Dra/AfaE Adhesin of Uropathogenic Dr/Afa⁺Escherichia coli Mediates Mortality in Pregnant Rats

AL858 (afaE afaD⁺).The pILL1131 recombinant plasmid that resulted from the cloning of the afaE3 gene of A30 interrupted by the cat cartridge, into the suicide plasmid pGP704, conferred ampicillin resistance to its host. Filter matings were performed between the kanamycin-resistant strain SM10 (λ pir), harboring pILL1131, as a donor and a rifampin mutant of A30 as a recipient. By screening rifampin- and chloramphenicol-resistant transconjugants for ampicillin and kanamycin susceptibility, one clone (AL858) was found to have lost the ability to agglutinate human red blood cells. Hybridization experiments demonstrated that AL858 resulted from a double-crossover event leading to the insertion of a single copy of the cat/BamHI cartridge into the afaE3 gene of A30.

AL861 (afaE⁺afaD).The pILL1238 recombinant plasmid resulted from the cloning of the 2.6-kb PstI fragment of pILL61 (21), containing the afaD gene of A30 interrupted by the insertion of the nonpolar kanamycin cassette (22) into the suicide plasmid pGP704 that conferred ampicillin resistance to its host. Conjugation was performed between SM10 (λ pir) harboring pILL1238 (donor) and a rifampin mutant of A30 (AL852; recipient). After several subcultures, rifampin- and ampicillin-resistant transconjugants were screened for the expression of kanamycin resistance and loss of ampicillin resistance. One such clone (AL861) was found to carry an interrupted afaD gene and an intact afaE gene by PCR. AL861 exhibited the same mannose-resistant hemagglutination and adhesin properties as the parental A30 strain.

AL863 (afaE afaD).The afaD3-mutated derivative of AL858 was obtained in a manner similar to that for AL861. PCR analyses confirmed that AL863 carried mutations in the afaE3 and afaD3 genes.

The growth rates were similar for all these strains.

Animal model.Timed-pregnancy Sprague-Dawley rats were obtained from Harlan Sprague Dawley, Houston, TX. Two days before experimental infection, each animal received one dose of streptomycin (7.0 mg/g of body weight) to eliminate possible sites of infection in the urogenital tract that may occur naturally.

Animals were infected with clinical E. coli strain A30 (afaE⁺afaD⁺) or one of its isogenic mutants, AL861 (afaE⁺afaD), AL858 (afaE afaD⁺), or AL863 (afaE afaD). An inoculum of E. coli in a volume of 200 μl was placed through the cervical os in the cavity of the left uterine horn of each pregnant rat on the 17th day of gestation using a blunted 16-gauge animal feeding stainless-steel needle (Popper & Sons, Inc., New Hyde Park, NY). To determine the relevance of the dose-versus-infection outcome, three different inoculum doses were used at an optical density at 600 nm (OD₆₀₀) of 2.7 (5 × 10¹² CFU/ml), an OD₆₀₀ of 2.0 (5 × 10¹⁰ CFU/ml), and an OD₆₀₀ of 0.8 (5 × 10⁸ CFU/ml). The maternal mortality was evaluated 24 h postinfection. Uteri and kidneys from surviving animals were collected for quantitative bacterial cultures. Tissues were homogenized in phosphate-buffered saline (PBS; pH 7.2), 10-fold dilutions were prepared, and 50-μl aliquots were cultured on L agar and MacConkey plates. The colonization rate was expressed as the number of CFU per gram of tissue.

Invasion assay.Bacterial invasion into HeLa cells was tested by gentamicin protection assay as described earlier (9, 28). Briefly, HeLa cells (∼10,000/well) were seeded in 12-well plates and allowed to grow to 75 to 80% subconfluency for 24 h in minimum essential medium (MEM) supplemented with 10% (vol/vol) fetal bovine serum. The monolayer was washed once with PBS, and fresh MEM without fetal bovine serum was added. Fresh 18-h bacterial cultures grown on L agar were used to prepare suspensions in PBS adjusted to an OD₆₀₀ of 2.0 (5 × 10¹⁰ CFU/ml), and 1 μl of bacterial suspension was added per 1 ml of MEM. After incubation for 3 h at 37°C with 5% CO₂, the bacterial suspension was removed, and the monolayer was washed twice with medium containing gentamicin (200 μg/ml) and incubated in the same medium for an additional 1 h to kill the extracellular bacteria and select for the internalized bacteria. The monolayers were then washed twice with sterile PBS and incubated for 10 min at room temperature with 1 ml of lysis buffer (0.9% NaCl, 0.1% sodium dodecyl sulfate). About 20 μl of the lysate was plated on L agar plates in duplicate and incubated overnight at 37°C with 5% CO₂. On the following day, the CFU were counted, and the results were expressed as CFU/well. The assay was run in triplicate and repeated three times.

Statistical analysis.Statistical analysis was performed with the GraphPad Prism 4 program using the Mann-Whitney U test and the chi-square test, and P values of less than 0.05 were considered significant.

Maternal mortality (animal model).In order to investigate the impact of Dr⁺E. coli virulence factors AfaE and AfaD on the outcome of intrauterine infection, we assessed maternal morbidity and mortality in pregnant rats infected with the isogenic mutants of E. coli expressing a single virulence factor (the afaE⁺afaD and afaE afaD⁺ strains). The clinical strain of E. coli expressing both virulence factors (the afaE⁺afaD⁺ strain) and the double mutant lacking both virulence factors (the afaE afaD mutant) were used as positive and negative controls, respectively.

