Computational methods for the analysis of protein/ligand complexes: applications to bacterial transcriptional regulation, drug discovery and plant immunology

Different numerical techniques can be used to study the binding of a protein/ligand complex, including Molecular Dynamics methods (more accurate but time consuming) and Molecular Docking methods (fast but more approximated). Molecular Dynamics are based on the numerical solution of the Classical Mechanics equations in a many-body scenario. Molecular Docking is essentially a static method where the protein atoms positions are kept fixed and the ligand is manipulated according to a certain algorithm in order to find the best pose that the ligand itself can assume in the receptor binding pocket. The binding energy is computed as a metric of the stability of the complex. Both these computational techniques have been used to address three different applications: 1) the activation of the DNA transcription by the bacterial transcriptional regulator GabR (with Molecular Dynamics) 2) the search for new natural compounds to be used as drugs for the treatment of Medulloblastoma disease (with Molecular Docking) 3) the study of a class of flavoenzymes known as Berberine Bridge Enzyme-like (BBE-like) proteins, which have a prominent role in plant immunology (with Molecular Docking and Molecular Dynamics). The GabR study contributes to understand the molecular mechanism underlying the GabR transcription activation upon GABA binding. Four variants of the GabR transcriptional regulator have been studied with Molecular Dynamics technique at classical level. It has been proven that only the presence of GABA cofactor in both the binding pockets stiffens the GabR and promotes the transition form open to closed form of the protein possibly enabling it to carry out its biological function. Moreover, by simulating an asymmetrical form of GabR, it has been found that the asymmetric perturbation of the active site residues may suggest the existence of a form of allosteric interference between the two active sites. These studies are among the first published theoretical/computational works available in literature on this protein. The virtual screening on the natural compounds suitable for Medulloblastoma treatment has been carried out with Molecular Docking. The results suggest that epigallocatechin gallate and curcumin are the most promising ligands to be further investigated in laboratory. Molecular Docking suffers from an inner unreliability, thus the results must be confirmed, for example, by experimental methods. To this end, the numerical results obtained have been submitted to our experimental collaborators in order to finalize the work. The main outcome of this activity was the design of a suite of software tools which has proved to be instrumental also for the study on the BBE enzymes. The work on the BBEs has been focused on the OGOX1 protein. Four structures of the OGOX1 have been taken into account (high/low pH and with/without a sulfur bridge between two helices) since it has been observed that the activity of this enzyme is dependent on the pH and sulphite sensitive. A first hybrid computational/experimental work has been published on an international journal. Molecular Dynamics-guided mutagenesis experiments revealed that two aspartic acids residues, involved in the formation of two salt bridges, are responsible of the pH-dependent activity of the OGOX1 enzyme. A further mostly computational work, to be submitted to an international journal, aims at clarifying the detailed molecular determinants of the substrate size-dependent specificity of the OGOX1 protein to oligogalacturonides. These are the first theoretical/computational studies on the OGOX1/OG complexes.