Cleavage of a C-terminal Peptide Is Essential for Heptamerization of Clostridium perfringens ε-Toxin in the Synaptosomal Membrane*

Activation of Clostridium perfringens ε-protoxin by tryptic digestion is accompanied by removal of the 13 N-terminal and 22 C-terminal amino acid residues. In this study, we examined the toxicity of four constructs: an ε-protoxin derivative (PD), in which a factor Xa cleavage site was generated at the C-terminal trypsin-sensitive site; PD without the 13 N-terminal residues (ΔN-PD); PD without the 23 C-terminal residues (ΔC-PD); and PD without either the N- or C-terminal residues (ΔNC-PD). A mouse lethality test showed that ΔN-PD was inactive, as is PD, whereas ΔC-PD and ΔNC-PD were equally active. ΔC-PD and ΔNC-PD, but not the other constructs formed a large SDS-resistant complex in rat synaptosomal membranes as demonstrated by SDS-polyacrylamide gel electrophoresis. When ΔNC-PD and ΔC-PD, both labeled with 32P and mixed in various ratios, were incubated with membranes, eight distinct high molecular weight bands corresponding to six heteropolymers and two homopolymers were detected on a SDS-polyacrylamide gel, indicating the active toxin forms a heptameric complex. These results indicate that C-terminal processing is responsible for activation of the toxin and that it is essential for its heptamerization within the membrane.

⑀-Toxin produced by Clostridium perfringens types B and D is the third most potent clostridial toxin after botulinum and tetanus toxins, and is responsible for the pathogenesis of fatal enterotoxemia in domestic animals caused by the organisms (1). This toxin exhibits toxicity toward neuronal cells via the glutamatergic system (2,3) or extravasation in the brain (4) after infection of experimental animals. It has been suggested to be a pore-forming toxin based on the following observations. (i) ⑀-Toxin can form a large complex in the membrane of MDCK 1 cells, and it permeabilizes them (5,6); (ii) the large complex formed by ⑀-toxin is not dissociated by SDS treatment (6), which is a common feature of pore-forming toxins such as aerolysin (7), Clostridium septicum ␣-toxin (8), and Pseudomonas aeruginosa cytotoxin (9); and (iii) the CD spectrum of ⑀-toxin shows it mainly consists of ␤-sheets (10), as is characteristically observed for pore-forming ␤-barrel toxins.
The structures of many bacterial pore-forming toxins or toxin components such as perfringolysin O (11), Bacillus thuringiensis ␦-toxin (12), aerolysin (13), staphylococcal ␣-toxin (14), and protective antigen of anthrax toxin (15) have been determined. These pore-forming toxins are believed to undergo a drastic conformational change upon interaction with a membrane. Since these toxins are inserted into the cytoplasmic membrane without the aid of other proteins such as chaperones or the translocation machinery, characterization of their metamorphosis has been regarded as a novel means for studying membrane-protein interactions (16). A characteristic feature of ⑀-toxin is its potent neurotoxicity, which is not seen for the structurally well defined pore-forming toxins. Thus, it could serve as a useful tool for extending our knowledge of how proteins gain entry into a membrane. Another characteristic feature of ⑀-toxin is that activation of the inactive precursor form (⑀-protoxin) by proteases such as trypsin, chymotrypsin (17), and -protease (18,19) is accompanied by removal of both N-and C-terminal peptides. In a previous study, we determined the N/C termini of ⑀-toxin activated by trypsin, trypsin plus chymotrypsin, and -protease to be Lys-14/Lys-274, Lys-14/Tyr-267, and Met-11/Tyr-267, respectively (19).
This study aimed to answer the following questions. (i) Which peptide(s), i.e. the N-, or C-, or both terminal peptides, regulates the activity of ⑀-toxin; ii) can the activated toxin form a large complex in the rat synaptosomal membrane; and iii) how many toxin monomers are present in the membrane complex?

