Fig 1.
Analysis of AUG triplets in regions flanking the enterovirus SL-VI AUG.
(A) Schematic representation of the enterovirus genome with indicated features: 5′ and 3′ untranslated regions (UTRs), internal ribosome entry site (IRES), and stemloop VI (SL-VI). Below are the most frequent non-SL-VI AUG patterns in the 20 nt 5′ and 20 nt 3′ of the SL-VI AUG. AUG configurations present in ≥5 of the 9347 sequences analyzed are shown; see S1–S3 Figs for the complete analysis. (B) Histograms of SL-VI AUG and non-SL-VI AUG ORF lengths. Histogram bins are in increments of 5 codons from 1–5 up to 136–140; the last two bins are for lengths 141–2000 and >2000 codons; ORF lengths in the last bin correspond to cases where the upstream-AUG-initiated ORF is contiguous with the ppORF (often rhinoviruses). ORF lengths are counted for each relevant AUG even if there are multiple in-frame AUGs.
Fig 2.
E. alphacoxsackie sequences that encode a UP-like protein initiated at an alternative upstream AUG codon.
(A) Nucleotide (left) and uuORF-encoded protein (right) sequences in enterovirus A89/A76/A90/A121/A91, where the uAUG-initiated uORF is truncated and the uuAUG-initiated uuORF can potentially rescue UP expression. Transmembrane helix (TMH) predictions are highlighted in yellow (50–80% confidence) or green (>80% confidence). Two ORFs are indicated in purple (uuORF) and blue (uORF) highlighting. (B) Schematic representation of the EV-A90 (JX390656) IRES dVI region with uORF (blue), uuORF (purple), and ppORF (orange) start and stop codons annotated.
Fig 3.
E. coxsackiepol sequences that potentially encode a UP-like protein from an alternative upstream AUG codon.
Nucleotide (left) and uuORF-encoded protein (right) sequences in enterovirus CVA19/CVA1/C113/C116/CVA22, where the uORF is truncated and the uuORF can potentially rescue UP expression. Transmembrane helix (TMH) predictions are highlighted in orange (20–50% confidence), yellow (50–80% confidence) or green (>80% confidence). Two ORFs are highlighted in purple (uuORF) and blue (uORF). uSTOP* indicates stop codons in the uORF for the last two sequences; the remaining ten have stop codons further downstream, resulting in a 17–18 aa peptide.
Fig 4.
Translation initiation in enterovirus CVA13.
(A) Schematic representation of the CVA13 IRES dVI region with the uORF (blue), uuORF (purple) and ppORF (orange) start and stop codons annotated. (B) Analysis of viral protein expression in HeLa cells infected with CVA13. Cells were infected at an MOI of 10, harvested at 0–18 hpi as indicated, and accumulated virus structural protein VP3 was analyzed by western blotting with the anti-VP3 antibody. Observed cytopathic effect (CPE) of virus infection is indicated by (−) absence, (+) <40% CPE, (++) 40–80% CPE, (+++) >80% CPE. (C) Ribosome profiling of CVA13-infected cells at 5 and 7 hpi in the presence of translation initiation inhibitor lactimidomycin. (D) Ribosome profiling of CVA13-infected cells at 5 and 7 hpi without lactimidomycin treatment. (C-D) Ribo-Seq RPF densities (mapping positions of 5′ ends of reads with a + 12 nt offset to indicate the approximate P site) in reads per million mapped reads (RPM). Colors indicate the triplet phase of reads relative to the genome start. Since most read 5′ ends map to the first nucleotide of codons (S5D Fig), reads deriving from translation in the uuORF (+1), uORF (+2) and ppORF (0) are expected to be predominantly purple, blue and orange respectively. The amino acid sequences of the predicted proteins/peptides encoded by the three ORFs are displayed underneath the profiles. The highlighted green region indicates the predicted transmembrane domain of the UP protein encoded by the uORF. Full-genome plots and Ribo-Seq quality control plots are provided in S5 Fig.
Fig 5.
Usage of both upstream AUGs in different enterovirus IRES reporters.
(A) Schematic representation of the modified pSGDluc expression vectors used to measure translation at the polyprotein (ppORF, orange) and two upstream (blue and purple) AUG codons. (B-D) Analysis of IRES activities for ppORF, uORF and uuORF expression relative to cap-dependent expression by dual-luciferase reporter assay in HeLa and HEK293T cells (FFLuc/RLuc) at 8 h post-transfection for CVA13 (B), CVA1 (C) and EV-A90 (D) enterovirus IRESes. IRES activities in each of the three frames were normalized to cap-dependent signal and presented as relative activities in the three frames (means ± SD, n = 3 biologically independent experiments). Each panel contains a schematic representation of the corresponding IRES reporters (left). The full annotated sequence of the region can be found in S6C Fig. Statistical analysis of the difference between the translation efficiency of frames +1 and +2 was conducted using two-tailed t-tests; * p value ≤ 0.05.
