RNA polymerase gate loop guides the nontemplate DNA strand in transcription complexes

Abstract

Upon RNA polymerase (RNAP) binding to a promoter, the σ factor initiates DNA strand separation and captures the melted nontemplate DNA, whereas the core enzyme establishes interactions with the duplex DNA in front of the active site that stabilize initiation complexes and persist throughout elongation. Among many core RNAP elements that participate in these interactions, the β' clamp domain plays the most prominent role. In this work, we investigate the role of the β gate loop, a conserved and essential structural element that lies across the DNA channel from the clamp, in transcription regulation. The gate loop was proposed to control DNA loading during initiation and to interact with NusG-like proteins to lock RNAP in a closed, processive state during elongation. We show that the removal of the gate loop has large effects on promoter complexes, trapping an unstable intermediate in which the RNAP contacts with the nontemplate strand discriminator region and the downstream duplex DNA are not yet fully established. We find that although RNAP lacking the gate loop displays moderate defects in pausing, transcript cleavage, and termination, it is fully responsive to the transcription elongation factor NusG. Together with the structural data, our results support a model in which the gate loop, acting in concert with initiation or elongation factors, guides the nontemplate DNA in transcription complexes, thereby modulating their regulatory properties.

Keywords: RNA polymerase; beta pincer; discriminator; promoter; transcription.

Fig. 1.

Bacterial transcription complexes. (A and B) An overview of the open promoter complex (RPo) with a 6-nt discriminator (A) and the TEC with bound NusG (B). The composite models (Datasets S1 and S2) were generated using T. thermophilus RPo (22) and TEC (38) and elements from other structures as described in SI Materials and Methods. Proteins are depicted by simplified differentially colored molecular surfaces; β, σ^A and NusG are rendered semitransparent. The positions of the N-terminal domain (NTD) of the σ1.1 region, α C-terminal domains (CTDs), and NusG CTDs were chosen arbitrarily within the volume permitted by the length of the flexible linkers; the cyan arrow in A indicates that the NTD of σ1.1 is predominantly located near the β lobe domain in E. coli RPo (39). Nucleic acids and βGL are shown as cartoons, two Mg²⁺ ions in the active site are shown as cyan spheres, and an incoming NTP is shown as red sticks. Selected DNA nucleotides are numbered relative to the TSS in A and from the RNA 3′ end in the posttranslocated TEC in B. (C) A zoomed-in view of the GL and the discriminator region of the NT DNA (the rectangular area outlined in A). The β lobe and σA are semitransparent. E. coli σ⁷⁰ Met102 was modeled into a homologous position of T. thermophilus σA and is shown in a balls-and-sticks configuration. (D) A side view of the TEC. Nucleic acids are depicted as surfaces. The view is clipped along the dashed line in B to expose NusG interactions with the GL and the β′ clamp domain. (Inset) NusG–RNAP contacts.

Fig. S1.

GL conservation and essentiality. (A) Multiple sequence alignment of GL sequences from divergent bacterial species. Species names are abbreviated as follows: Eco, E. coli; Hpy, Helicobacter pylori; Tth, T. thermophiles; Syn, Synechocystis sp. PCC 6803; Bsu, Bacillus subtilis; Mtu, Mycobacterium tuberculosis; Tma, Thermotoga maritima; Aae, Aquifex aeolicus; Mge, Mycoplasma genitalium. Amino acid residues within the sequence that were replaced with two glycines are shaded as follows: hydrophobic, green; polar, olive; Pro and Gly, yellow; Asp and Glu, red; Arg, Lys, and His, blue. Amino acid residues flanking the replaced region are shaded gray. Cartoons on the right depict the GL from E. coli RNAP (PDB ID code 4IGC) and outline its interactions with the neighboring TEC elements. The presence of the NT DNA near R371 and E374 is hypothesized based on the assumption that the single-stranded region of the NT-DNA follows a similar route in the TEC and the initiation complex. (B) RNAP lacking the GL does not support cell growth. The E. coli DH5α cells carrying plasmids with the rpoB gene (WT, D516V, or D516V+ΔGL) under the control of an isopropyl β-d-1-thiogalactopyranoside (IPTG)-inducible P_trc promoter were grown in LB to OD₆₀₀∼0.3, induced with 0.2 mM IPTG for 1 h, and then challenged with 50 µg/mL rifapentine (Rif). Cell density was measured at the indicated times after the addition of rifapentine; the averages (± SD) were determined from three independent experiments. (C) A decrease in RNA synthesis upon induction of the ΔGL rpoB. RNA samples were collected from strains grown as in B after 20-min and 60-min incubation with rifapentine. Total RNA was extracted using hot phenol chloroform and RNAlater solution (Thermo Fisher Scientific). Approximately 100 ng of total RNA was analyzed using the 6000 Nano Chip Kit on an Agilent 2100 Bioanalyzer. (Left) Representative electropherograms of RNA samples collected after 60-min incubation with rifapentine. The lower marker (a standard run with every sample) was used for the alignment of the rRNA peaks. (Right) Percentage of 16S rRNA and 23S rRNA species relative to the total RNA at 20 min and 60 min after the addition of rifapentine for WT (blue), WT D516V (red), and ΔGL D516V (green) RNAPs.

