In Vivo Cleavage Map Illuminates the Central Role of RNase E in Coding and Non-coding RNA Pathways

Abstract

Understanding RNA processing and turnover requires knowledge of cleavages by major endoribonucleases within a living cell. We have employed TIER-seq (transiently inactivating an endoribonuclease followed by RNA-seq) to profile cleavage products of the essential endoribonuclease RNase E in Salmonella enterica. A dominating cleavage signature is the location of a uridine two nucleotides downstream in a single-stranded segment, which we rationalize structurally as a key recognition determinant that may favor RNase E catalysis. Our results suggest a prominent biogenesis pathway for bacterial regulatory small RNAs whereby RNase E acts together with the RNA chaperone Hfq to liberate stable 3' fragments from various precursor RNAs. Recapitulating this process in vitro, Hfq guides RNase E cleavage of a representative small-RNA precursor for interaction with a mRNA target. In vivo, the processing is required for target regulation. Our findings reveal a general maturation mechanism for a major class of post-transcriptional regulators.

Keywords: 3′ UTR; ArcZ; Hfq; RNA degradome; RNase E; RprA; TIER-seq; non-coding RNA; sRNA maturation; uridine ruler.

Graphical abstract

Figure 1

Global Mapping of Endogenous RNase E Cleavage Sites in Salmonella using TIER-Seq (A) Schema of the TIER-seq approach. Endogenous cleavage sites were identified by analyzing the 5′ ends of RNase E cleavage products (purple) in the WT and rne^TS strains at the non-permissive temperature (44°C). Total RNA from WT and rne^TS was converted to cDNAs and sequenced; the 5′ ends depleted in the rne^TS libraries at 44°C indicate the RNase E cleavage sites (e.g., purple U). (B and C) Global analysis of 5′ end profile at the permissive temperature 28°C (B) and non-permissive temperature 44°C (C). The plots show the read counts for every 5′ base in WT samples and the relative fold change compared to rne^TS samples. Candidate RNase E cleavage sites that show >3-fold depletion in rne^TS samples (p < 0.05, FDR < 0.05) are colored in red. (D) TIER-seq captures known RNase E cleavage sites with single-nucleotide resolution. TS indicates the rne^TS samples. R1 and R2 are two biological replicates. The major RNase E sites are marked by red arrowheads and bold lettering; secondary cleavage sites are indicated by open arrowheads. The ORF or mature RNAs are shadowed by gray boxes. See also Figures S1 and S2.

Figure 2

Systems-wide Analysis of Cleavage Sites Reveals a Consensus for RNase E (A) Classification of all RNase E cleavage sites. The proportion (%) of all sites mapped within a category is shown. See also Table S1. (B) The number of cleavage sites mapped per mRNA gene. (C) The distribution of RNase E cleavage frequencies in mRNAs (RPKM > 10). See also Figure S1. (D) Sequences at the RNase E sites are less structured. Minimal folding energy (MFE) was calculated for each 25 nt using a sliding window and was compared to randomly shuffled sequences. Median Z score is shown as a bold line; dotted lines indicate the upper and lower quartile. (E) Distribution of AU content at the RNase E cleavage sites. Dashed line indicates the cleavage site (+1 nt). (F) The RNase E consensus motif based on alignment of all mapped cleavage sites. Error bars indicate 95% confidence intervals. See also Figures S1 and S2.

Figure 3

RNase E Cleaves mRNAs to Produce 3′ UTR-Derived sRNAs (A and B) Distribution of RNase E cleavage sites in mRNAs relative to their start codon (A) or stop codon (B). The gray lines in the lower panel indicate the distribution of consensus motif based on genomic sequence. (C) RNase E and Hfq are required for the biogenesis of 3′ UTR-derived sRNAs. WT and Δhfq strains were grown at 37°C to an OD₆₀₀ of 2. The location of sRNAs (red arrows) and host genes are shown in the lower panel. Promoters (where available) and terminators are shown. The 5S rRNA served as loading control (Figure S2C). See also Figures S2, S3, and S4.

Figure 4

RNase E Cleaves Non-coding RNAs to Release 3′ Mature sRNAs (A–C) RNase E cleavage sites are identified in the ArcZ sRNA (A), RprA (B), and 3′ETS^leuZ (C). The major sites are marked by red arrowheads and bold lettering, whereas the minor sites are indicated by open arrowheads. See also Figures S5 and S6. (D–F) RNase E is required for the processing of ArcZ (D), RprA (E), and 3′ETS^leuZ (F). Open arrowheads indicate precursor fragments and filled arrowheads indicate processed mature species. ^∗ indicates longer precursors of polycistronic LeuZ-tRNA fragments; 5S loading controls, see Figure S2C. (G) The maturation of ArcZ is dependent on RNase E activity. Expression of the full-length ArcZ precursor (pre-ArcZ) was induced by L-arabinose.

Figure 5

RNase E Mediates the Maturation of ArcZ sRNA In Vitro and In Vivo (A) Alignment of ArcZ sequence. Conservation scores are plotted below the sequences, and the conserved seed is colored in green. (B) Reconstitution of ArcZ maturation in vitro. Full-length pre-ArcZ RNA was incubated with RNase E in the presence or absence of Hfq. RNA was analyzed by northern blotting with an oligo antisense to the mature ArcZ. The lower set shows mature ArcZ signals with longer exposure. (C) Mutation of RNase E cleavage site. Variants of ArcZ precursors were incubated with Hfq, and then subjected to RNase E cleavage. The lower set shows mature ArcZ signals with longer exposure. (D) Primer extension to map the RNase E cleavage sites in ArcZ in vitro. (E) Validation of RNase E motif in ArcZ in vivo. See also Figures S5 and S6.

Figure 6

Maturation of ArcZ sRNA Is Essential for Target Regulation (A) Established base pair interactions between ArcZ and tpx mRNA (Papenfort et al., 2009). The major cleavage site in ArcZ is indicated. (B) Western blot detection of GFP levels. GFP was fused with tpx 5′ UTR; the introduced mutations are shown in (A). “WT” refers to WT full-length ArcZ, “mat” refers to mature ArcZ, and “GG” refers to the GAUGG variant of ArcZ. GroEL served as loading control. (C) Direct interaction of tpx with mature ArcZ by EMSA. Radiolabeled mature ArcZ was incubated with increasing concentration of tpx mRNA in the presence of Hfq (40 nM). The gel was resized; see Figure S6. (D) Direct interaction of tpx with pre-ArcZ by EMSA. (E) Mature ArcZ was co-shifted with tpx mRNA. Radiolabeled tpx mRNA was incubated with increasing concentration of pre-ArcZ or mature ArcZ (0, 6, 25, 100, 400, and 2,000 nM) in the presence of 40 nM Hfq. See also Figure S6.

Figure 7

Mechanism of RNase E Cleavage and an Alternative sRNA Biogenesis Pathway (A) RNase E cleavage constitutes a major sRNA biogenesis pathway in bacteria. (B) Proposed model for the +2 uridine ruler-and-cut mechanism of specific RNase E cleavage. The scissile phosphate is attacked hydrolytically by a water molecule (not shown) that is coordinated by the magnesium ion bound by the carboxylates of D₃₄₆ and D₃₀₃. Stacking interactions (between F₆₇ and K₁₁₂) and hydrogen bonding (with the K₁₁₂GAA loop) with the base at position +2 favor uridine at this position. The interactions are predicted to help orient the phosphate backbone into a geometry that would facilitate cleavage at the scissile phosphate. See also Figure S7.