Purification and Characterization of Procytotoxin of Pseudomonas aeruginosa

Cytotoxin of Pseudomonas aeruginosais a cytolytic toxin that forms a pore on the target membrane by oligomerizing into a pentamer. This toxin is produced as an inactive precursor (proCTX) and is converted to an active form by proteolytic cleavage at the C terminus. We purified proCTX to apparent homogeneity and characterized it in a comparison with the active toxin. ProCTX bound to the erythrocyte membrane but did not form an oligomer on the membrane, hence the lack of hemolytic activity in proCTX. Circular dichroic experiments showed that active and proCTX have similar β-sheet dominant structures. Intrinsic fluorescence analysis indicated that a molecule-buried tryptophan residue(s) of proCTX was exposed to the surface of the molecule as a result of conversion to the active form. In analytical gel filtration, chemical cross-linking, and analytical ultracentrifugation experiments, dimer to monomer conversion occurred with proteolytic activation.

Cytotoxin of Pseudomonas aeruginosa is a cytolytic toxin that forms a pore on the target membrane by oligomerizing into a pentamer. This toxin is produced as an inactive precursor (proCTX) and is converted to an active form by proteolytic cleavage at the C terminus. We purified proCTX to apparent homogeneity and characterized it in a comparison with the active toxin. ProCTX bound to the erythrocyte membrane but did not form an oligomer on the membrane, hence the lack of hemolytic activity in proCTX. Circular dichroic experiments showed that active and proCTX have similar ␤-sheet dominant structures. Intrinsic fluorescence analysis indicated that a molecule-buried tryptophan residue(s) of proCTX was exposed to the surface of the molecule as a result of conversion to the active form. In analytical gel filtration, chemical cross-linking, and analytical ultracentrifugation experiments, dimer to monomer conversion occurred with proteolytic activation.
Pseudomonas aeruginosa, a Gram-negative opportunistic pathogen, causes severe infections in compromised and immunosuppressed hosts. The bacterium produces various toxic substances that contribute to the virulence of this opportunistic pathogen (1,2). Some strains of P. aeruginosa produce a cytolytic toxin, cytotoxin (CTX), which is cytotoxic to a wide range of eukaryotic cells (3,4). The gene encoding the toxin (ctx) is carried by bacteriophages, and lysogenization of the phages converts CTX non-producing P. aeruginosa strains to CTX producers (5,6). The acquisition of the ctx gene increases the virulence of the bacterium (7).
CTX has been considered a channel-forming toxin, generating a pore of 2 nm in diameter (8,9). Although morphological identification has yet to be made, we recently found that CTX forms an oligomer on the target membrane (10). This formation correlated with the cytolytic activity; thus, CTX may form a pore by oligomerization on the membrane, as is the case for several cytolytic channel-forming toxins (11,12). Because the molecular size of the oligomer was estimated to be 145 kDa, using SDS-PAGE 1 (10), it likely consists of a pentamerized CTX.
CTX produced by P. aeruginosa is inactive and requires proteolytic action for activation. Analyses of the ctx gene and an active toxin purified from the trypsin-treated crude extract of P. aeruginosa (PACTX) showed that CTX is produced as a precursor (procytotoxin (proCTX)) of 286 amino acids and is activated by proteolytic cleavage at the C terminus (13). The cleavage occurs at the carboxyl side of Arg 266 (13). Among the channel-forming toxins, aerolysin from Aeromonas spp. (12) and the alpha toxin of Clostridium septicum (14,15) also require proteolytic cleavages at the C-terminal regions for conversion from inactive precursors to active forms. In the case of proaerolysin, the C-terminal region was seen to mask hydrophobic patches of the toxin molecule and to inhibit oligomerization of the toxin on the membrane (12,16). Concerning proCTX, nothing is known of the function of the C-terminal sequence or what changes occur on the CTX molecule by the C-terminal cleavage.