Fig. 2.

Footprinting of the ΔGL open complexes. (A) A linear λP_R fragment in which the NT DNA was labeled at the 5′ end with [γ³²P]-ATP. (B) Preformed open complexes were treated with 2 mM KMnO₄ for 30 or 60 s. After quenching, ethanol precipitation, and piperidine cleavage, the DNA was analyzed on an 8% denaturing gel. A representative of three independent experiments is shown; the 0 point is an untreated DNA control. Traces of the 60-s reactions were generated with ImageQuant; the band intensities normalized to the −10 signal (taken as 1) are shown in black (WT) and blue (ΔGL). (C) ExoIII was added to preformed promoter complexes. Aliquots were quenched at the indicated times (0 represents an untreated DNA control) and analyzed on a 6% denaturing gel; a representative of three independent experiments is shown. The positions of the modified residues (B) or the protection boundary (C) were identified using sequencing ladders.

Fig. 3.

Effects of the GL deletion on promoter complex properties. (A, Upper) Promoters used in this study. The −35 and −10 hexamers are boxed. The discriminator element is underlined, and its positions are numbered. The TSS and the D2 (−5 at λP_R) base that makes key contacts with σ1.2 are indicated. (Lower) The half-lives of open complexes (in minutes) were measured by assaying the fraction of transcriptionally competent complexes following the competitor challenge (SI Materials and Methods) in at least three independent repeats; data are shown as mean ± SD. The GL effect was defined as the ratio of the half-lives of the WT and ΔGL open complexes. (B, Upper) Open complexes assembled on λP_R with 4-thio-dT at the −4 NT position were exposed to 365-nm UV light. The reactions were separated on 4–12% Bis-Tris gels. (Lower) Relative cross-linking to σ and β (normalized to DNA) was calculated from four independent experiments; error bars indicate the SD. (C) Abortive initiation at λP_R. Open complexes were formed in the presence of [γ³²P]-ApU and were incubated with 250 μM NTPs (see Fig. S2). A representative gel and trace analysis of abortive products are shown. (D) TSS at rrnB P1 were mapped by primer extension of in vitro-transcribed RNAs. Extension products were analyzed on a denaturing 12% gel along with the sequencing ladder generated with the same primer. Positions corresponding to a TSS at 6 and 9 are indicated.

Fig. S2.

Effects of RNAP variants and promoter region swaps on abortive synthesis. (A) Abortive synthesis at the λP_R promoter. ApU, abortive RNAs, terminated (70 nt), and run-off (153 nt) RNAs and an unidentified arrested product are indicated. (B) A schematic representation of RNAP holoenzyme interactions with promoter elements. The sequences of the T7A1 and λP_R promoters that have a 7- and a 6-nt discriminator, respectively, are shown. (C) The pattern of abortive products depends on the discriminator region. Abortive transcription on synthetic hybrid templates that contain promoter regions from T7A1 (open boxes) or λP_R (hatched boxes). Reactions were carried out as in A with the WT RNAP. Variants with the λP_R discriminator (indicated by red dots) produce abundant abortive products, whereas those with the T7A1 discriminator do not.

Fig. S3.

The length but not the sequence of the discriminator influences promoter escape. (Upper) Variants of the λP_R promoter containing a 6-nt (red) or 7-nt (black) discriminator, including those from λP_R and T7A1. (Lower) Abortive synthesis by the WT and σY101A holoenzymes. For the WT RNAP, promoters with 6-nt discriminators give rise to more abundant abortive RNAs than those with a 7-nt discriminator. These differences disappear when RNAP–NT strand interactions are disrupted by the Y101A substitution.

Fig. 4.

The GL RNAP is responsive to NusG. (A) Single-round elongation assays were carried out on the pVS54 template; positions of pause sites, the hlyT terminator, and the run-off (RO) are shown. Halted [α³²P]-CMP–labeled A38 TECs formed with the WT or ΔGL RNAP were chased in the presence or absence of NusG. Reactions stopped at the indicated times were analyzed on 8% denaturing gels. (B) The average fraction (± SD) of run-off RNA determined from three independent experiments, including that shown in A. (C) Close-up view of NusG contacts with the GL and NT-DNA in the TEC model. The positions of residues replaced by alanine in NusG (7) and Spt5 (31) are shown in magenta.

