**Conflict of interest**

at the origin of modern tmRNAs [41]. Metazoan mt-*trnA* genes combine the highest levels of TAR10 and ATR49 triplets (>95% for each), but in the prokaryotic world, if the rate of TAG10

Studies strongly suggest that the tRNA cloverleaf structure unfolded prior to the appearance of a fully functional ribosomal core, making it one of the most ancient RNAs of the RNA world [70, 97] or even the oldest [98]. Though the "RNA-world" hypothesis is well accepted, the successive events leading to the emergence of different partners playing a role in translation and the involvement of tRNAs in this evolution are highly controversial coveted field [99]. However, some hypotheses as the "tRNA core" [88] strongly suggest that tRNAs would be at the origin of the primitive genetic material and gave rise to mRNA and rRNA, as well as the conformational structure of the first proto-ribozymes. The base module being a pleiofunctional RNA that can adopt the cloverleaf structure is found today in various sequences without direct link with translation. One may conclude that "one should not change a winning secondary structure." In a precellular context, a molecule with ss-tRNA characteristics (small ORF associated with cloverleaf structure) would be advantageous. Putatively, ss-tRNA-like molecules cumulating both tRNA and mRNA functions

is always higher than 91%, only one ATR49 occurs in Eubacteria and none in Archaea.

would have been the first molecules on Earth to support nonrandom protein synthesis.

The antiquity of ss-tRNAs can be discussed, and it is very likely that the TAR10 (and especially TAG) triplets played very early a critical role in the tertiary folding of some tRNAs. Their implication in translation termination would be an exaptation where firstly, they were part of a structural signal. Origin of ATR49 triplets is less clear perhaps tracing to the first endosymbiosis. Hence it would be apomorphic (derived character). Analyzes by taxa and tRNA species suggest a nonhomogeneous evolution. At the beginning of the RNA/protein world, it has quickly become essential to start peptide synthesis at particular codons and one cannot exclude that ATR49 was an ancestral state which would have not been retained as intergenic spaces increased. Analyzes of known tRNAs of α-proteobacteria and cyanobacteria could suggest that in organelles, ATR49 triplets would have been selected with genome reduction. Organelle genomes may be under increased pressure for size reduction with resulting overlaps (see, [100]). However, several features strongly suggest that overlapping genes are not a direct mechanism to substantially reduce genome size. Gene overlaps allow mtDNA genome compaction while avoiding the loss of tRNA genes [53]. Nevertheless, overlaps may allow a more efficient control in the regulation of gene expression, the regulatory pathways are simplified, and the number of proteins (and genes) required decreases [100]. Among others, short antiparallel overlaps may be involved in antisense regulatory mechanisms. Consequently, genomes with compact sizes enable putatively less flexible but more efficient physiologies.

The selection of tRNAs had to be done mainly on two seemingly opposite criteria, stability and plasticity, making it a kind of Swiss army knife of the RNA world. This explains that beyond their central role in protein synthesis, tRNAs have many other crucial functions. To date, it can be hypothesized that ss-tRNAs might regulate gene expression, stress responses, and metabolic processes. Indeed, *in silico* analyzes allowed to speculate that several overlapping sequences may code simultaneously for mRNAs and tRNAs in most of the metazoan

**4. Conclusions**

20 Mitochondrial DNA - New Insights

The authors declare no potential commercial or financial conflicts of interest.
