INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Chitin, a homopolymer of N-acetylglucosamine (GlcNAc), is a major component of the cell walls of fungi and the integuments of crustaceans and insects (38). The molecule is one of the most abundant biopolymers in nature and is used by many microorganisms as a source of carbon. Utilization of chitin as a nutrient usually requires degradation of the molecule to GlcNAc monomers. Complete degradation of chitin in both prokaryotes and eukaryotes is a two-step process which involves successive hydrolysis of the ,1-4 glycosidic bonds linking the GlcNAc subunits. In the first stage, endochitinase binds and degrades tetrameric and longer polymeric forms of GlcNAc to produce the disaccharide chitobiose. In the second step, chitobiase hydrolyzes chitobiose to GlcNAc monomers. The enzymes for chitin degradation are usually coordinately regulated and in several organisms are induced by chitosan, chitobiose, GlcNAc, or glucosamine (2, 7, 44).

Members of the family Vibrionaceae thrive in marine environments where chitin is abundant. It is not surprising that many Vibrionaceae evolved systems for utilizing chitin as a nutrient source. Chitinases have been identified in Vibrio vulnificus (56, 61), V. harveyi (57), and V. parahemolyticus (29, 30). Nucleotide sequence analysis indicated that the chitinase of V. harveyi is homologous with human hexosamindase, indicating that the two enzymes, as well as other chitinases, are members of a phylogenically related group (56).

V. cholerae is a human intestinal pathogen that resides in brackish and marine waters. In vitro experiments established that V. cholerae has the potential to use chitin as a sole source of carbon for growth (15). It is likely, therefore, that production of chitinase (29, 30, 42) by V. cholerae provides the bacterium with a readily available nutrient source in aquatic environments. Hydrolysis of chitin by V. cholerae is an extracellular process that requires expression of epsE, one of a cluster of genes in the eps locus (43, 46-48). Several proteins of V. cholerae are dependent on the eps system for extracellular transport, including cholera toxin (CT), an undefined protease, and a chitinase activity (43, 48). Although expression of chitinase activity has been reported for V. cholerae, the enzyme responsible for the activity has not been identified. To further characterize the extracellular chitinase of V. cholerae, we cloned a gene encoding chitinase activity. Here we report the nucleotide sequence of a cloned endochitinase gene and establish that the protein encoded by that gene is secreted to the extracellular environment by an eps-dependent mechanism.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, and reagents. Bacterial strains and plasmids used in this investigation are listed in Table 1. Escherichia coli strains were cultured in Luria-Bertani (LB) broth. Classical biotypes of V. cholerae were cultured in Syncase broth (26), El Tor biotypes of V. cholerae were grown in YEP broth (27), and Aeromonas hydrophila was cultured in Tryticase soy broth (Difco Laboratories, Detroit, Mich.). All strains were maintained on LB agar. Chemical reagents were obtained from Sigma Biochemicals (St. Louis, Mo.), Life Technologies, Inc. (Gaithersburg, Md.), and Fisher Scientific (Springfield, N.J.). Unless otherwise noted, ampicillin was used at 150 µg/ml, chloramphenicol was used at 10 µg/ml, tetracycline was used at 10 µg/ml, and kanamycin was used at 50 µg/ml. All antibiotics were purchased from Sigma Biochemicals.

Preparation of a genomic library of V. cholerae 569B. Preparation of a genomic library of 569B was previously reported (10). Chromosomal DNA from V. cholerae 569B was prepared by standard methods, partially digested with the restriction enzyme Sau3AI, and size fractionated by sucrose gradient centrifugation. DNA fragments 25 to 50 kbp in size were pooled and ligated to BamHI-digested cosmid vector pCOS5 (10). Ligated DNA was packaged into bacteriophage lambda capsids, using a Gigapack packaging extract (Stratagene Cloning Systems, La Jolla, Calif.), which were transfected into E. coli LE392. The average insert size of the cosmid clones was 37 kbp.

Preparation of EGC agar. LB agar medium containing ethylene glycol chitin (EGC) was prepared by mixing 600 µl of an aqueous solution of EGC (10 mg/ml; Sigma Biochemicals) and 160 µl of 1% aqueous solution of trypan blue with 16 ml of molten (56°C) LB agar. Antibiotics were added, as appropriate. EGC plates inoculated with chitinase-producing strains were incubated at 37°C until clear halos surrounding the colonies were detected.

