Self-interaction of pneumolysin, the pore-forming protein toxin of Streptococcus pneumoniae.

The pathogenically important cholesterol-binding pore-forming bacterial "thiol-activated" toxins (TATs) are commonly believed to be monomeric in solution and to undergo a transition on membrane binding mediated by cholesterol to an oligomeric pore. We present evidence, gained through the application of a number of biochemical and biophysical techniques with associated modelling, that the TAT from Streptococcus pneumoniae, pneumolysin, is in fact able to self-associate in solution to form the same oligomeric structures. The weak interaction leading to solution oligomerization is manifested at low concentrations in a dimeric toxin form. The inhibition of toxin self-interaction by derivatization of the single cysteine residue in pneumolysin with the thiol-active agent dithio (bis)nitrobenzoic acid indicates that self-interaction is mediated by the fourth domain of the protein, which has a fold similar to other proteins known to self-associate. This interaction is thought to have implications for the understanding of mechanisms of pore formation and complement activation by pneumolysin.

The molecular mechanism of pneumolysin, a virulence factor from Streptococcus pneumoniae.

Pneumolysin, a member of the thiol-activated cytolysin family of toxins, is a virulence factor from the Gram-positive bacterium Streptococcus pneumoniae. The toxin forms large oligomeric pores in cholesterol-containing membranes of eukaryotic cells. A plethora of biochemical and mutagenesis data have been published on pneumolysin, since its initial characterization in the 1930s. Here we present an homology model of the monomeric and oligomeric forms of pneumolysin based on the recently determined crystal structure of perfringolysin O and electron microscopy data. A feature of the model is a striking electronegative surface on parts of pneumolysin that may reflect its cytosolic location in the bacterial cell. The models provide a molecular basis for understanding the effects of published mutagenesis and biochemical modifications on the toxic activity of pneumolysin. In addition, spectroscopic data are presented that shed new light on pneumolysin activity and have guided us to hypothesise a detailed model of membrane insertion. These data show that the environment of some tryptophan residues changes on insertion and/or pore formation. In particular, spectroscopic analysis of a tryptophan mutant, W433F, suggests it is the residue mainly responsible for the observed effects. Furthermore, there is no change in the secondary structure content when the toxin inserts into membranes. Finally, the basis of the very low activity shown by a pneumolysin molecule from another strain of S. pneumoniae may be due to the movements of a key domain-domain interface. The molecular basis of pneumolysin-induced complement activation may be related to the structural similarity of one of the domains of pneumolysin to Fc, rather than the presumed homology of the toxin to C-reactive protein as previously suggested.

The cholesterol-dependent cytolysins.

In view of the recent studies on the CDCs, a reasonable schematic of the stages leading to membrane insertion of the CDCs can be assembled. As shown in Fig. 3, we propose that the CDC first binds to the membrane as a monomer. These monomers then diffuse laterally on the membrane surface to encounter other monomers or incomplete oligomeric complexes. Presumably, once the requisite oligomer size is reached, the prepore complex is converted into the pore complex and a large membrane channel is formed. During the conversion of the prepore complex to the pore complex, we predict that the TMHs of the subunits in the prepore complex insert into the bilayer in a concerted fashion to form the large transmembrane beta-barrel, although this still remains to be confirmed experimentally. Many intriguing problems concerning the cytolytic mechanism of the CDCs remain unsolved. The nature of the initial interaction of the CDC monomer with the membrane is currently one of the most controversial questions concerning the CDC mechanism. Is cholesterol involved in this interaction, as previously assumed, or do specific receptors exist for these toxins that remain to be discovered? Also, the trigger for membrane insertion and the regions of these toxins that facilitate the [figure: see text] interaction of the monomers during prepore complex formation are unknown. In addition, the temporal sequence of the multiple structural changes that accompany the conversion of the soluble CDC monomer into a membrane-inserted oligomer have yet to be defined or characterized kinetically.

Crystallization and preliminary X-ray analysis of a thiol-activated cytolysin.

We present the first reported crystallization of a member of the thiol-activated family of protein toxins. Perfringolysin O, a virulence factor of Clostridium perfringens, has been crystallized in two different forms by the hanging drop vapor diffusion method. In one form the toxin crystallizes with PEG 20000 in the orthorhombic space group C222(1) with cell dimensions of a = 47.8 A, b = 182.0 A and c = 175.5 A and the crystals diffract to beyond 2.5 A resolution. In the second form the toxin crystallizes in a large variety of organic solvents including malt whisky. This crystal form belongs to the orthorhombic space group P222(1) with unit cell dimensions a = 47.1 A, b = 166.1 A and c = 214.0 A and with diffraction observed to 2.4 A resolution.

