Activation of adenylate cyclase by heat-labile Escherichia coli enterotoxin. Evidence for ADP-ribosyltransferase activity similar to that of choleragen.

Highly purified, polymyxin-released, low molecular weight Escherichia coli heat-labile enterotoxin (LT) catalyzed the hydrolysis of NAD to ADP-ribose and nicotinamide. This NAD glycohydrolase activity was stimulated by dithiothreitol and was independent of cellular components. Nicotinamide formation was enhanced by arginine methyl ester > d-arginine congruent with l-arginine congruent with guanidine. A 20-fold increase in activity was noted with arginine methyl ester, and maximal activity again required dithiothreitol. When the reaction was initiated with toxin, a delay was observed before a constant rate was established. The reaction products found after incubation of [adenine-U-(14)C]NAD and l-[(3)H]arginine or unlabeled arginine methyl ester with the enterotoxin had mobilities on thin-layer chromatograms similar to the reaction products obtained after incubation of choleragen with these substrates and are consistent with the formation of ADP-ribose-l-arginine and ADP-ribose-l-arginine methyl ester, respectively. Both toxins, which catalyze the NAD-dependent activation of adenylate cyclase, thus appear to possess NAD glycohydrolase and ADP-ribosyltransferase activities. Although the activities of both toxins are dependent on dithiothreitol, Escherichia coli enterotoxin exhibited optimal activity in Tris (Cl(-)) (pH 7.5) and was inhibited by high concentrations of potassium phosphate (pH 7.0) or low pH (sodium acetate, pH 6.2). It appears that the optimal assay conditions as well as the kinetic constants for the reactants differ from those previously noted with choleragen. It is probable therefore that although the two toxins catalyze similar reactions, they differ in primary structure. The presence of transferase and glycohydrolase activities in structurally distinct toxins that activate adenylate cyclase strengthens our hypothesis that the ADP-ribosylation of arginine is a model for the NAD-dependent activation of adenylate cyclase; activation may result from ADP-ribosylation of the cyclase itself or of a protein that regulates its activity.

Gangliosides sensitize unresponsive fibroblasts to Escherichia coli heat-labile enterotoxin.

Chemically transformed mouse fibroblasts did not raise their cyclic AMP level in response to Escherichia coli heat-labile enterotoxin. These fibroblasts did, however, incorporate exogenous mono-, di-, and trisialogangliosides. After the uptake of monosialoganglioside galactosyl-N-acetylgalactosaminyl-[N-acetylneuraminyl]-galactosylglucosylceramide (GM1), the cells responded to E. coli heat-labile enterotoxin. The di- and trisialogangliosides were considerably less effective. GM1, the putative cholera toxin (choleragen) receptor, has been implicated previously as the receptor for E. coli heat-labile enterotoxin based on the ability of the free ganglioside to inhibit the effects of toxin. This investigation establishes that the ganglioside, when incorporated into fibroblasts, serves a functional role in mediating the responsiveness to the toxin.

Interaction of polymorphonuclear neutrophils with Escherichia coli. Effect of enterotoxin on phagocytosis, killing, chemotaxis, and cyclic AMP.

