Locking the nontemplate DNA to control transcription

Abstract

Universally conserved NusG/Spt5 factors reduce RNA polymerase pausing and arrest. In a widely accepted model, these proteins bridge the RNA polymerase clamp and lobe domains across the DNA channel, inhibiting the clamp opening to promote pause-free RNA synthesis. However, recent structures of paused transcription elongation complexes show that the clamp does not open and suggest alternative mechanisms of antipausing. Among these mechanisms, direct contacts of NusG/Spt5 proteins with the nontemplate DNA in the transcription bubble have been proposed to prevent unproductive DNA conformations and thus inhibit arrest. We used Escherichia coli RfaH, whose interactions with DNA are best characterized, to test this idea. We report that RfaH stabilizes the upstream edge of the transcription bubble, favoring forward translocation, and protects the upstream duplex DNA from exonuclease cleavage. Modeling suggests that RfaH loops the nontemplate DNA around its surface and restricts the upstream DNA duplex mobility. Strikingly, we show that RfaH-induced DNA protection and antipausing activity can be mimicked by shortening the nontemplate strand in elongation complexes assembled on synthetic scaffolds. We propose that remodeling of the nontemplate DNA controls recruitment of regulatory factors and R-loop formation during transcription elongation across all life.

Figure 1

RfaH interactions with the EC. A. A model of the RfaH-bound EC. RNAP α (grey), β (blue), and β′ (yellow) subunits and RfaH NTD (magenta) are depicted by simplified molecular surfaces; β (with the GL labeled) and RfaH NTD are rendered semi-transparent. Nucleic acids are shown as cartoons, two Mg²⁺ ions in the active site – as cyan spheres, and an incoming NTP – as red sticks. B and C. Footprinting of the upstream (B) and downstream (C) RNAP boundary in G8 and U11 ops ECs. Exo III was added to ECs assembled on synthetic scaffolds in which RNAP is halted at G8 and U11 positions in the ops element (NT sequence G₁GCGGTAG₈CGT₁₁G); the probed DNA strand (T, panel B; NT, panel C) was 5′-end labeled with [γ³²P]-ATP. RfaH was present at 100 nM where indicated. Aliquots were quenched at the indicated times (0 represents an untreated DNA control) and analyzed on a 12 % denaturing gel; a representative of three independent experiments is shown. Numbers indicate the distance from the RNAP active site.

Figure 2

Probing the upstream fork junction by crosslinking with 8-MP. G8 and U11 ECs containing the TA intercalation site (highlighted in yellow) upstream from the ops element were assembled with [γ³²P]-ATP labeled T DNA and RNA; C9 and G10 ECs were made by walking from G8. ECs were supplemented with 100 nM RfaH (where indicated) and illuminated with the 365 nm UV light. The products were analyzed on 12 % gels; U11 ECs were run on a different gel. The crosslinked T strand (TxNT) migrates much slower that free T DNA. Fractions of the crosslinked T DNA were determined in the absence (white bars) and presence (black bars) of RfaH. Error bars indicate the SDs of triplicate measurements.

Figure 3

Protection of the upstream DNA by RfaH and RNAP variants. A. U11 ECs were assembled with [γ³²P]-ATP labeled T strand with the WT or altered RNAP. Exo III was added following incubation with RfaH (or storage buffer) and the reactions were analyzed as above. B. The WT U11 ECs assembled as in A were probed with Exo III in the presence of selected RfaH variants. AAAA is a quadruple mutant in which HTTT residues 65–68 were replaced with four alanines.

Figure 4

Modeling Exo III approach to the EC. Models of the Exo III digesting the upstream duplex DNA in different ECs (see the main text and for details). The EC components, RfaH, and Exo III are depicted as in Figure 1. Numbers indicate distances from the Exo III active site to the upstream fork junction. Supplementary Information Fig. S6 shows alternative views of the models, which can also be viewed as 3D PDB models (see Supplementary Information).

Figure 5

Constraining NT DNA inhibits Exo III cleavage and promotes elongation. A. U11 ECs were assembled on scaffolds with the full-length (FL) NT DNA or NT DNAs containing deletions of four or five nucleotides. Exo III probing was carried out in the presence and in the absence of RfaH. Numbers indicate the distance from the bubble, to facilitate comparison with Fig. 4. B. ECs were assembled on scaffolds with the FL or Δ5 NT DNA and [γ³²P]-ATP labeled G14 RNA, and incubated with 10 μM GTP, 150 μM ATP, CTP, UTP for 5–240 sec before quenching. RNAs were analyzed on 12% gel. RNA fractions were determined from 3 replicates (±SD).

Figure 6

The upstream protection is independent of the ops DNA sequence. The scrambled ops ECs were assembled on a scaffold shown on top; the NT strand residues in magenta match those in ops. The wild-type full-length RfaH does not bind to this EC because contacts with the ops element are required to induce RfaH domain dissociation to expose the RNAP-binding site on the NTD. The isolated NTD binds to any EC and serves and a model of activated, post-recruitment conformation of RfaH. The assembled with the T DNA strand 5′-end labeled with [γ³²P]-ATP EC was incubated with RfaH or NTD (at 100 nM) and treated with Exo III. The reactions were analyzed on a 12% urea-acrylamide gel. Numbers indicate the distance from the RNAP active site.