Superantigens and pseudosuperantigens of gram-positive cocci

B cell superantigens in the human intestinal microbiota

IgA is prominently secreted at mucosal surfaces and coats a fraction of the commensal microbiota, a process that is critical for intestinal homeostasis. However, the mechanisms of IgA induction and the molecular targets of these antibodies remain poorly understood, particularly in humans. Here, we demonstrate that microbiota from a subset of human individuals encode two protein "superantigens" expressed on the surface of commensal bacteria of the family Lachnospiraceae such as Ruminococcus gnavus that bind IgA variable regions and stimulate potent IgA responses in mice. These superantigens stimulate B cells expressing human VH3 or murine VH5/6/7 variable regions and subsequently bind their antibodies, allowing these microbial organisms to become highly coated with IgA in vivo. These findings demonstrate a previously unappreciated role for commensal superantigens in host-microbiota interactions. Furthermore, as superantigen-expressing strains show an uneven distribution across human populations, they should be systematically considered in studies evaluating human B cell responses and microbiota during homeostasis and disease.

Human leukocyte antigen-DQ8 transgenic mice: a model to examine the toxicity of aerosolized staphylococcal enterotoxin B

Staphylococcal enterotoxins (SEs) belong to a large group of bacterial exotoxins that cause severe immunopathologies, especially when delivered as an aerosol. SEs elicit the release of lethal amounts of cytokines by binding to major histocompatibility complex (MHC) class II and cross-linking susceptible T-cell receptors. Efforts to develop effective therapeutic strategies to protect against SEs delivered as an aerosol have been hampered by the lack of small animal models that consistently emulate human responses to these toxins. Here, we report that human leukocyte antigen-DQ8 (HLA-DQ8) transgenic (Tg) mice, but not littermate controls, succumbed to lethal shock induced by SEB aerosols without potentiation. Substantial amounts of perivascular edema and inflammatory infiltrates were noted in the lungs of Tg mice, similar to the pathology observed in nonhuman primates exposed by aerosol to SEB. Furthermore, the observed pathologies and lethal shock correlated with an upsurge in proinflammatory cytokine mRNA gene expression in the lungs and spleens, as well as with marked increases in the levels of proinflammatory circulating cytokines in the Tg mice. Unlike the case for littermate controls, telemetric evaluation showed significant hypothermia in Tg mice exposed to lethal doses of SEB. Taken together, these results show that this murine model will allow for the examination of therapeutics and vaccines developed specifically against SEB aerosol exposure and possibly other bacterial superantigens in the context of human MHC class II receptors.

Interleukin-12 alone can not enhance the expression of the cutaneous lymphocyte associated antigen (CLA) by superantigen-stimulated T lymphocytes

It has been reported that bacterial superantigens induce interleukin (IL)-12 dependent expression of the cutaneous lymphocyte associated antigen (CLA) and that this may be relevant to the association between certain skin diseases and infections including psoriasis and streptococcal tonsillitis. We have confirmed that the streptococcal pyrogenic superantigen C (SpeC) increases CLA expression by both CD4+ and CD8+ T cells when PBMCs are incubated in medium enriched with fetal calf serum (FCS). However, such an increase could not be induced in medium enriched with human serum (HS) even when recombinant IL-12 was added to the PBMCs cultures. Strikingly, CD4+ T cells incubated with SpeC in HS showed a marked reduction in CLA expression, which was not due to apoptosis. In contrast, SpeC did induce T cell proliferation and expression of CD25, CD54 and CD103 in the presence of HS indicating that the absence of SpeC induced CLA expression in HS was not due to SpeC inhibitors. Although addition of low amounts of lipopolysaccharide endotoxin (LPS) caused a highly significant increase in CLA expression in the absence of SpeC in cultures enriched with HS, a combination of LPS and SpeC did not increase CLA expression beyond that induced by LPS alone. The superantigen-induced CLA expression in FCS was partially inhibited by anti-IL-12 but not by anti-IL-18 or antibodies to transforming growth factor (TGF)-beta. It is concluded that IL-12 alone can not increase CLA expression but requires the help of other factor(s) present in FCS but not in HS. Although LPS can induce CLA expression it does not seem to be the factor that interacts with IL-12 to induce superantigen-mediated CLA expression in cultures enriched with FCS.

