Bacterial Protein Toxins

Among the many regulatory schemes governing virulence potential, two-component systems have been found time and time again to play key roles. The structure-function relationships of the various domains, kinase, receiver, and HPt domains, of two-component systems are examined in this chapter, following the path of information from outside to inside a bacterial cell. For each domain the general state of knowledge are presented as it relates to two-component systems in general, and then illustrated, when appropriate, with examples from the BvgAS system governing expression of toxins and other virulence factors in B. pertussis. Two-component systems are found regulating not only the structural genes for toxins and other virulence genes but also other regulators. As more of previously dependable antibiotic therapies are becoming less effective against bacterial infections, other modalities for the control of infectious disease are being investigated. Two approaches to this problem are the development of new vaccines and the discovery or creation of new anti-infective compounds. One approach to vaccine development involves the attenuation of organisms to create a live vaccine strain that can colonize and engender a strong and effective immune response without causing a harmful infection. Historically, this has been accomplished simply by growth of the bacterium in vitro for many generations to allow it to lose virulence characteristics. Several features of two-component systems suggest the new anti-infective compounds as potential targets for hopeful drug designers.

Iron availability is one of the regulators of catalase and superoxide dismutase gene expression, and the oxidative stress-response regulators OxyR and SoxRS influence expression of the iron regulator gene fur in Escherichia coli. The discussion of three iron-regulated toxins, diphtheria toxin, Shiga toxin, and Pseudomonas exotoxin A, in this chapter illustrates the similarities and differences in the control mechanisms using bacterial toxin regulation. Pseudomonas exotoxin A has the most complex regulation of the three; regulation by an iron-binding repressor is one step in a cascade of positive and negative regulators that act in concert to control toxin expression. The functions of some of these iron-regulated promoter (irp) genes have not been determined as yet. Like for diphtheria toxin, iron was shown to be the component of media that reduced production of the S. dysenteriae Shiga toxin. The fur mutation resulted in loss of repression in high iron. Studies were done with the cloned toxin genes, and the construct lacked phage sequences that could influence regulation. Thus, the effect of iron and Fur, independent of potential prophage regulators, was determined in the studies. The S. dysenteriae Shiga toxin gene appears to have identical iron regulation as stx1. Repression of toxin synthesis by iron is observed in a number of unrelated pathogens. In each case, an iron-binding repressor protein provides the link between elevated iron levels in the environment and repression of toxin synthesis.

This chapter focuses on selected members of the AraC/XylS family of regulatory proteins as a potential model system to study the induction of genes related to bacterial virulence in human infections. The structural analysis of an AraC family member, Rob, demonstrated some similarities to MarA but also some significant differences, suggesting that alternative modes of DNA binding may influence transcriptional activation relative to different promoter contexts. The activation of the AraC family regulators by their cognate ligand becomes an interesting variable when considering the number of AraC family proteins likely to be involved in the induction of virulence-related genes. The modulation of DNA binding by small molecules specific for AraC family members is a key issue when considering whether disruption of the activation process could be a useful therapeutic strategy. Arabinose binding frees the arms from their interaction with the DNA-binding domain, allowing a preferential interaction with the dimerization domain. Infections with bacteria producing urease often result in cystitis and acute pyelonephritis and can progress to bacteremia. Structural analysis of the AraC proteins has been slowed because of the general aggregation properties of these proteins, but new information is accumulating through the recognition of different functional domains, the construction of chimeric proteins, and the use of genetic approaches. The AraC family includes many transcriptional activators that enhance the production of bacterial virulence determinants.

This chapter focuses on quorum sensing in gram-negative bacteria with a special emphasis on the well-studied intercellular communication network found in Pseudomonas aeruginosa. At least 22 gram-negative species have been shown to utilize an acyl-homoserine lactone based quorum sensing system to control various genes, and more than 50 different species have been shown to produce an acyl-homoserine lactone type of cell-to-cell signal. One of the more well-studied pathogens in this group is P. aeruginosa, which contains two separate quorum sensing systems. The study of quorum sensing in P. aeruginosa began when it was discovered that the production of the virulence factor elastase was controlled by LasR, a homolog of LuxR. The genetic organization of the rhl quorum sensing system is also similar to the las quorum sensing system. Evidence for the importance of quorum sensing in infections was also found by randomly mutagenizing the P. aeruginosa wild-type strain PA14 in search of virulence factors. A popular theory is that delaying the production of certain virulence factors may allow P. aeruginosa to face a lesser immune response while its population builds. Biofilm formation is believed to be a critical step in the disease produced when P. aeruginosa chronically infects the lungs of patients with cystic fibrosis. The authors hope the understanding of quorum sensing will provide the background for the development of new and effective antimicrobial therapies that will provide much needed options for the treatment of bacterial infections.

