Characterizing terminating ribosomes using Ribo-seq.

(A) Aggregation plots of mean Ribo-seq reads around the start (left) and stop (right) codons. Both 5’ end (green) and 3’ end (orange) of footprint reads (29 nt) were used for mapping. For bottom panels, the line plots show the distribution of reads with different lengths, whereas the bar plots show the read distribution of initiating (left) and terminating (right) ribosomes relative to elongation ribosomes. (B) Aggregation plots of mean Ribo-seq reads around the stop codon. The reads were stratified by the length followed by mapping using the 5’ end of footprint reads. (C) A histogram plot of ribosome pausing index at individual stop codon. The top shows the enriched sequence motif of mRNAs with strong ribosome pausing at stop codons. (D) Aggregation plots of mean Ribo-seq reads around the stop codon. “High” and “Low” refer to mRNAs with differential pausing indexes shown in (C). The 3’UTR read density was highlighted in the insert.

Characterizing terminating ribosomes using eRF1-seq.

(A) The left panel shows the schematic of eRF1-seq procedures. The right top panel shows the aggregation plots of Ribo-seq (green) and eRF1-seq (orange). The right bottom panel shows a representative mRNA (RPS2) with reads obtained from Ribo-seq and eRF1-seq. The 5’ end of reads were used for mapping. (B) Heatmaps of individual mRNAs with reads obtained from Ribo-seq (right) and eRF1-seq (left). (C) Aggregation plots of mean eRF1-seq reads around the stop codon. The reads were stratified by the length followed by mapping using the 5’ end of footprint reads. (D) Comparison of codon frequencies upstream of stop codons between mRNAs with and without eRF1 peaks. The right panel shows the enriched sequence motif for mRNAs with eRF1 peaks at stop codons.

eRF1-seq reveals prevailing termination sites.

(A) A pie chart shows fractions of eRF1 peaks mapped to different mRNA regions. “Inframe” and “Outframe” refer to the positions of eRF1 peaks within CDS relative to the annotated start codons. (B) The left panel shows mean eRF1-seq reads around the position of eRF1 peaks within 5’ UTR. The dot plot shows the frequency of A-site codons at eRF1 peaks. The right panel shows mean Ribo-seq reads around the position of eRF1 peaks within 5’ UTR. The bar graph shows the mean ribosome densities before (-) and after (+) the eRF1 peaks. (C) The left panel shows mean eRF1-seq reads around the position of out-of-frame eRF1 peaks within CDS. The dot plot shows the frequency of A-site codons at eRF1 peaks. The right panel shows mean Ribo-seq reads around the position of out-of-frame eRF1 peaks within CDS. The bar graph shows the mean ribosome densities before (-) and after (+) the eRF1 peaks. (D) The distribution of eRF1 peaks before the annotated stop codon. Only the out-of-frame eRF1 peaks were used for plotting. (E) A bar graph shows the HiBiT-based reporter assays in HEK293-Kb cells. HiBiT signals were measured from cells transfected with mRNA reporters bearing different sequences between GFP and HiBiT. Error bars, mean ± s.e.m. n = 3 biological replicates. ****P ≤ 0.0001 by unpaired two-tailed t-test.

Termination pausing is influenced by sequence contexts.

(A) The up panel shows the schematic of a massively parallel reporter assay, where the stop codon of uORF was replaced with a random 9-nt sequences. The uORF translation was monitored by the number of associated ribosomes separated by sucrose gradient (M, monosome; P, polysome). The heatmap shows the ratio of monosome fraction over polysome fraction (M/P) when different codons were placed at individual positions within the 9-nt random sequence (left). The stop codons UGA, UAG and UAA are highlighted. The bar graph (right) shows the mean M/P ratio averaged across all positions. (B) Similar as (A) except that the random 9-nt sequences were placed before the stop codon UAG. (C) Bar graphs show the HiBiT signals in HEK293-Kb cells transfected with mRNA reporters bearing C-rich or GA-rich sequences between the uORF and HiBiT-GFP. The up panel shows the control without the stop codon UAG. Error bars, mean ± s.e.m. n = 3 biological replicates. *P ≤ 0.05; **P ≤ 0.01 by unpaired two-tailed t-test. (D) A bar graph shows the HiBiT signals in HEK293-Kb cells transfected with mRNA reporters bearing C-rich or GA-rich sequence before the GFP stop codon UAG. The downstream HiBiT sequence was inserted into different reading frames. Error bars, mean ± s.e.m. n = 3 biological replicates. **P ≤ 0.01; ***P ≤ 0.001 by unpaired two-tailed t-test. (E) Representative Western blots of GFP and HiBiT in HEK293-Kb cells transfected with C-rich or GA-rich reporters as described in (D).

The 3’ end of 18S rRNA influences the dynamic of terminating ribosome.

