In the world plagued by the emergence of new diseases, it is essential that we accelerate the drug design process to develop new therapeutics against them. In recent years, deep learning-based methods have shown some success in ligand-based drug design. Yet, these methods face the problem of data scarcity while designing drugs against a novel target. In this work, the potential of deep learning and molecular modeling approaches was leveraged to develop a drug design pipeline, which can be useful for cases where there is limited or no availability of target-specific ligand datasets. Inhibitors of the homologues of the target protein were screened at the active site of the target protein to create an initial target-specific dataset. Transfer learning was used to learn the features of the target-specific dataset. A deep predictive model was utilized to predict the docking scores of newly designed molecules. Both these models were combined using reinforcement learning to design new chemical entities with an optimized docking score. The pipeline was validated by designing inhibitors against the human JAK2 protein, where none of the existing JAK2 inhibitors were used for training. The ability of the method to reproduce existing molecules from the validation dataset and design molecules with better binding energy demonstrates the potential of the proposed approach.

We investigate the probable proton-transfer pathways from the surface of human carbonic anhydrase II into the active site cavity through His-64 that has been widely implicated as a key residue along the proton-transfer path. A recursive analysis of hydrogen-bonded clusters in the static crystallographic structure shows that there is no complete path through His-64 in either of its experimentally detected conformations. Side chain conformational fluctuation of His-64 from its outward conformation toward the active site is found to provide a crucial dynamic connectivity needed to complete the path coupled to local reorganization of the protein structure and hydration. The energy and free energy barriers along the detected pathway have been estimated to derive the mechanism of His-64 rotation toward the active site. We also investigate a dynamical connectivity map that highlights networks of disordered water molecules that may promote a direct (and probably transient) access of the solvent to the active site. Our studies reveal how such solvent access channels may be related to the putative proton shuttle mediated by His-64. The paths thus identified can be potentially used as reaction coordinates for further studies on the molecular mechanism of enzyme action.

Fluorescent proteins (FPs) of the green fluorescent protein family blink and bleach like all fluorophores. However, contrary to organic dyes, the mechanisms by which transient losses of fluorescence occur in FPs have received little attention. Here, we focus on the photoactivatable IrisFP, for which a transient non-fluorescent chromophoric state with distorted geometry was recently reported (Adam, V.; et al. J. Am. Chem. Soc. 009, 131, 18063). We investigated the chemical nature of this blinked state by employing quantum chemical/molecular mechanical calculations. Our findings suggest two previously unidentified dark states that display similar distorted chromophores with a transiently ruptured π-electron system. Both are protonated at atom C(α) of the chromophore methylene bridge. Transient protonation may occur via proton transfer from the nearby Arg66 either in the triplet state T(1) after intersystem crossing or in an anionic radical (doublet) ground state. As Arg66 is conserved in green-to-red photoconvertible FPs, these dark states are predicted to be common to all these proteins. We also suggest that C(α) protonated dark states may accelerate photobleaching by favoring decarboxylation of the fully conserved Glu212.