Desmoglein as a target in autoimmunity and infection

Clinical phenotypes of most diseases are complex. However, once the mechanism behind the scene is clarified, the nature shows amazing beauty. There is a simple logic behind a complex disease. The exact molecular mechanism of the blister formation in staphylococcal scalded skin syndrome (SSSS) remained to be elucidated for 3 decades since exfoliative toxin was discovered by Melish and Glasgow in 1970. A knowledge accumulated to understand the pathogenesis of pemphigus and cell-cell adhesion of keratinocytes led us to solve this question. Desmoglein 1, which is a cadherin type cell-cell adhesion molecule in desmosomes, is targeted in two different skin diseases, pemphigus foliaceus, and SSSS. In pemphigus foliaceus IgG autoantibodies are developed against desmoglein 1 and inhibit its adhesive function with resultant blister formation in the superficial epidermis. In SSSS, exfoliative toxin produced by Staphylococcus aureus specifically binds and cleaves desmoglein 1 with resultant blister formation at the identical site.

Staphylococcal exfoliative toxin B specifically cleaves desmoglein 1

Staphylococcal scalded skin syndrome and its localized form, bullous impetigo, show superficial epidermal blister formation caused by exfoliative toxin A or B produced by Staphylococcus aureus. Recently we have demonstrated that exfoliative toxin A specifically cleaves desmoglein 1, a desmosomal adhesion molecule, that when inactivated results in blisters. In this study we determine the target molecule for exfoliative toxin B. Exfoliative toxin B injected in neonatal mice caused superficial epidermal blisters, abolished cell surface staining of desmoglein 1, and degraded desmoglein 1 without affecting desmoglein 3 or E-cadherin. When adenovirus-transduced cultured keratinocytes expressing exogenous mouse desmoglein 1 or desmoglein 3 were incubated with exfoliative toxin B, desmoglein 1, but not desmoglein 3, was cleaved. Furthermore, cell surface staining of desmoglein 1, but not that of desmoglein 3, was abolished when cryosections of normal human skin were incubated with exfoliative toxin B, suggesting that living cells were not necessary for exfoliative toxin B cleavage of desmoglein 1. Finally, in vitro incubation of the recombinant extracellular domains of desmoglein 1 and desmoglein 3 with exfoliative toxin B demonstrated that both mouse and human desmoglein 1, but not desmoglein 3, were directly cleaved by exfoliative toxin B in a dose-dependent fashion. These findings demonstrate that exfoliative toxin A and exfoliative toxin B cause blister formation in staphylococcal scalded skin syndrome and bullous impetigo by identical molecular pathophysiologic mechanisms.

Human endogenous retrovirus IDDMK(1,2)22 and mouse mammary tumor virus superantigens differ in their ability to stimulate murine T cell hybridomas

Recently, a newly identified human HERV-K18 like endogenous retrovirus (IDDMK(1,2)22) has been associated to the etiology of type I diabetes (IDDM). Although the exact mechanism remains unclear, it was postulated that the 3' end ORF product of the env gene of IDDMK(1,2)22 would trigger a V beta 7-specific human T cell expansion leading to their infiltration in the pancreas of afflicted patients and to the autoimmune destruction of the insulin-producing beta cells. Since then, such superantigen (SAg)-like activity as well as the association between the IDDMK(1,2)22 virus and IDDM pathogenesis have been challenged. To further characterize functionally the putative IDDMK(1,2)22-encoded SAg, we have cloned from human DNA the identical 462bp ORF sequence originally described. The IDDMK(1,2)22 ORF fragment was transfected in the same human B cell line (Raji) originally used as APC to demonstrate the V beta 7 specificity. The immunostimulatory potential of IDDM ORF was tested on murine T cell hybridomas and compared to the well-characterized mouse mammary tumor virus Mtv7 SAg transfected in the same conditions. A panel of 16 T cell hybridomas encompassing 14 different V betas was analyzed. We have failed to detect IDDMK(1,2)22-induced IL-2 production from any of these hybridomas, even those bearing the murine V beta 1 mV beta 1, V beta 4 or V beta 10 TcR beta chains which are most closely related to the human V beta 7 (hV beta 7). Our results suggest that IDDMK(1,2)22 ORF is devoid of superantigenic activity as defined by classical criteria.