Clostridium perfringens invasiveness is enhanced by effects of theta toxin upon PMNL structure and function: the roles of leukocytotoxicity and expression of CD11/CD18 adherence glycoprotein.

Clostridium perfringens infections are characterized by the lack of an inflammatory response at the site of infection and rapidly progressive margins of tissue necrosis. Studies presented here investigated the role of theta toxin from C. perfringens in the pathophysiology of these events. Mice passively immunized with neutralizing monoclonal antibody against theta toxin and challenged with an LD100 of log phase C. perfringens had significantly less mortality than untreated controls. Intramuscular injection of killed, washed C. perfringens in mice induced a massive time-dependent influx of polymorphonuclear leukocytes (PMNL) into tissue; injection of either viable, washed C. perfringens or killed organisms plus theta toxin dramatically attenuated PMNL influx although PMNL accumulated in adjacent vessels. The anti-inflammatory effects could not be attributed to an absence of chemoattractants since C. perfringens proteins had chemotactic factor activity, and killed bacilli generated serum-derived chemotactic factors. Scanning and transmission electron microscopy demonstrated the dramatic leukocidal effects of high doses of theta toxin on PMNL. In contrast, sublethal concentrations of theta toxin primed PMNL chemiluminescence, disrupted PMNL cytoskeletal actin polymerization/disassembly, and stimulated functional upregulation of CD11b/CD18 adherence glycoprotein. In summary, these results demonstrate that theta toxin is an important virulence factor in C. perfringens infection. In a concentration-dependent fashion, theta toxin contributes to the pathogenesis of clostridial gangrene by direct destruction of host inflammatory cells and tissues, and by promoting dysregulated PMNL/endothelial cell adhesive interactions.

Clostridium perfringens beta toxin and Clostridium septicum alpha toxin: their mechanisms and possible role in pathogenesis.

The Clostridium septicum alpha toxin and the Clostridium perfringens beta toxin are examples of pore-forming toxins that exhibit several different features. The cell types that are targeted by these toxins reflect the effect these toxins have on the host during infection with either organism. Alpha toxin elicits a rapid shock-like syndrome, whereas beta toxin appears to induce a variety of neurological effects. The effects of the purified toxins appear to mimic some of the features of the animal and human diseases caused by C. septicum and C. perfringens. This review, examines the current state of knowledge for the cytolytic mechanism, role in pathogenesis and structure of these two toxins.

Cloning and expression in Escherichia coli of the perfringolysin O (theta-toxin) gene from Clostridium perfringens and characterization of the gene product.

The gene encoding perfringolysin O, the thiol-activated hemolysin from Clostridium perfringens (ATCC 13124), was cloned and expressed in Escherichia coli. A gene library of C. perfringens chromosomal DNA was constructed in bacteriophage lambda EMBL3. A recombinant was identified that produced a hemolysin that was inhibited by cholesterol and was tentatively identified as perfringolysin O. Subcloning experiments localized the perfringolysin O gene (pfo) to a 1.8-kilobase region on the cloned chromosomal fragment. E. coli which carried a plasmid subclone of pfo (pRT1B) expressed perfringolysin O and secreted it into the periplasm. The amino-terminal sequence of the pfo gene product was identical with that determined for perfringolysin O purified from C. perfringens, indicating that E. coli correctly removed the signal peptide during secretion. Purification of the pfo product was accomplished by high-resolution gel filtration and anion-exchange chromatography. Analysis of the pfo product by sodium dodecyl sulfate gel electrophoresis showed that it comigrated with authentic perfringolysin O; both had an estimated molecular weight of 54,000. Two-dimensional tryptic peptide maps of the pfo product and of authentic perfringolysin O purified from C. perfringens were identical. The hemolytic activity of the pfo product was similar to that of authentic perfringolysin O; one hemolytic unit (HU) of the cloned gene product or authentic perfringolysin O corresponded to approximately 1 ng or a hemolytic activity of 10(6) HU per mg.

Nucleotide sequence of the gene for perfringolysin O (theta-toxin) from Clostridium perfringens: significant homology with the genes for streptolysin O and pneumolysin.

