Pore Formation of Thermostable Direct Hemolysin Secreted from Vibrio parahaemolyticus in Lipid Bilayers

Abstract

Vibrio parahaemolyticus secretes thermostable direct hemolysin (TDH), a major virulence factor. Earlier studies report that TDH is a pore-forming toxin. However, the characteristics of pores formed by TDH in the lipid bilayer, which is permeable to small ions, remain to be elucidated. Ion channel-like activities were observed in lipid bilayers containing TDH. Three types of conductance were identified. All the channels displayed relatively low ion selectivity, and similar ion permeability. The Cl⁻ channel inhibitors, DIDS, glybenclamide, and NPPB, did not affect the channel activity of pores formed by TDH. R7, a mutant toxin of TDH, also forms pores with channel-like activity in lipid bilayers. The ion permeability of these channels is similar to that of TDH. R7 binds cultured cells and liposomes to a lower extent, compared to TDH. R7 does not display significant hemolytic activity and cell cytotoxicity, possibly owing to the difficulty of insertion into lipid membranes. Once R7 is assembled within lipid membranes, it may assume the same structure as TDH. The authors propose that the single glycine at position 62, substituted with serine in the R7 mutant toxin, plays an important role in TDH insertion into the lipid bilayer.

Vibrio parahaemolyticus is a bacterium that causes food poisoning by polluting marine fish and shellfish (Joseph, Colwell, and Kaper 1982; Janda et al. 1988). Major clinical manifestations of infection by this organism include gastroenteritis symptoms, such as diarrhea. This disease is prevalent in Japan, where eating raw fish and shellfish is customary. Earlier studies demonstrate that thermostable direct hemolysin (TDH) secreted by V. parahaemolyticus is a major virulence factor (Honda and Iida 1993). Most clinical isolates of V. parahaemolyticus induce beta hemolysis on Wagatsuma agar, a special blood agar medium. TDH is responsible for this hemolysis, which is designated the ‘Kanagawa phenomenon’ (KP) (Honda et al. 1976; Miyamoto et al. 1969).

TDH, a pore-forming toxin (Honda et al. 1992), is a protein of 165 amino acid residues (Tsunasawa et al. 1987), which performs a variety of biological activities, including hemolytic activity, cytotoxicity, cardiotoxicity, and enterotoxicity. TDH binds erythrocyte membranes, and forms a pore with a functional diameter of approximately 1 to 2 nm, resulting in erythrocyte lysis owing to an increase in colloidal osmotic pressure (Miyamoto et al. 1969). The mechanism of pore formation by TDH remains to be elucidated.

To establish the relationship between TDH and hemolysis, several mutant toxins have been generated by in vitro mutagenesis (Iida et al. 1995). Among these, the R7 mutant contains a single amino acid substitution (glycine at position 62 replaced with serine) in its N-terminal region. It is reported that R7 retains the ability to bind erythrocytes, but loses hemolytic activity, and acts as a competitive inhibitor of wild-type TDH (Tang et al. 1994).

Previous studies using planar lipid bilayer methods report that other toxins, such as alpha toxin secreted by Staphylococcus aureus, El-Tor hemolysin secreted by Vibrio cholerae O1 bio-type El-Tor, aerolysin secreted by Aeromonas hydrophila, and VacA secreted by Helicobactor pylori, additionally form pores with channel-like activities in lipid bilayers (Korchev et al. 1995; Ikigai et al. 1997; Wilmsen, Pattus, and Buckley 1990; Tombola et al. 1999). Channel activities of these pores are important for the biological functions of toxins (van der Goot 2003). Thus, pore formation possibly contributes to the various activities of TDH. The characteristics of pores formed by TDH remain to be resolved.

In this study, we analyze the ion permeability of pores formed by TDH, using planar lipid bilayer methods. The mechanism of TDH binding to the lipid bilayer is additionally investigated in cultured human epithelial cells and liposomes. Furthermore, the functions of TDH and R7 are compared, and the structure-function relationship of TDH is discussed.

