Presence of the exoU Gene of Pseudomonas aeruginosa Is Correlated with Cytotoxicity in MDCK Cells but Not with Colonization in BALB/c Mice

Abstract

A total of 141 independent strains of Pseudomonas aeruginosa with different heterogeneities in the exo gene (exoS, exoT, exoU, and exoY) background were examined for their pathogenic roles. Results indicated that the exoU gene is the major contributor to cytotoxicity in Madin-Darby canine kidney cells but is not related to bacterial colonization in mice.

Pseudomonas aeruginosa harbors at least one or more exoS, exoT, exoU, and exoY genes that are translated into protein products related to type III secretion systems (TTSS). These products have been demonstrated to show a cytotoxic effect in vitro. Furthermore, in clinical studies, the presence of these toxins is associated with a dissatisfactory clinical outcome among patients with P. aeruginosa infection (6, 7, 9, 10). However, these TTSS-related toxin genes show heterogeneity in terms of their presence in the genome of individual P. aeruginosa strains. For example, the presence of exoU and that of exoS are mutually exclusive (2, 4, 5). Normally, more than 90% of clinical strains contain exoT and exoY genes, but clinical isolates, if they are isolated from urine, frequently have the exoY gene present at a relatively lower level—about 70% of strains (4, 5). Due to the complicated interactions that seem to occur during TTSS-related toxin gene expression, there is uncertainty about what the presence of these various different toxin genes means in terms of P. aeruginosa strain toxicity and intrinsic strain virulence (11). In this study, we examined which TTSS-related toxicity effectors are associated with cytotoxicity against mammalian cells or are involved in bacterial colonization of the spleen and liver of animals.

The isolates of P. aeruginosa were collected from different sources, either bacteremic patients (n = 100) or environments (n = 41; including water towers or conduits), in Kaoshiung, Taiwan. Their identification as P. aeruginosa was performed by biochemical characterization using an automatic system (BD Phoenix 100 Automated Microbiology System; Becton, Dickinson and Company, Franklin Lakes, NJ). Chromosomal DNA from each isolate was extracted using a DNeasy Miniprep kit (QIAGEN Co. Ltd., Chuo-Ku, Tokyo, Japan) and then used as a template for PCR. All isolates were demonstrated to be independent based on random amplified polymorphic DNA profiling (3). This profiling was analyzed using 6% polyacrylamide gel electrophoresis and Molecular Analyst fingerprinting software (Bio-Rad, Hercules, CA). The banding patterns were tested for reproducibility by profiling them in duplicate, and an independent strain was identified when two or more amplified bands were distinct from each other. The genotype in terms of TTSS genes was determined by amplifying specific TTSS amplicons using the methods of Ajayi et al. (1).

Madin-Darby canine kidney (MDCK) epithelial cell cytotoxicity was determined by the release of lactate dehydrogenase (LDH) into the supernatant. Once the MDCK cells (10⁵/well) had attached to 24-well plastic trays, the wells were washed three times and the culture medium replaced by fetal calf serum-free Dulbecco's modified Eagle's medium. After a 1-h incubation, individual bacterial isolates (10⁷ CFU/ml) were added to the wells. After another 3 h, the concentration of LDH was quantified using a cytotoxicity kit according to the manufacturer's instructions (Sigma). The proportion of cells lysed (percentage) in a sample well was calculated as follows: {[(LDH activity in sample well) − (spontaneous release)]/[(maximum amounts of LDH release) − (spontaneous release)]} × 100.

To determine the colonization of P. aeruginosa in vivo, BALB/c mice (7 to 8 weeks old) were inoculated via the tail vein with 10⁶ CFU of a series of representative P. aeruginosa strains suspended in sterile phosphate-buffered saline (50 μl). At 18 h postinfection, the inoculated mice were killed and their livers and spleens excised. These organs were weighed and separately homogenized in sterile phosphate-buffered saline, and then serial dilutions were plated on LB agar. The CFU were determined on the next day.