Structural basis for abrogated binding between staphylococcal enterotoxin A superantigen vaccine and MHC-IIalpha

Staphylococcal enterotoxins (SEs) are superantigenic protein toxins responsible for a number of life-threatening diseases. The X-ray structure of a staphylococcal enterotoxin A (SEA) triple-mutant (L48R, D70R, and Y92A) vaccine reveals a cascade of structural rearrangements located in three loop regions essential for binding the alpha subunit of major histocompatibility complex class II (MHC-II) molecules. A comparison of hypothetical model complexes between SEA and the SEA triple mutant with MHC-II HLA-DR1 clearly shows disruption of key ionic and hydrophobic interactions necessary for forming the complex. Extensive dislocation of the disulfide loop in particular interferes with MHC-IIalpha binding. The triple-mutant structure provides new insights into the loss of superantigenicity and toxicity of an engineered superantigen and provides a basis for further design of enterotoxin vaccines.

Basic streptococcal superantigens (SPEX/SMEZ or SPEC) are responsible for the mitogenic activity of the so-called mitogenic factor (MF)

The mitogenic factor (MF) of group A streptococci has been reported to be a superantigen stimulating human T cells carrying Vbeta2, 4 and 8 and has been designated streptococcal pyrogenic exotoxin F (SPEF). MF was also shown to possess DNase activity. Here we have purified MF from culture supernatants of different Streptococcus pyogenes strains. Surprisingly, the MF preparations from different strains showed different Vbeta specificities depending on the expression of SPEC or SMEZ3 by the producing strain. Their mitogenic activity decreased upon further purification. In addition, the mitogenic activity could be only neutralized by antibodies against the basic streptococcal superantigens SPEC or SPEX (SMEZ3) but not by antibodies against MF itself although the latter were able to neutralize completely the DNase activity of MF. We found that streptodornase type B (SDB) was expressed in two molecular forms (SDBI and SDBII), differing only by one additional N-terminal arginine at SDBI. MF was found identical to the enzyme SDBII but is devoid of superantigenic properties and should no longer be called a superantigen or a pyrogenic exotoxin.

Virulence of group A streptococci in fertile hens eggs is mainly effected by M protein and streptolysin O

In this study we have investigated whether streptolysin O contributes to the virulence of group A streptococci. For this purpose we generated M-negative and SLO-negative mutants by insertion mutagenesis into the chromosome of an M type 1 strain. The inactivation of M1 protein expression was achieved by the construction of the integrative plasmid pSFABS, which contains the internal fragment abs of the emm1 gene. Integration of pSFABS by homologous recombination into the chromosome of strain 38 541 resulted in the generation of mutant EMM1. Inactivation of slo with plasmid pFWSLOD resulted in two different mutant forms. The homologous recombination with plasmid pFWSLOD carrying the two slo fragments slo1 (899 base pairs in the 5' region) and slo2 (709 base pairs in the downstream part) resulted in mutants SLO3, SLO4 and SLO17. In SLO17 a double crossover event took place with insertion of the spectinomycin resistance gene aad9 between the slo fragments 1 and 2. In mutants SLO3 and SLO4 the homologous recombination with the same plasmid led to the integration of the whole plasmid construct into the chromosome of strain 38 541. Both forms of mutation failed to express SLO. In mutant SLO4 additionally M1 protein expression was significantly decreased. The mutants EMM1 (M-, SLO+) and SLO4 (M decreased, SLO-) showed a reduced binding to collagen-coated surfaces. In contrast the mutants SLO3 and SLO17 (both M+, SLO-) and the wild-type strain 38 541 (M+, SLO+) showed an affinity to collagen similar to purified M1 protein. All mutants were less virulent for chicken embryos compared to the wild-type strain after infection by intravenous injection as well as by application onto the chorioallantoic membrane. The results show that besides M protein SLO can also influence virulence of group A streptococci. Moreover, it became obvious that streptococci need more than one tool to fully develop their infectious potential.

Purification and biochemical characterization of a basic superantigen (SPEX/SMEZ3) from Streptococcus pyogenes

A potent basic superantigen (designated streptococcal pyrogenic exotoxin X, SPEX/SMEZ3) was purified to homogeneity from culture supernatants of a Streptococcus pyogenes scarlatina strain of type 12 (genotype speA(-), speC(-)) and characterized. Sequence alignments revealed SPEX to be an allele of the streptococcal mitogens type Z (SMEZ). The N-terminal amino acid sequence of SPEX was found with LEVDNNSLLR to be identical to the recently described acidic superantigen SMEZ. Although SPEX/SMEZ genes were present in all of the streptococcal strains tested, a toxin production could only be detected in a small number of strains. The produced toxin concentration in the culture supernatants of positive strains differed between 0 and 20 ng ml(-1). The purified SPEX stimulated human T-lymphocytes with Vbeta8 specificity at extremely low concentrations (lower than 100 pg ml(-1)).

