Cytotoxic factor influencing acquired antimicrobial resistance in Pseudomonas aeruginosa

The Gram-negative opportunistic bacterium Pseudomonas aeruginosa is associated with different types of human infections and because of emerging multidrug-resistant strains, these infections are of major global public health concern. Certain strains possess a unique cytotoxic effector protein ExoU, which contributes to the fitness of this organism in different ecological niches and is associated with acquired antibiotic resistance. This article summarises the current knowledge of the exoU gene in P. aeruginosa, including genetics, distribution in strains from different locations and association with antibiotic resistance. Understanding of this effector protein may have important implications for the understanding of pathogenesis and antimicrobial resistance in P. aeruginosa infections.

The type III secretion system (TTSS), which injects effector proteins into host cells, is an important determinant of virulence in the opportunistic pathogen Pseudomonas aeruginosa. The bacterium encodes four effector proteins that are secreted by the TTSS: ExoY, ExoS, ExoT and ExoU¹. Of note, a complete set of genes for these effector proteins may not be present in all isolates. Almost all strains carry both the exoY and exoT genes. However, a strain may possess either exoU or exoS but very rarely both^2,3. ExoU has phospholipase activity that rapidly kills cells and hence is associated with severe disease outcomes⁴.

The exoU gene was first identified in a highly cytotoxic P. aeruginosa strain PA103³. Using in vitro infection studies, Fleiszig et al. demonstrated that PA103 was highly cytotoxic despite the absence of the exoS gene³. It was subsequently found that the strain encodes a 72 kDa protein that was associated with the cytotoxicity and was named ExoU, after the nomenclature of ExoS and ExoT⁵. ExoU had been previously identified independently and called PepA (Pseudomonas exoprotein A), although ExoU is now the accepted name. Both studies demonstrated that an exoU knockout strain was non-cytotoxic in in vitro and in vivo infection models^5,6. Further studies indicated that exoU is a variable trait amongst clinical isolates of P. aeruginosa⁶. Several lines of evidence from earlier molecular studies have suggested that exoU and its cognate chaperon spcU are located within a region of the chromosome associated with genomic plasticity and the percent G+C of exoU/spcU (58.8) is less than the average percent G+C content (66.6) of P. aeruginosa^5,7. These findings implied that exoU may be a recently acquired gene in P. aeruginosa.

In phylogenetic comparisons, exoU+ strains form a separate clade, which contains less strains than the clade composed of exoS+ strains^8,9. There is an unequal distribution of exoU+ strains in different clinical and environmental settings, with the exoU+ strains being less abundant than exoS+ strains. In acute infections, exoU+ strains are isolated from 28–42% of samples whilst in chronic infections, such as lung infection of cystic fibrosis patients, the rate of isolation of exoU+ strains is less than 10%^2,4,10. Furthermore, exoU+ strains were less prevalent in environmental samples compared to clinical samples. However, within the samples from the environment, exoU+ strains were over-represented in domestic (e.g. sinks, drains, toilets, fountains, hoses)¹¹ and hospital (e.g. sinks and washtubs of intensive care units)¹² environments versus natural environments (e.g. soil and plants).

Perhaps the most intriguing difference in the distribution of exoU+ strains was noted in the strains isolated from corneal ulcers¹³. Strains possessing exoU have been isolated from corneal samples in 33% to 61% of cases^8,14–16,. These rates are higher than those commonly observed in other acute infections. The risk of keratitis caused by P. aeruginosa is linked to contact lens wear. ExoU+ strains are more commonly isolated from keratitis associated with contact lens wear^14,15. Furthermore, ulcers caused by exoU+ strains result in worse visual outcomes than those caused by strains lacking exoU¹⁶. The main reason for this could be that the exoU+ strains tend to be more prevalent in household tap water including sinks and drains, which could contaminate contact lens as many people insert lenses in the bathroom. However, there is lack of literature defining the roles of environmental reservoirs in dictating the prevalence of exoU+ versus exoS+ strains in keratitis. More studies are needed to confirm the disproportionate distribution of exoU+ and exoS+ strains in environmental reservoirs and fitness of different genotypes in various environments.

