Ultrastructure of influenza virus ribonucleoprotein complexes during viral RNA synthesis

Abstract

The single-stranded, negative-sense, viral genomic RNA (vRNA) of influenza A virus is encapsidated by viral nucleoproteins (NPs) and an RNA polymerase to form a ribonucleoprotein complex (vRNP) with a helical, rod-shaped structure. The vRNP is responsible for transcription and replication of the vRNA. However, the vRNP conformation during RNA synthesis is not well understood. Here, using high-speed atomic force microscopy and cryo-electron microscopy, we investigated the native structure of influenza A vRNPs during RNA synthesis in vitro. Two distinct types of vRNPs were observed in association with newly synthesized RNAs: an intact, helical rod-shaped vRNP connected with a folded RNA and a deformed vRNP associated with a looped RNA. Interestingly, the looped RNA was a double-stranded RNA, which likely comprises a nascent RNA and the template RNA detached from NPs of the vRNP. These results suggest that while some vRNPs keep their helical structures during RNA synthesis, for the repeated cycle of RNA synthesis, others accidentally become structurally deformed, which likely results in failure to commence or continue RNA synthesis. Thus, our findings provide the ultrastructural feature of vRNPs during RNA synthesis.

Fig. 1. HS-AFM observation of vRNPs during RNA synthesis.

a Primer-dependent in vitro RNA synthesis using virion-derived vRNPs. RNA was synthesized in vitro using ApG or globin mRNA as a primer with 30 min incubation. As a negative control, the reaction mixture was used without primer. A mixture of eight influenza A virus vRNA segments (Pols indicates 3 polymerases, PB2, PB1, and PA) transcribed by T7 RNA polymerase was loaded in the leftmost lane (T7) for evaluation of sizes of the newly synthesized RNAs. b Inhibition of in vitro RNA synthesis by T-705RTP. RNA was synthesized in vitro using ApG in the presence of the indicated concentration of T-705RTP. All purified RNA samples were analysed on a 4% polyacrylamide gel containing 7 M urea and detected by autoradiography. Uncropped autoradiograph images are shown in Supplementary Fig. 9. c As a negative control for the HS-AFM observation, the reaction mixture omitting a primer was used. d–h Virion-derived vRNPs were subjected to in vitro RNA synthesis using ApG as a primer. After incubation for 0 min (d) or 15 min (c, e–h), samples were observed with HS-AFM. Folded and looped RNAs associated with the helical (e, f) and deformed vRNPs (g, h), respectively, were observed as indicated by arrows at the different positions in the same samples. Scale bars on all images represent 50 nm.

Fig. 2. Cryo-EM observation of vRNPs during RNA synthesis.

An in vitro RNA-synthesis reaction was performed in the presence of ApG, and was observed with cryo-EM in vitreous ice. Folded RNAs (a, arrows) and looped RNAs (b, arrows) associated with vRNPs were observed. Scale bars represent 50 nm. c, d Cryo-ET analysis of vRNPs during RNA synthesis. c Cryo-ET observations of vRNP without RNA synthesis. d Cryo-ET observations of vRNP with RNA synthesis. Left panels: Consecutive Z-projections generated from tomograms; Thickness in Z is 44 nm (c) and 88 nm (d). Right panels: 3D reconstruction of vRNP segmented from the tomograms. The vRNP and RNA are coloured in blue and red, respectively. Scale bars on all images represent 20 nm.

Fig. 3. Incorporation of Br-UTP into newly synthesized RNAs.

a–d Br-UTP was used for in vitro RNA synthesis instead of UTP and HS-AFM images were taken without (a, b) or with (c, d) adding an antibody against Br-UTP. Binding of anti-Br-UTP antibodies was confirmed on folded RNAs (c, arrows) while no binding was observed on looped RNAs (d). Section analysis of the image (c) is shown in Supplementary Fig. 5. e, f vRNPs were in vitro transcribed using UTP and anti-Br-UTP antibody was added to the mixture. Each of these results was reproduced at least three times. Scale bars, 50 nm.

Fig. 4. Production of a double-stranded RNA by vRNP.

a, b Digestion of looped RNAs with RNases. During HS-AFM observation of looped RNA associated with vRNP, RNase A (a) or RNase III (b) was added to the liquid chamber at a final concentration of 0.5 μg mL⁻¹ or 0.02 U μL⁻¹, respectively. Five images were arbitrarily selected from each movie at the indicated times. One end of the looped RNA was detached from vRNP by adding RNase A at the position indicated by arrows (a). By contrast, RNase III digested looped RNA where the RNase bound (b, arrows). Scale bars represent 100 nm. c, d Binding of anti-dsRNA antibodies to RNA associated with the vRNP. Antibodies bound to looped RNA (c) and to folded RNA (d) are indicated by arrows. Results were reproduced at least five times. Scale bars in c, d represent 50 nm. e Detection of dsRNA in virus-infected cells by IFA. Vero cells were infected with influenza virus PR8 strain at MOI of 1. Infected cells were fixed at 10 h post-infection and double-stained with anti-NP and anti-dsRNA antibodies. Cell nuclei were stained with Hoechst. Scale bars, 20 μm.

Fig. 5. incorporation of EUTP into looped RNA.

a Confirmation of the incorporation of EUTP into looped RNA using Click chemistry. Streptavidin molecules binding to looped RNA are indicated by arrows. b Negative control of the Click reaction. The sample was prepared using UTP instead of EUTP. Each of these results was reproduced at least three times. Scale bars, 50 nm.

Fig. 6. Deformation of vRNP by releasing the residential vRNA.

a Deformation of vRNPs with the AFM probe tip. vRNP without a nascent RNA, with a folded RNA or with a looped RNA was deformed by applying force with the cantilever tip during the HS-AFM observation. As a control for the vRNP lacking its intact vRNA, the vRNP pre-treated with 0.05 μg mL⁻¹ of RNase A was also deformed. When the vRNP was confirmed as deformed (arrows), the force was measured as described in the Methods. Image sets are representative of 5 vRNPs of each sample and average forces required for deforming vRNPs are calculated. b Structural stability of vRNP during RNA synthesis. The force required for deforming the vRNP without nascent RNA was set as 100% and the relative force of each sample is shown. Significance was determined using the Tukey-Kramer multiple comparison test in R software. P < 0.05 was considered statistically significant. Error bars represent the standard deviation of five independent measurements.

Fig. 7. Model for synthesis of nascent viral RNAs by influenza vRNPs.

When folded viral RNA is synthesized, the vRNP keeps its helical rod-shaped structure and the vRNP is used in next round of RNA synthesis (upper). In contrast, when looped dsRNA is produced, the vRNP disrupts its helical rod-shaped structure because it loses the residential vRNA. As a result, such deformed vRNP cannot proceed to the next round of RNA-synthesis cycle (lower).