The superantigenic activity of streptococcal pyrogenic exotoxin B is independent of the protease activity

The nature of the mitogenic activity of pyrogenic streptococcal exotoxin B, also known as streptococcal cysteine protease, has been debated in the literature. Streptococcal exotoxin B has been shown to cleave interleukin-1beta precursor and create biologically active interleukin-1beta, a major cytokine mediating inflammation and shock. This activity could mimic the mitogenicity and cytokine release induced by superantigens in lymphocyte stimulating experiments. In this study, the protease activity of streptococcal exotoxin B was irreversibly inhibited by covalent binding of a tripeptide and the superantigenic properties of streptococcal exotoxin B were found not to be influenced by this inactivation. Native as well as protease-inactivated streptococcal exotoxin B was shown to stimulate T-cell proliferation without a need of metabolically active antigen presenting cells. Furthermore, streptococcal exotoxin B-induced T-cell proliferation was shown to require HLA-DQ since addition of HLA-DQ monoclonal antibodies totally inhibited the mitogenic activity of streptococcal exotoxin B, indicating that streptococcal exotoxin B, as other superantigens, makes direct contact with the T-cell receptor via HLA class II. The aim of this study was to characterize the relationship between the proteolytic and superantigenic properties of streptococcal exotoxin B.

Clinical, microbial, and biochemical aspects of the exfoliative toxins causing staphylococcal scalded-skin syndrome

The exfoliative (epidermolytic) toxins of Staphylococcus aureus are the causative agents of the staphylococcal scalded-skin syndrome (SSSS), a blistering skin disorder that predominantly affects children. Clinical features of SSSS vary along a spectrum, ranging from a few localized blisters to generalized exfoliation covering almost the entire body. The toxins act specifically at the zona granulosa of the epidermis to produce the characteristic exfoliation, although the mechanism by which this is achieved is still poorly understood. Despite the availability of antibiotics, SSSS carries a significant mortality rate, particularly among neonates with secondary complications of epidermal loss and among adults with underlying diseases. The aim of this article is to provide a comprehensive review of the literature spanning more than a century and to cover all aspects of the disease. The epidemiology, clinical features, potential complications, risk factors, susceptibility, diagnosis, differential diagnoses, investigations currently available, treatment options, and preventive measures are all discussed in detail. Recent crystallographic data on the toxins has provided us with a clearer and more defined approach to studying the disease. Understanding their mode of action has important implications in future treatment and prevention of SSSS and other diseases, and knowledge of their specific site of action may provide a useful tool for physiologists, dermatologists, and pharmacologists.

Understanding the mechanism of action of bacterial superantigens from a decade of research

In the face of the unique diversity and plasticity of the immune system pathogenic organisms have developed multiple mechanisms in adaptation to their hosts, including the expression of a particular class of molecules called superantigens. Bacterial superantigens are the most potent stimulators of T cells. The functional consequences of the expression of superantigens by bacteria can be extended not only to T lymphocytes, but also to B lymphocytes and to cells of the myeloid compartment, including antigen-presenting cells and phagocytes. The biological effects of bacterial superantigens as well as their molecular aspects have now been studied for a decade. Although there is still a long way to go to clearly understand the role these molecules play in the establishment of disease, recently acquired knowledge of their biochemistry now offers unique experimental opportunities in defining the molecular rules of T-cell activation. Here, we present some of the most recent functional and molecular aspects of the interaction of bacterial superantigens with MHC class II molecules and the T-cell receptor.

Acute Hepatotoxicity of Pseudomonas aeruginosa Exotoxin A in Mice Depends on T Cells and TNF

The most potent virulence factor of Pseudomonas aeruginosa, its exotoxin A (PEA), inhibits protein synthesis, especially in the liver, and is a weak T cell mitogen. This study was performed to correlate hepatotoxic and possible immunostimulatory features of PEA in vivo. Injection of PEA to mice caused hepatocyte apoptosis, an increase in plasma transaminase activities, and the release of TNF, IFN-γ, IL-2, and IL-6 into the circulation. Most strikingly, liver damage depended on T cells. Athymic nude mice or mice depleted of T cells by anti-Thy1.2 mAb pretreatment failed to develop acute hepatic failure, and survival was significantly prolonged following T cell depletion. Neutralization of TNF or lack of TNF receptors prevented liver injury. In the liver, TNF was produced by Kupffer cells before hepatocellular death occurred. After T cell depletion, Kupffer cells failed to produce TNF. Transaminase release was significantly reduced in perforin knockout mice, and it was even elevated in lpr/lpr mice. These results demonstrate that PEA induces liver damage not only by protein synthesis inhibition but also by TNF- and perforin-dependent, Fas-independent, apoptotic signals.

