An intronic regulatory element that overlaps with U1ex13-11 binding site regulates exon 13 splicing and forms a stem-loop RNA secondary structure
Among the CFTR exons evaluated, exon 13 was the most challenging to rescue. In fact, at variance to the other exons, we have identified only one ExSpeU1, U1ex13-11, with a reliable splicing rescue activity on both 1863 C>T and 1898+3A>G mutations (Fig. 1A). Thus, we decided to explore if there is any specific reason for this apparent lack of efficiency. To this aim we focussed on intronic regulatory elements within the ExSpeU1 binding region performing consecutive intronic deletions in WT, 1863C>T and 1898+3A>G (Fig. 3A) minigenes. Interestingly, deletion of bases in position 12-15 and 17-23 downstream of the 5’ss, but not in the other positions, improved splicing in both the WT and in the 1863C>T contexts (Fig. 3A, lanes 2,3 and 9,10, respectively). These deletions, however, did not affect splicing in the 1898+3A>G minigene. These results indicate that the binding sequence of the active U1ex13-11 (between positions +11 and +26) overlaps with an Intronic Splicing Silencer (ISS) located between position +11 and +23. In this context, the strength of the ISS is not sufficient to recover exon 13 skipping caused by the 5’ss mutation. In addition, analysis with RNA-mfold (Wang et al., 2017) showed that the ISS is located within a 7bp-long stem-loop that include the 5’ss (Fig. 3B). This configuration suggests that the peculiar RNA secondary structure where the ISS is located might limit the accessibility of the U1snRNP to the 5’ss, as previously reported in other cases (Buratti & Baralle, 2004). To clarify the ISS mechanism in detail we performed site directed mutagenesis experiments assisted by RNA secondary structure analysis in the loop (positions +12 and +13), in the stem (positions +14 to +19) and outside the structure (positions +20 to +23). The resulting effect on splicing expressed as changes in percentage of exon inclusion was then related to the predicted effect on the secondary structure expressed as ΔG, that is the quantity of energy needed to fully break the secondary structure (Fig. 3D and Table S1). Mutagenesis from the +14 to +19 position are predicted to open the stem of the structure and accordingly in splicing assay they increase exon inclusion (Fig. 3C, lanes 8-18). In contrast, the mutants predicted to close the loop, namely +13C, +12C13C and +12A13C, had a negative effect on exon 13 inclusion and induced nearly complete skipping (Fig. 3C, lanes 5, 25, 26). All other mutants in the loop (+12T, +12A, +12C, +13T, +13G +12A13T, +12C13G) and those outside of the structure (+20A, +20G, +21G, +22T, +22A, +23G), that do not significantly affect the secondary structure, had no significant effect on splicing. Based on the predicted changes in ΔG (Table S1) and the resulting change in percentage of exon inclusion we modelled association between the change in stability of the secondary structure and the impact on splicing (Fig. 3D). This analysis showed a clear sigmoidal association between splicing of exon 13 and ΔG, supporting the hypothesis that the stem-loop structure at CFTR exon 13 5’ss is involved in splicing regulation and that this might restrict the effect of ExSpeU1s to specific intronic binding sites.