The nucleotide sequence was determined for the gene encoding the thiol-activated cytolysin, perfringolysin O (theta-toxin), from Clostridium perfringens. The nucleotide-sequence-derived primary structure of perfringolysin O is 499 residues long and exhibits a 27-amino-acid signal peptide. The calculated molecular weight of the secreted (mature) form of perfringolysin O is 52,469. The deduced amino-terminal sequence of perfringolysin O is identical to that determined for purified perfringolysin O. Hydropathy analysis indicated that, except for the signal peptide, no major stretches of hydrophobic residues are present. Extensive amino acid sequence homology (65%) was detected with the low-molecular-weight form of streptolysin O, and a lesser amount (42%) was detected with pneumolysin. The nucleotide sequence of the perfringolysin O gene (pfo) exhibits approximately 60% homology with the streptolysin O gene (slo) and 48% homology with the pneumolysin gene (ply). All three toxins contain an identical region of 12 amino acids, which includes the essential cysteine of all three toxins. The location of these 12 residues was conserved in all three toxins when the primary sequences were aligned for maximum homology.

Purification and characterization of toxin B from Clostridium difficile.

Toxin B from Clostridium difficile was purified to homogeneity and characterized. Purification of toxin B was achieved by gel filtration, chromatography on two consecutive anion-exchange columns, and chromatography on a high-resolution anion-exchange column in the presence of 50 mM CaCl2. The molecular weight of toxin B was estimated to be 250,000 by denaturing sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and 500,000 by gel filtration. No subunits were apparent when the toxin was reduced and analyzed by SDS-PAGE. The estimated molecular weight of native toxin B indicated that dimers may form in solution. Toxin B was homogeneous by SDS-PAGE, native PAGE, and high-resolution anion-exchange chromatography. No secondary sequences were detected when the amino terminus of the toxin was sequenced, which also indicated that contaminating peptides were absent from the preparation. The amino terminus of toxin B was determined to be NH3-Trp-Leu-Val-Asn-Arg-Lys-Gln-Leu-Glu-Lys-Met-Ala-Asn-Val-ARg-Phe-Arg. One cytotoxic unit of toxin B was estimated to be 0.2 to 0.8 pg.

Molecular cloning and expression of gene fragments from corynebacteriophage beta encoding enzymatically active peptides of diphtheria toxin.

Two restriction fragments from corynebacteriophage beta vir tox+ that encode peptides similar to diphtheria toxin fragment A and the chain termination fragment, CRM45, have been cloned into Escherichia coli in plasmid pBR322. Clones containing the recombinant plasmids produced gene products that were active in catalyzing the ADP ribosylation of elongation factor 2 and were reactive with diphtheria toxin antiserum. Toxin-related peptides were found primarily in the periplasmic compartment and were degraded to nonimmunoreactive forms within 1 to 2 h of synthesis. The expression of both gene fragments appears to have originated from the diphtheria toxin promoter.

Transport and processing of staphylococcal enterotoxin B.

A larger, membrane-bound form of staphylococcal enterotoxin B was shown by in vivo pulse-chase analysis to be the kinetic precursor to extracellular enterotoxin B. Processing of the enterotoxin B precursor molecules can apparently occur either cotranslationally or posttranslationally. Subcellular fractionation of cells revealed that all of the precursor toxin was associated with the membrane fraction. Once processed and released from the membrane, it was transiently associated with the cell wall before being released into the extracellular environment. The cell-wall-associated enterotoxin B was completely resistant to protease treatment and to extraction by high- or low-salt solutions at 0 to 2 degrees C, although it could be easily released from the cell by removal of the cell wall with lysostaphin. These data imply that newly formed enterotoxin B may be temporarily sequestered in specialized regions that require cell wall integrity before being released into the extracellular environment.

The propeptide of Clostridium septicum alpha toxin functions as an intramolecular chaperone and is a potent inhibitor of alpha toxin-dependent cytolysis.