Using a method developed for the purification of the recombinant CTX, we have now purified proCTX for the first time. Purified proCTX bound to the erythrocyte membrane but did not form the oligomer on the membrane. In circular dichroic experiments, significant changes in the secondary structure were not observed between the active CTX and proCTX. However, a striking difference was found in the oligomeric states of the proteins. We describe here a novel activation mechanism of the pore-forming toxin, involving a dimer-to-monomer conversion.

EXPERIMENTAL PROCEDURES
Materials-Trypsin (Type XIII: L-1-tosylamido-2-phenylethylchloromethyl ketone-treated), bovine serum albumin, and apomyoglobin were from Sigma. The molecular size standards for SDS-PAGE were obtained from Bio-Rad and New England Biolabs. All columns used for protein purification were purchased from Pharmacia Biotech.
Site-directed Mutagenesis and Overproduction of the Toxins in Escherichia coli-A codon for threonine at position 267 of the ctx gene was replaced with a premature stop codon by changing ACA to TAA. The mutagenesis was performed by the gapped duplex DNA method of Kramer and Frits (17) using a Mutan-G system (Takara). The mutagenized ctx gene was inserted into a broad host range expression plasmid, pMMB22 (18), to construct pMMB(⌬C20), using the same strategy as described for pMMB(CTX) for the overproduction of procytotoxin (13).
E. coli SM32 was transformed with pMMB(⌬C20) or pMMB(CTX) and grown at 37°C in Luria-Bertani broth supplemented with 50 g/ml ampicillin. When cell cultures reached an A 600 of 0.5, isopropyl-1-thio-␤-D-galactopyranoside was added at a final concentration of 1 mM, and incubation was continued for 4 h. The cells were collected by centrifugation, washed twice with saline, and stored at Ϫ80°C.
Purification Procedure-All purification procedures were performed at 4°C, except that fast protein liquid chromatography procedures (Pharmacia Biotech Inc.) were carried out at room temperature.
For purification of ⌬C20, the cells derived from a 10-liter culture were suspended in 400 ml of Buffer A (50 mM sodium phosphate buffer, pH 7.2) and disrupted by ten 10-s bursts of sonication with 50 s interval between each burst at 4°C, using a tip sonicator (Branson). The cell lysate was then centrifuged at 20,000 ϫ g at 4°C for 30 min. Because overexpressed ⌬C20 was aggregated in E. coli cells, more than 95% of ⌬C20 was recovered in the pellet fraction. The pellet was dissolved in 100 ml of Buffer A containing 2 M urea while stirring for 2 h. After removing unsolubilized materials by centrifugation (20,000 ϫ g at 4°C for 30 min), the soluble fraction was dialyzed four times against a 20-fold volume of Buffer B (Buffer A with 0.2 M NaCl). The dialyzate was clarified by centrifugation at 100,000 ϫ g for 1 h, and concentrated by ultrafiltration through a YM10 membrane (Amicon). The concentrated preparation was applied to a HiLoad 26/60 Superdex 200 pg column equilibrated with Buffer B by fast protein liquid chromatography. Proteins were eluted with the same buffer at a flow rate of 2 ml/min. Fractions with toxic activity were pooled and dialyzed against Buffer C (Buffer A containing 2 M NaCl). The dialyzate was applied to a phenyl-Superose HR 10/10 column equilibrated with Buffer C. Proteins were eluted with a linear NaCl gradient (2.0 -0 M) at a flow rate of 1 ml/min.
The procedure used for purification of proCTX was the same as that used for ⌬C20 up to the step of hydrophobic chromatography, except for two modifications: proCTX was solubilized with 3 M urea, and Buffer A containing 1 M NaCl was used for the gel filtration. After hydrophobic chromatography, fractions containing proCTX were subjected to a Mono Q HR 5/5 column equilibrated with Buffer A. Proteins were eluted with a linear NaCl gradient (0 -1.0 M) at a flow rate of 1 ml/min. During the purification, proCTX was monitored by immunoblotting with anti-CTX serum (4).
Purified proteins were stored at Ϫ20°C in Buffer A containing 20% (v/v) glycerol.