MATERIALS AND METHODS

Construction and Purification of TDH and R7

The cDNA of tdh was cloned from the Vibrio parahaemolyticus T4750 strain using two primers, 5′-GGAT CCATCGAAGGTCGTTTTGAGCTTCCATCTGTC-3′ and 5′-GTCGACTTATTGTTGATGTTTACA-3′. A nucleotide sequence corresponding to amino acids sensitive to Factor Xa was inserted into the signal peptide of the tdh gene. A BamHI restriction site was cloned into the N-terminus, and a SalI restriction site inserted into the C-terminus of tdh. Structural tdh was cloned into pET28a (Novagen) for gene expression. A recombinant plasmid harboring the tdh gene was introduced into Escherichia coli BL21 (DE3) cells by transformation used by heat shock. Transformants were cultivated at 37°C with rotary shaking in Luria-Bertani Broth (1% Bacto tryptone [Difco], 0.5% Bacto yeast extract [Difco], 1% NaCl) containing 50 μg/ml kanamycin. At OD₅₅₀ nm of 0.4 to 0.6, 1 mM isopropyl-β-D(−)-thiogalactopyranoside was added to induce protein, and the culture was further incubated for 3 h at 37°C. Bacterial cells were harvested by centrifugation at 4000 × g for 30 min, and the pellet lysed with CelLytic B Bacterial Cell Lysis/Extraction reagent (Sigma), according to the manufacturer’s instructions. After centrifugation at 30, 000 × g for 30 min, the supernatant was applied to a His.bind resin column (Novagen), and recombinant TDH (His fusion protein) was purified, using the manufacturer’s protocol. Following dialysis with Tris buffer saline (TBS; 20 mM Tris, 137 mM NaCl, pH 7.2), the sample was concentrated to 1 to 4 mg/ml in TBS plus 2 mM CaCl₂. Factor Xa (New England Biolabs) was added (Factor Xa:protein = 1:200), and the sample was incubated with shaking for 15 to 17 h at room temperature. To remove Factor Xa, benzamidine-sepharose (Amersham Pharmacia Biotech) equilibrated with TBS was added, and the sample was shaken overnight at 4°C. For the removal of benzamidine-sepharose, the sample was filtered using the Minisart syringe filter (0.45 μl; Sartorius UK Products). Gel filtration of the sample was monitored by high-performance liquid chromatography. Each compartment was examined by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) (using 12% polyacrylamide gels), according to the method of Laemmli (1970), and the appropriate compartments selected. Compartments were concentrated using Amicon Ultra4 (Millipore), according to the manufacturer’s protocol.

R7, a mutant toxin of TDH with a single amino acid substitution (glycine-62 to serine involving a nucleotide change from GGT to AGT), was generated by site-directed mutagenesis using the GeneTailor System (Invitrogen life technologies). The purification procedure for R7 was similar to that for TDH. The purity of samples was examined by SDS-PAGE in polyacrylamide 12% gels with 2 μg of protein per lane.

Planar Lipid Bilayer Experiments

Planar bilayers were formed at the tips of silanized borosilicate glass pipettes by hydrophobic apposition of two lipid monolayers initially formed at the air-water interface. Di-phytanoylphosphatidylcholine (DPhPC) (Avanti Polar Lipids, Alabaster, AL, USA) plus 50% (w/w) cholesterol (Sigma) was used as membrane lipid. Glass pipettes (World Precision Instruments) with impedance of 2 to 4 MΩ were prepared in an electrode puller (Model p-97; Sutter Instrument). The pipette and bath solution contained 150 mM KCl, 10 mM HEPES, with the pH adjusted to 7.2 with KOH. Following 10 μg/ml TDH insertion into the lipid bilayers at the top of the pipette, ion-channel like currents were measured.

Channel currents of pores formed by TDH were recorded with an electrical amplifier (CEZ2200; Nihon Khoden, Tokyo Japan), and stored in a personal computer with an analog to digital converter (DigiData 1200; Axon Instruments, Foster, CA). The sampling frequency of single channel data was 5 kHz with a low-pass filter (1 kHz). The pClamp version 7 software (Axon Instruments) was employed for data acquisition, and BIO-PATCH version 3.42 software (BIO-LOGIC) for data analysis.