Enterotoxigenic Escherichia coli are associated with noninflammatory diarrhea and stimulate adenylate cyclase activity of mammalian cells, thereby increasing intracellular cyclic adenosine 3',5'-monophosphate (cyclic AMP). Increased concentrations of cyclic AMP in polymorphonuclear neutrophils (PMN) inhibit phagocytosis, candidacidal activity, granule discharge, and chemotactic responsiveness. We examined the effect of enterotoxin on the interaction of human PMN with E. coli. Enterotoxigenic and nonenterotoxigenic strains, including serotypes of E. coli identical except for the presence or absence of the plasmid coding for enterotoxin production, were utilized. Enterotoxigenic and nonenterotoxigenic E. coli, tumbled with PMN, were phagocytized and killed (>97%) equally well, and these strains stimulated PMN hexose monophosphate shunt activity equivalently.However, a chemotaxis assay under agarose demonstrated that filtrates of 10 enterotoxigenic strains were less chemotactic for PMN by 15+/-2% total migration or 46+/-1% directed migration, when compared with 6 non-enterotoxigenic strains (P < 0.001). Inactivation of the enterotoxin by heat (65 degrees C for 30 min) or antibodies formed to E. coli enterotoxin eliminated the inhibitory effect of the enterotoxic filtrates for PMN chemotaxis. Addition of purified E. coli enterotoxin directly to the PMN decreased chemotaxis to E. coli filtrates by 32+/-2% (P < 0.001). These data suggest that the effect was due to the heat-labile enterotoxin. The phosphodiesterase inhibitor, 1-methyl-3-isobutylxanthine (0.1 mM), which potentiates effects due to an increase in intracellular cyclic AMP, further decreased total PMN migration (random plus directed) toward enterotoxic filtrates to 46% of that to nonenterotoxic filtrates (P < 0.001). Addition of cholera toxin (1 mug/ml), which is similar to E. coli enterotoxin, to the PMN inhibited total migration toward nonenterotoxic filtrates by 16+/-2% (P < 0.001). Exogenous dibutyryl cyclic AMP (2 mM) inhibited total PMN migration toward E. coli filtrates by 32% (P < 0.001). PMN intracellular cyclic AMP levels increased by 220% after 2 h of incubation with purified E. coli enterotoxin. The decreased chemotactic attractiveness of enterotoxic E. coli filtrates appears to be related to the ability of enterotoxin to increase cyclic AMP in PMN. Enterotoxin production by E. coli may be advantageous to the microbe by decreasing its chemotactic appeal for PMN.

Amplified Assay for Specific Dual-Labeled DNA Using the Coagulation Cascade (EDNA-ELCA).

Detection of the products of the PCR reaction using nonisotopically labeled DNA molecules containing biotin, fluorescein, or digoxigenin has become a popular method for identification of specific products of polymerase chain reaction (PCR) (1,3). These labeled molecules are prepared either by PCR synthesis in the presence of labeled deoxyuridine triphosphate (1,3) or by hybridization of labeled probes to the unlabeled PCR product (1,2). Detection is then in a format very similar to enzyme-linked immunosorbent assays (ELISA) using similarly labeled antigens and antibodies, i.e., by capture on the receptor for one ligand (streptavidin or antibody) and using the other ligand for detection.

Factors influencing in vitro skin permeability factor production by Vibrio cholerae.

The development of a new semisynthetic medium which stimulates in vitro production of the skin permeability factor (PF) by Vibrio cholerae is described. The effects of pH, aeration, temperature, and length of incubation on PF formation or release in strain 569B and several other strains, or both, are compared. Data are presented which show that maximal PF accumulation occurs during a transitional period of growth joining the exponential and stationary phases of the growth cycle. PF elaboration is completed well ahead of any visible signs of lysis in the culture and the PF activity appears to be proportional to the length of the linear growth phase. Possible mechanisms of toxigenicity and the nature of PF are discussed.

Adenosine diphosphate-ribosylation of adenylate cyclase catalyzed by heat-labile enterotoxin of Escherichia coli: comparison with cholera toxin.

The heat-labile enterotoxin of Escherichia coli, like cholera toxin, activates adenylate cyclase by catalyzing the transfer of adenosine diphosphate-ribose from HAD+ (oxidized nicotinamide adenine dinucleotide) to the guanyl nucleotide-dependent regulatory component of the cyclase. A preparation of enterotoxin that had been released from E. coli following exposure to polymyxin B and then partially purified was found to contain two enzymatically active peptides, one of about 29,000 and the other of about 24,000 daltons, which correspond in molecular size to the enzymatically active subunit A and fragment A1 of cholera toxin, respectively. As with cholera toxin, the enzymatic activity of E. coli enterotoxin was elevated by incubation with sodium dodecyl sulfate to release active peptides. Treatment with dithiothreitol, however, had no effect. Dithiothreitol activates subunit A of cholera toxin by reducing an internal disulfide bond, but no corresponding bond appears to be present in the partially purified E. coli enterotoxin.