Measurement of Cytotoxic Activity and Oligomer Formation-Cytotoxic activities of the toxins were determined by means of a hemolytic activity assay (10). Oligomer formation of toxins on the rat erythrocyte membrane was detected by immunoblotting as described previously (10). Activation of proCTX was done by incubation with trypsin at 37°C for 60 min at 1:20 trypsin per proCTX (w/w) ratio. The reaction was quenched by adding phenylmethylsulfonyl fluoride (final concentration, 1 mM).
Preparation of Antisera and Affinity Purification of Antibody-Rabbit anti-⌬C20 serum was prepared as described (4). For the specific detection for proCTX, a 9-mer peptide corresponding to the C-terminal sequence of proCTX (LETRVRSAE) with a cysteine residue at the N terminus was synthesized by the Fmoc (9-fluorenylmethoxycarbonyl) strategy using a peptide synthesizer (Applied Biosystems Model 431A). The peptide (0.1 mg) was cross-linked to 0.1 mg of bovine serum albumin with m-maleimidobenzoyl-N-hydrosuccinimide ester and used to immunize a rabbit.
The antibody specific to the C terminus of proCTX was purified by affinity chromatography using 2-fluoro-1-methylpyridinium toluene-4sulfonate-activated cellulofine (Seikagaku) coupled with the C-terminal peptide. The antibody obtained reacted to proCTX and not to ⌬C20.
Immunoadsorption of Toxins with Immobilized Antibody-The affinity purified antibody was immobilized on Affi-Gel 10 (Bio-Rad) in coupling buffer (0.1 M HEPES, pH 8.0). Toxins (0.5 g) were incubated with 10 l of the antibody-coupled gel (1.1 g of IgG/l gel) in 40 l of binding buffer (coupling buffer with 0.1 M NaCl) for 1 h at 37°C. After unbound materials were separated by brief centrifugation, pellets were washed five times with the binding buffer (200 l). Bound materials were eluted with SDS loading buffer, and the bound and unbound fractions were analyzed by SDS-PAGE. For immunostaining, rabbit anti-⌬C20 serum and peroxidase-conjugated anti-rabbit IgG antibody specific to the Fc region (Promega) were used.
Measurement of Circular Dichroism (CD) spectra-CD spectra of toxins were measured at 25°C in a Jasco J-600 spectrometer with a quartz cell of 0.2 mm path length. The protein concentration was 150 g/ml in 20 mM sodium phosphate buffer, pH 7.2. Contents were calculated using the reference described by Yang et al. (19). Estimates of the secondary structure were made using a SSE-338 program (Jasco).
Measurement of Tryptophan Fluorescence-Fluorescence measurements were made using a Hitachi fluorometer F3000 with a quartz cuvette. The excitation wavelength was 290 nm, and the range of emission wavelength was from 300 to 400 nm. The protein concentration was adjusted to 0.6 M in Buffer B.
Mass Spectrometry-Samples for spectrometry were prepared according to Nakanishi et al. (20). Briefly, 33 pmol of toxins in 10 l of Buffer A was incubated with an equal volume of anti ⌬C20 serum for 3 h at room temperature, the samples were centrifuged, and the pellets were washed once with 0.6 ml of 0.9% NaCl and twice with 0.6 ml distilled water and lyophilized. Pellets were dissolved in 3.3 l of distilled water and mixed with an equal volume of saturated sinapinic acid matrix solution in 33% (v/v) acetonitrile/distilled water. Mass measurements were made on a matrix-assisted laser desorption ionization time-of-flight mass spectrometry (Voyager ELITE XL, PerSeptive Biosystems) according to the manufacturer's instructions. Bovine serum albumin and apomyoglobin were used as standards.
Analytical Gel Filtration Analysis-Analytical gel filtration was performed using a Superose TM12 HR 16/60 column equilibrated with Buffer B. Two hundred l of toxin solution (150 g/ml) was applied to the column and then eluted at a flow rate of 0.5 ml/min. Molecular weight standards used were ovotransferrin (76,000 -78,000), ovalbumin (45,000), carbonic anhydrase (30,000), and myoglobin (17,200).