Isolation and manipulation of plasmid DNA. Plasmids were obtained from E. coli by using a modified alkaline lysis method (25) and ethidium bromide extraction (58) or by use of PlasmidPure spin filters (Sigma Biochemicals). Restriction enzymes and DNA-modifying enzymes were purchased from Life Technologies. Agarose was purchased from J. T. Baker (Phillipsburg, N.J.). DNA fragments used for subcloning were isolated from agarose gel slices by using a GeneCleanII kit (Bio 101, Inc., La Jolla, Calif.). Plasmids were transformed into E. coli by osmotic shock (35), and transformants were selected on LB agar containing appropriate antibiotics.

DNA sequencing. Double-stranded DNA sequencing was performed at the Sequencing Facility of the Center for Advanced Molecular Biology and Immunology at the State University of New York at Buffalo, using fluorescent dye primer or dye terminator chemistry. Synthetic single-stranded oligonucleotide primers used to initiate sequencing and for subsequent PCR were purchased from Integrated DNA Technologies, Inc. (Coralville, Iowa). The nucleotide sequences of the synthetic oligonucleotides used for sequencing the insert in pTDCC2 were Chi-1 (5'-[741]GCTGTTCACTGCCCGTTG-3'), Chi-2 (5'-[732]CAACGGGCAGTGACAGC-3'), Chi-3 (5'-[1504]GGCCAAACCGTTGGTCTA-3'), Chi-4 (5'-[1474]TAGACCAACGGTTTGGCC-3'), Chi-5 (5'-[1936]CAAGCAAGATATGAAAGC-3'), Chi-6 (5'-[1994]CCTTCAGCGCCACC-3'), Chi-7 (5'-[252]CGGAGTGCCAGTGTG-3'), Chi-8 (5'-[304]GAACTTTATCACTGCCAA-3'), Chi-9 (5'-[2167]AAAAACATCGGTGGTGAT-3'), Chi-10 (5'-[2319]ACATGCAGAATATCAAG-3'), Chi-11 (5'-[2253]GTGGAATTTGGGGCGC-3'), Chi-12 (5'-[2618]CTGCATAATTGGGGTAG-3'), and Chi-13 (5'-[2477]GCCATTGGTTTGCCATCAGG-3'). The numbers in brackets denote the position of the first nucleotide relative to the sequence in Fig. 2. Positive numbers are positions on the sense strand, while negative numbers indicate positions on the antisense strand. The T7 (5'-TAATACGACTCACTATAGGG-3') and T3 (5'-ATTAACCCTCACTAAAGGGA-3') primers are homologous to 5' and 3' regions of pBluescriptKS and pBluescriptSKII+ flanking the multicloning site (Stratagene).

SDS-PAGE and Western (immunoblot) assays. Isoelectric focusing-grade acrylamide and bisacrylamide were purchased from Pharmacia Biotechnology (Piscataway, N.J.). Unless otherwise noted, samples were prepared by solubilizing at 100°C for 10 min in a buffer containing 31 mM Tris (pH 6.8), 1% sodium dodecyl sulfate (SDS), 2.5% 2-mercaptoethanol, and 5% glycerol. Proteins in the solubilized samples were resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) on 8.75% gels (33) which were stained for proteins with silver (37) or colloidal Coomassie brilliant blue (39). Proteins for immunoblotting were resolved by SDS-PAGE and electrophoretically transferred to nitrocellulose paper. Blots were blocked for 30 min in a 5% (wt/vol) solution of skim milk in phosphate-buffered saline and incubated in phosphate-buffered saline-5% skim milk to which mouse monoclonal antibody to the E-tag epitope (1:200; Pharmacia Biotech), rabbit anti-maltose binding protein (MBP) antiserum (1:2,000; New England Biolabs, Inc., La Jolla, Calif.), or rabbit anti-MBP-ChiA-E-tag antiserum (1:15,000) had been added. Affinity-purified rabbit anti-mouse immunoglobulin G antibodies (1:2,000; Sigma Biochemicals) were used as a second antibody when immunoblots had been initially probed with the mouse anti-E-tag monoclonal antibody. To detect antibody-bound polypeptides, immunoblots were probed with a horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (1:10,000; Sigma Biochemicals) and developed with 4-chloro-1-naphthol (20). In some cases, immunoblots were developed with a luminol reagent kit (DuPont/New England Nuclear, Wilmington, Del.). Fluorescent signals from the luminol blots were detected by exposure on Biomar Blue-sensitive autoradiographic film (Marsh Biomedical Products, Inc., Rochester, N.Y.). A Vista-S6E scanner (UMAX Technologies, Inc., Fremont, Calif.) and Molecular Analyst/PC software (Bio-Rad Laboratories, Hercules, Calif.) were used for semiquantitative densitometric analysis of the exposed films.

