Molecular features of the cytolytic pore-forming bacterial protein toxins

The repertoire of the cytolytic pore-forming protein toxins (PFT) comprises 81 identified members. The essential feature of these cytolysins is their capacity to provoke the formation of hydrophilic pores in the cytoplasmic membranes of target eukaryotic cells. This process results from the binding of the proteins on the cell surface, followed by their oligomerization which leads to the insertion of the oligomers into the membrane and formation of protein-lined channels. It impairs the osmotic balance of the cell and causes cytolysis. In this review the molecular aspects of a number of important PFT and their respective encoding structural genes will be briefly described.

Superantigen bacterial toxins: state of the art

Superantigens (SAgs) are viral and bacterial proteins exhibiting a highly potent polyclonal lymphocyte-proliferating activity for CD4(+), CD8(+) and sometimes gammadelta(+) T cells of human and (or) various animal species. Unlike conventional antigens, SAgs bind as unprocessed proteins to invariant regions of major histocompatibility complex (MHC) class II molecules on the surface of antigen-presenting cells (APCs) and to particular motifs of the variable region of the beta chain (Vbeta) of T-cell receptor (TcR) outside the antigen-binding groove. As a consequence, SAgs stimulate at nano-to picogram concentrations up to 10 to 30% of host T-cell repertoire while only one in 10(5)-10(6) T cells (0.01-0.0001%) are activated upon conventional antigenic peptide binding to TcR. SAg activation of an unusually high percentage of T lymphocytes initiates massive release of pro-inflammatory and other cytokines which play a pivotal role in the pathogenesis of the diseases provoked by SAg-producing microorganisms. We briefly describe in this review the molecular and biological properties of the bacterial superantigen toxins and mitogens identified in the past decade.

Pyrogenicity and cytokine-inducing properties of Streptococcus pyogenes superantigens: comparative study of streptococcal mitogenic exotoxin Z and pyrogenic exotoxin A

Streptococcal mitogenic exotoxin Z (SMEZ), a superantigen derived from Streptococcus pyogenes, provoked expansion of human lymphocytes expressing the Vbeta 2, 4, 7 and 8 motifs of T-cell receptor. SMEZ was pyrogenic in rabbits and stimulated the expression of the T-cell activation markers CD69 and cutaneous lymphocyte-associated antigen. A variety of cytokines was released by human mononuclear leukocytes stimulated with SMEZ, which was 10-fold more active than streptococcal pyrogenic exotoxin A. Th2-derived cytokines were elicited only by superantigens and not by streptococcal cells.

Cholesterol-binding cytolytic protein toxins

Cholesterol-binding cytolysins (CBCs) are a large family of 50- to 60-kDa single-chain proteins produced by 23 taxonomically different species of Gram-positive bacteria from the genera Streptococcus, Bacillus, Clostridium, Listeria and Arcanobacterium. Apart pneumolysin, which is an intracytoplasmic toxin, all the other toxins are secreted in the extracellular medium. Among the species producing CBCs, only L. monocytogenes and L. ivanovii are intracellular pathogens which grow and release their toxins in the phagocytic cells of the host. CBCs are lethal to animals and highly lytic toward eukaryotic cells, including erythrocytes. Their lytic and lethal properties are suppressed by sulfhydryl-group-blocking agents and reversibly restored by thiols or other reducing agents. These properties are irreversibly abrogated by very low concentrations of cholesterol and other 3beta-hydroxysterols. Membrane cholesterol is thought to be the toxin-binding site at the surface of eukaryotic cells. Toxins molecules bind as monomers to the membrane surface with subsequent oligomerization into arc-and ring-shaped structures surrounding large pores generated by this process. Thirteen structural genes of the toxins (all chromosomal) have been cloned and sequenced to date. The deduced primary structure of the proteins shows obvious sequence homology particularly in the C-terminal part and a characteristic common consensus sequence containing a unique Cys residue (ECTGLAWEWWR) near the C-terminus of the molecules (except pyolysin and intermedilysin). However, another Cys residue outside this undecapeptide and closer to the C-terminus occurs in ivanolysin. Genetic replacement of the Cys residue in the consensus undecapeptide by certain amino acids demonstrated that this residue was not essential for toxin function. Other residues in the undecapeptide have been mutagenized, particularly the Trp residues. One of these Trp appeared critical for lytic activity. The recent elucidation of the 3-D structure of perfringolysin O provided interesting information on the structure-activity relationship. The molecule was divided into four domains. Three domains are arranged in a row, giving an elongated shape. Domain 3 is covalently connected to the N-terminal domain 1 and packed laterally against domain 2. Membrane interaction of the monomer appears to be mediated by domain 4, while, oligomerization involves several sites scattered throughout the sequence. The Trp-rich region around the conserved Cys residue within domain 4 is assumed to conformationally adapt to cholesterol, and domain 3 is envisaged to move across the "hinge" by which it is connected to domain 1.

Streptococcal pyrogenic exotoxin A (SPE A) superantigen induced production of hematopoietic cytokines, IL-12 and IL-13 by human peripheral blood mononuclear cells

A quantitative and kinetic study of the release of the hematopoietic cytokines IL-3, IL-5 and GM-CSF, the immunoregulatory cytokine IL-12 heterodimer (and its p40 subunit) and IL-13 by human peripheral blood mononuclear cells (PBMC) stimulated in vitro with the superantigen streptococcal pyrogenic (erythrogenic) exotoxin A (SPE A) from Streptococcus pyogenes is reported. PBMC were stimulated in parallel with heat-killed group A streptococcal cells, E. coli lipopolysaccharide (LPS) and with concanavalin A (Con A) in certain experiments for comparative purposes. The cytokines were assayed in the supernatant fluids by ELISA. IL-13 expression was also determined by a quantitative competitive PCR. IL-3, IL-5, GM-CSF, IL-12 p40, IL-12 heterodimer and IL-13 expression was induced by SPE A in a time- and dose-dependent manner in rather substantial amounts except the IL-12 heterodimer, which was released in small quantities. In contrast to SPE A, IL-3, IL-5 and IL-13 were not or poorly elicited by streptococcal cells or LPS whereas these two stimulants induced relatively high amounts of GM-CSF. Interestingly, both IL-12 p40 and IL-12 heterodimer were released in much higher amounts by streptococcal cells. Con A induced IL-3, IL-5, GM-CSF and IL-13 production in amounts comparable to those elicited by SPE A. The possible pathophysiological relevance of the elicitation by SPE A and streptococcal cells of these cytokines is discussed.