Cross-linking-Purified proteins (50 g/ml in 50 mM triethanolamine, pH 8.5) were incubated with 2.5 mM or10 mM dimethyl suberimidate for 10 min at 37°C. Reaction was quenched by incubation with Tris-HCl, pH 8.0 (final concentration, 100 mM), and then SDS loading buffer was added. Aliquots of the reaction mixture were boiled and analyzed by SDS-PAGE.
Analytical Ultracentrifugation-Analytical ultracentrifugation was carried out using a Beckman Optima XL-A ultracentrifuge equipped with an optical scanning detector. Proteins in Buffer B at a concentration of 150 g/ml were analyzed. Data were collected in a Beckman An-60Ti rotor with double sector charcoal-filled Epon cells and quartz windows.
A sedimentation velocity run was carried out at 50,000 rpm at 25°C. Boundaries were recorded at 280 nm. Apparent sedimentation coefficients were standardized to water at 20°C.
For sedimentation equilibrium experiment, samples were brought to equilibrium at 4°C for 20 h at 24,000 rpm. Partial specific volumes for proCTX and ⌬C20 were calculated for amino acids composition, using data for specific volumes of amino acids as described by Edsall (21).
Other Methods-N-terminal amino acid sequences of the purified proteins were determined by Edman degradation using an automated protein sequencer (Applied Biosystems Model 476A). SDS-PAGE was performed as described by Laemmli (22). Proteins were stained with Coomassie Brilliant Blue R-250. Immunoblotting was done as described previously (13). Protein concentrations were assessed according to Lowry et al. (23) for unpurified materials and by absorbance at 280 nm for purified preparations. Absorption coefficients of ⌬C20 and proCTX (2.24 and 2.32, respectively) were determined from amino acid compositions of the proteins.

RESULTS
Purification of an Active Form of Toxin-We earlier reported a method for purification of an active form of CTX from the crude extract of P. aeruginosa (4). The procedure involved five steps, including trypsin treatment for activation of proCTX in the crude extract. Purification procedures described by other researchers were complicated and included an activation step (24,25). Because these methods were not applicable to the purification of CTX or its mutants expressed in E. coli, we tried to design a simple purification method for recombinant CTX.
For direct expression of an active form of CTX (⌬C20), we first introduced a premature stop codon at position of Thr 267 in the ctx gene by site-directed mutagenesis. The C terminus of ⌬C20 is Arg 266 , the same as the active toxin derived from trypsin treatment (13). Although ⌬C20 was overproduced in E. coli SM32, a Lon protease-deficient strain, most of the protein produced was insoluble. We attempted to solubilize the aggregated toxin using different concentrations of guanidine hydrochloride or urea (2, 3, 4, and 6 M). The toxin solubilized by guanidine hydrochloride at each concentration was reaggregated after dialysis against Buffer B. Here, a significant portion of the toxin solubilized by 2 M urea remained soluble after dialysis and the subsequent centrifugation at 100,000 ϫ g. Higher concentrations of urea reduced the amounts of soluble toxin obtained.
The toxin solubilized by 2 M urea was purified using HiLoad Superdex pg200 and phenyl-Superose columns. From 10 liters of cultured cells, 8.5 mg of ⌬C20 was obtained (Fig. 1A). The elution profiles of ⌬C20 on both columns were the same as those of PACTX. Purified ⌬C20 showed a wide range of cytotoxicity with specific activity similar to that of PACTX. As shown for PACTX (10), ⌬C20 formed an oligomer of 145 kDa on the rat erythrocyte membrane in a manner correlating with the hemolytic activity (Figs. 2 and 3). The N-terminal sequence of the purified ⌬C20 was identical to the sequence deduced from the ctx gene (MNDID) and a mass of ⌬C20 assessed by mass spectrometry was 29,341 Ϯ 2 Da, in agreement with a calculated value of 29,367.1 Da. In addition, ⌬C20 showed a marked resistance to various proteases, including trypsin, as observed for PACTX (data not shown). Thus, the purified ⌬C20 had the same biological and physicochemical properties as PACTX and degradation on the toxin molecules was negligible during the purification process.
