Intramolecular chaperone activity of the pro-region of Vibrio cholerae El Tor cytolysin.

Vibrio cholerae synthesizes a toxin named El Tor cytolysin/hemolysin, which lyses erythrocytes and other mammalian cells. This toxin is encoded by the hlyA gene and is synthesized as a precursor form, prepro-HlyA. Prepro-HlyA consists of, from the amino terminus of this protein, a signal peptide, a pro-region, and a mature region. The pro-region is cleaved off extracellularly resulting in activation. To analyze the role of the pro-region, we substituted the native hlyA gene with the pro-region-deleted hlyA gene (hlyAΔpro). The hemolytic activity of the mutant organism was markedly decreased; the product of the hlyAΔpro gene, secreted in the periplasm, was degraded. To compare their abilities to form tertiary structure, the purified mature- and pro-HlyA were denatured and then renatured by reducing the concentration of denaturant; the denatured pro-HlyA recovered almost all activity while the mature-HlyA was not renatured. The sequences of the pro-region and a molecular chaperone, Hsp90, were similar. The pro-region expressed in Escherichia coli containing the hlyAΔpro gene increased the cytolytic activity. The purified pro-region peptide also facilitated renaturation of the denatured mature HlyA. These results suggest that the pro-region possibly guides the folding of the cytolysin similar to a molecular chaperone; the pro-region and molecular chaperones share common function and structure.

Vibrio cholerae O1 biotype El Tor, which is the causative agent of the current seventh cholera pandemic, secretes a cytolytic toxin (V. cholerae El Tor cytolysin/hemolysin). The cytolysin lyses erythrocytes and other mammalian cells and exhibits enterotoxicity in experimental diarrhea models (1). Thus, the cytolysin may contribute to the pathogenesis of gastroenteritis caused by V. cholerae strains, especially the strains not producing cholera toxin, the major cause of cholera diarrhea.
El Tor cytolysin/hemolysin is encoded by hlyA and synthesized as an 82-kDa precursor form (prepro-HlyA) (2). The prepro-HlyA consists of, from the amino terminus, a signal peptide (25 residues), a pro-region (132 residues), and a mature region (584 residues) (2). The signal peptide is cleaved during secretion through the bacterial inner membrane, resulting in pro-HlyA. Following secretion through the outer membrane, a 15-kDa pro-region of pro-HlyA is cleaved, generating mature HlyA. Removal of the pro-region to yield mature HlyA results in a more than ten-fold increase in cytolytic activity (2). The role of the pro-region, however, in production and the secretory process is not known.
It is well known that proteases are produced in a precursor form and processed to active enzymes by releasing the proregion of the precursor protein. Recently, the pro-regions of a few serine proteases, namely, subtilisin E of Bacillus subtilis (3), ␣-lytic protease of Lysobacter enzymogenes (4), and carboxypeptidase Y of Saccharomyces cerevisiae (5), have been shown to have a chaperone-like activity with their mature regions in vitro. To distinguish the pro-regions of these proteases from molecular chaperones, they are called intramolecular chaperones; of importance is that these N-terminal peptides possess chaperone-like activity that only functions in folding of the precursor protein of which they are a part (6,7). The roles of intramolecular chaperones in vivo, where and how they work in the cell, however, are still not well understood. Further, it has been reported that the isolated pro-regions of subtilisin E and ␣-lytic protease guide the folding of the matured enzyme intermolecularly as well as intramolecularly (3,4,7).
In this paper, we show that the pro-region of El Tor cytolysin works as a chaperone for the cytolysin, a protein that is biologically and physically different from the serine proteases. The pro-region seems to act like a molecular chaperone (e.g. heat shock protein 90 (Hsp90)). Moreover, we present evidence suggesting that the isolated pro-region of HlyA acts like an intermolecular chaperone in facilitating the folding of cytolysin.

EXPERIMENTAL PROCEDURES
Bacterial Strain, Plasmids-Bacterial strains and plasmids used in this study are listed in Table I. V. cholerae N86 was the original strain used in this study. Escherichia coli HB101, SM10pir and BL21(DE3) were used for general manipulation of plasmids, mobilization of plasmid into V. cholerae, and expression of fusion protein, respectively.
