Extended Data Fig. 1: Heat shock and oxidative stress illicit similar but distinct ISR translation programs.

(A) Quantification of ribo-spike. To confirm that ribo-spikes functioned as intended, cytoplasmic extract from HCT116 cells was serially diluted to achieve the indicated relative concentrations, and then a constant amount of ribo-spike was added to each sample before ribosome profiling and RNA-seq analyses. Plotted are the relative concentrations of total mRNA (dashed line) and RPFs (solid line) as calculated from the 1000 most highly expressed genes, using no-spike controls (blue) and ribo-spike normalized samples (red) as a function of the actual relative concentrations. Spike-normalized RPF and mRNA concentrations closely matched the known dilutions. (B) Relative TE measurements before adjusting for ribo-spike. Plotted with circles are the average log₂ TE values (line, median; notch, 95% confidence interval; box, quartiles; whiskers, 1.5 x IQR) calculated for mRNAs from either untreated cells (black, three replicates), cells treated with 500 µM sodium arsenite for 1 h (green, four replicates), or cells treated with heat shock at 45 °C for 25 min (pink, 2 replicates). Red Xs denote the ribo-spike TE values (one for each replicate). Statistical significance was calculated based on ribo-spike values alone (***, p \(\le \,\)0.001; Welch’s two-sample t-test). (C) TE measurements of panel B after adjusting for ribo-spike; otherwise, as in (B). Note that statistical significance is determined in panel B, using ribo-spike values prior to adjustment (***, p \(\le \,\)0.001). (D) Preferential translation of mRNAs containing uORFs during arsenite stress. The plots indicate log₂ FC in TE caused by arsenite for mRNAs containing at least one uORF and those containing no uORFs, according to annotations from Chen et al.⁴² (line, median; notch, 95% confidence interval; box, quartiles; whiskers, 1.5 x IQR); ***, p \(\le \,\)0.001; Welch’s two-sample t-test). (E) Relationship between translation during arsenite stress and mRNA half-lives. The plots indicate the log₂ FC in TE due to arsenite as a function of the mean-normalized half-life (PC1) of mRNAs as calculated from consensus of multiple human cell data sets by Agarwal et al.⁴⁵ n indicates number of unique mRNAs. (F) Preferential translation of mRNAs containing uORFs during heat shock. Plots indicate log₂ FC in TE caused by treatment with 45 °C heat shock; otherwise, as in (D). (G) Relationship between translation during heat shock and mRNA half-lives. The plots indicate the log₂ FC in TE due to heat shock as a function of the mean-normalized half-life of mRNAs; otherwise, as in (E). (H – J) Comparison of arsenite stress and heat shock ISR translation programs. Plotted for each mRNA are changes in TE caused by arsenite as a function of changes in TE caused by heat shock. Points for mRNAs classified as either stress-enhanced, stress-resistant, or stress-hypersensitive mRNAs in either arsenite stress (D), heat shock (E), or both (F) are coloured. n indicates number of unique mRNAs; R_s indicates Spearman coefficient.