Abstract :
[en] Transcription factors (TFs) are usually defined as sequence-specific DNA-binding proteins that control transcription as a first step of gene expression. However, more and more evidence indicates that they are also involved in the regulation of posttranscriptional steps, in particular the splicing process. In our laboratory, we previously demonstrated that ERG protein, a member of Ets TFs has a role in the regulation of decay [1], a downstream mRNA processing event. In addition, it was suggested that TFs act directly on splicing through different mechanisms, such as by modifying RNA polymerase II elongation rates and altering the kinetics of exposure of splice sites, by the recruitment of transcriptional coactivators involved in splicing, or through the modulation of direct splicing factors expression. However, currently it was proposed that some TFs can control alternative splicing (AS) directly through their binding onto pre-mRNA, thus controlling AS via an unknown but direct mechanism [2]. In this work, we hypothesized that ERG TF can regulate AS process and this can be affected in cancers such as Ewing’s sarcoma (EwS). Thus, we reported for the first time that ERG can regulate AS of numerous splicing targets. This is probably by a direct mechanism. Indeed, first we showed that ERG associates with spliceosomal components, it is found on nascent pre-mRNA and induces AS through its CTAD domain, by inclusion or exclusion of an alternative exon when it is recruited onto a reporter. Second, transcriptomic analysis demonstrated that depletion of ERG affects a large number of AS events, mainly cassette exons, in a tissue-specific manner. Third, we observed that sequences of ERG-regulated cassette exons and their 200 adjacent intronic base pairs are enriched in binding motifs for RBFOX2, a tissue-specific splicing regulator and an important functional interactor of the LASR splicing complex. We demonstrated that ERG and RBFOX2 interact via their CTAD and C-terminal domains respectively, and both proteins collaborate to regulate a large set of cassette exons. Finally, our observations suggested that the splicing function of ERG is independent of its DNA-binding and transcriptional activity since we did not find any significant overlap between differentially expressed and differentially spliced genes in ERG-depleted cells, and a transcriptionally inactive ERG variant lacking the DNA-binding domain (ERG-ETS) presented full splicing activity in our reporter assay. Perturbation of the splicing program is a feature of EwS. This was previously attributed to the presence of FET-Ets fusions [3] and the splicing function of EWS-FLI1 was only attributed to its EWS moiety. In addition, the functional relevance of EWS-FLI1 in the EwS oncogenic process is still unknown. We have studied the role of the EWS-FLI1 fusion in the reporter assay and observed that its AS function is linked to its recruitment onto the reporter mRNA. In addition, our observations suggested that the FLI1-derived moiety can also be a major contributor to the fusion protein’s splicing function in EwS and that EWS-FLI1 also interacts with RFBOX2, suggesting that the EWS-FLI1 and RFBOX2 interaction could play a major role in EwS development. Altogether, our results support a model in which the ERG protein controls AS processes beyond its role as a TF. This new function of ERG in AS regulation should be considered for future cancer therapies.
