Interplay between the RNA Chaperone Hfq, Small RNAs and Transcriptional Regulator OmpR Modulates Iron Homeostasis in the Enteropathogen Yersinia enterocolitica

Abstract

Iron is both essential for and potentially toxic to bacteria, so the precise maintenance of iron homeostasis is necessary for their survival. Our previous study indicated that in the human enteropathogen Yersinia enterocolitica, the regulator OmpR directly controls the transcription of the fur, fecA and fepA genes, encoding the ferric uptake repressor and two transporters of ferric siderophores, respectively. This study was undertaken to determine the significance of the RNA chaperone Hfq and the small RNAs OmrA and RyhB1 in the post-transcriptional control of the expression of these OmpR targets. We show that Hfq silences fur, fecA and fepA expression post-transcriptionally and negatively affects the production of FLAG-tagged Fur, FecA and FepA proteins. In addition, we found that the fur gene is under the negative control of the sRNA RyhB1, while fecA and fepA are negatively regulated by the sRNA OmrA. Finally, our data revealed that the role of OmrA results from a complex interplay of transcriptional and post-transcriptional effects in the feedback circuit between the regulator OmpR and the sRNA OmrA. Thus, the expression of fur, fecA and fepA is subject to complex transcriptional and post-transcriptional regulation in order to maintain iron homeostasis in Y. enterocolitica.

Keywords: Fur repressor; RNA chaperone Hfq; Yersinia enterocolitica; fecA; fepA; fur; iron homeostasis; regulator OmpR; sRNA OmrA; sRNA RyhB1.

Figure 1

Post-transcriptional regulation of fur expression. (A,B) Influence of Hfq on fur expression. Green fluorescence of strain Ye9N and its derivatives carrying a fur′-′gfp translational fusion (pFX-fur) was measured in cultures grown to the stationary phase in LB or LBD. The presented data are the mean relative fluorescence values normalized to the OD₆₀₀ of the culture (±standard deviation) from at least three independent experiments, each performed with three separate cultures per strain. Significance was calculated by one-way ANOVA (**** p < 0.0001, *** p < 0.001). (C) Influence of Hfq on the abundance of Fur-3×FLAG protein. Levels of Fur-3×FLAG were analyzed in Ye9Fflag cells carrying either the empty vector pBAD24Cm or the plasmid pBAD-Hfq, grown to the stationary phase in LB or LBD. Western blotting of cell lysates was performed with a monoclonal mouse anti-FLAG epitope antibody. The upper panel shows the immunoblot, and the lower panel shows part of the Bio-Rad Stain-Free gel as a loading control. The percentage values indicate the immunostained protein band intensities relative to Ye9Fflag/pBAD24Cm grown in LB or LBD. MW—color prestained protein marker. (D) Predicted base pairing between sRNAs RyhB1/RyhB2 and the fur transcript. The mRNA sequences are in bold, with the translation start codons marked in red. Nucleotides are numbered from the first base of the start codon. (E) Northern blot showing the level of ryhB1 mRNA in strains Ye9N (wt) and Ye9N/pBR-RyhB1 grown in LB or LBD. As a loading control, the level of 5S rRNA was examined. (F) The level of fur transcripts assessed by RT-qPCR in Ye9N (wt) and Ye9N/pBR-RyhB1 grown to early stationary phase in LBD. Relative fur transcript levels, normalized to the amount of 16S rRNA, are shown, taking the mRNA level in Ye9N/pBR1 as 1. The presented data are the mean values (±standard deviation) obtained from at least three independent experiments. Significance was calculated using Student’s unpaired t-test (**** p < 0.0001).

Figure 2

Influence of Hfq on fecA and fepA expression. (A,B) Green fluorescence of strain Ye9N and its derivatives carrying a fecA’-’gfp translational fusion (pFX-fecA) or fepA’-’gfp translational fusion (pFX-fepA) measured in cultures grown to stationary phase in LB or LBD medium. The presented data are the mean relative fluorescence values normalized to the OD₆₀₀ of the culture (±standard deviation) from at least three independent experiments, each performed with three separate cultures per strain. Significance was calculated by one-way ANOVA (**** p < 0.0001, ** p < 0.01). (C) Influence of Hfq on FecA and FepA abundance in Y. enterocolitica strains. Levels of FecA-3×FLAG (left panel) and FepA-3×FLAG (right panel) in strains Ye9N and Ye9hfq carrying plasmid pBAD-Hfq or empty vector pBAD24Cm grown to stationary phase in LBD. The percentage values indicate the immunostained protein band intensities relative to the Ye9hfq/pBAD24Cm. Western blotting of cell lysates was performed with a monoclonal mouse anti-FLAG epitope antibody. Part of the Bio-Rad Stain-Free gel is shown as a loading control. These results are representative of two independent experiments. MW—3-color prestained protein marker.

