Genetic studies on uropathogenic Escherichia coli: identification and characterisation of virulence determinants

Scriptie/masterproef: Doctoral Thesis


Urinary tract infections (UTIs) are common in human beings. UTIs include cystitis, pyelonephritis and asymptomatic bacteriuria. Although UTIs are frequently relatively benign, these are often recurrent and cause major discomfort to the patient. Some complications can be severe and even life-threatening. UTIs cause a huge cost to the healthcare system and are a heavy burden for the hospital and the physician.
Most known Escherichia. coli are commensal or cause intestinal diseases. However, some E. coli infect other organs, these are named extra-intestinal pathogenic E. coli (ExPEC). ExPEC causing UTIs are known as uropathogenic E. coli (UPEC). Most community-acquired UTIs are caused by UPEC. Although UPEC have been intensively studied, so far there are no vaccines or effective treatments, other than antibiotics, available to treat for UPEC infections. A detailed understanding of the molecular genetic basis of the virulence of UPEC may contribute to the development of more effective therapeutics and prophylactics.
Comparison of complete genome sequences and epidemiological investigations demonstrated that UPEC share more than 50% of their genes with commensal E. coli and other ExPEC and that the clinical isolates belong to various phylogenetic groups and serotypes.
Virulence factors are the ammunition of the bacterial pathogens to attach to the epithelial cells of the organs of the host, escape from the host defenses, invade the host tissues, survive in the host and/or cause disease. The molecular mechanism of pathogenesis and the regulation of the discovered virulence factors have been intensively investigated. However, the interaction of UPEC with their host is rather complex. UPEC with variety of virulence factors stimulate several responses of the host, including apoptosis, immune and inflammatory responses, endocytosis of UPEC and dispersal of the bacteria in intracellular bacterial communities (IBCs) in bladder epithelial cells.
During the present study, we first deciphered the molecular mechanism of the NAD dependence of UTI89, which is an acute cystitis isolate and of serotype O18:K1:H7. A mutation in the nadB gene, encoding L-aspartate oxidase, was shown to be responsible for the nicotinamide-requirement of UTI89. This was further confirmed by the complementation assay. Construction of site-specific point mutations, selection and analysis of spontaneous prototrophic revertants and sequencing of nadB of other NAD-dependent and independent O18 isolates demonstrated that the Ala28Val mutation in NadB completely impedes de novo synthesis of nicotinamide. However, no significant differences were observed between the virulence of isogenic NAD-auxotrophic and prototrophic strains in the murine ascending urinary tract infection model. We applied the nadB locus as a neutral site for DNA insertions in the bacterial chromosome, that is, we stably inserted other genes into the middle region of nadB to complement the genes inactivated on the chromosome. Such complementation is required to confirm the identification of these genes as virulence factors and is often difficult to perform using plasmids, as these are often lost at an excessive rate in vivo.
The use of the alkaline phosphatase gene phoA as a reporter in UTI89 is hampered by the presence of an indigenous phoA gene, causing a high background activity. Therefore, we decided to first construct a UTI89?phoA mutant. Since UTI89 is a refractory strain for using the Red recombination system, we first constructed the mutant in the E. coli K-12 strain BW25141. It turned out that the transducing phage P1 did not produced high titer lysates on BW25141?phoA::cat. Therefore we first transferred the ?phoA::cat mutation to the wild type E. coli K-12 strain MG1655, and then to UTI89 via P1vir transduction, to produce the UTI89?phoA::cat mutant DV7728. At the same time, we observed that the phoA flanking sequences are almost identical between UTI89 and E. coli K-12. Therefore, we used longer flanking sequences to transfer the phoA deletion to UTI89, by Red-mediated recombination of a linear fragment that was amplified from DV7728 by PCR. The resulting UTI89?phoA mutant, DV7748, was used as a recipient strain for the construction of a transposon mutant library by conjugation with the donor strain S17-1 (?pir) harboring pGV4220, which contains a phoA reporter gene within the miniTn5phoA2 transposon.
We took advantage of the difference of nutrient requirements of DV7748 and S17-1(?pir) on minimal A medium to construct a transposon insertion mutant library. The transposon insertion mutants were selected on minimal A medium containing nicotinamide and kanamycin, and were then further tested on LB containing the PhoA substrate 5-bromo-4-chloro-3-indolyl phosphate (X-P). A set of 103 miniTn5phoA2 insertion mutants were analyzed and the insertions of 96 mutants were localized. In total, 91 discrete insertions were identified. The phenotype of these mutants was tested in vitro and in vivo. We found that the survival ratio of Caenorhabditis elegans feeding on mutant 26 with a disrupted wciP gene is statistically significantly higher than that of C. elegans feeding on the wild type strain. Further, screening for mutants defective in colonization in the bladder and kidneys of C3H/HeN mice, demonstrated tatA and pstC, nhaA as well as degP, yohJ and plasmid-encoded genes p060 and p135 potentially are virulence determinants of UTI89. Comparison of the data obtained in C. elegans and mice showed that several mutations affecting virulence in the mouse do not increase the lifespan of infected C. elegans and would not be identified if the screening is only performed in nematodes.
Using linear fragments amplified from the chromosome, the degP and yohJ transposon insertion mutations were transferred to wild type UTI89. Such construction of isogenic strains in a clean genetic background is required to prove the linkage between a mutant phenotype and an inactivated gene. The results demonstrated that the mucous colony morphology of the yohJ transposon insertion mutant resulted from some unlinked mutation but not from inactivation of yohJ. Comparison of the infection in the bladder and kidneys by degP mutants and the wild type strains in the murine UTI model, confirmed that degP plays an important role both in the cystitis isolate UTI89 and the pyelonephritis isolate CFT073. However, the contribution of degP to the colonization in the pyelonephritis isolate CFT073 is earlier than in the cystitis isolate UTI89.
Several mutations disrupted genes in the indigenous plasmid pUTI89. The plasmid-encoded protein ParB was shown to contribute to the virulence of UTI89. It is one of the proteins encoded by the parAB plasmid partitioning system. We isolated plasmid-free cells when only one of the genes, parB or parA, was replaced by a resistance marker. However, we could not isolate plasmid-free cells from the parAB mutant missing both genes. In addition, the plasmid pCP20, encoding the yeast FLP recombinase, cannot be harbored stably in the parB mutant, which implies some interaction between gene products in the UTI89 parB mutant and the plasmid pCP20.
In conclusion, the present study discovered the molecular mechanism of NAD dependence of O18 UPEC strains, developed a generalized method to transfer mutations between E. coli strains and disclosed a new neutral insertion site in the chromosome for complementation assays. Plasmid-free UTI89 derivatives were isolated by specific targeting of a plasmid partitioning gene, parB or parA. The plasmid partitioning gene parB was first confirmed as a virulence determinant of UPEC in the mouse model.
Datum Prijs26 feb 2013
BegeleiderJean-Pierre Hernalsteens (Promotor), Luc Leyns (Jury), Henri De Greve (Jury), Pierre Cornelis (Jury), Frank Pasmans (Jury) & Herman Goossens (Jury)

Citeer dit