Streptococcal pyrogenic exotoxin A, streptolysin O, exoenzymes, serotype and biotype profiles of Streptococcus pyogenes isolates from patients with toxic shock syndrome and other severe infections

The determination of protein M and T serotypes, biotypes and pyrogenic (erythrogenic) exotoxin A (SPE A), streptolysin O (SLO), streptokinase (SK), hyaluronidase (HA) and cysteine proteinase release by 212 S. pyogenes isolates from patients with severe invasive group A streptococcal (GAS) infections, among them 74 cases of streptococcal toxic shock syndrome (STSS) has been investigated. M1 or M3 serotypes were expressed by 25% of the isolates (53/212), whereas 59% (125/212) belonged to 15 other different serotypes and 16% (34/212) were untypeable. Of the 74 isolates from STSS patients, 42% (31/74) expressed M1 and to a lesser extent M3 serotypes versus 19% of the non STSS isolates (26/138). Among the ten different biotypes known, biotypes 1 and 3 were prevalent, particularly the former in the case of STSS isolates. SPE A was detectably produced by about 25% (54/212) of the strains. However, as high as 40.5% of the STSS isolates (30/74) versus 17.4% of non STSS isolates (24/138) released SPE A. Moreover, 67% of the SPE A producing strains were of serotype M1 or M3. SK and HA were released by 71% and 10% of the isolates respectively. All strains released SLO (4 to 256 HU/ml) and 85% cysteine proteinase. No relationship between toxin or enzyme titer and the type of disease or clinical origin of the strains was found. Culture supernatants of all isolates showed moderate to high lymphocyte transforming activity with index values ranging from 14.5 to 50.3 including those strains which did not release detectable amounts of SPE A suggesting that SPE C and other mitogenic factor(s) are released by the isolates investigated.

Human pro- and anti-inflammatory cytokine patterns induced by Streptococcus pyogenes erythrogenic (pyrogenic) exotoxin A and C superantigens

The superantigenic streptococcal erythrogenic toxins A and C (ETA/SPEA and ETC/SPEC) elicit the production by human peripheral blood mononuclear cells of substantial amounts of Th1-derived cytokines (interleukin-2 [IL-2] and gamma interferon) as well as anti-inflammatory cytokines (IL-10 and IL-1 receptor antagonist). In contrast, very low levels of IL-4 and no alpha interferon were induced. The production of these cytokines after stimulation with Streptococcus pyogenes heat-killed bacteria and lipopolysaccharide from gram negative bacteria differed qualitatively and quantitatively from that elicited by the superantigens.

A survey of bacterial toxins involved in food poisoning: a suggestion for bacterial food poisoning toxin nomenclature

There is at present no accepted nomenclature for bacterial protein toxins, although there have been several attempts at dividing them into groups by their mode of action. In this paper we will not try to describe all known bacterial protein toxins, but concentrate on the toxins involved in food poisoning. Although most of these toxins are enterotoxins (protein exotoxins with the site of action on the mucosal cells of the intestinal tract) there are also other toxins involved in food poisoning, like the neurotoxins. In Table 1 the most important food pathogens in Europe are listed. For most, but not all, of these food pathogens, toxins are virulence factors. Generally, we divide food poisoning into infections and intoxications, where Salmonella spp. and Shigella spp. are typical examples of infections and Clostridium botulinum and Staphylococcus aureus for intoxications. We consider it better to make four different groups of food pathogenic bacteria, according to Table 2. Today the first three groups are all defined as infections, although for both group 2 and 3 the bacterium itself does not harm the host directly. The bacterium in such locations is like an 'enterotoxin factory'. The bacteria belonging to group 3 do not even interact with the epithelial cells in the intestine, while the bacteria of group 2 must colonise the epithelial cells prior to enterotoxin production.

The sulphydryl-activated cytolysin and a sphingomyelinase C are the major membrane-damaging factors involved in cooperative (CAMP-like) haemolysis of Listeria spp

The negative mutant approach was used in this study to identify listerial cytolytic factors involved in cooperative haemolysis (CAMP-like phenomenon) with Staphylococcus aureus and Rhodococcus equi. A Listeria monocytogenes non-haemolytic mutant specifically impaired in listeriolysin O (LLO) production gave no CAMP reaction with S. aureus, and was virtually CAMP-negative with R. equi, indicating that the listerial sulphydryl-activated toxin played a major role in cooperative haemolysis. This was confirmed by direct evidence using purified LLO and alveolysin (from Bacillus alvei) in diffusion CAMP assays. To our knowledge, this is the first evidence of involvement of a sulphydryl-activated toxin in cooperative lytic processes. Phosphatidylcholine- and phosphatidylinositol-specific phospholipases C from L. monocytogenes did not seem to significantly contribute to cooperative haemolysis, as the corresponding mutants displayed wild-type CAMP reactions. In contrast, the sphingomyelinase C from Listeria iva-novii was the cytolytic factor responsible for the characteristic shovel-shaped CAMP reaction shown by this listerial species with R. equi. Possible mechanisms of lytic cooperation are discussed.

Activation of human effector cells by different bacterial toxins (leukocidin, alveolysin, and erythrogenic toxin A): generation of interleukin-8

We analyzed the transcription and release of interleukin-8 (IL-8) from human polymorphonuclear granulocytes (PMNs) and a lymphocyte-monocyte-basophil (LMB) cell population stimulated for different time periods (30 min to 16 h) with pore-forming bacterial toxins (Panton-Valentine leukocidin [Luk-PV] and alveolysin [Alv]) as well as with the erythrogenic toxin A (ETA) as a superantigen. At high toxin concentrations (500 ng/10(7) cells), Luk-PV and Alv led to a decreased IL-8 generation from LMBs within the first 30 min; with PMNs, a slight increase in IL-8 release was observed. Under these conditions, stimulation with ETA did not lead to an altered cellular IL-8 release. At lower concentrations (5 and 0.5 ng/10(7) cells), all three toxins led to a continuous increase (over 16 h) in IL-8 release and IL-8 mRNA expression of PMNs and LMBs. Preincubation of the cells with the protein tyrosine kinase inhibitors lavendustin A and tyrphostin 25 led to a reduction of the toxin-mediated effects on IL-8 release and IL-8 mRNA expression when Luk-PV and Alv were used as stimuli. In contrast, IL-8 synthesis in cells which were stimulated with ETA was not influenced by protein tyrosine kinase inhibition. From our data, one may suggest that multiple pathways for IL-8 production are operative in human leukocytes.

