Identification of a conserved branched RNA structure that functions as a factor-independent terminator

Abstract

Anti-Q is a small RNA encoded on pCF10, an antibiotic resistance plasmid of Enterococcus faecalis, which negatively regulates conjugation of the plasmid. In this study we sought to understand how Anti-Q is generated relative to larger transcripts of the same operon. We found that Anti-Q folds into a branched structure that functions as a factor-independent terminator. In vitro and in vivo, termination is dependent on the integrity of this structure as well as the presence of a 3' polyuridine tract, but is not dependent on other downstream sequences. In vitro, terminated transcripts are released from RNA polymerase after synthesis. In vivo, a mutant with reduced termination efficiency demonstrated loss of tight control of conjugation function. A search of bacterial genomes revealed the presence of sequences that encode Anti-Q-like RNA structures. In vitro and in vivo experiments demonstrated that one of these functions as a terminator. This work reveals a previously unappreciated flexibility in the structure of factor-independent terminators and identifies a mechanism for generation of functional small RNAs; it should also inform annotation of bacterial sequence features, such as terminators, functional sRNAs, and operons.

Keywords: antisense; attenuation; cell–cell signaling; pheromone; transcription.

Fig. 1.

A map of the P_X-P_Q region. DNA is shown at the top as a dark line with features indicated. Transcriptional start sites are shown as bent arrows; terminators are shown as stem-loops; the prgQ and prgX ORFs are shown as open arrows labeled Q and X, respectively. The PrgX primary and secondary binding sites are shown as double and single triangles, respectively. Wavy lines above and below the map indicate RNAs with transcription initiated at P_X and P_Q, respectively. The region of the prgX mRNA targeted by RNase III is bracketed. The sequence of the Anti-Q/prgX sense strand is shown below. The two mapped transcriptional start sites are indicated (+1, +4). Open diamonds (◇) indicate the 3′ termini of Anti-Q in vivo as determined by S1 endonuclease mapping (+102) and 3′ RACE (+107). Closed circles (●) indicate in vitro termination points (+111, +119). Arrows indicate the right terminus of various pCF10 fragments tested for Anti-Q formation, as described in the text (+107, +117, +223). Brackets indicate sequence removed by defined deletions. The proposed polyuridine tract is underlined. IRSX and IRS1 denote stem-loop structures that function in transcription termination.

Fig. 2.

Anti-Q is generated by transcription termination. (A) Northern blot showing RNA harvested from OG1Sp carrying plasmids with P_X and Anti-Q sequences through +117 (pCJ15, lane 1) or +107 (pCJ14, lane 2) with an ectopic terminator fused downstream. The membrane was hybridized with a RNA probe specific for Anti-Q, E. faecalis 5S is shown as a loading control. (B) Northern blot of RNA harvested from either wild-type E. faecalis or an isogenic RNase III deletion mutant strain (OG1RFΔ3097) carrying pBK2 and probed for Anti-Q or 5S RNA as in A. (C) In vitro transcription reactions performed using the plasmids pCJ15 (lane 1) and pCJ14 (lane 2) as templates. (D) An in vitro transcription reaction performed using RNAP that was adsorbed to Nickel-agarose beads. The template was a linear PCR product containing P_X −167 to +400 with P_Q inactivated by a mutation to its −10 region. Lane 1, complete reaction; lane 2, RNAP associated products; lane 3, free products.

Fig. 3.

Disrupting Anti-Q structure lowers frequency of termination. (A) In vitro transcription reactions performed using plasmid templates that contained P_Q fused to Anti-Q with deletions to different regions of Anti-Q: lane 1, wild-type; lane 2, Δ34–65; lane 3, Δ66–95. The region in which wild-type and truncated Anti-Q sized products appear is indicated. The bands at the top of the gel are the loading wells. (B) RNA was isolated from strains carrying plasmids with P_Q driving different Anti-Q alleles: lane 1, wild-type Anti-Q; lane 2, a vector control; lane 3, Δ34–65; lane 4, Δ66–95. Blots were probed for Anti-Q (Upper) or 5S (Lower). (C) In vitro transcription reactions using PCR products as templates that contained P_X and Anti-Q sequences through +117 with an ectopic terminator (IRS1) downstream. The frequency of termination at Anti-Q was calculated by dividing by the intensity of the Anti-Q band by the intensity of the IRS1 band. Values were normalized to termination at Anti-Q in a reaction with the wild-type template. Bars are labeled with the mutations carried on each template. White bars indicate templates that carry mutations predicted to restore secondary structure. Asterisks indicate a significant difference in termination frequency (P < 0.02). (D) The proposed structure of Anti-Q as a functional terminator. The different stems are labeled. The in vitro termination points, as described in the text, are labeled with asterisks. Experimentally tested mutations are labeled. Regions removed by the stem deletions are boxed.

Fig. 4.

Anti-Q termination mutations affect Q_L, Q_S, and Anti-Q, but not prgX RNA levels. Total RNA was isolated from strains containing the following plasmids: lane 1, pBK2; lane 2, pBK2-D1 (Δ107–113); lane 3, pBK2-D2 (Δ116–120); lane 4, pBK2-S112 (U112C); lane 5, pBK2-S119 (U119C). (A) Northern blots in which the same RNA blot was hybridized to probes specific for Anti-Q, Q_L and 5S (Materials and Methods and Table S4). (B) The prgX blot was hybridized to a probe specific for the prgX ORF. Full length (1.4 kb) and RNase III-processed (1.2 kb) X RNA transcripts are indicated. 5S RNA levels were used as loading controls. (C) PrgX protein levels in the above strains were determined by Western blotting. Averages and SDs were determined from ≥3 independent experiments. Asterisks indicate a significant difference, relative to wild-type; P < 0.05, t test.

Fig. 5.

The prgQ operon responds to a lower concentration of pheromone in strains with reduced Anti-Q termination. OG1Sp containing pBK2 and pBK2-D1 (10⁸ cells/mL) were induced with 0.1 ng/mL or 1 ng/mL of cCF10. Samples were taken at 0, 15, 30, and 90 min after induction. Total RNA was isolated and analyzed by real-time RT-PCR. Q_L RNA levels were normalized to gyrB control. Data represents the mean and range of two independent experiments.

Fig. 6.

Branched terminators can be found in bacterial chromosomes. (A) The chromosomal context of t11562 is shown. ORFs for OG1RF_11562 and OG1RF_11563 are indicated by open arrows. t11562 is shown as a branched stem-loop. The RNA sequence and a predicted structure for t11562 are depicted below. Nucleotide pairs indicated by a filled circle (●) are not predicted by mfold. (B and C). In vitro transcription reactions using free RNAP (B) or RNAP adsorbed to nickel-agarose beads (C) using a PCR template with P_Q followed by t11562, amplified from pCY1790322. The reaction was performed as detailed in Materials and Methods except reactions were allowed to proceed for 30 min at room temperature. The products of t11562 are predicted to be ∼175-bp long and are indicated by an asterisk. (D) RNA from stationary phase E. faecalis OG1RF cells with or without pCY1790322 carrying the P_Q fused to the t11562 terminator was used for a Northern blot. The blot was hybridized to a probe specific for the t11562 terminator.