Multifaceted Interactions of Bacterial Toxins With the Gastrointestinal Mucosa

MR Popoff


Future Microbiol. 2011;6(7):763-797. 

In This Article

Enterotoxins Active on the Cell Surface

The first step of toxin activity is the recognition of a specific receptor on the target cell surface. When bound to the receptor, toxins can act at the cell membrane by interfering with signal transduction pathways, pore formation or enzymatic activities on membrane compounds. Various enterotoxins interact with cells of the digestive mucosa by using a primarily membrane-based activity, including signal transduction in enterocytes or neuronal extensions in the intestinal mucosa, enzymatic degradation of specific enterocyte membrane compounds or pore formation through enterocyte membrane, leading to change in enterocyte homeostasis, stimulation of certain afferent neurons or necrosis and cell death (Table 2).

Enterotoxins Acting by Hormone-like Signal Transduction

Of the enterotoxins active on cell membranes, some modulate transmembrane signal-transduction pathways in a hormone-like manner. Certain enterotoxins interfere with signal transduction in enterocytes, triggering an efflux of electrolytes and water, which results in the clinical symptom of diarrhea, or stimulates, similar to a neuromediator, nerve cell endings in the intestinal mucosa and, thus, causes vomiting. This strategy allows efficient signal amplification, thus yielding potent diarrheagenic or emetic toxins.

Enterotoxins Modulating Signal Transduction in Enterocytes

Heat-stable enterotoxins Some enterotoxigenic Escherichia coli and other Gram-negative enteropathogens (Yersinia enterocolitica, Vibrio cholerae) secrete heat-stable enterotoxins (STs) that can cause acute diarrhea in humans and animals. These toxins are small peptides that fall into two subgroups: methanol soluble (STa or ST-I) and methanol insoluble (STb or ST-II) toxins. Analysis of STs shows they possess a similar structure, containing three segments joined by three disulfide bridges. Ala13 in the flexible central segment plays a key role in the activity of the toxin. This residue is probably involved in the interaction of the toxin with its receptor. In the case of STa, the secreted protein encompasses 18–19 amino acids, including six cysteines and is capable of forming three disulfide bridges to create a highly stable molecule. The carboxy-terminal segment of STs shares similarities with ionophores and, therefore, are expected to interact with metal ions. Enteroaggregative E. coli (EAggEC) strains also produce a ST termed EAST1, which is related to STa with similar pathological effects.[5,6]

STa induces watery diarrhea without causing obvious histologic morphological damage. The toxin binds to the extracellular domain of guanylate cyclase (GC–C) localized on the apical membrane of enterocytes. GC–C consists of four domains: an extracellular domain, a transmembrane segment, a kinase-like domain, and an enzymatic domain, which catalyzes the formation of cGMP (Figure 3). The kinase-like domain exerts an inhibitory effect on the catalytic activity. Binding of STa to the extracellular domain of GC–C has been suggested to induce a conformational change in the protein kinase-like domain, resulting in an uncontrolled increase of GC–C activity. Elevation of intracellular cGMP activates protein kinase II (cGKII, also referred as protein kinase G) and protein kinase A, which, in turn, phosphorylate and stimulate the cystic fibrosis transmembrane conductance regulator (CFTR) Cl channels. This results in a net fluid secretion through activation of apical Cl channels in parallel with the inhibition of coupled NaCl transporters presumably by inhibiting the Na/H exchanger (NHE)-3. These findings have been confirmed in GC–C-knockout mice, which have a lower intestinal GC–C activity and do not exhibit a secretory response to STa treatment (reviewed in [7–10]). In addition, STa can stimulate duodenal bicarbonate secretion in a CFTR-independent manner via a tyrosine kinase activity resulting in apical Cl/HCO3 exchange.[11,12]

Figure 3.

Mode of action of representative enterotoxins active on enterocyte membrane.
Escherichia coli thermostable toxin (STa), a hormone-like enterotoxin, recognizes the extracellular domain of GC–C and blocks the enzyme in its active form, leading to accumulation of cGMP, activation of protein kinase II (or PK-G), reduction of Na+ absorption by inhibition of the Na/H exchanger (NHE-3), opening of CFTR channels and secretion of water and electrolytes. CPE is a pore-forming toxin that binds to claudin to form small complexes and then to occludin (large complexes) resulting in the leakage of water and electrolytes, as well as in disassembly of tight junctions and the induction of cell death. Vibrio cholerae ZOT activates a signal cascade leading to activation of intracellular kinases, such as PAR2, and subsequent phosphorylation of ZO-1 and opening of tight junctions.
CFTR: Cystic fibrosis transmembrane conductance regulator; CPE: Clostridium perfringens enterotoxin; GC–C: Guanylate cyclase; PK-G: Protein kinase G; ZOT: Zonula occludens toxin.

