Lectin-binding properties of Aeromonas caviae strains.

The cell surface carbohydrates of four strains of Aeromonas caviae were analyzed by agglutination and lectin-binding assays employing twenty highly purified lectins encompassing all sugar specificities. With the exception of L-fucose and sialic acid, the sugar residues were detected in A. caviae strains. A marked difference, however, in the pattern of cell surface carbohydrates in different A. caviae isolates was observed. Specific receptors for Tritricum vulgaris (WGA), Lycopersicon esculentum (LEL) and Solanum tuberosum (STA) (D-GlcNAc-binding lectins) were found only in ATCC 15468 strain, whereas Euonymus europaeus (EEL, D-Gal-binding lectin) sites were present exclusively in AeQ32 strain, those for Helix pomatia (HPA, D-GalNAc-binding lectin) in AeC398 and AeV11 strains, and for Canavalia ensiformes (Con A, D-Man-binding lectin) in ATCC 15468, AeC398, AeQ32 and AeV11 strains, after bacterial growing at 37°C. On the other hand, specific receptors for WGA and EEL were completely abrogated growing the bacteria at 22°C. Binding studies with (125)I- labeled lectins from WGA, EEL and Con A were performed. These assays essentially confirmed the selectivity, demonstrated in the agglutination assays of these lectins for the A. caviae strains.

INTRODUCTION

Aeromonas species are important opportunistic pathogens, causing diarrhoea, dysentery and extra-intestinal infections (1,2,3). Recent studies revealed that some filamentous structures, outer-membrane proteins (OMPs), capsule and lipopolysaccharides (LPS) of Aeromonas spp. could be intestinal adhesins (4,5,6). Some studies demonstrated that A. caviae isolates show a high percentage of adherence to and invasion in different host cell lines (6,7,8). A. caviae isolates showed higher adherence levels when grown at 22°C than at 37°C. Furthermore, these high adherence and invasive proprieties correlated to higher filamentous structures and OMPs expression at 22°C (6). Similar results were observed with A. veronii biovar sobria strains (9). In addition, the high percentage of A. caviae isolation from children with gastroenteritis associated with the adhesive and invasive characteristics of the isolates may partially reflect the enteropathogenic potential of the A. Caviae (1).

However, the problems with speciation of the Aeromonas spp. may account for much of the confusion surrounding the role of these microorganisms in disease (10). On the basis of the colonization capacity and the problems with speculation of the Aeromonas spp., we looked for a study up on the bacterial cell surface carbohydrates in A. caviae. Carbohydrates in the forms of glycoproteins and glycolipids have been reported to play an important role in the binding of bacteria to receptors on host epithelial cells (11,12). In recent years, carbohydrate structures have been implicated in cell-to-cell or cell-to-substrate interaction, such as cell adhesion or aggregation (13-15). The specificity, affinity and concentrations of the interacting molecules will determine the degree of success of the bacterial colonization. Much of the works concerning structure and function of cell surface carbohydrates are derived from studies with lectins (16,17). The purpose of this investigation was to evaluate the influence of bacterial growth temperature on lectin binding properties of four A. caviae strains.

MATERIALS AND METHODS

Bacterial strains and growth conditions

Three nonfimbriated A. caviae strains, confirmed by transmission electron microscopy under fimbrial expression conditions (6) and the reference strain (ATCC 15468) obtained from American Type Culture Collection, were used in this study. We previously demonstrated that these three nonfimbriated strains grown at 22°C show a high percentage of adherence to and invasion in different host cell lines (6,7). The strain AeC398 was isolated from stools of children with gastroenteritis in the Laboratory of Bacteriology, Hospital Pedro Ernesto, Universidade do Estado do Rio de Janeiro (UERJ) (18). The lettuce (AeV11) and the Brazilian white cheese (AeQ32) A. caviae isolates were gently supplied by Dra M.L. Pena-Queiróz, Faculdade de Ciências Médicas (UERJ). The strains were cultivated onto blood agar and incubated at 37°C overnight. A sweep of colonies was inoculated into Tryptic Soy Broth (TSB) without glucose (Difco Lab., Detroit, MI, USA) and incubated at 37°C for 24h or at 22°C for 48h (6).

Lectins

Twenty lectins (Sigma Chemical Co., St. Louis, MO, USA) of various specificities were used for this study. All lectins were tested for the presence of glycosidase using a- and b-p-nitrophenyl glycosides of D-Gal, D-GalNAc and D-GlcNAc (19).