Reversal potential measurements were taken by establishing a salt gradient across membranes containing the toxin channels. The KCl gradient solution comprised bath solution (500 mM KCl, 10 mM HEPES, pH adjusted to 7.2 with KOH), and pipette solution (150 mM KCl, 10 mM HEPES, pH adjusted to 7.2 with KOH). From the reversal potentials (V _m ) and the salt concentrations (c′ and c″) in the pipette and bath solutions, the ratio of the P _c (cation) and P _a (anion) permeabilities was determined, using the Goldman-Hodgkin-Katz equation (Benz, Janko, and Lauger 1979).

$$V_m=RT/Fln((P_c\hspace{0.17em}c″+P_a\hspace{0.17em}c\prime )/(P_{c}c\prime +P_{a}c″))$$

(R is the gas constant, T the absolute temperature, and F the Faraday constant). The values represent the mean obtained from at least three membranes.

Ion channel inhibitors, such as 4, 4′-diisothiocyanatostilbene-2, 2′-disulphonic acid (DIDS) (Sigma), glybenclamide (Wako), and 5-nitro-2-(3-phenylpropylamino)benzoate (NPPB), (Calbiochem), were added in bath solution.

Cell Culture Conditions

Caco-2 cells originally derived from human colon adenocarcinoma were purchased from a Health Science Research Bank, and grown in Dulbecco’s modified Eagle’s medium (Sigma) supplemented with 10% fetal bovine serum (INC Biomedicals, Aurora, OH) and 50 μg/ml gentamicine (Sigma). Cells were cultured on 60-mm cell culture dishes at 37°C and 5% CO₂ until further analysis.

Hemolytic Activity

Hemolytic activity was assayed, as described in a previous report (Miyake, Honda, and Miwatani 1989). Briefly, rabbit erythrocytes were washed five times with TBS, and adjusted to a concentration of 10% (v/v) suspension in TBS. Purified TDH or R7 diluted with TBS (5 μl), was added to 0.1 ml erythrocyte suspension, and incubated at 37°C for 60 min. The mixture was centrifuged at 1000 × g for 2 min, and the OD at 540 nm of the supernatant was determined. The hemolysis percentage was calculated relative to 100% erythrocyte lysis with Triton X-100 as the standard.

Cell Viability

Caco-2 cells were trypsinized and suspended at a concentration of 1 × 10⁵ cells/ml in medium. Cells were cultured in 96-well plates for 6 days. Thirty minutes after the addition of toxin, cell viability was measured with a Cell Counting Kit (Dojindo, Japan), according to the manufacturer’s instructions.

Preparation of Anti-TDH Antiserum

Anti-TDH antiserum was prepared as described previously (Honda et al. 1980). Briefly, 200 μg TDH in 2.5 ml TBS (pH 7.2) was emulsified in an equal volume of Freund’s incomplete adjuvant (Difco). The emulsion was inoculated intramuscularly (5 ml per rabbit) into young New Zealand white rabbits. After five booster injections, antiserum was obtained.

Binding of TDH and R7 to the Cell Membrane

Caco-2 cells were trypsinized, and suspended at a concentration of 1 × 10⁵ cells/ml in medium. Cells were cultured on 60-mm dishes for 6 days, incubated with 10 μg/ml TDH or R7 for 30 min at 37°C, washed twice with ice-cold PBS (−), and solubilized in 1 mM cell lysis buffer containing 20 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA, 1% (w/v) sodium deoxycholate, 0.1% SDS, 1% Nonidet-P40, and 1 mM phenylmethylsulfonyl fluoride (PMSF; Sigma). After 30 min on ice, the cell suspension was transferred to a 1.5-ml microcentrifuge tube, and centrifuged at 30, 000 × g for 30 min. The supernatant was subjected to SDS-PAGE (on a 12% polyacrylamide gel). For Western blot analysis, proteins were transferred to a PVDF membrane (PROTRAN Nitro cellulose Transfer Membrane; Schleicher & Schuell Bio-Science GmbH). The membrane was incubated for 1 h at room temperature (RT) in blocking buffer (3% (w/v) skimmed milk in Tris buffer saline, pH 7.6), and treated with wash buffer (TBS-0.05%, Tween-20). Next, the membrane was incubated with primary antibody diluted with blocking buffer (1:1000) for 30 min at 37°C. Following treatment with wash buffer, the membrane was visualized with peroxidase-conjugated secondary antibody (1:1000 dilution) (peroxidase labeled anti-rabbit antibody) using an enhanced chemiluminescence (ECL) plus detection system (Amersham Pharmacia).

