HLA-DR on the T-cell response was examined in parallel (Table 1). Substantial levels of IL-2 production were seen in the presence of DR-transfected 8124 L cells but not in the presence of control 8400 L cells. Anti-DR MAb markedly inhibited the SElR-induced IL-2 production from T cells in the presence of 8124 L cells. Anti-mouse MHC class I MAb (anti-H-2K^k) did not influence the T-cell response induced by SElR.

TABLE 1.

Requirement for the presence of MHC class II molecules on ACs in T-cell activation by SElR^a

Expt no.	AC line	Antibody	IL-2 production (U/ml)
			With SElR at:			With SEA at 10 ng/ml
			0 μg/ml	5 μg/ml	10 μg/ml
1	None	None	<0.1	<0.1
	8400	None	<0.1	<0.1
	8124	None	<0.1	17.2
	8124	Anti-HLA-DR	<0.1	8.3
	8124	Anti-H-2Kk	<0.1	19.4
2	None	None	<0.1		<0.1
	8400	None	<0.1		<0.1
	8124	None	<0.1		7.8
	8124	Anti-HLA-DR	<0.1		<0.1
	8124	Anti-H-2Kk	<0.1		8.4
3	None	None	<0.1	<0.1		<0.1
	8400	None	<0.1	<0.1		<0.1
	8124	None	<0.1	40.6		73.4
	8124	Anti-HLA-DR	<0.1	7.9		9.3
	8124	Anti-H-2Kk	<0.1	42.8		79.1

^aMAb I2C3 (reactive with HLA-DR molecules) or MAb P3-83P (reactive with H-2K^k) was added as a 1:100 dilution of ascites fluid to culture before addition of SElR or SEA. 8400, control L cells; 8124, HLA-DR4⁺ L cells.

Analysis of the TCR Vβ repertoire of SElR-reactive human T cells.

SElR- or anti-CD3-induced T-cell blasts were obtained by stimulating PBMCs with 10 μg of SElR/ml or 1 μg of MAb to anti-CD3 (OKT3)/ml for 3 days and expanding harvested blasts for 4 days in the presence of 100 U of recombinant human IL-2 (Shionogi, Osaka, Japan)/ml. T-cell blasts stimulated with SElR or anti-CD3 MAb were stained with MAbs to TCR Vβ elements (IOTest Beta Mark kit; Beckman Coulter, Miami, Fla.). Samples were analyzed on an EPICS XL flow cytometer (Beckman Coulter) with FlowJo software, as described previously (22). Figure 1 shows the representative results of three experiments. It was observed that T cells bearing TCR Vβ 3, 11, 12, 13.2, and 14 were preferentially activated by SElR. T cells bearing TCR Vβ 14 responded with the highest level of expansion (approximately 30% of T cells in three donors). The primary structure of SElR has homology to that of SEG (65.9% of identity) (16). Previously, Jarraud et al. examined the TCR Vβ composition of T cells stimulated by recombinant SEG using the immunoscope technique. They showed that T cells bearing TCR Vβ 3, 12, 13a, 14, and 15 were selectively expanded by SEG stimulation (8). Furthermore, T cells bearing TCR Vβ 14 showed the highest response to both SEG and SElR. These similarities in the TCR Vβ skewing of SEG and SElR reflect the similarity of the primary structures of these superantigens.

FIG. 1.

TCR Vβ profile of SER. Human PBMCs were stimulated with either anti-CD3 antibody or SER. Cells were stained with MAbs and analyzed by flow cytometry. Data are presented as percentages of the T-cell fractions expressing distinct Vβ elements in whole T cells. Increases in comparison with anti-CD3 stimulation observed in three donors are indicated with an asterisk.

Development of sandwich enzyme-linked immunosorbent assay (ELISA) for detection of SElR.

Rabbit anti-SElR immunoglobulin G was purified from hyperimmune serum (16) by use of an immobilized protein G-Sepharose column (Amersham Pharmacia Biotech Inc., Piscataway, N.J.) and was used as capture antibody. Monospecific rabbit anti-SElR antibody was affinity purified from hyperimmune serum by use of an SElR-coupled Sepharose column. One milligram of monospecific antibody was conjugated to EZ-Link Plus horseradish peroxidase (Pierce, Rockford, Ill.) according to the manufacturer's instructions. ELISA was performed in 96-well Nunc microplates (Nalge Nunc International, Rochester, N.Y.) according to the work of Omoe et al. (15). The concentration of each toxin was determined by converting the absorbance values to the corresponding concentrations via the standard curve. Straight lines were observed at concentrations of SElR between 1.0 and 10 ng/ml (Fig. 2). SElR was not detectable at 0.5 ng/ml. Additionally, 100 ng of purified SEA, SEB, SEC, SED, SEE, SEG, SEH, and SEI was subjected to sandwich ELISA, and no cross-reactivity was observed (data not shown).

FIG. 2.

Standard curve for detection of SElR by ELISA. A linear increase was obtained between 1.0 and 10 ng of SElR/ml.

Productivities of SElR in S. aureus strains harboring the selr gene.

