Investigating molecular mechanisms of 2A-stimulated ribosomal pausing and frameshifting in Theilovirus

Figure 1.. TMEV 2A adopts the beta-shell fold. (A) SDS-PAGE analysis of TMEV 2A after Ni-NTA, heparin affinity and size-exclusion chromatography. The gel was stained with Coomassie blue. (B) SEC-MALS analysis of 2A in high-salt buffer at 3.1 mg/ml. The differential refractive index curve (blue) is shown and the weight-averaged molar masses for indicated peaks are listed. (C) X-ray crystal structure of TMEV 2A in two orthogonal views. N- and C- termini are indicated, and amino acids introduced as a result of the cloning strategy are labelled in purple. <Inset >Electrostatic surface potential calculated at pH 7.4, coloured between +3 (blue) and −3 kT/e– (red). (D) Locations of mutations are colour-coded and shown as sticks: M1 (K24A / R28A, green); M2 (R45A, blue); M3 (R85A / R95A, red); M4 (K90A / K91A, pink). (E) Superposition of TMEV 2A (blue) with EMCV 2A (orange). Insets:Close-up views of regions of interest including the divergent structure at the C-terminus, the longer α1-β4 loop present in EMCV 2A and the conserved arginine loop.

Figure 2.. Comparison of the TMEV 2A structure with the EMCV orthologue reveals a divergent RNA-binding surface and C-terminal region. (A) Amino acid sequence alignment of selected divergent TMEV-like (upper) and EMCV-like (lower) cardiovirus 2A protein sequences, guided by the structural alignment of EMCV and TMEV 2A proteins. Known secondary structures are indicated above the corresponding sequence for EMCV and TMEV proteins. N- and C-terminal residues not observed in the structures are greyed out. Conservation is highlighted in blue. Local motifs of functional significance are highlighted and annotated. (B) Comparison of RNA-binding residues in EMCV 2A (28) (beige) with equivalent surface in TMEV 2A (pale blue). Residues involved in RNA binding come from three regions of the EMCV protein as indicated in (A) (shown as red, crimson and brown sticks). TMEV residues that may be functionally-equivalent are shown as sticks (cyan). (C) Surface of TMEV 2A, coloured by conservation from highly conserved (purple) to variable (teal) amongst the divergent TMEV-like sequences listed in A. Highly conserved, surface-accessible residues are labelled and shown as sticks. (D) Comparison of the YxxxxLΦ…

Figure 3.. 2A recognises a minimal 37 nt stimulatory element in the viral RNA. (A, B) Sequences and schematic diagrams of the TMEV 1–6 constructs used to assay 2A binding. (C) EMSA analysis conducted with 50 nM Cy5-labelled RNAs and 2A concentrations between 0 and 28 μM. Following non-denaturing electrophoresis, fluorescence was imaged using a Typhoon scanner. (D) MST binding curves and apparent KD values using the same constructs. (E) Experiment showing the effects of titrating excess short RNAs (TMEV 1–6) as competitors into an in vitro frameshift reporter assay. The concentrations of the reporter mRNA and 2A were kept constant in the RRL and short RNAs were added in 10- and 100-fold molar excess relative to the reporter mRNA, as indicated. Translation products were visualised by using autoradiography, and % frameshifting was calculated following densitometry and correction for the number of methionines present in 0 frame and −1 frame products.

Figure 4.. A conserved AAGG motif present in the 5′ extension is required for 2A binding. (A) Logo plot showing conservation of aligned cardiovirus RNA sequences (see Supplementary Figure S3) spanning the experimentally defined minimal element. Asterisks indicate invariant nucleotides. (B) Schematic diagrams of the TMEV 7–12 sequences used to assay 2A binding. (C) EMSA analysis conducted with 50 nM Cy5-labelled RNAs and 2A concentrations between zero and 28 μM. Following non-denaturing electrophoresis, fluorescence was imaged using a Typhoon scanner. (D) Frameshift reporter assays showing that mutation of the loop CCC motif inhibits frameshifting, but complementary mutations in the 5′ AAGG motif that allow base pairing between positions 17 and 38 restore frameshifting. (E) EMSA analysis showing attenuation of 2A binding to the minimal TMEV 9 RNA by introducing point mutations to either the 5′ AAGG motif or loop CCC motif. (F) Hypothetical equilibrium between predicted stem-loop and alternate pseudoknot-like conformations, colour-coded as in (A). The pseudoknot-like conformation involves a base-pair between G17 in the 5′ extension, and C38 in the loop (shown as sticks). In this predicted…

