Cholera vibrios produce a single polymeric protein that (i) causes hemagglutination; (ii) appears to participate in their attachment to gut epithelium; (iii) may mediate their detachment from gut epithelium; and (iv) is a protease that hydrolyzes fibronectin and mucin, cleaves lactoferrin, and nicks the A subunit of the choleragen-related heat-labile enterotoxin of Escherchia coli.In 1947, F. M. Burnet reported the discovery of a mucinase, elaborated by Vibrio cholerae, which participated in the desquamation of epithelium from pieces of guinea pig intestine in vitro (1, 2). Even though Burnet meticulously avoided overstating his conclusions, the observations inspired a flurry of activity among cholera researchers, which led to recommendations that cholera mucinase be included in cholera vaccines (3-6). The concept that the cholera mucinase played a significant direct role in the pathogenesis of choleraic diarrhea was, however, short-lived. Gangarosa et al. (7) showed that the integrity of the intestinaltepithelium was unaltered during the disease in human beings; Gordon (8) found that the disease was not an "exudative enteropathy"; Phillips (9) observed. that the intestinal secretions were low in protein and isotonic; and Finkelstein et al. (10) found that the enterotoxin (choleragen) which caused the diarrhea of cholera could be separated from the cholera mucinase and that the mucinase was not diarrheagenic. Our current observations, however, indicate that Burnet's mucinase may still be involved in the pathogenesis of cholera even though it is not responsible directly for the diarrhea of cholera.Recent efforts in our laboratory have been directed toward understanding the mechanism(s) by which the cholera vibrios elude the host intestinal clearance mechanisms of mucus secretion and peristalsis (11). We have isolated (12), from broth cultures of V. cholerae, a hemagglutinin (HA), which we previously called cholera lectin (13). This protein, a polymer of 32,000 Mr subunits, which appears to participate in the attachment of cholera vibrios to intestinal epithelium, has inherent protease activity (12). Specific antibody against the purified HA inhibits its protease function and also inhibits attachment of V. cholerae to intestinal epithelium (12).The present report shows that the HA/lectin/protease is produced in both a cell-associated and a soluble form in vivo and that it hydrolyzes fibronectin, mucin, -and lactoferrin-three proteins that may participate in host defense against cholera. It also nicks the A subunit of the choleragen-related heat-labile enterotoxin (LT) of Escherichia coli (14) 18)]. The infant rabbits (nine were inoculated with CA 401, three with 569B, and five with 3083) were sacrificed approximately 18 hr later and intestinal fluids ranging between 6 and 17.5 ml and containing 2.0-42 x 107 live cholera vibrios per ml, were harvested and lightly centrifuged (200 x g for 5 min) to sediment the nonbacterial solids. The supernates were subjected to higher speed (12,000 x g for 20 min) cent...

The structural gene, hap, for the secreted hemagglutinin/protease (HA/protease), a putative virulence factor of Vibrio cholerae, has recently been cloned and sequenced (C. C. Hase and R. A. Finkelstein, J. Bacteriol. 173:3311-3317, 1991). The availability of the null mutant, HAP-1, and HAP-1 complemented with pCH2 (which expresses HA/protease), enabled an examination of the role of HA/protease in the virulence of V. cholerae in an animal model. However, the mutants exhibited reversible colonial variation similar but not identical to that which was previously associated with dramatic changes in virulence of parental strain 3083. Regardless of colonial morphology, the mutants were found to be fully virulent in infant rabbits. Thus, the HA/protease is not a primary virulence factor (for infant rabbits). Observations using cultured human intestinal cells indicated, instead, that the HA/protease is responsible for detachment of the vibrios from the cultured cells by digestion of several putative receptors for V. cholerae adhesins.

No abstract

Unnicked cholera enterotoxin was isolated from culture supernatants of Vibrio cholerae 569B by either rapid processing of flask-grown cultures or by growing and processing fermentor cultures in the presence of ethylene glycol-bis(O-aminoethyl ether)-N,N,N',N'-tetra acetic acid, an inhibitor of the previously described V. cholerae hemagglutinin/protease. When unnicked cholera enterotoxin was incubated with purified hemagglutinin/protease, the unnicked A subunit was converted to a molecular weight consistent with that of the Al subunit as demonstrated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and its specific activity for Y1 adrenal cells increased.

