Background

A substantial portion of disease-associated mutations (at least 15%) are predicted to result in aberrant mRNA splicing (Cartegni et al., 2002; Singh and Cooper, 2012; Sterne-Weiler and Sanford, 2014). The ‘classical’ splice mutations are those affecting the canonical sequences defining the 5’- and 3’-splice sites (donor and acceptor splice sites, respectively). However, splice mutations can also occur in other non-coding and coding regions (Wang and Cooper, 2007; Scotti and Swanson, 2016). There is growing evidence that the frequency of splice mutations in coding regions (exonic mutations) has been underestimated (Julien et al., 2016; Soukarieh et al., 2016). Exonic splice mutations (i.e., point mutations, insertions or deletions) can induce exon skipping, intron retention or lead to the generation of novel donor or acceptor splice sites. Depending on the gene and exon composition, these mechanisms can have different impacts on the protein level ranging from haploinsufficiency to gain of function. Nevertheless, the exact knowledge of molecular mechanisms underpinning the disease-causing mutations is essential for developing optimal treatments.

mRNA splicing occurs in a highly cell type-specific manner, highlighting the need to analyze the impact of potential splice mutations in the tissue which is primarily affected by the mutation (Wang et al., 2008). Consequently, in the optimal case, mRNA splicing should be analyzed on the native gene and in the native tissue. This option, however, is rather demanding for several reasons:
1) It might require the elaborate generation of genetically modified human cell lines. This impedes a more systematic analysis of splice mutations for a single gene.
2) Many native cell types are highly specialized and their cultured counterparts (if available at all) do not reflect every morphological and molecular hallmark of the native cells including the composition and activity of the splicing machinery.
3) The generation of humanized animal models expressing the respective splice mutation in a given tissue is not only technically challenging, but also time-consuming and costly. Therefore, this approach also appears rather unsuitable for systematic testing of splice mutations for a given gene.
4) Often, native genes are too large to be cloned into classical expression vectors.

One alternative to circumvent a number of these obstacles is to use human minigenes designed for expression in appropriate animal models (e.g., mouse). We have evaluated this approach in recent studies addressing the effects of disease-associated mutations in different genes, e.g., PRPH2, on mRNA splicing (Becirovic et al., 2016b; Nguyen et al., 2016; Khan et al., 2017; Petersen-Jones et al., 2017). For stable and specific ectopic expression of the minigenes, we took advantage of rAAV vectors. These vectors are capable of transducing a variety of different cell types in vivo (Zincarelli et al., 2008; Lisowski et al., 2014). Furthermore, the design, cloning, production and purification of rAAV vectors can be completed in a few weeks and does not require elaborate technical equipment (Becirovic et al., 2016a).

Most native genes including PRPH2 exceed the limited packaging capacity of AAVs (approx. 4.7 kb) (Wu et al., 2010). Therefore, we designed PRPH2 minigenes lacking large intronic parts, which usually do not contain information required for correct mRNA splicing. For genes which do not contain large exon numbers or sizes, shortening of the intronic sections also allows for introducing the entire protein coding region into the rAAV vector-based minigenes.

This strategy (cf. Figure 1) was developed and evaluated to analyze the impact of known disease-linked mutations on mRNA splicing and protein expression in photoreceptor-specific genes, but should in principle also be transferable to other cell types.


Figure 1. Workflow schematic for construction and cloning of minigenes