Measurement of virulence of aeromonads using a suckling mouse model of infection.

Virulent and avirulent strains of Aeromonas spp. were identified and virulence quantified using an animal model. Virulence was measured by determining a 50% lethal dose (LD50) 43 h after oral administration of live bacteria. The LD50 of virulent Aeromonas isolates ranged from log10 7.53 (mean) organisms to log10 8.88 (mean). Some isolates were avirulent in this model. Detection of cytotoxic activity in culture supernatants correlated with virulence (Fisher exact test, P = 0.0029). There was no correlation between LD50 and the source of the isolate, beta-haemolysis or lipopolysaccharide (LPS) banding profile on SDS-PAGE. In this animal model, virulence was multifactorial in that: (i) bacterial multiplication in the gut was associated with fatal infection; (ii) the increase in bacterial numbers in the gut of mice administered a lethal dose of bacteria was accompanied by accumulation of fluid; and (iii) there was evidence of extraintestinal spread of infection. Protection of suckling mice by rabbit antiserum to Aeromonas cell envelopes was observed.

Hemagglutination properties and adherence ability of Aeromonas hydrophila.

Human diarrheal isolates and enterotoxigenic strains of Aeromonas hydrophila are strong hemagglutinators with human blood cells. Sugar inhibition studies and yeast coagglutination tests with 11 selected strains revealed six different hemagglutination mechanisms for this species. These were characterized by inhibition by L-fucose, inhibition by D-galactose, inhibition by D-mannose, and two distinguishable mechanisms which were inhibited by either L-fucose or D-mannose, one being pilus mediated. Inhibition of hemagglutination by another strain required a combination of D-galactose and D-mannose. The hemagglutinating strains also attached well to human blood cells and buccal epithelial cells, with as many as 55% of the cells of a culture attaching successfully. In some cases the attachment to buccal epithelial cells appeared to involve mechanisms different from those used for hemagglutination.

Aeromonas hydrophila typing scheme based on patterns of agglutination with erythrocytes and yeast cells.

An agglutination typing scheme has been developed for strains of Aeromonas hydrophila. Primary agglutination typing is based on testing agar-grown A. hydrophila cells with human, horse, rat, and guinea pig erythrocytes and Saccharomyces cerevisiae cells. Further subdivision of primary groups is based firstly on whether yeast cell agglutination is inhibited by a D-mannose polymer, yeast mannan, and secondly on patterns of inhibition of hemagglutination by yeast mannan and the monomeric sugars L-fucose, D-galactose, and D-mannose. A total of 320 isolates were tested, and these were divisible into 39 distinct types on the basis of this scheme. Application of this typing scheme in the future to isolates of A. hydrophila known to be associated with human infection may enable correlations to be made between particular agglutination types and human pathogenicity.

Biochemical characteristics of enterotoxigenic Aeromonas spp.

Biotypes of Aeromonas spp. correlated well with enterotoxin production in a study of 174 strains. Using biochemical characteristics determined by conventional methods and multitest systems, we correctly classified 93% of the strains with regard to enterotoxin production. Most of the enterotoxigenic strains were Voges-Proskauer (VP) positive and did not hydrolyze arabinose, but VP-positive strains which hydrolyzed arabinose were mainly non-enterotoxigenic. Aeromonas punctata subsp. caviae, which is VP negative and does not oxidize gluconate or produce gas from glucose, was non-enterotoxigenic. Although the number of other VP-negative strains was small, most were enterotoxigenic. Discrimination was improved so that 97% of the strains were correctly classified if the hemolysin assay was used either for all strains or for only the VP-positive, arabinose-positive and VP-negative, non-A. punctata subsp. caviae strains. Because the proposed classification system does not require facilities for carrying out in vivo assays such as suckling mouse or ileal loop methods, the identification of enterotoxigenic Aeromonas strains should be possible in diagnostic laboratories and will facilitate epidemiological studies of the role of these organisms in acute diarrhea.

Exotoxins of Aeromonas hydrophila.

Eighty of 103 strains of Aeromonas hydrophila cultured at 100 rev./min produced heat-labile enterotoxins detected using the suckling mouse assay. Results in intestinal perfusion agreed with the suckling mouse test in all strains tested by both methods. Enterotoxic activity correlated with haemolysin and cytotoxin production, but 4% of strains would have been wrongly classified using haemolysin assay in place of the suckling mouse test and 11% misclassified on the basis of cytotoxin assay. There was a significant association between haemolytic and cytotoxic activity, but 15% of strains produced only one of these toxins. Haemolysin, cytotoxin and enterotoxin were not always associated in a given isolate. The time of appearance of exotoxins during bacterial growth and the effects of dialysis, heating and proteolytic enzymes also suggest that haemolysins, cytotoxins and enterotoxins of Aeromonas hydrophila are separate toxins and not different manifestations of the same toxin.

Isolation of carbohydrate-reactive outer membrane proteins of Aeromonas hydrophila.