Preparation of culture supernatants and periplasmic extracts. Culture supernatants and periplasmic extracts of bacterial cultures having an optical density at 600 nm of between 4.1 and 5.6 for late-log-phase cultures and between 4.5 and 5.4 for stationary-phase cultures were obtained as previously described (9, 11).

GM₁ ELISA for CT. Methods to measure CT in samples using ganglioside GM₁-dependent enzyme-linked immunosorbent assay (ELISA) and a rabbit anti-CT antiserum were previously described (11). Purified CT (a gift from R. K. Holmes) was used as a standard.

Biochemical assays for chitinase activity. Chitinase activity was assayed by using chromogenic analogues of chitin (Sigma Biochemicals). For assays using p-nitrophenyl-N-acetyl--D-glucosaminide (p-NAG) (61), fresh colonies cultured overnight on LB agar were transferred by using a sterile toothpick onto small squares of Whatman 3MM filter paper (Schleicher & Schuell, Keene, N.H.). Fifteen microliters of p-NAG diluted to 3.6 mg/ml in 0.1 M sodium phosphate buffer (pH 7.4) was pipetted onto each colony. After being wetted with water, filter papers were incubated for 30 min at 37°C and observed for production of a bright yellow color which indicated chitinase (chitobiase) activity. 4-Methylumbelliferyl (4-MU)-N-acetyl--D-glucosaminide, 4-MU-N-acetyl--D-glucosaminide, 4-MU-N-acetyl--D-galactosamide, 4-MU--D-N,N'-diacetylchitobioside, and 4-MU--D-N,N',N"-triacetylchitotrioside were used for detection of chitinase activity (41). Conjugates dissolved initially in a small amount of dimethylformamide were diluted in 0.1 M sodium phosphate buffer (pH 7.4) to a final concentration of 1.43 mM. Fresh colonies from overnight LB agar plates were transferred to squares of filter paper, and 15 µl of a 4-MU conjugate was pipetted to the side of the colony and allowed to diffuse into the cells by capillary transfer. After 15 min of incubation at 37°C, a few drops of an aqueous saturated sodium bicarbonate solution was pipetted onto the colony to enhance fluorescence. Colonies were observed under 366-nm UV light for blue fluorescence which indicated chitinase-mediated hydrolysis of the conjugates to 4-methylumbelliferone.

Construction of a frameshift mutation in chiA. NheI-digested and Klenow-repaired pTDCC2 was self-ligated under conditions that favored recircularization of the linearized, blunt-ended plasmid. The ligation mixture was transformed into E. coli DH5F'tet, and transformants were selected by plating onto LB agar containing ampicillin and tetracycline. A plasmid that had lost the NheI site was isolated and designated pTDCC2.3.

Construction of promoter probe plasmids. DNA fragments were obtained by PCR from regions upstream of the chiA gene in pTDCC2. Synthetic oligonucleotides used as PCR primers were Chi-15 (5'-GGAATT[29]CCGAAAAAGAATTGAAAGC[47]-3'), Chi-17 (5'-GGAATTCC[99]ATTGATTAACATGCA[113]-3'), Chi-18 (5'-GGAA[123]TTCACTCTGGAGTATTAG[140]-3'), Chi-19 (5'GGAAT[157]TCTAATACAGGAGAGAAACA[176]-3'), and Chi-20 (5'-CG[465]TTGGCGTTATTGAGTG[-449]-3'). Brackets denote nucleotide positions, as noted above; restriction sites for EcoRI are denoted by single underlining, while the site for BamHI is denoted by a double underline. pTDCC2 was used as a template in PCR using the following reaction conditions: 30 s at 92°C, 45 s at 45°C, and 60 s at 72°C (30 cycles). DNA fragments obtained from the reactions were digested with EcoRI and BamHI and ligated into the same sites of the multicloning site of the promoter probe vector pRS415 (53).

Engineering a chiA mutant of 569B. The suicide plasmid pKAS32 (55) was used for allelic exchange mutagenesis. The chiA gene was insertionally inactivated by ligating an end-repaired EcoRI fragment of pUC4K (Pharmacia LKB Biotechnology) into the unique intragenic HpaI site of pTDCC2. This intermediate plasmid was designated pATK-1. A KpnI/SstI fragment of pATK-1 was subsequently ligated into equivalent sites of the suicide plasmid pKAS32 to produce pJPF3. The suicide plasmid was then transformed into E. coli S17pir (36) and conjugated into 569Bstp10rif, a spontaneous streptomycin- and rifampin-resistant mutant of 569B. A single ampicillin-resistant transconjugant having pJPF3 integrated into the genome was grown overnight in LB broth without antibiotic selection at 37°C. Overnight cultures were plated on LB agar containing streptomycin (30 µg/ml) to select for clones in which the plasmid sequences had been deleted from the chromosome by recombination between the wild-type and mutant copies of the chiA genes. Streptomycin-resistant clones were screened for loss of plasmid-encoded ampicillin resistance and for loss of chitinase activity on EGC agar. A transconjugant that was ampicillin sensitive and chitinase negative was designated 569B(chiA::Kan^r). PCR analysis using the oligonucleotides Chi-2 and Chi-4 and Southern hybridization using pKAS32 and a PCR-derived fragment of chiA as hybridization probes were used to confirm the allelic exchange of chiA in 569B(chiA::Kan^r).

