Degree Name

Doctor of Philosophy


School of Biological Sciences


Escherichia coli are one of the most diverse species of bacteria found in the gastrointestinal tract (GIT) of vertebrates. E. coli usually reside harmlessly, confined to the intestinal lumen of mammals, however certain groups of E. coli can cause a wide spectrum of intestinal and extraintestinal diseases in humans and animals. Based on clinical pathogenesis and genetic make up E. coli is broadly classified into three major groups: extraintestinal pathogenic E. coli (ExPEC), intestinal pathogenic E. coli (enteric or diarrheagenic) and commensal E. coli. Diseases caused by E. coli result in considerable economic loss around the world. Consequently, considerable focus has been given to understand pathogenic E. coli. The resultant is a reduced focus on the commensal form of E. coli and therefore a deficit in the knowledge base. To improve our knowledge of commensal E. coli and the different microenvironmental niches it inhabits, this study focuses on understanding the genetic constituents of all three groups of E. coli. In this study we utilised a human model for the understanding of ExPECs by focusing on urinary tract infection causing E. coli. These strains are readily available from human clinical cases. For investigating commensal and pathogenic E. coli that colonise the gastrointestinal tract we utilised a porcine model as we were able to access different intestinal regions for analysis. Enterotoxigenic E. coli (ETEC) that cause porcine neonatal and post-weaning diarrhoea and commensal E. coli that colonises different regions of swine GIT were utilised in this study to understand the genetic constituent of different groups of E. coli.

Initially, this study investigated the genetic diversity of 159 E. coli strains isolated from patients with symptomatic and asymptomatic bacteriuria (ABU) in catheterized (CA-ABU) and non-catheterized patients using plasmid replicon typing, bacteriocin, virulence gene (VG) profiling and random amplified polymorphic DNA (RAPD). Plasmid replicon typing revealed fourteen different plasmid types, with 84% of the strains carrying at least one plasmid. Replicon FIB was the most common replicon identified regardless of clinical syndrome. Bacteriocin profiling identified carriage in 56% of strains, with 29% of strains containing colicin genes and 38% of strains containing microcin genes. Colicin Ia/Ib, E1 and colicin V were the most common colicin genes in the collection; microcin M and H47 were the most common microcin genes. Comparative analysis of bacteriocins and VGs identified a sub group of strains with a high number of VGs (median score of 9) and microcins M/H47.

Analysis of RAPD data revealed that there was no significant difference between the E. coli strains associated with each UTI syndrome. This study identified associations between microcin genes, VGs and specific plasmid types in ABU, CA-ABU and symptomatic UTI E. coli strains.

This study also investigated the carriage of plasmids, bacteriocins, VGs, ARGs and integrons in 73 Australian porcine ETEC that cause neonatal, pre and post-weaning diarrhoea in pigs. Plasmid typing of these isolates identified twelve different types with all isolates carrying at least one or more plasmids. IncI1 plasmids were the most common type identified, followed by Inc FIB, HI1 and FIC. The bacteriocin analysis demonstrated that colicins are significantly more prevalent in ETEC than microcins, at a rate of 60% to 7% respectively. Colicins Ia/Ib, B & M and E7 & E3 were the most commonly identified colicins. An overwhelming presence of ARGs and integrons were detected in the ETEC isolates. These included genes encoding resistance to streptomycin (aadA), sulphonamides (sulI), β-lactam (blaTEM), AmpC-family β- lactam group EBC (blaMIR1 or blaACT1) and class 1 integron. Cluster analysis successfully differentiated the different serogroups due to differences in their carriage of VGs, ARGs, bacteriocins, plasmids and integrons.

One of the most important steps in the diagnosis and epidemiological understanding of neonatal, pre and post-weaning diarrhoea by porcine ETEC is accurate serogrouping. In many instances, however, conventional serogrouping fails to produce accurate identification of serogroups. In this study, we developed a modified and simplified molecular serogrouping method (rfb-RFLP) for the accurate identification of the most common porcine ETEC strains that cause neonatal, pre and post weaning diarrhoea in Australia. This method will be of value in routine diagnosis and epidemiology of ETEC infections in pigs.

By utilizing commensal E. coli isolates from different compartments of the pig gut (n=146), the present study has sought to compare the E. coli intestinal community firstly via characterization of adaptation genes by plasmid and bacteriocin typing and secondly via the genome structures by serogrouping (conventional and molecular serogrouping) and RAPD. Plasmid and bacteriocin profiling revealed significant differences in their carriage among E. coli from different intestinal compartments. The most frequently detected plasmid types were: duodenum FIB (53.3%) and I1 (16.7%); ileum N (55.6%), HI1 (47.2%) and FIB (47.2%); colon FIB (30.6%) and I1 (22.2%); faeces FIB (59.1%) and FIA (43.2%). The bacteriocin analysis revealed carriage in 45% of the isolates (colicin 43% and microcin 3%). The commonly detected colicins were: duodenum Ia/Ib, B and M; ileum Ia/Ib, B, M and E7; colon E7 and E1; faeces E7 and E2. Conventional and molecular serogrouping (rfb-RFLP) resolved 26 serogroups. The most common serogroups were O40, O100 and O71. RAPD analysis using primers 1247, 1254 and 1290 revealed 87 different patterns but did not exclusively separate the isolates on the basis of intestinal compartmentalization. RAPD patterns of some isolates and certain serogroups were unique to each compartment. This study shows that communities of E. coli which have adapted to different intestinal compartments have done so by the acquisitions of specific genome structures and combinations of specific bacteriocin genes as well as the extended gene repertoire from plasmids.

This study provides a comprehensive understanding of bacteriocins, plasmids and genetic relatedness of commensal and pathogenic E. coli that colonise the human urinary tract and porcine gastrointestinal tract. The bacteriocin analysis revealed differences in its carriage among different groups of E. coli. The plasmid typing demonstrated that certain plasmid types such as Inc FIB are found in all groups of E. coli; however there are differences in plasmid profiles among each group of E. coli. The bacteriocin and plasmid status of these commensal and pathogenic E. coli is of value in screening potential probiotic strains which have antagonistic effects on ExPECs and ETEC that causes UTI in humans and diarrhoea in pigs. The development of the modified molecular serogrouping facilitates the use of this method in routine for diagnostics and epidemiological studies.

Furthermore, this study has developed a whole cell biosensor to detect and quantify the induction of the SOS response activated by DNA-degrading colicins. This biosensor utilises the SOS-responsive cda promoter to regulate the expression of green fluorescent protein. The biosensor assay revealed induction of stress for all DNA-degrading reference colicins (E2, E7 and E8). In addition, it was successfully used to screen and identify previously uncharacterized wild type colicin producers. This would be of value in visualizing colicin mediated bacterial interactions.



Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.