Molecular modeling of functional domain of ABCA4: towards understanding the genotype-to-phenotype relationships in Stargardt’s disease

Abstract

Purpose: ABCA4 is an ATP binding cassette protein transporter. Mutations affecting the ABCA4 transporter protein are known to cause Stargardt’s macular degeneration, but very little is known about the structure of ABCA4, making it difficult to understand how these mutations affect the function of ABCA4 as a protein transporter. Here we performed molecular modeling of the functional domain which includes trans-membrane and nucleotide-binding domains, a part of ABCA4 polypeptide responsible for the flipping of N-retinylidene PE over the disc membrane.

Methods: The functional domain was modeled by homology, refined and equilibrated using molecular dynamics. This model was used to predict structural perturbations caused by mutations found in affected human subjects. Overall genotype scoring function was derived from free energy changes for the ABCA4 open/close conformers was calculated for each mutation (Sergeev et al, Hum Mol Genetics, 2010). We then evaluated possible correlations of this modeling with phenotypes of patients from the NEI, EyeGENE database, Italian and Kellogg Eye Center Stargardt’s patients (42 genotypes total).

Results: In the selected group of 15 patients we found an association between the overall genotype score and the age of disease onset known to be related to the severity of retinal disease (Pearson’s r = - 0.71, Adj. R-square = 0.46). Similar association between these parameters was observed for 7 NEI patients enrolled in EyeGENE (Pearson’s r = - 0.76, Adj. R-square = 0.53). Finally, these data were combined with data from Italian Stargardt’s patients. The total dataset of 28 patients showed a linear association between predicted severities and the age of disease onset (Pearson’s r = - 0.67, Adj. R-square = 0.42). Similar result was obtained for 12 patients from the Kellogg Eye Center (Pearson’s r = - 0.76, Adj. R-square=0.54).

Conclusions: The overall impact score for each individual analyzed was associated with the age of retinal disease onset. The functional activity suggested by the model is in agreement with recent experimental data on ABCA4 as a ‘flippase’ protein. Thus, molecular modeling of the ABCA4 functional domains and the scoring system derived in our work has a potential to be used for a study of genotype-to-phenotype relationships in Stargardt’s disease.

Keywords: 539 genetics • 696 retinal degenerations: hereditary • 473 computational modeling