Engineering a malE-chiA fusion. PCR was used to facilitate engineering a malE-chiA chimera by fusing the malE gene to a fragment encoding the entire structural gene of chiA excluding the putative signal peptide. A DNA fragment containing the chiA gene was amplified from pTDCC2.6, a recombinant plasmid in which sequences encoding a 13-amino-acid E-tag epitope (Pharmacia Biotech) was fused in frame to the 3' end of chiA. Oligonucleotide primers used in the PCR were Chi-Mal (5'-GGTCTAGA[241]TATAACTGTGCCGGAGTG-3') and Blue-619 (5'-GTAAAACGACGGCCAGTGA-3'). The numbers in brackets denote the position of the first nucleotide in the sense strand relative to the sequence in Fig. 2. Use of Chi-Mal in the PCR incorporated an XbaI site (underlined) into the amplified DNA fragment at a position immediately 5' to the triplet codon for tyrosine-23, the predicted NH₂-terminal amino acid of the mature ChiA protein. Blue-619 is homologous to a region of the multicloning site in pTDCC2.6 downstream of the 3' ligation joint of the DNA insert. The amplified DNA was digested with the restriction enzymes XbaI and HindIII, and the resulting 2.5-kbp fragment was ligated to the expression vector pmal-p2 in the XbaI-HindIII sites to produce an in-frame hybrid of malE to chiA. After transformation of the ligation mixture into E. coli DH5F'tet, clones were analyzed by restriction mapping and one clone harboring a plasmid with an appropriate restriction map was screened on EGC agar to confirm a ChiA⁺ phenotype. The junction sites of plasmid pJL1 were sequenced to confirm the in-frame fusion.

Amylose affinity chromatography. The MBP-ChiA-E-tag fusion protein was purified by established methods using sucrose extraction (18, 19) and amylose affinity chromatography (New England Biolabs). The purity of the MBP-ChiA fusion protein was determined by SDS-PAGE and immunoblotting using a rabbit anti-MBP antiserum (New England Biolabs) and an anti-E-tag mouse monoclonal antibody (Pharmacia Biotech).

Preparation of rabbit antiserum. Hyperimmune antiserum to the affinity-purified MBP-ChiA fusion protein was obtained from subcutaneous immunization of a New Zealand White rabbit (5). Immunizations and serum collection were performed by the Monoclonal Antibody Center at the State University of New York at Buffalo.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Cloning of chiA. Chitinases are commonly produced by members of the family Vibrionaceae (61). To determine if chitinase was produced by V. cholerae 569B, EGC agar was stab inoculated with the strain. After 24 h of incubation, the site around the colony was surrounded by a large zone of clearing, indicating hydrolysis of the EGC (data not shown). To isolate the gene or genes that encoded the chitinase activity, approximately 800 cosmid clones of a 569B genomic library were screened for the ability to hydrolyze EGC. Three cosmid clones were identified by the ability to elicit a zone of clearing in the EGC agar. To determine if the cosmid clones had common DNA inserts, cosmid DNA prepared from each clone was digested with the restriction enzyme BamHI and the fragments were resolved by agarose gel electrophoresis. All three cosmids had identical BamHI restriction patterns (data not shown). Similar results were obtained when the cosmids were digested with SalI, providing strong evidence that the inserts of the three cosmids were identical. The cosmid from one of the clones was designated pTDCC1 (Fig. 1, streak 2).

View larger version (60K): [in this window] [in a new window]   FIG. 1.   Detection of chitinase activity produced by chiA clones grown on EGC agar. (A) No IPTG induction; (B) induction with 0.1 mM IPTG. Streaks: 1, pCOS5; 2, pTDCC1; 3, pBluescriptKS; 4, pTDCC2; 5, pTDCC2.6; 6, pmal-p2; 7, pJL1. E. coli DH5F'tet was used as a host strain for all plasmids except pTDCC1, which was expressed in LE392.