STa was the first ligand found to bind GC–C, and later studies demonstrated that the hormones guanylin and uroguanylin are the natural ligands for this receptor. These hormones have been shown to be involved in the regulation of fluid and electrolyte transport in intestinal cells and other tissues. Guanylin and uroguanylin consist of 15 amino acids and are highly homologous to STa.[9]

Zonula occludens toxin In addition to the potent enterotoxin, cholera toxin (CT; discussed later), V. cholerae produces other toxins, which interact with the intestinal barrier and contribute to acute dehydrating diarrhea, such as zonula occludens toxin (ZOT). The zot gene is located in the chromosomally integrated genome of a filamentous phage called CTXΦ and the 45-kDa ZOT protein has a role in phage morphogenesis. ZOT is located in the outer membrane of V. cholerae, and a 12-kDa C-terminal peptide, referred to as ΔG, which is proteolytically cleaved and released in the intestinal medium, supports its biological activity. ΔG contains an octapeptide (GRLCVQDG), which is responsible for the binding to a specific receptor in intestinal cell membrane mainly localized at cell–cell contacts. ZOT and the active fragment ΔG induce a reversible opening of tight junctions between cells and increase the paracellular permeability in a nontoxic manner. ZOT preferentially binds to the jejunum and late ileum, and more to intestinal villus than crypt cells, indicating a regional activity in the intestinal tract.

The proposed model of ZOT activity (Figure 3) includes phospholipase C activation subsequently to ZOT binding to cell surface receptor, leading to phosphatidyl inositol hydrolysis with release of inositol 1,4,5-triphosphate (IP3) and diacylglycerol (DAG). PKCα activation resulting from DAG or IP3-dependent release of Ca2+ from internal store, stimulates actin polymerization. Rearrangement of the apical actin cytoskeleton induces a redistribution or reorganization of proteins in tight junction complexes, which is accompanied by a loosening of tight junctions.[13] In addition, ZOT, through its N-terminal hexapeptide, FCIGRL, activates Src kinase and MAPK pathways, leading to tyrosine phosphorylation of ZO-1, possibly in a PKCα-dependent manner, and its redistribution away from tight junctions and rearrangement of actin filaments.[14] ZOT hexapeptide is structurally similar to the activating ligand of proteinase-activated receptor (PAR2), which is involved in tight-junction regulation. Thus, it seems that ZOT hexapeptide directly activates PAR2 leading to PKCα stimulation and subsequent phosphorylation of ZO-1 and also myosin-1C, which is accompanied by loss of ZO-1 association with other tight junction components, such as occludin and claudins and removal of myosin-1C from junctional complexes.[15]

The specific activity of ZOT on tight-junction regulation suggests that this toxin mimics a related physiological modulator of the intestinal epithelial barrier. Using immunopurification based on specific anti-ZOT antibodies and subsequent cloning, a human intestinal ZOT homolog has been identified, which is called zonulin. Zonulin is related to ZOT at the amino acid sequence level. In particular, both proteins retain a similar octapeptide (GGVLVQPG), which is involved in binding to an intestinal receptor. In addition, zonulin shows the same intestinal sites of activity and comparable modulation of intestinal permeability to ZOT, indicating that zonulin recognizes the same receptor on enterocyte and uses a similar mechanism of action. It has been found that zonulin belongs to a serine protease family, including growth hormones, such as EGF.[13]

Accessory cholera enterotoxin Accessory cholera enterotoxin (ACE), the third V. cholerae enterotoxin, is an 11.3-kDa peptide that shows amino acid sequence similarity with eukaryotic ion transporting ATPases, including human plasma membrane calcium pump, calcium-transporting ATPase from rat brain and CFTR. ACE induces fluid secretion in experimental intestinal loop and increases the permeability of intestinal cell monolayers. Its mode of action is still poorly understood. ACE seems not to stimulate cAMP or cGMP, but rather a Ca2+ secondary messenger, since ACE-mediated secretory activity is dependent of extracellular and intracellular Ca2+.[16] ACE is predominantly a hydrophobic protein and the recombinant protein is in a dimeric form. It is presumed that ACE oligomerizes and inserts into cell membrane through its helical C-terminal end, which has an amphipathic character.[17] It is not yet known whether ACE forms functional transmembrane pores.