In gram-negative bacteria such as Escherichia coli, proteins destined for the cell surface or surrounding medium must cross the cytoplasmic (inner) and outer membranes and the periplasm between them. Several mechanisms employ periplasmic intermediates, e.g., for assembly of adhesion pili, or utilize a large number of proteins to span the envelope, e.g., in the assembly of flagella. The type I export mechanism contrasts with these as it does not generate periplasmic intermediates and employs a dedicated secretory apparatus of just three proteins. Although acylation is essential for toxin interaction with mammalian cell target membranes, it is not required for export across the prokaryotic envelope. The type I export machinery requires only three export components. These are all integral membrane proteins, a traffic ATPase, an accessory or ‘‘adaptor’’ protein (HlyD), and the outer membrane protein TolC. Cell membrane traffic ATPases provide energy from ATP hydrolysis for movement of various molecules, large polypeptides to small ions, across membranes. The hemolysin export ATPase HlyB has 707 residues and is assumed to function as a homodimer. Putatively, six transmembrane helices between amino acids 158 and 432 interact with the bacterial membrane, whereas the C-terminal ca. 200 residues form the ATPase domain located in the cytoplasm. Each TolC monomer contributes four antiparallel β-strands and four antiparallel α-helical strands to form the channel and tunnel domains, respectively. Whereas a β-barrel is a typical feature of outer membrane proteins, the TolC channel domain is different in that the three monomers form a single β-barrel.

The type II secretion pathway is responsible for secretion of toxins and a variety of hydrolytic enzymes that include proteases, lipases, phospholipases, cellulases, and pectinases, which contribute to tissue damage and disease of animals and plants. Cholera toxin, exotoxin A, and aerolysin produced by Vibrio cholerae, Pseudomonas aeruginosa, and Aeromonas hydrophila, respectively, are examples of three toxins that utilize the type II pathway for extracellular secretion. The biogenesis of all three toxins has been analyzed, and their crystal structures have been solved. The secretion of CT across the bacterial cell envelope occurs in two distinct steps that have different requirements; transport across the cytoplasmic membrane is followed by outer membrane translocation. A current hypothesis postulates that molecules secreted by the type II system may encode information critical to their secretion within their tertiary and quarternary structures. Restriction of proaerolysin to the poles could be evidence of periplasmic compartmentalization with concomitant secretion at the poles. The green fluorescent protein (GFP) was fused with Eps proteins to determine the cellular location, in living cells, of the type II secretion apparatus in V. cholerae. Recent intriguing data based on experiments accomplished with an active blue fluorescent protein (BFP) fused with ribosomal protein L1 in B. subtilis showed that ribosomes are located around the periphery of nucleoids, predominantly at the cell poles.

Type III secretion system (TTSS) is believed to have originally evolved from the flagellar export system and is now dispersed among a number of both animal- and plant-interacting gram-negative bacteria. The aim of much current research on TTSS is to understand the mechanisms involved in effector secretion/injection and what the effectors are doing inside the host cell. Bacterial pathogens use several different protein secretion pathways to export virulence proteins from the bacterial cytoplasm to their site of action. Chaperones bind to effector proteins in the bacterial cytosol and remain cytosolic following export of their cognate substrate. Secreted proteins exhibiting significant structural similarities to YopB and YopD are present in all TTSSs of animal pathogens but not plant pathogens. Secretion of LcrQ upon cell contact depletes LcrQ from the cytoplasmic compartment and triggers increased transcription of type III genes. Secretion and polymerization of PrgI are required to complete the assembly of the needle complex and type III export apparatus. Two flagellar components, FlhB, an inner membrane protein with a substantial C-terminal cytoplasmic domain, and FliK, the secreted hook-length control protein, are proposed to be involved in switches substrate specificity process. An important question is how far knowledge of flagellar biogenesis can be extrapolated to understand the structure and function of the type III export apparatus. It is obvious that there will be features unique to each system; however, the basic process of transporting substrates across the bacterial membranes appears to be relatively well conserved.