EXPERIMENTAL PROCEDURES
Construction of Expression Vectors-Escherichia coli NovaBlue (Novagen, Madison, WI) was used for the construction of all recombinant plasmids. DNA fragments encoding ⑀-protoxin and an N-terminally truncated form of it were obtained by PCR amplification using total DNA from C. perfringens type B NCIB 10691 (19) as the template and the following pairs of synthetic primers: 5Ј-GGCCAAGGAAATATCTA-ATACAGTATCTAATGAA-3Ј (etx 1F primer) and 5Ј-GAATTCTTATTT-TATTCCTGGTGCCTTAATAGAAAG-3Ј (etx 1R primer) for ⑀-protoxin (Lys-1 to Lys-296); and 5Ј-GGCCAAAGCTTCTTATGATAATGTA-GATACATTA-3Ј (etx 2F primer) and etx 1R primer for ⌬ N-⑀-protoxin (Lys-14 to Lys-296). These DNA fragments were cloned into the pT7Blue T-vector (Novagen), and then inserted into the MscI and EcoRI sites of pET22b(ϩ) (Novagen) so that recombinant toxins could be directed to the periplasmic space of E. coli by a signal peptide encoded in the vector. The resultant plasmids, which expressed ⑀-protoxin and ⌬N-⑀-protoxin, were designated as pEP1 and pEN1, respectively.
To construct an ⑀-protoxin derivative (PD) and an N-terminally truncated form of it (⌬N-PD), in which a factor Xa recognition sequence, IEGR, was inserted between Lys-273 and Lys-274 of ⑀-protoxin, sequential PCR amplification was carried out using pEP1 as the template and the following pairs of primers: 5Ј-TAATACGACTCACTATAGGG-3Ј (T7 primer) and 5Ј-TTTACGACCTTCGATTTTATCTACAGGTATTAC-ATATTCTTG-3Ј (antisense, sequence corresponding to IEGR is underlined); and 5Ј-GATAAAATCGAAGGTCGTAAAGAAAAAAGTAAT-GATTCAAAT-3Ј (sense, sequence corresponding to IEGR is underlined) and 5Ј-GCTAGTTATTGCTCAGCGGTGG-3Ј (T7ter primer). The resulting PCR products were used as mixed templates, and T7 and T7ter were used as primers for a second PCR amplification. The specific PCR products were digested with AvrII and EcoRI, and then cloned into pEP1 and pEN1. The resultant plasmids, which expressed PD and ⌬N-PD, were named pEP4 and pEN4, respectively.
The MscI-NotI insert DNA fragments from pEP4 and pEN4 were subcloned into the SmaI and NotI sites of pBluescript II KSϩ (Stratagene, La Jolla, CA). The EcoRI fragments from these plasmids were inserted into pGEX-2TK (Amersham Pharmacia Biotech, Uppsala, Sweden). The resultant plasmids, named pEP9 and pEN9, enabled the expression of GST-PD and GST-⌬N-PD fusion proteins, respectively, which contained a pentapeptide recognized by the cyclic-AMP-dependent protein kinase (20). The accuracy of all the final DNA constructions was confirmed by DNA sequencing.
Expression and Purification of ⑀-Protoxin Derivatives-Transformants of E. coli BL21(DE3)pLysS (Novagen) carrying plasmids pEP1, pEN1, pEP4, and pEN4 were used to prepare ⑀-protoxin, ⌬N-⑀-protoxin, PD, and ⌬N-PD, respectively. Each transformant was grown in 1 liter of LB broth supplemented with ampicillin (100 g/ml) and chloramphenicol (33.4 g/ml) to an optical density at 600 nm of 0.7. Isopropyl ␤-D-thiogalactopyranoside was added to a final concentration of 1 mM, and then the cultures were grown for another 3.5 h. The cells were collected and treated with polymyxin B to obtain the periplasmic fraction, as described previously (21). Proteins in this fraction were precipitated with ammonium sulfate (60% saturation), dialyzed against 5 mM sodium phosphate buffer (pH 7.0), and then applied to a DEAE-Sephadex A-25 column (1.5 ϫ 12 cm; Amersham Pharmacia Biotech), which had been equilibrated with the buffer used for dialysis. The flowthrough fraction was collected and dialyzed against 50 mM Tris-HCl (pH 7.5) containing 1 M ammonium sulfate. This fraction was then subjected to hydrophobic interaction high performance liquid chromatography on a Phenyl-Superose HR 5/5 column (1 ml; Amersham Pharmacia Biotech). Proteins were eluted with a 20-ml linear gradient of ammonium sulfate (1-0 M) in 50 mM Tris-HCl (pH 7.5). Fractions containing ⑀-protoxin or its derivatives, as determined by SDS-PAGE, were pooled and dialyzed against 20 mM Tris-HCl (pH 7.5). Purified toxins were stored at Ϫ80°C.
