In-silico methods applied on druggable proteins to identify transient pockets: new approaches for studying drug-target molecular mechanisms. A case study on CFTR.

Cystic Fibrosis is the most common genetical lethal disorder in Caucasians and it is caused by the mutation of the Cystic Fibrosis Transmembrane conductance Regulator (CFTR) protein. Up to now, for the treatment of cystic fibrosis patients carrying at least one copy of CFTR deleted of the phenylalanine 508 (F508del-CFTR), the worldwide most frequent mutation, only four drugs have been approved to be used in combination or alone. All the approved compounds have been developed, studied, and are currently commercialized by Vertex Pharmaceuticals. Despite the benefits of these marketed drugs, they are still too expensive for many countries, and they cannot be prescribed to all patients. Thus remains a pressing need to better understand the CFTR structure-function relationship, and the binding site and molecular mechanism of already approved drugs, to identify other CFTR modulators for the rescue of the mutated protein, in particular, F508del-CFTR. On these bases my research activity has been focused on a deep study of the protein function, investigating its three-dimensional structure and dynamics in complex with the already approved CF drug lumacaftor and new possible CFTR modulators by means of drug repositioning. An optimized model, obtained before the starting of my PhD, of the F508del-CFTR protein and a library of pockets, in which an interesting large druggable pocket (DP1) was identified using lumacaftor as a template, has been used for the following drug repositioning strategy. An in-house database which included 846 drugs and nutraceuticals approved by AIFA (actually implemented to more than 10000 molecules from AIFA and Drugbank database) was built, drawing their 3D structure with the right protonation state of the drugs, and then screened by docking against F508del-CFTR. Among the best eleven repositioned compounds identified within this procedure, tadalafil was one that has been already taken into consideration for cystic fibrosis therapy, confirming the goodness of this approach. Quercetin emerged as the best ligand among the eleven selected, suggesting that small molecules could give a consistent contribution in the search for new CFTR modulators. Focusing on this concept, the several DP1 sub-pockets surrounding the lumacaftor binding region were explored, searching for the most druggable ones and in the meantime scouting small molecules able to fill the transient druggable DP1 sub-pockets and synergize with lumacaftor. At the end of this procedure, NAM was found as a possible hit. Moreover, during my PhD project, my computational studies have been also focused on two proteins of therapeutical interest, which mutations are causative of rare genetical disorders: the Leucine-Rich and ImmunoGlobulin-like domains 2 (LRIG2) and the Nucleoporin 98 (NUP98), whose mutations lead to the Urofacial Syndrome and a phenotype resembling the Rothmund-Thomson syndrome, respectively. The study of LRIG2 involved the investigation of the role of the first immunoglobulin-like domain (Ig1) of the LRIG2 protein, and its deletion and mutations, in the LRIG2 homodimerization. The LRIG2 homodimerization was predicted in silico and its dimerization interface was computationally characterized. Then, by means of accelerated molecular dynamic simulations, the central role of the Ig1 domain in the LRIG2 dimerization was furthermore validated by studying the impact of the Ig1 domain mutations, described in the literature as pathogenic, in the context of the monomeric LRIG2. This advanced molecular dynamic technique allowed to clarify the role of these mutations in the impairment of the LRIG2 homodimerization. Eventually, regarding the study of NUP98, a novel germline alteration (G28D) located in the unstructured N-terminal of the NUP98, which is characterized by phenylalanine-glycine (FG) repeats, was computationally evaluated. Differences in the dynamic behavior between the wild type and G28D mutated protein were observed, which are produced from a dispersion of the intramolecular cohesion elements (FG repeats) leading to more elongated conformational states of the unstructured N-terminal of the NUP98 mutant in comparison to the wild type. Those differences may affect the role of NUP98 as a multi-docking station for RNA and proteins, and its folding process when a specific interaction is required.