Enterotoxins Modulating Signal Transduction in Afferent Nerves

CereulideBacillus cereus is a common food-borne pathogen, which causes two types of gastrointestinal diseases in humans, the diarrhoeal and the emetic syndromes. The diarrhoeal disease is due to enterotoxins, which are multicomponent toxins produced in situ subsequent to an overproliferation of B. cereus in the small intestine, whereas the emesis syndrome results from ingestion of a B. cereus emetic toxin, named cereulide, which accumulates in contaminated foods.[18]

Cereulide is a cyclic dodecadepsipeptide (1.2 kDa) that is synthesized by a nonribosomal peptide synthetase.[19–21] Cereulide is produced by only certain B. cereus strains, which are associated with the emetic syndrome.[22] The toxin is stable in acidic conditions, proteolysis and heat and thus, it is not degraded by gastric acid and digestive proteases. Cereulide is cytotoxic towards various primary or cultured cell lines. It induces mitochondria swelling (reported first as vacuole formation) in Hep2 cells and necrotic cell death in porcine pancreatic Langerhans cells, as well as causing emesis in experimental animal models.[23,24] The mechanism of action of cereulide is only partially known. Cereulide is a K+ ionophore similar to valinomycin, as evidenced by selective increase in K+ permeability of phospholipid bilayers induced by the toxin.[25] Both toxins are also structurally related.[20] While valinomycin is more potent at high K+ concentration (120 mM), cereulide is more active than valinomycin at low K+ concentrations (1–3 mM), which corresponds to the physiological level in blood serum. Cereulide promotes K+ uptake by mitochondria, efflux of H+, and drop in the inner transmembrane potential, leading to mitochondria swelling, arrest of respiratory function, inhibition of ATP synthesis and release of proapoptotic or necrotic factors.[26,27] The emetic effects of cereulide seem to be dependent on stimulation of 5-HT3 receptors on vagal afferent neurons, since 5-HT3 receptor antagonists, such as ondanserom hydrochloride or vagotony, inhibit the emetic effect in Suncus murinus.[20] It is not known whether cereulide directly interacts with vagal sensory endings on the intestinal mucosa or releases 5-HT (serotonin) from enterochromaffin cells.

Staphylococcal enterotoxinsStaphylococcus aureus produces a myriad of 25–30-kDa single-chain proteins called staphylococcal enterotoxins (SEs).[28] These potent toxins cause a prevalent form of food poisoning found throughout the world, and these molecules possess 'superantigenic' properties. By classic definition, this latter property involves binding to both major histocompatibility complex class II (MHC II) on antigen-presenting cells and V-specific T-cell receptors, which ultimately cause massive T-cell proliferation with a concomitant uncontrolled production/release of various proinflammatory cytokines. To date, it is uncertain if superantigenicity plays a direct role in SE-induced food poisoning. However, levels of inflammatory mediators, such as prostaglandins and leukotrienes, are increasingly evident in the circulatory system of nonhuman primates shortly after an oral dose of SE type B (SEB).[29] Mast cells may also play a role in SE-induced food poisoning, which perhaps involves not only inflammatory mediators, but also neuropeptides, such as substance P released from sensory neurons by SEB.[30] Another study reveals that SEB-induced effects in mice (intraperitoneal injection) are abrogated by capsaicin, the active ingredient of hot chili peppers, which depletes peptidergic sensory nerve fibers and TNF production.[31] An intraperitoneal injection of SEB into rats induces expression of Fos (a cell activator) throughout the brain via vagus nerve stimulation, thus suggesting that the peripheral presence of an SE has profound effects upon the brain.[32] More recently, it has been reported that SE type A (SEA)-induced emesis is mediated by 5-HT. Indeed, emesis caused by SEA in an animal model (house musk shrew) is inhibited by 5-HT synthesis inhibitor and a 5-HT3 receptor antagonist; in addition, SEA increases 5-HT release in the intestine. The mechanism of 5-HT release mediated by SEA is not yet known. SEA might interact directly with enterochromaffin cells or neurons to trigger 5-HT release, or could act indirectly through the release of proinflammatory molecules or free-radical formation.[33] Mucosal terminals of vagal sensory afferent neurons, which project to the emetic center in the brainstem, contain 5-HT3 receptors.[34] The proposed model includes a SEA-dependent excess release of 5-HT in the intestine and subsequent stimulation of 5-HT3 receptors on vagal afferent neurons triggering the emesis.[33]