C-terminally truncated PD (⌬C-PD), and N-and C-terminally truncated PD (⌬NC-PD) were obtained from purified PD and ⌬N-PD, respectively, by cleaving them at the factor Xa-sensitive site. Digestion with factor Xa (restriction grade, Novagen) and its removal on a Xarrest-agarose column (Novagen) were carried out according to the manufacturer's instructions. GST-PD and GST-⌬N-PD were prepared from cultures of E. coli BL21 carrying pEP9 and pEN9, respectively, which were grown and induced as described above. The fusion proteins were purified by affinity chromatography on a glutathione-Sepharose 4B column (Amersham Pharmacia Biotech) according to the manufacturer's instructions.
Mouse Lethality Test and Cytotoxicy Assay-The lethality of ⑀-protoxin and its derivatives toward mice were determined as described previously (19). A group of 15 male ddY mice weighing 37 g each were injected intravenously with 0.25 ml of the toxin solution, and deaths occurring within 24 h were recorded.
MDCK cells were cultured in Eagle's minimum essential medium containing Earle's salts, penicillin (100 units/ml), and streptomycin (100 g/ml), supplemented with 10% heat-inactivated fetal bovine serum, in a cell culture incubator (under 5% CO 2 at 37°C). Freshly trypsinized cells were cultured in 96-well microculture plates for 48 h to give monolayers. The medium was exchanged for 200 l of minimum essential medium, followed by the addition of 50 l of PBS or PBS containing one of the purified ⑀-protoxin derivatives. After a 6-h incubation, the cells in each well were washed with PBS containing Mg 2ϩ and Ca 2ϩ . Cell viability was determined by the 3-(4, 5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium inner salt (Promega, Madison, WI) conversion assay. The absorbance, which is proportional to the number of viable cells, was read at 490 nm using an enzyme-linked immunosorbent assay plate reader. Percentage of cell viability was calculated as follows: the mean absorbance value of a toxin-group/that of a control.
Preparation of Rat Brain Synaptosomal Membranes-Synaptosomes were prepared from rat brains as reported previously (22) with some modifications. Briefly, rat brains (7.5 g) were homogenized in nine volumes (w/v) of buffer A (10 mM Tris-HCl (pH 7.0), 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, 5 mM benzamidine HCl, 1 M pepstatin A, 20 M leupeptin) containing 0.32 M sucrose, by means of 10 passes through a glass homogenizer with a Teflon plunger. The homogenate was centrifuged at 1,000 ϫ g for 10 min, and the supernatant was centrifuged at 15,000 ϫ g for 30 min. Then the pellet was re-suspended in a small volume of buffer A containing 0.32 M sucrose, loaded onto a 0.8 M sucrose solution layered on a 1.1 M sucrose solution in a centrifuge tube, and finally centrifuged at 63,000 ϫ g for 2 h. The synaptosomal membrane fraction at the 0.8 M/1.1 M interface was isolated and collected by centrifugation at 100,000 ϫ g for 1 h. The fraction was re-suspended in buffer A containing 0.14 M NaCl. Preparation of Radiolabeled ⑀-Protoxin Derivatives-⑀-Protoxin derivatives were radiolabeled with 125 I as follows. The purified toxin derivatives (8 g) were incubated with 36.2 GBq of 125 I (643.8 GBq/mg; PerkinElmer Life Sciences) and IOAD-BEADS iodination reagent (Pierce) in 60 l of PBS for 15 min at room temperature. The radioiodinated proteins were separated from free iodine by gel filtration on a Sephadex G-25 column (0.75 ϫ 5.5 cm; Amersham Pharmacia Biotech), equilibrated with Tris-buffered saline (TBS; 150 mM NaCl, 50 mM Tris-HCl (pH 7.4)). The specific activities of 125 I-labeled PD, ⌬N-PD, ⌬C-PD, and ⌬NC-PD were 430, 630, 540, and 390 kcpm/g of protein, respectively.