Figure 3

Nucleotide sequence conservation of OmrA and OmrB sRNAs from Gammaproteobacteria. (A) Multiple sequence alignment of omrA and omrB genes from 12 species (13 strains) of Gammaproteobacteria (Escherichia coli str. K-12 substr. MG1655 (E.col.; NCBI taxid: 511145); Shigella flexneri serotype 2a str. 301 (S.fle.; NCBI taxid: 198214); Salmonella enterica subsp. enterica serovar Typhimurium str. 14028S (S.ent.; NCBI taxid: 588858); Klebsiella pneumoniae subsp. pneumoniae HS11286 (K.pne.; NCBI taxid: 1125630); Yersinia enterocolitica subsp. palearctica Ye9N bioserotype 2/O:9 (Y.ent. Ye9N; a shotgun genome sequence: Accession number NZ_JAALCX010000053.1); Yersinia enterocolitica subsp. enterocolitica 8081 (Y.ent. 8081; NCBI taxid: 393305); Yersinia intermedia (Y.int.; NCBI taxid: 631); Yersinia ruckeri ATCC 29473 (Y.ruc.; NCBI taxid: 527005); Yersinia pseudotuberculosis IP 32953 (Y.pse.; NCBI taxid: 273123); Yersinia pestis str. A1122 (Y.pes.; NCBI taxid: 1035377); Serratia marcescens strain KS10 (S.mar.; NCBI taxid: 615); Erwinia carotovora subsp. atroseptica SCRI1043 (E.car.; NCBI taxid: 218491); and Dickeya dadantii 3937 (D.dad.; NCBI taxid: 198628)). The alignment was performed using the T-Coffee tool and visualized in Jalview version 2.8.2 [79,80]. The nucleotides are colored by percentage identity; the most highly conserved nucleotides have the darkest color (dark blue signifies more than 80% nucleotide identity; medium blue represents more than 60% identity; light blue corresponds to more than 40% identity, and no color indicates identity below 40%). The level of identity is shown by a histogram at the bottom. The coding region lengths are given. (B) Heatmap indicating the percentage identity of omrA/omrB genes shared by two species. The identity matrix was calculated based on comparisons using the T-Coffee tool. The yellow–blue gradient bar represents the identity percentage scale. (C) Phylogenetic tree of omrA and omrB sequences in Gammaproteobacteria. The diagram was created with Jalview version 2.8.2 [80] based on matrices of average distances for the selected species.

Figure 4

The Y. enterocolitica fecA and fepA genes are negative regulatory targets of sRNA OmrA. (A) Predicted base pairing between sRNA OmrA and the fecA and fepA transcripts. The mRNA sequences are in bold, with the translation start codons marked in red. Nucleotides are numbered from the first base of the start codon. (B) The levels of fecA and fepA transcripts assessed by RT-qPCR in Ye9N (wt) and Ye9omrA (∆omrA). The analysis was performed using RNA prepared from cells grown to the early stationary phase in LBD (left panel) and LBD supplemented with 20% sucrose (right panel). Relative fecA and fepA transcript levels, normalized to the amount of 16S rRNA, are shown, taking the mRNA level in Ye9N as 1. The presented data are the mean values (±standard deviation) obtained from at least three independent experiments. Significance was calculated using Student’s unpaired t-test (** p < 0.01, * p < 0.05).

Figure 5

Influence of OmpR on omrA promoter function. (A) Putative OmpR-binding site (OmpR box) identified in the omrA regulatory region of Y. enterocolitica strain Ye9N (boxed). The 5′ end of the omrA transcript is shaded blue. Below, the OmpR boxes identified in the promoters of omrA and omrB in E. coli [24] are aligned with the OmpR box recognized in the Ye9N omrA promoter (% identity is shown). Nucleotides crucial for OmpR interaction are shown in bold. Identical nucleotides are shaded gray. (B) Analysis of omrA expression in the wild type (Ye9N), ∆ompR mutant (AR4) and strain AR4 complemented with pHR4, using a P_omrA::gfp transcriptional fusion. Strains were cultivated in LB at 26 °C or 37 °C to stationary phase. The data represent mean fluorescence activity values normalized to the OD₆₀₀ of the culture (±standard deviation) from a representative experiment performed using at least triplicate cultures of each strain. Significance was calculated using Student’s unpaired t-test (**** p < 0.0001, *** p < 0.001). (C) Fluorescence intensity of the wild-type strain Ye9N harboring the P_omrA::gfp fusion plasmid grown in LB to exponential phase then exposed to high osmolarity or different pH conditions for 1 h at 26 °C or 37 °C. Fold change relative to strain Ye9N grown in LB is shown. Significance was calculated using Student’s unpaired t-test (**** p < 0.0001, ** p < 0.01, * p < 0.05). (D) Northern blot showing the levels of omrA mRNA in the wild-type strain Ye9N, ∆ompR mutant (strain AR4) and complemented strain AR4/pOmpR grown to exponential phase in LB with or without the addition of procaine (10 mM). As a loading control, the level of 5S rRNA was examined. (E) EMSA analysis to study the ability of His-tagged OmpR-P to bind to the promoter region of omrA in vitro. The 250 bp omrA promoter fragment was incubated without protein (lane 2) or with 12.4 μM (lane 3), 24.9 μM (lane 4), 37.3 μM (lane 5) or 49.7 μM (lane 6) His-tagged OmpR-P. Lane 1, MassRuler DNA Ladder Low Range. A 304 bp 16S rDNA fragment was included in the reaction mixtures as a competitor and negative control. Unbound DNA and protein/DNA complexes are indicated.