Confirming the ORF by frameshift mutagenesis. To confirm that the ORF in pTDCC2 encoded the chitinase activity observed on EGC agar, we used a unique NheI site to engineer a 1 frameshift mutation in the putative gene (Fig. 2). Chitinase activity was not produced by E. coli DH5F'tet cells which had been transformed with pTDCC2.3, the plasmid encoding the mutated ORF (Table 2). Based on the analysis of the nucleotide sequence and phenotypic characterization of the mutated plasmid, the 2,538-bp ORF in pTDCC2 was designated chiA.

View larger version (77K): [in this window] [in a new window]   FIG. 2.   Nucleotide sequence of the DNA insert in pTDCC2. A promoter region at positions 29 to 97 is denoted by double underlining; single underlining indicates a putative ribosomal binding site (RBS). A potential signal peptidase I cleavage site is marked by //. A putative transcription termination sequence located at 2744 to 2774 is double underlined. Nucleotides comprising the ORF of chiA are in uppercase letters; flanking regions are in lowercase letters. The predicted amino acid sequence of ChiA is designated in single-letter amino acid code.

Use of allelic exchange to engineer a chiA mutant of 569B. To demonstrate that the chiA gene encoded the major chitinase activity of 569B, the wild-type copy of the gene in the chromosome was replaced by an insertionally inactivated copy of chiA. The mutant, 569B(chiA::Kan^r), exhibited little or no endochitinase activity on EGC agar (Fig. 3). In addition, Western blotting of 569B(chiA::Kan^r) with the anti-ChiA antiserum confirmed that the mutant did not produce immunoreactive protein (data not shown). We surmise that the slight residual chitinolytic activity evident in the EGC agar assay was due to nonspecific chitin degradation. An alternative explanation is that V. cholerae likely expresses, in addition to endochitiase, a chitobiase which degrades chitobiose, an intermediate molecule in the degradative pathway of chitin. It is possible that chitobiase has some minor reactivity against EGC.

View larger version (49K): [in this window] [in a new window]   FIG. 3.   An insertional mutation in chiA abolishes endochitinase activity in V. cholerae 569B. Production of chitinase activity was measured by culturing wild-type (569B) and mutant [569B(chiA::Kanr)] V. cholerae strains on EGC agar. E. coli DH5F'tet cells containing pTDCC2 and the phagemid vector pBluescriptKS served as positive and negative controls, respectively.

Mapping the chiA promoter. Expression of chiA was likely a result of the activity of an endogenous promoter within the 3.0-kbp insert of pTDCC2, since nucleotide sequence analysis of the pTDCC2 showed that chiA was positioned in the orientation opposite that of the lac promoter in the vector. Rather than having no effect on expression, addition of isopropyl--D-thiogalactopyranoside (IPTG) to cultures of E. coli DH5F'tet(pTDCC2) caused a decrease in chitinase activity (Fig. 1, streak 4). This observation suggested that the promoter was present in the 174 nucleotides that preceded the ATG initiation codon of the gene. To map the promoter of chiA more precisely, four DNA fragments isolated by PCR from the 5' end of the insert of pTDCC2 (nucleotides 31 to 465, 99 to 465, 123 to 465, and 157 to 465) were cloned into pRS415 (53), a vector engineered for promoter analysis. When introduced into 569B, only pZ1 expressed high levels of -galactosidase activity; pZ3 containing nucleotides 98 to 465 expressed very little -galactosidase, as did plasmids containing shorter segments of the DNA upstream of chiA (Table 3). These results strongly suggested that promoter activity for chiA was located within a 69-bp region bounded by nucleotides 29 and 97 (Fig. 2). Sequences homologous to consensus 35 and 10 hexamers which are components of most ⁷⁰-like promoters were not evident in the 68-bp region. It should be noted that 569B exhibited a low level of endogenous -galactosidase activity (pRS415 [Table 3]), but this background did not interfere with the measurements of promoter activity.