(A) A cryoEM structures of mammalian initiating ribosomes (PDB: 6ZMW) showing the proximity of the 3’ terminus of 18S rRNA and mRNA. (B) Superposition of different conformations sampled by normal mode analysis (NMA) of the region mentioned in (A). The analysis was carried out with the mRNA segment from the -10 nt to the +5 nt related to the A-site. The superposition is outlined only for the Rps26 protein (green surface), the mRNA segment (blue surface) and the last 10 nucleotides at the 3’ rRNA extremity (dark red surface), with the remaining proteins and rRNA segments at the site described only by the average structure (silver transparent surface and cartoon). The tRNA anticodon loop at the P ribosomal site is shown in dark cyan. (C) HEK293 cells were transfected with plasmids encoding 18S rRNA WT or mutant followed by Ribo-seq. Aggregation plots show the mean reads around stop codons of mRNAs with the GA sequence motif or C-rich sequence element. (D) HEK293 cells were transfected with plasmids encoding 18S rRNA WT or mutant followed by transfection with mRNA reporters shown in (4D). Bar graphs show the HiBiT signals in transfected cells with HiBiT at different reading frames. Error bars, mean ± s.e.m. n = 3 biological replicates. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001 by unpaired two-tailed t-test.

Differential termination pausing in mouse tissues.

(A) Different mouse tissues were collected followed by Ribo-seq. Metagene analysis shows the distribution of mean ribosome reads across the transcriptome aligned to start and stop codons. (B) Representative Western blots of different mouse tissues using antibodies indicated. The experiment was independently repeated three times with similar results. (C) Mouse liver and testis were subjected to polysome profiling using sucrose gradient. Representative Western blots of ribosome fractions were conducted using antibodies indicated. The experiment was independently repeated three times with similar results.

Rps26 modulates termination pausing.

(A) A cryo-EM structure of the human 48S pre-initiation complex (PDB: 6YAL) depicted in cartoon and colored in yellow. The region at the 5’ exit of the mRNA tunnel centered in Rps26 and extended around 30 Å of this protein is highlighted in transparent surface and in dark colors. (B) Conformational sampling by normal mode analysis (NMA) of the region mentioned in (A). The analysis was carried out with the mRNA segment from the -10 nt related to the P-site to the +5 nt related to the A-site. The superposition is outlined only for the Rps26 protein (green surface), the mRNA segment (blue cartoon) and the last 10 nucleotides at the 3’ rRNA extremity (dark red cartoon), with the remaining proteins and rRNA segments at the site described only by the average structure (silver transparent surface and cartoon). The tRNA anticodon loop at the P ribosomal site is omitted on the image. (C) HEK293 cells with or without Rps26 knockdown were subjected to Ribo-seq. Aggregation plots show the mean reads around stop codons of mRNAs with the GA sequence motif or C-rich sequence element. (D) HEK293 cells with or without Rps26 knockdown were transfected with mRNA reporters shown in 4D. Bar graph shows the HiBiT signals at different reading frames upon normalization to upstream GFP levels Error bars, mean ± s.e.m. n = 3 biological replicates. ns, nonsignificant; **P ≤ 0.01; ***P ≤ 0.001 by unpaired two-tailed t-test. (E) HEK293 cells with or without Rps26 overexpression were subjected to Ribo-seq. Aggregation plots show the mean reads around stop codons of mRNAs with the GA sequence motif or C-rich sequence element. (F) HEK293 cells with or without Rps26 overexpression were transfected with mRNA reporters shown in 4(D). Bar graph shows the HiBiT signals at different reading frames. Error bars, mean ± s.e.m. n = 3 biological replicates. *P ≤ 0.05; **P ≤ 0.01 by unpaired two-tailed t-test.

Characterizing terminating ribosomes.

(A) HEK293 cells were treated with cycloheximide (CHX, 100 μg/mL) for varied times followed by Ribo-seq. Aggregation plots show the distribution of mean reads across the transcriptome aligned at start and stop codons. (B) A representative mRNA (RPS2) with reads obtained from ribo-seq in cells as (A). The 5’ end of reads were used for mapping. (C) The ratio of ribosome density in 3’ UTR over the density in CDS. The mRNAs were stratified into different groups based on the identify of stop codons and the nucleotide immediately after the stop codon. (D) Relative fractions of Ribo-seq reads mapped to different reading frames. Reads in CDS and 3’UTR were shown. (E) HEK293 cells with or without crosslinking by formaldehyde were subjected to eRF1 immunoprecipitation followed by Western blotting using antibodies indicated. Representative results of three independent experiments are shown. (F) A scatter plot shows the correlation of read counts between two biological replicates of eRF1-seq (Rho = 0.89). (G) Comparison of 3’UTR length (left) and folding free energy (right) between mRNAs with and without eRF1 peaks at the stop codon.

eRF1-seq reveals prevailing termination sites in 3’UTR and CDS.