Clostridium septicum alpha toxin is activated by a proteolytic cleavage at Arg-398 in its carboxy terminus, which yields a 41.3-kDa cytolytically active toxin and a 5.1-kDa propeptide. Studies were performed to determine when the propeptide dissociated from the toxin after proteolytic activation of the protoxin (AT(pro)) and to demonstrate the chaperone activity of the propeptide. The propeptide was found to remain associated with the toxin after activation with trypsin (AT(act)) when analysed by gel filtration or affinity chromatography of a polyhistidine-tagged derivative that contained the polyhistidine tag on the propeptide. The affinity of the propeptide for the toxin was decreased significantly when a mutation was introduced in which Val-400 was converted to a cysteine residue. This mutation destabilized the interaction of the propeptide with the toxin and the propeptide was found to dissociate from the toxin under the same gel-filtration conditions used for the wild-type toxin. The separation of the propeptide in the V400C mutant did not affect the cytolytic activity of the toxin and therefore the propeptide was not necessary for cytolytic activity. These data suggested that the propeptide did not dissociate from the main body of the toxin after proteolysis. Further analysis demonstrated that purified propeptide was a potent inhibitor of alpha toxin activity, which inhibited the oligomerization of alpha toxin into a functional pore. These data suggest that the propeptide does not participate in the final oligomerized complex and that oligomerization appears to displace the propeptide from AT(act). The importance of the propeptide to the solution stability of alpha toxin was also demonstrated. When AT(pro) was activated in solution with trypsin a significant level (approximately 50%) of inactive aggregate formed. This aggregate, which could be removed by centrifugation at 14,000 x g, was made up of both SDS-sensitive and -resistant aggregates, suggesting that a variety of inactive aggregates formed when the monomers interacted in solution. Significantly higher levels of haemolytic activity (approximately 16-fold) were observed when alpha toxin was proteolytically activated after membrane binding instead of in solution. These results support the role of the propeptide as an intramolecular chaperone that stabilizes the monomeric AT(pro) and shuttles it to the membrane where it is activated by protease, oligomerizes into a pre-pore complex and forms a pore. The data suggest that oligomerization of the toxin displaces the propeptide from the monomer form of alpha toxin and that the propeptide does not participate in, and is not necessary to, the final cytolytic complex.

Purification and partial characterization of a putative precursor to staphylococcal enterotoxin B.

A putative precursor to staphylococcal enterotoxin B (SEB) has been identified as a component of purified membranes from Staphylococcus aureus S6. Agarose gel immunodiffusion analysis of the solubilized membranes demonstrated an immunoreactive protein that formed complete lines of identity with purified extracellular SEB. This putative precursor (pSEB) also had a different electrophoretic mobility from that of extracellular SEB when analyzed by immunoelectrophoresis. When membrane proteins from S6 were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and then transferred to nitrocellulose sheets and probed with I-125 labeled, affinity-purified anti-SEB, the pSEB band was identified. The pSEB was approximately 3,500 daltons larger than extracellular SEB. This component was purified by immunoprecipitation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Two-dimensional peptide maps of the putative SEB precursor revealed that most of the tryptic peptides were identical to those of mature extracellular SEB. When purified membranes of other SEB+ (DU4916 and 10-275) and SEB- (RN450, RN451, S6R, and FR1100) S. aureus strains were analyzed by the nitrocellulose blot procedure, only the SEB+ strains contained this putative SEB precursor on their membranes.

Evidence that Clostridium perfringens theta-toxin induces colloid-osmotic lysis of erythrocytes.

Clostridium perfringens theta-toxin was shown to lyse target erythrocytes by a colloid-osmotic mechanism. Analysis showed the onset of lysis of erythrocytes by theta-toxin could be temporarily stabilized with 0.3 M sucrose. Flow cytometry analysis of the size distribution of theta-toxin-treated erythrocytes showed swelling of the erythrocytes prior to lysis.

Kinetic aspects of the aggregation of Clostridium perfringens theta-toxin on erythrocyte membranes. A fluorescence energy transfer study.

Fluorescence resonance energy transfer was used to monitor aggregation kinetics of the "thiol-activated" cytolysin (perfringolysin O (PFO) or theta-toxin) of Clostridium perfringens on erythrocyte membranes. PFO was labeled with the isothiocyanate derivatives of either fluorescein or tetramethylrhodamine. No detectable change in the hemolytic activity of PFO was detected after modification with either fluorophore at a ratio of 1:2 fluorophore molecules/cytolysin molecule. Fluorescence energy transfer (FET) between the donor (fluorescein.PFO or PFOD) and the acceptor (tetramethylrhodamine.PFO or PFOA) was detected by both quenching of donor fluorescence (520 nm) and by enhancement of acceptor fluorescence (575 nm) upon aggregation of labeled cytolysin molecules. FET was only observed when PFOD and PFOA were incubated in the presence of membranes. FET was not observed when PFOD and PFOA were incubated in a membrane-free solution or when unlabeled toxin was substituted for PFOA. FET was also found to be temperature-dependent. The temperature-dependent rates of change in FET upon mixing labeled toxin with erythrocyte membranes proceeded without a lag phase and displayed an activation energy of 18.7 kcal/mol. At all temperatures aggregation of PFO was virtually complete before the onset of hemolysis, the latter exhibiting a distinct lag phase. The lag period before onset of hemolysis was temperature-dependent and exhibited an activation energy of 23.2 kcal/mol. These results suggest that the aggregation of membrane-associated PFO is necessary to initiate the hemolytic process, and the lag phase which occurs before onset of hemolysis reflects the kinetics of PFO monomer to polymer conversion.

