Ribosome Decision Graphs for the Representation of Eukaryotic RNA Translation Complexity

Figures

• 

  Figure 1.. ORFs do not adequately represent translational complexity. A. Two formal definitions of ORFs. Three reading frames are shown as horizontal bars with vertical bars corresponding to AUG (green) or stop codons (black). Several examples of Start-Stop (green arcs) and Stop-Stop (black arcs) ORFs are shown. B-D. The relationship between ORFs (top) and expressed proteoforms (bottom) for mRNAs with different locations of starts and stops (middle). Only two relevant reading frames are shown for simplicity. B. An RNA encoding two proteoforms with alternative N-termini due to utilization of two start codons. Due to multiple potential AUG codons, there are many Start-Stop ORFs whose conceptual translation does not correspond to encoded proteoforms. Stop-Stop ORF does not reflect the existence of alternative proteoforms. C. In the case of stop codon readthrough or selenocysteine insertion, ribosomes may read through specific stop codons by incorporating an amino acid yielding a product (purple) that cannot be described as a product of a single ORF. D. Similarly, ribosomal frameshifting generates a trans-frame protein (blue) that does not represent a product of a single ORF.

• 

  Figure 2.. Relationship between translated segments within the same RNA. A. A schematic of metabolite-dependent translation regulation via ribosome arrest at uORF, such as in AMD1 mRNA. In the presence of high concentration of polyamines ribosome with MAGDIS peptide stalls at the end of mRNA. B. A schematic of delayed re-initiation mechanism enabling translation of selected mRNAs during global translation suppression caused by eIF2alpha phosphorylation during Integrated Stress Response. When eIF2 is phosphorylated the concentration of the eIF2*tRNAi*GTP ternary complex decreases and it requires a longer time and distance for the scanning ribosome complex to acquire it. As a result the long uORF is bypassed and initiation occurs at CDS.

• 

  Figure 3.. Ribosome paths through mRNA. Top: An example of an RNA with three start codons (green for AUG and light green for CUG) located in two different reading frames depicted as differentially shaded horizontal bars. The translation of this mRNA is represented as a set of translons below, Ribosome Paths further below and Ribosome Decision Graphs at the bottom. RNA regions scanned by the ribosome are shown in dark grey, vertical path in light grey represents the post-terminating small ribosome subunit that continues scanning and remains initiation-competent.

• 

  Figure 4:. Representation of the effect of a genomic variant on CDS translation. Representations of the NF2 mRNA for the reference sequence (A) and in the presence of the insertion variant (B). ORF organisation is shown at the Top with reading frames shaded differentially according to the reference sequence. AUG and stop codons are represented as green and black vertical bars, respectively. RDGs are shown at the bottom. Given the low probability of leaky scanning through the first two AUGs (see text), it is expected that the translon T3 corresponding to CDS cannot be translated in the pathogenic variant sequence.

• 

  Figure 5:. Representation of translation of human mRNAs. A. The top plots are A-site subcodon footprint densities for NXT1 and NRAS mRNAs coloured to match the best-supported reading frame below in the ORF plots, in which AUGs are in white and stops are in black. Further below are RDG representations with translons coloured to match the translated frame. The CDS starts are treated deterministically and unproductive RiboPaths are not shown. B. Densities of ribosome A-site footprints obtained from ribosome profiling under different conditions (NRAS) or from different cells (NXT1). RDGs are shown below each density plot with translons coloured as a heatmap reflecting relative translation efficiencies. Branching points (starts) are also coloured as heatmaps reflecting the inferred probability of their initiation. The probabilities of translation initiation are shown as fractions decimals, while the relative translation efficiencies of each path are shown as percentages (%).

• 

  Figure 6:. RDG representations of special cases. (A) mRNA encoding selenoprotein GPX4 and (B) PEG10 mRNA requiring ribosomal frameshifting for its expression. Trips-Viz screenshots of ribosomal profiling density for these mRNAs are at the top with ORF plots beneath and RDG representations further below. C. Top: An ORF plot containing three translons (protein-coding CDS and two uORFs, short and long), which is a minimal requirement for the mechanism of delayed re-initiation. Below: The corresponding RDG with green and red representing the predominant ribosome paths for normal and stress conditions, respectively. Arrows indicate the distance sufficient for scanning ribosomes to bind ternary complex under normal (green) or stress (red) conditions.