Enterotoxins With an Enzyme Activity on Cell Surface

Bacteroides fragilis is a commensal bacteria of the intestine, and occasionally it is responsible for abdominal abscesses and septicemia. Some strains produce an enterotoxin (BFT) and can cause diarrhea in humans, lambs and, more rarely, in other young animals. The bft gene is localized on a 6-kbp pathogenicity island in the chromosome of enterotoxigenic B. fragilis strains. BFT is synthesized as a 397-amino acid protein and is secreted by means of a signal peptide (18 amino acids) as an inactive precursor. The 193 N-terminal residues (propeptide) are linked to the 186 C-terminal residues (mature protein) by a trypsin cleavage site. The mature protein contains a characteristic zinc-binding metalloprotease motif (HExxHxxGxxH). In vitro, BFT is able to proteolyze actin, gelatin, casein and azocoll. Three highly related (92–96% identity) isoforms have been characterized. BFT2, but not BFT1 or BFT3, contains an additional C-terminal extension of 20 amino acids forming an amphipathic structure. This suggests that BFT2 could oligomerize and insert into the cell membrane forming a pore.

Bacteroides fragilis toxin induces morphological changes in cultured intestinal and renal cells (HT29, T84, MDCK and Caco-2) that are able to form intercellular junctions. Cell rounding, increased cell volume, and effacement of microvilli as well as apical junctional complexes are observed. BFT does not enter cells, but cleaves the extracellular domain of E-cadherin in an ATP-independent manner. This step is followed by the degradation of the intracellular domain of E-cadherin probably by ATP-dependent cellular proteases. A reorganization of actin filaments occurs without decrease of the content of actin filaments. In polarized T84 cells treated with BFT, the apical actin ring and microvilli actin filaments are lost, whereas actin filaments accumulate in the basal pole. The permeability of polarized cell monolayers increases in a time- and concentration-dependent manner. BFT is more active when applied in the basolateral side of polarized cells. The cell effects are reversible, since cells regain a normal shape by 2–3 days after toxin treatment that correlates with resynthesis of E-cadherin. In addition, BFT stimulates the secretion of IL-8 by intestinal cells, which promotes an inflammatory response with recruitment of polymorphonuclear leucocytes to the intestinal mucosa. The inflammatory response is supposed to contribute to the intestinal secretory effects. BFT is the only known bacterial toxin that modifies the actin cytoskeleton by cleaving a cell surface molecule. The BFT-dependent proteolysis of the extracellular domain of E-cadherin leads to the loss of the intracellular domain. Since the actin filaments are connected to E-cadherin via catenins, BFT induces a relocalization of the actin filaments from the apical to the basal side, which is accompanied by the loss of microvilli and disorganization of the adherens and tight junctions. This is followed by a decrease of the intestinal cell barrier function (Figure 4).[35,36]

Figure 4.

Examples of intracellularly active enterotoxins.
CT or Escherichia coli thermolabile enterotoxin (LT) recognize GM1 at the enterocyte surface and enter through a long endocytic pathway until reaching the ER. The enzymatic domain (A1) is delivered into the cytosol and blocks G by ADP ribosylation leading to a permanent stimulation of adenyl cyclase, accumulation of cAMP and active secretion of water and electrolytes. ST transits in a similar endocytic pathway. The enzymatic domain (ST-A1) hydrolyzes a glycosidic bond of rRNA, thus inhibiting protein synthesis. Cytolethal distending toxins consist of three subunits. The enzymatic subunit, CdtB, is internalized into the nucleus where it cleaves the DNA and induces cell cycle arrest.
ADPR: ADP-ribose; CFTR: Cystic fibrosis transmembrane conductance regulator; CT: Cholera toxin; EF: Elongation factor; ER: Endoplasmic reticulum; ST: Shiga toxin.