Type IV transporters differ from other transporter systems in that type IV systems are used not only to transport proteins but also to mobilize DNA. In fact, it is likely that type IV transporters first evolved as conjugation systems that functioned to transfer genetic information between bacteria and only later were these systems modified by pathogenic bacteria to transfer critical virulence factors across the bacterial membranes and into the host eukaryotic cell. Since transfer of genetic information from one bacterium to another via bacterial conjugation is an ancient system that likely predates the evolution of pathogens and therefore predates the necessity to transport virulence factors across bacterial membranes, it seems likely that the first type IV transporter was a conjugation system that transferred DNA from one bacterium to another. A slight variation of type IV DNA transport systems appears to have occurred when type IV systems evolved that solely transport proteins without any DNA attached. Pathogens that produce type IV transporters include Bordetella pertussis, Brucella spp., Bartonella henselae, Helicobacter pylori, Rickettsia prowazekii, and Legionella pneumophila. B. pertussis is the causative agent of the disease pertussis, or whooping cough. This pathogen secretes one of its important virulence factors, pertussis toxin (PT), using a type IV transporter. Several intracellular pathogens utilize type IV transporters to export important virulence factors from the bacterium to the cellular milieu of the host cell.

A review of Shiga toxin receptor, diphtheria toxin (DT) receptor, Pseudomonas exotoxin A (PEA) receptors will illustrate the different ways these molecules are studied and that bacterial toxins have harnessed a variety of processes to enter target cells. Evidence presented by researchers that low-density lipoprotein receptor-related protein (LRP) serves as the receptor for PEA is several-fold. First, both the toxin-binding protein purified from LM cells or mouse liver and LRP have similar mobility on SDS-PAGE and are indistinguishable immunologically. Second, native PEA, but not a mutant toxin defective in its ability to bind to LM cells, binds to purified LRP that is immobilized on polystyrene or on nitrocellulose; the toxin interacts with the 515-kDa heavy chain of LRP on ligand blots, not with the 85-kDa light chain. Third, receptor-associated protein (RAP) both blocks binding of PEA to mouse LM cells and abolishes toxicity. Cells expressing receptors with different-length juxtamembrane domains bind DT normally; however, they exhibit reduced sensitivity to DT when compared to wild-type cells. The three receptors have functions essential to the normal physiologic properties of mammalian cells. Nevertheless, they represent molecules usurped by different bacterial toxins as the first step in the intoxication process. Cells lacking functional cell surface receptor are resistant to the toxin, expression of receptor correlates with the tissue specificity of toxin damage, and this in turn correlates with the disease symptoms seen in animal models and in patients.

Adenylate (or adenylyl) cyclase (AC) toxin is an essential virulence factor for Bordetella pertussis, as mutants lacking this molecule are virtually avirulent in an animal infection model. Receptor-mediated endocytosis (RME) is characterized by several features that can be used to determine whether a particular toxin is entering by that pathway. First, it is generally true that ligand binding can occur at 4 and 37°C, but internalization only occurs at 37°C. Second, binding is saturable and limited by the number of receptor molecules on the target cell. Third, toxins that enter target cells by RME exhibit a lag phase between toxin addition and biological effects, due to the time necessary for uptake into an endosome, acidification of the endosome, and resultant translocation of the catalytic portion into the cytoplasm. All toxins that act on intracellular targets must have a mechanism by which at least a part of their structure can traverse the plasma membrane of the host cell. Divalent metal binding to AC toxin and the membrane potential of the target cell are clearly critical factors in the processes of insertion and translocation of the toxin, but the signals that initiate this sequence of events from the cell surface remain unknown. Most of the understanding of the structure and function of AC toxin comes from work utilizing soluble material obtained by urea extraction of B. pertussis organisms or recombinant Escherichia coli expressing AC toxin. Infection with B. pertussis is not systemic; organisms remain localized to the respiratory mucosa.

