Fibronectin binding by Salmonella strains: evaluation of a particle agglutination assay

Abstract

Thirty-five Salmonella strains isolated from human cases of salmonellosis were tested and compared for their fibronectin (fn) binding capacities by using two fn-particle agglutination assays (fn-PAAs) prepared by coating with human fn either (i) latex beads (Difco; 0.81-micron diameter) (L-fn-PAA) or (ii) heat-killed formalin-treated Staphylococcus aureus Cowan 1 cells (C-fn-PAA). Six S. aureus strains were also included in this study as controls. The strains were cultured on colonization factor antigen agar and blood agar and in tryptic soy broth and brain heart infusion broth. The Salmonella and S. aureus strains were cultured at 33 and 37 degrees C, respectively, for optimal expression of fn-binding proteins. Bacterial cells (approximately 10(10) cells per ml) harvested from growth in various culture media and suspended in 0.02 M potassium phosphate buffer (pH 6.8) agglutinated the fn-PAA reagents. These reactions were scored semiquantitatively from + to + depending on the speed or intensity of the reactions within 2 min. Maximum agglutination in fn-PAA systems was observed when the cells were grown in brain heart infusion broth, while tryptic soy broth proved to be least suitable media for culturing cells for fn-PAAS. Although a statistically highly significant correlation was obtained between results of assays of radiolabeled fn and 29-kDa fragment binding, no significant correlation was observed (i) between the results of strains cultured in different media or (ii) when semiquantitative score results of the two fn-PAA systems were compared with those of the conventional radiolabeled fn assay. To enhance the efficiency of the test system, the C-fn-PAA reagent was stained with methylene blue (2% in 0.17 M glycine-NaOH buffer [pH 6.8]). This facilitated easy interpretation of results, which could be performed on hydrophobic paper instead of glass slides. The results obtained with both unstained C-fn-PAA and stained C-fn-PAA were comparable to each other and reproducible. Although the fn-PAAs are simple and easy to perform, the results did not differentiate between negative, low, moderate, and high binding abilities when Salmonella strains were evaluated for fn binding, and the results were not comparable to those obtained by the conventional radiolabeling method.