Purification of proCTX-We applied the method developed here to purify proCTX because this approach needed no activation process. ProCTX produced in E. coli was also insoluble and could be solubilized with urea. The optimal concentration of urea for solubilization of proCTX was 3 M. Solubilized proCTX was purified to apparent homogeneity by the procedure used for the purification of ⌬C20 (Fig. 1B), although the extra step of anion exchange chromatography was necessary. From 10 liters of culture, 5.5 mg of proCTX was obtained.
Purified proCTX migrated slightly more slowly than ⌬C20 on a SDS-PAGE gel, and after activation by trypsin treatment, it migrated at the same position as ⌬C20 (data not shown). The N-terminal sequence was the same as that of ⌬C20. The molecular mass determined by mass spectrometry was 31,644 Ϯ 2 Da, close to 31,680.6 Da, a value calculated from the amino acid sequence (13). These data indicated that the purified protein consisted of the full length of the polypeptide encoded by the ctx gene.
ProCTX Does Not Form a 145-kDa Oligomer on the Target Membrane-After trypsin treatment, the purified proCTX exhibited hemolytic activity on rat erythrocytes comparable to the activity of PACTX, as well as that of ⌬C20 (Fig. 2). Without activation, proCTX had no hemolytic activity (up to 4 M). When erythrocytes incubated with proCTX were analyzed by immunoblotting, proCTX at the concentration of 0.06 M bound to the erythrocytes. Binding of proCTX was in a dose-dependent manner. ProCTX, however, did not form the oligomer of 145 kDa on the erythrocyte membrane (Fig. 3). Formation of the oligomer by proCTX was not detected even at the higher concentration (up to 4 M; data not shown). Trypsin-treated proCTX formed the oligomer as seen for ⌬C20. These data suggested that the inactive nature of proCTX was due to the lack of potential to form a 145 kDa oligomer on the target membrane.
In Fig. 3, in addition to the 145-kDa band of the oligomer, 60-kDa bands were seen for the active toxins. This was also the case for proCTX, albeit in a reduced amount. It appears that CTX forms a dimer on the erythrocyte membrane, even though the dimeric form of active CTX is not seen in solution, as described below. The oligomeric state of CTX may change once CTX encounters the hydrophobic environment of the membrane and the dimer could be an intermediate in the formation of the 145-kDa oligomer. However, because a 60-kDa band was also detected by immunoblotting for the heat-inactivated toxin and the toxin not incubated with cells (10), we cannot exclude the possibility that it was formed after solubilization with SDS.
C-terminal Peptide Is Dissociated from the Active Toxin-Proteolytic cleavage in the C terminus is required for activation of proCTX. It remained to be determined whether the proteolytic cleavage removes the C-terminal peptide from the toxin molecule or simply introduces a nick in the molecule. Using an antibody specific to the C-terminal peptide immobilized on Affi-Gel 10 beads, we estimated the fate of the C-terminal peptide generated by trypsin treatment. As shown in Fig. 4A,  5 g in lanes S and G and 0.7 g in lanes H and Q) were separated on 11.5% SDS-PAGE gels. Molecular size markers (Bio-Rad) are indicated on the left. FIG. 2. Hemolytic activities of ⌬C20 and proCTX with or without trypsin treatment. Rat erythrocytes were incubated with serially diluted toxins at 37°C for 60 min in 50 mM phosphate buffer, pH 7.2, 0.1 M NaCl. After centrifugation, hemolysis was determined by measuring hemoglobin release spectrophotometrically at 412 nm. Osmotic lysis of erythrocytes in distilled water was taken as 100% hemolysis.