Chemicals, Recombinant DNA Techniques and Protein Purification-Enzymes and chemicals were purchased from standard commercial sources. Manipulation of DNA was carried out as described previously (8). Site-directed mutagenesis was performed by the method of Kunkel (11,12). El Tor cytolysin (mature HlyA) was purified from the supernatant of V. cholerae N86, as previously reported (10). Antibodies against purified mature HlyA were affinity purified from rabbit antiserum, as described previously (10). Purification of fusion peptides was performed using pET Expression System 25b (Novagen, Madison, WI). * This work was supported by a grant-in-aid for scientific research from the Ministry of Education, Science, and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Gene Replacement-Replacement of a native gene on V. cholerae chromosome with a mutated gene was carried out by double-crossover homologous recombination as described previously (13). Briefly, E. coli SM10pir harboring a recombinant plasmid having a mutant gene on a suicide vector, pKY719, was conjugated with V. cholerae. Conjugation was performed on nitrocellulose membranes overlaid on LB agar plates. Following incubation for 2 h at 37°C, the mated bacteria were washed from the membranes and inoculated onto TCBS agar medium plates containing 5 g/ml chloramphenicol. Following incubation overnight at 37°C, the chloramphenicol-resistant V. cholerae colonies, resulting from single-crossover events, were isolated. One of the isolates was cultured in LB broth without antibiotics. Colonies resulting from the second crossover were tested for chloramphenicol sensitivity on LB agar plates with 5 g/ml of chloramphenicol. Chloramphenicol-sensitive colonies were confirmed for the presence of about a 400-base pair deletion by Southern blot hybridization.
SDS-Polyacrylamide Gel Electrophoresis and Western Blotting Analysis-One hundred l of an overnight culture in LB broth were inoculated into 10 ml of syncase medium (9) supplemented with 3% glycerol in Petri dishes. After a 9-h stationary incubation at 30°C, cells were collected by centrifugation at 5,000 ϫ g for 5 min. The precipitated cell pellet was suspended with 1 ml of 2,000 unit/ml polymyxin B for 30 min at 4°C (14) to stimulate the release of periplasmic proteins. The bacterial cells were removed from the supernatant by centrifugation. The culture supernatant, separated from the cell fraction, was precipitated by addition of 1 ml of 100% (w/v) trichloroacetic acid. Following centrifugation for 15 min at 1,500 ϫ g, the precipitate was suspended in 1 ml of 10 mM phosphate-buffered saline (PBS), 1 pH 7.0. Electrophoresis on 10% polyacrylamide gel and Western blot analysis were carried out as described previously (2).
Pulse-Chase Experiment-A pulse-chase experiment was carried out basically by the method described previously (2). In brief, 50 l of an overnight culture in LB broth were inoculated into 5 ml of M9 minimal medium supplemented with 0.1% casamino acids. After incubation for 6 h at 37°C, cells were precipitated by centrifugation and suspended in sulfur-free M9 medium. Cells were then incubated for 1 h to induce sulfur starvation, and 1 mCi of a [ 35 S]methionine and [ 35 S]cysteine mixture was then added, followed by further incubation. After 3 min, 500 l of a mixture containing nonradioactive methionine (50 mM) and cysteine (100 mM) were added. At appropriate periods, 1.5 ml of the culture medium were transferred into ice-chilled centrifugation tubes containing 30 l of 0.5 M EDTA and centrifuged immediately. To extract the periplasmic proteins, the precipitated cell pellet was immediately suspended in 100 l of Buffer A (20 mM Tris-HCl, pH 7.5, 0.1 M NaCl, 10 mM EDTA) with 2,000 unit/ml of polymyxin B and incubated for 30 min at 4°C. The solution was then centrifuged to remove the precipitated cells, and the supernatant was used as the periplasmic sample.