Defining the substrate specificity ChiA. To determine the substrate specificity of the enzyme encoded by chiA, we employed a rapid test method that used a series of 4-MU conjugates (41). Hydrolysis of the glycosidic bonds of the conjugates releases 4-umbelliferone, which fluorescences with a strong blue light when illuminated with UV irradiation. Five different 4-MU conjugates were used in these experiments: 4-MU-N-acetyl--D-glucosaminide, 4-MU-N-acetyl--D-glucosaminide, 4-MU-N-acetyl--D-galactosaminide, 4-MU--D-N,N'-diace-tylchitobiose, and 4-MU--D-N,N',N"-triacetylchitotriose. When expressed in DHF'tet, pTDCC2 directed the hydrolysis of 4-MU--D-N,N'-diacetylchitobiose and 4-MU--D-N,N',N"-triacetylchitotriose (Table 2). DH5F'tet(pTDCC2) did not react with 4-MU-N-acetyl--D-glucosaminide, 4-MU-N-acetyl--D-glucosaminide, or 4-MU-N-acetyl--D-galactosaminide. DH5F'tet(pTDCC2) also did not hydrolyze p-NAG, a conjugate that releases a yellow pigment when the glycosidic bonds are cleaved (61). A minor amount of hydrolytic activity for 4-MU-N-acetyl--D-glucosaminide was evident after prolonged incubation of the substrate on DH5F'tet(pTDCC2) cells. It was concluded from the pattern of degradation of the synthetic substrates that ChiA encoded by pTDCC2 was likely an endochitinase capable of hydrolyzing ,1-4 glycosidic bonds.

Production of anti-ChiA hyperimmune serum. To expedite the process of obtaining purified ChiA for raising antibodies, a malE-chiA fusion was constructed by using the expression vector pmal-p2 (New England Biolabs) and a PCR-amplified fragment of pTDCC2.6. Ligation of the amplified fragment into pmal-p2 by directional cloning positioned the chiA gene into the vector such that an in-frame fusion to the 3' end of malE fusion was produced. Introduction of the fusion plasmid pJL1 into DH5F'tet produced a strain that expressed strong IPTG-inducible chitinase activity on EGC agar (Fig. 1, streak 7).

View larger version (73K): [in this window] [in a new window]   FIG. 4.   Expression of recombinant ChiA. Proteins in the immunoblot were detected with a rabbit anti-ChiA hyperimmune antiserum. All plasmids were expressed in E. coli DH5F'tet with the exception that pTDCC1 was expressed in E. coli LE392. Molecular mass standards are in kilodaltons.

Expression of ChiA in classical and El Tor biotypes of V. cholerae and in A. hydrophila. Bacterial cultures of two classical biotypes (569B and 0395) and two El Tor biotypes (U1 and JBK70) of V. cholerae were analyzed for reactivity to the anti-ChiA antiserum. All four strains synthesized immunoreactive polypeptides that were similar in size to the recombinant polypeptide encoded by pTDCC2 (Fig. 5). A smaller immunoreactive polypeptide having an apparent molecular mass of 65 kDa was also present in all four strains. No protein of V. cholerae corresponding to the 42-kDa MalE of E. coli was detected in the anti-ChiA immunoblots.

View larger version (67K): [in this window] [in a new window]   FIG. 5.   Expression of chitinase by classical (569B and 0395) and El Tor (JBK70 and U1) strains of V. cholerae and by A. hydrophila. Bacterial cultures comprised cells and culture supernatants were used for the immunoblotting analysis. Proteins in the immunoblot were detected with a rabbit anti-ChiA hyperimmune antiserum. The antiserum did not react with proteins in fresh culture media (data not shown). pTDCC2 was expressed in E. coli DH5F'tet. Molecular mass standards are in kilodaltons.

Extracellular transport of ChiA. Chitinases in other species are typically extracellular proteins. To establish whether ChiA was transported to the extracellular medium by V. cholerae, culture supernatants and periplasmic extracts isolated from late-log-phase and stationary-phase cultures of 569B were analyzed by SDS-PAGE and immunoblotting for ChiA (Fig. 6A). In late-log-phase cultures of 569B, two immunoreactive polypeptides were found in the culture supernatant. The major polypeptide had an apparent molecular mass of 90 kDa, which was equivalent in size to recombinant ChiA. A minor, faster-migrating polypeptide of 76 kDa was also evident. Periplasmic extracts of 569B cultures contained only the 90-kDa polypeptide. Semiquantitative densitometric analysis of the immunoblots demonstrated that the majority of the total immunoreactive protein (69.4%) of 569B was located in the culture supernatant of late-log-phase cultures (Fig. 6).

View larger version (61K): [in this window] [in a new window]   FIG. 6.   Secretion of ChiA by V. cholerae depends on epsE. (A) Late-log-phase culture; (B) stationary-phase culture. M14 is an epsE mutant derived from 569B. pTDCepsE is a clone of the wild-type epsE gene from 569B. pTDCC2 and pBluescriptKS were expressed in E. coli DH5F'tet. Proteins in the immunoblot were detected with a rabbit anti-ChiA hyperimmune serum. S, culture supernatant; P, periplasmic extract. Molecular mass standards are in kilodaltons.

