Abstract
Purpose:
Autosomal recessive Stargardt disease is the most common form of juvenile macular dystrophy and results from mutations in the ABCA4 gene. Approximately 49% of pathogenic ABCA4 missense mutations occur in the trans-membrane (TM) or nucleotide binding domains (NBD). Our previous ABCA4 molecular modeling predicted the destabilizing effects of 20 missense changes in the TM and NBD functional domains<br /> in a small patient cohort. Here we report an analysis of the missense mutations found in a larger patient cohort.
Methods:
Patient genotypes came from the National Eye Institute and EyeGene (518 patients), Kellogg Eye Center (University of Michigan) (28 patients), Catholic Medical University (Rome, Italy) (29 patients), and Moorfields Eye Hospital/UCL (London UK) (164 patients). 532 patients with 2 affected alleles carrying 1153 ABCA4 mutations were analyzed, along with age of onset and current visual acuities. We excluded those patients with only a single allele mutation identified. Atomic models of functional domains were modeled by homology which was refined and equilibrated using molecular dynamics. Severities of missense changes localized to TM-NBD’s were evaluated in similar fashion to our previous work (Sergeev et al., HMG 2010, 2013).
Results:
About half of the selected cohort, 265 genotypes, had mutations in the TM or NBD. Other mutations were missense changes affecting parts of ABCA4 of unknown atomic structure (~26%) or were nonsense changes (~24%). Preliminary shown that the mutation impact for each individual analyzed was associated with the age of retinal disease onset. The majority of significant hotspot mutations with >10 genotypes each were located in nucleotide-binding domains (78%). Outstanding mutation hotspots were mapped to the TM (G863A, with 64 genotypes), the NBD1 domain (G1961E, 111; A1038V, 48; R1108C, 22; N965S, 13; and E1087K, 10 genotypes), and the NBD2 domain (L2027F, 31; R2107H, 24; R2030W, 17; and V2050L, 11 genotypes). Molecular modeling demonstrated that hotspots located at domain surface were severe and perturbed the ATP-hydrolysis. Mutation G863A affected protein flippase activity.
Conclusions:
The large-scale mutational analysis and predictions of mutation severities from atomic level could be useful for the functional annotation of genetic variants from next-generation sequencing data and establishing the genotype-to-phenotype relationships in genetic disease.