Design of small molecules interacting with wild-type and mutant rhodopsins

Rhodopsin is a class A member of the G protein-coupled receptor (GPCR) family that detects light in the rod photoreceptor cells, and initiates the visual transduction cascade through the isomerization of 11-cis to all-trans retinal. Rhodopsin mutations and altered concentration of the retinal cofactor are associated with retinitis pigmentosa (RP), a group of degenerative diseases that lead to progressive loss of photoreceptor cells and blindness. Mutations at the N-terminal tail of rhodopsin are the most common mutations associated with RP, causing receptor misfolding and mislocalization, while mutations at the C-terminal tail of rhodopsin cause mistrafficking from Golgi to the rod outer segment, and are notably associated with more aggressive RP forms, thus becoming of particular interests for the development of therapeutic strategies. Levels of endogenous retinal cofactor are finely controlled by a series of enzymes, including the aldehyde dehydrogenase (ALDH) 1A3 isoform. Although its role still unclear, ALDH1A3 enzymatic activity may correlate with retinal concentration, and its modulation may have therapeutic implications in RP 2. However, limited experimental evidence and the lack of structural details that may clarify the role of rhodopsin mutations and ALDH1A3 in RP currently hamper the drug design process. In this work, these limitations have been addressed using three different computational modelling approaches: The first approach is based on the evidence that the correct folding and localization of mutant rhodopsin can be partially rescued by small chaperone ligands, which are able to bind rhodopsin in the 11-cis retinal binding site. To identify new small molecules that could act as a molecular chaperone for misfolded P23H mutant rhodopsin, the homology model of P23H mutant rhodopsin was generated and used as rigid receptor for a docking-based virtual screening carried out on different libraries of compounds. Following the virtual screening procedure, candidate hits were selected based on a combination of docking score, docking pose, and polar/hydrophobic interactions with the receptor and they were further investigated for their ability to restore the correct localization of mutant rhodopsin on the cell membrane by experimental validation in cell-based assay. Interestingly, one molecule (i.e., compound A3) was found to restore the correct protein folding and membrane localization. The second approach describes a computational workflow for studying clinically relevant C-terminal rhodopsin mutants in RP and to provide structural insights into the molecular mechanisms involved in the trafficking of WT and P347 mutant rhodopsin to the rod membrane. The conformational space of full-length WT and P347 mutant rhodopsin was explored by molecular dynamics (MD) simulations. The WT VAPA-COOH motif adopts a unique conformation that is not found in pathological mutants, suggesting that structural features could better explain the pathogenicity of P347 rhodopsin mutants than physicochemical or steric determinants. Interestingly, MD simulations of isolated and N-capped deca-peptides corresponding to the C-terminal tail of WT and P347 mutant rhodopsin showed results that are highly comparable to the full-length system, suggesting that deca-peptides are reliable model systems for studying C-terminal rhodopsin mutations in silico. The last approach focuses on the role of ALDH1A3 in the metabolism of retinal cofactor. Here we hypothesize that inhibition of ALDH1A3 may increase the level of endogenous 11-cis retinal in RP, given that one of the causes of the disease is related to the low cellular levels of the retinal cofactor. Starting from the reference ALDH inhibitor GA11, previously developed by our group, a novel series of imidazo[1,2-a]pyridines was developed and optimized by means of a structure-based approach. These novel compounds were evaluated in vitro for their activity and selectivity against the ALDH1A family and investigated through X-ray crystallography and molecular modeling studies for their ability to interact with the catalytic site of the 1A3 isoform. Overall, computational approach provided structural and thermodynamics information that are essential for the design and optimization of small molecules ALDH1A3 inhibitors. As a side project of this thesis, the main sequence and structural features of the RNA-dependent RNA polymerase of emerging RNA viruses, including the most recent SARS-CoV-2, were reviewed and the main limitation that prevent the successful application of the structure-based drug design approach in antiviral drug discovery process were discussed 8. In this context, the repositioning of sofosbuvir, approved for HCV infections, as an anti-Flaviviruses drug, was proposed through experimental and computational evaluations. In general, the data collected in this thesis enhance the use of computational techniques in a context of multidisciplinary effort to improve drug design and drug discovery approach in different research fields.