Agglutination assays

Agglutination assays were performed with a Terasaki microtitrator, using 25-ml loops and suspensions containing 2 × 108 bacteria ml-1. Equal volumes of the cell suspension and the lectin dilution were mixed, incubated at room temperature (24oC) for 1 h. Bacterial agglutination was always scored visually with a hand lens after gently resuspending the settled cells and by observations using a phase contrast microscope. Agglutination inhibition assays were done at room temperature with 4 agglutinating units of lectin and 0.1 M concentration of each specific carbohydrate. In certain experiments washed bacteria suspended in phosphate-buffered saline (PBS) 10 mM, pH 6.0, have been treated with 0,1 U of Clostridium perfringens sialidase for 60 min at 37°C, the cells washed four times with PBS, pH 7.2, and ressuspended at suitable concentrations for agglutination or binding assays. (20).

Lectin radioiodination and binding studies

Triticum vulgaris (WGA), Euonymus europaeus (EEL) and Canavalia ensiformes (Con A) lectins were labeled with 125I by the method of Hunter (21), in the presence of 0.1 M of the specific inhibitory sugar. Specific activities ranged from 2 to 4x104 cpm μg-1 of lectin. The lectin binding studies were performed as follows: 106 bacterial cells were incubated with increasing concentrations of iodinated lectin in 150 μl of PBS- 0.5% BSA. The amount of iodinated lectin bound to the cells was determined using a gamma counter (Beckman Instruments, Inc., Palo Alto, CA, USA). Specificity of binding was ascertained by performing parallel binding determinations in the presence of 0.1 M of each of the specific sugar inhibitor for each concentration of lectin used (20).

RESULTS AND DISCUSSION

In our effort to find potential differences in surface carbohydrates of the four A. caviae strains grown at different temperatures, we have analyzed lectin binding properties by agglutination and lectin-binding assays. The lectin agglutination reaction of A. caviae (ATCC 15468, AeC398, AeQ32 and AeV11 strains) was considered most specific with bacterial cells that were agglutinated at the lowest lectin concentration (Table 1). Visual observation of agglutination showed roughly the same titters as those determined by phase contrast microscopy. The two L-fucose-binding lectins, Lotus tetragonolobus (LTL) and Ulex europaeus I (UEL-I), and the two sialic acid-binding lectins, Aaptos papillata II (APL-II) and Limulus polyphemus (LPL), did not agglutinate the A. caviae strains. The results shown in Table 1 suggest the occurrence of qualitative and quantitative differences in bacterial surface carbohydrate composition, as detected by lectins, between A. caviae strains grown at 37°C and 22°C. Four lectins with affinity for D-GlcNAc were tested. Lycopersicon esculentum (LEL) and Solanum tuberosum (STA) lectins agglutinated strongly ATCC 15468 strain of A. caviae grown at 37°C or at 22°C but were unreactive with the three A. caviae isolates. Triticum vulgaris (WGA), however, interacted selectively and strongly with ATCC 15468 strain of A. caviae grown at 37°C. Agglutination was inhibited by 0.1 Mα-methyl-D-GlcNAc. Among the five lectins with affinity for D-GalNAc tested, only Helix pomatia (HPA) lectin showed activity. HPA recognised A. caviae AeC398 strain grown at 37°C and A. caviae AeV11 strain grown at both temperatures. Agglutination was inhibited by 0.1 M D-GalNAc. Six D-Gal-binding lectins were tested. Euonymus europaeus (EEL) selectively agglutinated A. caviae AeQ32 strain grown at 37°C. Agglutination was inhibited by 0.1 M D-Gal. Among the two lectins tested (Con A and Lens culinaris), which have binding sites for D-Man-like residues, only Con A reacted with all A. caviae strains grown at 37°C and A. caviae ATCC 15468 and AeQ32 strains grown at 22°C. Agglutination was inhibited by 0.1 M a-methyl-D-mannoside. Because of the dual specificity of the WGA binding site, with affinities for both D-GlcNAc polymers and sialic acid residues, neuraminidase treatment of A. caviae ATCC 15468 strain was performed. Such treatment did not abolish the strong agglutinating activity of WGA. To verify if the differences detected in the cell agglutination pattern with lectins (Table 1) were compatible with the expression of their respective cell surface receptors, the binding of 125I- labeled WGA, EEL and Con A to A. caviae strains was determined. All three iodine-labeled lectins discriminated strains of A. caviae grown at 37°C (Fig. 1). WGA interacted more strongly with A. caviae ATCC 15468 strain, whereas the binding of EEL was higher with A. caviae AeQ32 strain. Con A preferentially bound to two strains of A. caviae grown at 37°C. Although the AeV11 strain bound 2 times less radioactivity than AeC398 strain, the amount of [125I]-Con A bound was nevertheless specific. The binding was specific because it could be reversed with 0.1 M of the corresponding specific saccharide. The binding data in Fig. 1 demonstrate that the differential agglutination of the A. caviae strains by WGA, EEL and Con A (Table 1) correlated with the number of exposed [125I]-lectin receptor sites on them. When the binding data were plotted according to the method of Steck and Wallach 1965 (22), as shown in Fig. 2, the average number of lectin-binding sites (n) and the lectin association constant (Ko) were determined. WGA-binding data for A. caviae ATCC 15468 strain was n ~ 6.2 × 105/bacteria, Ko ~ 2.6 × 107 M-1. EEL-binding data for A. caviae AeQ32 strain was n ~ 2.0 × 105/ bacteria, Ko ~ 3.6 × 107 M-1. Con A-binding data for A. caviae AeC398 and AeV11 strains was n ~ 3.4 × 105/bacteria, Ko ~ 2.2 × 107 M-1 and n ~ 5.5 × 105/bacteria, Ko ~ 1.3 × 107 M-1 respectively. Based on these findings, a new approach for the isolation of cell surface receptor sites can be envisaged. Furthermore, the analysis of the lectin receptors should allow deeper insight in how the A. caviae strains exert their biologically relevant functions. The differential expression of carbohydrate structures on the bacterial surface could be in part associated with host specificity that, as is well known, depends on microorganism recognition based on surface saccharide components (17,23). Finally, lectin reactivity in A. caviae isolates, as opposed to other organisms, might be used as a diagnostic method that may be eventually useful in epidemiological studies (16,24). It also opens the possibility of using lectins to isolate cell surface glycoconjugates by affinity chromatography for analysis of their chemical structure and biological function (25,26). Furthermore, the ability to vary quantitative and qualitatively the surface carbohydrate composition under different environmental conditions could be advantageous to A. caviae during different stages of host cell colonization. There are strong evidences that bacterial growth temperature influence Aeromonas spp. adherence to host cells and the fimbrial structure, OMPs and smooth-LPS phenotype expression (4,6,9). A. hydrophila was found to produce a 43kDa haemagglutinin that binds to H-antigen expressed on the surface of most human erythrocytes and is inhibited by L-fucose (27). Previous studies also demonstrated that differences in A. caviae adherence levels correlated with higher expression of the 43kDa OMP at 22°C (6). Aeromonas strains isolated from feces also showed higher mannose-resistant adherence than strains isolated from samples of food and river water (28). Interestingly, in this study, Con A reacted with all A. caviae strains grown at 37°C and was remarkably specific for AeC398 and AeV11 strains of A. caviae. The lectin­like reactivity could be associated with the ability of these bacteria to adhere to carbohydrate-rich surfaces such as the human gut, contributing to enteric diseases. Thus, complementary studies are required in order to better understand the role of cell-surface carbohydrates in pathogenesis of A. caviae.

