Search

The deregulation of GSK-3β (glycogen synthase kinase-3β) is involved in the pathogenesis of diverse diseases, such as cancer, diabetes and neurodegenerative disorders. Therefore, GSK-3β has become an attractive target for the design and development of new inhibitors for pathologies that present many limitations in therapeutic treatment. In the present thesis, we report a fast and efficient protocol for the expression of GSK-3β in insect cells with the baculovirus system, and employ purified GSK-3β in two different drug discovery projects. The first project describes the perspective validation of a novel computational method based on adiabatic bias molecular dynamics (ABMD) that aims to simulate protein-ligand unbinding events. Surface plasmon resonance (SPR) experiments performed on a series of GSK-3β inhibitors confirmed a coherence between predicted and experimental koff, highlighting the potential of this method for the calculation of residence times in hit-to-lead and lead optimization phases of drug discovery programs. Three novel X-ray crystal structures of pyrazine inhibitors in complex with GSK-3β are reported, providing further details in the binding mode of ATP-competitive GSK-3β inhibitors. The second project involves the structural and biophysical characterization of known GSK-3β allosteric inhibitors, and, in parallel, the discovery of novel allosteric modulators through computational and biophysical techniques. Herein, we describe the identification of promising molecules that have been selected by virtual screening and microscale thermophoresis (MST) analysis that display affinity to GSK-3β in the micromolar range. Taken together, our results provide useful insights for future rational drug design and discovery of small selective GSK-3β inhibitors.

Virtually all biochemical activities are mediated by the organization and recognition of biological macromolecules. An accurate characterization of the thermodynamics and kinetics governing the formation of supramolecular complexes is required to deeply understand the molecular principles driving all biological interactions. Thermodynamics provides the driving force of protein-ligand binding and is quantified by the binding free energies or the equilibrium dissociation constants. Since the interacting partners are out of equilibrium in vivo, the thermodynamic description of binding needs to be complemented by the knowledge of the kinetic rates. Nowadays, various biophysical experimental techniques can determine thermodynamic and kinetic properties, which are still difficult to be efficiently predicted by computational methods mainly because of the limited force field accuracy and the high computational cost. During my Ph.D., I applied molecular dynamics (MD)-based methods to characterize the thermodynamics and kinetics of inter-molecular interactions. First, I worked on a new enhanced MD-based protocol to simulate protein-ligand dissociation events. This approach provides a realistic description of the evolution of the system to an external perturbation accounting for the natural forces driving the dissociation mechanisms. By applying this computational approach to two pharmaceutically relevant kinases, I was able to rank two series of compounds on unbinding kinetics and to get qualitative mechanistic and path information on the underlying unbinding events, providing additional valuable information to be used in the optimization of lead compounds. Then, I developed an innovative computational method to estimate free energies applicable to systems of arbitrary complexity. Despite the number of challenges to be overcome, the method is very promising being able to provide accurate free energy estimates. Therefore, computer simulations emerged as a valuable tool to obta

High-mobility Group Box 1 (HMGB1) is an abundant protein present in all mammalian cells and involved in several processes. During inflammation or tissue damage, HMGB1 is released in the extracellular space and, depending on its redox state, can form a heterocomplex with CXCL12. The heterocomplex acts exclusively via the chemokine receptor CXCR4 enhancing leukocyte recruitment. Here, we used multi-microsecond molecular dynamics (MD) simulations to elucidate the effect of the disulfide bond on the structure and dynamics of HMGB1. The results of the MD simulations show that the presence or lack of the disulfide bond between Cys23 and Cys45 modulates the conformational space explored by HMGB1, making the reduced protein more suitable to form a complex with CXCL12.

Modeling a complex system is almost invariably a challenging task. The incorporation of experimental observations can be used to improve the quality of a model and thus to obtain better predictions about the behavior of the corresponding system. This approach, however, is affected by a variety of different errors, especially when a system simultaneously populates an ensemble of different states and experimental data are measured as averages over such states. To address this problem, we present a Bayesian inference method, called “metainference,” that is able to deal with errors in experimental measurements and with experimental measurements averaged over multiple states. To achieve this goal, metainference models a finite sample of the distribution of models using a replica approach, in the spirit of the replica-averaging modeling based on the maximum entropy principle. To illustrate the method, we present its application to a heterogeneous model system and to the determination of an ensemble of structures corresponding to the thermal fluctuations of a protein molecule. Metainference thus provides an approach to modeling complex systems with heterogeneous components and interconverting between different states by taking into account all possible sources of errors.

Transcription factors are proteins able to bind DNA and induce the transcription of specific genes. Consequently, they play a pivotal role in multiple cellular pathways and are frequently over-expressed or dysregulated in cancer. Here, we will focus on a specific “signal transducer and activator of transcription” (STAT3) factor that is involved in several pathologies, including cancer. For long time, the mechanism by which STAT3 exerts its cellular functions has been summarized by a three steps process: (1) Protein phosphorylation by specific kinases, (2) dimerization promoted by phosphorylation, (3) activation of gene expression by the phosphorylated dimer. Consequently, most of the inhibitors reported in literature aimed at blocking phosphorylation and dimerization. However, recent observations reopened the debate and the entire functional mechanism has been revisited stimulating the scientific community to pursue new inhibition strategies. In particular, the dimerization of the unphosphorylated species has been experimentally demonstrated and specific roles proposed also for these dimers. Despite difficulties in the expression and purification of the full length STAT3, structural biology investigations allowed the determination of atomistic structures of STAT3 dimers and several protein domains. Starting from this information, computational methods have been used both to improve the understanding of the STAT3 functional mechanism and to design new inhibitors to be used as anticancer drugs. In this review, we will focus on the contribution of structural biology to understand the roles of STAT3, to design new inhibitors and to suggest new strategies of pharmacological intervention.