⌬C20 and the trypsin-treated proCTX were not adsorbed by the immobilized antibody and recovered in the supernatants, whereas only a small portion of proCTX remained in the supernatant fraction. Although toxins in the pellet fractions were not clearly visible by Coomassie Blue staining because of the presence of broad bands of the IgG light chain co-migrating with toxins, immunostaining revealed that proCTX was recovered in the pellet fraction, whereas trypsin-treated proCTX, as well as ⌬C20, was not (Fig. 4B). These observations suggested that the C-terminal peptide left the toxin molecule once it was generated by proteolytic cleavage. This was further support for the notion that ⌬C20 could serve as a model molecule for the activated toxin.
CD Spectra of proCTX and ⌬C20 in the Far-ultraviolet-As a first step to elucidate molecular events involved in the activation process, we wanted to know whether activation of proCTX would alter the structure. For this purpose, we carried out CD spectroscopic analysis of proCTX and ⌬C20. The far-UV CD spectra of both proteins, presented in Fig. 5, were similar. ProCTX was estimated to contain 14.3% ␣-helices, 52.7% ␤-sheet, 14.2% ␤-turn, and 18.8% random structure, and ⌬C20 contained 12.5% ␣-helices, 54.4% ␤-sheet, 10.3% ␤-turn, and 22.9% random structure. These findings indicated that the secondary structure of proCTX was not affected by proteolytic removal of the C-terminal 20 amino acid residues and that CTX was a ␤-sheet predominant protein.
Tryptophan Fluorescence-In some cases, intrinsic fluorescence is sensitive for microenvironment transition around tryptophan residues. Because there was no tryptophan residue in the C-terminal region on proCTX, we compared fluorescence spectra of the active toxin and proCTX. As shown in Fig. 6, the maximum of the fluorescence emission spectrum of proCTX was displayed at 337.5 nm, whereas a maximum emission wavelength of ⌬C20 spectrum shifted to 342 nm and the intensity was decreased. The maximum wavelengths of both proteins were not affected by shifted salt concentration in the range of 0 -1.0 M NaCl (data not shown). Trypsin-treated proCTX without fractionation of the C-terminal peptide also showed a red-shifted spectrum with a maximum wavelength the same as that of ⌬C20 (Fig. 6). Although there was a small difference in intensity between ⌬C20 and trypsin-treated proCTX, this was attributed to slightly different concentrations of the proteins. These data suggested that an environment around the tryptophan residue(s) became hydrophilic during the activation process of proCTX.
Dimer to Monomer Conversion of proCTX by Proteolytic Ac-tivation-The active CTX was suggested to exist as a monomer in the solution (4), and the oligomeric state of proCTX remained to be characterized. We first did analytical gel filtration experiments. ProCTX was eluted much earlier than active toxins ⌬C20 and PACTX, and the molecular weights of proCTX and active toxins were estimated to be 52,000 and 24,000, respectively. Furthermore, the molecular weight of proCTX was reduced to 24,000 by trypsin treatment (data not shown). These findings suggested that proCTX existed as a homodimer in the solution and was converted to monomers by proteolytic activation.
To confirm the results, chemical cross-linking experiments were done. As shown in Fig. 7, a protein band migrating at the position of the dimer was detected for proCTX following incubation with dimethyl suberimidate. On the contrary, the dimer band was observed for neither ⌬C20 nor trypsin-treated proCTX.
Finally, the exact oligomeric states of the toxins were determined in analytical ultracentrifugation experiments (Table I). Sedimentation coefficient was 5.04 S for proCTX or 2.61 S for ⌬C20. Equilibrium ultracentrifugation of the toxins showed molecular weights of 63,000 for the protoxin and 29,000 for the active toxin. Each value was in good agreement with a calculated molecular mass of the dimer of proCTX or the monomer of ⌬C20.

DISCUSSION
To purify proCTX of P. aeruginosa, we used a new method, which is simple but includes a step of solubilizing the aggregated CTX by urea. This may raise the possibility that the purified proteins lost their native structures. However, the biological and physicochemical properties of the purified recombinant active CTX, ⌬C20, were indistinguishable from those of an active toxin obtained from P. aeruginosa. Purified proCTX was also active after trypsin treatment. Thus, the protein refolded correctly during dialysis after being denatured by urea.