Table 1

Activity of lectins of various specificities for strains grown at 37°C and 22°C.

	Minimum lectin concentration (μg ml-1) required to agglutinate
	ATCC 15468 strain		AeC398 strain		AeQ32 strain		AeV11 strain
	37°C	22°C	37°C	22°C	37°C	22°C	37°C	22°C
D-GlcNAc-binding lectins
Triticum vulgaris (WGA)	15.6	_b	_	_	_	_	_	_
Lycopersicon esculentum (LEL)	15.6	15.6	_	_	_	_	_	_
Solanum tuberosum (STA)	15.6	15.6	_	_	_	_	_	_
D-GalNAc-binding lectin
Helix pomatia (HPA)	-	-	125	_	_	_	15.6	15.6
D-Gal-binding lectin
Euonymus europaeus (EEL)	-	-	_	_	31.2	_	_	_
D-Man-binding lectin
Canavalia ensiformis (Con A)	125	125	62.5	_	125	62.5	125	_

Average of three experiments.

Agglutination was not observed even at the highest concentrations tested (500. μg ml-1).

Figure 1

Binding of iodinated lectins Triticum vulgaris (WGA), Euonymus europaeus (EEL) and Canavalia ensiformis (Con A) to A. caviae strains grown at 37°C. WGA-binding to ATCC 15468 strain (▪); EEL-binding to AeQ32 strain (□); Con A-binding to AeC398 strain (●) and to AeV11 strain (○). Each point is the mean ± S.E.M. derived from three separate experiments.

Figure 2

Binding of [125I]-labeled Triticum vulgaris (WGA) to ATCC 15468 strain (▪), [125I]-labeled Euonymus europaeus (EEL) to AeQ32 strain (□) and [125I]-labeled Canavalia ensiformis (Con A) to AeC398 strain (●) and to AeV11 strain (○). The data have been plotted by the method of Steck and Wallach (11) according to the equation: Where C is the concentration of bacteria, n is the number of lectin molecules bound/bacteria, K is the lectin association constant and Lectin is the concentration of lectin in molar. All points represent the average of triplicate experiments and the standard deviation is lower than 10%.