Culture supernatant (0.3 ml) was mixed with 30 l of 100% (w/v) trichloroacetic acid and centrifuged. The precipitate was suspended in 100 l of Buffer A, followed by heating at 95°C for 3 min to ensure solubilization, and used as the supernatant sample.
Both samples (100 l) were diluted with 900 l of Buffer B (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.5% Triton X-100, 0.05% bovine serum albumin and 5 mM EDTA), and undissolved materials were removed by centrifugation. The sample was mixed with 10 l of the antibodies and kept at 0°C overnight. The mixture was added to 20 l of a Sepharose CL4B gel coupled with protein A (50% (v/v) suspension) and incubated for 2 h at 0°C. The mixture was centrifuged, and the supernatant was discarded. The precipitate was washed five times with Buffer B, suspended in 20 l of SDS-polyacrylamide gel electrophoresis sample buffer, heated for 3 min at 95°C, and centrifuged (5,000 ϫ g). 10 l of the supernatant were applied to SDS-polyacrylamide gels.
Denaturation and Renaturation Procedure-Denaturation and rena-   Assay for Hemolytic Activity-Hemolytic activity was determined as described previously (16). Briefly, sheep erythrocytes were washed in PBS three times and adjusted to a hematocrit of 2% with PBS. Sixty l of serially diluted samples with PBS were mixed with an equal volume of the erythrocyte suspension and incubated for 2 h at 37°C. After centrifugation at 2,000 ϫ g for 30 s, the supernatant (100 l) of the reaction mixtures was taken for spectrophotometric measurement of released hemoglobin at 540 nm on a 96-well plate by Multiskan Mcc/340 (Labosystems, Tokyo, Japan). One hemolytic unit (HLU) is defined as the activity required to lyse 50% of cells in 0.1 ml of 1% sheep erythrocytes in phosphate-buffered saline at 37°C for 2 h.

Construction of a Pro-region-deleted Mutant-
To investigate the role of the pro-region of El Tor cytolysin of V. cholerae, we constructed the pro-region-deleted gene (hlyA⌬pro) from the structural gene (hlyA) for El Tor cytolysin (2,17,18) (Fig. 1). Restriction enzyme MscI sites were created at both ends of the pro-region by site-directed mutagenesis, resulting in pNK100. The DNA encoding the pro-region (amino acids 26 -157) was removed by digestion with MscI and self-ligation of pNK100, resulting in pNK101. This deletion does not change the coding frame, and the amino acid residues constituting the pre-mature junction (Ala-Asn) are the same as those of the pre-pro and pro-mature junctions. The NruI-AccI fragment of pNK101 was then exchanged with the NruI-AccI fragment of hlyA in pNK105, resulting in pNK106. This exchange was carried out because pNK105 has longer upstream and downstream adjacent regions of V. cholerae chromosome than pNK101 does, which would give it an advantage in the homologous recombination studies. The SacI-BamHI fragment of pNK106, having hlyA⌬pro with adjacent regions, was subcloned into the SacI and BamHI sites of an R6K-ori suicide vector, pKY719 (19), resulting in pNK109.
The plasmid, pNK109, was introduced into E. coli SM10pir, which was mated with V. cholerae N86. V. cholerae containing a single-crossover recombinant of hlyA were isolated first; second crossover recombination events were confirmed by Southern blot hybridization. We selected two isogenic clones generated from a single-crossover intermediate clone, one having the native hlyA gene (N86-RV, revertant to the same genotype as the parent) and the other lacking the coding sequence for the pro-region (N86hlyA⌬pro). N86hlyA⌬pro showed no significant differences in growth compared with the parent N86 or the isogenic revertant, N86-RV. The hemolytic activity of N86hlyA⌬pro on the SB-LBA medium was significantly decreased compared with N86 and N86-RV. The hemolytic activity in the supernatant of N86hlyA⌬pro was only 7% of N86 and N86-RV (data not shown).
