How are circRNAs translated by non-canonical initiation mechanisms?

Circular RNAs (circRNAs) are covalently closed RNA loops produced by a very large number of expressed eukaryotic genes. Initially considered as splicing background and/or splicing side products, recent studies have shown that they are evolutionary conserved and abundant in cells. Yet, their functions remain largely unknown. Because of their circular shape, they were initially categorized as non-coding RNAs. However, recent studies based on mass spectrometry analysis indicate that some cytoplasmic circRNAs are effectively translated into detectable peptides. This raises the interesting question of which mechanisms regulate the translation initiation of those circular transcripts, i.e. unable to recruit the small ribosome subunit through the 5' cap. A possible mechanism for alternative translation initiation is the presence of an IRES (Internal Ribosome Entry Site) that allows direct recruitment of initiation factors and ribosomes on the RNA independently from the cap. This is the case for several circRNAs that exhibit IRESs upstream from the start codon. Yet, another process seems to be involved in initiating the translation of circRNAs: the presence of N6-methyladenosine (m6A) residues. These m6A can promote cap-independent translation and have been shown to be enriched in circRNAs. Interestingly, these two alternative translation initiation processes are generally activated under cellular stress to allow expression of specific stress response genes. These discoveries therefore link circRNA translation to cellular response to stress conditions, raising new enquiries about the regulation of circRNA expression under stress conditions and their functions. This review provides a state of the art on this emerging area.