Phosphorylation of ⑀-protoxin derivatives with 32 P was performed as follows. Approximately 300 g of GST-PD or GST-⌬N-PD were loaded onto a glutathione-Sepharose 4B column (bed volume, 100 l; Amersham Pharmacia Biotech), and then phosphorylated using the catalytic subunit of bovine heart protein kinase (Sigma) and [␥-32 P]ATP (167 TBq/mmol; ICN Biochemicals, Costa Mesa, CA) according to the protocol recommended by the manufacturer. The specific activities of 32 Plabeled GST-PD and GST-⌬N-PD were about 21 and 48 kcpm/g of protein, respectively. [ 32 P]⌬C-PD and [ 32 P]⌬NC-PD were prepared from 32 P-labeled GST-PD and GST-⌬N-PD, respectively, by cleavage at the factor Xa-specific site within the C-terminal portion and also at a nonspecific but sensitive site in a GST linker sequence (see "Results").
Formation of an SDS-resistant Complex in Synaptosomal Membranes-Fifteen nanograms of each 125 I-labeled toxin derivative were added to 2.2 g of synaptosomal membranes in 12 l of TBS containing 0.1% BSA. After incubation at 37°C for 90 min, the reaction mixture was centrifuged at 17,000 ϫ g for 5 min. The membrane pellet was washed three times with 200 l of TBS at 4°C, and then dissolved by heating in SDS sample buffer (62.5 mM Tris-HCl (pH 6.8), 2% ␤-mercaptoethanol, 1% SDS, 0.01% bromphenol blue) at 95°C for 5 min. Samples were electrophoresed on a SDS-PAGE gel, followed by exposure to an imaging plate (Fuji Photo Film, Kanagawa, Japan) for autoradiography.
Heteromeric polymerization of ⑀-protoxin derivatives was carried out using [ 32 P]⌬C-PD and [ 32 P]⌬NC-PD in various ratios. A total of 20 -40 ng of each labeled toxin derivative was added to 2.2 g of synaptosomal membranes in 10 l of TBS. After incubation at 37°C for 90 min, SDS sample buffer was added to the reaction mixtures, followed by heating at 95°C for 5 min. To separate heteromeric polymers, SDS-PAGE was performed on a 5% polyacrylamide gel prepared in a DNA sequencing gel apparatus (800 ϫ 170 ϫ 0.4 mm; Bio-Rad) at 15 mA and room temperature for 20 h.
Analytical Methods-N-terminal amino acid sequencing of the purified toxins was performed with a protein sequencer (model 492 Procise; Applied Biosystems), as described previously (23). Protein concentrations were measured using Pierce bicinchoninic acid protein assay reagent (Pierce) with bovine serum albumin as the standard.
SDS-PAGE was performed as described by Laemmli (24). High and low molecular weight marker proteins were obtained from Amersham Pharmacia Biotech.
Mass determinations were carried out with a MALDI-TOF mass spectrometry instrument (Voyager DE PRO; Applied Biosystems) in the positive linear mode at an acceleration voltage of 15 kV with delayed extraction. Samples were mixed with sinapinic acid (10 mg/ml in 0.1% trifluoroacetic acid, 50% acetonitrile), which was used as the absorbing matrix. Horse skeletal muscle myoglobin was used as the standard.
