Bioinformatic model. Clusters of recurrent GVs facilitate aberrant splicing of HAS1 gene in patients with MM to create the intronic HAS1Vb splice variants. In this analysis we used the web-based bioinformatic tool ESE finder V2. Results were evaluated using ESE V3 (http://rulai.cshl.edu/cgi-bin/tools/ESE3/esefinder.cgi). For more detailed analysis we used ASD (The Alternative Splicing Database; workbench bioinformatics tools). Using these tools, we evaluated the distribution of splicing elements in HAS1 exons 3 and 4 and introns 3 and 4 of wild-type and mutated sequences. (A) Relative distribution of recurrent mutations detected in patients with MM and WM is shown, and the accumulation of 2 important splicing cofactors, hnRNP I (PTB) and hnTNP A, is shown in exon 4 and introns 3 and 4. (B-D) The location on the HAS1 gene where the aberrations occur. (E) Predicts the effect of recurrent GVs on HAS1 splicing. (D), The red letters “A” and “Y” represent activated splicing branch point (BP) and polypyrimidine tract (PPT) of splicing, respectively, and gray letters “A” and “Y” represent native BP and PPT. Description of the model. No differences were found between wild-type and mutated exon 3 with respect to the accumulation of hnRNPs which bind manly splicing suppressors and promote exon exclusion. However, in mutated exon 4, compared with wild-type exon 4 and in mutated exon 3, bioinformatic analysis predicted a massive accumulation of hnRNPs, including hnRNP I (PTB, polypyrimidine tract binding protein), which is distributed across the entire mutated exon 4 (A,B). As suggested in the diagram, the binding of PTBs at several sites of an exon could cause a loopout of this exon, and subsequently these types of exons become inaccessible for the assembly of the spliceosome (B,C,E). The analysis did not predict any significant differences between wild-type and mutated intron 3 with respect to Serin/Arginine-rich proteins (SRs) or the distribution of hnRNP binding motifs. In addition, no significant difference was found when BP and PPT were mapped on wild-type and mutated intron 3. However, for mutated intron 4, the existence of alternative splicing branch points were predicted. These alternative BPs are located upstream of the alternative PPT (D). In addition, splicing element analysis of wild-type and mutated intron 4 showed an accumulation of a significant number of SR and hnRNP binding motifs in mutated intron 4. Among them, the most significant predicted difference that contributes to intronic splicing of HAS1 is recruitment of U2AF65 protein by the alternative PPTs (D). These predicted PPT sequences overlap with the 1st and 2nd “T” stretches and TTTA repeats of mutated intron 4 (the common motif) where the MM clusters of GVs are located. The protein U2SF65 is known to be responsible for the recruitment of SFs to splicing BP. Subsequently, this protein acts as a “bridge” between BP and PPT and stabilizes the spliceosomal complex necessary for the first stage of the splicing reaction. In addition, our analysis of wild-type and mutated intron 4 predicted the loss of a significant number of binding motifs for hnRNP proteins from mutated intron 4 compared with wild type. However, mutated intron 4 maintained ability to recruit hnRNP, a protein which most likely contributes to the exclusion of exon 4 through its ability to dimerise with other molecules of hnRNP A located within and on adjacent introns (E). (F) Splicing of aberrant HAS1 Vb transcripts in transfected HeLa cells. The diagram shows the expression cassettes for HAS1 minigene constructs. HAS1 sequences were flanked by a mammalian CMV promoter at the 5′ end of HAS1 gene and the bovine growth hormone polyadenylation signal, poly A, at the 3′ end. mRNA splicing was analyzed by transfecting HeLa cells with HAS1 minigene cassettes. RT-PCR was performed 24 hours after transfection, using specific primers for HAS1 full-length (FL), HAS1Vb or B2m (β-2 microglobulin). On the gel, Ø indicates the result obtained from the cells transfected with cassette without HAS1 gene; FL, the result obtained from the cells transfected with pcDNA3-HAS1-FL cDNA construct, which is already spliced; Lanes 1 to 4 represent Hela cells transfected with pcDNA3-HAS1-g3–4-5 construct. We tested 4 subclones of the HAS1 minigene cassette; transfection of all 4 subclones gave identical results. Product identity was confirmed by sequencing. For this experiment, Ø and FL were used as controls to verify specificity of the in vitro splicing assay.