Identification of domains involved in superantigenicity of streptococcal pyrogenic exotoxin F (SpeF)

A series of 11 synthetic peptides of 30 amino acids, each with 10 amino acids overlap which spanned the entire sequence of streptococcal pyrogenic exotoxin F (SpeF), were employed in proliferation studies on human peripheral blood mononuclear cells (PBMCs). Regions 41-70, 141-170 and 181-210 were identified as important for SpeF-induced lymphocyte activation. Secondary structure predictions of these peptides showed similarities to regions in other superantigens known to be important for T cell mitogenicity. Furthermore, antisera specific to peptides covering amino acids 1-70 and 181-228 were able to inhibit SpeF-induced mitogenicity by 25% when pre-incubated with SpeF prior to PBMC activation.

Neurohormonal host defense in endotoxin shock

Lipopolysaccharide (LPS) of gram-negative bacteria is capable of activating the immune system of higher animals, which may lead to cytokine-induced lethal shock and death. LPS has little toxicity for the frog and fish, but it kills the horseshoe crab instantly by causing intravascular blood coagulation. The response to LPS evolved from simple reactions in lower animals into an intense reaction in mammals that involves a massive immune activation leading to a profound neuroendocrine and metabolic response. This is now known as the acute-phase response (APR). During APR, LPS-binding proteins (LBP) are produced by the liver in rapidly increasing quantities under the influence of interleukin-6, glucocorticoids, and catecholamines. After combination with LPS, LPB is capable of activating monocyte-macrophages and granulocytes via the CD14 surface receptor. Other receptors (CD18, 80-kDa receptor) allow for direct action by LPS of phagocytes, B and T lymphocytes, and other cells. Numerous other acute-phase proteins are produced in the liver, including C-reactive protein, complement components, fibrinogen, enzyme inhibitors, and anti-inflammatory proteins. Similar responses may be stimulated by subtoxic doses of LPS or by detoxified LPS, which manifest in endotoxin tolerance. Tolerant animals and man show increased resistance to LPS, to infections, and to various noxious insults. Infection and various forms of tissue injury are also capable of causing APR. There is much evidence to indicate that APR, which manifests in febrile illness, is an efficient host defense reaction. It is an emergency response in cases where specific immunity fails to protect the host. Therefore, the neuroimmunoregulatory network converts the immune system to a less specific, but rapid and more efficient response, APR. The hypothesis is presented that intestinal LPS serves to amplify the APR in response to various insults, which contribute to host defense, regeneration, and healing.

Exclusion of bioactive contaminations in Streptococcus pyogenes erythrogenic toxin A preparations by recombinant expression in Escherichia coli

The streptococcal erythrogenic exotoxin A (SPEA) belongs to the family of bacterial superantigens and has been implicated in the pathogenesis of a toxic shock-like syndrome and scarlet fever. Concerning its biological activity, mainly T-cell-stimulatory properties, conflicting data exist. In this study, we show that most of the SPEA preparations used so far contain biologically active contaminations. Natural SPEA from the culture supernatant of Streptococcus pyogenes NY-5 and recombinant SPEA purified from the culture filtrate of S. sanguis are strongly contaminated with DNases. We show that natural SPEA induces more tumor necrosis factor alpha (TNF-alpha) than recombinant SPEA, but we also show that DNases are able to induce TNF-alpha. In commercial SPEA preparations, we identified a highly active protease, which was shown not to be SPEB. To exclude these contaminations, we overexpressed SPEA cloned in the effective high-level expression vector pIN-III-ompA2 in Escherichia coli. The expressed SPEA shows the same amino acid composition as natural SPEA, whereas functional studies reported so far were carried out with toxins containing an incorrect amino terminus. We describe the rapid purification of lipopolysaccharide-, DNase-, and protease-free SPEA in two steps from the host's periplasm and its structural characterization by circular dichroism. Our results represent for the first time the production in E. coli of recombinant SPEA with the authentic N-terminal sequence and a proven superantigenic activity. Collectively, our results indicate that immunological studies of superantigens require highly purified substances free of biologically active contaminations.