A number of protein toxins with a cytosolic target act on cells by first binding to cell surface receptors, then the toxins are endocytosed, and subsequently they are transported to the organelle from where they are translocated to the cytosol. In spite of the structural similarities between these toxins, they use different strategies to enter the cytosol. Most protein toxins have to be endocytosed before being translocated to the cytosol. Only few exceptions are known; one is the adenylate cyclase from Bordetella pertussis, which enters the cytosol directly through the plasma membrane. A number of protein toxins enter the cytosol either from acidic endosomes or from the endoplasmic reticulum (ER) after retrograde transport from endosomes to the Golgi apparatus and to the ER. Endocytosis of protein toxins can occur by several mechanisms. Cholera toxin was the first bacterial toxin visualized in the Golgi apparatus. Later, toxin transport to the Golgi apparatus has been demonstrated for several other protein toxins as well. Both bacterial toxins and plant toxins such as ricin are transported from endosomes to the Golgi apparatus. There is a lively transport of newly synthesized proteins from the cytosol and into the ER lumen through the Sec61p complex. It is also known that this protein complex is used for transport of misfolded protein in the other direction, back into the cytosol, where these proteins can be ubiquitinylated, deglycosylated, and degraded by proteasomes.

This chapter focuses on the biology and mechanisms of transcytosis for dIgA and IgG, and explains in detail the mechanism of transcytosis for cholera toxin. Reference is made to Shiga and Shiga-like enterotoxins that move across epithelial barriers by opportunistically exploiting similar mechanisms of vesicular transport Fcγ receptor (FcRn) is expressed in the polarized epithelial cells lining the intestine of adult humans. Cholera toxin (CT) binds a ganglioside receptor on the apical cell surface, enters the intestinal cell by non-clathrin-mediated endocytosis, and then moves retrograde into Golgi cisternae and endoplasmic reticulum (ER) to enter the cytoplasm and induce disease. To induce disease, CT must move from the cell surface into the endosome and then retrograde into the ER. To distinguish the transcytotic pathway that involves transit through the Golgi apparatus from that defined by transcytosis of the pIgR and the FcRn, the authors have termed the process of CT trafficking across epithelial cells as indirect transcytosis. It is well documented that CT and LTI represent the most potent mucosal immunogens and adjuvants recognized to date. Such efficiency in eliciting inductive immunity after application to mucosal surfaces, not seen with other ingested proteins of comparable size, implies that CT exhibits an ability to encounter and perhaps act on antigen-presenting cells of the mucosal immune system. Transcytosis of both dIgA and IgG depends on sorting motifs embedded within the protein structures of the pIgR and the FcRn.

The chapter discusses bona fide pore-forming toxins (PFTs) with the exception of RTX toxins. The role of PFTs in bacterial pathogenesis are discussed first; then their general mode of action are outlined, and the events that lead to pore formation are described at the structural level, using two examples, the PFTs from Staphylococcus aureus and the cholesterol-dependent toxins (CDTs). Finally, some of the consequences of pore formation are reviewed. Most PFTs are able to form pores in artificial membranes such as liposomes. This allowed researchers to study the mechanisms that lead to pore formation in great detail using in vitro approaches. Staphylococcus aureus secretes a variety of membrane-damaging toxins, including the α-hemolysin and the bicomponent leukotoxins, the active toxin of which comprises the combination of two similar subtype proteins. VacA was recently found to form small anion-selective, voltage-dependent channels in biological membranes at acidic pH. A variety of PFTs were also shown to trigger the release of calcium from intracellular stores (e.g., staphylococcal PFTs, aerolysin, and streptolysin O). By unknown mechanisms, these toxins lead to activation of G proteins, production of inositol(1,4,5)-triphosphate, and opening of calcium channels in the endoplasmic reticulum. Pore formation in the plasma membrane also allows entry of extracellular calcium.

