Molecular modeling and global computational mutagenesis of ABCA4 protein show a role of missense changes in Stargardt’s disease.

Abstract

Purpose : Autosomal recessive Stargardt’s disease is the most common form of juvenile macular dystrophy and results from mutations in the ABCA4 gene. Understanding of genetic mutations mechanisms at the atomic level of protein structure is harmed by the absence of ABCA4 atomic crystal structure. Here we model a full-length ABCA4 atomic structure and implicate a global computational mutagenesis to analyze effects of missense changes on protein folding, stability, and association with Stargardt’s disease phenotype.

Methods : The protein stability changes-to-phenotype evaluations were performed for a subgroup of ~400 patient's data (genotypes, age of onset, BCVA) affected by missense mutations only. ABCA4 sequence containing 2273 amino acids (Uniprot, P78363) was used for homology modeling based on a structural template of human ABCA1 (RCSB File: 5XJY). The structure was incorporated in a lipid membrane and the protein-membrane complex was equilibrated by 4ns molecular dynamics in water. The global computational mutagenesis (McCafferty & Sergeev, Sci. Reports, 2016, PLOS One 2017) was applied to the structure.

Results : Effects of patient’s 253 missense variants were evaluated from the changes in ABCA4 protein stability. Within these variants, 36% mutations expected to be associated with a loss of native conformation (unfolding fraction > 0.9) and 27% have a stabilizing effect on protein fold (free energy change < 0). Predicted protein stability changes due to the mutation ~80% agree with published experimental data (Garces et al, 2018). Age of onset was analyzed as a function of protein stability change. Patient’s genotypes fall in 3 groups by changes in protein stability. The genotypes, which were not changing or improving protein stability with the age of onset increase (free energy change <0 kcal/mol) were included in a 1st group. The second group was related to severe changes (>10 kcal/mol) and the 3d group of genotypes had intermediate changes in protein stability. In the last two groups was demonstrated that the changes in protein stability caused by genotype mutations playing a more significant role at earlier ages of disease development.

Conclusions : The analysis of Stargardt’s patient genotypes from the atomic level of protein structure could provide a computational tool to understand the role of protein stability in the disease.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.