Comparative study of cytokine release by human peripheral blood mononuclear cells stimulated with Streptococcus pyogenes superantigenic erythrogenic toxins, heat-killed streptococci, and lipopolysaccharide

The differences between toxic or septic shocks in humans during infections by streptococci and gram-negative bacteria remain to be fully characterized. For this purpose, a quantitative study of the cytokine-inducing capacity of Streptococcus pyogenes erythrogenic (pyrogenic) exotoxins (ETs) A and C, heat-killed S. pyogenes bacteria, and Neisseria meningitidis endotoxin (lipopolysaccharide [LPS]) on human peripheral blood mononuclear cells (PBMC) and monocytes has been undertaken. The levels of interleukin-1 alpha (IL-1 alpha), IL-1 beta, IL-6, IL-8, tumor necrosis factor alpha (TNF-alpha), and TNF-beta induced by these bacterial products and bacteria were determined by using cell supernatants. The capacity of ETs to elicit the monocyte-derived cytokines IL-1 alpha, IL-1 beta, IL-6, and TNF-alpha was found to depend on the presence of T lymphocytes, because of the failure of purified monocytes to produce significant amounts of these cytokines in response to ETs. PMBC elicited large amounts of these cytokines, as well as IL-8 and TNF-beta, with an optimal release after 48 to 96 h. The most abundant cytokine produced in response to ETA was IL-8. In contrast to the superantigens ETA and ETC, LPS and heat-killed streptococci stimulated the production of significant amounts of IL-1 alpha, IL-1 beta, IL-6, and TNF-alpha, with optimal production after 24 to 48 h in monocytes, indicating no significant involvement of T cells in the process. ETs, but neither LPS nor streptococci, were potent inducers of TNF-beta in PBMC. This study outlines the differences in the pathophysiological features of shock evoked by endotoxins and superantigens during infection by gram-negative bacteria and group A streptococci, respectively. The production of TNF-alpha was a common pathway for LPS, streptococcal cells, and ETs, although cell requirements and kinetics of cytokine release were different.

Bifactorial versus monofactorial molecular status of Staphylococcus aureus gamma-toxin

The extracellular Staphylococcus aureus gamma-toxin (hemolysin) released by the Smith 5R strain has been purified (M(r) 38 kDa, pl 9.55). We established that this cytolysin is a single polypeptide fully lytic on rabbit erythrocytes. In contrast, this toxin alone was unable to lyse other cells and was required to act jointly with an accessory 58 kDa protein released by the same strain. This protein, named sensitizing protein (SP), was required in order to damage the cytoplasmic membranes of other red blood cells including human erythrocytes as well as that of other eukaryotic cells (Jurkat and Hep-2). The lytic process can be referred to as conditional synergistic or cooperative lysis. gamma-toxin, and SP were also found able to disrupt phospholipid/cholesterol containing liposomes. We demonstrated that a minor membrane phospholipid, phosphatidylinositol, is crucial for gamma-toxin binding to cells and/or channel formation through membrane lipid bilayer.

Staphylococcus aureus toxic shock syndrome toxin 1 and Streptococcus pyogenes erythrogenic toxin A modulate inflammatory mediator release from human neutrophils

We studied the influence of staphylococcal toxic shock syndrome toxin 1 and streptococcal erythrogenic (pyrogenic) toxin A (ETA) on intact and digitonin-permeabilized human polymorphonuclear granulocytes (PMNs). As was shown by reversed-phase high-performance liquid chromatography analysis, toxic shock syndrome toxin 1 or ETA alone, in the absence of any additional stimulus, did not induce the generation of the chemoattractant leukotriene B4 (LTB4) from PMNs in a wide range of concentrations. In addition, pretreatment of intact PMNs with either toxin potentiated formyl-methionyl-leucyl-phenylalanine (fMLP)- and washed Staphylococcus aureus cell-induced generation of LTB4 in a time- and dose-dependent manner. This increase included LTB4 as well as its inactive omega-oxidated compounds. Further studies revealed evidence that toxin exposure was accompanied by enhanced cellular receptor expression for fMLP as well as for LTB4. The intrinsic GTPase activity of membrane fractions was modulated by both toxins. Short-term incubation with ETA increased the GTPase activity of PMNs up to 141%. Inhibitory effects were obtained when GTP-binding protein functions were stimulated with sodium fluoride (NaF). In addition, specific binding of Gpp(NH)p to GTP-binding protein was inhibited by both toxins during the first 10 min of incubation and was restored at later times of incubation. Our data therefore suggest that both toxins significantly affect the signal transduction pathways of human PMNs, which results in immunomodulatory functions.

Erythrogenic toxins A, B and C: occurrence of the genes and exotoxin formation from clinical Streptococcus pyogenes strains associated with streptococcal toxic shock-like syndrome

We report the study of 53 clinical isolates of group A streptococci, all from patients with streptococcal toxic shock-like syndrome. The strains were analysed for the occurrence of the genes of erythrogenic toxins (pyrogenic exotoxins) types A, B and C and in vitro production of these toxins. In contrast to reports indicating that 85% of the toxic shock-like syndrome-associated isolates contained the erythrogenic toxin A gene, only 58.5% of our strains harboured this gene. The erythrogenic toxin C gene was detected in 22.6% of the isolates. Erythrogenic toxin A and erythrogenic toxin B were produced by 68.7% and 58.3% of the strains containing either gene. For all group A streptococci, irrespective of clinical association, the erythrogenic toxin B gene was detected in all the isolates tested. Thus, it is difficult to define a specific role for erythrogenic toxin B in toxic shock-like syndrome as there was no clear correlation between this disease and the presence of toxin genes. Our results suggest the existence of other pathogenic factor(s) produced by group A streptococci which may stimulate human peripheral T lymphocytes in a manner similar to that of erythrogenic toxins, thus explaining different observations in previous epidemiological genetic studies.