Figure 6

OmrA inhibits ompR expression at the post-transcriptional level. (A) Predicted base pairing between sRNA OmrA and the ompR mRNA of Y. enterocolitica Ye9N. mRNA sequences are shown in bold. The AUG translation start codon and putative Shine–Dalgarno sequences are in bold and marked in red and green, respectively. (B) Northern blot showing the level of omrA mRNA in the wild-type strain Ye9N and Ye9N carrying pBR-OmrA. Strains were grown in LB to stationary phase. As a loading control, the level of 5S rRNA was examined. (C) Analysis of ompR expression in strain Ye9N carrying the ompR’-’gfp translational fusion (pFX-ompR) after growth in LB to stationary phase in the presence or absence of plasmid pBR-OmrA. The data represent mean fluorescence activity values normalized to the OD₆₀₀ of the culture (±standard deviation) from at least three independent experiments, each with three separate cultures per strain. Significance was calculated using Student’s unpaired t-test (*** p < 0.001). (D) The level of ompR transcripts assessed by RT-qPCR in Ye9N, Ye9N/pBR1, Ye9N/pBR-OmrA and the ΔomrA mutant. The analysis was performed using RNA prepared from cells grown to early stationary phase in LB + 20% sucrose. Relative ompR transcript levels, normalized to the amount of 16S rRNA, are shown, taking the mRNA level in Ye9N as 1. The presented data are the mean values (±standard deviation) obtained from at least three independent experiments. Significance was calculated using one-way ANOVA (** p < 0.01, * p < 0.05).

Figure 7

Levels of fecA and fepA transcripts in the wild-type (Ye9N) and ∆ompR strains (AR4) transformed with pBR-OmrA or the empty vector pBR1. The abundance of fecA and fepA mRNAs was assessed by RT-qPCR after the growth of these strains in LBD. Relative fecA and fepA transcript levels, normalized to the amount of 16S rRNA, are shown, taking the mRNA level in Ye9N/pBR1 as 1. The presented data are the mean values (±standard deviation) from at least two independent experiments. Significance was calculated using one-way ANOVA (**** p < 0.0001).

Figure 8

Model for the interplay between regulatory proteins OmpR, Fur, Hfq and sRNAs RyhB1 and OmrA in the regulation of iron acquisition in Y. enterocolitica. The production of FecA and FepA, the OM transporters for ferric citrate and ferric enterobactin, is controlled by OmpR via direct regulation of the transcription of the fecA and fepA genes, plus fur, encoding the Fur protein, the major transcriptional repressor of fecA and fepA [22]. We hypothesize that the RNA chaperone Hfq acts in concert with different sRNAs to exert its post-transcriptional effects on the Fur repressor and transporters FecA and FepA. Hfq silences the expression of fur at the post-transcriptional level in stationary-phase cells grown in iron-depleted medium, with the participation of putative cofactors (sRNAs RyhB1/RyhB2?). Production of FecA and FepA is negatively controlled by Hfq at the post-transcriptional level. The silencing of fecA and fepA depends on the sRNA OmrA, and Hfq is important in this regulation. Finally, the production of the sRNA OmrA is controlled by OmpR through its direct positive effect on omrA transcription. In turn, OmrA post-transcriptionally controls the expression of OmpR; thus, a coherent feed-forward regulatory loop exists. Dashed blue lines—previously reported regulation of transcription; red and blue lines—regulatory pathways identified in this study (post-transcriptional and transcriptional regulation, respectively). Lines ending with a perpendicular bar indicate a negative interaction; lines ending with an arrowhead indicate a positive effect.