Transposable element exonization by non-canonical splicing generates a diversity reservoir of functional protein isoforms

mRNA splicing enlarges the diversity of the protein repertoire, mainly through alternative exon retention or skipping. Although non-canonical splicing variants that exonize intronic regions were recently identified at the mRNA level, their contribution to the cellular proteome remains unclear. Here, we describe a population of 1300 unannotated protein isoforms generated by non-canonical splicing between exons and transposable elements (TE) using a combination of transcriptome assembly, ribosome profiling, and mass spectrometry. Despite being shorter and expressed at lower levels, their translation efficiency is similar to that of canonical isoforms, and they are shared between individuals. Functional analyses of 5 different non-canonical isoforms show stable expression, specific intracellular localization, and modified functions, as compared to the corresponding canonical isoforms. Non-canonical isoforms derive mainly from evolutionarily ancient genes and are a preferential source of alternative splicing upon which natural selection can act. We conclude that recently exonized TE isoforms are potentially functional and represent a diversity reservoir of novel protein isoforms. RNA-seq from from human cancer cell lines (treated or untreated with puromycin) used for transcriptome assembly. Differential expression analysis on cell lines overexpressing PTEN and WWOX non-canonical isoforms. Ribosome profiling in lung cancer cell lines. Differential expression analysis on cell lines for the study of non-canonical isoforms.