Cytokine production by murine cells activated by erythrogenic toxin type A superantigen of Streptococcus pyogenes

The mode of pathogenic action of the Steptococcus pyogenes superantigen erythrogenic toxin type A (ETA) in causing toxic shock-like syndrome in humans is thought to be mediated by massive release of cytokines by patients immune cells. The cytokine-inducing capacity of ETA as an extracellular protein was compared with that of lipopolysaccharide (LPS), a component of cell wall of gram-negative bacteria. Peritoneal macrophages and splenocytes of BALB/c and C3H/HeJ mice were stimulated by ETA and LPS. Tumor necrosis factor (TNF), interleukin 3 (IL-3) and interleukin 6 (IL-6) activities in the supernatants of stimulated cells were evaluated. In contrast to LPS, ETA induced only low amounts of IL-6 and no detectable TNF activities in peritoneal macrophage supernatants. ETA-triggered BALB/c and C3H/HeJ splenocytes produced great amounts of IL-6. ETA triggered the production of IL-3 by both mice strains splenocytes in a dose dependent manner. The amounts of IL-3 in supernatants were comparable to those induced by concanavalin A. The simultaneous presence of ETA and LPS in macrophage and splenocyte cultures induced a slight enhancement above an additive value after 72-96 h. Challenge of BALB/c mice with ETA 6 h before the harvest of peritoneal macrophages led to an enhanced production of IL-6 upon stimulation with ETA as well as with LPS. Splenocytes of nude BALB/c mice did not produce IL-6 upon stimulation with ETA, whereas LPS-induced IL-6 production was similar in these mice and in their littermates. The pathogenic effect of ETA on host's immune cells could most likely be explained as a consequence of T cell activation. The results confirm also that LPS- and ETA-induced shock is mediated by different cell types.

The phosphoinositide pathway of lymphoid cells: labeling after permeabilization by alveolysin, a bacterial sulfhydryl-activated cytolysin

An improved method allowing incorporation of [3H]myo-inositol into the phosphoinositide pool of human lymphoid cells is described. The procedure devised involves cell permeabilization with a thiol-activated membranolytic toxin, alveolysin, and optimization of the phosphoinositide labeling and extraction. In these conditions 4 to 10% of the added [3H]myo-inositol is found intracellularly and half of this amount (2-5%) is incorporated into the phosphoinositide pool in only 1 h as compared to the classical 0.2 to 0.3% incorporation obtained after 10 to 20 h. The integrity of coupling between receptors and phospholipase C was assessed by the inositol phosphate production after cell stimulation by various agonists.

One-step purification of the serotonin transporter located at the human platelet plasma membrane

A 68-kDa glycoprotein bearing the biological activity of the plasma membrane serotonin (5-hydroxytryptamine, 5-HT) transporter has been purified from human blood platelets, a classical cell model for the study of 5-HT uptake. After treatment of the whole platelet population or its plasma membrane fraction by sulfhydryl-dependent bacterial protein toxins or by digitonin, purification was reproducibly obtained by a one-step affinity chromatography using two different columns with 5-HT or 6-fluorotryptamine as ligands and elution by 5-HT or Na(+)-free buffer. The purified fraction migrated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis as a single band with an apparent molecular mass of 68 kDa and exhibited an apparent isoelectric point of 5.6-6.2. Two sialic acid residues were detected in the purified material. The purified glycoprotein bound the 5-HT uptake blocker [3H]paroxetine with a Kd (0.25 nM) similar to the one observed for intact human platelets. It also bound [3H] 5-HT but neither [3H]hydroxytetrabenazine nor [3H] ouabain, the respective markers of the granular monoamine transporter and of the Na+,K(+)-ATPase associated to the plasma membrane 5-HT transporter. 5-HT derivatives and 5-HT uptake inhibitors exhibited similar Ki values for 5-HT uptake and paroxetine binding in intact human platelets and in the purified glycoprotein. Under laser UV irradiation, 40% of this purified glycoprotein could be labeled by either [3H]paroxetine or [3H]cyanoimipramine. No labeling was detected with either [3H] gamma-aminobutyric acid or [3H]GBR 12783, the respective markers of gamma-aminobutyric acid and dopamine carriers. The purified 68-kDa protein is therefore likely to correspond at least to the binding domain of the 5-HT transporter located at the human platelet plasma membrane.

Reduced virulence of a Listeria monocytogenes phospholipase-deficient mutant obtained by transposon insertion into the zinc metalloprotease gene

A phospholipase-deficient mutant, termed JL762, was obtained from a virulent strain of Listeria monocytogenes by screening a bank of 5,000 Tn1545 transposon-induced mutants on 2.5% egg yolk brain heart infusion agar. As previously shown (J. Mengaud, C. Geoffroy, and P. Cossart, Infect. Immun. 59:1043-1049, 1991), the transposon insertion took place inside the gene mpl, which encodes a zinc metalloprotease. By Western blot (immunoblot) analysis, we showed that loss of phospholipase activity was associated with loss of a 29-kDa zinc-dependent phosphatidylcholine-phospholipase C (PC-PLC) in culture supernatant of JL762 and of EGD-SmR incubated with ion chelator. As the parental strain, JL762 still produced in supernatants approximately 33-kDa proteins antigenically closely related to the 29-kDa PC-PLC. These results strongly suggest that the zinc metalloprotease of L. monocytogenes might play a role in the maturation of the 29-kDa PC-PLC. Although the uptake and the intracellular growth of bacteria were not affected in vitro, we found that the virulence of mutant JL762 was strongly impaired in the mouse.

Bacterial toxins induce heat shock proteins in human neutrophils

We studied the influence of different bacterial toxins (alveolysin; toxic shock syndrome toxin 1, TSST-1 and erythrogenic toxin A, ETA) on the expression of heat shock proteins (hsps) in isolated human polymorphonuclear granulocytes (PMNs). As was shown by Western blotting (anti-hsp72) ETA and TSST-1 were potent inducers of hsps at low toxin concentrations (10 ng/ml). Alveolysin led to the expression of hsps at hemolytic concentrations (1 HU; 700 ng/ml) whereas at subhemolytic concentrations (7 ng/ml) no heat shock response was observed. The induction of heat shock proteins was also accompanied by increased mRNA levels for hsp70 as was determined by PCR-analysis.

