RAN translation and frameshifting as translational challenges at simple repeats of human neurodegenerative disorders

Wojciechowska, Marzena; Olejniczak, Marta; Galka-Marciniak, Paulina; Jazurek, Magdalena; Krzyzosiak, Wlodzimierz J.

The translation of most eukaryotic mRNAs is initiated via the cap-dependent scanning mechanism, which employs the 43S pre-initiation complex loaded with several protein factors. The complex proceeds in a 3 direction until it encounters the first AUG codon in a Kozak consensus sequence (24) . These canonical rules seem not to apply to RAN translation, and it currently remains unclear what forces the synthesis of homoproteins by ribosome and exactly where the process starts. One hypothesized trigger is an RNA structure formed by the repeats that, in fact, features in all RANtranslated sequences reported thus far ( Figure 1 ). As described, the hairpin structure-forming CAG repeats are efficiently RAN translated, in contrast to CAA repeats that do not form hairpins (14) . CGG repeats of the FMR1 transcript are also RAN translated; these repeats form hairpins (44) that are able to fold into G-quadruplex structures under specific conditions (45, 46) . The sense G 4 C 2 repeats of C9ALS/FTD also fold into G-quadruplexes that are not detected at the antisense C 4 G 2 repeats of the C9, which, however, may fold into hairpins (47) . As these structures show different stability, and because hairpins are less stable than G-quadruplexes, this phenomenon may affect the expression levels of RAN-translation products and their variety from repeat-harboring transcripts (15, 17, 19) .