With the availability of metagenome data, the genomic context of the exoU carrying genomic island has been elucidated. The exoU island is derived from the integrative plasmid pKCL102 and is integrated at a tRNA-Lys gene adjacent to the locus PA0976 of PAO1^17,18. The pKCL102 contains XerC/XerD-integrase, which is responsible for site-specific integration and because of possession of origin of replication (oriV), it can replicate autonomously. Thereby, pKCL102 can occur in multiple copies¹⁷. We have previously demonstrated that in the keratitis strain PA34, the exoU gene is found on a 7.5 kb island flanked by tRNA-Lys and homologs of the PAO1 genes PA0976 and PA0988¹⁹. However, there is a wide variation in the size of the exoU island between strains, for example, it is 14 kb (PAPI-2) in strain PA14 (highly virulent reference strain), 81 kb (exoU island A) in ocular isolate 6077, 29.8 kb (exoU island B) in ocular isolate 19660, 3.9 kb (exoU island C) in blood isolate X13273 and 3.5 kb in an environmental isolate (Figure 1)^20,21. Despite the differences in the size of exoU carrying islands between strains, the exoU gene appears to encode a functional cytotoxin¹⁹. The difference in island size is attributed to the presence of several mobile elements including insertion sequences and transposons, which can be subjected to recombination, deletion or elimination in response to the environmental selective pressure. It has been hypothesised that such mobile elements potentially lead to excision of exoS leaving the strain as exoU+ (cytotoxic)²¹. However, exoU and exoS are not closely located on the chromosome and no explanation has emerged for strains that carry either both exoU and exoS or neither. More importantly, understanding the selection pressures such as growth in the presence of predators (e.g. protozoa) or antimicrobials that might favour the acquisition of exoU in different environmental settings will help to better understand pathogenesis and epidemiology of different types of P. aeruginosa infections.

Figure 1.  Graphical representation of various exoU carrying islands. The phylogenetic tree is based on core genome SNPs, constructed using Parsnp v1.2²². Protein coding regions are represented by arrows and key features/associated genes are shown in different colours (as labelled in the figure). Strains were selected based on previous literature and the name of the genomic islands given are as originally published.

Not only are exoU carrying strains able to cause more severe infections, co-selection of exoU gene and antimicrobial resistance, including disinfectant resistance, has been widely reported in exoU+ ocular and non-ocular strains^14,23–26. This has raised concerns that antibiotic resistance may be a factor for the evolution of more virulent strains of P. aeruginosa. Most studies have shown a significant correlation of exoU with fluoroquinolone resistance in P. aeruginosa. The majority of exoU+ strains have mutations in DNA gyrase (gyrA and gyrB) and topoisomerase (parC and parE), which are responsible for fluoroquinolones resistance. However, deletion or acquisition of exoU alone (not the pathogenicity island) in experimental models does not affect fluoroquinolone susceptibility¹⁶, which suggests that other genes in the pathogenicity island may be important in the development of this resistance phenotype. Limited studies have also shown a correlation between exoU carriage and resistance to beta-lactams and aminoglycosides^16,27. Our previous study examined the correlation between carriage of exoU or exoS and mutation and expression of beta-lactamase genes. We observed that exoU+ strains were usually more resistant to beta-lactams than exoS+ strains (Figure 2)²⁸, and this may have been due to exoU+ strains having more mutations in genes associated with beta-lactam resistance (mexR, ampC and ampR). Gene expression analysis suggested that such mutations generally lead to antibiotic resistance²⁸. However, the reason for a higher mutation rate in above-mentioned resistance genes in exoU+ strains remains unanswered. Perhaps selection pressure, which favours acquisition of exoU+, is associated with higher mutation rates in exoU strains.

Figure 2.  Antibiotic susceptibility patterns of strains carrying exoU (n = 8) or exoS (n = 14). Minimum inhibitory concentrations are represented as dots with median values shown. Bars represent the range of observed MICs (data taken from Subedi et al.²⁸).

Taking all the information together, it is clear that exoU in P. aeruginosa is ecologically important and an important determinant of virulence and antimicrobial resistance. Perhaps screening for exoU might help predict clinical outcomes and resistance patterns, which in turn could lead to development of strategies of improved therapies. Furthermore, given that exoU+ strains are distributed differently in the different environment, the exoU profile of strains may help to track the epidemiology of P. aeruginosa infections. Further studies on the incompatibility of exoU and exoS within the same P. aeruginosa genome, and factors favouring acquisition of the exoU might help to understand acquired resistance in P. aeruginosa.

The authors declare no conflicts of interest.