Antibody levels and in vitro lymphoproliferative responses to Streptococcus pyogenes erythrogenic toxin A and mitogen of patients with rheumatic fever

We investigated the in vitro lymphoproliferative responses to a streptococcal mitogen and erythrogenic toxin A of children with acute rheumatic fever (ARF) and patients with chronic rheumatic heart disease (CRHD). Antibody levels to the streptococcal products were also analyzed in the sera of those with ARF or chronic rheumatic heart disease as well as in the sera of children with streptococcal pharyngitis or poststreptococcal glomerulonephritis. Our results demonstrated that the individuals had depressed lymphoproliferative responses during the active stage of rheumatic fever. The depressed responses were not found either to be induced by time-sensitive mitogen-specific suppressor cells or to be related to a dose-response phenomenon. On the other hand, antibody levels to the extracellular mitogens were significantly elevated in the sera of children with ARF compared with the levels in the rest of the groups. The hyperresponsiveness noted among children with ARF was found to be at a quantitative level and was not due to recognition of more epitopes, as determined by Western blotting (immunoblotting). The profile of immune responsiveness in children with ARF to the streptococcal extracellular mitogens is discussed in relation to the pathogenesis of disease.

Purification and characterization of an extracellular 29-kilodalton phospholipase C from Listeria monocytogenes

We purified and characterized an extracellular phospholipase produced by Listeria monocytogenes. This enzyme was separated as a homogeneous protein of 29 kDa by chromatography on DEAE-52 cellulose and Bio-Gel P100 columns. It is a zinc-dependent phospholipase C (PLC) that is mainly active at pH 6 to 7 and expresses lecithinase activity and a weaker sphingomyelinase activity. The exoenzyme also hydrolyzed phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, and sphingomyelin but not phosphatidylinositol. It was distinct from the 36-kDa phosphatidylinositol PLC produced by L. monocytogenes and from the L. ivanovii sphingomyelinase. The pure protein expressed a weak, calcium-independent hemolytic activity and was not toxic in mice. Western immunoblot analysis using a rabbit immune serum raised against the enzyme showed that all virulent strains of L. monocytogenes tested produced in the culture supernatant a 29-kDa PLC. In contrast, no proteins antigenically related to the 29-kDa PLC were detected in supernatants of L. ivanovii, L. seeligeri, L. innocua, or L. welshimeri. The role in virulence of the 29-kDa PLC specifically produced by L. monocytogenes remains to be established.

The amphiphilic alpha-helix concept. Consequences on the structure of staphylococcal delta-toxin in solution and bound to lipids

Staphylococcal delta-toxin, a synthetic analogue and a fragment were studied in order to determine their structure in solution and bound in lipids. In solution, a self-association process is observed. Analytical ultracentrifuge and quasi-elastic light-scattering experiments suggest an isodesmic aggregation in the high concentration domain above 2 microM up to very large asymmetrical species. Decreasing concentrations below 2 microM of delta-toxin and the analogue allows dissociation, probably into monomers. The self-associated species are essentially alpha-helical (70%) with buried and highly immobilized Trp either at position 15 for natural delta-toxin or 16 for the analogue. At the lowest concentration studied, the alpha-helix content severely decreases down to 35% while Trp fluorescence shows that these residues are exposed to buffer. The fragment 11-26 is always monomeric and structureless. From all the data, a structural model of aggregated species is proposed with stacked antiparallel amphipathic rods. When bound to lipids, whatever their initial structure in solution, 26-residue long peptides mainly adopt an alpha-helix conformation (80%) while fragment 11-26 exhibits about 50% alpha-helix. The lipid-peptide interactions were quantitatively analysed. For fragment 11-26, a single-step mechanism fits the spectroscopic changes and defines a single monomeric bound structure. On the other hand, for the 26-residue-long analogue, multiple-step processes must occur. The data were analysed with a partition of tetramers into lipids followed by a partial dissociation. Finally, the affinity of fragment 11-26 severely decreases from micelles to fluid and gel-state bilayers. The partition coefficient of the delta-toxin analogue is higher than those of other more apolar peptides, such as melittin and alamethicin, correlating with Eisenberg's hydrophobic moments. It is therefore proposed that delta-toxin probably lies parallel to the surface, only penetrating weakly in lipids, depending on their packing.

Alveolysin, the thiol-activated toxin of Bacillus alvei, is homologous to listeriolysin O, perfringolysin O, pneumolysin, and streptolysin O and contains a single cysteine

The gene coding for alveolysin, the thiol-activated toxin produced by Bacillus alvei, has been cloned by means of an oligonucleotide based on the known N-terminal sequence of the secreted protein. The complete nucleotide sequence of the gene has been determined. The deduced amino acid sequence of alveolysin shows that alveolysin shares homologies with listeriolysin O, perfringolysin O, pneumolysin, and streptolysin O. Alveolysin, like the other members of the family, contains a single cysteine in the conserved peptide sequence ECTGLA WEWWR.

Effect of staphylococcal delta-toxin and bee venom peptide melittin on leukotriene induction and metabolism of human polymorphonuclear granulocytes

The abilities of delta-toxin from Staphylococcus aureus and melittin to induce and modulate the generation of leukotriene from human polymorphonuclear granulocytes (PMNs) were studied. Stimulation of PMNs with melittin (10 micrograms) induced leukotriene formation, whereas stimulation with delta-toxin did not. Preincubation of the PMNs with delta-toxin modulated the subsequent generation of leukotriene from PMNs induced by Ca ionophore A23187 or opsonized zymosan. The generation of leukotriene B4 (LTB4), induced by the Ca ionophore A23187, was increased when the PMNs were preincubated with delta-toxin for 5 min. When opsonized zymosan was used as a secondary stimulus to activate the delta-toxin-pretreated PMNs, LTB4 generation decreased. In contrast, melittin showed no significant modulatory effect on the generation of leukotriene from PMNs. In addition, preincubation of PMNs with delta-toxin inhibited the conversion of LTB4 to omega-oxidation products. Our data suggest that peptides with similar structures, e.g., delta-toxin and melittin, induce and modify leukotriene generation in different manners.

