Disome-seq reveals widespread ribosome collisions that promote cotranslational protein folding

Figures

• 

  Fig. 1. Disome-seq detects ribosome collisions. a A schematic explaining why traditional ribo-seq could miss the information on ribosome collisions and why the investigation of disome footprints may provide unique information on translation. b The disome persisted after RNase I digestion. Sucrose gradient profiles of the ribosome-mRNA complexes without (black) and with (green) RNase I digestion are shown. The x-axis displays the positions in the 5–50% sucrose gradient. The y-axis indicates the RNA abundance inferred from UV absorption (OD254). RNP, free ribonucleoprotein. c The schematic of disome-seq for 3-AT treated yeast cells. Briefly, ribosomes were extracted in a lysis buffer and digested with RNase I. Extracted RNA was separated on a polyacrylamide gel. RNA fragments with the length of approximate 20–30 nts or 50–80 nts were subjected to high-throughput sequencing. d The length distribution of disome footprints obtained from 3-AT treated yeast cells. Footprints are shown in different colors according to the coding frame of its 5′-end (the top panel). The average footprint abundance of two replicates is shown, in the unit of reads per million (RPM). e Determination of the…

• 

  Fig. 2. Disome-seq detects the translational pauses missed in monosome-seq. a The schematic of disome-seq for yeast cells cultivated in the rich medium. b The length distribution of disome footprints obtained from yeast cells cultivated in the rich medium. Similar to Fig. 1d. c Widespread ribosome collisions were detected in the yeast genome. Numbers of translated genes (genes with at least one monosome footprint) with and without disome footprints are shown in orange and gray, respectively. Footprints in two biological replicates were combined. 589,461 monosome footprints were used to define translated genes, and the disome-seq footprints were down-sampled to 18,082 footprints in order to match the population ratio between monosomes and disomes (32.6:1). d More frequent ribosome collision was observed in the gene with higher monosome density. e The translational pauses detected from monosome footprints (blue) and disome footprints (orange) rarely overlapped. Only the codon site with the footprint abundance greater than the mean of the corresponding gene was considered as a translational pause. If the A-site of a monosome footprint overlaps the A-site of the leading ribosome of a disome…

• 

  Fig. 3. Ribosome collisions occur at stop codons. a Two genes (RPS14A and RPL15A) exemplify ribosome collisions at stop codons. b The schematic of the calculation of the A-site pausing score. We counted the numbers of disome (or monosome) footprints with the concerning codon (GAA as an example here) or the other codons (all 63 non-GAA codons) at their A-sites, respectively. The mRNA-seq reads were used to control for the codon frequency in the transcript. The A-site pausing score was defined as the common odds ratio among genes, and the P value was given by the Mantel-Haenszel test. c The scatter plot shows the A-site pausing scores estimated from cells treated with 3-AT or growing in the rich medium, for the 61 codons that encode amino acids. The dashed line shows y = x. d The scatter plot shows the A-site pausing scores estimated from disome-seq or monosome-seq, for cells growing in the rich medium. The dashed line shows y = x. e Aggregated profiles of footprint densities over 3527 (58, 59, 61, and 62-nt footprints in disome-seq, in orange) and 5230 (28 and 29-nt footprints in monosome-seq, in blue) genes, normalized against the CDS length, are shown

• 

  Fig. 4. Ribosome collisions occur at the amino acids that can terminate α-helices. a Volcano plots show the P-site pausing score (based on disome and monosome footprints, respectively) of each amino acid and the corresponding Q value (false discovery rate). Six amino acids with > 1 P-site pausing scores in disome-seq are labeled. P, proline; G, glycine; N, asparagine; K, lysine; C, cysteine; R, arginine. b The scatter plot shows the P-site pausing scores estimated from disome-seq or monosome-seq. The dashed line shows y = x. c The scatter plots show the P-site pausing score of an amino acid and the ratio of the whole-genome occurrence of the amino acid in gaps vs. in α-helices. The P value was given by linear regression. The dashed line represents the linear regression line (if statistically significant). The residuals of both linear models followed the normal distribution (P = 0.63 and P = 0.80, the Kolmogorov-Smirnov tests). d The scatter plots show the P-site pausing score of an amino acid and the propensity to terminate an α-helix by the amino acid (ΔG relative to alanine, in the unit of kJ/mol). The P value was given by linear regression. The dashed line represents the linear…

• 

  Fig. 5. Ribosome collisions are enriched at polylysine in the exit tunnel. a The schematic of calculating the exit-tunnel pausing score. For each gene, we counted the numbers of the concerning amino-acid 3-mer (KKK as an example here) or the other 3-mers (non-KKK) in the 20 amino-acid region upstream of the P-site of the paused ribosome. mRNA-seq data were used as the background. The exit-tunnel pausing score was defined as the common odds ratio among genes, and the P value was given by the Mantel-Haenszel test. b The volcano plots show the exit-tunnel pausing score (based on disome or monosome footprints) of each of the 8000 amino-acid 3-mers and the corresponding Q value. Some 3-mers of consecutive positively charged amino acids are labeled in orange. c Determination of elongation rate with the dual-luciferase assay, in which the variants of different codon combinations for triple-lysine were inserted between Renilla and firefly-luciferase. Data are shown as the mean ± standard error (N = 11 biological replicates for each variant). ANOVA stands for the analysis of variance. d Similar to Fig. 5b, the volcano plot shows the exit-tunnel pausing score for disome footprints of each of the…

• 

  Fig. 6. The cryo-EM structure of disomes. a A composite map of the cryo-EM structure of disomes. The disomes were collected without cycloheximide in the lysis buffer. Both the leading and the trailing ribosomes are in the rotated state harboring hybrid A/P (blue) and P/E tRNAs (green). Small subunits are labeled by orange and yellow, and large subunits are labeled by cyan and gray. A model of mRNA is highlighted in red. b The top views on disomes (top) and di-ribosomes (middle). The structure of di-ribosomes was from Ikeuchi et al. [9]. An overlay is shown (bottom) by docking di-ribosomes to disomes, aligned according to the trailing ribosomes; the conformational difference in the leading ribosomes is highlighted by the arrow. c–e The zoomed-in detail of the contact interface between the 40S subunits of the leading and the trailing ribosomes, in particular, the 40S head-to-head contact site of disomes (d) and that of di-ribosomes (e). The blue star indicates a strong interaction between the two ribosomes in di-ribosomes, and the red crosses indicate the missing interactions in disomes

• 

  Fig. 7. Collided ribosomes are associated with specific chaperones. a The schematic of detecting the disome-enriched proteins through SILAC. Two biological replicates were performed with reciprocal labeling (the left panel). Because equal amount of total protein extracted from the monosome fraction and disome fraction were mixed, proteins having the same abundance per ribosome between monosomes and disomes should be located on the diagonal (the right panel). Proteins locating above or below the diagonal should be those enriched in the disomes or the monosomes, respectively. b The protein intensities per ribosome in disomes were plotted against those in monosomes. Chaperones (green) are enriched in the disome, whereas most constitutive ribosomal proteins (pink) show the same abundance per ribosome between monosomes and disomes. c The m/z (mass divided by charge number) spectrum of Ssb1p and its fragments detected in MS/MS. d The enrichment of chaperones in disomes. Immunoblotting was performed using anti-FLAG antibody, with Ponceau S stained total protein as the loading control. Mono and Di represent proteins extracted from the monosome and disome fractions, respectively. e A schematic of…