The highest mortality rate was observed in groups infected with E. coli expressing the AfaE adhesin in combination with AfaD, strain A30 (afaE⁺afaD⁺), or expressing the AfaE adhesin alone, strain AL 861 (afaE⁺afaD) (Fig. 1). The differences in mortality between groups infected with the clinical strain and mutants lacking AfaE (the afaE afaD⁺ and afaE afaD mutants) were significant in animals infected with the highest concentration of bacteria at an OD₆₀₀ of 2.7 (A30 versus AL858, P = 0.0187; A30 versus AL863, P = 0.0143). The difference in mortality between the AL861 (afaE⁺afaD) group and the group infected with the afaE⁺afaD⁺ clinical strain was not significant, irrespective of the concentration of bacteria used for infection, suggesting that AfaE is a major lethal virulence factor of Afa/Dr⁺E. coli.

• Open in new tab
• Download powerpoint

FIG. 1.

Maternal mortality in animals infected with clinical strain E. coli A30 and its isogenic mutants. Three different concentrations of bacteria, 5 × 10¹² CFU/ml, 5 × 10¹⁰ CFU/ml, and 5 × 10⁸ CFU/ml, were used for infection. Bars with asterisks are significantly different from others with the same OD (P < 0.01; chi-square test).

The significance of the dose of infecting agent in the experimental infection outcome was evaluated using three different inocula of bacteria at an OD₆₀₀ of 2.7 (5 × 10¹² CFU/ml), an OD₆₀₀ of 2.0 (5 × 10¹⁰ CFU/ml), and an OD₆₀₀ of 0.8 (5 × 10⁸ CFU/ml). Maternal mortality was observed in animals infected with the higher-dose inocula (those with OD₆₀₀s of 2.7 and 2.0) of bacteria expressing AfaE and/or AfaD and was dose dependent. There was no mortality in the group infected with the afaE afaD double mutant, AL863, irrespective of the concentration of bacteria used for infection (Fig. 1).

Quantitative analysis of tissue infection rate.In order to investigate the contribution of E. coli virulence factors AfaE and/or AfaD to intrauterine infection during pregnancy, we quantitated the infection in the uteri and kidneys of the infected animals. Due to postmortem effects on bacterial growth, quantitative tissue cultures of dead animals were not performed.

In all tissues, irrespective of the concentration of bacteria used for infection, the clinical strain of E. coli (A30) expressing both virulence factors AfaE and AfaD had the highest infection rate. The differences in infection rate between the groups were most pronounced in animals infected with the dose corresponding to an OD₆₀₀ of 2.0 (Fig. 2). In rats infected with single mutants AL861 (afaE⁺afaD) and AL858 (afaE afaD⁺) and the double mutant AL863 (afaE afaD), the infection rate was much lower, and the difference reached statistical significance in the uteri but not in the kidneys. However, in the kidneys of animals infected with the double mutant, the infection rate was lower than in those infected with single mutant infections and reached statistical significance compared to animals infected with the afaE⁺afaD strain but not the afaE afaD⁺ strain.

• Open in new tab
• Download powerpoint

FIG. 2.

(A) Infection rate in uteri of animals infected with the clinical strain E. coli A30 and its isogenic mutants (at an OD₆₀₀ of 2.0 [5 × 10¹⁰ CFU/ml]). (B) Infection rate in kidneys of animals infected with the clinical strain E. coli A30 and its isogenic mutants (at an OD₆₀₀ of 2.0 [5 × 10¹⁰ CFU/ml]). Results for groups with asterisks are significantly different from those for the A30 group (P < 0.01; Mann-Whitney test).

Invasion assay.To better understand the role of AfaE and AfaD virulence factors in the infectious process in vivo, we attempted to evaluate the potential invasive capacity of the clinical strain of E. coli and its isogenic mutants. We used a gentamicin protection assay, which allows the assessment of E. coli invasiveness since gentamicin is unable to kill intracellular bacteria as opposed to extracellular bacteria, which are killed by bactericidal concentrations of this antibiotic (2, 15). All strains used in the experiment were unable to grow on agar plates after 1 h of incubation in MEM containing gentamicin at a working concentration of 200 μg/ml. There was no bacterial growth, and only a few colonies were found after plating the last PBS washings collected prior to the lysis of the monolayer. The results of bacterial internalization into HeLa cells were expressed as the number of recovered colonies per well.

The internalization rates of both AfaE-positive strains (the afaE⁺afaD⁺ and afaE⁺afaD strains) were at the level of 10⁴ CFU/well and were significantly higher than those in strains which did not express AfaE (P < 0.01). This might suggest that AfaE is mostly responsible for the binding and internalization of bacteria and that mutation of the afaE gene results in the loss of attachment/invasion properties of this strain. It is interesting to note, however, that the highest internalization rate was observed in the AL861 (afaE⁺afaD) strain and that the difference was significant compared to that in the afaE⁺afaD⁺ clinical strain (P < 0.01) (Fig. 3).

• Open in new tab
• Download powerpoint

FIG. 3.

Invasion of wild-type E. coli A30 expressing Dr/Afa adhesins and its isogenic mutants into HeLa epithelial cells in a gentamicin protection assay. Bars marked with different letters are significantly different from one another (P < 0.01; Mann-Whitney test).