Pore-forming Enterotoxins

Pore formation is a common mode of toxin activity. More than 30% of bacterial toxins are pore-forming toxins (PFTs). Several enterotoxins are specific enterotocyte PFTs. Pores through enterocyte membranes lead to a passive efflux of water and electrolytes. Hence, enterotoxins with pore-formation activity (C. perfringens entrotoxin, aerolysin and Bacillus enterotoxins) usually cause milder diarrhea than enterotoxins, which activate a physiological process of water and ion secretion. Moreover, PFTs might induce intracellular signalings, leading to additional effects, such as disorganization of intercellular junctions, cell apoptosis or necrosis, and, thus, increased epithelial permeability and necrotic lesions of the intestinal mucosa. An inflammatory response usually accompanies the cell necrosis process, thus contributing to the development of intestinal mucosal lesions. C. perfringens β toxins, NetB and Clostridium septicum α-toxin induce severe necrosis and hemorrhage in the intestinal mucosa. Some toxins, such as VacA, have both a pore-forming activity in the plasma membrane and an intracellular activity.

C. Perfringens Enterotoxin Enterotoxigenic C. perfringens strains are responsible for food poisoning and, more rarely, for antibiotic-associated diarrhea and chronic non-food-borne diarrhea in humans. This pathogen has also been suspected in infant death syndrome in humans, as well as in diarrhea in foals and piglets.[37]

Clostridium perfringens enterotoxin (CPE) is a 319-amino acid protein (35.5 kDa), which is synthesized during the sporulation phase and does not contain any signal peptide. The toxin is released from the bacteria during mother cell lysis. Trypsin or α-chymotrypsin increase CPE activity two- to three-fold by removing a putative propeptide.[37]

Clostridium perfringens enterotoxin is cytotoxic for Vero cells and intestinal epithelial cells, including Caco-2, I407 and Hep3b. The first step consists of CPE binding to a cell membrane receptor, which is only present in the CPE-sensitive cells. The CPE receptor has been identified as claudin (isoform 3, 4, 6, 7, 8 or 14), which is an essential component of tight junctions.[38,39] The structure of the CPE receptor-binding domain is rich in β-strands (nine-strand β-sandwich) similarly to receptor-binding domains of other clostridial toxins.[40] In contrast to many other PFTs, CPE receptor does not localize in detergent-resistant membrane microdomains.[41] This results in the formation of a 90–100-kDa complex (small complex) in membrane. At this step, the toxin is largely accessible to antibodies and proteases, such as pronase, indicating that it is exposed to the cell surface. Subsequently, postbinding maturation occurs when cells are incubated at 37°C, consisting of the formation of an intermediate (135 kDa) and then large (160 or 200 kDa) complex by association with a membrane protein. The large complex is significantly resistant to SDS and pronase, probably by insertion of CPE into the membrane. Occludin, a major structural protein of tight junctions, is part of the 200-kDa complex.[42] The CPE large complex has been identified as a prepore, which inserts into the membrane via a β-barrel, formed by the association of β-hairpins (amino acids 81–106) from each CPE monomer (Figure 3). CPE β-hairpin forming the pore consists of two antiparallel β-strands with alternating hydrophobic–hydrophilic residues, which resembles the transmembrane domains of β-barrel PFTs (β-PFTs).[43] The CPE stoichiometry in the pore structure is still poorly understood, however the functional pore consists of CPE hexamer.[41] In addition, the 200-kDa complex formation in polarized Caco-2 cells results in the removal of occludin from tight junctions and increased paracellular permeability.[44] This mechanism is probably responsible for the intestinal fluid accumulation and epithelium desquamation observed in vivo. CPE induces cell death by a mechanism not yet well understood. At high concentrations, CPE seems to trigger cell necrosis and, at low concentrations, cell apoptosis subsequent to Ca2+ entry into cells.[45]

In the small intestine of experimental animals, CPE induces a desquamation of the intestinal cells, particularly those of the villus tips, and a rapid loss of fluid and electrolytes. The ileum is the most sensitive segment of the intestine, whereas little or no effects are observed in the colon. CPE binds to rabbit villus tips where claudin-4 is abundant, suggesting that claudin-4 is an intestinal receptor. Histopathological changes have a major role in the fluid and electrolyte perturbations.[46,47]

Aerolysin Aerolysin was discovered and partially purified from the ubiquitous Gram-negative fresh-water bacteria Aeromonas hydrophila, but it was later found that it is also produced by various other species of the genus Aeromonas. Members of Aeromonas can be pathogenic to a broad spectrum of different hosts, including fish, amphibians, reptiles and mammals. In humans, aerolysin is mainly associated with gastrointestinal diseases, but it is also involved in wound infections, septicemia and meningitis, especially in immunocompromised individuals.[48,49] Pathogenicity to humans is also a potential problem for the food industry, since an investigation in poultry, fish and shrimps revealed that more than 50% of commercial raw food samples, and between 20 and 30% of processed and ready-to-eat food, is contaminated with hemolytic and multiple resistant A. hydrophila.[50]