Binding of TDH and R7 to Liposomes

Liposomes (Coatsome EL C-01, N-01, and A-01; NOF) were prepared in keeping with the manufacturer’s instructions. TDH or R7 (10 μg/ml) and liposomes (2.5 μmol) were incubated for 30 min at 37°C. The toxin-liposome complexes formed were collected by centrifugation at 20,000 ×g for 30 min at 4°C, and washed twice with PBS (−) to remove unbound toxin. Toxin-liposome complexes were suspended in 30 μl PBS (−) containing 5% Triton X-100, and incubated for 1 h at 37°C. Samples were centrifuged at 20,000 ×g for 30 min, and the supernatant was subjected to SDS-PAGE and Western blotting, as described above.

RESULTS

Analysis of Purified TDH and R7 by SDS-PAGE

Purified TDH and R7 displayed a single band on SDS-PAGE (Figure 1A), indicating homogeneity. Samples were stored in TBS at 4°C until use in assays.

Hemolytic Activity of TDH and R7

To determine whether recombinant TDH and R7 perform the correct biological functions, we analyzed their hemolytic activities (Figure 1B). TDH displayed dose-dependent hemolytic activity. However, R7 did not induce hemolysis, consistent with earlier reports (Iida et al. 1995; Tang et al. 1994).

Cytotoxicity of TDH and R7 to Caco-2 cells

Next, we examined the cytotoxicity of TDH and R7 to Caco-2 cells (Figure 2). Following treatment with TDH, Caco-2 cells exhibited a dose-dependent loss of viability. In contrast, R7 did not affect the viability of Caco-2 cells. This result induced the hypothesis that the mutant toxin might not have ability of pore formation.

Characterization of the Channels Formed by TDH and R7

We analyzed channel formation by TDH in planar lipid bilayer membranes using single channel recordings. The ion current was measured with a KCl solution (Figure 3). TDH formed pores displaying channel-like activity with three types of conductance (large, intermediate, and small). Upon substitution of KCl in the bath and pipette solution with NaCl or N-methyl-d-glucamine (NMDG)-Cl, similar currents were observed (Figure 3). The current voltage (I-V) relationship (Figure 3C), and channel conductance at +100 mV and −100 mV pipette potentials were measured (Table 1).

The relative permeabilities of TDH channels were determined under asymmetric ionic conditions (see Materials and Methods). Figure 4 depicts the current voltage (I-V) relationship, and the reversal potentials obtained (Table 2). Selectivity (P_K⁺ /P_Cl⁻ ) was calculated using the Goldman-Hodgkin-Katz equation (Table 2), as described in Materials and Methods. Little selectivity was observed for each channel.

To determine the characteristics of these channels in more detail, we examined the effects of Cl⁻ channel inhibitors, including DIDS (inhibitor of the Ca²⁺-activated Cl⁻ channel) (Cunningham et al. 1992; Inoue et al. 1997), glybenclamide (inhibitor of the cAMP-dependent Cl⁻ channel) (Schultz et al. 1999; Sheppard et al. 1993), and NPPB (inhibitor of the Cl⁻channel) (Cunningham et al. 1992; Schultz et al. 1999). Inhibitors added to the bath solution at the specified concentrations (Figure 5A ) did not affect the open probability (NPo) of TDH channels (Table 3).

Next, we analyzed R7 using planar lipid bilayer membranes (Figure 6). Surprisingly, R7 displayed channel-like activity similar to TDH (Figures 4 to 6), indicative of pore-forming ability.

Binding of TDH and R7 to Lipid Membranes

We had the question why R7 showed channel activity in lipid bilayers though R7 did not show the hemolysis or cell cytotoxicity. The binding of toxins to lipid membranes is important for hemolysis. Accordingly, we analyzed the ability of TDH or R7 to bind to Caco-2 cell membranes by Western blotting (Figure 7). R7 bound the cells and liposomes, but to a lower extent than TDH. The results support the theory R7 insertion into lipid membranes is difficult. Because the surfaces of cell membranes are negatively charged, the same analysis was performed using three types of liposomes with different electric charges, specifically, cationic, anionic, and nonionic liposomes (Figure 8). Notably, the electric charge of the membranes did not affect the toxin binding.

