Mycobacterial HelD is a nucleic acids-clearing factor for RNA polymerase

Abstract

RNA synthesis is central to life, and RNA polymerase (RNAP) depends on accessory factors for recovery from stalled states and adaptation to environmental changes. Here, we investigated the mechanism by which a helicase-like factor HelD recycles RNAP. We report a cryo-EM structure of a complex between the Mycobacterium smegmatis RNAP and HelD. The crescent-shaped HelD simultaneously penetrates deep into two RNAP channels that are responsible for nucleic acids binding and substrate delivery to the active site, thereby locking RNAP in an inactive state. We show that HelD prevents non-specific interactions between RNAP and DNA and dissociates stalled transcription elongation complexes. The liberated RNAP can either stay dormant, sequestered by HelD, or upon HelD release, restart transcription. Our results provide insights into the architecture and regulation of the highly medically-relevant mycobacterial transcription machinery and define HelD as a clearing factor that releases RNAP from nonfunctional complexes with nucleic acids.

Fig. 1. Cryo-EM structures of Msm HelD–RNAP complexes.

a Description of Msm RNAP core (PDB ID 6F6W) subunits and domains; RNAP subunits are color-coded according to the inset legend. b–d Atomic model surface representation of three identified Msm HelD–RNAP complexes: State I – PCh-engaged, State II – PCh-engaged AS-interfering, and State III – PCh-dis-engaged AS-interfering. When fully ordered in State I and II (b, c), the HelD protein (color-coded as in e) forms a crescent-like shape, ends of which protrude to the primary and secondary channels of the RNAP core. The partly ordered HelD protein in State III (d) vacates most of the RNAP primary channel. e Schematic linear representation of the domain structure of the HelD protein. The 1A domain (two shades of yellow) is split in aa sequence into two parts, separated by a large HelD-specific insertion (hues of blue and orange). The nucleotide-binding motifs are marked as vertical thick black lines. Aa numbering (Msm) is shown below. f–h Three states of HelD as observed in b–d color-coded according to the domain structure (e); secondary structure elements are marked as in Supplementary Fig. 7a.

Fig. 2. The HelD N-terminal domain inserts into the RNAP secondary channel; domains 1A–2A comprise the NTPase unit.

a, b Ribbon (State I) and surface (State II) representation of the HelD N-terminal domain interaction with the secondary channel of RNAP core (gray). The HelD coiled-coil domain (NCC-domain, firebrick) and the distinct loop (NG-loop, red) of the HelD globular domain (NG-domain, salmon) are inserted between β′-funnel, shelf, and jaw. The NCC-domain reaches only the boundary line of the β′ bridge helix (β′-BH, cyan) and leaves a passageway to the RNAP core active site (MgA, magenta sphere). The HelD NCC also restricts the trigger loop (TL, yellow) movement. The linker (NG-linker, violet) connects the N-terminal domain with domain 1A-1. c The two Msm HelD Rossman fold domains (1A yellow and 2A green) form a canonical NTPase unit heterodimer with respect to structurally described SF1 helicases. Domain 1A tightly packs with β-lobe (dark gray) and its extension (brown) is clamped in between one β-turn (β/184–187) of β-lobe and the tip of the β′ subunit jaw (light gray). d Model of ATP binding to the conserved nucleotide-binding site of motifs Q (blue), I (brown), II (pink), ~III (orange), IIIa (red), Va (pale green), and VI (deep blue). ATP (green) and Mg²⁺ (magenta sphere) are added based on superposition with the ternary complex of UvrD (PDB ID 2IS4). e HelD exhibits ATPase and GTPase activities but does not hydrolyze CTP. The apparent negative value of CTP hydrolysis was caused by high background readings. The bars show mean values, the error bars indicate ±SD and the individual symbols represent values from three independent replicates. The data were analyzed and the graphics created with GraphPad Prism 7.02.

Fig. 3. The Msm HelD-specific domain interactions with the RNAP primary channel.