Microbial superantigens stimulate T cells by the superantigen bridge and independently by a cytokine pathway

Superantigens cross-link the MHC II molecule on accessory cells with the Vbeta region of the T cell receptor (TCR). In this study, we compared the capacity of established superantigens for inducing cytokine release. The experimental protocol was generated to answer the question whether all superantigen effects are transmitted by the MHC/TCR cross-linkage and induce mainly a T cell response. We found that TSST-1, ExFTA, and SEC3 differed from all other superantigens tested because they stimulated a stronger monokine release. T cell proliferation after challenge with these superantigens was mainly mediated by a cytokine pathway and not by the cross-linkage of MHC and TCR. For the other superantigens, we were able to demonstrate that major immunomodulatory effect is mediated by the superantigen bridge. With the exception of these three superantigens, the proliferative response of superantigens correlated with their Vbeta specificity. Interleukin-1 (IL-1) and IL-6 were induced in monocytes by all superantigens, whereas tumor necrosis factor-alpha (TNF-alpha) was induced in T cells and by some superantigens, also in monocytes. IL-2 was always induced by the superantigen bridge, whereas interferon-gamma (IFN-gamma) was also induced indirectly by monokines. Collectively, our results indicate that not all superantigens are suitable for investigating superantigen-specific effects, as they show indirect (mitogenic) side effects. Observations for an individual superantigen are, therefore, not transferable to all other superantigens.

Common themes in microbial pathogenicity revisited

Bacterial pathogens employ a number of genetic strategies to cause infection and, occasionally, disease in their hosts. Many of these virulence factors and their regulatory elements can be divided into a smaller number of groups based on the conservation of similar mechanisms. These common themes are found throughout bacterial virulence factors. For example, there are only a few general types of toxins, despite a large number of host targets. Similarly, there are only a few conserved ways to build the bacterial pilus and nonpilus adhesins used by pathogens to adhere to host substrates. Bacterial entry into host cells (invasion) is a complex mechanism. However, several common invasion themes exist in diverse microorganisms. Similarly, once inside a host cell, pathogens have a limited number of ways to ensure their survival, whether remaining within a host vacuole or by escaping into the cytoplasm. Avoidance of the host immune defenses is key to the success of a pathogen. Several common themes again are employed, including antigenic variation, camouflage by binding host molecules, and enzymatic degradation of host immune components. Most virulence factors are found on the bacterial surface or secreted into their immediate environment, yet virulence factors operate through a relatively small number of microbial secretion systems. The expression of bacterial pathogenicity is dependent upon complex regulatory circuits. However, pathogens use only a small number of biochemical families to express distinct functional factors at the appropriate time that causes infection. Finally, virulence factors maintained on mobile genetic elements and pathogenicity islands ensure that new strains of pathogens evolve constantly. Comprehension of these common themes in microbial pathogenicity is critical to the understanding and study of bacterial virulence mechanisms and to the development of new "anti-virulence" agents, which are so desperately needed to replace antibiotics.

Analysis of human T cell responses to group A streptococci using fractionated Streptococcus pyogenes proteins

Cell extract and spent culture supernatant proteins from Streptococcus pyogenes Manfredo strain (type M5) were each separated to give 22 narrow range molecular weight fractions by blot-elution from SDS-polyacrylamide gels. Eluted samples and unfractionated proteins were screened for T cell stimulatory activity using human peripheral blood mononuclear cells (PBMC) from healthy adults in proliferation assays. Responses were measured in 4- and 7d cultures. Responses to a wide range of cell extract proteins were revealed by fractionation, the degree of response to each fraction varying between donors. Unfractionated culture supernatant proteins elicited proliferative responses by PBMC from all individuals examined. Responses to culture supernatant fractions containing 25-33 kDa proteins could be attributed to known superantigens. Furthermore, samples from culture supernatants containing higher molecular weight fractions (> 45 kDa) elicited responses in 50% of donors in 7d cultures, suggesting that these fractions contained common recall antigens. The efficacy of using electroeluted samples to identify T lymphocyte stimulatory proteins was confirmed by demonstrating that a known superantigen of S. pyogenes Manfredo strain, streptococcal pyrogenic exotoxin C (SPEC), could be fractionated successfully using this method and its activity recovered. Our results show that human T cell responses to group A streptococci involve a remarkably wide range of both cell-associated and released streptococcal proteins.