The major cellular effect of RTX toxins has been ascribed to their pore-forming capacity, resulting in plasma membrane lesions and osmotic lysis. One group of RTX toxins consists of hemolysins, such as Escherichia coli α-hemolysin (HlyA) and Actinobacillus pleuropneumoniae (ApxIA), that are toxic for a wide range of cell types from various species including humans and ruminants. The other category embraces leukotoxins produced by A. actinomycetemcomitans (LtxA) and Pasteurella haemolytica (LktA), which display a more restricted target cell cytolytic activity. Both the number and the length of the fatty acyl groups differ among the RTX toxins. The characteristic feature of RTX toxins is a Ca2+-binding repeat domain (GGXGXDXUX, where U represents a large hydrophobic residue and X represents any amino acid) located in the C-terminal part of the protein. More recently, data have been presented that suggest that RTX toxins also affect intracellular signaling pathways without causing lysis of the cells. High doses of RTX toxins result in oligomerization and formation of large transmembrane pores that could be responsible for an extremely rapid destruction of the target cell membrane, providing no time for the host cell to defend itself by induction of the inflammatory response or apoptosis. The role of two RTX toxins in the urinary and respiratory tract, respectively, is finally reviewed in the chapter.

The study of ADP-ribosylating toxins has advanced our understanding of bacterial pathogenesis and provided insight into the molecular basis of eukaryotic physiology, especially G-protein-coupled signal transduction. The pathology associated with an ADP-ribosylating toxin is due to alterations in the activity of specific eukaryotic proteins. To date, essentially all of the eukaryotic proteins that are ADP-ribosylated by bacterial toxins are nucleotide-binding proteins, most often GTP-binding proteins. Two types of activation have been observed, covalent modification and association with eukaryotic proteins or cofactors. The study of bacterial toxins involves a sequential evolution of knowledge and strategies. First, the toxin is isolated and its catalytic and intoxication mechanisms determined; next, strategies are developed to determine whether the toxin is a useful vaccine candidate by empirically attempting to chemically or genetically inactivate the toxin while retaining its immunogenicity. The final stage in toxin research is to determine whether it can be used as a pharmacological reagent. Continued studies on the eukaryotic mono-ADP ribosylating enzymes should provide new insight into how the posttranslational modification regulates cell physiology, as studies on bacterial ADP-ribosylating toxins have contributed to our understanding of bacterial pathogenesis. Polyacrylamide gel electrophoresis techniques have been used for the measurement of the kinetics of in vivo ADP-ribosylation of eukaryotic proteins by other bacterial toxins. Although ADP-ribosylation of eukaryotic proteins represents the first covalent modification attributed to a bacterial toxin, subsequent studies have identified additional covalent and noncovalent modifications catalyzed by bacterial toxins.

The family of large clostridial cytotoxins comprises Clostridium difficile toxins A and B, the lethal and the hemorrhagic toxins from Clostridium sordellii, the α-toxin from Clostridium novyi, and various toxin isoforms mainly produced by C. difficile. Differential glucosylation is used to measure the modification of Rho GTPases in intact cells. Toxin-modified Rho GTPase in intact cells blocks subsequent toxin-catalyzed labeling of Rho proteins in the cell lysate upon addition of UDP[14C]glucose. Covalent modification by bacterial protein toxins not only inhibits but also activates Rho GTPases. Measurement of the GTP hydrolysis by cytotoxic necrotizing factor (CNF)1-treated RhoA reveals an inhibition of the intrinsic and GTPase-activating protein (GAP)-stimulated GTPase activity, indicating that deamidation of Gln-63 forms a constitutively active Rho protein. The modification of Rho GTPases by dermonecrotizing toxin (DNT) occurs predominantly with the GDP-bound form of Rho, while the modification of Rho by CNFs is not nucleotide dependent. Like deamidation of Gln-63, transglutamination of Gln-63 inhibits GTP hydrolase activity of Rho proteins, although the precise function of the addition of primary amines onto Rho GTPases is not completely understood. Deamidation and transglutamination by CNFs or DNT change the migration of GTPases by SDS-PAGE. This change in migration occurs with RhoA but not Rac or Cdc42 and depends on the deamidation of Glu-63 and, therefore, is also observed with the recombinant RhoA-Q63E. Mass spectrometric analysis of proteolytic peptides of Rho GTPases allows the detection of the 1-Da shift caused by CNF1-induced deamidation of Gln-63 of RhoA (Gln-61 of Rac and Cdc42).