Analysis of Pro-region-deleted Cytolysin Produced by N86hlyA⌬pro-The specific activity of the pro-region-cleaved mature HlyA is much higher than that of pro-HlyA (2). However, the hlyA⌬pro, which lacked the sequences coding for the pro-region, expressed much weaker hemolytic activity than did the wild-type parent. Therefore, we investigated the fate of

TABLE II
Hemolytic activity of pro-and mature HlyA after denaturation and renaturation 1.25 nmol/ml purified pro-and mature HlyA were mixed into 6 M guanidine HCl and incubated for 2 h at 20°C for denaturation. After incubation, the denatured samples were diluted 60 times with renaturing buffer (10 mM sodium phosphate buffer, 0.15 M NaCl, 0.1% bovine serum albumin, and 5 mM 2-mercaptoethanol, pH 7.0) and incubated for 30 min at 20°C. Samples were then mixed with an equal volume of 100 ng/ml trypsin and incubated at 37°C for 30 min. the product of this mutant gene in N86hlyA⌬pro. Pulse-chase analysis showed no HlyA⌬Pro in the culture medium of N86hlyA⌬pro although native HlyA of N86-RV was secreted within a few minutes ( Fig. 2A). On the other hand, in the periplasmic fraction of N86hlyA⌬pro, several 30 -40-kDa degradation products, which were not seen in N86-RV, appeared rapidly ( Fig. 2A). Western blot analysis gave similar results. HlyA⌬Pro was not secreted into culture medium although some degradation products were evident (Fig. 2B). These data indicate that the HlyA pro-region is essential for secretion and that HlyA⌬Pro stays in the periplasmic space where it is degraded. Denatuation of Pro-and Mature HlyA-As misfolded or denatured proteins are considered to be more susceptible to proteases and degraded more easily than folded proteins (20), we suspected that HlyA⌬Pro might be defective with regard to folding. We compared the renaturation abilities of denatured pro-and mature HlyA proteins. First, to test the influence of denaturant on activities of both toxins, guanidine HCl was added to each toxin. Following incubation for 2 h at 20°C, hemolytic activities were determined. Hemolytic activities of both toxins decreased similarly with increasing concentration of guanidine HCl; both toxins lost almost all in the activities when the concentration of guanidine HCl was 1 M (data not shown).
Renaturation of Denatured Pro-and Mature HlyA-To denature the pro-and mature HlyA completely, both toxins were incubated with 6 M guanidine HCl for 2 h. After denaturation, the mixtures were quickly diluted with renaturing buffer. After incubation for 30 min at 20°C, hemolytic activity was determined (Fig. 3). After dilution of guanidine HCl in the samples, the specific activities of the pro-HlyA increased. When the final concentration of guanidine HCl was decreased to 0.1 M, the activity of the denatured pro-HlyA was recovered completely. On the contrary, the activity of denatured mature HlyA was not recovered even when the concentration of denaturant was decreased to 0.1 M.
As trypsin activates pro-HlyA by removal of the pro-region (21), the influence of the denaturation and renaturation on activation was investigated. Pro-HlyA and mature HlyA were denatured in 6 M guanidine HCl. These solutions were then diluted to give a final concentration of guanidine HCl of 0.1 M, and treated with trypsin to activate pro-HlyA. The activity of the renatured and activated pro-HlyA was 222 HLU/nmol, which was comparable with untreated pro-HlyA (232 HLU/ nmol) (Table II). However, the recovery of activity with the mature HlyA was poor (Ͻ10 HLU/nmol, less than 3% of undenatured control). These results suggest that the mature HlyA is unable to fold itself and that the HlyA pro-region may contribute to the refolding of the denatured toxin.
Homology between Pro-region and Molecular Chaperones-Molecular chaperones (22,23) are considered to mediate the folding of newly synthesized proteins in vivo and are also known to promote folding of denatured proteins in vitro. If the pro-region and molecular chaperones have similar effects, some structural homology among these proteins would be expected. We compared the amino acid sequence of the pro-region with those of chaperones in EBI and SWISS-PROT data bases. A highly homologous region was observed between the pro-region and a heat shock protein, Hsp90, which is a member of a family of molecular chaperones (15) found in Escherichia coli from humans (24 -26) (Fig. 4). The homologous region found in Hsp90 is highly conserved among the Hsp90 family proteins (24 -26). In the HlyA pro-region, this homologous region constitutes about one-half (68 residues) of the entire pro-region  (24 -26) are aligned with that of the pro-region of El Tor cytolysin (2) (pro). Identical (dark shading) and similar (light shading) amino acids between the pro-region sequence and Hsp90 protein sequences are indicated. Horizontal bars in the sequences represent gaps to maximize the alignment. The numbers following each line indicate the amino acids at the beginning and end of each sequence. Note that the sequence of HtpG of V. cholerae is not available after the 123rd amino acid.