Figure 5.. 2A binds to the RNA stimulatory element with equimolar stoichiometry and nanomolar affinity. (A) EMSA analysis showing the effects of salt concentration (indicated) on RNA binding. Experiments were conducted with 50 nM Cy5-labelled TMEV 6 and 2A concentrations between zero and 28 μM. Following non-denaturing electrophoresis, fluorescence was imaged using a Typhoon scanner. (B) ITC isotherm for a titration of TMEV 6 RNA into 2A protein in the presence of 400 mM NaCl. (C) Binding curve from titration in (B), showing approximately 1:1 molar ratio and nanomolar affinity. Inset: Histogram showing relative contributions of ΔH and TΔS terms to the overall exergonic interaction. (D) SEC-MALS analysis of 2A-TMEV 6 RNA complex in buffer containing 200 mM NaCl. The 280 nm UV absorbance trace is shown (green). Weight-averaged molar masses across the indicated peaks are listed, along with mass contributions from the protein (blue) and RNA (pink) components, following a protein conjugate analysis. The two peaks correspond to the RNA-protein complex (peak 1; Mw 33.6 kDa) and the excess RNA (peak 2; Mw 11.0 kDa).

Figure 6.. 2A stimulates −1 PRF with 85% efficiency and leads to ribosomal pausing in infected cells. (A) Metabolic labelling of BSR cells infected with WT, SS, or M3 TMEV and harvested at 10 hpi. Positions of TMEV proteins are indicated, with those downstream of the frameshift site written in red. (B) Frameshift (FS) efficiency in infected cells, calculated from metabolic labelling (blue bars) and ribosome profiling (green bars). Bars represent the mean of three replicates, with values for each replicate indicated by crosses. (C) Schematic of ribosome profiling methodology. Cells are flash frozen before lysis. RNase I is added to digest regions of unprotected RNA, then ribosomes and enclosed ribosome-protected fragments (RPFs) of RNA are purified. RPFs are released and prepared for high-throughput sequencing to determine the positions of translating ribosomes at the time of cell lysis. (D) RPF densities in reads per million mapped reads (RPM) on WT, SS and M3 viral genomes (top panel), after application of a 15 nt running mean filter, from cells harvested at 10 hpi. Positive sense reads are plotted in green (above the horizontal axis), negative sense in red (below the horizontal axis;…

Figure 7.. Investigation of the region around the frameshift site by ribosome profiling. (A) Percentage of reads (all read lengths) attributed to each phase in the regions upstream and downstream of 2B* and in the short 2B/2B* overlap region (‘transframe’). (B) Density of RPFs in the 200 nt up to and including the frameshift site, plotted at inferred P site positions of the ribosomes, and coloured according to RPF phase as in Figure 6D. The encoded amino acid sequence in this region is underlaid beneath the data in the top panel, and all libraries are set to the same scale on the y axis, with no running mean filter applied. The last amino acid of 2A—after which StopGo-mediated peptide release occurs – is annotated in brown, the P site of frameshifting (FS) in grey, and the RNA-binding arginines mutated to alanine in M3 in black. (C) Histogram of relative adaptiveness values of each sense codon to the cellular tRNA pool, defined based on a combination of intracellular tRNA abundance and the strength of the codon-anticodon interaction. Values for the Chinese hamster Cricetulus griseus were taken from the Species-Specific tRNA Adaptive Index Compendium (57) as M. auratus values were not…

Figure 8.. Profiling of disome-protected fragments reveals ribosome stacking at the frameshift site. (A) Length distribution of reads mapping within host CDSs for broad spectrum libraries. (B) Total number of reads attributed to each phase, for reads of each specified length, in the broad-spectrum libraries. Read lengths which were selected for inclusion in the ‘disome-protected fragment’ plots (C and D) are indicated by square brackets, and correspond to peaks in (A). (C) Density of disome-protected fragments on the WT and M3 viral genomes, derived from the same lysate as the monosome-protected fragment data plotted in Figure 6D–G. Reads of lengths 51–52, 54–55 and 57–64 nt were selected for inclusion, and read densities are plotted as RPM after application of a 15 nt running mean filter. Positive-sense reads are plotted in green (above the horizontal axis), negative sense in red (below the horizontal axis). In all disome plots, reads are plotted at the inferred P site position of the colliding ribosome. (D) Density of disome-protected fragments, plotted at inferred P site positions of colliding ribosomes involved in disomes upstream of the 2B* frameshift site, coloured according to phase…