(A) A Venn diagram shows the number of uORFs identified by Ribo-seq and eRF1-seq. (B) The left panel shows mean eRF1-seq reads around the position of eRF1 peaks within 3’ UTR. The dot plot shows the frequency of A-site codons at eRF1 peaks. The right panel shows mean Ribo-seq reads around the position of eRF1 peaks within 3’ UTR. The bar graph shows the mean ribosome densities before (-) and after (+) the eRF1 peaks. (C) The left panel shows mean eRF1-seq reads around the position of in-frame eRF1 peaks within CDS. The dot plot shows the frequency of A-site codons at eRF1 peaks. The right panel shows mean Ribo-seq reads around the position of in-frame eRF1 peaks within CDS. The bar graph shows the mean ribosome densities before (-) and after (+) the eRF1 peaks.

MPRA assays dissect the sequence context of stop codons.

(A) The up panel shows the schematic of a massively parallel reporter assay, where the stop codon of uORF was replaced with a random 9-nt sequences. The uORF translation was monitored by the number of associated ribosomes separated by sucrose gradient (M, monosome; P, polysome). The heatmap shows the codon frequency within the 9-nt random sequence in monosome (M) and polysome (P) fractions. (B) Bar graphs show the top 10 codons enriched in monosome (M) over polysome (P) at each reading frame. (C) The up panel shows the schematic of a massively parallel reporter assay, where a random 9-nt sequence was replaced after the uORF stop codon UAG. The uORF translation was monitored by the number of associated ribosomes separated by sucrose gradient (M, monosome; P, polysome). The heatmap shows the codon frequency within the 9-nt random sequence in monosome (M) and polysome (P) fractions. (D) A bar graph (right) shows the mean M/P ratio averaged across all positions in (C). (E) The up panel shows the schematic of a massively parallel reporter assay, where a random 9-nt sequence was replaced before the uORF stop codon UAG. The uORF translation was monitored by the number of associated ribosomes separated by sucrose gradient (M, monosome; P, polysome). The heatmap shows the codon frequency within the 9-nt random sequence in monosome (M) and polysome (P) fractions. (F) A bar graph shows the HiBiT signals in HEK293-Kb cells transfected with mRNA reporters bearing C-rich or GA-rich sequence before the GFP stop codon UAG. Error bars, mean ± s.e.m. n = 3 biological replicates. **P < 0.01 by unpaired two-tailed t-test. (G) A bar graph shows the HiBiT signals in HEK293-Kb cells transfected with mRNA reporters bearing C-rich or GA-rich sequence before the GFP stop codon UAG. Error bars, mean ± s.e.m. n = 3 biological replicates. *P < 0.05; **P ≤ 0.01; ***P < 0.001 by unpaired two-tailed t-test for two groups and ordinary one-way ANOVA test followed by Dunnett’s multiple comparisons for more than two groups.

ABCE1 modulates termination pausing independent of the mRNA sequence context.

(A) Western blots of HEK293-Kb cells with or without ABCE1 knockdown. (B) Comparison of polysome profiles of cells with or without ABCE1 knockdown. (C) Aggregation plots show ribosome density around stop codons in cells with or without ABCE1 knockdown. (D) Aggregation plots of mean Ribo-seq reads around the stop codon in cells with or without ABCE1 knockdown. The reads were stratified by the length followed by mapping using the 5’ end of footprint reads. (E) A bar graph shows the HiBiT signals in HEK293-Kb cells with or without ABCE1 knockdown after transfection with mRNA reporters bearing C-rich or GA-rich sequence before the GFP stop codon UAG. Error bars, mean ± s.e.m. n = 3 biological replicates. **P < 0.01 by unpaired two-tailed t-test.

The putative role of 3’ terminus of 18S rRNA in termination pausing.

(A) A schematic shows the secondary structure of 18S rRNA with the 3’ end sequence highlighted. (B) Comparison of polysome profiles of cells expressing the 18S rRNA WT or mutant. The inserts depict the nucleotide sequence of 18S rRNA WT or mutant. (C) HEK293 cells were transfected with plasmids encoding 18S rRNA WT or mutant followed by Ribo-seq. Aggregation plots show the distribution of mean reads across the transcriptome aligned at start and stop codons.

Tissue specificity of termination pausing.

(A) Different mouse tissues were collected followed by polysome profiling on sucrose gradients. (B) Mouse liver and testis were subjected to Ribo-seq. Aggregation plots show the distribution of mean reads across the common or tissue-specific transcripts aligned at start and stop codons. (C) Mouse liver and testis were subjected to Ribo-seq. Aggregation plots show the mean reads around stop codons of mRNAs with the GA sequence motif.

The role of Rps26 in termination pausing.

(A) Western blots of HEK293 cells with or without Rps26 knockdown. (B) Comparison of polysome profiles of cells with or without Rps26 knockdown. (C) HEK293 cells with or without Rps26 knockdown were subjected to Ribo-seq. Aggregation plots show the distribution of mean reads across the transcriptome aligned at start and stop codons. (D) Western blots of HEK293 cells with or without Rps26 overexpression. (E) Comparison of polysome profiles of cells with or without Rps26 overexpression. (F) HEK293 cells with or without Rps26 overexpression were subjected to Ribo-seq. Aggregation plots show the distribution of mean reads across the transcriptome aligned at start and stop codons.