A novel antisense RNA regulates at transcriptional level the virulence gene icsA of Shigella flexneri

Abstract

The virulence gene icsA of Shigella flexneri encodes an invasion protein crucial for host colonization by pathogenic bacteria. Within the intergenic region virA-icsA, we have discovered a new gene that encodes a non-translated antisense RNA (named RnaG), transcribed in cis on the complementary strand of icsA. In vitro transcription assays show that RnaG promotes premature termination of transcription of icsA mRNA. Transcriptional inhibition is also observed in vivo by monitoring the expression profile in Shigella by real-time polymerase chain reaction and when RnaG is provided in trans. Chemical and enzymatic probing of the leader region of icsA mRNA either free or bound to RnaG indicate that upon hetero-duplex formation an intrinsic terminator, leading to transcription block, is generated on the nascent icsA mRNA. Mutations in the hairpin structure of the proposed terminator impair the RnaG mediated-regulation of icsA transcription. This study represents the first evidence of transcriptional attenuation mechanism caused by a small RNA in Gram-negative bacteria. We also present data on the secondary structure of the antisense region of RnaG. In addition, alternatively silencing icsA and RnaG promoters, we find that transcription from the strong RnaG promoter reduces the activity of the weak convergent icsA promoter through the transcriptional interference regulation.

Figure 1.

Identification of the RnaG promoter. (A) Primer extension analysis was carried out using the oligo G+59 on 10 µg of total RNA from the E. coli wt strain HMG11, a MC1029 derivative, transformed with plasmid pMYSH6601 (lane 2) or pKG673 (lane 3). Lane 1 is the control sample in the absence of RNA. Lanes G, A, T and C represent the dideoxy sequencing reactions carried out with the same primer. The positions of the two transcriptional start sites are indicated. Smears below the specific signals are due to the low quality of some commercial oligonucleotides. (B) The relevant nucleotide sequence of the regulatory region of the icsA gene. The first nine coding triplets of icsA, the transcriptional start point, the –10 and –35 consensus sequences of the icsA and RnaG promoters are indicated. Mutated TATA boxes created in P_icsA and P_RnaG to construct plasmids pGT1083 and pGT1129, respectively, are reported above and below the sequence. (C) Schematic representation of plasmids pGT1127 (P_icsA/_RnaG), pGT1129 (P_icsA/_RnaGmut), pGT1083 (P_icsAmut/_RnaG). (D) Primer extension analysis was performed on 15 µg of total RNA from the E. coli strain HMG11 transformed with plasmids pGT1127, pGT1129 or pGT1083 using a mixture of primers G+110 and G+50 to detect icsA mRNA and RnaG, respectively. Lanes G, A, T and C represent the sequencing reactions using the primer G+110.

Figure 2.

Transcriptional Interference regulates icsA and RnaG promoter activity. (A) About 200 ng of supercoiled plasmids pGT1127 (P_icsA/RnaG), pGT1129 (P_icsA/_RnaGmut) and pGT1083 (P_icsAmut/RnaG), schematically represented in Figure 1C, were used as templates in in vitro transcription assays carried out at 30 and 37°C. RnaG and icsA transcripts were detected by primer extension using a mixture of oligos G+50 and G+110. Lanes marked G and C are the dideoxy sequencing reactions performed on pGT1127 with the oligo G+110 as primer. (B) Schematic organization of the E. coli chromosome (strain P90C) carrying the different ULS lacZ fusions. (C) β-Galactosidase activity of ULS transcriptional fusions was determined on strains grown in LB broth (A₆₀₀ = 0.4) at 37°C. Values represent the average of four independent experiments and the standard deviation is indicated.

Figure 3.

The RnaG downregulates icsA transcription. Transcription was investigated in vitro as function of increasing amounts of purified RnaG using as template a 331-bp DNA fragment (from position –117 to +214), corresponding to the wt icsA promoter (A), the same DNA fragment carrying the mutation 81/4 (see text and Figure 7A) (B) and a 480-bp DNA fragment (from position –314 to +166) corresponding to the E. coli hns gene as control (57) (C). The wt and mutated icsA DNA fragments were generated by PCR amplification using the primer pair G-100/ACC9 and plasmids pGT1129 and pGT1129M as templates, respectively. F and T indicate full-length run-off and terminated icsA transcripts, respectively. Transcripts were labeled by incorporating [α⁻³²P]-UTP and lanes marked with M are different RNA ladders.

Figure 4.