Effect of Staphylococcus aureus delta-toxin on human granulocyte functions and platelet-activating-factor metabolism

The production of delta-toxin is supposed to be responsible for various pathophysiological effects during infection with Staphylococcus aureus. We compared the effects of delta-toxin with the structurally related bee venom toxin melittin on granulocyte functions and inflammatory mediator release. Delta-toxin and melittin induced a rapid Ca2+ influx, as was shown by fluorescence detection. Furthermore, oxygen radical production, as determined by luminol-enhanced chemiluminescence, was triggered by delta-toxin (0.15 to 15 micrograms/ml), whereas melittin showed only marginal effects. Release of lysozyme and beta-glucuronidase was observed only at high concentrations of 15 micrograms of melittin and delta-toxin per ml. Preincubation (15 min) of neutrophils with both toxins resulted in the formation of 3H-platelet-activating factor (3H-PAF) from 3H-lyso-PAF. After 5 min of incubation, the exogenously added lyso-PAF was converted to PAF (delta-toxin, 80 +/- 2%; melittin, 27 +/- 12% of total radioactivity; n = 3, mean +/- standard error of the mean) and 1-O-alkyl-2-acyl-glycerophosphorylcholine (alkyl-acyl-GPC) (corresponding values, 20 +/- 3% and 51 +/- 14% of total radioactivity). The newly generated PAF was rapidly metabolized to lyso-PAF and alkyl-acyl-GPC during the subsequent incubation period of 60 min. In the absence of any toxin, no formation of PAF from lyso-PAF was observed. Further studies indicated that the metabolism of PAF into lyso-PAF and alkyl-acyl-GPC was inhibited in the presence of delta-toxin. Melittin had no significant effects on PAF metabolism. Neither delta-toxin nor melittin modulated the uptake of PAF and lyso-PAF significantly. Our data provide evidence that delta-toxin has an effect on the activity of neutrophil granulocytes with regard to its proinflammatory capacity.

Reaction of staphylococcal alpha-toxin with peptide-induced antibodies

Two peptides representing separate 13-amino-acid sequences of staphylococcal alpha-toxin have been synthesized and acrylamide gel-purified alpha-toxin monomer and hexamer forms have been prepared and used to produce antisera in rabbits. We report here that each synthetic peptide, P-I and P-II, induces the formation of a specific precipitating antiserum. Moreover, these sera also react with the toxin monomer and sometimes with the hexamer, indicating that each peptide has more than one epitope. The purified toxin monomer can induce antibodies to fragments of toxin but is significantly less potent than the hexamer in inducing antibodies to the toxin monomer and almost not effective in inducing a response to the toxin hexamer. The purified toxin hexamer induces responses that are almost the reciprocals of the monomers, with the antihexamer and -monomer responses dominating and almost no responses to fragments of toxin being induced. These responses are interpreted in terms of the stability of the toxin hexamer to proteolytic degradation, compared with the relative sensitivity of the monomer to proteases. In assays of toxin-neutralization activity, only those sera containing antihexamer antibodies can block toxin hemolytic activity. This is true for both peptide- and toxin-induced antisera. The basis for this apparent association between toxin-neutralizing potency and antihexamer reactivity is being studied. Peptide P-I contains the uniquely reactive tyrosine residue and may be involved in monomer-to-monomer associations required to form hexamers. Peptide P-II is near the carboxyl terminus of alpha-toxin and may be involved in the binding of toxin to membranes. In a study of the ability of each peptide to inhibit the rate of hexamer formation induced by membrane lipoprotein, peptide P-I (as expected) proves to be more efficient than peptide P-II. Finally, one rabbit immunized with the toxin hexamer produces antibodies to peptides P-I and P-II. This finding suggests that the two synthetic peptides selected for study are relevant to the in vivo immunoprocessing of staphylococcal alpha-toxin.

Modulation of leukotriene generation by pertussis toxin

The purpose of our study was to characterize the properties of the interaction of pertussis toxin with human polymorphonuclear leukocytes for the modulation of leukotriene generation and metabolism. The cells were stimulated with either the Ca ionophore A23187, opsonized zymosan, or the bacterial peptide formyl-methionyl-leucyl phenylalanine. Incubation of the cells with pertussis toxin led to a rapid inhibition of LTB4 generation when formyl-methionyl-leucyl phenylalanine was used as the stimulus, whereas there was no effect with the Ca ionophore and just a low effect with opsonized zymosan. The inhibition of leukotriene generation was dependent on the incubation time, temperature, and pertussis toxin concentration. The effect was not dependent on the presence of calcium. Incubation of the cells with guanosine 5'-O-(3-thiotriphosphate) the stable analog of GTP, led to a time-dependent increase in leukotriene generation induced by formyl-methionyl-leucyl phenylalanine which was abolished by the simultaneous addition of pertussis toxin. Our data suggest that the formyl-methionyl-leucyl phenylalanine-induced generation of leukotrienes is dependent on a GTP-binding protein. The participation of the various G proteins has yet to be elucidated.

Staphylococcal alpha-toxin: a study of membrane penetration and pore formation

Cell lysis by staphylococcal alpha-toxin, a potent virulence factor of most pathogenic strains of Staphylococcus aureus, follows a three-step sequence: binding of toxin to the membrane, leaking of ions caused by membrane injury, and rupturing of the membrane caused by osmotic swelling. The membrane injury step is composed of two separate events, membrane penetration and membrane perturbation. The membrane penetration event involves conversion of the soluble toxin monomer into an amphipathic molecule, which inserts into the lipid bilayer of the membrane. The membrane perturbation event involves association of the toxin monomers, in the plane of the membrane, to form hexameric transmembrane pores. In this study, we demonstrate that, in an asolectin liposome system, controlling the pH of the external buffer permits these two events to be temporally resolved. Using Controlled-Pore Glass bead-purified alpha-toxin, four events are measured as a function of pH: (a) release of potassium from prelabeled asolectin vesicles, (b) conversion of the toxin to a globally hydrophobic molecule, (c) binding of detergent by the toxin, and (d) labeling of the toxin with photoactivable, radiolabeled, hydrophobic probes. Two of these events, potassium release and conversion to a net hydrophobic state, are paired in that, for the event to occur, each requires a pH of 4.6 or less. In contrast, photolabeling with the membrane probes PC I and PC II (where PC represents phosphatidylcholine) is easily detectable at pH values as high as 5.0 and 6.0. These results demonstrate that, as the pH is lowered, two distinct changes in the physical properties of alpha-toxin occur. The first, which occurs under mild acidic conditions, converts the toxin from a water-soluble molecule into an amphipathic molecule. The second, requiring relatively more acidic conditions, converts the amphipathic toxin molecule into a globally hydrophobic molecule. Correlated with these physical changes in the alpha-toxin molecule is the acquisition of two new biological properties. The conversion of alpha-toxin into an amphipathic conformation correlates with the acquisition of the biological property of the reversible penetration into the bilayer of the asolectin liposome membrane, as evidenced by labeling with the photoactivable probes. At lower pH, the conversion of the toxin into a globally hydrophobic molecule correlates with the biological property of causing damage to the cell membrane, as measured by the release of internal potassium ions, presumably by the formation of transmembrane hexamer pores.(ABSTRACT TRUNCATED AT 400 WORDS)