MDCK cell lysis was found to be correlated with the presence or absence of the exoU (P < 0.001) or exoY (P < 0.001) gene in the P. aeruginosa isolates after a 3-h treatment, but the exoS gene showed a reverse correlation (P < 0.001). After consideration of the relationship between the different exo genes and clinical/environmental isolates by multiple regression analysis, the levels of cytotoxicity were found to be significantly related to the presence of the exoU gene (β = 0.595; P = 0.024), more than that of the exoY gene (β = 0.241; P = 0.001). No significant effect was found for the presence of the exoS gene (P = 0.946). In our collection of 141 isolates, 5 TTSS-related toxin genotypes were present, namely exoT⁺ exoU⁺, exoT⁺ exoU⁺ exoY⁺, exoS⁺ exoT⁺, exoS⁺ exoT⁺ exoY⁺, and exoS⁺ exoT⁺ exoU⁺ exoY⁺. On comparing the interaction of each individual gene within the exo gene complex, the mean cytotoxic effect of P. aeruginosa strains with the genotype exoT⁺ exoU⁺ exoY⁺ (42.8% ± 10.0%) was greater than that of exoS⁺ exoT⁺ exoY⁺ (29.2% ± 6.7%) or exoS⁺ exoT⁺ (24.7% ± 6.2%) strains (P < 0.001). The 18.1% difference between exoT⁺ exoU⁺ exoY⁺ strains and exoS⁺ exoT⁺ strains was higher than that between exoT⁺ exoU⁺ exoY⁺ strains and exoS⁺ exoT⁺ exoY⁺ strains (13.6%), suggesting that the exoU gene is the “major gene” and that the exoY gene does contribute to some extent to MDCK cytotoxicity caused by P. aeruginosa (Table 1).

TABLE 1.

Comparison of cytotoxic effects of bacteria on MDCK cells and of bacterial colonization in BALB/c mice

Group no. (TTSS genotype)a	Cytotoxicity (%)b	Colonization (log CFU/g)c
		Spleen	Liver
1 (T+U+)	31.0 ± 8.5	3.00 ± 0.58	3.48 ± 0.41
2 (T+U+Y+)	42.8 ± 10.0	3.89 ± 0.85	3.88 ± 0.44
3 (S+T+)	24.7 ± 6.2	3.76 ± 0.20	3.83 ± 0.70
4 (S+T+Y+)	29.2 ± 6.7	3.69 ± 0.09	3.63 ± 0.24
5 (S+T+U+Y+)	43.1 ± 9.9

^aBacterial genotypes are as follows: T⁺U⁺, exoT⁺ exoU⁺; T⁺U⁺Y⁺, exoT⁺ exoU⁺ exoY⁺; S⁺T⁺, exoS⁺ exoT⁺; S⁺T⁺Y⁺, exoS⁺ exoT⁺ exoY⁺; S⁺T⁺U⁺Y⁺, exoS⁺ exoT⁺ exoU⁺ exoY⁺.

^bPercentage of MDCK cells lysed. For bacterial genotype groups 1 to 5, results are means ± standard deviations for 5, 35, 20, 78, or 2 strains, respectively. Significant differences (P < 0.05) were as follows: cytotoxicity of genotype group 2 was significantly greater than that of groups 1, 3, and 4; cytotoxicity of group 5 was significantly greater than that of group 3.

^cResults are means ± standard deviations for four mice (each) for bacterial genotype groups 1 to 4. No significant differences were found.

BALB/c mice mimicking bacteremia were used to estimate the colonizing ability of P. aeruginosa with or without the exoU gene in vivo. Subsequent to tail injection with randomized representative isolates having the genotype exoT⁺ exoU⁺, exoT⁺ exoU⁺ exoY⁺, exoS⁺ exoT⁺, or exoS⁺ exoT⁺ exoY⁺, the concentration of P. aeruginosa in the blood was determined to be <10 CFU/ml but bacterial colonization of the spleen and liver was detectable. When the bacterial loads in the spleen and liver were correlated with the various TTSS genotypes, no significant difference was found (Table 1). The results indicated that the genes producing the TTSS-related toxins did not seem to facilitate bacterial colonization after infection with P. aeruginosa.

Schulert et al. reported that the P. aeruginosa strains isolated from hospital-acquired pneumonia patients were highly virulent if the isolate harbored the exoU gene versus isolates that did not (11). Shaver and Hauser reported that isogenic strains containing the exoU gene had greater virulence against BALB/c mice than strains containing the exoS or exoT gene (12). This effect was not synergistic if P. aeruginosa secreted the ExoU protein in combination with the ExoS or ExoT protein (13). However, P. aeruginosa causes a variety of serious infections involving multiple virulence determinants and accessory factors (8). Multiple processes, such as bacterial colonization, target site invasion, and cytotoxicity, determine the establishment of a P. aeruginosa infection. With 141 independent strains, we have demonstrated that the presence or absence of the exoU gene alone or the exoU gene together with other TTSS-related toxin genes supports the hypothesis that the exoU gene is the major contributor to potential virulence. However, the presence of exoU genes did not act as an indicator for enhanced bacterial loads in the spleen or liver in BALB/c mice. It is likely that other accessory factors alone or combined with the exoU gene were involved in bacterial colonization. The complicated interaction of multiple genes might mask a significant role in the predisposition that the exoU⁺ strains of P. aeruginosa may have to cytotoxicity, which may be relevant for bacterial colonization in vivo. Whether the exoU gene plays a role in virulence after bacterial colonization by P. aeruginosa still needs to be studied.