Analysis of O-island deletions in Escherichia coli O157:H7

Abstract

Escherichia coli (E. coli) are a diverse species of bacteria that reside, often harmoniously and beneficially, in the gastrointestinal tracts of humans and other mammals. However, some strains are associated with serious intestinal and extra-intestinal disease and are considered pathogens. The main differences between strains of these different E. coli pathotypes can be explained by the acquisition of genetic information introduced by mobile genetic elements, in particular bacteriophage. In enterohaemorrhagic E. coli (EHEC) O157:H7 strain EDL933, a pathotype of E. coli containing prophage-encoded Shiga toxins associated with severe gastrointestinal and systemic disease in humans, these horizontally acquired elements have been termed O-islands (OIs) and include both fully functional and cryptic prophages. The overall aim of this research was to try and determine what these OIs are actually doing for the bacteria. Systems pertinent in the life cycle and virulence of this pathogen were therefore investigated by phenotypically screening a large library of OI deletions in EHEC strain TUV93-0, a Shiga toxin-negative derivative strain of EDL933, and then comparing these with the parent strain. These analyses highlighted a subset of OIs with the potential to regulate motility and type III secretion (T3S), the latter being an essential colonisation factor for EHEC that is encoded by the locus of enterocyte effacement (LEE). Deletion of OI-51, a 14.93 Kb cryptic prophage designated as CP-933C, significantly reduced persistence of faecal shedding in sheep and levels of T3S expression in vitro. Cloning and complementation together with targeted allelic replacements in OI-51 identified a novel positive regulator of the LEE, encoded by ecs1581 in the sequenced E. coli O157:H7 strain Sakai that is present but not annotated in the EDL933 sequence. Functionally important residues of ECs1581 were identified by site-directed mutagenesis based on phenotypic variants present in strains from different E. coli pathotypes, including strains not harbouring a LEE-encoded T3S system. This regulator was subsequently termed RgdR based on a motif demonstrated to be important for stimulation of gene expression from LEE1. Purified RgdR protein was able to form multiple complexes on a PCR generated LEE1 promoter fragment, and activation of this operon appeared to require this DNA binding capacity as a non-T3S inducing variant was unable to bind this same LEE1 promoter fragment. RgdR did not directly activate LEE1 transcription in vitro, nor did it activate transcription by relieving H-NS repression as proposed for the global regulator Ler (LEE-encoded regulator). However, RgdR activation did require a wild type LEE1 promoter and the Ler auto-induction cycle to induce LEE2-5 expression and T3S. RgdR was able to increase binding to Congo red and was capable of repressing bacterial motility. Further analyses demonstrated that RgdR did not regulate T3S and cell motility via GrlA (global regulator of LEE activator) and QseC (quorum sensing E. coli regulator C), two established regulators in E. coli that control LEE gene expression and motility in conjunction with their partners, GrlR (global regulator of LEE repressor) and QseB (quorum sensing E. coli regulator B) respectively. RgdR is therefore identified as a novel regulator able to co-ordinate T3S and motility expression. This research has identified OI-51 as being important for EHEC O157:H7 colonisation in sheep and has identified a completely new family of small bacterial regulators that control surface factor expression in E. coli.