When characterizing the purified proCTX, there were four major findings: (a) proCTX binds to the erythrocytes membrane but cannot form the oligomer on the membrane; (b) proteolytic action removes the C-terminal peptide from the toxin molecule; (c) activation of proCTX by the C-terminal cleavage does not induce change in the secondary structure; and (d) by proteolytic removal of the C-terminal peptide, the homodimer of proCTX is converted to monomers. One of the crucial steps in intoxication by CTX is oligomerization on the target membrane. The oligomer formed by CTX is most likely in a pentameric form (10). From our findings, the activation process of proCTX is proposed to be as follows: proCTX existing in a dimeric form can bind to the membrane but is cytolytically inactive because it does not oligomerize into a pentamer. Once the C-terminal peptide is removed by proteolytic action, proCTX is converted into monomers, which form a pentamer on the membrane and thus are active in cytolytic events. This is in clear contrast with the activation process described for aerolysin, another pore-forming toxin that requires the C-terminal processing for activation. Aerolysin is also produced as an inactive precursor and exists as a homodimer. After activation by C-terminal cleavage, aerolysin exists as a homodimer (26). To form a heptameric oligomer on the membrane, conversion from dimer to monomer is thought to be essential for aerolysin as well. Although such an intermediate has not been detected, the conversion may be induced by a structural change upon active dimeric toxin after binding to the target membrane (12). The process of oligomerization on the membrane is one point that remains to be elucidated for pore-forming toxins.
The C terminus of proCTX contributes to stable dimer formation but the mechanism remains obscure. A synthesized 21-mer peptide corresponding to the C terminus of proCTX did not form the dimer. 2 A direct interaction between the C-terminal regions may not be an event in dimer formation, or such an interaction may be stabilized by other interactions between the two proteins. Another possibility is that the C-terminal region, which contains a stretch of 10 hydrophobic amino acids followed by a sequence potentially forming a coiled-coil helix (residues 277-286), may involve an intermolecular interaction with other parts of the toxin molecule to stabilize the dimeric form of proCTX.
It also remains unclear how the change in intrinsic fluorescence of the toxin occurred by activation. The results suggest that a tryptophan residue(s) buried in the proCTX molecule was exposed to the solvent after activation; it might have been covered directly by the C-terminal peptide and appeared on the surface of the protein by removal of the peptide. Alternatively, the residue(s) might exist on the intersurface of the dimeric proCTX and be exposed to the solvent by dissociation of the toxin molecule into monomers. Identification of the tryptophan residue(s) responsible for change in intrinsic fluorescence will be informative for understanding the structural features of proCTX.
Another important finding on the structure of CTX is that the toxin consists predominantly of the ␤-sheet. Increasing numbers of bacterial toxins that form a pore by oligomerization 2 M. Ohnishi, unpublished data.
FIG. 7. Chemical cross-linking analysis. ProCTX (P), trypsintreated proCTX (T), and ⌬C20 (⌬) (100 g/ml) were incubated with the indicated concentration of dimethyl suberimidate in 50 mM triethanolamine, pH 8.5. As a control, proCTX was incubated with inactivated dimethyl suberimidate by incubation with 100 mM Tris-HCl (lane C). Proteins were separated on 11% SDS-PAGE gels and stained with Coomassie Blue. have been reported to have structures rich in the ␤-sheet. This was documented most clearly for aerolysin (16) and staphylococcal ␣ toxin (27). The structure of oligomerized ␣ toxin at a resolution of 1.9 Å revealed that a transmembrane domain of the channel is a 14-stranded ␤-barrel. The ␤-sheet predominant feature may be essential for oligomerization and channel formation of CTX as well. To understand the structural roles of the ␤-sheet in CTX, the oligomerized toxin will have to be isolated, and then it will be possible to examine the exact number of CTX protomers constituting the oligomer.