Circularization of flavivirus genomic RNA inhibits de novo translation initiation

Figures

• 

  Figure 1.. 48S complex assembly on ZIKV RNA is cap-dependent. (A) Schematic of the ZIKV RNA in a linear (upper) and circular (lower) conformation. The nucleotides corresponding to the UAR, DAR and CS are shaded in green, blue and purple, respectively. The initiating AUG and upstream near-cognate codons discussed in the text are in bold. (B) Schematic representation of the in vitro reconstitution and toeprinting assay described in the text. (C and D) Toeprinting analysis of 48S complex assembly on (C) uncapped or capped ZIKV5′utr+ RNA and (D) capped ZIKV5′utr+ RNA in the absence of the indicated factors. Selected codons are labelled on the left and toeprints caused by 48S complex assembly are marked with a closed arrowhead on the right. To balance the intensities in panel D, different exposures of the sequencing and toeprint portions of the gel are shown, indicated by the black line. FL, full-length. See also Supplementary Figure S1.

• 

  Figure 2.. Genome circularization is required for ZIKV replication. (A) Schematic illustration of ZIKVNluc. The first copy of the capsid gene is fused to a Nluc reporter and is separated from a second capsid sequence and the rest of the polyprotein by the FMDV 2A StopGo (40). ZIKVNluc-GAA and ZIKVNluc-Δ3′ CS polymerase mutations are indicated. (B andC) Time course of (B) ZIKV genome copy number quantified by qPCR and (C) luciferase activity normalized to total protein amount after electroporation of capped ZIKVNluc RNA into Vero cells. Data are mean ± SEM for three independent experiments. (D) Luciferase activity normalized to total protein amount at 6 h post-electroporation relative to the WT control (full time course shown in C). Data are mean ± SEM of three independent experiments. ZIKVNluc-GAA and ZIKVNluc-Δ3′ CS mutants were compared by Mann–Whitney test (P = 0.40). NS, not significant. (E) SHAPE reactivity in the 5′ CS of ZIKVfl (green) and ZIKVfl-Δ3′ CS (purple). Nucleotides within the 5′ CS are enclosed by dotted brackets. Data are normalized SHAPE reactivities from three experiments, mean ± SD at each base. SHAPE reactivities in the 5′ CS region of ZIKVfl and ZIKVfl-Δ3′ CS mutant…

• 

  Figure 3.. ZIKVmini RNA is circularized. (A) Schematic illustration of the ZIKVmini mutants examined. The dark and light purple represent the 5′ CS and 3′ CS, respectively. The black dotted line represents cyclization potential and the red bar represents the anti-sense oligo (ASO). (B and C) Native PAGE analysis of (B) ZIKVmini containing mutations in either the CS or UAR elements and (C) WT ZIKVmini after heating and snap-cooling in the presence or absence of ASO or scrambled oligo (SCR). (D and E) SHAPE reactivity in the 5′ CS of the WT and mutant ZIKVmini RNA. Nucleotides within the 5′ CS are enclosed by dotted brackets. WT SHAPE reactivity in light green is compared to (D) CS swap (dark green) and Δ3′ CS (purple) or to (E) WT annealed to SCR (dark green) or ASO (purple) RNA. Data are mean ± SD of normalized SHAPE reactivities at each base from three experiments. The same WT dataset is used in each panel as all conditions were examined simultaneously. SHAPE reactivities in the 5′ CS region of the WT and mutant minigenomes were compared by the Student's t-test. P-values are shown in Supplementary Table S2. (F) Heat map of SHAPE reactivities from a representative experiment overlaid onto…

• 

  Figure 4.. ZIKV genome circularization inhibits translation initiation. Toeprinting analysis of 48S complex assembly on WT capped ZIKVmini RNA and (A) ΔCS mutants, (B) ΔUAR mutants, (C) in the presence of scrambled (SCR) or anti-sense (ASO) RNA oligos and (D) upstream near-cognate codon mutants highlighted in red in the schematic on the left. Selected codons are labelled on the left of each gel and toeprints caused by 48S complex assembly are marked with a closed arrowhead on the right. Open arrowheads mark reverse transcriptase stops caused by hybridization of the 5′ and 3′ cyclization elements. FL, full-length. See also Supplementary Figure S3.

• 

  Figure 5.. Circularization inhibits translation initiation on DENV. (A and C) Schematics of the (A) DENV4 and (C) DENV1 genomes in the circular conformation. Nucleotides of the UAR, DAR and CS are shown in green, blue and purple, respectively. The initiating AUG and upstream near-cognate codons discussed in the text are in bold. (B and D) Toeprinting analysis of 48S complex assembly on capped (B) WT and Δ3′ CS DENV4mini or (D) WT and Δ3′ CS DENV1mini RNAs. Selected codons are labelled on the left and toeprints caused by 48S complex assembly are marked with a closed arrowhead on the right. Open arrowheads mark reverse transcriptase stops caused by hybridization of the 5′ and 3′ cyclization elements. Scrambled (SCR) or anti-sense (ASO) RNA oligos were annealed to DENV1/4mini RNAs prior to the assembly reaction as indicated. FL, full-length.

• 

  Figure 6.. Model for the control of translation initiation by genome circularization. Linear genomic RNA recruits eIFs and 40S ribosomal subunits (pale grey) in a cap-dependent manner. Secondary structure in the 5′ UTR is effectively resolved by helicase eIF4A to allow translation initiation from the authentic initiation codon, indicated by a green 48S complex. Following genome circularization, promoted by viral and host factors (64–68) and likely occurring at the ER membrane, scanning ribosomes are impeded by extensive RNA structure formed by hybridization of 5′ and 3′ cyclization elements. Decreased translation from the authentic AUG permits efficient genome replication by NS5pol (purple).