Aerolysin is produced as a 52-kDa inactive protoxin, called proaerolysin, and secreted in the extracellular medium through a type II secretion system. The monomers are L-shaped molecules with a small N-terminal lobe (domain 1) rich in β-structures and low α-helical structure; the monomer has a bilobal structure and can be divided into a little lobe, consisting of domain 1 (residues 1–82) and a big, elongated lobe, split into three more domains 2–4 (residues 83–470), with the characteristic feature of the presence of long β-strands.[48,49] Aerolysin monomers, but not oligomers, have the ability to bind to specific cell-surface receptors, which have been identified as glycosyl-phosphatidyl-inositol (GPI)-anchored proteins. Activation of the secreted precursor form consists of the proteolytical removal of a 41–43 C-terminal amino acid-long signal peptide. Proteolytic cleavage occurs, thereby, either in solution or in its receptor-bound state within a flexible loop in domain 4. Through the removal of the C-terminal peptide, aerolysin becomes activated and is able to undergo spontaneously oligomerization. Binding to GPI-anchored proteins, which are mainly localized in lipid rafts, facilitating a high concentration of toxin monomers in these restricted cell surface areas, thus favoring their clustering and subsequent oligomerization.[48,49,51] Proteolytic cleavage occurs after receptor binding, and approximately 40 residues are removed from the aerolysin C-terminus. Activated monomers assemble into heptamers through interactions between domains 3 and 4, and form channels through insertion into lipid bilayers. As a consequence of the formation of anion-selective transmembrane channels the cell loses predominantly ions, water, electrolytes and small molecules but no proteins. Besides the breakdown of the Na+/K+ gradient, an influx of extracellular calcium can be observed and seems to be mediating cell signaling.[52] The loss of K+ facilitates the formation and activation of the inflammasome, a cytoplasmic multiprotein complex harboring Nod-like receptors, such as IPAF and NALP3, as well as caspase-1.[53] Finally, the K+ efflux results, via caspase-1 activation, in the mobilization of sterol regulatory element binding protein (SREBP)-1 and -2. Both proteins are involved in the biogenesis of membranes and their activation enhances cell survival.[53] Caspase-1 is also responsible for the processing and activation of cytokines IL-1b and IL-18,[54] which is also observed during aerolysin intoxications. Furthermore, it has been shown that IL-8 is produced as a consequence of activation of the NF-κB pathway upon treatment of macrophages with aerolysin.[55] An additional effect, besides pore formation, is the decreasing host cell membrane integrity, depending on the number of formed pores, and finally resulting in cell lysis. Both events; formation of intrinsic ion channels and loss of membrane integrity, induce a number of cells signaling pathways resulting in the release of secondary messengers. In granulocytes, secondary messengers activate G-proteins, leading to production of IP3 and release of Ca2+ from the endoplasmic reticulum (ER) through IP3-dependent channels. Apoptotic effects have also been shown in murine macrophages, and the causative factor seems to be the influx of extracellular calcium through the aerolysin channels or a secondary effect linked to the toxin-induced TNF-α production.[56,57] It is thought that this effect is caused by a massive influx of calcium. Alteration of MAPK pathways could be involved in the aerolysin-dependent cytotoxic effects.

B. cereus EnterotoxinsBacillus cereus enterotoxins encompass tripartite toxins (Hbl and Nhe) and a single-protein cytotoxin (CytK), which are responsible for the diarrheal food intoxication due to this microorganism. Hbl and Nhe consist of three proteins (L1, L2 and B, and NheA, NheB and NheC, respectively) ranging from 38 to 46 kDa, which show amino acid sequence identity between 18 and 44%. Hbl B contains five long α-helix bundles and a unique subdomain consisting of a hydrophobic β-hairpin flanked by two short α-helices, and is structurally related to the pore-forming toxin cytolysin A produced by E. coli, Shigella flexneri and Salmonella. NheB and NheC probably share a similar structural organization with Hbl B based on their sequence identity. Hbl and Nhe induce fluid accumulation in intestinal loop tests; they are hemolytic and cytotoxic towards various eukaryotic cells. The proposed model of Hbl and Nhe activity includes a binding to a specific cell surface receptor, subsequent oligomerization, pore formation and cell death through colloid osmotic lysis. All three components of Hbl toxin (L1, L2 and B) are able to bind to the cell membrane, whereas only NheB binds to the cell surface. Stoichiometry of oligomers and the structure of the pores are not yet known. The putative oligomeric pores seem to be formed by the α-helices, since the β-hairpins would be too short to span transmembrane β-barrels.[18]