Along with identification of the enzymatic function of ribosome-inactivating proteins (RIPs) came the recognition that certain bacterial toxins, called Shiga toxins (Stxs), fit into the category of plant proteins. We now know that these plant and bacterial poisons have the broader specificity of polynucleotide:adenosine glycosidases, at least in vitro. This chapter refers to the enzymes interchangeably as RIPs or RNA N-glycosidases. Recently, additional enzymatic activities have been attributed to the RIPs. However, when four different RIPs were highly purified, the only enzymatic activity they retained was the capacity to remove adenine from rRNA or DNA. This observation indicates that the RIPs, originally defined as RNA N-glycosidases, are actually polynucleotide:adenosine glycosidases, at least in vitro. One unusual feature of the RIPs is that, although they share the same enzymatic activity and their preferred target is intact mammalian ribosomes, they are not equally active on target ribosomes in vitro. There appear to be at least two explanations for this phenomenon. One reason that some RNA N-glycosidases do not appear to be as active as others in cell-free assays is that they require cofactors such as ATP. The second reason that some RIPs have varying specificities is that they appear to interact with ribosomal proteins. Consistent with a protective function for the plant RNA N-glycosidases is that the RIPs are both lethal and often stored in large quantities.

This chapter examines the molecular biology and biochemistry of cytolethal distending toxin (CDT) and its molecular mode of action. Transcript analysis of the Haemophilus ducreyi cdt gene cluster (Hd-cdtABC) indicated that the cdt genes are organized in an operon. The cdtABC operon encodes three CDT polypeptides, CdtA, CdtB, and CdtC. All three proteins bear apparent signal peptide coding sequences consistent with secretion across the inner membrane by the general export pathway. Biochemical evidence for the requirement of CdtA, CdtB, and CdtC in biological activity has recently been unequivocally demonstrated. Recent findings on the role of CdtB in the mechanism of CDT action, the similarity of CDT action to that of IR, the apparent pathway of CDT internalization, and the reconstitution of biological activity from pure CDT subunits constitute major advances in the understanding of CDT. The apparent cell type specificity for the induction of apoptosis versus cell cycle arrest and the activation of p53 following CDT treatment represent additional parallels between toxin action and radiation treatment. The resolution of this observation will yield not only important information regarding CDT but perhaps also the cellular response to DNA damage in general.

This chapter focuses on bacterial protein toxins that target eukaryotic cells acting as zinc-metalloproteases. Additional understanding of the mode of action of proteases that contribute to pathogenicity could lead to the development of inhibitors that could prevent or interrupt the disease process. The seven botulinum neurotoxins (BoNTs, types A to G) and the single tetanus neurotoxin (TeNT) are responsible for the clinical manifestations of botulism and tetanus, respectively, and constitute a group of metalloproteases endowed with unique properties. These bacterial metalloproteases do not act on the cell surface but exert their enzymatic action in the cell cytosol on selected proteins that form the core of the neuroexocytosis machinery. The SNARE complex is insensitive to clostridial neurotoxins (CNT) proteolysis, as expected on the basis that proteases are known to attack predominantly unstructured exposed loops. Bacteroides fragilis toxin (BFT) is the first bacterial toxin known to remodel the intestinal epithelial cytoskeleton and F-actin architecture via cleavage of a cell surface molecule and represents the prototype of a novel class of bacterial toxins that act without cell internalization and covalent modification of intracellular substrates. In analogy with tetanus, botulism, and anthrax, one is tempted to suggest that BFT is an essential virulence factor.

This chapter describes a distinct class of toxins that mimic endogenous regulatory factors of Rho GTPases. The actions of bacterial toxins are noncovalent and are therefore reversible. These ‘‘modulating toxins’’ encode domains that function as guanine nucleotide exchange factors (GEFs) or as GTPase-activating proteins (GAPs). The study of bacterial GEFs and GAPs is shedding new light on mechanisms of bacterial pathogenesis. Binding of GTP to the nucleotide-free form of the GTPase is favored by a high cytoplasmic concentration of GTP relative to GDP. Three types of endogenous eukaryotic proteins regulate GTPase cycling between GTP-bound and GDPbound forms. The rate of GTP hydrolysis is accelerated by interaction with GAPs. GAPs act in two ways to increase GTPase activity. The first involves the positioning of a nucleophilic water molecule that attacks theγ-phosphate of GTP. Second, GAPs donate a catalytic arginine residue that is needed to complete the active site of the GTPase. Several toxins secreted by type III systems have GAP activity for Rho GTPases. These are SptP from Salmonella enterica, ExoS and ExoT from Pseudomonas aeruginosa, and YopE from Yersinia pestis, Y. pseudotuberculosis, and Y. enterocolitica. A major difference between them is that ExoT possesses only about 0.2% of the catalytic ADP-ribosyltransferase activity of ExoS. The three-dimensional structures of ExoS bound to Rac1 and SptP bound to Rac1 in the presence of GDP have been determined.

