The Post-Transcriptional Regulatory Protein CsrA Amplifies Its Targetome through Direct Interactions with Stress-Response Regulatory Hubs: The EvgA and AcnA Cases

Abstract

Global rewiring of bacterial gene expressions in response to environmental cues is mediated by regulatory proteins such as the CsrA global regulator from E. coli. Several direct mRNA and sRNA targets of this protein have been identified; however, high-throughput studies suggest an expanded RNA targetome for this protein. In this work, we demonstrate that CsrA can extend its network by directly binding and regulating the evgA and acnA transcripts, encoding for regulatory proteins. CsrA represses EvgA and AcnA expression and disrupting the CsrA binding sites of evgA and acnA, results in broader gene expression changes to stress response networks. Specifically, altering CsrA-evgA binding impacts the genes related to acidic stress adaptation, and disrupting the CsrA-acnA interaction affects the genes involved in metal-induced oxidative stress responses. We show that these interactions are biologically relevant, as evidenced by the improved tolerance of evgA and acnA genomic mutants depleted of CsrA binding sites when challenged with acid and metal ions, respectively. We conclude that EvgA and AcnA are intermediate regulatory hubs through which CsrA can expand its regulatory role. The indirect CsrA regulation of gene networks coordinated by EvgA and AcnA likely contributes to optimizing cellular resources to promote exponential growth in the absence of stress.

Keywords: Csr network; CsrA; RNA–protein interactions; bacterial stress responses; post-transcriptional regulation.

Figure 1

CsrA directly interacts with the evgA and acnA transcripts in vitro. 0.5 nM of P³²-radiolabeled evgA and acnA were individually incubated with increasing concentrations of purified CsrA. All binding assays are performed in excess of yeast total RNA to inhibit the non-specific association of CsrA with labeled mRNA. (A) Electrophoretic mobility shift analysis (EMSA) of CsrA-binding to the evgA leader sequence. K_D was estimated from the fitted binding curve shown right. The standard deviation and 95% confidence intervals were determined from the non-linear fit of the individual gel measurements for each CsrA concentration. (B) EMSA of CsrA-binding to acnA and its respective fitted binding curve. Lanes 5 and 6, corresponding to 88 and 75 nM CsrA, respectively, were loaded in reverse order of the concentration gradient. These lanes are denoted with a bracket “{” for clarity. Supershiftted complexes that form at higher CsrA concentrations are indicated with an (*).

Figure 2

High-affinity GGA motifs mediate the CsrA-evgA interaction. (A) Secondary structure of the evgA leader sequence (5′ UTR + first 100 nt of coding sequence). The predicted binding sites from Leistra et al. (2018) [23] are shown in blue (PBS-I & PBS-II). PBS-I contains a high-affinity GGA motif. Additional GGA motifs that were considered for analysis are colored pink (GGA2, GGA3, and GGA4). The start codon (labeled “SC” in gray) and the coding sequence nucleotides are outlined in dark gray. Mutations introduced to test each individual binding site are shown in red. (B) Binding curves were generated via EMSAs for mutant versions of evgA to assess the contribution of each site to CsrA-binding, with (C) estimated K_D values determined from the binding curves. In this figure, “All pred. sites mutant” refers to a mutant of both PBS-I and PBS-II) and “no GGAs mutant” refers to a mutant of all GGA motifs.

Figure 3

One high-affinity GGA motif and a degenerate ANGGN site mediate the CsrA-acnA interaction. (A) Secondary structure of the acnA leader sequence (5′ UTR + first 100 nt of coding sequence). The predicted binding sites from Leistra et al. (2018) [23] are shown in blue (PBS-I, PBS-II, and PBS-III). PBS-I and PBS-II contain high-affinity GGA motifs. The start codon (labeled “SC” in gray) and the coding sequence nucleotides are outlined in dark gray. (B) Binding curves were generated via EMSAs for mutant versions of acnA to assess their contribution of CsrA-binding, with (C) estimated K_D values determined from the binding curves.

Figure 4

CsrA mediates evgA repression. (A) Diagram of the evgA-gfp in vivo reporter system. The first plasmid has CsrA expressed under the control of an IPTG-inducible promoter, while the second plasmid contains a constitutively expressed evgA leader sequence fused in-frame to gfp. (B) The secondary structure of the evgA leader with the predicted binding sites (blue) and additional putative GGA motifs (pink) is shown for reference. (C) Fluorescence ratios were calculated by dividing evgA-gfp fluorescence in the presence of CsrA by the fluorescence in the absence of CsrA. The results are representative of three independent biological replicates. Statistically significant values (p-value ≤ 0.05) are indicated by an asterisk (*) and were determined by comparing the negative control values to those of each mRNA variant tested using an unpaired t-test. (D) In vitro transcription–translation reactions were performed using the PURExpress kit with evgA wild-type and evgA no GGAs mutant translational fusions expressed from a T7 promoter. Increasing concentrations of purified CsrA were added prior to the start of each reaction. Values are shown for at least five replicates collected across three independent experiments. Statistically significant values are denoted with an asterisk (*) and indicate differences in fluorescence upon the addition of the respective CsrA concentration relative to the fluorescence when no CsrA was added to the reaction (0 nM). Figure created with BioRender.com.

