Ligand specificity in Saccaromices cerevisiae G-Protein coupled receptor binding site design

Developing an algorithm to rationally modify the amino acid sequence of a protein to construct a binding site for a user-selected ligand could prove to be a powerful tool for future biotechnological applications. In a recent paper, we constructed a computational protocol capable of creating a binding site for epinephrine on the Ste2 protein of Saccharomyces cerevisiae. Modifying this protein, as it is the only GPCR expressed by this species, could be used for the development of whole-cell biosensors, where this yeast could react to the presence of a specific molecule in the environment. Although the strength of the binding between the mutant species of Ste2 and epinephrine had already been computationally validated, in this work, we focus on testing its specificity. We have collected a large structural dataset of experimentally known protein-ligand complexes and studied their similarity to a known complex involving epinephrine. We observed how the design algorithm can construct a pocket on Ste2 with a similarity to the experimental pocket recognising epinephrine that is very unlikely to occur by chance, both morphologically and in terms of hydrophobicity. Furthermore, it was found that our framework is fine enough to recognise that pockets recognising similar ligands are similar to each other, as intuitively expected. Although our results regarding the specificity of the designed binding site are satisfactory, we believe that including other physicochemical aspects to optimise during the design process could further improve the algorithm's performance.