This chapter discusses the superantigens (SAgs) of Staphylococcus aureus and Streptococcus pyogenes. The classic SAgs form a large family of related proteins that lack detectable enzymatic function but rather function to cross-link major histocompatibility complex (MHC) class II molecules on antigen-presenting cells (APCs) with certain T-cell receptors (TCRs), primarily those of the CD4 lineage. Staphylococcal and streptococcal SAgs are single-polypeptide-chain, nonglycosylated proteins of 22 to 30 kDa. TSST-1 is encoded on SaPIs 1, 2, and bovine. SaPIs are approximately 15.2 kb in size and are present in a single, but not necessarily the same, site in the chromosome. Mutagenesis studies provided evidence that the cystine loop of SE C1 is required for emesis. SAgs interact with TCR outside the typical TCR region for contact with antigenic peptide and MHC class II. The massive release of these cytokines can cause hypotension through capillary leak mediated by the nitric oxide pathway, with consequent effects on the endothelium. SAgs are most often associated with TSS illnesses. Staphylococci and streptococci are common pathogens of humans, and as such have myriad cell surface virulence factors that contribute to colonization and immune avoidance. Most of the SAgs are made during the postexponential phase of growth (excluding SE A, K, and Q, which are exponential phase regulated), when cell numbers are highest and the organisms are most likely to spread to new hosts.

Several bacteria produce potent toxins that are entirely, or in part, responsible for the severe diseases caused by the microorganisms. This chapter describes those toxins classically used to prevent disease and the approaches for their future use. Today, it is possible to inactivate toxins and microorganisms by using genetic tools. Therefore, in addition to traditional diphtheria and tetanus vaccines, the acellular pertussis vaccine is also described in the chapter . This represents the first vaccine produced by genetic inactivation of a bacterial toxin. Several methods have been described for the purification of diphtheria and tetanus toxins and are generally based on diafiltration of culture supernatant, precipitation by ammonium sulfate, and, if necessary, purification by gel filtration or ion-exchange chromatography. With these methods, diphtheria and tetanus toxins can be purified to 85 to 95% purity, representing approximately 2,300 and 1,800-2,000 Lf/mg of protein nitrogen for tetanus and diphtheria, respectively, by ammonium sulfate precipitation, whereas the conventional vaccines have a purity of approximately 60%. Pertussis toxin (PT) plays a central role in the pathogenesis of whooping cough and induces protective immunity against infection. As for the other toxins, to be included in vaccines, PT needs to be detoxified. The most powerful mucosal immunogens and adjuvants recognized to date are cholera toxin (CT) and Escherichia coli heat-labile enterotoxin (LT). To study the structure-function of CT and LT and to find molecules that are nontoxic but still active as mucosal adjuvants and immunogens, more than 50 site-directed mutations have been generated within these toxins.

Bacterial toxins represent some of the deadliest molecules known to man. Over the ages, evolution has led to precision engineering of these toxins such that only minute quantities of certain bacterial toxins are capable of eliciting lethal outcomes for humans. Many vaccines are inactivated bacterial toxins, which stimulate an immune response to protect the recipient from that toxin-producing pathogen. Regrettably, the power of bacterial toxins has also been harnessed for destructive purposes. Toxins and toxin-producing bacteria that are thought to pose major risks to the public are Bacillus anthracis, which causes anthrax in humans and animals, Yersinia pestis, which is the causative agent of plague, and botulinum toxin. Continued studies on the basic molecular properties of bacterial toxins will provide insight to develop novel antitoxin therapies, such as the recent recognition that nontoxic forms of bacterial toxins can act in a dominant-negative manner to neutralize toxin action. The potential utilization of pathogenic bacteria and their toxins for criminal acts is a formidable problem that requires a concerted effort on the part of the scientific community. Scientific research has the promise of leading to the development of vaccines and therapeutics that will make biological agents useless in the hands of terrorists.