Classification of Rhodopsin mutations by integrated in silico and in vitro analyses for screening of chaperon molecules to rescue misfolding.

Purpose: About 140 point mutations were identified in the rhodopsin gene (RHO) as cause of Autosomal Dominant Retinitis Pigmentosa (ADRP), a genetic degenerative disease causing blindness in later life. A recent analysis indicates that 89% of the biochemically characterized RHO mutants are misfolded, supporting the protein-misfolding disease model suitable for treatments with pharmacological chaperones. Characterization of the structural and molecular features of such mutants will support the development of rational drug design. Methods: Wild type RHO and 33 different RHO mutations were analyzed in silico either in the rhodopsin form bound to retinal or in the opsin form by thermal unfolding simulations combined with the graph-based Protein Structure Network analysis. In parallel, the same mutants were cloned in expression vectors and in vitro expressed in COS-7 cells either in the absence or presence of 9-cis retinal in the culture medium. The subcellular localization was analyzed with two monoclonal antibodies recognizing either the extracellular N-terminal or the intracellular C-terminal of RHO. Retention in the endoplasmic reticulum (ER) was assessed by analysis of co-localization with calnexin and calculation of the Pearson Correlation Coefficient (PCC) of co-localization. Results: In silico studies revealed that the selected ADRP RHO mutations share marked abilities to impair highly connected nodes in the protein structure network, i.e. hubs, essentially located in the retinal binding site, which participates to the stability of the protein. We defined computational indices whose combination led to a structural classification of the mutants. The in vitro level of analysis revealed increased ER retention and reduction of plasma membrane localization of most of the mutants compared to wild type RHO. We found a strong correlation of the perturbation indexes calculated by in silico analyses with PCC calculated by in vitro analyses. We could also characterize different abilities of the mutated proteins to be affected by treatment with 9-cis retinal. Conclusions: These two levels of analysis allowed a novel characterization of the different mutants to generate the first classification of ADRP RHO mutants based on a multiscale approach, i.e. at the cellular and atomic levels of detail. We also developed a PCC index to evaluate the effect of retinal on protein folding and protein localization at the plasma membrane. This knowledge will be our starting point for in silico screening of compounds able to bind the retinal site and act as chaperones. The in vitro studies have developed a quantitative analysis to assess therapeutic effects of chaperone molecules.