Molecular mechanisms of circular RNA translation

Abstract

Circular RNAs (circRNAs) are covalently closed single-stranded RNAs without a 5' cap structure and a 3' poly(A) tail typically present in linear mRNAs of eukaryotic cells. CircRNAs are predominantly generated through a back-splicing process within the nucleus. CircRNAs have long been considered non-coding RNAs seemingly devoid of protein-coding potential. However, many recent studies have challenged this idea and have provided substantial evidence that a subset of circRNAs can associate with polysomes and indeed be translated. Therefore, in this review, we primarily highlight the 5' cap-independent internal initiation of translation that occurs on circular RNAs. Several molecular features of circRNAs, including the internal ribosome entry site, N⁶-methyladenosine modification, and the exon junction complex deposited around the back-splicing junction after back-splicing event, play pivotal roles in their efficient internal translation. We also propose a possible relationship between the translatability of circRNAs and their stability, with a focus on nonsense-mediated mRNA decay and nonstop decay, both of which are well-characterized mRNA surveillance mechanisms. An in-depth understanding of circRNA translation will reshape and expand our current knowledge of proteomics.

Fig. 1. Biogenesis of circRNAs.

CircRNA biogenesis from pre-mRNAs. EcircRNAs, EIciRNAs, and ciRNAs are generated via canonical splicing or back-splicing of pre-mRNAs.

Fig. 2. Molecular mechanisms for canonical cap-dependent translation initiation and internal initiation of translation.

a Canonical cap-dependent translation initiation. The molecular axis of eIF4E-eIF4G-eIF3-40S ribosome is critical for 5′-cap-dependent translation initiation. The interaction between eIF4G and PABP1 results in the formation of a looped structure, which is favorable for translation initiation. b Various modes of internal translation initiation of circRNAs. Various molecular features of circRNAs contribute to internal entry of the ribosome to initiate translation. These features include (i) canonical IRESs; (ii) IRES-like sequences or elements such as m⁶A, poly(U) sequence motifs, complementary sequences to 18 S rRNA, and tandem repeats of short nucleotides that can induce RAN translation; and (iii) the EJC deposited on the circRNA after back-splicing. The features for which the molecular mechanism underlying their involvement in ribosome recruitment is not clear are marked with a question mark.

Fig. 3. Various types of ORFs in circRNAs and their roles in circRNA stability.

a NMD of linear mRNA and circRNA. The presence of the EJCs sufficiently downstream of the stop codon induces rapid degradation of linear mRNAs via the NMD pathway (left). Similarly, when a circRNA contains an EJC (at the BSJ) sufficiently downstream of the stop codon, it would be targeted for NMD (left). b Diverse ORFs in circRNAs depending on the relative positions of translation start codons, stop codons, and the BSJ. When the EJC is deposited onto the ORF in a circRNA, it would be displaced from the circRNA by elongating ribosomes. Consequently, the circRNA would evade NMD and be stabilized. c Rolling circle translation of circRNAs lacking stop codons. Whereas a linear mRNA lacking stop codons is rapidly degraded via the NSD pathway (left), a circRNA lacking stop codons would be subject to rolling circle translation (right).