Figure 5

CsrA-mediated repression of acnA. (A) Diagram of the acnA-gfp in vivo reporter system. One plasmid contains CsrA expressed under the control of an IPTG-inducible promoter, while the second plasmid contains a constitutively expressed acnA leader sequence fused in-frame to gfp. (B) The secondary structure of the acnA leader with predicted binding sites (blue) considered for our mutational analysis is shown for reference. (C) Fluorescence ratios were calculated by dividing the acnA-gfp fluorescence in the presence of CsrA by the fluorescence in the absence of CsrA. The results are representative of three independent biological replicates. Statistically significant values (p-value ≤ 0.05) are indicated by an asterisk (*) and were determined by comparing the negative control values to those of each mRNA variant tested using an unpaired t-test. (D) In vitro transcription–translation reactions were performed using the PURExpress kit with acnA wild-type and acnA double mutant of PBS-I and PBS-III translational fusions expressed from a T7 promoter. Increasing concentrations of purified CsrA were added prior to the start of each reaction. Values are shown for at least five replicates collected across three independent experiments. Statistically significant values are denoted with an asterisk (*) and indicate differences in fluorescence upon the addition of the respective CsrA concentration relative to the fluorescence when no CsrA was added to the reaction (0 nM). Figure created with BioRender.com.

Figure 6

The expression of specialized gene clusters is influenced by the CsrA-evgA interaction. (A) Number of differentially expressed genes previously associated with CsrA. (B) Volcano plot of the differentially expressed genes in the evgA no GGAs genomic mutant relative to wild-type E. coli. Genes have colored outlines based on their cluster association in the gene network shown below (green: polyamine transport, purple: formate oxidation, and red: acid resistance). (C) EnrichGO analysis results showing the top 5 most upregulated and downregulated genes in the evgA genomic mutant at Early Exponential. (D) Network representation of the differentially expressed genes upon disruption of the CsrA-evgA interaction. Node size represents the number of growth conditions in which a gene was differentially expressed. Outlined colors denote genes clustered together in our network analysis. Genes with similar functions based on their GO annotations are filled with the same color.

Figure 7

CsrA indirectly coordinates acid stress responses through its interaction with evgA. Schematic of known EvgA-regulated acidic stress-related genes. Solid lines denote known interactions. Light gray dotted lines indicate genes regulated by EvgA for which direct interactions are not known. Purple dotted lines indicate genes that are known to be functionally regulated by CsrA. Created with BioRender.com.

Figure 8

The CsrA-acnA interaction affects specialized gene clusters. (A) Number of differentially expressed genes previously associated with CsrA. (B) Volcano plot of the differentially expressed genes in the acnA genomic mutant (with mutated CsrA binding sites) relative to wild-type E. coli. Genes have colored outlines based on their cluster association in the gene network shown below (green: metal ion stress and DNA damage; blue: putrescine transport; purple: poxB and tktB cluster; red: iron–sulfur cluster; and yellow: toxin–antitoxin). (C) EnrichGO analysis results show the top 5 most upregulated and downregulated genes in the acnA genomic mutant at Early Exponential. (D) Network representation of the differentially expressed genes upon disruption of the CsrA-acnA interaction. Node size represents the number of growth conditions in which a gene was differentially expressed. Outlined colors denote genes clustered together in our network analysis. Genes with similar functions based on their GO annotations are filled with the same color.

Figure 9

The CsrA-acnA interaction coordinates response systems for metal-induced oxidative stress. This simplified diagram illustrates the different pathways affected by CsrA through its interaction with acnA based on the observed differential expression of genes upon breaking the CsrA- acnA interaction. Solid lines denote known interactions. Light gray dotted lines indicate genes regulated by AcnA (inferred from RNA-seq) for which direct interactions are not known. Purple dotted lines indicate genes that are known to be functionally regulated by CsrA. Created with BioRender.com.

Figure 10

The interaction between CsrA and evgA impacts tolerance to acidic stress in E. coli. (A) Percentage of cell survival after a 2 h challenge at different pH values. The percentage was calculated by dividing the number of CFU/mL that grew post-stress treatment by the number of CFU/mL in an unstressed control. Error bars indicate percentage variations between independent biological triplicates. Asterisks denote statistically significant (p-value ≤ 0.05) differences between the percent survival of the evgA mutant (referred to as evgA genomic no GGAs mutant in the table) and that of wild-type E. coli. Significance was determined by contrasting the percent survival of these strains using an unpaired t-test. (B) Serial dilutions of the cultures post-stress were spot-plated to observe differences in acid stress tolerance. Images were collected for independent biological duplicates for each strain. Here, evgA mutant refers to a genomic mutant of all GGA motifs present in the evgA leader.

Figure 11

The CsrA-acnA interaction influences bacterial tolerance to heavy metal stress. (A) Growth curves of wild-type E. coli and acnA mutant (genomic mutant of CsrA binding sites in the acnA sequence) strains in LB media only (left) and LB media supplemented with increasing concentrations of CuSO₄ (middle) and CuCl₂ (right). Shading denotes the standard deviation between biological duplicates. Cells were challenged with increasing concentrations of (B) CuSO₄ and (C) CuCl₂ for 8 h. Serial dilutions of the cultures post-stress were spot-plated to observe differences in metal stress tolerance. Images represent the observed cell growth for independent biological duplicates of each strain.