The RNA repair proteins RtcAB regulate transcription activator RtcR via its CRISPR-associated Rossmann fold domain

Figures

• 

  Figure 1. Schematic representation of the Rtc system (A) The rtcBA operon is under the control of the sigma 54 dependent rtcBA promoter whose activator is encoded by the oppositely oriented rtcR gene. RtcA is an RNA 3′-terminal phosphate cyclase that converts 2′-P and 3′-P ends into 2′,3′-cyclic P ends. RtcB is an RNA ligase that ligates 3′-P and 2′,3′-cyclic P ends to 5′-OH ends. (B) The sigma 54 transcriptional activator RtcR has three functional domains including a N-terminal, regulatory CARF domain. Under normal conditions, the CARF domain represses the AAA + domain so that no transcription of rtcBA takes place from the rtcBA promoter (left). Conversely, RtcR mutants lacking the CARF domain are constitutively active (middle). Under stress, an unknown signal ligand interferes with the CARF domain mediated repression of the AAA + domain and the rtcBA operon is expressed (right).

• 

  Figure 2. The RtcRAB proteins interact via the RtcR CARF domain (A and B) Rtc proteins were overexpressed and purified (Figure S2A), and their physical interactions were investigated using (A) MST and (B) gel filtration chromatography. (A) Changes in thermophoresis (y axis of graph) of fluorescently labeled RtcRWT and RtcRΔCARF was used to quantify their binding to RtcA or RtcB in titration experiments. RtcRWT, but not RtcRΔCARF, interacts in vitro with RtcA (left) and RtcB (right), as shown by MST. The dissociation constant (Kd) of RtcA and RtcB with RtcR is in the μM range. (B) Alterations in the fractionation and apparent molecular weight of RtcRWT and RtcRΔCARF were used to show their complex formation with RtcA or RtcB. The peak of MBP-RtcR, but not MBP-RtcRΔCARF, potentially forming a hexamer, shifts towards a higher molecular weight in the presence of His-RtcA and His-RtcB, as shown by gel filtration chromatography. Co-localization of MBP-RtcR with His-RtcA and His-RtcB was confirmed with SDS-PAGE and immunoblotting. The proteins were present at equimolar concentrations (20 μM). (C) Pairwise combinations of N-terminal and C-terminal T18/T25 fusions between full length or truncated…

• 

  Figure 3. The RtcR CARF domain is repressive, and RtcR is not activated by RtcAB or oligoadenylates in vitro The transcriptional activity of purified RtcR and its CARF deletion variant was measured in an in vitro transcription assay leading to synthesis of radioactive RNA UpGpGpG following extension of UpG by addition of radioactive GTP from the super-coiled PrtcBA-nifH hybrid promoter template. (A–C) The RtcRΔCARF variant, in the presence of in trans CARF domain, (B) full length RtcR and cyclic tetra/hexa (4/6) adenylates (cOA), or (C) full length RtcR, RtcA and/or RtcB and cOA. The UpGpGpG was separated from unincorporated GTP by electrophoresis on a 20% (w/v) urea-PAG. In panels (A–C), the %A is percentage of transcription activity versus that seen with the RtcRΔCARF variant.

• 

  Figure 4. The Rtc system is induced in the Hpx- strain The transcriptional activation of the Rtc system in the Hpx- strain was assessed using β-galactosidase assays with cells harbouring a plasmid-encoded PrtcBA-lacZ reporter and RT-qPCR. In parallel, the growth of the Hpx- stain was monitored. (A) The rtcBA promoter activity is induced in LB broth, LB broth supplemented with extra 5 g/L NaCl, LB broth supplemented with 50 μM Fe-citrate and nutrient broth, after 24 h under aerobic but not anaerobic conditions (N = 5). (B) The rtcBA promoter activity is induced in the Hpx- strain because of gene deletion ΔahpC; RtcR, RtcB and RtcA are required for rtcBA promoter activity in the gene deletion mutants ΔahpC (N = 4). (C and D) The rtcBA promoter activity is induced and (D) the rtc mRNA levels are increased, as shown by β-galactosidase reporter assays and RT-qPCR, respectively (N = 3). (E) The presence of the PrtcBA-lacZ reporter plasmid leads to inhibited growth of the Hpx- compared to the wild-type (N = 3). (F) In the absence of the PrtcBA-lacZ reporter plasmid, the inhibitory growth effect is not as evident (N = 3). (G) Proposed model of the molecular events occurring in the Hpx- cells in…

