Position-dependent termination and widespread obligatory frameshifting in Euplotes translation

Abstract

The ribosome can change its reading frame during translation in a process known as programmed ribosomal frameshifting. These rare events are supported by complex mRNA signals. However, we found that the ciliates Euplotes crassus and Euplotes focardii exhibit widespread frameshifting at stop codons. 47 different codons preceding stop signals resulted in either +1 or +2 frameshifts, and +1 frameshifting at AAA was the most frequent. The frameshifts showed unusual plasticity and rapid evolution, and had little influence on translation rates. The proximity of a stop codon to the 3' mRNA end, rather than its occurrence or sequence context, appeared to designate termination. Thus, a 'stop codon' is not a sufficient signal for translation termination, and the default function of stop codons in Euplotes is frameshifting, whereas termination is specific to certain mRNA positions and probably requires additional factors.

Figure 1. Frequent frameshifting in Euplotes

Ribo-seq profiles of individual mRNAs are shown in the upper panels, RNA-seq in the middle panels, and features of reading frames in the lower panels. Start (ATG, green vertical lines) and stop codons (TAA, TAG, red lines) are shown in each of the three reading frames for chromosomes (a, d) and transcripts (b, c). Inferred translated regions are highlighted in blue. ATG codons corresponding to translation initiation sites are indicated beneath each plot. Stop codons (and adjacent upstream codons) where termination or frameshifting occur are also indicated. (a) Example of +1 ribosomal frameshifting at AAA_TAA. (b) Example of mRNA with several ribosomal frameshifting sites. (c) Example of +2 frameshifting at the ATA_TAA. (d) Example of +1 frameshifting at AAC_TAA.

Figure 2. Identification of amino acids inserted at frameshift sites

(a) Lysine (K) and asparagine (N) are inserted at the AAA_TAA_C heptamer. Nucleotide sequence surrounding the AAA_TAA +1 frameshift site is shown in the middle. Amino acid sequence is shown above for the zero frame and below for the +1 frame. (b) Recorded MS/MS spectrum confirming the presence of a peptide derived from predicted frameshifting. (c) Peptides detected by MS/MS analysis that were derived from the translation of frameshift sites are shown along with the corresponding nucleotide templates. Nucleotides “skipped” as a result of frameshifting are highlighted in gray. Codons preceding stop codons are shown in red, and the amino acids inserted at frameshifting sites are indicated.

Figure 3. Distribution of codons upstream of stop codons at the frameshift sites and at the sites of translation termination

(a) Frameshift sites. The plot on the left shows absolute frequency of each sense codon ranked based on its frequency. Identity of codons is given by Codon in the middle table. GC content and the inferred mechanism of frameshifting (+1 or +2) are also indicated (nr indicates that the mechanism was not resolved). The absolute number of frameshift sites is listed in Count. Plot on the right shows frequency of codons relative to their expected occurrence based on their usage in internal positions of coding regions. Rows are colored according to codon type. (b) Sites of translation termination. See panel (a) for details. Broken lines indicate average values for absolute frequencies and expected values for normalized frequencies.

Figure 4. Metagene analysis of ribosome profiling and distribution of frameshifting according to transcript levels

(a) Metagene analysis of ribosome density in the vicinity of frameshift sites. First nucleotide of a stop codon is shown as zero coordinate. Note that while ribosome density upstream and downstream of frameshift sites is similar, there is a peak of density at the frameshift sites and this is accompanied by another peak 30 nucleotides upstream. A sequence logo below represents the information content of sequences used for metagene alignment. The sequence AAA_TAA is predominant, and there are no other position-specific signals associated with frameshifting. (b) Metagene analysis of ribosome density in the vicinity of translation termination sites. A drop in ribosome density is evident downstream of stop codons. A sequence logo representing information content in the sequences used for metagene analysis is given below. Only mRNAs with 3’UTRs longer than 90 nts (polyA is not included) were used. (c) Frequency of transcripts with the sites of ribosomal frameshifting (axis X) versus the transcripts ranked based on the levels of protein synthesis (Ribo-seq density), axis Y. (d) Similar to (c), but ranking is based on RNA levels (RNA-seq density). (e) Distribution of transcripts with different Ribo-seq to RNA-seq ratios containing frameshift sites (red) and not containing frameshift sites (black).

Figure 5. Comparison of ribosomal frameshifting at AAA vs non-AAA frameshifting sites and TAA vs TAG frameshifting sites

Aggregated densities of ribosome footprints around frameshift sites containing AAA codon preceding stop (a), non-AAA codons (b), TAA stop codons (d) and TAG stop codons (e). Comparison of footprint density changes observed at frameshift sites at each mRNA (D3 region) and downstream of frameshift sites (D2) relative to footprint density upstream of frameshift sites (D1). D1 and D3 regions were chosen 60 nts upstream and downstream of frameshift sites in order to avoid aberrant densities inflicted by ribosome pauses at frameshifting sites. Box plots represent ratio distributions with horizontal line corresponding to the median, box representing 25th and 75th percentiles and whiskers 5^th and 95^th percentiles. The comparison was carried out for AAA (n=1368) vs non-AAA (n=397) containing frameshift sites (e) and TAA (n=1488) vs TAG (n=277) containing frameshifting sites (f). P-values were calculated using unpaired Wicloxon rank-sum test on log ratios. The data suggest that the frameshifting efficiencies are similar at all frameshift sites, but strong pauses (D3/D1) are less frequent in non-AAA and TAG containing sites.

Figure 6. Cross-species comparison and frequency of nucleotide deletions in different hexamers

(a) Two typical pairwise alignments containing single nucleotide gaps in one of two orthologous sequences in E. crassus and E. focardii. (b) Frequency analysis of all possible hexamer patterns corresponding to deletions (as highlighted in yellow in a) in pairwise alignments for E. crassus (left) and E. focardii (right). The Y axis shows the frequency of each hexamer found in the pairwise alignments with a gap corresponding to the fourth position of the hexamer. Hexamers that end with either TAA or TAG are shown in red. Two most frequent hexamers, AAATAA and AAATAG, are indicated.