In order to characterise the dynamics of proteins, a well-established method is to incorporate experimental parameters as replica-averaged structural restraints into molecular dynamics simulations. Here, we justify this approach in the case of interproton distance information provided by nuclear Overhauser effects by showing that it generates ensembles of conformations according to the maximum entropy principle. These results indicate that the use of replica-averaged structural restraints in molecular dynamics simulations, given a force field and a set of experimental data, can provide an accurate approximation of the unknown Boltzmann distribution of a system.

Based on the development of new algorithms and growth of sequence databases, it has recently become possible to build robust higher-order sequence models based on sets of aligned protein sequences. Such models have proven useful in de novo structure prediction, where the sequence models are used to find pairs of residues that co-vary during evolution, and hence are likely to be in spatial proximity in the native protein. The accuracy of these algorithms, however, drop dramatically when the number of sequences in the alignment is small. We have developed a method that we termed CE-YAPP (CoEvolution-YAPP), that is based on YAPP (Yet Another Peak Processor), which has been shown to solve a similar problem in NMR spectroscopy. By simultaneously performing structure prediction and contact assignment, CE-YAPP uses structural self-consistency as a filter to remove false positive contacts. Furthermore, CE-YAPP solves another problem, namely how many contacts to choose from the ordered list of covarying amino acid pairs. We show that CE-YAPP consistently improves contact prediction from multiple sequence alignments, in particular for proteins that are difficult targets. We further show that the structures determined from CE- YAPP are also in better agreement with those determined using traditional methods in structural biology.

Genome integrity is constantly affected by DNA damage and replication errors due to exogenous and endogenous sources. Cells have developed several strategies to counteract these threats detecting DNA damages and mediating its repair through different mechanisms depending on the specificity of the DNA lesion. Cancer cells are more sensitive to these insults compared to normal cells and most therapeutic strategy indeed rely on inducing specific lethal DNA damages. The urgent need for new therapeutic strategies in cancer treatment is due to the frequent development of resistance to therapeutic induced DNA damage. Through a multidisciplinary project, comprising a computational, a synthetic, a biology as well as a biophysical contribution, we aim at applying a synthetic lethality approach in cancer cell therapy by preventing, with small molecules, the viable hRAD51-BRCA2 interaction. The latter interaction plays a key role in the repair of double strand breaks (DSBs) within the homologous recombination (HR) pathway. The leading idea of our project is mimicking the synthetic lethality induced in BRCA2 defective oncology patients by olaparib, a poly ADP ribose polymerase (PARP) inhibitor. We aim at increasing cancer cells sensitivity to PARPi and at broadening PARPi applicability by administering, in combination with the latter inhibitors, hRAD51-BRCA2 small molecules disruptors also to individuals without BRCA2 mutations. The goal of my PhD research project is the characterization of the hRAD51-BRCA2 interaction through a biophysical approach, in order to gain kinetic and thermodynamic insights on this critical junction of the HR pathway, and through a structural approach, to elucidate the atomic details of the interaction. My research work is concurrently a critical support for the overall project, which aims at the development of a new drug and at the validation of a broaden applicability of the synthetic lethality concept.

Characterizing thermodynamics and kinetics of molecular systems is the ultimate goal of biophysics. In drug discovery, this information becomes essential. Understanding local and global rearrangements, how formation and disruption of biomolecular complexes occur, the molecular determinants involved and the preferred pathways followed, contribute to forming a solid ground for the development of new drugs. Quantifying specific kinetic parameters, such as off rates and the closely related residence time, is increasingly being incorporated in the drug optimization phase. Several experimental techniques established to study and quantify kinetic features. Conversely, the computational counterpart still faces severe challenges, such as accessing the time scales at which these slow events occur, while acquiring acceptable statistics. During this PhD program, we explored current, state-of-the-art computational methods, and combination thereof, to study kinetic properties of pharmaceutically relevant biomolecules. In particular, we applied different protocols to three test systems. In the first case, we reconstructed the free energy surface of an intrinsically disordered protein and calculated interconversion rates between the differently folded states identified. In the second application, Markov State Models were employed to identify relevant states along the protein-ligand binding pathway. Using these states as a template, a putative pathway on which computing the free energy profile associated with the binding process was determined. As for the third test case, we performed unbinding simulations on a series of ligands and prioritized them according to their average computational unbinding time. The obtained ranking was subsequently confirmed by performing experimental assays. Despite clear limitations, the picture arising from the studies was encouraging. Computer simulations emerged undoubtedly as a valuable instrument for assessing kinetic properties of biomolecular sys