Circular dichroism (CD) spectra in the far UV region (260 -205 nm) were obtained with a J720WI spectropolarimeter (Jasco, Tokyo, Japan) using a cell with a 1-mm light path at 25°C. All samples were dialyzed against PBS at 4°C, and the dialysate was also measured as the background. The measurements were repeated four times, and the results were averaged.

RESULTS
Construction and Characterization of ⑀-Protoxin Derivatives-Both an N-terminal peptide of 13 amino acids and a C-terminal peptide of 22 amino acids are removed upon tryptic activation of ⑀-protoxin. To determine which peptide inactivates ⑀-protoxin, we constructed recombinant ⑀-protoxin derivatives without the N-or C-terminal peptide using an E. coli expression system. The recombinant ⑀-protoxins with and without the N-terminal peptide were inactive but activated by treatment with trypsin plus chymotrypsin, as shown by the mouse lethality test (Table I). The LD 50 of their activated forms was 50 ng/kg of body weight, which coincided with the value reported for ⑀-toxin from C. perfringens cultures. When the recombinant ⑀-protoxin without the C-terminal peptide was expressed in E. coli, it precipitated as inclusion bodies, which could be solubilized with 6 M urea, but it could not be recovered in a soluble form after dialysis (data not shown). Thus, we constructed two ⑀-protoxin derivatives, with and without the N-terminal peptide, respectively, in which IEGR, a coagulation factor Xa recognition sequence, was inserted between Lys-273 and Lys-274 (trypsin cleavage site), as shown in Fig. 1A. This construction enabled the recovery of these proteins, i.e. PD and ⌬N-PD, from the periplasmic space and also allowed cleavage at the relevant site with factor Xa.
PD and ⌬N-PD were purified from E. coli cultures to homogeneity, as demonstrated by SDS-PAGE (Fig. 1B). Another two derivatives, ⌬C-PD and ⌬NC-PD, were obtained by treatment of the purified PD and ⌬N-PD, respectively, with factor Xa (Fig.  1B). The identities of all the constructs were confirmed by nucleotide sequencing of the recombinant plasmids, N-terminal amino acid sequencing (data not shown), and molecular mass determination by MALDI-TOF mass spectrometry (Fig.  1A). CD analysis revealed that all the derivatives gave a negative ellipticity band at 215 nm, indicating a predominance of ␤-structure. Some minor differences were observed in the CD spectra, which may be due to a minor change in the conformation arising from the insertion of the factor Xa recognition sequence or removal of the N-and/or C-terminal sequence(s). However, there were no remarkable differences in the CD spectra, strongly suggesting neither the insertion nor the removal(s) affects the overall structure of ⑀-protoxin (Fig. 2).
Toxicity of ⑀-Protoxin Derivatives-The toxicity of the ⑀-protoxin derivatives was determined by means of a mouse lethality test (Table I). The LD 50 of PD with an insertion of the IEGR sequence was 31,000 ng/kg body weight, slightly lower than that previously reported for ⑀-protoxin purified from C. perfringens cultures (70,000 ng/kg of body weight). The LD 50 of ⌬N-PD was the same as that of PD, being higher than that of ⌬C-PD by a factor of about 60. The LD 50 of ⌬NC-PD was almost the same as that of ⌬C-PD, although both were slightly higher than that reported for trypsin-activated ⑀-toxin (320 ng/kg). This might be due to the difference in the C-terminal region: ⌬NC-PD and ⌬C-PD have four extra amino acids after the trypsin cleavage site. To eliminate the possibility that N-terminal processing has an additional activating effect, the toxicities of the four derivatives were determined by means of an in vitro assay method using MDCK cells, which is less sensitive but more accurate than the mouse lethality test (Fig. 3). ⌬C-PD and ⌬NC-PD exhibited almost the same toxicity toward MDCK cells, and there was no significant difference in the 50% cytotoxicity dose between the two derivatives. No cytotoxicity was detectable for ⌬N-PD and PD at the highest concentration (600 ng/ml) used in this study. These results clearly indicate that only cleavage at the C-terminal region is responsible for activation of ⑀-protoxin.