FIG. 5. Construction of the pro-region expression system. The
NruI and HpaI sites, which are located at the 30th and 634th nucleotide of the hlyA gene on pNK105, respectively, were converted to BamHI and HindIII sites, respectively, with linkers. This BamHI-HindIII fragment encoding the pro-region was ligated to the BamHI and HindIII sites of an expression vector, pET25b(ϩ) in frame, resulting in pNK118. The solid black lines represent hlyA, and the gray-shaded portions (pro) indicate the pro-region coding domain. The horizontal line shading denotes the polypeptide coding domain that was present in the vector beforehand. Parentheses indicate restriction sites broken by linker ligation. N, Hp, B, and H denote NruI, HpaI, BamHI, and HindIII sites, respectively.
(132 residues) and is located in the central section. This homologous region had 34% identity and 63% similarity to Hsp90.
Expression System of Pro-peptide Fusion Protein-To examine whether the HlyA pro-region can interact intermolecularly with its target, mature region, similar to native molecular chaperones, we constructed a pro-region expression system (Fig. 5). The pro-region of the hlyA gene was subcloned into the BamHI and HindIII sites of an expression vector, pET25b(ϩ) in frame, resulting in pNK118. E. coli BL21(DE3) harboring pNK118 expressed a 30-kDa protein (Pro-peptide) after induction with 1 mM isopropyl-1-thio-␤-D-galactopyranoside. This molecular size agreed with the estimated size (31 kDa) of the pro-region fusion product of pNK118 (data not shown).
In Vivo Complementation of HlyA⌬Pro by the Pro-peptide in E. coli-pNK118 (pro) was introduced into E. coli BL21(DE3) harboring pNK108 (hlyA⌬pro) by transformation. The transformants were cultured in LB broth containing 100 g/ml of ampicillin and 50 g/ml of kanamycin, and the Pro-peptide encoded on pNK118 was induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside. After a 30-min induction, the cells were sonicated and centrifuged, and the hemolytic activities of the supernatants were determined. The hemolytic activity of E. coli harboring pNK107 (hlyA) and pNK118 (pro) simultaneously was 0.1 HLU/ml. E. coli harboring both pNK108 (hlyA⌬pro) and the vector plasmid, pET25b(ϩ), gave no detectable hemolytic activity. E. coli harboring pNK108 (hlyA⌬pro) and pNK118 (pro), however, had about 40% of the activity (0.04 HLU/ml) compared with the control, E. coli harboring pNK107 (hlyA) and pNK118. This indicates that the expression of Propeptide complemented the defective cytolysin (HlyA⌬Pro) in vivo.
In Vitro Complementation of Renaturation of Denatured Cytolysin by Pro-peptide-To investigate if the isolated pro-region may support formation of active cytolysin in vitro, the Propeptide was overexpressed and purified from E. coli BL21(DE3) harboring pNK118 using a pET expression system. To rule out possible influence of the Pro-peptide on hemolytic activity of native toxin, mature HlyA was mixed with Pro-peptide, incubated for 1 h, and the hemolytic activity was determined. The hemolytic activity of the mature HlyA was unaffected even when 100-fold concentration of the Pro-peptide molecules was added (Fig. 6).
mature HlyA was denatured with 6 M guanidine HCl and diluted 60 times with renaturing buffer with and without the Pro-peptide and incubated for 1 h (Fig. 6). When denatured cytolysin was diluted without Pro-peptide, little activity was recovered. Depending on the concentration of the Pro-peptide in the renaturation solution, however, hemolytic activity was recovered. More than 80% activity was restored when the molar ratio of Pro-peptide to mature HlyA was 40 or higher. Bovine serum albumin, used as a control instead of the Propeptide, had no effect, even at a concentration 1200 times higher than that of mature HlyA (data not shown). Incubation for shorter and longer times following dilution did not affect the recovery (data not shown). These findings indicate that the isolated pro-region facilitates the formation of denatured mature HlyA in vitro.