• 

  Figure 5. Transcriptome profiling of E. coli cells under Rtc-inducing conditions Transcriptome profiling of E. coli wild-type cells, cells lacking gor, mazF or srmB, and Hpx-[RtcON] and Hpx-[RtcOFF] cells was performed by NGS and followed by differential gene expression analysis. Transcriptome profiling of E. coli cells lacking rtcA or rtcB was described previously by Engl et al. (2016).15 (A) Venn diagram of the genes differentially expressed in cells lacking gor or mazF compared to the wild-type and associated GO terms (Tables S7 and S8). (B) Principal component analysis of the transcriptional profiling of the wild-type, Hpx-[RtcON] and Hpx-[RtcOFF] strains. (C) Venn diagram of the genes differentially expressed in Hpx-[RtcON[ and Hpx-[RtcOFF] as compared to wild-type cells at 8 and 24 hpi, and associated GO terms and KEGG pathways. (D) Heat maps of 5S (left; log2fold expression ±5), 23S (left; log2fold expression ±7) and 16S (left; log2fold expression ±6) ribosomal (r) RNA abundance levels in a range of mutants, as shown by NGS; rows and columns have been grouped based on a hierarchical clustering algorithm. Limma/DESeq2 ∗ adjusted p-value < 0.05; ∗∗ adjusted p-value < 0.01; ∗∗∗∗…

• 

  Figure 6. tRNAs are cleaved under oxidative stress Damaged RNAs in transcriptomes of E. coli Hpx-[RtcON] and Hpx-[RtcOFF] cells were identified by pairwise calculation of the Pearson’s Correlation Coefficient (PCC) for each individual gene. (A) Globally, more transcripts are damaged in Hpx-[RtcOFF] than Hpx-[RtcON] cells compared to wild-type cells 8 h post-inoculation, as illustrated by plotting PCC values of all genes in descending order from high to low. (B) More tRNAs are damaged in Hpx-[RtcOFF] (top right) than Hpx-[RtcON] (top left) compared to wild-type cells, as illustrated by plotting PCC values of pairwise comparisons for selected tRNAs. Distribution of all reads mapped to gltU in Hpx-[RtcOFF] and wild-type cells reveals no obvious cleavage sites (bottom).

• 

  Figure 7. RtcB-mediated repair targets the tmRNA ssrA under oxidative stress Targets of the RtcB ligase, including the tmRNA ssrA, were identified by RtcB mediated ligation using an adapter and followed by NGS. (A) Number of genes whose RNA is an RtcB ligase target (left) and events of RtcB-mediated ligations (right). (B) Ligation events for ssrA in wild-type (green arrows) and Hpx- cells (red arrows). Different colours indicate distinct ssrA structural domains, including the tRNA-like domain (TLD), the mRNA-like domain (MLD), helices 2 (H2) and 5 (H5), and pseudoknots 1 (PK1), 2 (PK2), 3 (PK3) and 4 (PK4). The nucleotides of the open reading frame (ORF) are shown in a larger font, whereas the resume and stop codons are highlighted. Ligation events were significantly more abundant in PK3 (p-value < 0.01) for wild-type cells, and in both PK3 (p-value < 0.01) and PK4 (p-value < 0.0001) for Hpx- cells.

• 

  Figure 8. RtcB increases translation elongation speed in vivo The impact of RtcB on the functioning of the translation apparatus was measured in vivo using β-galactosidase based assays and growth assays. RtcB (A) increases translation elongation speed, (B) has no effect on translation fidelity, (C and D) increases cell growth and (D) increases cell growth rate in early exponential phase. Data are shown as mean and error bars represent standard deviation from the mean of triplicates. ns = non-significant (p-value > 0.05).

• 

  Figure 9. Regulation of Rtc RNA repair system expression and function Under conditions that activate of the Rtc RNA repair system that helps maintain translation, RtcA and RtcB interact with the RtcR CARF domain together with an unknown ligand, leading to derepression of the RtcR AAA + domain by the RtcR CARF domain. We propose that the complex has dual functionality: transcriptional activation of rtcBA by RtcR and RNA repair by RtcA and RtcB. The RtcR CARF domain likely does not use cOAs as activating ligands (this work), rather it binds particular linear RNA sequences with a 2’-3’ cyclic P end (this work and Hughes et al., 202018).