Translation is a core cellular process carried out by a highly conserved macromolecular machine, the ribosome. There has been remarkable evolutionary adaptation of this machine through the addition of eukaryote-specific ribosomal proteins whose individual effects on ribosome function are largely unknown. Here we show that eukaryote-specific Asc1/RACK1 is required for efficient translation of mRNAs with short open reading frames that show greater than average translational efficiency in diverse eukaryotes. ASC1 mutants in S. cerevisiae display compromised translation of specific functional groups, including cytoplasmic and mitochondrial ribosomal proteins, and display cellular phenotypes consistent with their gene-specific translation defects. Asc1-sensitive mRNAs are preferentially associated with the translational ‘closed loop’ complex comprised of eIF4E, eIF4G and Pab1, and depletion of eIF4G mimics the translational defects of ASC1 mutants. Together our results reveal a role for Asc1/RACK1 in a length-dependent initiation mechanism optimized for efficient translation of genes with important housekeeping functions.
Overall design: Ribosome footprint profiling and matched RNA-Seq was performed on S. cerevisiae mutants with altered Asc1 function or expression. Data from additional ribosomal protein mutants, rpl23b∆, rpp1a∆, rps0b∆, and rps16b∆, was collected for comparison. Data was also collected from an ASC1 protein null mutant, asc1-M1X, in a stress condition (growth in glycerol-containing media). In each case, data was collected from two biological replicates.
Supplementary processed data files linked below:
abbreviations:
FP= ribosome footprint profiling
total= total RNA-Seq
TE= translation efficiency measurements
glycerol= yeast grown in rich media containing glycerol instead of glucose
rz= libraries made using Ribo-Zero rRNA subtraction rather than polyA selection
reprep= library made from the same biological replicate starting material, as noted, in order to obtain higher read coverage. For downstream analyses, the original and reprep libraries were pooled.
*counts.txt:
Counts per feature for each library. Unique-mapping reads only included.
*scaledcounts.txt
Counts scaled (per experiment, including replicates) using the DESeq R package.
*rpkms.txt:
Reads per kilobase per million mapped reads for each feature.
*foldchanges.csv:
log2 fold changes for all mutants (mutant vs. wild type, average values derived from two biological replicates).
*filtered_foldchanges.csv:
log2 fold changes of all mutants (mutant vs. wild type, average values derived from two biological replicates), filtered such that each comparison (i.e. mutant vs. WT total RNA-Seq) has a minimum of 128 reads. Features too lowly expressed to meet this cutoff were removed from the relevant comparison.
Less...