Large Complex Formation in Synaptosomal Membranes-Activated ⑀-toxin has been shown to form an SDS-resistant 174-kDa complex in the membranes of MDCK cells (6), and the complex has also been reported in brain homogenates incubated with the toxin (5). Such a complex may also be formed in synaptosomal membranes, which could be responsible for the neuronal cell death in the hippocampus observed in ⑀-toxin injected mice (2). In order to assess this possibility, membranes were incubated with 125 I-labeled toxin derivatives and ana-  lyzed by SDS-PAGE (Fig. 4A). In the samples incubated with 125 I-labeled ⌬C-PD and ⌬NC-PD, large SDS-resistant complexes were detected, with apparent molecular masses of ϳ200 and 180 kDa, respectively. No SDS-resistant complex was detected in membranes incubated with 125 I-labeled PD and ⌬N-PD. When equal amounts of 125 I-labeled ⌬C-PD and unlabeled ⌬NC-PD or vice versa were incubated with the membranes, a smeared band corresponding to 200 -180-kDa complexes was detected on a SDS-PAGE gel (Fig. 4B), indicating that both active forms are equally capable of forming the complex.
Heptamerization of Active ⑀-Toxin-The molecular masses of ⌬C-PD and ⌬NC-PD are 31 and 29 kDa, respectively. Therefore, the 200-or 180-kDa large complex should consist of five toxin molecules and an intrinsic membrane protein, or at maximum seven toxin molecules, if they are unprocessed in the membrane. The result of an experiment involving a mixture of the two active forms suggested that the smeared band corresponds to heteromultimers consisting of the two forms in various ratios, and that the number of toxin molecules per complex is independent of the ratio. If this is the case, complexes of different molecular sizes will form, which can be separated to determine the number of toxin molecules in a complex. The heteromeric complexes formed by 125 I-labeled toxins were not well separated even with a SDS-PAGE gel system for large polypeptides. To circumvent this problem, we constructed GST fusion proteins containing a protein kinase recognition site (RRXSV) at the N termini of the two toxin derivatives (Fig. 5A). These constructs were cleaved by factor Xa not only at the C-terminal specific site but also between the two arginine residues in the RRXSV sequence. Thus, the fusion proteins were purified (Fig. 5B, lanes 1 and 2), phosphorylated, and cleaved with factor Xa (Fig. 5B, lanes 3 and 4). The resultant phosphorylated ⌬C-PD and ⌬NC-PD were designated as P-⌬C-PD and P-⌬NC-PD, respectively. The identities of these protein derivatives were verified by N-terminal amino acid sequencing (data not shown), and molecular mass determination by MALDI-TOF mass spectrometry (Fig. 5A). There was no significant difference in mouse lethality between these two constructs, although their lethalities were slightly lower than those of ⌬N-PD and ⌬NC-PD (Table I). They also retained the ability to form an SDS-resistant complex in synaptosomal membranes (Fig. 5C). The two 32 P-labeled proteins were mixed in appropriate ratios, incubated with membranes, and then electrophoresed on a 5% SDS-PAGE gel in a DNA sequencing gel apparatus (Fig. 5D). Six distinct bands were observed between those of the P-⌬C-PD and P-⌬NC-PD homopolymers, indicating that the large complex contains seven monomers.