Fig. S4.

ΔGL TECs are resistant to GreB-facilitated transcript cleavage. Linear pIA226 DNA template (100 nM), holo RNAP (200 nM; WT or ΔGL), ApU (100 µM), and starting NTPs (1 µM GTP, 5 µM ATP and UTP, 10 µCi [α³²P]-GTP, 3,000 Ci/mmol) were mixed in 30 μL of TGA-2 (20 mM Tris-acetate, 20 mM Na-acetate, 2 mM Mg-acetate, 5% glycerol, 1 mM DTT, 0.1 mM EDTA, pH 7.9) and were incubated for 15 min at 37 °C. Halted A26 complexes were purified by gel filtration through a G-50 spin column (GE Healthcare) equilibrated in TGA-2, diluted fourfold, and stored on ice. Reactions were initiated by shifting samples to 37 °C. GreB (20 nM) was added where indicated. Samples were removed at the times shown and were quenched by the addition of an equal volume of STOP buffer (10 M urea, 20 mM EDTA, 45 mM Tris-borate; pH 8.3). (Upper) A schematic of the experimental setup and the sequence of the A26 RNA transcript, with the positions of [α³²P]-GTP indicated in bold. (Lower) RNA products were separated on an 8% denaturing urea-acrylamide gel. The percent of A26 remaining was calculated relative to A26 RNA at time 0 (before GreB addition).

Fig. S5.

Effects of GL deletion on hairpin-dependent pausing and termination. (A) Single-round assays on the pIA171 template encoding the hisP signal. Halted [α³²P]-CMP–labeled A29 TECs were formed with the WT or ΔGL RNAP and chased with NTPs (10 μM GTP, 150 μM ATP, CTP, UTP) and rifapentine in the presence or absence of 100 nM NusA. Samples were analyzed on 8% denaturing gels. The hisP half-life calculated as described in ref. is shown below each panel. (B) Termination at a NusA-dependent rsxC terminator. (Upper) Transcript generated from the λP_R promoter on a linear pIA1239 DNA; the TSS (bent arrow), C-less region (residues 1–26), rsxC terminator (release at 107), and transcript end (179) are indicated. (Lower) Halted A26 TECs were formed at 50 nM with WT or ΔGL RNAP. Termination was assayed in single-round A26 RNA extension by the addition of all four NTPs (to 150 µM) and heparin (at 10 µg/mL) in the absence or presence of 100 nM NusA. The reactions were incubated for 10 min at 37 °C and were quenched. Products were analyzed on a 6% denaturing gel. Positions of terminated (Term) and run-off (RO) RNAs are shown on the left. Termination efficiency (terminated transcript as a fraction of total RNA) was determined in three independent experiments. (C) Efficiency of termination at selected terminators. Halted [α³²P]-labeled complexes were chased with 150 μM NTPs and rifapentine. Termination efficiencies for the WT (white bars) and ΔGL (black bars) RNAPs were calculated from three independent experiments. The average values are shown inside the bars; the error bars indicate SD. All intrinsic termination signals are composed of an RNA hairpin followed by a U-rich region, but their mechanisms of RNA release may differ. Three models of termination have been proposed (52). In the forward translocation model, formation of the hairpin pushes RNAP forward without concomitant RNA extension, shortening the RNA:DNA hybrid and thereby destabilizing the TEC. In the hybrid-shearing model, the hairpin pulls the nascent RNA out. In the allosteric model, the hairpin induces a structural change in the RNAP that facilitates melting of the hybrid. Among the terminators tested here, t500 relies on forward translocation, but tR2 and hisT do not (52); the mechanisms of others are not known. The ΔGL RNAP terminated somewhat less efficiently at five of six terminators. The lack of defect at hisT may be explained by its unusual structure: It is composed of an exceptionally stable hairpin followed by a run of nine U residues. At other terminators the hairpin is weaker, and U-tracks are interrupted by G and C residues. It is possible that the mild pausing defect of ΔGL RNAP is sufficient to reduce pausing at an interrupted U-track, thereby reducing termination, whereas the perfect U would overcome this defect.

Fig. S6.

The GL alters the NT DNA path in transcription complexes. (A) In elongation complexes, the GL contacts with the NT strand weaken base pairing at the upstream fork junction while stabilizing the downstream base pairs. In the absence of the GL, the upstream junction is strengthened, and backtracking is inhibited. (B) In prescrunched rrnB P1 ICs, the GL binds the NT DNA just upstream from the point of scrunching, and transcription initiates at a distal (9A) site. In the absence of the GL, the NT DNA is more relaxed, allowing utilization of a proximal TSS at 6A.