DISCUSSION
Alternative splicing, which is present in over 90% of mammalian genes, plays an important role in biodiversity and biocomplexity, and the misregulation of splicing is implicated in an increasing number of human diseases. About 10% of HGMD reported human pathogenic variants have been found to influence pre-mRNA splicing (Soemedi et al., 2017), which is likely an underestimate. Therefore, it is necessary to study the effects of exonic variants in the SLC12A3 gene of GS on potential splicing, with the discovery of novel variants. In the absence of RNA samples to study alternative splicing, minigene analysis, the effectiveness of which has been confirmed by different studies (Fraile-Bethencourt et al., 2019; Suarez-Artiles, Perdomo-Ramirez, Ramos-Trujillo, & Claverie-Martin, 2018; Takeuchi et al., 2015; Tournier et al., 2008), is still the most direct and credible experimental approach to assess whether a variant influences recognition of an exon and potentially causes phenotypic changes (Bao, Moakley, & Zhang, 2019). In previous studies, we have used this approach to assess the consequences on pre-mRNA splicing of presumed mutations associated with different diseases (Han et al., 2019; Wang et al., 2020; Zhang et al., 2021).
Gene splicing diversity is dramatically affected not only by canonical splice signals such as splice sites recognized by the core spliceosomal components, but also by a large number of regulatory elements in the exon or flanking intronic sequences. Based on the above, we empirically chose BDGP and HSF as bioinformatics tools to predict the effect of missense variants of SLC12A3 gene, and obtained some satisfactory results through minigene assays. Variants located at or near to classical splice sites (DS or AS) could result in skipping of the corresponding exon either by generating new splice sites or by inducing a significant reduction of splice site strength, as confirmed by the minigene analysis of variants c.602G>A, c.602G>T, c.1925G>A, c.2548G>C and c.2549G>C in the study. Additionally, variants in the exon regulated by a weak donor or acceptor splicing site could modify splicing regulatory sequences (disrupt ESEs or create ESSs), leading to the corresponding exon skipping, as revealed by the minigene analysis of variants c.1667C>T in our study. Among the 10 candidate variants selected by BDGP and HSF following the screening criteria, 6 variants causing exon skipping were predicted successfully, which indicated a degree of concordance betweenin silico predictions and the experimental results. Despite a considerable degree of inaccuracy, it is still valuable to use these two bioinformatics tools to predict the impact of SLC12A3 exonic variation on splicing based on this or more stringent criteria.
According to the resulting data, among six variants that predicted successfully, c.602G>T, c.1925G>A and c.2548G>C were complete splicing mutations just generating the corresponding exon-excluded transcripts, while the other variants (c.602G>A, c.1667C>T and c.2549G>C) still produced a little part of transcripts that contained the corresponding variant and were the same size as WT products. Therefore, these three variants probably have a dual damaging effect: on the one hand, a portion of SLC12A3 transcripts are deleterious due to the exon skipping, which would result in loss of the corresponding proteins, and on the other hand, the remaining mRNA is destructive due to the replacement of a single amino acid, which could lead to the malfunctioning of NCCT. The true splicing effect of these variants needs to be confirmed by analyzing the mRNA from the patient, which we did not have yet.
However, it is worth noting that bioinformatics software and minigene assays, as effective tools for identifying splicing malfunctions, have the limitation in detecting and simulating all splicing patterns compared to the in vivo situation. Variants of c.2755A>T and c.2863A>T, predicted by BDGP and HSF to have a significant impact on ESE / ESS motifs, did not cause exon 24 skipping, and the WT minigene of exon 12 only produced the exon 12-excluded transcript. Coincidentally, the same results that candidate variants did not alter pre-mRNA splicing were revealed in other studies (Suarez-Artiles et al., 2018; Wang et al., 2020; Zhang et al., 2021) and there was a previous study that suggested only 18 of the 20 wild-type minigene constructs produced the desired wild-type transcripts in the pSPL3 context, improving the awareness of the limitations of minigene assays and emphasizing the importance of sequence context in regulating splicing (Lin et al., 2021). The reasons for these results may be related to the defects of software, the limited transferability of minigenes, the differential expression of splicing factors in cells, the interference of the mRNA secondary structure, etc. The complete skipping of wild-type exon 12 might also be associated with the inactivation of splicing sites and the difficulty in recognizing splicing sites. Moreover, it has been reported that the role of splicing regulatory elements could be highly context dependent (Goren et al., 2006), but the underlying mechanism has been unclear, and our minigene splicing assays might disrupt this sequence context effect leading to exon 12 skipping.
In addition, we have not yet prepared the expected cDNA constructs based on the prediction of the single amino acid substitution and the corresponding exon deletion, nor have we examined the functional activities and the cell surface expression of these mutant NCCTs, which need further investigation.
In order to verify the pathogenicity of exonic variants of theSLC12A3 gene in GS, variation analyses at both DNA and RNA levels should be performed to predict the impact on mRNA and protein. Minigene splicing assays have revealed new insights and changed the misinterpretation about the functional consequences of certain variants. It is worth mentioning that exon-skipping methods to correct variants that disturb normal pre-mRNA splicing have been effectively assessed in rare diseases (Veltrop & Aartsma-Rus, 2014). Knowledge of the results of exonic splicing variants may have potential therapeutic significance for GS, further emphasizing the importance of minigene splicing analysis.