Dual coding in alternative reading frames correlates with intrinsic protein disorder

摘要

Numerous human genes display dual coding within alternatively spliced regions, which give rise to distinct protein products that include segments translated in more than one reading frame. To resolve the ensuing protein structural puzzle, we identified 67 human genes with alternative splice variants comprising a dual-coding region at least 75 nucleotides in length and analyzed the structural status of the protein segments they encode. The inspection of their amino acid composition and predictions by the lUPred and PONDR? VSL2 algorithms suggest a high propensity for structural disorder in dual-coding regions. In the case of +1 frameshifts, the average level of disorder in the two frames is similarly high (47.2% in the ancestral frame, 58.2% in the derived frame, with the average level of disorder in human proteins being approximately 30%), whereas in the case of -1 frameshifts, there is a significant tendency to become more disordered upon shifting the frame (16.7% in the ancestral frame, 56.3% in the derived frame). The regions encoded by the derived frame are mostly disordered (disorder percentage ＞50%) in 39 out of 62 cases, which strongly suggests that structural disorder enables these protein products to exist and function without the need of a highly evolved 3D fold. The potential advantages are also demonstrated by the appearance of novel functions and the high incidence of transcripts escaping nonsense-mediated decay. By discussing several examples, we demonstrate that dual coding may be an effective mechanism for the evolutionary appearance of novel intrinsically disordered regions with new functions.