In both experiments, we observed bands at different positions, in addition to the TDH monomer (Figures 7, 8, open arrows).

DISCUSSION

The characteristics of pore formation by TDH are yet to be resolved. In our experiments, TDH formed pores with ion channel–like activities in lipid bilayers, and three types of conductance were identified. Ion channel–like activities of most other reported pore-forming toxins display unitary channel conductance (Chakraborty et al. 1990; Menzl et al. 1996; Wilmsen, Pattus, and Buckley 1990). To date, no ion channel-like activities of proteins with high homology to TDH have been reported. Recently, Hardy, Nakano, and Iida (2004) also observed that the TDH had ion channel–like activities and the channel conductance ranged 30 to 450 pS in 0.5 M KCl. They speculated the variability of the conductance might reflect variations in the number of toxin monomers that oligomerize to form the transmembrane channel. We propose that TDH is a novel type of channel-forming protein. There are two possible reasons for the several conductances displayed by TDH. One is the diversity of oligomer formation. It is necessary for TDH to form an oligomer for hemolysin activity (Chakraborty et al. 1990; Füssle et al. 1981; Menzl et al. 1996). Aerolysin secreted from A. hydrophila forms pores in the membranes of mammal cells (Parker et al. 1994), and the pore is a set with seven molecules of the toxin (Wilmsen et al. 1992). Moreover, α-hemolysin secreted from S. aureus that binds erythrocyte and liposome membranes is heptameric (Bhakdi et al. 1992; Gouaux et al. 1994). This heptamer formation acts as a membrane channel (Korchev et al. 1995). TDH is a band of approximately 21 kDa on an SDS gel. However, the molecular weight of TDH is 42 kDa on a gel filtration column (Takeda, Taga, and Miwatani 1978). Accordingly, it is proposed that TDH forms a dimer in liquid. Our current data from Western blotting analyses reveal that TDH has a higher molecular weight than the 21-kDa monomer. These data support the hypothesis that TDH forms several types of structures in lipid bilayers, resulting in several conductances. The other reason for multiconductance is the appearance of subconductance as a result of conformational changes inside the pore (a part that the ion passes). Cystic fibrosis transmembrane conductance regulator (CFTR), a Cl⁻ channel, forms one channel with one molecule. However, a number of reports using planar lipid bilayer methods show that the conformational changes of CFTR induce subconductance (Gunderson et al. 1995; Tao et al. 1996; Xie et al. 1995). Similarly, the conformational changes inside the pore formed by TDH may induce subconductance. TDH displays low homology with other pore-forming toxins, and its structure is yet to be solved. Thus, to clarify the form of TDH in the lipid bilayers, further analyses are necessary.

It is proposed that hemolysin participates in various biological functions via the ion channel–like activities of the pore formed by the toxin (van der Goot 2003). TDH is a major virulence factor of V. parahaemolyticus (Honda and Iida 1993). A mouse ileal loop assay shows that TDH contributes to enterotoxigenicity (Nishibuchi et al. 1992). In a study with Ussing chambers, TDH induced Cl⁻ secretion from Ca²⁺-activated Cl⁻channels via influx of Ca²⁺ to human colonic epithelial cells, and caused diarrhea (Takahashi et al. 2000). However, ion channel–like activity of the pore formed by TDH was not inhibited by DIDS, an inhibitor of the Ca²⁺-activated Cl⁻ channel. These results indicate that Cl⁻ secretion by TDH is not due to passing of Cl⁻ through the pore. Although R7 formed a pore with ion channel–like activity in lipid bilayers, no cytotoxicity was observed with the mutant. Analysis of TDH and R7 binding to lipid membranes using Caco-2 cells and liposomes revealed that the extent of R7 binding to membranes was significantly lower than that of TDH, and insertion of the mutant toxin into the lipid membrane was difficult. Based on these results, we propose that the cytotoxicity of R7 decreases significantly because the amount of toxin inserted into lipid bilayers is small, albeit sufficient to form a pore. R7 contains a single amino acid substitution (glycine 62 to serine) (Honda et al. 1992). The data collectively suggest that the glycine residue at position 62 from the N-terminus of TDH is important for binding of the toxin to lipid bilayers.