CytK (34 kDa) has been associated with severe symptoms of B. cereus food-borne disease, characterized by bloody diarrhea. CytK belongs to the family of β-PFTs, which also includes C. perfringens β-toxin (discussed later) and S. aureus α-hemolysin. These toxins are secreted as water-soluble monomers, bind to the cell membrane and associate in oligomeric prepores, which subsequently insert into the lipid bilayer, forming transmembrane pores.[18,58]

Clostridium Perfringens β-toxin β-toxin is an essential virulence factor of C. perfringens type B and C,[59–61] which are involved in necrotic enteritis in young animals and in humans (Pig-bel and Darmbrand), and in sheep enterotoxemia. In humans, the disease is characterized by a necrotizing hemorrhagic jejunitis accompanied by abdominal pain, abdominal distension, vomiting and passage of bright blood. Pig-bel was a major cause of death in children in Papua New Guinea, where the population has a low-protein diet and occasionally consumes pork meat, often contaminated with C. perfringens. The incidence of the disease decreased since the 1980s after a vaccination program.[62] However, sporadic cases still occur in parts of Asia, Africa and the South Pacific. Pig-bel is rare in developed countries, some cases have been observed in diabetic patients.[63,64] A similar disease, called Darmbrand, occurred in Germany shortly after World War II.

β-toxin is very labile and sensitive to protease degradation. For this reason, the β-toxin-induced pathology is only observed in particular circumstances, such as in newborns, in which the protease activity of the digestive tract is low, for example piglet necrotizing enterocolitis. The risk factors involved in human disease are a low-protein diet inducing low trypsic activity in the intestine and consumption of sweet potatoes, which contain a trypsin inhibitor. The low protease activity permits a high level of active toxin into the intestinal lumen.[65]

β-toxin is synthesized as a 336-amino acid protein, the first 27 amino acids of which constitute a signal peptide. The secreted protein has a molecular weight of 34,861 Da and a pI of 5.5.[66] The toxin is dermonecrotic and lethal, but it is not hemolytic. When injected intravenously in rat, β-toxin induces an elevated blood pressure and vascular contraction, possibly by interacting with the autonomic nervous system and stimulating catecholamine release.[67] When injected intradermally, β-toxin induces edema and dermonecrosis, which seem to be mediated by stimulation of sensory nerves containing tachykinins, such as substance P and release of TNF-α.[68,69]

β-toxin is related to pore-forming cytolysins produced by S. aureus at the amino acid level: 28% similarity with α-toxin, 18–28% with the A, B and C components of γ-toxin, 17 and 28% with the S and F components of leukocidin and leukocidin R, respectively.[66] S. aureus α-toxin is a single-component toxin that forms hexamers and inserts into the cell membrane, leading to pore formation. In contrast to S. aureus toxins, β-toxin is only cytotoxic for some cell types and is not hemolytic.

β-toxin is a PFT. It was first reported that β-toxin associates with human umbilical vein endothelial cell membranes in multimeric complexes[70] and forms cation-selective channels in artificial phospholipids bilayers.[71] β-toxin induces swelling and lysis of the lymphocytic HL60 cell line, which are preceded by toxin oligomer formation (hexamer or heptamer) in membrane lipid rafts, K+ efflux, and Ca2+, Na+ and Cl influxes. β-toxin pore formation has also been observed in phosphatidyl choline–cholesterol liposomes.[72]

Experimental inoculation of β-toxin with trypsin inhibitor in rabbit ileal or jejunal loop causes a severe hemorrhagic necrosis of the mucosa accompanied by an inflammatory response and bleeding to the lumen, whereas less damage was observed in the duodenum, and no modification in the colon. Administration of β-toxin alone without trypsin inhibitor was ineffective.[73] This confirms that β-toxin is a necrotic toxin for the intestinal mucosa, preferentially for the small intestine segments, and that it is rapidly inactivated by digestive proteases. Interestingly, coculture of β-toxin-producing Clostridium with intestinal Caco-2 cells induces an overproduction of β-toxin as well as other toxins, such as perfringolysin and α-toxin. Overexpression of the β-toxin gene seems to involve the regulatory two-component system VirR–VirS of C. perfringens.[74] Interaction of C. perfringens with intestinal host cells, directly through bacterium–cell membrane contact, or via soluble factors released from cells by bacterial exoenzymes, likely controls toxin production and might account for the high pathogenicity of C. perfringens in the intestine.

