SummaryThe production of exoenzyme S is correlated with the ability of Pseudomonas aeruginosa to disseminate from epithelial colonization sites and cause a fatal sepsis in burn injury and acute lung infection models. Exoenzyme S is purified from culture supernatants as a non-covalent aggregate of two polypeptides, ExoS and ExoT. ExoS and ExoT are encoded by separate but highly similar genes, exoS and exoT. Clinical isolates that injure lung epithelium in vivo and that are cytotoxic in vitro possess exoT but lack exoS, suggesting that ExoS is not the cytotoxin responsible for the pathology and cell death measured in these assays. We constructed a specific mutation in exoT and showed that this strain, PA103 exoT::Tc, was cytotoxic in vitro and caused epithelial injury in vivo, indicating that another cytotoxin was responsible for the observed pathology. To identify the protein associated with acute cytotoxicity, we compared extracellular protein profiles of PA103, its isogenic non-cytotoxic derivative PA103 exsA::⍀ and several cytotoxic and non-cytotoxic P. aeruginosa clinical isolates. This analysis indicated that, in addition to expression of ExoT, expression of a 70-kDa protein correlated with the cytotoxic phenotype. Specific antibodies to the 70-kDa protein bound to extracellular proteins from cytotoxic isolates but failed to bind to similar antigen preparations from non-cytotoxic strains or PA103 exsA::⍀. To clone the gene encoding this potential cytotoxin we used Tn5 Tc mutagenesis and immunoblot screening to isolate an insertional mutant, PA103exoU :: Tn5 Tc, which no longer expressed the 70-kDa extracellular protein but maintained expression of ExoT. PA103 exoU ::Tn5 Tc was non-cytotoxic and failed to injure the epithelium in an acute lung infection model. Complementation of PA103exoU ::Tn5 Tc with exoU restored cytotoxicity and epithelial injury. ExoU, ExoS and ExoT share similar promoter structures and an identical binding site for the transcriptional activator, ExsA, data consistent with their co-ordinate regulation. In addition, all three proteins are nearly identical in the first six amino acids, suggesting a common amino terminal motif that may be involved in the recognition of the type III secretory apparatus of P. aeruginosa.

Pseudomonas aeruginosa, an opportunistic pathogen, is capable of establishing both chronic and acute infections in compromised hosts. Previous studies indicated that P. aeruginosa displays either a cytotoxic or an invasive phenotype in corneal epithelial cells. In this study, we used polarized MDCK cells for in vitro infection studies and confirmed that P. aeruginosa isolates can be broadly differentiated into two groups, expressing either a cytotoxic or an invasive phenotype. In vivo infection studies were performed to determine if cytotoxic and invasive strains displayed differential pathology. Invasion was assayed in vivo by in situ infection of mouse tracheal tissue followed by electron microscopy. Both cytotoxic and invasive strains entered mouse tracheal cells in situ; however, more necrosis was associated with the cytotoxic strain. In an acute lung infection model in rats, cytotoxic strains were found to damage lung epithelium more than invasive strains during the short infection period of this assay. The expression of cytotoxicity requires a functional exsA allele. In the strains tested, the ability to invade epithelial cells in vitro appears to be independent of exsA expression. Since ExsA is a transcriptional regulator of the exoenzyme S regulon, chromosomal preparations from invasive and cytotoxic strains were screened for their complement of exoenzyme S structural genes, exoS, encoding the 49-kDa ADP-ribosyltransferase (ExoS), and exoT, encoding the 53-kDa form of the enzyme (Exo53). Invasive strains possess both exoS and exoT, while cytotoxic strains appear to have lost exoS and retained exoT. These data indicate that the expression of cytotoxicity may be linked to the expression of Exo53, deletion of exoS and perhaps other linked loci, or expression of other ExsA-dependent virulence determinants. In the absence of a functional cytotoxicity pathway (exsA::⍀ strains), invasion of eukaryotic cells is detectable.

