Identification of a New Class of Exonic Splicing Enhancers by In Vivo Selection

In vitro selection strategies have typically been used to identify a preferred ligand, usually an RNA, for an identified protein. Ideally, one would like to know RNA consensus sequences preferred in vivo for as-yet-unidentified factors. The ability to select RNA-processing signals would be particularly beneficial in the analysis of exon enhancer sequences that function in exon recognition during pre-mRNA splicing. Exon enhancers represent a class of potentially ubiquitous RNA-processing signals whose actual prevalence is unknown. To establish an approach for in vivo selection, we developed an iterative scheme to select for exon sequences that enhance exon inclusion. This approach is modeled on the in vitro SELEX procedure and uses transient transfection in an iterative procedure to enrich RNA-processing signals in cultured vertebrate cells. Two predominant sequence motifs were enriched after three rounds of selection: a purine-rich motif that resembles previously identified splicing enhancers and a class of A/C-rich splicing enhancers (ACEs). Individual selected ACEs enhanced splicing in vivo and in vitro. ACE splicing activity was competed by RNAs containing the purine-rich splicing enhancer from cardiac troponin T exon 5. Thus, ACE activity is likely to require a subset of the SR splicing factors previously shown to mediate activity of this purine-rich enhancer. ACE motifs are found in two vertebrate exons previously demonstrated to contain splicing enhancer activity as well as in the well-characterized Drosophila doublesex (dsx) splicing enhancer. We demonstrate that one copy of the dsx repeat enhances splicing of a vertebrate exon in vertebrate cells and that this enhancer activity requires the ACE motif. We suggest the possibility that the dsx enhancer is a member of a previously unrecognized family of ACEs.