SummaryEnteropathogenic Escherichia coli (EPEC) produces attaching and effacing lesions (AE) on epithelial cells. The genes involved in the formation of the AE lesions are contained within a pathogenicity island named the locus of enterocyte effacement (LEE). The LEE comprises 41 open reading frames organized in five major operons: LEE1, LEE2, LEE3, LEE4 and tir. The first gene of the LEE1 operon encodes a transcription activator of the other LEE operons that is called the LEE-encoded regulator (Ler). The LEE2 and LEE3 operons are divergently transcribed with overlapping 210 promoter regions, and gene fusion studies have shown that they are both activated by Ler. Deletion analysis, using lacZ reporter fusions, of the LEE2 and LEE3 promoters demonstrated that deletions extending closer to the LEE2 transcription start site than 2247 bp lead to loss of activation by Ler, whereas only 70 bp upstream of the LEE3 transcription start site is required for Ler-mediated activation. We have purified Ler as a His-tagged protein and used it to perform DNA-binding assays with LEE2 and LEE3. We observed that Ler bound to a DNA fragment containing the 2300 to 11 region of LEE2; however, it failed to bind to a DNA fragment containing the 2300 to 11 region of LEE3, suggesting that Ler activates both operons by only binding to the regulatory region upstream of LEE2. The Ler-activatable LEE3::lacZ fusions extended to what would be 2246 bp of the LEE2 operon. A lacZ fusion from the 2300 to 11 region of LEE3 failed to be activated by Ler, consistent with our hypothesis that Ler activates the expression of LEE2 and LEE3 by binding to a region located downstream of the LEE3 transcription start site. DNase I footprinting revealed that Ler protected a region of 121 bp upstream of LEE2. Purified Ler mutated in the coiled-coil domain was unable to activate transcription and to bind to the LEE2 regulatory region. These data indicate that Ler may bind as a multimer to LEE2 and activate both divergent operons by a novel mechanism potentially involving changes in the DNA structure.

The type III secretion system (TTSS) of enteropathogenic Escherichia coli (EPEC) has been associated with the ability of these bacteria to induce secretion of proinflammatory cytokines, including interleukin-8 (IL-8), in cultured epithelial cells. However, the identity of the effector molecule directly involved in this event is unknown. In this

SummaryThe membrane proteins ToxR and ToxS regulate a variety of genes associated with the virulence of Vibrio cholerae, the agent of human cholera. One of the ToxRS-regulated genes is the ompU gene, which encodes a porin that may also act as an adhesin. To begin to understand the mechanism of ompU transcription activation by ToxRS, we performed genetic and biochemical studies on the ompU promoter. Deletions with a 5Ј end-point at or downstream of ¹128, relative to the start site for transcription, did not direct expression of a lacZ reporter gene in wild-type V. cholerae, although the ¹128 promoter fragment did direct ToxRS-dependent reporter gene activity under conditions of ToxR overexpression in E. coli. Consistent with the activation data is that membranes containing ToxR and ToxS caused a gel electrophoretic mobility shift when mixed at low concentrations with deletion fragments whose end-point is at ¹211, but not with ¹128 or ¹68 fragments. ToxRS membranes did shift the ¹128 fragment when added at higher concentrations. DNase I footprinting analysis of ompU promoter DNA complexed with ToxRS membranes demonstrated protection of three sites: an upstream site ranging from ¹238 to ¹139, and two downstream sites ranging from ¹116 to ¹58 and ¹53 to ¹24. Within the DNA protected from DNase I digestion by ToxRS membranes, there are no elements bearing similarity to those identified previously within the promoters of two other ToxR-dependent genes, ctxA and toxT. We suggest a model for transcription activation that involves sequential ToxR-binding events to distinct regions in the ompU promoter.

SummaryIn pathogenic Vibrio cholerae, at least 17 genes are co-ordinately regulated by ToxR. Most of these genes, including those that encode cholera toxin (CT), toxin co-regulated pilus (TCP), accessory colonization factor (ACF) and OmpU, are positively regulated. OmpT is the only identi®ed protein under negative regulation of ToxR. To understand the molecular mechanism by which ToxR represses OmpT expression, we cloned ompT and characterized the ompT promoter and its interaction with ToxR. Sequence analysis revealed that ompT encodes a predicted 35.8 kDa outer membrane porin of V. cholerae. Primer extension analysis identi®ed a transcriptional start site 104 bp upstream of the translational start codon. Both primer extension analysis and promoter fusion studies showed that ToxR represses OmpT expression at the transcriptional level. Promoter fusion studies also suggest that cyclic AMP receptor protein (CRP) is involved in ompT activation. Gel mobility shift assays combined with DNase I footprinting analysis demonstrated that ToxR mediates repression of ompT transcription by directly binding to an A/T-rich region between À95 and À30 of the ompT promoter. To further understand how the interaction of ToxR with different promoters results in its function as an activator or repressor, we have also mapped the regions on the ctxAB and toxT promoters to which ToxR binds. The regions protected by ToxR on each of these promoters are all A/T rich and large in size, although they are positioned differently relative to each transcriptional start site.

We have constructed an improved recombination-based in vivo expression technology (RIVET) and used it as a screening method to identify Vibrio cholerae genes that are transcriptionally induced during infection of infant mice. The improvements include the introduction of modified substrate cassettes for resolvase that can be positively and negatively selected for, allowing selection of resolved strains from intestinal homogenates, and three different tnpR alleles that cover a range of translation initiation efficiencies, allowing identification of infection-induced genes that have low-to-moderate basal levels of transcription during growth in vitro. A transcriptional fusion library of 8,734 isolates of a V. cholerae El Tor strain that remain unresolved when the vibrios are grown in vitro was passed through infant mice, and 40 infection-induced genes were identified. Nine of these genes were inactivated by in-frame deletions, and their roles in growth in vitro and fitness during infection were measured by competition assays. Four mutant strains were attenuated >10-fold in vivo compared with the parental strain, demonstrating that infection-induced genes are enriched in genes essential for virulence.Much remains to be learned about genes that the facultative pathogen Vibrio cholerae induces during infection and how their protein products function during the complex and dynamic process in which this pathogen adapts to the human small intestine. Several new methods have been developed that are helping us to explore this process, including in vivo expression technology (IVET), signature-tagged mutagenesis, microarray technology, differential fluorescence induction, in vivo-induced antigen technology, and real-time reverse transcription-PCR, among others. A specific IVET method, recombination-based IVET (RIVET), has been used previously to identify V. cholerae genes that are induced during infection of infant mice (1, 4). RIVET is very sensitive to low or transient expression of in vivo-induced (ivi) genes during infection and is therefore capable of identifying members of this potentially interesting class of genes. However, this sensitivity is also a double-edged sword, as some ivi genes have low-to-moderate levels of expression in vitro and will therefore be lost during library construction, i.e., there will be premature excision (resolution) of the selectable cassette in such strains. In the present study, we have developed a modified RIVET that can overcome this main disadvantage and used it as a large-scale screening method to identify V. cholerae genes that are transcriptionally induced during infant mouse infection. Briefly, the new system incorporates two different resolvable cassettes that differ in the efficiency of excision, as well as three different alleles of the resolvase-encoding gene tnpR that have different efficiencies of translation initiation. These modifications extend the number of V. cholerae strains that are unresolved in vitro that can be generated in the final library. Additional modificatio...