a Surface of the Msm RNAP core (PDB ID 6F6W), color-coded as in Fig. 1a with the description of individual domains and functional parts. b, c Surface representation of States I and II of the Msm HelD–RNAP complex with RNAP color-coding as in a and marked domain names; HelD color-coding as in Fig. 1e. d, e Ribbon representation of the HelD-specific domain inserting into the RNAP primary channel in State I (d) and State II (e). In State I (d), the clamp opening (CO, blue) HelD-specific domain is projected from the HelD 1A domain (yellow) towards the β′-clamp (gray). At one end, the CO is bonded to the 1A domain by the CO-linker (cyan), and stabilized by β-turn 561-563 and α19 (yellow). On the other end, the CO-domain tip abuts towards the β′-NCD three-stranded sheet. Concomitantly, the HelD helix α16 (part of peptide HelD/449–473, orange) butts against the β′/1164–1210 three-stranded sheet. The connection between α16 and the 1A-extension is disordered (dotted line). In State II (e), The CO interaction with the 1A domain remains similar to State I (d). The CO-domain tip, however, shifts towards the β′-rudder (green) and β′/122–133 α-helix. Concomitantly, the HelD PCh-loop (orange) folds towards the active site (MgA, magenta sphere) and folds back towards the 1A-extension (brick) and 1A domain. f The PCh-loop folds into the RNAP active site. The HelD loop 473–494 and the two adjacent α-helices (α16 and α17, orange) fold alongside the RNAP bridge helix (BH, cyan) towards the RNAP active site and HelD/Asp482 directly contacts the MgA (magenta sphere, details in the inset, coordination of MgA is marked with blue dotted lines). The RNAP trigger loop (TL, yellow) is restricted and folded between the HelD PCh-loop helix α17, the HelD NCC-domain (ruby), β′-BH, and the β-core domain (dark gray). g Detail of the β′-BH interaction with HelD α16 and α17. BH β′/Arg874 and Arg875 sandwich HelD/Tyr466, and β′/Tyr871 stacks on HelD/Phe502. The stacking interactions are marked with yellow dotted lines. h, i The HelD PCh-loop binding in the active site chamber is mutually exclusive with the presence of the transcription bubble. Two perpendicular views of superposition of the Tt RNAP elongation complex (PDB ID 2O5J, pale colors) and HelD State II (solid colors) are shown. The folded TL in pre-translocated EC would sterically clash with the HelD NCC-domain. The HelD PCh-loop tip would sterically clash with RNA/DNA hybrid at positions +1 to −2, and the HelD α16 and α17 helices would clash with downstream DNA duplex. Color code as in f, template DNA in pink, non-template DNA in gray, product RNA and incoming NTP at position +1 in green.

Fig. 4. Binding of Msm HelD to RNAP and its effects on DNA–RNAP interactions.

a DNA binding to RNAP - EMSA - binding of 300 bp DNA to the Msm RNAP core and the effect of HelD. b The same gel as above but stained for proteins. The dotted line shows where the gel was electronically assembled. c Quantitation of EMSA – the bars here (the amount of unshifted DNA) and in e are mean values from at least three independent experiments, the error bars show ±SD, the individual symbols show values of individual independent replicates. The leftmost bars were set as 1 and the other values within each graph were normalized relative to this bar. The turquoise bars here and in e indicate the addition of HelD. d EC disassembly - scheme: ECs were assembled on DNA:RNA scaffolds and challenged with HelD and/or NTPs. RNAP released into buffer was quantitated by western dot blots. e Quantitation of EC disassembly from five independent experiments. Representative primary data are shown below the graph. Presence/absence of individual components is indicated. +° indicates heat-inactivated HelD. The statistical significance in e for the indicated combinations was p < 0.05 (one-sided Student’s t-test; exact p-values are written in the graph). f Representative SDS-PAGE of immunoprecipitations of Msm RNAP (β), σ^A, and HelD. All proteins were FLAG fusions, the antibody was anti-FLAG. Wt, a strain without any FLAG fusion. The identity of the bands was confirmed by mass spectrometry. IP, immunoprecipitation; M, markers. The experiment (biological replicates) was performed 3× with the same result. g Representative western blot of IPs of FLAG-tagged Msm RNAP (β), σ^A, and HelD. Antibodies against RNAP β and σ^A were used to detect the presence of proteins in complexes. M, marker – purified σ^A. The experiment was performed twice with the same result. h In vitro protein interactions - EMSA. Proteins were detected by Simply blue SafeStain. In all cases, RNAP was first reconstituted with HelD and then with RbpA and/or σ^A. A small, but a reproducible shift was observed after the addition of both RbpA and σ^A to RNAP–HelD, indicating the presence of all proteins in one complex. Numbered arrows indicate complexes with different protein composition (determined by mass spectrometry). In some cases, complexes with different protein compositions displayed the same migration in the gel: 1. RNAP, RNAP-RbpA; RNAP-σ^A; 2. RNAP–HelD, RNAP–HelD-RbpA; 3. RNAP–HelD-σ^A; 4. RNAP–HelD-σ^A-RbpA. The experiment (biological replicates) was performed 3× with the same result.

Fig. 5. A model of the HelD functioning in RNAP recycling.

a When EC stalls, it needs to be disassembled. b The HelD N-terminal domain (pink) first approaches the RNAP secondary channel and then induces changes in RNAP likely destabilizing the RNAP-dwDNA interaction. c Subsequent interactions of the HelD PCh-loop (orange) and the whole HelD-specific domain (cyan) in the RNAP primary channel open the RNAP cleft, widen the RNA exit channel and mechanically interfere with dwDNA. d An even broader cleft/RNA exit opening together with the PCh-loop intervening deep in the AS (MgA, magenta sphere) displace dwDNA and the RNA/DNA hybrid from the active site cavity. e The HelD–RNAP nucleic acid-free complex binds σ^A factor and RbpA, and all factors can bind RNAP core simultaneously. f The complex binds to DNA promoter via the σ^A factor with a concomitant displacement of HelD from RNAP by an unknown mechanism, possibly dependent on NTP hydrolysis by HelD and a new round of σ^A-dependent transcription cycle can initiate.