Folding free energies of 5'-UTRs impact post-transcriptional regulation on a genomic scale in yeast

Abstract

Using high-throughput technologies, abundances and other features of genes and proteins have been measured on a genome-wide scale in Saccharomyces cerevisiae. In contrast, secondary structure in 5'-untranslated regions (UTRs) of mRNA has only been investigated for a limited number of genes. Here, the aim is to study genome-wide regulatory effects of mRNA 5'-UTR folding free energies. We performed computations of secondary structures in 5'-UTRs and their folding free energies for all verified genes in S. cerevisiae. We found significant correlations between folding free energies of 5'-UTRs and various transcript features measured in genome-wide studies of yeast. In particular, mRNAs with weakly folded 5'-UTRs have higher translation rates, higher abundances of the corresponding proteins, longer half-lives, and higher numbers of transcripts, and are upregulated after heat shock. Furthermore, 5'-UTRs have significantly higher folding free energies than other genomic regions and randomized sequences. We also found a positive correlation between transcript half-life and ribosome occupancy that is more pronounced for short-lived transcripts, which supports a picture of competition between translation and degradation. Among the genes with strongly folded 5'-UTRs, there is a huge overrepresentation of uncharacterized open reading frames. Based on our analysis, we conclude that (i) there is a widespread bias for 5'-UTRs to be weakly folded, (ii) folding free energies of 5'-UTRs are correlated with mRNA translation and turnover on a genomic scale, and (iii) transcripts with strongly folded 5'-UTRs are often rare and hard to find experimentally.

Figure 1. Structure of Yeast mRNA

(A) The mRNA has a tripartite structure consisting of a 5′-UTR, a coding region, and a 3′-UTR. CRE, cis-acting regulatory element; m7G, 7-methyl-guanosine cap; SS, secondary structure. (B) The computed minimum free energy secondary structure for the 5′-UTR of the gene YBR296C-A.

Figure 2. Folding Free Energies of 5′-UTRs

(A) Cumulative distributions of folding free energies, ΔG, are shown for 5,888 ORFs for 5′-UTRs (50 nt upstream of the ORF; solid line), 3′-UTRs (50 nt downstream of the ORF; dashed-dotted line), coding sequences (50-nt sequences following downstream of the start codon of each ORF; dotted line), and 5,888 sequences of length 50 nt selected randomly from intergenic regions (dashed line). (B) Distribution of Z-scores for 5′-UTRs of 5,888 ORFs. Each 5′-UTR sequence was shuffled 100 times and a Z-score was calculated for each to compare the folding free energy of the native sequence to the shuffled sequences. A histogram of these Z-scores is shown together with a standard normal distribution (dashed line).

Figure 3. Comparison between Ribosome Densities and Folding Free Energies of 5′-UTRs

(A) Scatter plot of mRNA ribosome density and folding free energy of the 5′-UTR (ΔG) for 5,888 ORFs. (B) ORFs were grouped based on the change in free energy (ΔG). For each energy group, the average ribosome density (±SEM) is shown. From left to right, the number of ORFs in each energy group used to calculate the average density was 573, 796, 1,214, 1,438, and 1,187.

Figure 4. Comparison between mRNA Half-Lives and Folding Free Energies of 5′-UTRs

(A) Scatter plot of mRNA half-life and folding free energy of the 5′-UTR (ΔG) for 5,888 ORFs. (B) ORFs were grouped based on the folding free energy (ΔG). For each energy group, the average mRNA half-life (±SEM) is shown. From left to right, the number of ORFs in each energy group used to calculate the average density was 467, 657, 982, 1,169, and 983.

Figure 5. Correlations between Decay and Translation Rates

Pearson correlations together with corresponding p-values are shown for mRNA half-life versus (A) ribosome density and (B) ribosome occupancy. ORFs were, depending on mRNA half-life, grouped into all ORFs, 1,058 slowly decaying ORFs with t _1/2 ≥ 33 min, and 1,013 fast decaying ORFs with t _1/2 ≤ 13 min.