Proteins From Hereditary Eye Diseases: In silico and Experimental Studies

National Eye Institute

Investigators

Linked publications & trials

Abstract

To understand how a pathogenic mutation causes inherited eye disease, it is necessary to recognize how pathogenic mutations can affect protein structure-function and metabolic pathways, and how these perturbations may be associated with clinical parameters describing the disease phenotype. Tyrosinase serves as the key enzyme in melanin biosynthesis, catalyzing the initial steps of the pathway. Previously, we aimed to develop a global computational mutagenesis method to predict protein stability changes caused by genetic perturbations (McCafferty & Sergeev, 2016, 2017; Ortiz & Sergeev, 2019). This method was successfully applied to analyze the effect of single genetic mutations. For further development in evaluating the effect of multiple mutations occurring in the same protein molecule, we need a more advanced grading algorithm. Recently, we conducted two studies associated with oculocutaneous albinism, an autosomal recessive inherited disorder linked to mutations in the TYR gene. The first study began several years ago when our colleagues from the University of Alberta (Canada) were unable to correctly determine the severity of a mutant variant in a 1-year-old patient. A single missense change in tyrosinase (Tyr) can result in partial or complete loss of catalytic activity. The effect of two genetic mutations in the same Tyr molecule is less studied. We expressed and purified human tyrosinase in a larval system, including single mutation variants R217Q and R402Q, as well as a double mutant variant, R217Q/R402Q. We also applied computational methods to establish a link between alterations at the atomic model level of the protein and the disease phenotype (Toay S., Sheri N., MacDonald I., Sergeev Y., 2025). Human recombinant intra-melanosomal Tyr domains of Tyr and the three mutant variants were expressed in T. ni larvae, purified using a combination of IMAC and SEC, and their diphenolase activities were measured. The Tyr homology model was equilibrated using 100 ns molecular dynamics and analyzed using computational methods. The purified R217Q and R217Q/R402Q variants showed decreased catalytic activities compared to those of the Tyr and R402Q variants. The R217Q/R402Q variant exhibited the lowest protein activity, and its protein yield was significantly reduced. The effect of the R402Q mutation, well known as a polymorphism, is less pronounced compared to the R217Q mutation. Principal component analysis (PCA) of the tyrosinase collective motions was performed. Oculocutaneous albinism type 1 is a genetic disorder caused by the disruption of tyrosinase activity in the melanogenesis pathway. The tyrosinase's intramelanosomal domain can be subdivided into the catalytic and Cys-rich subdomains, integral for protein stability and catalytic activity. To understand the movement in the tyrosinase intra-melanosomal subdomains and their link to its catalytic activity, we performed essential dynamics on homology models for tyrosinase and the mutant variants R217Q, R402Q, and R217Q/R402Q. Dimensional reduction techniques, such as PCA, are fundamental to systematically comprehending collective movements in protein structure. The alpha-carbon atomic coordinates for all residues across a 100-ns molecular dynamics trajectory were input into the PCA function, and the results were analyzed alongside correlated movements and free energy profiles for each protein structure. The PCA identified coordinated movement underlying the stable conformations of wild-type tyrosinase that arises within the H9 and H10 helices, which are proximal to the flexible tunnel system and the interface of the catalytic and Cys-rich subdomains. In contrast, genetic mutations R217Q and R217Q/R402Q disrupt the coordinated movement of the tyrosinase intra-melanosomal domain, indicating a cause of mutant variant instability. Linking Protein Stability to Pathogenicity: Predicting Clinical Significance of Single-Missense Mutations in Ocular Proteins Using Machine Learning" (Majid I., Sergeev Y., 2024). This paper is the first article to establish a link between mutation pathogenicity and protein stability for seven proteins from inherited eye diseases. Understanding the effect of single-missense mutations on protein stability is crucial for clinical decision-making and therapeutic development. The impact of these mutations on protein stability and 3D structure remains underexplored. Here, we developed a program to investigate the relationship between pathogenic mutations and protein unfolding and compared seven machine learning (ML) models to predict the clinical significance of single-missense mutations with unknown impacts based on protein stability parameters. We analyzed seven proteins from the ocular proteome website at NEI Data COMMONS associated with ocular disease-causing genes. The program revealed an R-squared value of 0.846 using Decision Tree Regression between pathogenic mutations and decreased protein stability, with 96.20% of pathogenic mutations in RPE65 leading to protein instability. Among the ML models, Random Forest achieved the highest AUC (0.922) and PR AUC (0.879) in predicting the clinical significance of mutations with unknown effects. Our findings indicate that most pathogenic mutations affecting protein stability occur in alpha-helices, beta-pleated sheets, and active sites. This study suggests that protein stability can serve as a valuable parameter for interpreting the clinical significance of single-missense mutations in ocular proteins. Molecular modeling of NR6A1. NR6A1 is an orphan member of the nuclear hormone receptor family of transcription factors, often acting as a transcriptional repressor. We performed computational analysis to demonstrate that the NR6A1 variants were pathogenic or likely pathogenic for OVR syndrome (Neelathi et al., 2025). A structural model of NR6A1 was created using the AlphaFold server. DNA binding to the ZFD of NR6A1 was modeled using a single ZFD domain of the retinoid X receptor alpha-liver X receptor beta in complex with DNA. Two missense variants (R92W and R436C) were generated using YASARA. The variant models were optimized and minimized using gradient descent. All two minimized mutants and the two WT, ZFD, and NR_LBD models underwent 10 ns of Molecular Dynamics simulations. Molecular modeling suggests missense variants disrupt important intramolecular interactions. The NR6A1 amino acid sequence is well-conserved between humans, mice, and zebrafish. To understand the effects missense variants had on protein stability and function, we created an in-silico model of a complex of NR6A1 with DNA. The AlphaFold model of NR6A1 was analyzed in the composition of Zn-finger (residues 60–172) and NR_LBD (residues 246–480) domains. The rest of the model is predicted as an irregular structure by AlphaFold. In wild-type (WT) NR6A1, a positively charged arginine residue 92 is predicted to interact with negatively charged DNA based on a zinc-finger protein model. The R92W variant replaces the R92 residue with hydrophobic tryptophan (W), possibly interrupting the electrostatic interaction with DNA. The R436C variant affects the putative nuclear receptor ligand-binding domain NR_LBD. In NR6A1, hydrogen atom 1HH2 of arginine R436 closely interacts with the oxygen atom of glutamic acid E388. The variant R436C breaks this bond and creates a cysteine residue which could form abnormal disulfide bridges in the variant protein, since residues C443, C391, and C422 are distanced at 8–12 Å from C436 in this variant domain as compared to 14–19 Å (C443–C391), 11.09 Å (C443–C422) and 4.52 Å (C91–C422) in the WT protein model.