DISCUSSION
In this study, we investigated the possible folding ability of the pro-region of El Tor cytolysin (HlyA). The guanidine HCldenatured pro-HlyA recovered hemolytic activity following dilution of the denaturant; the mature HlyA lacking the proregion was not activated (Fig. 3 and Table II). This indicates that the pro-region may support the folding of HlyA. The proregion-deleted HlyA, HlyA⌬Pro, was not secreted and was degraded in the periplasmic space (Fig. 2). These results suggest that the degradation of HlyA⌬Pro in vivo may be due to its failure to fold in the periplasmic space.
Molecular chaperones (22,23) are known to enhance the formation of the native conformation of other proteins. In the organelles (e.g. mitochondria and endoplasmic reticulum), molecular chaperones facilitate the folding of proteins exported across membranes. In bacterial cells, although several cytoplasmic molecular chaperones have been found and well characterized (22,23), few chaperones have been reported in the periplasm (20). In the absence of chaperone, the HlyA proregion is likely to function in a manner similar to molecular chaperones and be effective in the periplasm. Of interest, V. vulnificus cytolysin, which is evolutionarily close to El Tor cytolysin, does not have a pro-region (27). Why El Tor cytolysin has a special domain for folding remains unclear.
The HlyA pro-region has a domain highly homologous to Hsp90, a family of molecular chaperones that stimulates protein folding (15). This chaperone is widely distributed and found in bacteria from humans (24 -26). The region in Hsp90, which is similar to the pro-region of El Tor cytolysin, is one of the most conserved regions among the Hsp90 family proteins (24 -26). The pro-region was found to act like a chaperone not only intramolecularly but also intermolecularly. These results suggest that the HlyA pro-region and molecular chaperone Hsp90 may share a common mechanism for supporting protein folding. These findings raise the intriguing possibility that the HlyA pro-region may have structural and functional similarities to molecular chaperones.
Recently, the pro-regions of precursor proteins for a few FIG. 6. Influence of the Pro-peptide on the hemolytic activity of HlyA. A, influence of the Pro-peptide on the hemolytic activity of native mature HlyA. 1.25 nmol/ml (40 l) purified mature HlyA were incubated with and without 125 nmol/ml purified Pro-peptide for 1 h at 20°C and then hemolytic activity was determined. Each value is the mean of three experiments. B, renaturation of denatured mature HlyA with the Pro-peptide. 1.25 nmol/ml of mature HlyA was denatured and renatured as described under "Experimental Procedures," and hemolytic activity was determined. Renaturation was carried out in renaturation solution containing the indicated concentration of Pro-peptide. The activity of samples subjected to the same treatment without denaturation is taken to be 100%. Each value is the mean of three experiments. serine proteases, subtilisin (3), ␣-lytic protease (4) and carboxypeptidase Y (5), have been shown to have chaperone-like ability in vitro. To distinguish them from molecular chaperones, these pro-regions are called intramolecular chaperones. All these chaperone-like peptides are part of precursor proteins and the peptides possessing chaperone-like function only for the folding of the precursors (6, 7). All the intramolecular chaperones known so far, including the HlyA pro-region, are found in the N terminus of the precursor proteins.
The pro-regions of subtilisin and ␣-lytic protease have been reported to act not only as intramolecular chaperones but also like molecular chaperones (3,4,7); that is, these pro-regions fold their target peptides when these pro-and the mature regions are expressed in trans. The pro-regions of these proteases have no significant homology to the known molecular chaperones, including Hsp90. The HlyA pro-region has no significant homology with the pro-regions of other proteases either. HlyA pro-region may function in a manner different from the pro-regions of these proteases.