The molecular basis for control of ETEC enterotoxin expression in response to environment and host

Abstract

Enterotoxigenic Escherichia coli (ETEC) cause severe diarrhoea in humans and neonatal farm animals. Annually, 380,000 human deaths, and multi-million dollar losses in the farming industry, can be attributed to ETEC infections. Illness results from the action of enterotoxins, which disrupt signalling pathways that manage water and electrolyte homeostasis in the mammalian gut. The resulting fluid loss is treated by oral rehydration. Hence, aqueous solutions of glucose and salt are ingested by the patient. Given the central role of enterotoxins in disease, we have characterised the regulatory trigger that controls toxin production. We show that, at the molecular level, the trigger is comprised of two gene regulatory proteins, CRP and H-NS. Strikingly, this renders toxin expression sensitive to both conditions encountered on host cell attachment and the components of oral rehydration therapy. For example, enterotoxin expression is induced by salt in an H-NS dependent manner. Furthermore, depending on the toxin gene, expression is activated or repressed by glucose. The precise sensitivity of the regulatory trigger to glucose differs because of variations in the regulatory setup for each toxin encoding gene.

Figure 1. Distribution of CRP and H-NS across the ETEC H10407 genome.

A) The panel shows maps of the ETEC H10407 chromosome (i) and associated plasmids (ii). In each plot, tracks 1 and 2 (blue lines) show the position of genes, track 3 (purple and cyan graph) is a plot of DNA GC content, track 4 (green) is the H-NS binding profile and track 5 (orange) is the CRP binding profile. B) A DNA sequence motif generated by aligning regions of the ETEC H10407 chromosome bound by CRP.

Figure 2. Unoccupied CRP sites on p666 and p948 align with H-NS bound regions.

A) A histogram showing the number of putative CRP binding sites in each of 7 discrete bins. Each bin is delineated by the “score” of the putative CRP site. A high score indicates a better match to the Position Weight Matrix that represents the consensus for CRP binding. B) The graph illustrates binding of CRP to a target from each of the bins shown in Panel A. CRP was used at concentrations of 0, 175, 350 or 700 nM. C) ChIP-seq data for CRP and H-NS binding at five regions of plasmids p666 and p948 that contain unoccupied CRP targets bound by CRP in vitro. The CRP and H-NS binding profiles are plots of sequence read counts at each position of the genome on both the top (above the central line) and bottom (below the central line) strand of the DNA. The y-axis scale is the same in each panel. The scale for H-NS binding is 1,785 reads on each strand and for CRP binding is 14,000 reads on each strand.

Figure 3. The estA2 promoter is activated by a Class I CRP dependent mechanism.

A) Sequence of the estA2 gene regulatory region. The CRP binding site is shown in orange, the UP element is blue and the promoter -10 and -35 elements are shown in purple. The different promoter positions are numbered relative to the transcription start site (+1). B) Location of the PestA2 transcription start site. The gel shows the product of an mRNA primer extension analysis to determine the estA2 transcription start site (Lane 5). The gel was calibrated using arbitrary size standards (A, C, G and T in Lanes 1–4). C) Binding of CRP to PestA2. The panel shows the result of a DNAse I footprint to monitor binding of CRP to the 93 bp PestA2 DNA fragment. The gel is calibrated with a Maxim-Gilbert DNA sequencing reaction. CRP was added at concentrations of 0.35–2.1 µM. D) CRP is required for transcription from PestA2 in vivo. The panel shows a cartoon representation of the 93 bp PestA2::lacZ fusion and a bar chart illustrates LacZ activity in lysates of cells carrying this fusion. Assays were done in LB medium. E) i) Stimulation of PestA2 by CRP in vitro. The figure shows the results of an in vitro transcription reaction. The 112 nt transcript initiates from PestA2 and the 108 nt RNAI transcript is an internal control. CRP was added at a concentration of 350 nM and RNA polymerase was added at a concentration of 400 nM. ii) quantification of band intensities from the in vitro transcription analysis.

Figure 4. The estA2 promoter is repressed by H-NS.

A) The panel shows different PestA2::lacZ fusions. The lacZ gene is shown as a red arrow and the estA2 gene is shown as a blue arrow. PestA2 is illustrated using a bent arrow and the CRP binding site is shown as an orange box. B) H-NS binds to PestA2 only in the presence of flanking DNA. ChIP-PCR was used to measure binding of H-NS to the different PestA2 derivatives cloned in pRW50. PCR products were generated using primers that could detect PestA2 in the context of both the 93 bp fragment and the longer 460 bp fragment. C) The values are β-galactsidase activity values for lysates of M182, or M182Δhns, carrying the different PestA2 derivatives. Assays were done in LB medium.