Outer membrane proteins of Aeromonas hydrophila A6 were isolated by affinity chromatography on the basis of their reactivity with trisaccharide structures analogous to the terminal trisaccharide of the H antigen of the human ABO(H) blood group system and were characterized by using antisera raised against the isolate. The outer membrane extract for affinity chromatography was prepared from pressure-disrupted outer membranes by differential centrifugation, followed by solubilization of outer membrane components in a nondenaturing, nonionic detergent. Carbohydrate-reactive outer membrane proteins (CROMPs) were then purified by affinity chromatography on two different affinity matrices composed of trisaccharides resembling the terminal trisaccharide of the H antigen, attached to inert silica beads. The relative efficiencies of H type 1 and 2 terminal trisaccharides as affinity adsorbents were established. Reactive proteins were eluted under alkaline conditions (pH 11.0) and in the presence of soluble H substance prepared from group O secretor saliva, but not by 60 mM alpha-L-fucose or under acid conditions (pH 3.0). The eluate contained at least three components (M(r)s, 43,000, 40,000, and < 14,000), as detected by immunoblot analysis with a polyvalent, polyspecific rabbit antiserum to A. hydrophila A6 (serum 3/83). A specific antiserum (serum 3/91) prepared in a rabbit by repeated immunizations with nitrocellulose containing the 43,000-Da band reacted with three bands (M(r)s, 43,000, 40,000, and < 14,000) in immunoblot analysis of solubilized outer membranes of A. hydrophila A6, suggesting that the 40,000- and < 14,000-Da elements are immunologically related to components of the 43,000-Da protein. Furthermore, pretreatment of A. hydrophila A6 with serum 3/91 reduced the strength of bacterial hemagglutination. The purified CROMPs did not agglutinate human group O erythrocytes. The reactivity of isolated CROMPs with a second CROMP-specific antibody (lipopolysaccharide-absorbed serum 3/83) was investigated. CROMPs, proteinase K-treated CROMPs, and bovine serum albumin were bound to latex beads and reacted with lipopolysaccharide-absorbed serum 3/83. Antibodies eluted from CROMP-latex inhibited hemagglutination of human erythrocytes by A. hydrophila A6 to a titer of 4. Antibody eluted from proteinase K-treated CROMP-latex beads showed hemagglutination inhibition activity only when undiluted. There was no hemagglutination inhibition antibody activity detectable in the eluate from bovine serum albumin-latex beads. These results show that antibodies which react with the isolated CROMPs also react with an H-antigen-reactive hemagglutinin of A. hydrophila A6. The possibility that CROMPs act as an adhesin, or adhesins, and contribute to the virulence of this organism is discussed.

Localization of a 23,000 MW antigen of Cryptosporidium by immunoelectron microscopy.

Rabbit antiserum was raised against a 23,000 molecular weight (MW) antigen prepared from Cryptosporidium oocysts by electro-elution from polyacrylamide gels. The antiserum was tested for specificity by immunoblotting against solubilized oocyst preparations. Several antigens including the 23,000 MW antigen were recognized suggesting that it shared common epitopes with higher MW proteins. The antiserum was then used in conjunction with a protein A-colloidal gold conjugate to locate antigenic sites within exogenous and endogenous developmental stages of Cryptosporidium. The pellicles of both sporozoites and merozoites exhibited specific labelling, particularly around their anterior ends. No specific labelling was observed for any other membrane determinants or organelles in these or other life cycle stages.

Carbohydrate-reactive, pore-forming outer membrane proteins of Aeromonas hydrophila.

Two outer membrane proteins of Aeromonas hydrophila A6, isolated in a one-step affinity chromatography process based on carbohydrate reactivity, were found to be pore-forming molecules in artificial planar bilayer membranes. These carbohydrate-reactive outer membrane proteins (CROMPs; M(r)s, 40,000 and 43,000) were subjected to amino acid analysis. The amino acid profiles for these two outer membrane proteins were almost identical. A partial protein sequence of a 14-amino-acid fragment of the 40,000-Da protein revealed homology with outer membrane porins of Escherichia coli and A. hydrophila. CROMPs were compared with carbohydrate-reactive porins also extracted from outer membranes of A. hydrophila A6. These porins were isolated by using standard porin purification techniques (insolubility in 2% sodium dodecyl sulfate, solubility in 0.4 M NaCl, and Sephacryl S-200 gel filtration), and then Synsorb H type 2 affinity chromatography was done. The physical and functional properties of the carbohydrate-reactive porins and CROMPs were found to be identical. On the basis of pore-forming properties in planar lipid bilayers and channel inhibition with maltotriose solutions, a nonspecific, general diffusion porin and a LamB-like maltoporin were identified in both CROMP and carbohydrate-reactive porin preparations. To our knowledge, the use of carbohydrate reactivity to isolate channel-forming proteins from bacterial outer membranes has not been reported previously.