Studies on the genetic and cellular control of sensitivity to enterotoxins in the sealed adult mouse model.

A sealed adult mouse (SAM) model was developed for studies on the effects of cholera enterotoxin (CT). With this system, 38 strains of outbred, inbred, congenic, recombinant, and mutant mice were starved for 24 h, anorectally occluded with cyanoacrylamide ester glue, given CT per os, and sacrificed at 6 h. Fluid accumulation (FA) values were calculated as gut weight to body weight ratios. At a saturating dose of CT (24 micrograms per mouse), FA responses were found to be independent of body weight and gut length. It was found, using recombinant and congenic mice, that mice which possess the H-2k haplotype (homozygous or heterozygous) are 2.5 to 3 times less responsive to CT than animals with the H-2b haplotype. The allele(s) responsible for this affect is located near the K end of the H-2 complex. Inbred and congenic mice given CT intravenously exhibited the same (b = responder, k = nonresponder) pattern in terms of weight loss and death, thus indicating that the H-2 effect is not limited just to the small intestinal epithelium. Mice given sublethal doses of CT intravenously and challenged after conversion to SAM 14 days later showed an immune response inversely related to weight loss (i.e., b haplotypes lost 10 to 15% body weight, recovered, but were not protected against challenge; k haplotypes lost little or no weight but were protected). To examine the possibility of a cellular basis for control of innate responses to CT, responder C57BL/10 (B10) mice were irradiated with 950 rads and immediately reconstituted with bone marrow from (B10 X B10.BR)F1 (nonresponder) mice. The chimeras became nonresponsive to CT when challenged 5 weeks after reconstitution. Reconstituted B10 controls responded normally. Outbred and inbred nude athymic mice also were nonresponsive when compared with normal responder controls. These data demonstrate a genetic basis for resistance to CT and that response and nonresponse is mediated, at least in part, by cells derived from bone marrow.

In vitro production of choleragen and vascular permeability factor by Vibrio cholerae.

The in vitro production of significant amounts of extracellular choleragen and vascular permeability factor (PF) by Vibrio cholerae strain VC-12 (Ogawa) in a basal peptone medium required forced aeration, low incubation temperature, and a low initial pH. Filtrates of alkaline peptone cultures of VC-12 grown at 37 C contained an ion translocase inhibitory activity but neither choleragen nor PF activity, Sterile filtrates of pH 6.5 peptone cultures of VC-12 grown at 29 C contained no ion translocase inhibitory activity but possessed choleragen activity (lethality for the suckling rabbit) and PF activity to the extent that the intradermal inoculation of 0.1 ml of a 1:12,288 dilution of such a filtrate gave rise to a vascular permeability reaction (8 by 8 mm in diameter) in the guinea pig. PF excretion occurred during the late logarithmic phase of growth but did not appear to be the consequence of cell lysis. The PF activities of strains VC-12 and 569B (Inaba) were neutralized to the same extent by anticholeragen antiserum prepared against crude 569B choleragen.

Fine structure of Vibrio cholerae during toxin production.

The fine structural changes associated with cell growth and toxin production have been examined in Vibrio cholerae strain 569B. No morphological alterations in the cell envelope are apparent during logarithmic growth with thin-section techniques. However, internal swelling, suggesting alteration of cell envelope permeability, is evident particularly during the late logarithmic and early stationary phases of growth. Certain extracellular material demonstrable with negative-stain techniques does appear during the period of toxin production. The possible origin of this material is discussed. The effects of high temperature (37 C) and aeration on cell structure are also examined.

Biochemistry of Vibrio cholerae virulence. 3. Nutritional requirements for toxin production and the effects of pH on toxin elaboration in chemically defined media.