Fig 6.
Development of the reverse genetics (RG) plasmid for CVA13 and properties of the uuORF knockout virus in HeLa cells.
(A) Schematic representation of the CVA13 infectious clone with the zoomed-in SL-VI/uORF region indicating the uuKO mutation (AUG → GUG). (B) Multistep growth curves of CVA13 viruses in HeLa cells (MOI 0.1). (C) Analysis of VP3 and UP expression in CVA13-infected HeLa cells. Cells were infected at MOI 10, harvested at the indicated time post-infection, and analyzed by western blotting. (D) CVA13 WT and uuKO virus competition was performed in HeLa cells. Cells were infected with WT and uuKO mutant viruses at the indicated proportions in four independent experiments and passaged six times at MOI 0.01. The virus RNA from passages 1 and 6 was isolated, RT-PCR amplified, and sequenced. The proportion of A579(WT)/G579(uuKO) nucleotide abundance was determined for each sample.
Fig 7.
Properties of the uuORF knockout virus in the neuronal and intestinal organoid infection models.
(A) Differentiated iPSC-derived neurons were infected in quadruplicate at MOI 0.1, and the released virus was analyzed by plaque assay in HeLa cells. (B) CVA13 WT and uuKO virus competition was performed in differentiated iPSC-derived neurons. Neurons were infected at MOI 0.1 with WT and uuKO mutant viruses at the indicated proportions in 12 independent experiments and incubated for 6 days with media changes every other day. The virus RNA from input and post-competition samples was isolated, RT-PCR amplified, and sequenced. The proportion of A579(WT)/G579(uuKO) nucleotide abundance was determined for each sample and presented as a doughnut plot (left, mean values) and box and whiskers plot with boxes extending from the 25th to 75th percentiles, and whiskers ranging from the smallest to largest values (right, median values). (C) Monolayers of differentiated cultured intestinal organoids were infected in quadruplicate with CVA13 viruses at MOI 10, aliquots of culture media were collected at the indicated time points, and viral titers were analyzed by plaque assay in HeLa cells. (D) CVA13 WT and uuKO virus competition was performed in differentiated human intestinal organoids. Organoid monolayers were infected at MOI 1 with WT and uuKO mutant viruses at the indicated proportions in 10 independent experiments and incubated for 4 days with media changes after 24 and 48 hpi. The virus RNA from input and post-competition (96 hpi) samples was isolated, RT-PCR amplified, and sequenced. The proportion of A579(WT)/G579(uuKO) nucleotide abundance was determined for each sample and presented as a doughnut plot (left, mean values) and box and whiskers plot with boxes extending from the 25th to 75th percentiles, and whiskers ranging from the smallest to largest values (right, median values). The data in (A) and (C) represent two biologically independent experiments; ns, nonsignificant using nonlinear regression analysis. The sequencing of the RT-PCR product derived from virus collected at 96 hpi (neuron) and 72 hpi (organoid) is provided in S7 Fig. Statistical analysis (B, D) was performed using two-tailed t-tests; ** p value ≤ 0.01; * p value ≤ 0.05; ns, nonsignificant.
Fig 8.
Mechanistic roles of 5′ and 3′ elements in IRES-dependent translation.
(A) Schematic representation of the 5′ UTR and 5′-3′ UTR reporters used to measure translation in the three frames. (B) Analysis of IRES activities for CVA13 reporters in the three frames in the 5′ UTR and 5′-3′ UTR reporters in HeLa and intestinal HIEC6 cells. Fold changes between reporter activities are indicated above each pair. (C) Schematic representation of the three CVA13 ORFs and the uuORF knock-out introduced via AUG → GUG mutation. (D) Analysis of IRES activities for CVA13 5′-3′ UTR reporters in the three frames in HeLa and intestinal HIEC6 cells with and without virus RNA at 8 h.p.t. The full set of 5′ UTR and 5′-3′ UTR reporter data can be found in S8. (E-F) Fig Schematic representation of the CVA1 (E) and EV-A90 (F) ORFs and AUU mutants used to measure translation in three frames in HeLa and HIEC6 cells. Statistical analysis was conducted using two-tailed t-tests; * p value ≤ 0.05; ns, nonsignificant. The full annotated sequence of the region can be found in S10 Fig. (G) Schematic representation of uORF-dependent translation in eukaryotic ATF4 mRNA. (H) Analysis of IRES activities for CVA13 reporters in the ppORF and +1 uORF frames in response to sodium arsenite-mediated stress. The normalized cap-dependent translation corresponds to the Renilla luciferase signal for the indicated reporter. Statistical analysis (E, F, H) was performed using two-tailed t-tests; * p value ≤ 0.05; ns, nonsignificant. (I) Schematic representation of uORF-preferred translation in CVA13 during stress conditions. (E,F,I) The SL-VI structure is shown for the context of the start and stop codon positions; it is unwound during translation [29].