Interaction of staphylococcal delta-toxin and synthetic analogues with erythrocytes and phospholipid vesicles. Biological and physical properties of the amphipathic peptides

Staphylococcal delta-toxin, a 26-residue amphiphilic peptide is lytic for cells and phospholipid vesicles and is assumed to insert as an amphipathic helix and oligomerize in membranes. For the first time, the relationship between these properties and toxin structure is investigated by means of eight synthetic peptides, one identical in sequence to the natural toxin, five 26-residue analogues and two shorter peptides corresponding to residues 1-11 and 11-26. These peptides were designed by the Edmundson wheel axial projection in order to maintain: (a) the hydrophilic/hydrophobic balance while rationalizing the sequence, (b) the alpha-helical configuration and (c) the common epitopic structure. The fluorescence of the single Trp residue was used to monitor the behaviour of the natural toxin and analogues. All 26-residue analogues were hemolytically active although to a lesser extent than natural toxin. The peptide of residues 11-26 bound lipids weakly and was hemolytic at high concentration. The peptide of residues 1-11 did not bind lipids and was hemolytically inactive. All peptides except the latter cross-reacted in immunoprecipitation tests with the natural toxin. The study of a 26-residue analogue by circular dichroism revealed an alpha-helical configuration in both the free and lipid-bound state. Changes in the fluorescence of the peptides in the presence of lipid micelles and bilayers varied according to the position of the reporter group. When bound to lipids, Trp5, Trp16 and the Fmoc-1 positions of the analogues became buried while Trp15 of the natural toxin and its synthetic replicate remained more exposed. All changes are rationalized by the proposal of an amphipathic helix whose hydrophobic face is embedded within the apolar core of bilayers while the hydrophilic and charged face remains more exposed to solvent.

Induction of murine hemopoietic growth factors by toxic shock syndrome toxin-1

Toxic shock syndrome toxin-1 (TSST-1), an extra-cellular 22 kDa single chain protein produced by most Staphylococcus aureus strains isolated from patients with toxic shock syndrome (TSS), induces modifications of blood cell values similar to those observed during TSS. We therefore analyzed the effects of TSST-1 on the proliferation and differentiation of murine granulocyte-macrophage progenitor cells (CFU-culture) and the eventual role of endotoxin in this response. TSST-1 had no direct effect on the proliferation of CFU-culture and was unable to influence the CSF-induced proliferation and differentiation of these progenitors. In contrast, TSST-1 was a potent inducer in spleen cell cultures of a factor with an ability to induce both colony formation by bone marrow cells and proliferation of an IL-3-dependent cell line. Nanogram amounts of TSST-1 were able to induce the release of CSF activity in spleen cell cultures from both normal and LPS-hyporesponsive mice. Cells from C3H/HeJ mice were as responsive as cells from C3H/He Pas mice. Furthermore, in spleen cell cultures from normal mice, TSST-1 and LPS did not act synergistically to induce CSF activity. Nanogram amounts of TSST-1 were also able to induce CSF activity in vivo but failed to induce IL-3 activity in the serum and organ-conditioned media from TSST-1-treated mice.

Induction of murine hemopoietic growth factors by toxic shock syndrome toxin-1

Toxic shock syndrome toxin-1 (TSST-1), an extra-cellular 22 kDa single chain protein produced by most Staphylococcus aureus strains isolated from patients with toxic shock syndrome (TSS), induces modifications of blood cell values similar to those observed during TSS. We therefore analyzed the effects of TSST-1 on the proliferation and differentiation of murine granulocyte-macrophage progenitor cells (CFU-culture) and the eventual role of endotoxin in this response. TSST-1 had no direct effect on the proliferation of CFU-culture and was unable to influence the CSF-induced proliferation and differentiation of these progenitors. In contrast, TSST-1 was a potent inducer in spleen cell cultures of a factor with an ability to induce both colony formation by bone marrow cells and proliferation of an IL-3-dependent cell line. Nanogram amounts of TSST-1 were able to induce the release of CSF activity in spleen cell cultures from both normal and LPS-hyporesponsive mice. Cells from C3H/HeJ mice were as responsive as cells from C3H/He Pas mice. Furthermore, in spleen cell cultures from normal mice, TSST-1 and LPS did not act synergistically to induce CSF activity. Nanogram amounts of TSST-1 were also able to induce CSF activity in vivo but failed to induce IL-3 activity in the serum and organ-conditioned media from TSST-1-treated mice.

Production of thiol-dependent haemolysins by Listeria monocytogenes and related species

Twenty-six strains belonging to the five main species of the genus Listeria were examined for production of thiol-dependent exotoxins. All strains of L. monocytogenes cultured in charcoal-treated broth secreted a haemolytic factor at a level ranging from 200 to 800 haemolytic units (HU) ml-1, except for the strain EGD (1500 HU ml-1) and the type strain CIP 82110T (10 HU ml-1). The haemolytic activity reached a maximum level by 8-10 h and then rapidly declined as soon as bacterial exponential growth ceased. The titres of haemolytic activity were markedly reduced when bacteria were grown in charcoal-untreated broth. The haemolytic factor produced by L. monocytogenes strains was characterized as listeriolysin O (Mr about 60,000), a member of the group of thiol-dependent exotoxins. Strains of Listeria ivanovii also produced high levels of thiol-dependent exotoxin (about 2500 HU ml-1), in both charcoal-treated and untreated broth. Small amounts of haemolytic factor (about 9-30 HU ml-1) were also produced by Listeria seeligeri in charcoal-treated broth. The haemolysin produced by L. seeligeri was identified for the first time as a thiol-dependent exotoxin of Mr about 60,000, antigenically related to listeriolysin O. As expected, we failed to detect thiol-dependent exotoxin in the two nonhaemolytic species, Listeria innocua and Listeria welshimeri.