The investigations described in this report concern the metabolic and physiologic parameters governing cholera enterotoxin production in chemically defined growth media. The results indicate that the minimal nutritional requirements for growth of pathogenic Vibrio cholerae are the same as those for toxin production, and that toxin production parallels growth of the organisms. Studies of the relationship between toxin accumulation and pH reveal that toxin biosynthesis can be separated from toxin release. Toxin is synthesized below pH 7.0, but release and accumulation of extracellular toxin occur only at neutral or alkaline pH values.

Enterotoxigenicity of clinical isolates of non-O1 Vibrio cholerae.

Whole cultures, but not culture supernatant fluids, of 21 isolates of non-O1 V. cholerae from patients with diarrhea were shown to induce positive fluid accumulation in infant mice. CHO cell assays demonstrated the elaboration of heat-labile cytotonic, cytotoxic or both factors from most isolates when grown under optimal culture conditions. These factors were not neutralized by anti-cholera toxin serum. Also genetic studies performed on 9 vibrio isolates using a DNA hybridization probe failed to detect gene sequences homologous with cholera toxin. ELISA assays recognized six isolates which produced a cell-associated substance which immunologically cross-reacted with cholera toxin. Enzymatic profiles of the vibrio isolates did not correlate with the production of any toxic factor. The results indicate that mainly heat-labile and cell-associated cytotonic and cytotoxic factors appear to influence the enterotoxigenic potential of this heterogenous group of non-O1 vibrios.

Fluid accumulation in infant mice caused by Vibrio hollisae and its extracellular enterotoxin.

Vibrio hollisae, a halophilic bacterium isolated from patients with diarrhea, was examined for virulence factor production. Intragastric administration of 2 X 10(7) CFU per mouse elicited fluid accumulation which peaked at ca. 6 h postchallenge in infant mice. An enterotoxin which elongated Chinese hamster ovary (CHO) cells was detected in extracts of infected-mouse intestines and in culture fluids from various growth media. The yield of the enterotoxin was maximal beginning at the onset of the stationary phase of growth in heart infusion broth supplemented with 0.5% NaCl. A concentrated preparation obtained by ammonium sulfate precipitation of culture supernatant fluids induced intestinal fluid accumulation which peaked at 2 h postchallenge in infant mice. The abilities of the enterotoxin preparation to elongate CHO cells and to elicit fluid accumulation in infant mice were inseparable by gel filtration, isoelectric focusing, and hydrophobic interaction chromatography. The enterotoxin has a molecular weight of ca. 33,000 by gel filtration and an isoelectric point of ca. 4 and is sensitive to heat.

Biochemical and genetic basis for the ETS (enterotoxin sensitivity) phenotype in mice.

Evidence has been accumulating that humans are genetically predisposed to cholera gravis. Using the sealed adult mouse model, we found that certain inbred mice were also hypersensitive to cholera toxin (CT). Such mice were designated ETS+ (enterotoxin sensitive), and the trait was linked to the K end of the mouse H-2 histocompatibility complex. Cells isolated from ETS+ mice bound more CT and accumulated more cyclic adenosine monophosphate (cAMP) after intoxication. Analysis of ETS+ cells showed that they express lesser amounts of the non-GM1 gangliosides that block or compete for relevant CT binding sites in ETS- cells. Conversion of ETS- non-GM1 gangliosides to GM1 with neuraminidase increased CT binding and cAMP responses. Reconstitution of nonreactive ganglioside-deficient cells with ETS+ or ETS- gangliosides caused them to bind CT like the original ETS+ or ETS- cells. Ganglioside expression genes known to map to the same H-2-linked region as the ETS phenotype seem to be involved in controlling murine susceptibility to CT.

Virulence mechanisms associated with clinical isolates of non-O1 Vibrio cholerae.

