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 could affect protein structure-function and metabolic pathways, and how these perturbations could be associated with clinical parameters describing the disease phenotype. Currently, no biochemical studies in vitro reported the catalytic activity of Tyrp2 protein. This year we purified and characterized the intramelanosomal domain of human recombinant tyrosinase-related protein 2 (Tyrp2) (residues 1-474) and missense variants C40S and C61W mimicking the alterations found in genetic studies in patients with oculocutaneous albinism type 8 (OCA8) (Dolinska et al., 2022). Recombinant proteins were produced in the Ti. ni larvae were purified by a combination of affinity and size-exclusion chromatography (SEC) and characterized by protein biochemistry methods. Tyrp2 is involved in the melanogenesis pathway, catalyzing the tautomerization of dopachrome to 5,6-dihydroxyindole-2-carboxylic acid (DHICA). The testing of Tyrp2 activity is a difficult task because of the absence of commercial dopachrome which performs the role of substrate in the reaction of tautomerization. To avoid this issue, we used tyrosinase immobilized magnetic beads (Tyr-MB) to isolate a native dopachrome in the tyrosinase diphenol oxidase reaction with L-DOPA (Isabella Osuna). The native dopachrome was used as a substrate in the testing of Tyrp2 catalytic activity. In this reaction, a significant level of DHICA was detected at 325 nm, while Tyrp2 was incubated in the presence of native dopachrome. Thus, native dopachrome could be a good substrate for testing Tyrp2 and mutant variants' catalytic activity. To investigate the effect of genetic mutations on the human Tyrp2 structure and stability, we expressed and purified two OCA8-related mutants (C40S and C61W). In contrast to the Tyrp2 wild type, mutants showed no peak when eluted from SEC and indicated no bands at the proper position on the Western blots suggesting that both mutant variants showed no soluble protein. Finally, C40S and C61W mutant variants appeared to be misfolded and aggregated. Disease-related mutant variants were analyzed computationally to understand their protein stability (Taariq Woods). First, the atomic model of Tyrp2 was obtained by homology modeling using the Tyrp1 crystal structure as a structural template. The model was refined using molecular dynamics in water. Second, we imply global computational mutagenesis for the prediction of mutation effect from the level of protein structure. Each Tyrp2 residue had a foldability parameter or an aggregated sum of the propensities of the unfolding parameters of that residue's mutation. The foldability parameter highlights the most critical residues involved in the thermodynamic stability of the protein. Residues critical for protein stability were localized in the central part of the protein molecule. Alterations at these positions could potentially destabilize a protein fold. Both C40 and C61 were expected to contribute the most to protein stability, with foldability values of 19. These values were consistent with the unfolding parameters of 1 and 0.99 for the C40S and C61W mutations, respectively, suggesting a complete misfolding of the mutant. Third, results from global computational mutagenesis were matching the pathogenic clinical significance evaluated for the residues in ClinVar. The additional third disease-related mutation, G59V, was also expected to be a misfolding-causing mutation. Unfortunately, we currently do not have clinical and biochemical data about the effect of this mutation. The mutation G59V resulted in incomplete protein maturation and targeting in vitro compatible with a partial or total loss of function. In the future, it will be interesting to test this mutation in vitro. Also, all cysteine residues, which were forming disulfide bridges, had high foldability parameters and were conserved from an evolutionary perspective. The importance of cysteine mutations agreed with phenotypic changes in homozygous mice with C40S and C61W missense mutations in the Tyrp2 gene (equivalent to those found in patients with OCA8), exhibiting dark gray hair and significantly less pigmentation of the RPE. In summary, we showed that using in vitro and computational methods, we proved that the localized in the Cys-rich domain of Tyrp2 mutations strongly impacted protein folding and stability. This result suggests the importance of the Cys-rich domain for tyrosinases and confirms our previous results on Tyrp1 protein (Patel et al., 2001). We also demonstrated a direct link between mutation severity and the pathogenic clinical significance of the mutation. Our experimental and computational work contributes to the understanding of inherited disease from the atomic level of protein structure and analysis of the impact of genetic mutations on disease phenotype. Therefore, we are trying to improve our methods of molecular modeling and developed in our lab global computational mutagenesis to imply these methods to different clinical cases. The possible association between CDHR1 variants and clinical findings was analyzed for 4 disease-causing missense mutations (Malechka et al, 2022). The effect of a missense mutation on structural changes is evaluated in silico by molecular modeling. This suggests the possible association between predicted alteration or damage to the structure of CDHR1 protein and the severity of retinal phenotype. Molecular modeling of the interaction of integral membrane protein ADIPOR1 and CTRP5 was performed using crystallographic monomeric structures and refined by molecular dynamics (Miyagishima et al., 2021). The structure of CTRP5 was modeled as a trimer. To further support that native CTPR5 interacts with ADIPOR1, we modeled in silico interactions using existing structural models for both proteins. The heterocomplex of two structures refined in water demonstrates that a CTRP5 trimer can be docked to the binding domain of ADIPOR1 positioned in a lipid membrane and may compete with ADIPOR1 ligand ADIPONECTIN for the receptor's binding cavity. Computer simulation of CTRP5 binding to ADIPOR1 reveals that the Arg163 residue lies on the interface of CTRP5 expected to dock with ADIPOR1 and predicts amino acid residues that are most likely to show interaction. The serine to arginine change in CTRP5 adds a positive charge that repels the Arg122 on ADIPOR1s surface, increasing their C-C distance by 87.5% (native: 4.8; mutant: 9.0). Similar changes are predicted for other recently reported disease-associated CTRP5 variants (p.Gly216Cys, p.Prp188Thr) that are also close to its surface and are expected to repel ADIPOR1, reducing the likelihood of interaction compared to the WT protein. Overall, these results suggest that the mutant protein not only reduces the secretion of the WT counterpart but likely also reduces its binding affinity to the ADIPOR1 receptor. Results of our computational study were incorporated into the ocular proteome website at the NEI Commons (https://neicommons.nei.nih.gov/#/proteomeData). The latest version of the ocular proteome website contains in-silico predictions for 1,411,100 missense mutations associated with 112 protein structures from 164 inherited eye diseases.