No abstract

Glycogen synthase kinase 3 (GSK-3) is an important drug target for human severe unmet diseases. Discovery and/or design of allosteric kinase modulators are gaining importance in this field not only for the increased selectivity of this kind of compounds but also for the subtle modulation of the target. This last point is of utmost importance for the GSK-3 inhibition as a therapeutic approach. GSK-3 activity is completely necessary for life, and only the aberrant overactivity found in the pathologies should be inhibited with its inhibitors treatment. We performed here a search for the druggable sites on the enzyme using the fpocket algorithm with the aim to provide allosteric potential binding sites on it and new clues for further drug discoveries. Moreover, our results allowed us to determine the binding sites of different GSK-3 ATP noncompetitive inhibitors, such as manzamine A and the new small molecule VP 0.7, providing evidence for potential allosteric inhibition of GSK-3.

Cumulative evidence strongly supports that glycogen synthase kinase-3 (GSK-3) is a pathogenic molecule when it is up-dysregulated, emerging as an important therapeutic target in severe unmet human diseases. GSK-3 specific inhibitors might be promising effective drugs for the treatment of devastating pathologies such as neurodegenerative diseases, stroke, and mood disorders. As GSK-3 has the ability to phosphorylate primed substrates, small molecules able to bind to this site should be perfect drug candidates, able to partially block the activity of the enzyme over some specific substrates. Here, we report substituted 5-imino-1,2,4-thiadiazoles as the first small molecules able to inhibit GSK-3 in a substrate competitive manner. These compounds are cell permeable, able to decrease inflammatory activation and to selectively differentiate neural stem cells. Overall, 5-imino-1,2,4-thiadiazoles are presented here as new molecules able to decrease neuronal cell death and to increase endogenous neurogenesis blocking the GSK-3 substrate site.

The role of polarization in the stabilization of a series of biologically relevant alkali and alkaline-earth metal cations (Li(+), Na(+), K(+), Mg(2+) and Ca(2+)) with the pi-electron distribution of benzene is examined by means of MP2 computations using Sadlej's basis set. In all cases a full description of the energy profile for the approach of the metal cation along the axis normal to the molecular plane of benzene has been performed. Analysis of the different contributions to the interaction energy, performed within the framework of the symmetry-adapted perturbation theory (SAPT), illustrates the important role of the induction component in the definition of the geometrical and energetic properties of the pathway leading to the formation of cation-pi complex. Finally, the ability of classical polarization models based on models of implicitly and explicitly interacting distributed isotropic polarizabilities to describe the induction term has been examined and discussed in the context of the generation of new polarizable force fields.

Different classical models for monovalent ions (typically used to neutralize proteins or nucleic acids) are available in the literature and are widely used in molecular dynamics simulations without a great knowledge of their quality, consistency with the macromolecular force field and impact on the global simulation results. In this paper the ability of several of the most popular ion models to reproduce both quantum mechanics and experimental results is examined. Artefacts due to the use of incorrect ion models in molecular dynamics simulations of concentrated solutions of NaCl and KCl in water and of a short DNA duplex in 500 mM aqueous solutions of NaCl and KCl have been analyzed. Our results allow us to discuss the robustness and reliability of different ion models and to highlight the source of potential errors arising from non-optimal models. However, it is also found that the structural and dynamic characteristics of DNA (as an example of a heavily charged macromolecule) in near-physiological conditions are quite independent of the ion model used, providing support to most already-published simulations of macromolecules.