PhoB activates Escherichia coli O157:H7 virulence factors in response to inorganic phosphate limitation

Abstract

Enterohemorrhagic Escherichia coli (EHEC), an emerging food- and water-borne hazard, is highly pathogenic to humans. In the environment, EHEC must survive phosphate (Pi) limitation. The response to such Pi starvation is an induction of the Pho regulon including the Pst system that senses Pi variation. The interplay between the virulence of EHEC, Pho-Pst system and environmental Pi remains unknown. To understand the effects of Pi deprivation on the molecular mechanisms involved in EHEC survival and virulence under Pho regulon control, we undertook transcriptome profiling of the EDL933 wild-type strain grown under high Pi and low Pi conditions and its isogenic ΔphoB mutant grown in low Pi conditions. The differentially expressed genes included 1067 Pi-dependent genes and 603 PhoB-dependent genes. Of these 131 genes were both Pi and PhoB-dependent. Differentially expressed genes that were selected included those involved in Pi homeostasis, cellular metabolism, acid stress, oxidative stress and RpoS-dependent stress responses. Differentially expressed virulence systems included the locus of enterocyte effacement (LEE) encoding the type-3 secretion system (T3SS) and its effectors, as well as BP-933W prophage encoded Shiga toxin 2 genes. Moreover, PhoB directly regulated LEE and stx2 gene expression through binding to specific Pho boxes. However, in Pi-rich medium, constitutive activation of the Pho regulon decreased LEE gene expression and reduced adherence to HeLa cells. Together, these findings reveal that EHEC has evolved a sophisticated response to Pi limitation involving multiple biochemical strategies that contribute to its ability to respond to variations in environmental Pi and to coordinating the virulence response.

Figure 1. Global analysis of differentially expressed genes in response to Pi starvation or phoB inactivation in EDL933 strain.

A. Classification of genes whose expression levels were altered in Pi-dependent and in PhoB-dependent manners. Left and right circles indicate the differentially expressed genes of wild-type and ΔphoB-mutant strains with expression levels that were altered over 2-fold under Pi limitation. Venn diagram: Group 1 includes 936 PhoB-independent genes that are differentially expressed under Pi limitation in the wild-type strain, but that did not change in the ΔphoB mutant. Group 3 includes 472 PhoB-dependent genes differentially expressed in the ΔphoB mutant but did not change under Pi-limitation in the wild-type strain. Group 2 included 131 PhoB-dependent Pi response genes that are differentially expressed under Pi limitation in the wild-type strain and between the wild-type and the ΔphoB strains. B. Functional classification of genes with altered expression in strain EDL933 grown in Pi-limited conditions compared to cells grown in Pi-rich conditions (Pi-dependent (white bars)) and EDL933 incubated in Pi-limited conditions compared to ΔphoB mutant cells grown in the Pi-limited conditions (PhoB-dependent (gray bars)).

Figure 2. Low Pi conditions and deletion of phoB gene increased LEE gene expression and T3SS secretion of EspB in EDL933 strain.

A. A heat map of the expression levels of LEE genes between wild-type strain grown in low or high Pi concentrations, and between wild-type and ΔphoB strain grown in low Pi condition. Expression values were determined from the variance analysis by the EB (Wright & Simon) algorithm and are represented colorimetrically, with red representing up regulation (ratio of +2.3) and blue representing downregulation (ratio of -1) on a log₂ scale. The data are represented as the means from three biological replicates. B. Expression level of LEE genes ler, sepZ, escV, tir and espB was analyzed by RT-qPCR in the wild-type, the ΔphoB mutant and its complement (ΔphoB+C) and the Δpst mutant grown in low or high Pi media as indicated. C. Western-blot using anti-EspB antibody against the supernatant of the wild-type, the ΔphoB mutant and its complement and the Δpst mutant grown in low or high Pi media as indicated. The ΔescN mutant grown in DMEM was used as a negative control. One μg/mL of Gro-El was added as a protein precipitation control. Asterisks represent the significant ANOVA p value (*<0.05, **<0.01, ***<0.001). ns: Not significant.

Figure 3. Effect of Pi and PhoB on stx2 gene expression and toxin production.

Low Pi and PhoB increased transcription of stx2 and its toxin production and of BP933W genes except for repressor gene cI that was repressed. A. Heat map of the expression levels of BP933W genes between wild-type EDL933 strain grown in low or high Pi conditions and between wild-type and ΔphoB strain grown in low Pi conditions. B. The expression levels of the BP933W genes cI, cro and stx2AB were analyzed by RT-qPCR in the wild-type strain and the ΔphoB mutant grown in low or high Pi media as indicated. The wild-type strain induced by mitomycin C (WT+MitC) was used as a positive control C. Fluorescence of the wild-type EDL933 strain carrying a chromosomal fusion reporting stx2 transcription level grown in Pi+ or Pi- conditions. D. The production of Stx2 measured by ELISA in extra-cellular protein (ECP) and whole cell protein (WCP) fractions of the EDL933 wild-type strain grown in Pi+ or Pi- conditions and in the ΔphoB mutant and its complemented derivative. Asterisks represent the significant ANOVA P value (*<0.05, **<0.01, ***<0.001).

Figure 4. Bio-informatic analyses of PhoB-binding sites.

Sequence logos determined from 12 putative PhoB-binding sites in EDL933 are indicated (upper panel). Potential consensus sequences identified in the promoter regions of LEE 1, LEE 2 operons and upstream stx2AB genes are shown with their statistical scores and genomic positions (lower panel). Pho box prediction probabilities were determined using the matrix frequencies of Table S4 that were uploaded into the Gibbs software algorithm. http://ccmbweb.ccv.brown.edu/gibbs/gibbs.html.

Figure 5. PhoB binds in vitro to LEE1, LEE2 and stx2 promoter regions.

A. Schematic representation of LEE and lambdoid prophage BP933W DNA regions. Arrows indicate the orientation of transcription. The blue lines/arrows indicate the probes used for EMSA assays. The square symbol indicates the predicted Pho box. B. Increasing amounts of GST-purified recombinant PhoB^CA were used in the EMSA assay to shift the 6-FAM labeled DNA probes amplified from LEE1 (−253 to +64 bp), LEE2 (−228 to +81 bp), LEE3 (−109 to −259 bp) and stx2AB (−402 to −3 bp) promoter regions.

Figure 6. Regulation of adhesion of EDL933 to human epithelial cells by the Pho regulon.

HeLa cells were infected with EDL933 and indicated mutants. After 6-type and ΔphoB strains (A and B). The adhesion decreased in the Δpst mutant (P <0.001) to levels similar to that of the ΔescN mutant which lacks functional TTSS (C and D). The attachment phenotype was restored in the trans complemented Δpst mutant (E). The number of adherent bacteria per HeLa cell was determined from 25 cells (F). Magnification 64 ×.