Coordinated regulation of acid resistance in Escherichia coli

Abstract

Background: Enteric Escherichia coli survives the highly acidic environment of the stomach through multiple acid resistance (AR) mechanisms. The most effective system, AR2, decarboxylates externally-derived glutamate to remove cytoplasmic protons and excrete GABA. The first described system, AR1, does not require an external amino acid. Its mechanism has not been determined. The regulation of the multiple AR systems and their coordination with broader cellular metabolism has not been fully explored.

Results: We utilized a combination of ChIP-Seq and gene expression analysis to experimentally map the regulatory interactions of four TFs: nac, ntrC, ompR, and csiR. Our data identified all previously in vivo confirmed direct interactions and revealed several others previously inferred from gene expression data. Our data demonstrate that nac and csiR directly modulate AR, and leads to a regulatory network model in which all four TFs participate in coordinating acid resistance, glutamate metabolism, and nitrogen metabolism. This model predicts a novel mechanism for AR1 by which the decarboxylation enzymes of AR2 are used with internally derived glutamate. This hypothesis makes several testable predictions that we confirmed experimentally.

Conclusions: Our data suggest that the regulatory network underlying AR is complex and deeply interconnected with the regulation of GABA and glutamate metabolism, nitrogen metabolism. These connections underlie and experimentally validated model of AR1 in which the decarboxylation enzymes of AR2 are used with internally derived glutamate.

Keywords: Acid resistance; Regulatory network modeling; Systems biology/ChIP-Seq.

Fig. 1

Mapping E. coli transcriptional regulatory interactions using ChIP-Seq. Examples of identified binding sites for csiR, nac, ntrC, and ompR. Each panel plots the total read coverage (blue), forward read coverage (green), and reverse read coverage (red). The maximum coverage for each plot is given by the number on the y-axis in units of coverage normalized to mean coverage. Multiple biological replicate experiments are shown for 3 TFs as noted on leftmost y-axes. ChIP-Seq coverage plots are shown for 8 separate genomic regions. The start location of each region is provided at the bottom left x-axes. The tick marks on the bottom x-axes are spaced 500 bp apart. Different regions are plotted at different scales for clarity. Previously described binding sites from EcoCyc are shown as black ticks below the coverage plot in each panel

Fig. 2

ChIP-Seq mapping and transcriptomics reveal regulatory links between AR systems and cellular metabolic pathways. Map of selected direct binding sites potentially associated with AR. Novel TF binding is displayed as colored dashed lines. Novel regulatory links confirmed with gene expression data are shown as solid colored lines. Black lines signify previously reported known binding and regulation. Circle terminators indicate unconfirmed or indeterminate regulatory effect

Fig. 3

Validation of a Proposed Mechanism for AR1. We hypothesized that AR1 may be mediated by the AR2 machinery using an internal source of glutamate. Our regulatory network implicates both nac and csiR in this process. We tested this hypothesis by examining the phenotype of several deletion mutants in acid stress assays using published protocols for inducing AR1 or AR2, along with positive and negative controls (Castanie-Cornet et al. [11]; Lin et al [4]). Acid stress assays consisted of overnight culture, acid challenge at pH 2.5 for 2 h, followed by plating, overnight incubation, and colony counting (Methods). a Example plates for one experiment for selected mutants comparing AR1 conditions to AR2 conditions. b Summary of colony counts averages for all mutants across all experiments for AR1, AR2, and for two non-acidic control growth conditions (for which strains were plated directly after overnight incubation without acid challenge) for 3 replicates (n = 3). Colony counts provided to allow comparison to control WT data. Resulting counts were tested at a significance level of α = 0.05 (* p-value < 0.05). Plots of % survival for AR1 and AR2 are provide in Additional file 1: Figure S6 c RT-PCR of gadE in WT, ΔcsiR, and Δnac from colonies recovered after acid challenge following AR2 induction (n = 3 for all). d AR Rescue of KO strains via induction of gadE showing the summary of colony counts averages for WT, ΔcsiR, Δnac and ΔgadE with gadE induced in AR1 and AR2 conditions for 3 replicate experiments (n =3). Numbers on the x-axis above strain names indicate amount of aTc added during AR challenge in ng/μL. Resulting counts were tested at a significance level of α = 0.05 (* p-value < 0.05)