Twenty one isolates of non-O1 V. cholerae from patients with diarrheal illness were examined for the presence of potential virulence mechanisms. The motile strains (90%) produced cell-associated mannose-sensitive hemagglutinins which reacted with human group O, chicken, sheep and rabbit erythrocytes. Motile isolates also attached to embryonic intestinal epithelial cells (ATCC 407), and the adherence was not inhibited by the presence of 1% D-mannose. All vibrio isolates hemolyzed sheep erythrocytes. Three vibrio isolates (14%) harbored two or three plasmids which ranged in size between 1.7 and 5.2 megadaltons. The presence of the plasmid did not correlate with the presence of hemolysin, hemagglutinins, adhesions or antibiotic resistance in any of the isolates. Thus, it appears that multiple factors associated with bacterial cell surfaces influence adhesin and apparently pathogenic potential of the non-O1 vibrio isolates in the host intestine.

Biochemistry of vibrio cholerae virulence. II. Skin permeability factor-cholera enterotoxin production in a chemically defined medium.

A chemically defined medium capable of eliciting high titers of skin permeability factor (PF)/enterotoxin from three different strains of Vibrio cholerae has been developed. Toxin/antigen elaboration in synthetic and in complex media was monitored by a specific passive hemagglutination-inhibition test. A distinct temporal difference in the pattern of toxin/antigen elaboration was noted when the two types of media were compared. In complex media, PF activity and corresponding antigen release were coincident, whereas in the defined medium a biphasic pattern resulting in elaboration of nontoxic antigen during the second phase was seen. Possible reasons for the latter observation are discussed, and several experiments illustrating the unique utility of the defined medium are presented.

Production of Bacteriophage-Associated Materials by Vibrio cholerae: Possible Correlation with Pathogenicity.

Classical and El Tor strains of Vibrio cholerae were examined for production of bacteriophage or bacteriophage-related material by electron microscopy of mitomycin C-induced cultures. The strains were also tested for pathogenicity by using the ligated ileal loops of adult rabbits. Of the 27 strains tested, 25 showed a correlation between the production of bacteriophage-related material and pathogenicity. Both P(+) and P(-) strains produced bacteriophage tails. The bacteriophage tails of strains 569B and ATCC 9168 and other strains were shown to occur in an extended form with tail appendages.

Biochemistry of Vibrio cholerae Virulence I. Purification and Biochemical Properties of PF/Cholera Enterotoxin.

A simple method for concentrating, isolating, and purifying PF/cholera enterotoxin from culture supernatant fluids is presented. Precipitation with dextran sulfate and ammonium sulfate, followed by gel filtration and ion-exchange chromatography, yielded PF/enterotoxin of high biological potency. Although not completely free from traces of vibriocidal antibody-stimulating activity, the resultant purified toxic immunogen appeared homogeneous as judged by immunoelectrophoresis, immunodiffusion, and disc electrophoresis. Advantages and disadvantages of this purification method are discussed in comparison with those of previously published techniques.

Adaptation of the staphylococcal coagglutination technique for detection of heat-labile enterotoxin of Escherichia coli.

Protein A-containing staphylococci coated with specific antiserum were tested for heat-labile enterotoxin of Escherichia coli. The immunological cross-reactivity of E. coli heat-labile enterotoxin with Vibrio cholerae toxin (choleragen) was the basis for sensitizing stabilized suspensions of the Cowan I strain of Staphylococcus aureus with anticholeragen. Unconcentrated culture supernatant fluid containing E. coli heat-labile enterotoxin produced macroscopic agglutination when mixed with sensitized staphylococci in capillary tubes. A total of 15 toxigenic and 61 nontoxigenic isolates were tested by the staphylococcal coagglutination technique in a coded fashion and found to be in agreement with previous results of the Chinese hamster ovary cell assay and the passive immune hemolysis test. The staphylococcal coagglutination technique is simple, relatively inexpensive to perform, and requires the immunoglobulin fraction of anticholeragen as the only specific reagent. The staphylococcal coagglutination technique appears to have potential for routine use in diagnostic microbiology laboratories.