View larger version (33K): [in this window] [in a new window]   FIG. 7.   Localization of ChiA and CT in wild-type and mutant V. cholerae. Percentage values represent the relative amounts of ChiA and CT in the respective samples. (A) Total immunoreactive ChiA in culture supernatants and periplasmic extracts. Semiquantitative data were obtained from densitometric analysis of the immunoblots in Fig. 6. (B) Total CT in the culture supernatants and periplasmic extracts used for the immunoblot shown in Fig. 6. CT in the samples was measured by GM1 ELISA (4, 11).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Mechanisms for extracellular transport of proteins have been identified in a diverse number of bacteria. Klebsiella oxytoca (44), Erwinia chrysanthemi (22), Erwinia carotovora (34), Pseudomonas aeruginosa (1), Xanthomonas campestris (13), and A. hydrophila (3) contain a cluster of genes that encode the main terminal branch of the general secretory pathway (GSP). Recently, it was determined that V. cholerae has a cluster of genes that are homologous to the secretory clusters in those bacteria. At least 12 genes are encoded by the eps cluster of V. cholerae (43, 47, 48). Extracellular transport of CT, an oligomeric enterotoxin produced by the bacterium, is dependent on expression of the eps genes (11, 43, 47, 48). Here we provide strong evidence that the eps system of V. cholerae is also involved in extracellular transport of ChiA.

The degree of shared homology observed among the secretory systems of the diverse bacteria does not enable the systems to be interchanged. In most cases, extracellular proteins are secreted only by their cognate secretory systems. For example, E. chrysanthemi secretes a pectate lyase, but the pectate lyase of E. chrysanthemi is not secreted by K. oxytoca (22). The type I heat-labile enterotoxin LT-I, a protein that is highly homologous to CT, is secreted by V. cholerae and seven species of the family Vibrionaceae but not by E. chysanthemi, Xanthomonas maltophilia, or K. pneumoniae (43). Of the secretory systems that have been well described, it appears that the eps system of V. cholerae is the most promiscuous. Previous investigations established that the eps system of V. cholerae promotes secretion not only of CT and LT-I but also of the B polypeptides of LT-IIa and LT-IIb, two members of the type II heat-labile enterotoxins of enterotoxigenic E. coli (11). ChiA can now be added to the list of extracellular proteins secreted by the eps system. The ability of CT, LT-I, LT-IIa, LT-IIb, and ChiA to traffic through the same secretory pathway suggests that each of the proteins has an extracellular transport signal that is recognized by the eps system. The molecular structures of the signals in the five proteins have not been identified, but there is growing evidence that the signals are not composed of a linear array of conserved amino acids. CT and the type II enterotoxins have little if any amino acid homology yet are secreted with equal efficiency by V. cholerae (11). From those observations, it was hypothesized that the transport signals in these three proteins were likely conserved conformation-dependent motifs. Since ChiA is not homologous to CT or to the type II enterotoxins, it is likely that the extracellular transport signal in ChiA is also a conformation-dependent motif. A similar situation was observed in K. oxytoca, where two nonadjacent regions of pullulanase were required to promote translocation of a -lactamase fusion protein across the outer membrane (50, 51). It is possible that many if not all extracellular transport signals in proteins transported by eps-like secretory systems are comprised of conformation-dependent domains. If that is indeed the case, experiments to define the structures of the transport signals may require high-resolution crystal structures of the proteins.

When expressed in E. coli, CT and the type II heat-labile enterotoxins LT-IIa and LT-IIb accumulate in the periplasm of the cell. A similar pattern of periplasmic accumulation of extracellular proteins was observed when the genes for amylase (17), aerolysin (24), and protease (45) of A. hydrophila and the gene for pullulanase of K. pneumoniae (12) were expressed in E. coli. For pullulanase, mobilization of the proteins out of the periplasm of E. coli required coexpression of the pul cluster, the cognate secretory system (12). However, when the genes for CT, LT-IIa, and LT-IIb were expressed in E. coli, low but significant amounts of the proteins were found in the extracellular medium (11). Hydrolysis of EGC by DH5(pTDCC2) suggested that recombinant ChiA was also released to the extracellular medium. Using -lactamase as a marker for periplasmic proteins, we found that recombinant ChiA, CT, LT-IIa, and LT-IIb in the culture supernatants of E. coli could be attributed to passive release by natural autolysis of the cells (data not shown). Results from these control experiments did not rule out the possibility that small amounts of ChiA are actively transported in E. coli. The chitinases of Serratia marcescens (28) and A. hydrophila (6) were mostly found in the extracellular medium when the genes for these proteins were cloned in E. coli. The mechanism by which these chitinases were transported to the medium has not been elucidated. An intriguing possibility is that E. coli transports the chitinases across the outer membrane by an intrinsic secretory system. Genes homologous to those encoding the main terminal branch of the GSP have been identified in E. coli (14). Complementation experiments using the pul system of K. oxytoca showed that at least two of the E. coli genes, gspO and gspG, were functional (14). Although it has not been demonstrated that the gsp secretory system of E. coli is expressed under most laboratory conditions, it is an interesting proposition that gsp genes may be involved in transport of the chitinases of S. marcescens and A. hydrophila across the E. coli outer membrane. Low-level transport of ChiA by the gsp system could be responsible for the extracellular chitinase activity of E. coli harboring chiA genes. Expression of chiA in E. coli gsp mutants will be required to test this rather speculative hypothesis.

