Quantitative profiling of basal and stress-induced ribosome collisions

Figures

• 

  Figure 1.. Spike-in incorporated Ribo-seq/Disome-seq estimates disome percentage in yeast and human cells. (A) Scheme of the spike-in incorporated Ribo-seq/Disome-seq experiment. Yeast and human cell lysates were digested with RNase I, resulting in monosome and disome footprints. After these footprints were recovered, a mixture of synthetic spike-in oligos matching the sizes of monosome and disome footprints (30mers and 60mers, shown in purple) were added prior to size selection. The footprints were then separated during size selection and Ribo-seq/Disome-seq libraries were prepared in parallel. (B) Computational pipeline used to map footprint and spike-in reads. The reads were adapter-trimmed and demultiplexed. When indicated, the reads were downsampled to a defined number of reads and were subsequently mapped to non-coding RNAs. Unmapped reads were either deduplicated and mapped to genome (yeast) or transcriptome (human), or were mapped to spike-ins with 2 mismatches (mm) (pipeline 3). The mapped reads were then deduplicated to generate spike-in reads of pipeline 2. These reads were trimmed to 30 or 60 nt (Ribo-seq and Disome-seq spike-ins, respectively) and then mapped to spike-ins with…

• 

  Figure 2.. Analysis of high-resolution Disome-seq reveals constitutive disome peaks in human cells. (A) Metagene analysis of Ribo-seq (gray) and Disome-seq (black) average normalized reads corresponding to region surrounding the start (top) and stop (bottom) codons. The shaded error bar indicates mean ± standard deviation from 3 biological replicates. Y axes indicate the distance of start or stop codon from the 5’ end of the footprints. (B) Correlation of Ribo-seq and Disome-seq reads (rpkm) per each gene (Pearson R2=0.55). The transcripts with more enriched disomes are indicated in red. (C) Ribo-seq (top, gray) and Disome-seq (bottom, black) snapshots of RPL39, SELENOK, and XBP1-unspliced transcripts. The stalling sequence spanning the disome peak is indicated as codons and corresponding amino acids under each snapshot. (D) The scheme and formula of disome peak score calculation, which is the ratio of disome peak reads to the average reads of the corresponding gene. Average reads exclude 0 values. An example of mildly enriched (disome peak score>2), and strongly enriched (disome peak score>10) peaks are indicated with red dashed line. (E) pLogo analysis showing the consensus amino acid…

• 

  Figure 3.. Disome-inducing stress in yeast changes distribution of disomes in a stress-specific manner (A) Western blotting of yeast cells treated with 0.1% MMS for 30 minutes. (B) Ribo-seq expression of GCN4 ORF in WT, hel2Δ, or WT+0.1% MMS cells. The error bar indicates mean ± standard deviation of two biological replicates. One-way ANOVA with uncorrected Fisher’s LSD was used for statistical analysis, p-values are indicated. (C) Normalized disome percentages in WT, hel2Δ, or WT+0.1% MMS cells (all downsampled to 10x106). The error bar indicates mean ± standard deviation of two biological replicates. One-way ANOVA with uncorrected Fisher’s LSD was used for statistical analysis, p-values are indicated. (D) The scheme and formula of pause score calculation for tri-amino acid motifs consisting of −1, 0, and +1 positions (E, P, A sites, respectively). Pause scores are calculated by dividing the average of reads at a specific motif to the background reads within a defined window (50 nt surrounding the motif of interest). (E) Disome-seq pause score distributions comparing WT and hel2Δ (left), or WT with WT+0.1% MMS (right). Average of two biological replicates was plotted for n=6267 motifs.…

• 

  Figure 4.. Anisomycin-induced disome formation causes loss of baseline disomes and new disome formation in a context-specific manner (A) Western blotting of human HEK293 cells treated with 1 μg/mL ANS for 15 min. Vinculin serves as a loading control. (B) Normalized disome percentages in mock- or ANS-treated HEK293 cells (downsampled to 10x106). The error bar indicates mean ± standard deviation of 2 biological replicates. One-way ANOVA with uncorrected Fisher’s LSD was used for statistical analysis, p-values are indicated. (C) Disome-seq read length distribution from mock- and ANS-treated HEK293 cells. The error bar indicates mean ± standard deviation of 2 biological replicates. (D) Metagene analysis of Disome-seq average normalized reads corresponding to region surrounding the start (top) and stop (bottom) codons in mock- and ANS-treated cells. The shaded error bar indicates mean ± standard deviation from 2 biological replicates. Y axes indicate the distance of the start or stop codon from the 5’ end of the footprints. (E) Disome-seq pause score distributions comparing mock and ANS. Average of two biological replicates was plotted for n=8000 motifs. Select sequence motifs are indicated in…

• 

  Figure 5.. Model of altered distribution of disomes upon cellular stress At basal conditions, there is low level of baseline disome formation, which we estimate to be ~2% in yeast and ~10% in human cells, though this value could be affected by several technical factors described in the Discussion section. Some of these constitutive disomes are strongly enriched and occur at prominent ribosome stalling motifs. In the presence of disome-inducing stress, the global disome levels increase. This alters distribution of disome peaks, where the heights of the constitutive disome peaks at the most prominent stalling sites are reduced. Instead, new stress-induced disome peaks appear, either at specific sequences if the stressor has a sequence context, such as 3-AT or ANS, or transcriptome-wide with no context, such as MMS. Spike-in incorporated Ribo-seq/Disome-seq method enables simultaneous detection of global disome levels and local disome signatures at transcript level.