Size matters: ABCA4 midigenes facilitate accurate RNA splicing analysis of non-canonical splice site variants in Stargardt disease

Abstract

Purpose : Stargardt disease is characterized by a conspicuously high proportion of non-canonical splice site (NCSS) variants in ABCA4 that have unknown functional consequences. In vitro minigene splice assays enable a cost-effective assessment of putative splice defects. While using an in-house Gateway-adapted minigene splice assay vector containing RHO exons 3 and 5, we found a splicing artifact when testing an exon 40 non-canonical splice variant. The strong nature of the RHO exon splice sites resulted in aberrant wild-type and mutant exon 40 skipping. We therefore tested the hypothesis that larger constructs, containing multiple exons and coined ‘midigenes’, would represent a more reliable in vitro system. In addition, we aimed to generate a complete library of wild-type midigenes for the ABCA4 gene which would allow us to systematically test all published NCSS variants for splice abnormalities.

Methods : Wild-type midigenes were generated by PCR amplification of overlapping ABCA4 fragments from a bacterial artificial chromosome clone containing ABCA4. We collected all NCSS ABCA4 variants mentioned in peer-reviewed papers from 1997 to 2015. Their putative effect on splicing was assessed by using 5 different splicing algorithms via Alamut Visual software. Through site-directed mutagenesis, the wild-type midigenes were converted into mutant versions and subsequently used in HEK293T splice assays.

Results : 28 wild-type ABCA4 midigenes (insert sizes: 4.7 – 11.7 kb) were generated, each of which containing the exon or intron of interest plus as many flanking exons (ranging from at least 2 to 6) as possible. Altogether they cover ~80% of the 128-kb ABCA4 gene and 46 of 50 exons. 39 ABCA4 NCSS variants were tested and abnormal splicing was observed in 65% of these. We found exon skipping, exon elongation and multiple exon abnormalities, which enabled us to determine the severity of the different NCSS variants.

Conclusions : The creation of a library of ABCA4 wild-type midigenes allowed us to generate mutant constructs containing NCSS variants. Similarly, this midigene collection will also enable us to quickly test potential splicing defects due to any other unclassified variant, and thereby can become a cost-effective diagnostic tool.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.