The predicted molecular mass of the recombinant ChiA was in good agreement with the size of the largest immunoreactive protein detected in the anti-ChiA immunoblots of 569B. Time course experiments demonstrated that while a 90-kDa immunoreactive protein was present in the culture supernatants from late-log-phase cultures of 569B, little if any 90-kDa polypeptide was evident in culture supernatants from stationary-phase cultures. In those culture supernatants, the predominant immunoreactive molecule was a 65-kDa polypeptide. Results from preliminary experiments using EGC zymograms indicated that the 65-kDa polypeptide was enzymatically active (data not shown). While it is possible that the 65-kDa polypeptide was simply a degradative product of the 90-kDa polypeptide, it is equally possible that ChiA is purposely processed to the smaller polypeptide by factors that are expressed only when the cells enter stationary phase.

Amino acid sequence analysis showed that ChiA has homology to chitinases produced by the soil bacteria Bacillus licheniformis (GenBank entry U71214), Janthinobacterium lividum (16), and Alteromonas sp. (21), as well as the marine species V. harveyi (GenBank entry U81496), Vibrio furnissii (32), Aeromonas sp. strain 10S-24 (52), and Aeromonas caviae (54) (Fig. 8). Significant homology to ChiA was confined to a 38-amino-acid domain. The degree of sequence similarity observed among the chitinases suggests an importance for the domain in enzymatic activity. Structure-function studies of the chitinase of Aeromonas sp. strain 10S-24 (52) and of related chitinases indicated that this domain may have chitin-binding activity. Although conserved in structure, the location of the putative chitin-binding domain is divergent. In ChiA of V. cholerae and the chitinases of V. furnissii, Aeromonas sp. strain 10S-24, and J. lividum, the domain is located in the amino-terminal third of the proteins. In the chitinases of V. harveyi, A. caviae, and B. licheniformis, the domain is located in the carboxyl-terminal third of the proteins. Mutational analysis will be needed to confirm the role of the domain in enzymatic activity of ChiA.

View larger version (25K): [in this window] [in a new window]   FIG. 8.   Homology of ChiA to chitinases of other bacterial species. Amino acids are provided in single-letter code. The position of the first amino acid in the sequence of the respective protein is denoted in brackets. Amino acids of ChiA which are conserved in the other proteins are indicated (white letters on a black background).

Degradation of chitin by free-living V. cholerae is thought to begin with binding of the bacterium to the homopolymer by a chitin-binding surface receptor. In vivo experiments demonstrated that the chitin-binding receptor also has affinity for ligands on the surfaces of rabbit epithelial cells and chicken erythrocytes (49). Binding to the cell surface was inhibited by GlcNAc, the monomeric unit of chitin. ,1-4-linked glycosidic bonds occur frequently in glycosylated molecules found on the surfaces of many epithelial cells. It is tempting to speculate that V. cholerae may bind to intestinal cells by interaction between the chitin-binding receptor and an unidentified glycosylated cell surface molecule. Furthermore, it is conceivable that the ,1-4-linked glycosidic bonds in these cell surface molecules are cleaved by chitinases produced by V. cholerae. It will be interesting to determine if ChiA has hydrolytic activity for glycosylated molecules on the intestinal epithelial cell surface and whether mutations in chiA reduce the virulence of the pathogen.

Translocation of proteins across membranes is a fundamental property of prokaryotic and eukaryotic cells. To better understand the process of protein translocation, it will be necessary to elucidate the mechanisms by which the secretory systems recognize and transport extracellular proteins. Investigations into extracellular secretion of ChiA and other proteins by V. cholerae will facilitate experiments to discover the cognate transport signals and how those signals interact with components of the secretory machinery. Current experiments are focused on the use of genetics to delimit the regions of CT and ChiA that are required for extracellular transport.