Figure 5. Comparison of PestA1 and PestA2 reveals differential activity and regulation by CRP.

A) Comparison of PestA1 and PestA2. The panel shows the DNA sequences of PestA1 and PestA2. Bases that are identical are highlighted by a solid vertical line. The CRP sites are shown in orange, the UP element in blue and the core promoter elements in purple. The sequences are numbered with respect to the transcription start site (+1). B) Location of the PestA1 transcription start site. The gel shows products from an mRNA primer extension analysis (Lanes 5 and 6). The gel was calibrated using arbitrary size standards (A, C, G and T in Lanes 1–4). C) Sequences of hybrid estA promoters. The sequences labelled estA2.1 through estA2.7 are derivatives of the 93 bp PestA2 DNA fragment where different sequence elements have been replaced with the equivalent sequence from PestA1. D) The bar chart shows β-galactosidase activity measurements for lysates obtained from cultures of M182, or the Δcrp derivative, containing the indicated hybrid promoter fragment was fused to lacZ. E) The panel shows different PestA1::lacZ fusions. The lacZ gene is shown as a red arrow and the estA1 gene is shown as a blue arrow. PestA1 is illustrated using a bent arrow and the CRP binding site is shown as an orange box. Assays were done in LB medium.

Figure 6. The eltAB promoter is indirectly repressed by CRP.

A) The Panel shows ChIP-seq data for CRP and H-NS binding at the eltAB locus. The sequence of 3 putative CRP binding sites proposed by Bodero and Munson (2009) are shown. The CRP and H-NS binding profiles are plots of sequence read counts at each position of the genome on both the top (above the central line) and bottom (below the central line) strand of the DNA. The y-axis scale for H-NS binding is 1,785 reads on each strand and for CRP binding is 14,000 reads on each strand. B) Results of an Electorphoretic Mobility Shift Assay to measure binding of CRP to the 93 bp PestA2 fragment (Lanes 1–7) or the 359 bp PeltAB fragment (Lanes 8–14). Specific and non-specific binding of CRP is indicated to the left and right of the gel. CRP was added at a concentration of 0.2–7.0 µM. C) Panel (i) shows different PeltAB::lacZ fusions. The lacZ gene is shown as a red arrow and the eltAB operon is shown in purple. PeltAB is illustrated using a bent arrow and the putative CRP binding sites are shown as open orange boxes. In panel (ii) the values are β-galactsidase activity measurements taken in M182 or the Δcrp derivative. Assays were done in LB medium.

Figure 7. The eltAB promoter is directly repressed by H-NS.

A) The panel shows two PeltAB::lacZ fusions. B) The results of a ChIP-PCR analysis used to measure binding of H-NS to the two different PeltAB derivatives shown in Panel A. C) The values are β-galactsidase activity measurements for lysates of M182, or the Δhns derivative, carrying the different PestA2::lacZ fusions. Assays were done in LB medium.

Figure 8. H-NS and CRP integrate signals of osmolarity and metabolism to control expression of LT and ST.

The figure shows β-galactosidase activity measurements for lysates obtained from cultures of M182 (i) or the Δhns derivative (ii) containing A) PestA2 or B) PeltAB fused to lacZ in plasmid pRW50. Cultures were grown in the presence and absence of 2% glucose and/or salt (60 mM NaCl and 20 mM KCl). Assays were done in M9 minimal medium so that the glucose and salt concentrations could be more accurately controlled.

Figure 9. Modulation of estA2 and eltA transcription during attachment of ETEC E24377A to gut epithelial cells.

A) The figure shows β-galactosidase activity measurements for lysates obtained from cultures of M182 or the Δhns and Δcrp derivatives containing PestA2 (460 bp fragment) or PeltAB (1126 bp fragment) from ETEC 24377A fused to lacZ in plasmid pRW50. B) The panel shows log₂ fold changes in the transcription of crp, hns, eltA and estA in ETEC E24377A cells over a two hour incubation with a Caco-2 intestinal epithelial cell culture (29). The log₂ values represent the fold change in transcription between ETEC cells attached and unattached to Caco-2 cells at each time point. C) The panel shows a scatter plot of absolute crp and estA2 mRNA levels in ETEC E24377A attached to Caco-2 intestinal epithelial cells. Each data point represents a different biological replicate. For each data point the absolute level of hns mRNA is shown in parenthesis. D) The panel shows the survival rate of BALB/C mice (n = 30) after intranasal inoculation with wild type ETEC H10407 or the Δcrp derivative.

Figure 10. An osmo-metabolic gene regulatory circuit comprised of CRP and H-NS controls expression of LT and ST.

The diagram illustrates the regulatory effects of salt, cAMP and glucose on transcription from the various ST and LT promoter regions.