Structure of the giant RNA polymerase ejected from coliphage N4

Abstract

Schitoviruses are widespread prokaryotic viruses that encapsidate a giant (~3,500-residue) virion-associated RNA polymerase (vRNAP). During infection, vRNAP is expelled into Gram-negative bacteria, along with two additional ejection proteins, to assemble a transient DNA-ejectosome that becomes transcriptionally active, initiating viral replication. Here, we present an integrative structural analysis of the coliphage N4 vRNAP (gp50). We find that this 383 kDa enzyme is a multi-domain, single-chain RNA polymerase, structurally distinct from both compact single-chain RNAPs and large multi-subunit holoenzymes. vRNAP is composed of loosely connected domains and exhibits an intramolecular mode of allosteric regulation through its C-terminal domain. Comparative analysis of intact and genome-released virions identified gp51, which forms an outer-membrane complex, and gp52, which assembles a periplasmic tunnel. These proteins generate heterogeneous pores that facilitate the release of vRNAP. We further uncover a signaling hub in the phage tail, composed of the receptor-binding protein, tail tube, and tail plug, that detects receptor engagement and orchestrates the release of ejection proteins. We propose that the beads-on-a-string architecture of vRNAP enables the translocation of megadalton-scale protein complexes through the ~35 Å channel formed by the tail and ejection proteins. These findings establish N4 as a distinctive model for protein translocation through biological channels.

Figure 1:. Cryo-EM reconstruction of N4 vRNAP.

(A) Linear schematics of vRNAP domain organization. (B) Experimental density of ΔN-vRNAP tight complex colored by domain. (C) Ribbon diagram of ΔN-vRNAP tight complex colored by domain. (D) Ribbon diagram of ΔN-vRNAP loose complex.

Figure 2:. Plastic quaternary organization of vRNAP domains.

(A) Map of binding interfaces and interactions that stabilize vRNAP tight conformation. Yellow indicates salt bridges with the corresponding residue linked by a dashed line. Other colored residues indicate hydrogen bonding residues to subunits of the same color. (B) General schematics of all putative domain organizations of vRNAP observed on grid with supporting 2D classes. (C) Putative model of beads-on-a-string pre-ejection conformation. NTD subdomains were predicted by AlphaFold3; RNAP and CTD subdomains were predicted by the protein peeling structural decomposition algorithm SWORD2.

Figure 3:. CTD allosterically regulates vRNAP transcriptional activity.

(A) ΔN-vRNAP tight complex atomic model with RNAP colored by subdomains. (B) Surface model of the tight complex with the DNA hairpin P1 promoter bound. (C) Surface model of isolated RNAP with P1 promoter bound. (D) Conformational changes in RNAP subdomains due to P1 and CTD binding. (E) Radiolabeled transcription runoff assay of FL-vRNAP (res. 1–3500) and recombinant mini-vRNAP (res. 998–2103) shows less processivity for full-length protein compared to mini-vRNAP.

Figure 4:. Asymmetric cryo-EM reconstructions of N4 full virion and empty particle identify ejection proteins gp51 and gp52.

(A) Linear schematic of genomic organization of the N4 ejection cassette (gp50-gp51-gp52) with adjacent ejection-related proteins (gp53-gp54). (B) Cross-section of full virion density with major differences from empty particles boxed in red. (C) Isolated density of α-helical hairpin surrounding the portal is assigned to gp51. (D) Isolated density of gp52 fragment (res. 1–16) localized to the tail lumen. (E) Cross-section of empty particle density with major differences from full virions boxed in red.

Figure 5:. Channel activity of N4 ejection proteins gp51 and gp52.

(A) Comparison of the cryo-EM structure of the nonameric Pseudomonas DEV phage gp73/gp72 ejection channel with the AlphaFold3 prediction of the homologous gp52 and gp51 proteins of the Escherichia coli N4 phage. The outer membrane complex proteins (OMC) are gp73 and gp52. The inner diameters (Ø_i) of the predicted pores of gp52 and gp51 are indicated. (B-G) Lipid bilayer experiments were performed using diphytanoyl phosphatidylcholine (DphPC) membranes bathed in 10 mM HEPES pH 7.4, 1 M KCl as an electrolyte, and at −10 mV applied potential. All samples were added to both sides of the cuvette. After formation of the membrane across the aperture, each experiment was recorded for 15 minutes or less in the event of membrane rupture. Blue arrows indicate open channels, and red arrows indicate channel closure events. Current traces of gp52 in DDM buffer. Channel-forming activity was observed at protein concentrations of 8 ng mL⁻¹ (B), 10 ng mL⁻¹ (C), and 4 ng mL⁻¹ (D). Current traces of gp51 in DDM buffer. Channel-forming activity was observed at protein concentrations of 320 ng mL⁻¹ (E), 80 ng mL⁻¹ (F), and 240 ng mL⁻¹ (G).

Figure 6:. Structure of the N4 asymmetric tail-gating complex.

(A) Ribbon diagram of asymmetric N4 tail with all sheath heterodimers (gp64-gp65) shown in light gray except for the gp53 interacting sheath heterodimer, which forms the tail-gating complex. (B) Atomic model of asymmetric tail-gating complex. (C) Cross-section of experimental density colored by component showing unique internal features. Density for sheaths and front tail-tube protomers has been removed. (D) Densities showing the different conformations of the N4 sheath. Detached conformation shown with an additional viewpoint to show weak density of the sheath (boxed in red).

Figure 7:. Composite model of N4 ejection protein-facilitated genome delivery.

(A) Diagram of the infectious N4 virion with ejection proteins gp50, gp51, and gp52 poised for expulsion. In pre-ejection conformation, gp52 localizes inside the tail, gp51 surrounds the portal, gp50-NTD associates with gp51, and gp50 RNAP-CTD folds as beads-on-a-string inside the capsid. (B) Schematic diagram of phage N4 receptors at the E. coli surface: the novel glycan receptor (NGR), an exopolysaccharide exported by the proteinaceous receptor NfrA. (C) N4 appendages (gp66) reversibly bind NGR. (D) Irreversible binding of the RBP gp65 with the receptor NfrA, which displaces the plug (gp53), letting several luminal copies of gp52 to implant into the membrane as the OMC. (E) The ejection cascade continues, with gp51 being expelled through the tail and OMC to form a tunnel through the periplasm. (F) Gp50 follows in ejection through the interaction between gp51 and gp50-NTD. The inner membrane complex (IMC) is likely formed by both the gp51 C-terminus and the gp50-NTD. gp50-RNAP and CTD refold within the bacterial cytoplasm but stay anchored to the bacterial membrane by the linker. (G) Once transcription of early promoters is complete, the autoinhibition of transcription by gp50-CTD prevents unnecessary transcription from happening.