Ten S. aureus strains harboring the selr gene, four selr-negative S. aureus strains, two Staphylococcus epidermidis strains, and two Escherichia coli strains were cultured in brain heart infusion broth supplemented with 1% yeast extract with shaking for 40 h, and the supernatants were subjected to sandwich ELISA. Culture supernatants were preincubated with 20% (vol/vol) normal rabbit serum at 4°C overnight and then diluted 10-, 100-, and 1,000-fold in phosphate-buffered saline containing 0.05% Tween 20 to avoid any nonspecific reaction caused by protein A (5, 15). Thus, the minimum detectable concentration of SElR in culture supernatant was 10 ng/ml. Nine out of 10 S. aureus isolates produced significant levels of SElR (38.2 to 324.8 ng/ml), while only one human nasal swab isolate did not produce a detectable level (10 ng/ml) of SElR (Table 2). The selr gene exists on two kinds of plasmids, a pIB485-related plasmid and a pF5-related plasmid. The SElR production levels did not differ between S. aureus strains harboring a pIB485-related plasmid and those harboring a pF5-related plasmid (P = 0.3, Mann-Whitney U test). All culture supernatants of selr-negative S. aureus strains, S. epidermidis strains, and E. coli strains were negative for SElR production (Table 2).

TABLE 2.

Specificity of sandwich ELISA and SElR productivities of S. aureus isolates harboring the selr gene

Strain	SE genotypea (SElR-encoding plasmid)	SElR productivity (ng/ml)	Source	Reference(s)
S. aureus
selr positive
FRI-361	sec sed sej seg sei selr (pIB485-like)	52.9	Reference strain	15, 16
196E	sea sed sej selr (pIB485-like)	45.1	Reference strain	16
11727	sea sed sej selr (pIB485-like)	323.6	Food poisoning	16
11740	sea sed sej selr (pIB485-like)	324.8	Food poisoning	16
IVM7	sed sej seg sei selr (pIB485-like)	<10	Human nasal swab	16
IVM16	sed sej seg sei selr (pIB485-like)	161.9	Human nasal swab	16
IVM46	sed sej seg sei selr (pIB485-like)	38.2	Human nasal swab	16
Fukuoka 5	sej selr (pF5)	211.8	Food poisoning	16
Fukuoka 6	sej selr (pF5-like)	198.0	Food poisoning	16
Fukuoka 7	sej selr (pF5-like)	209.4	Food poisoning	16
selr negative
S6	sea seb	<10	Reference strain	15
FRI-326	see	<10	Reference strain	15
FRI-569	seh	<10	Reference strain	21
834	sec seg sei tst-1	<10	Hospital	14
S. epidermidisb
HUH-2	SE negative	<10	Hospital	This work
HUH-3	SE negative	<10	Hospital	This work
E. coli
DH5α	SE negative	<10	Toyobo	20
BL21	SE negative	<10	Stratagene	20

^aSE genotype of staphylococcal strains was determined by multiplex PCR (sea to see, seg to sei, and tst-1) (15) and Southern blotting (sed, sej, and selr) (16).

^bS. epidermidis strains were identified by API-Staph (Biomerieux, Marcy l'Etoile, France).

In conclusion, SElR acts as a superantigen, and the biological activities of SElR most resemble those of SEG. Most S. aureus strains harboring the selr gene produce a significant amount of SElR. We have not tested the emetic activity of SElR using the primate model. The International Nomenclature Committee for Staphylococcal Superantigen Nomenclature has proposed that only staphylococcal superantigens that induce emesis following oral administration in a primate model be designated SEs and that other, related toxins that either lack emetic properties in this model or have not been tested should be designated staphylococcal enterotoxin-like (SEl) superantigens to indicate that their potential role in staphylococcal food poisoning has not been confirmed (personal communication from Keiichi Hiramatsu). Based on this criterion, we rename this staphylococcal enterotoxin-like superantigen as SElR, instead of SER. Further analysis with the primate model is needed to clarify the role of SElR in staphylococcal food poisoning.

As described above, the selr gene is carried by two kinds of plasmids, and the selr gene coexists with sed and sej (pIB485-related plasmid) or sej (pF5-related plasmid). In addition, a recent full genome sequencing analysis (1, 9) and genetic analysis of staphylococcal pathogenicity islands (4, 11, 24) showed that pathogenic factors including SEs and TSS toxin 1 were encoded by mobile genetic elements and several toxin genes linked to particular mobile genetic elements. Orwin et al. suggested that the presence of multiple expressed toxins in a single S. aureus strain makes it less clear whether any single toxin is responsible for diseases and that it is better to assume that these S. aureus strains express multiple toxins with similar activities and that all could contribute to disease (17). In other words, staphylococcal diseases related to superantigenic and emetic toxins, such as food poisoning and TSS, may be assumed to be a complex phenotype caused by these toxins with similar activities. To clarify the nature of SE- and enterotoxin-like toxin-related diseases, it is important to study the biological properties of newly identified SEs and SE-like toxins and to assess the toxin production levels of the S. aureus strains harboring the genes for the new SEs and SE-like toxins in vitro, in vivo, and in foods.