RnaG negatively affects icsA expression. (A) Expression of icsA-lacZ fusion USL1129 was monitored in cells transformed with pGT1083, pGT1127, pGT1129 or pGEMT vectors. (B) The in vivo level of icsA and RnaG transcripts was monitored during the growth of Shigella by real-time PCR. The expression trend of either template, normalized to the corresponding maximum (100%), is shown.

Figure 5.

RNA probing of the icsA mRNA leader region either alone or in combination with RnaG. (A) Total RNA (15 μg) was primer-extended in the absence or in the presence of the indicated amounts of purified RnaG using the oligo ACC9. (B) In vitro transcribed icsA RNA (2 pmol) was treated with increasing amounts of DMS (lanes 1, 0%; lane 2, 0.3%; lanes 3, 0.6%; lane 4, 1.2%) and CMCT (lanes 5, 0 mg/ml; lane 6, 0.25 mg/ml; lane 7, 0.5 mg/ml; lane 8, 1.0 mg/ml). (C) The icsA mRNA (2 pmol) was treated with CMCT (3 mg/ml) either without (lanes 1 and 2) or with different amounts of RnaG: 1 pmol (lanes 3 and 4), 2 pmol (lanes 5 and 6) and 4 pmol (lanes 7 and 8). Relatively to both (B) and (C), modified nucleotides were detected by primer extension using the oligo G+187 and accessible sites were evaluated comparing samples incubated in the absence (–) and in the presence of either DMS or CMCT. Lanes C, T, A and G correspond to DNA sequencing ladders made with the same primer. (D) About 300 fmol of [³²P]-end-labeled icsA mRNA were digested with RNases T1 (2.5 × 10^–2 U, lanes 2, 7 and 12; 5 × 10^–2 U, lanes 3, 8 and 13) or T2 (1.1 × 10^–3 U, lanes 4, 9 and 14; 2.2 × 10^–3 U, lanes 5, 10 and 15) in the absence (lanes 1–5) and in the presence of 290 (lanes 6–10) and 580 fmol (lanes 11–15) of RnaG as indicated. M₁ and M₂ represent the OH– and ΔT1 ladders, respectively. (E) Schematic representation of the secondary structure of icsA mRNA (nucleotides from +1 to +145). The AH1 and AH2 motifs, nucleotides reactive to DMS or CMCT (circles), cleavage points by T1 and T2 RNases (triangles), the initiation triplet and the S.D. sequence are indicated.

Figure 6.

Model showing the RnaG-mediated transcriptional attenuation mechanism. (A) The interactions of GH3 and GH2 with AH1 possibly provide the initial nucleation points leading to duplex formation between RnaG and icsA mRNA. The GH1 hairpin is not represented contacting AH2 because at this stage the latter motif is not completely transcribed. (B) RNA–RNA pairing causes modifications of the secondary structure of the icsA mRNA resulting in the formation of a potential intrinsic terminator between nucleotides +78 and +105 (see also Figures 7A and 8).

Figure 7.

The mutation icsA81/4 abolishes the RnaG-mediated transcriptional termination. (A) The hairpin followed by a polyU sequence typical of Rho-independent transcription terminators and base exchanges to create the icsA81/4 mutated mRNA are shown. The underlined nucleotides represent the putative 3′ terminus (positions +104 and +105) of the abortive icsA mRNA predicted according to the TransTerm algorithm (58). In vitro transcription was carried out at 37°C on 300 ng of supercoiled plasmid DNA using pGT1129 (B) and pGT1129M (C) as templates. The icsA and icsA81/4 mRNAs were detected by primer extension using the oligo G+110 (from position +97 to +116). Lanes G and A represent the sequencing reactions using the same primer. After imager quantization, the radioactivity associated with mRNA in (B) and (C) was used to calculate the ratio B/E at the same RnaG concentration. These values have been plotted as percentage (D) assuming 100% transcription in the absence of RnaG.

Figure 8.

Secondary structure of the antisense region of RnaG. (A) Chemical probing of RnaG has been carried out essentially as described in Figure 5B, using the oligo G + 1H, in the presence of increasing amounts of DMS (lane 1, 0%; lane 2, 0.3%; lane 0.6%) and CMCT (lane 4, 0 mg/ml; lane 5, 1.5 mg/ml; lane 6, 3 mg/ml; lane 7, 6 mg/ml). (B) Schematic representation of the secondary structure of RnaG (nucleotides from +1 to +120). Numbering is according to 5′ end of antisense RNA (position +120 on icsA sequence) and nucleotides reactive to DMS or CMCT in (A) are circled in (B) except those between positions +97 and +108 that have been visualized in an additional experiment performed with an other primer (data not shown). The same secondary structure was obtained using RnaG120 (data not shown).