C. perfringens NetB A novel toxin, NetB, has recently been observed in C. perfringens strains isolated from avian necrotic enteritis. NetB (33 kDa) is cytotoxic and, since it is related to PFTs, such as δ-toxin, β-toxin and S. aureus α-toxin (38, 40 and 31% amino acid identity, respectively), this toxin is probably a PFT.[75,76] However, NetB might not be the only C. perfringens virulence factor involved in avian necrotic enteritis, since not all C. perfringens isolates from birds with necrotic enteritis contain the netB gene.[77]

A genetic approach has been used to identify NetB as the causative toxin of avian necrotic enteritis. It was first reported that C. perfringens strains isolated from avian necrotic enteritis and able to experimentally reproduce the disease only secrete α-toxin as the major toxin among the already known C. perfringens toxins.[78] However, α-toxin gene-knockout mutants from a C. perfringens strain did not produce α-toxin, but still induced experimental necrotic enteritis in chickens, showing that α toxin was not responsible for the intestinal lesions.[79] Genome sequencing of one of these mutants reveals the presence of a β-toxin-related toxin gene, which was termed NetB. Negative netB mutants did not secrete the cytotoxic NetB, and did not produce the disease in chickens. In addition, complementation with the wild-type netB gene restores pathogenicity.[75] This suggests that NetB is an essential toxin causing avian necrotic enteritis.

Clostridium Septicum α-toxin Clostridium septicum is responsible for fulminate traumatic and nontraumatic gas gangrene, as well as necrotizing enterocolitis. A strong association is observed between C. septicum infection and malignancy (most often colorectal carcinoma and leukemia) and neutropenia. Mucosal ulceration at the tumor surface offers a portal of entry for the bacteria. Anaerobic glycolysis of the tumor is presumed to provide a hypoxic and acidic environment favorable for spore germination and subsequent infection.[80–82] The reasons why C. septicum, which is a rare inhabitant of the human digestive tract, and not another Clostridium species can develop in the intestinal tumor lesions are not known. C. septicum enterocolitis shows necrosis of the intestinal wall, abdominal pain, fever, vomiting and/or diarrhea, which progress rapidly to peritonitis, septicemia and shock. The rate of mortality is high.[80–82]

Clostridium septicum α-toxin is the only lethal virulence factor of C. septicum and is absolutely necessary for pathogenesis.[83]C. septicum α-toxin shows 27% identity with aerolysin at the amino acid level and is structurally related to aerolysin and C. perfringens ɛ-toxin, but domain 1 of aerolysin is missing in C. septicum α-toxin.[84] α-toxin is secreted by means of a signal peptide (32 N-terminal amino acids) via the type II secretion pathway as an inactive prototoxin (46.5 kDa).[84] α-toxin precursor binds to GPI-anchored membrane proteins, similarly to aerolysin. However, the receptors for the two toxins are different, those of C. septicum α-toxin include folate receptor, contactin and Thy1.[85] In contrast to aerolysin, the N-glycan of anchored proteins is not important for the binding to α-toxin.[86] After binding to its receptor, located in lipid rafts, by domain 1, the prototoxin is cleaved to the active form (41 kDa) by cell proteases, such as furin, which cleaves a C-terminal 45 amino acid peptide. The cleaved propeptide functions as an intramolecular chaperone, which hinders the toxin, undergoing incorrect folding or oligomerization in solution. Activated monomers interact with each other to initiate an oligomeric prepore complex. Based on the structural homology of α-toxin with aerolysin and C. perfringens ɛ-toxin, it assumes that α-toxin forms heptameric pores.[84] α-toxin induces a programmed cellular necrosis characterized by rapid cell depolarization, ATP depletion, mirochondria deregulation with release of reactive oxygen species, translocation of proinflammatory histone-binding protein high-mobility group box 1 (HMGB1), and caspase-3-independent cell death.[87,88] The rapid loss of intracellular K+ seems to represent the early signal of pore-forming toxin-mediated cell necrosis.[88]