Inhibition of Complex Formation by ⑀-Protoxin-The effect of ⑀-protoxin on the ability of ⌬C-PD to form an SDS-resistant complex was examined. Complex formation was inhibited in a dose-dependent fashion by ⑀-prototoxin and completely blocked   by a 10-fold excess of ⑀-protoxin (Fig. 6A). Such dose-dependent inhibition by ⑀-protoxin was also observed for complex formation by ⌬NC-PD (Fig. 6B). The inhibition of SDS-resistant complex formation seems to be due to competition for receptor binding, but it may also involve a subsequent complex formation step. The labeled toxins were each detected as a monomer in the membranes to a similar extent, independent of the amount of ⑀-protoxin added. If they represent monomers dis-sociated from a receptor but associated with the membrane, ⑀-protoxin would inhibit complex formation by these monomers. DISCUSSION This paper demonstrates that removal of the C-but not the N-terminal peptide is responsible for the activation of ⑀-toxin. In a previous study, we showed that ⑀-protoxin is cleaved not only at an N-terminal region but also a C-terminal region by trypsin, chymotrypsin, and -protease, and that ⑀-protoxin activated by such proteases differs in both its N and C termini (19). The different forms also differ in toxicity, with the trypsin plus chymotrypsin-activated ⑀-toxin being the most potent. Since the presence or absence of the N-terminal peptide did not affect toxicity, the potency of ⑀-toxin is likely to be determined solely by the difference in the C-terminal cleavage site. This paper also provides evidence of the ability of active ⑀-toxin to form a large complex in rat synaptosomal membranes. We previously demonstrated that the toxin is neurotoxic to the brains of mice and rats (2,3). The finding that the toxin forms a large complex in the synaptosomal membrane strongly suggests that the neurotoxicity is due to large complex formation, which also is implicated in membrane permeabilization and cell death of ⑀-toxin-treated MDCK cells.
Aerolysin has been proposed to form heptameric oligomers, based on the results of SDS-PAGE, scanning transmission electron microscopic mass measurement of single oligomers, and image analysis of two-dimensional crystals, which were obtained by reconstitution of purified aerolysin and E. coli phospholipids (13,25). To solve the problem that low resolution electron crystallography gave artifactual data for other poreforming toxins, Moniatte et al. (26) determined the molecular masses of oligomers formed in a solution by MALDI-TOF mass spectrometry, providing further evidence of heptamerization. However, analysis of the complex in biological membranes is a prerequisite for proving heptamerization and for elucidating its molecular mechanism, since aerolysin polymerizes at higher concentrations (27). Taking advantage of the fact that the N-terminally processed and unprocessed ⑀-toxins both form the large complex, we have demonstrated that seven monomers are present in the synaptosomal membrane complex. This is the first demonstration of a heptameric toxin complex in a biological membrane, and also supports the validity of the heptermerization suggested for aerolysin and other pore-forming toxins.
The inhibition of SDS-resistant complex formation by ⑀-pro-  toxin shown in this study can be explained simply by competition for receptor binding. Alternatively, ⑀-protoxin associated with the membrane may exert its inhibitory effect by sequestering active ⑀-toxin, making it unavailable to assemble into a productive complex. Whichever of these possibilities is true, the results unequivocally show the ability of ⑀-protoxin to bind to a receptor and its inability to form the productive large complex. Thus, it seems very likely that a receptor binding site is exposed on the surface of ⑀-protoxin, and that the C-terminal peptide sterically hinders exposure of a site or conformational change required for complex formation.
⑀-Protoxin exerts no toxicity unless it encounters proteases such as trypsin, chymotrypsin, and -protease. This may imply that the C-terminal peptide functions as an intramolecular chaperone. It should be noted that C-terminally truncated forms of ⑀-protoxin were produced as inclusion bodies, and that several attempts to obtain active forms through solubilization of the precipitate with urea or guanidine HCl and subsequent dilution or dialysis were unsuccessful. However, all ⑀-toxin constructs containing the C-terminal peptide were obtained in a soluble form. It has been reported that unfolding of proaerolysin in 7 M urea was reversible upon dilution in urea-free buffer (28). These observations suggest that these C-terminal peptides aid proper folding of toxin molecules. Another likely chaperonic function of the C-terminal peptide is to prevent the active ⑀-toxin from aggregating in solution. The 5.1-kDa Cterminal propeptide of C. septicum ␣-toxin has been shown to be associated with the toxin even after proteolytic activation, preventing the toxin from aggregating (8). Further study, of the association/dissociation of the C-terminal peptide with the active ⑀-toxin, and its effect on unfolding/refolding of the toxin, should be undertaken to elucidate the roles of the C-terminal peptide in the structure and function of ⑀-toxin.