A soluble hemagglutinin/protease produced by Vibrio cholerae, which has previously been shown to hydrolyze fibronectin and ovomucin and to cleave lactoferrin and the A subunit of the heat-labile enterotoxin of Escherichia coli, appears to be a zinc metalloendopeptidase. Both its hemagglutinative and protease functions are inhibited by chelating agents, including Zincov, a hydroxamic acid derivative specifically designed to inhibit zinc metalloproteases. Thermolysin, a known zinc-containing protease, also causes hemagglutination of responder chicken erythrocytes. This activity is inhibited by Zincov, which does not affect the hemagglutination activity of trypsin and pronase. The hemagglutinin/protease is active on furylacryloyl-Gly-Leu-NH2, a synthetic substrate for thermolysin and other similar proteases. The hemagglutination activity of V. cholerae-infected or cholera toxin-treated infant rabbit intestinal fluid is not inhibited by Zincov, which suggests that this activity is not due to the hemagglutinin/protease, as formerly proposed.

The emergence and rapid rise to dominance of Vibrio cholerae 0139 in India and Bangladesh in 1992 led to the consideration that choleraphage might serve as both a selective mechanism and a means for horizontal transmission of genetic information. A filamentous phage '493' from 0139 strain AJ27-493 has been purified and partially characterized. The phage was inactive on classical biotype V. cholerae 01 but it was active on El Tor biotype strains isolated prior to 1994 when E l Tor re-emerged in Bangladesh. More recent E l Tor isolates were all resistant to the phage. The phage was also active on 0139 strains. Unlike the filamentous -4, the receptor for 493 is not TcpA. The phage genome was a 9 3 kb closed circular single-stranded molecule containing a 0.4 kb double-stranded stem supporting a 2 kb single-stranded loop. A 283 bp fragment was cloned and used as a probe in Southern hybridization, in parallel with total phage 493 DNA. These probes hybridized both chromosomally and extrachromosomally with most 0139 strains, but not with 01 strains. Infection of hybridization-negative E l Tor or 0139 strains resulted in the presence of hybridizing loci (both plasmid and chromosomal), in the appearance of an 18 kDa protein, and in marked alterations in colonial morphology. Phage 493 is clearly distinct from other 0139 choleraphages which have been described. Phage 493 DNA hybridized with an encapsulated non-01 (031) strain (NRT36S) which was isolated before 0139 was recognized. NRT36S also produces a phage which can infect El Tor strains with low efficiency. Further studies may reveal whether bacteriophage play a role in the emergence and the territoriality of new choleragenic vibrios.

We previously isolated from a 1994 isolate of Vibrio cholerae O139 a filamentous lysogenic bacteriophage, choleraphage 493, which inhibits pre-O139 but not post-O139 El Tor biotype V. cholerae strains in plaque assays. We investigated the role of the mannose-sensitive hemagglutinin (MSHA) type IV pilus as a receptor in phage 493 infection. Spontaneous, Tn5 insertion, andmshA deletion mutants are resistant to 493 infection. Susceptibility is restored by mshA complementation of deletion mutants. Additionally, the 493 phage titer is reduced by adsorption with MSHA-positive strains but not with a ΔmshA1 strain. Monoclonal antibody against MSHA inhibits plaque formation. We conclude that MSHA is the receptor for phage 493. The emergence and decline of O139 in India and Bangladesh are correlated with the susceptibility and resistance of El Tor strains to 493. However, mshA gene sequences of post-O139 strains are identical to those of susceptible pre-O139 isolates, indicating that phage resistance of El Tor is not due to a change in mshA. Classical biotype strains are (with rare exceptions) hemagglutinin negative and resistant to 493 in plaque assays. Nevertheless, they express the mshA pilin gene. They can be infected with 493 and produce low levels of phage DNA, like post-O139 El Tor strains. Resistance to 493 in plaque assays is thus not equivalent to resistance to infection. The ability of filamentous phages, such as 493, to transfer large amounts of DNA provides them, additionally, with the potential for quantum leaps in both identity and pathogenicity, such as the conversion of El Tor to O139.

Protease and soluble hemagglutinating activities produced by a non-O1 Vibrio cholerae strain isolated from a patient with diarrhea were compared with similar activities produced by V. cholerae 01. The soluble protease activities were indistinguishable in heat stability, immunodiffusion, inhibition by antiserum, and electrophoretic analysis. On the other hand, the soluble hemagglutinating activities of both strains were not completely identical. The hemagglutinating activity of the non-O1 V. chokrae strain was not inhibited by Zincov; it was more sensitive to inhibition by normal serum, and it had an unusual pattern of heat stability. Heating at 100°C resulted in some recovery of activity of a sample previously inactivated by heating at 600C.