Isolation of a tryptic fragment from Clostridium perfringens theta-toxin that contains sites for membrane binding and self-aggregation.

Trypsin cleaves Clostridium perfringens theta-toxin (perfringolysin O or PFO) at a single site between residues 303 and 304 (Ohno-Iwashita, Y., Iwamoto, M., Mitsui, K., Kawasaki, H., and Ando, S. (1986) Biochemistry 25, 6048-6053; Tweten, R. K. (1988b) Infect. Immun. 56, 3228-3234) and yields an amino-terminal fragment of 30,208 Da (T1) and a carboxyl-terminal fragment of 22,268 Da (T2). Both peptides were purified by reverse phase chromatography of trypsin-nicked PFO. Neither peptide retained hemolytic activity. Peptide T1 had no apparent effect on the hemolytic activity of PFO, whereas T2 was found to inhibit the hemolytic activity of PFO and was analyzed further. The order of binding of T2 and PFO to membranes did not alter the inhibitory effect of T2 on PFO-induced hemolysis, indicating that competitive binding by T2 for PFO membrane binding sites was not the basis for the observed inhibition. Further analysis showed that T2 could inhibit membrane-dependent fluorescence energy transfer (FET) between PFO molecules labeled with fluorescein (fluorescent donor) or tetramethylrhodamine (fluorescent acceptor). This provided evidence that T2 could complex with PFO. T2 was also found to be incapable of self-aggregation (as opposed to PFO), since preincubation of T2 with either erythrocytes or erythrocyte ghost membranes did not affect the T2-dependent inhibition of hemolysis or FET. These data indicate that T2 inhibits PFO-dependent hemolysis by forming a complex with PFO, which inhibits aggregation and that the membrane binding site and a single aggregation site remain intact on T2.

Purification and characterization of the lethal toxin (alpha-toxin) of Clostridium septicum.

Clostridium septicum lethal (alpha-toxin) was purified and found to be a basic protein (pI 8.4) of approximately 48 kDa that is both lethal and hemolytic. The alpha-toxin had a hemolytic activity of approximately 2 x 10(7) hemolytic units per mg and a 50% lethal dose of approximately 10 micrograms/kg of body weight for mice. The alpha-toxin formed concentration-dependent, sodium dodecyl sulfate-resistant aggregates of approximately 230 kDa. Mice immunized with alpha-toxin showed a significant increase in survival time over mock-immunized mice when challenged with C. septicum. Rabbit polyclonal antibody was generated against the purified toxin and was used to confirm that toxin with the same molecular weight was present in seven different C. septicum isolates. No proteins in the supernatants from cultures of Clostridium perfringens, Clostridium histolyticum, Clostridium chauvoei, or Clostridium difficile were found to react with the C. septicum alpha-toxin-specific antibody.

Generation of a membrane-bound, oligomerized pre-pore complex is necessary for pore formation by Clostridium septicum alpha toxin.

Low-temperature inhibition of the cytolytic activity of alpha toxin has facilitated the identification of an important step in the cytolytic mechanism of this toxin. When alpha toxin-dependent haemolysis was measured on erythrocytes at various temperatures it was clear that at temperatures < or = 15 degrees C the haemolysis rate was significantly inhibited with little or no haemolysis occurring at 4 degrees C. Alpha toxin appeared to bind to and oligomerize on erythrocyte membranes with similar kinetics at 4 degrees C and 37 degrees C. The slight differences in these two processes at 4 degrees C and 37 degrees C could not account for the loss of cytolytic activity at low temperature. At 4 degrees C alpha toxin neither stimulated potassium release from erythrocytes nor formed pores in planar membranes. In contrast, at temperatures > or = 25 degrees C both processes proceeded rapidly. Pores that were opened in osmotically stabilized erythrocytes could not be closed by low temperature. Therefore, low temperature appeared to prevent the oligomerized complex from forming a pore in the membrane. These data support the hypothesis that alpha toxin oligomerizes into a membrane-bound, pre-pore complex prior to formation of a pore in a lipid bilayer.

Transport and processing of staphylococcal alpha-toxin.

Two larger precursors to staphylococcal alpha-toxin were identified and partially characterized. Both precursor proteins were present on the cell membrane at very low levels and appeared to be rapidly processed to the mature form. Dinitrophenol inhibited processing such that the two precursors accumulated in the membranes, whereas little extracellular (mature) alpha-toxin is formed. The peptide maps of the 35S-labeled peptides from extracellular alpha-toxin and the two precursors were almost identical. The larger precursor protein contained four additional peptides and the smaller precursor protein contained three additional peptides not found in the extracellular toxin.