Pseudomonas aeruginosa is internalized into multiple types of epithelial cell in vitro and in vivo and yet is often regarded as an exclusively extracellular pathogen. Paradoxically, ExoS, a type three secretion system (T3SS) effector, has antiphagocytic activities but is required for intracellular survival of P. aeruginosa and its occupation of bleb niches in epithelial cells. Here, we addressed mechanisms for this dichotomy using invasive (ExoS-expressing) P. aeruginosa and corresponding effector-null isogenic T3SS mutants, effector-null mutants of cytotoxic P. aeruginosa with and without ExoS transformation, antibiotic exclusion assays, and imaging using a T3SS-GFP reporter. Except for effector-null PA103, all strains were internalized while encoding ExoS. Intracellular bacteria showed T3SS activation that continued in replicating daughter cells. Correcting the fleQ mutation in effector-null PA103 promoted internalization by >10-fold with or without ExoS. Conversely, mutating fleQ in PAO1 reduced internalization by >10-fold, also with or without ExoS. Effector-null PA103 remained less well internalized than PAO1 matched for fleQ status, but only with ExoS expression, suggesting additional differences between these strains. Quantifying T3SS activation using GFP fluorescence and quantitative reverse transcription-PCR (qRT-PCR) showed that T3SS expression was hyperinducible for strain PA103ΔexoUT versus other isolates and was unrelated to fleQ status. These findings support the principle that P. aeruginosa is not exclusively an extracellular pathogen, with internalization influenced by the relative proportions of T3SS-positive and T3SS-negative bacteria in the population during host cell interaction. These data also challenge current thinking about T3SS effector delivery into host cells and suggest that T3SS bistability is an important consideration in studying P. aeruginosa pathogenesis.

Pseudomonas aeruginosa is considered an extracellular pathogen. Using assays to determine intracellular survival in the presence of gentamicin, we have demonstrated that some strains of P. aeruginosa are able to invade corneal cells during experimental bacterial keratitis in mice. Although intracellular bacteria were detectable 15 min after inoculation, the number of intracellular bacteria increased in a time-dependent manner over a 24-h period. Levels of invasion were similar when bacteria were grown as a biofilm on solid medium and when they were grown in suspension. Intracellular bacteria survived in vitro for at least 24 h, although only minimal bacterial multiplication within cells was observed. P. aeruginosa PAK and Escherichia coli HB101 did not cause disease in this model and were not isolated from corneas after 24 h even when an inoculum of 108 CFU was applied. Transmission electron microscopy of corneal epithelium from eyes infected for 8 h revealed that intracellular bacteria were present within membrane-bound vacuoles, which suggests that bacterial entry was an endocytic process. At 24 h, the observation of many bacteria free in the cytoplasm indicated that P.

Pseudomonas aeruginosa is usually considered an extracellular pathogen. Using assays to determine intracellular survival in the presence of gentamicin, we have previously demonstrated that P. aeruginosa is able to invade corneal cells during infectious keratitis in mice. In vitro, P. aeruginosa was found to enter the following cells: human corneal cells removed by irrigation; epithelial cells in the cornea of rats, mice, and rabbits; and primary corneal epithelial cells cultured from rat and rabbit eyes. The level of invasion was related to the level of adherent or associated bacteria. In general, invasion was more efficient with cultured epithelial cells than with cells tested in situ. Invasion did not occur when assays were performed at 4؇C. Cytochalasin D but not colchicine inhibited bacterial invasion, suggesting that bacterial entry was an endocytic process dependent on actin microfilaments but not microtubules. Bacteria that invaded cultured corneal epithelial cells were found to multiply within cells. The ability of P. aeruginosa to invade and multiply within corneal epithelial cells may contribute to the virulence of this organism during infectious keratitis, since intracellular bacteria can evade host immune effectors and antibiotics commonly used to treat infection.