Interaction of Clostridium perfringens delta toxin with erythrocyte and liposome membranes and relation with the specific binding to the ganglioside GM2

The specific interaction of the cytolytic Clostridium perfringens delta toxin with membrane GM2 was indicated by: (i) characterization of this glycolipid in the membrane of sheep and goat erythrocytes, which are lysed by the toxin, whereas GM2 was undetectable in insensitive rabbit erythrocytes, (ii) demonstration of 125I-toxin binding to GM2, by autoradiography, following incubation with thin-layer chromatograms containing separated neuroblastoma gangliosides, and (iii) toxin fixation by phospholipid-cholesterol unilamellar vesicles containing either sheep gangliosides or GM2. In order to investigate the intramembrane events leading to membrane disruption following toxin binding, the photoreactive probe 12(4-azido-2-nitrophenoxy)stearoyl 1-14C glucosamine, which inserts into the outer layer and labels integral membrane proteins, was used to establish whether delta toxin penetrates into target cell membrane. No toxin labeling was found, suggesting that toxin action takes place at the membrane surface. This contention is supported by the observation that despite toxin binding, GM2 liposomes did not release entrapped 14C-glucose. Treatment of toxin with carboxypeptidases, but not aminopeptidases, abolished both toxin binding capacity onto erythrocytes and its combination with antitoxin neutralizing antibodies, suggesting that the carboxy terminal end of the toxin is critical for binding to cell membrane.

Molecular cloning and characterization of Bacillus alvei thiol-dependent cytolytic toxin expressed in Escherichia coli

A chromosomal DNA fragment from Bacillus alvei, encoding a thiol-dependent haemolytic product known as alveolysin (Mr 60,000, pI 5.0) was cloned in Escherichia coli SK1592, using pBR322 as the vector plasmid. Only a single haemolysin-positive clone was identified, by testing for haemolysis on blood agar plates. The haemolytic material was associated with the host bacterial cell. It was released by ultrasonic disruption and purified 267-fold. A 64 kDa polypeptide of pI 8.2 cofractionated with haemolytic activity during gel filtration chromatography and isoelectric focusing. It behaved identically to alveolysin in its activation by thiols, inactivation by thiol group reagents, inhibition by cholesterol, and neutralization, immunoprecipitation and immunoblotting by immune sera raised against alveolysin and streptolysin O.

Induction of tumour necrosis factor by staphylococcal toxic shock toxin 1

Staphylococcal toxic shock syndrome toxin 1 (TSST1) induced the release of tumour necrosis factor (TNF) from human and rabbit monocytes in vitro. Nanogram amounts of TSST1 were sufficient to induce TNF release. There was considerable variation in response between cells from different rabbits and different donors. Rabbit monocytes were slightly more sensitive to TSST1 than were human monocytes. Release of TNF in vivo could explain many of the symptoms of toxic shock syndrome.

A new kinetic approach for studying phospholipase C (Clostridium perfringens alpha toxin) activity on phospholipid monolayers

The enzymatic activity of purified phospholipase C (alpha toxin) from Clostridium perfringens was investigated with various phospholipid monolayers. A two-step reaction was used. Enzymatic hydrolysis of insoluble lecithin films by phospholipase C, generating 1,2-diacylglycerol and water-soluble phosphocholine, was coupled with the action of pancreatic lipase in order to give rise to fatty acid and 2-monoacylglycerol, which are rapidly desorbed from the interface. With this new procedure, it is possible to obtain continuous and accurate kinetic measurements of the phospholipase C catalyzed reaction with phospholipid monolayers as the substrate. It is thus possible to avoid the use of radiolabeled substrates as necessary in previous studies, and the difficulties caused by diacylglycerol accumulation in the lipid film are minimized. No hydrolysis was detected when either phosphatidylethanolamine, phosphatidylserine, or phosphatidylglycerol films were used as substrates. By means of a film transfer technique, Ca2+ and Zn2+ ions were found to play a specific and critical role. The present study demonstrates clearly for the first time that Ca2+ is essential for enzyme binding to lipid films, whereas Zn2+ is specifically involved in the catalytic hydrolysis of the substrate.

Damaging effects of Clostridium perfringens delta toxin on blood platelets and their relevance to ganglioside GM2

The lytic effect of Clostridium perfringens delta toxin was investigated on goat, human, rabbit, and guinea pig platelets. In contrast to erythrocytes from the latter three species, which are insensitive to the toxin, the platelets were equally lysed by the same amount of toxin. These results suggest the presence of GM2 or GM2-like ganglioside(s) as a specific recognition site of the toxin on platelet plasmic membrane as previously established for sensitive erythrocytes. Plasmic membrane damage of human platelets was evidenced by the release of entrapped alpha-[14C]aminoisobutyric acid used as a cytoplasmic marker. The specific binding of hemolytically active 125I-delta toxin by human and rabbit platelets was practically identical, dose dependent, and inhibitable by GM2. Labeled toxin was also bound by various subcellular organelles separated from rabbit platelets except the 5-hydroxytryptamine (5-HT)-containing dense bodies, suggesting the absence or inaccessibility of GM2 on the surface of the latter organelles. This result correlates with the low amounts of 5-[3H]HT liberated after platelet challenge with delta toxin whereas this mediator was massively liberated upon lysis by the sulfhydryl-activated toxin alveolysin. The levels of M and P forms of phenol sulfotransferase (PST), involved in 5-HT catabolism, were determined in human platelet lysates after challenge with delta toxin, alveolysin, and other disruptive treatments. The low PST-M activities detected after lysis by delta toxin suggest that this isoenzyme is very likely associated to dense bodies in contrast to PST-P which is cytoplasmic. Platelet lysis by the toxin allows easy separation of these organelles.

Toxic shock syndrome toxin 1 as an inducer of human tumor necrosis factors and gamma interferon

We present evidence that toxic shock syndrome toxin 1 (TSST-1) induces the production of high levels of TNF by human blood monocytes. Enriched lymphocyte preparations incubated with the staphylococcal toxin produced significant levels of TNF-like activity that is not neutralized by anti-rHuTNF antibodies and is likely to be lymphotoxin (LT or TNF-beta). We demonstrate also that TSST-1 is a potent inducer of IFN-gamma. When lymphocyte preparations were costimulated with PMA, the TSST-1 effect was strongly potentiated and the levels of cytotoxic factors, IFN-gamma, and IL-2 present in supernatant fluids were comparable to those observed after treatment with PMA and PHA. Thus, TSST-1, which is also known as an inducer of IL-1 and IL-2, stimulates the production of endogenous mediators that could play a role in the physiopathological processes of toxic shock syndrome (TSS). The described results suggest that the discrepancies in the clinical features between TSS and endotoxin shock may be related to qualitative differences in cytokine production.