Fabiana Caporuscio

Structure-Based Design of Potent Aromatase Inhibitors by High-Throughput Docking

Cytochrome P450 aromatase catalyzes the conversion of androgen substrates into estrogens. Aromatase inhibitors (AIs) have been used as first-line drugs in the treatment of estrogen-dependent breast cancer in postmenopausal women. However, the search for new, more potent, and selective AIs still remains necessary to avoid the risk of possible resistances and reduce toxicity and side effects of current available drugs. The publication of a high resolution X-ray structure of human aromatase has opened the way to structure-based virtual screening to identify new small-molecule inhibitors with structural motifs different from all known AIs. In this context, a high-throughput docking protocol was set up and led to the identification of nanomolar AIs with new core structures.

Application of a post-docking procedure based on MM-PBSA and MM-GBSA on single and multiple protein conformations

In the last decades, molecular docking has emerged as an increasingly useful tool in the modern drug discovery process, but it still needs to overcome many hurdles and limitations such as how to account for protein flexibility and poor scoring function performance. For this reason, it has been recognized that in many cases docking results need to be post-processed to achieve a significant agreement with experimental activities. In this study, we have evaluated the performance of MM-PBSA and MM-GBSA scoring functions, implemented in our post-docking procedure BEAR, in rescoring docking solutions. For the first time, the performance of this post-docking procedure has been evaluated on six different biological targets (namely estrogen receptor, thymidine kinase, factor Xa, adenosine deaminase, aldose reductase, and enoyl ACP reductase) by using i) both a single and a multiple protein conformation approach, and ii) two different software, namely AutoDock and LibDock. The assessment has been based on two of the most important criteria for the evaluation of docking methods, i.e., the ability of known ligands to enrich the top positions of a ranked database with respect to molecular decoys, and the consistency of the docking poses with crystallographic binding modes. We found that, in many cases, MM-PBSA and MM-GBSA are able to yield higher enrichment factors compared to those obtained with the docking scoring functions alone. However, for only a minority of the cases, the enrichment factors obtained by using multiple protein conformations were higher than those obtained by using only one protein conformation.

Structure-based design of small-molecule protein–protein interaction modulators: the story so far

As the pivotal role of protein-protein interactions in cell growth, transcriptional activity, intracellular trafficking, signal transduction and pathological conditions has been assessed, experimental and in silico strategies have been developed to design protein-protein interaction modulators. State-of-the-art structure-based design methods, mainly pharmacophore modeling and docking, which have succeeded in the identification of enzyme inhibitors, receptor agonists and antagonists, and new tools specifically conceived to target protein-protein interfaces (e.g., hot-spot and druggable pocket prediction methods) have been applied in the search for small-molecule protein-protein interaction modulators. Many successful applications of structure-based design approaches that, despite the challenge of targeting protein-protein interfaces with small molecules, have led to the identification of micromolar and submicromolar hits are reviewed here.

1-[(3-Aryloxy-3-aryl)propyl]-1H-imidazoles, New Imidazoles with Potent Activity against Candida albicans and Dermatophytes. Synthesis, Structure−Activity Relationship, and Molecular Modeling Studies

New 1-[(3-aryloxy-3-aryl)propyl]-1 H-imidazoles were synthesized and evaluated against Candida albicans and dermatophytes in order to develop structure-activity relationships (SARs). Against C. albicans the new imidazoles showed minimal inhibitory concentrations (MICs) comparable to those of ketoconazole, miconazole, and econazole, and were more potent than fluconazole. Several derivatives ( 10, 12, 14, 18- 20, 24, 28, 29, 30, and 34) turned out to be potent inhibitors of C. albicans strains resistant to fluconazole, with MIC values less than 10 microg/mL. Against dermatophytes strains, compounds 20, 25, and 33 (MIC <or= 5 microg/mL) were equipotent to ketoconazole, econazole, and miconazole. SARs of imidazoles 10- 44 were rationalized with reasonable accuracy by a previously developed quantitative pharmacophore for antifungal agents.

Discovery of Multitarget Antivirals Acting on Both the Dengue Virus NS5-NS3 Interaction and the Host Src/Fyn Kinases

This study describes the discovery of novel dengue virus inhibitors targeting both a crucial viral protein-protein interaction and an essential host cell factor as a strategy to reduce the emergence of drug resistance. Starting from known c-Src inhibitors, a virtual screening was performed to identify molecules able to interact with a recently discovered allosteric pocket on the dengue virus NS5 polymerase. The selection of cheap-to-produce scaffolds and the exploration of the biologically relevant chemical space around them suggested promising candidates for chemical synthesis. A series of purines emerged as the most interesting candidates able to inhibit virus replication at low micromolar concentrations with no significant toxicity to the host cell. Among the identified antivirals, compound 16i proved to be 10 times more potent than ribavirin, showed a better selectivity index and represents the first-in-class DENV-NS5 allosteric inhibitor able to target both the virus NS5-NS3 interaction and the host kinases c-Src/Fyn.

Molecular Dynamics Simulations and Classical Multidimensional Scaling Unveil New Metastable States in the Conformational Landscape of CDK2.

Protein kinases are key regulatory nodes in cellular networks and their function has been shown to be intimately coupled with their structural flexibility. However, understanding the key structural mechanisms of large conformational transitions remains a difficult task. CDK2 is a crucial regulator of cell cycle. Its activity is finely tuned by Cyclin E/A and the catalytic segment phosphorylation, whereas its deregulation occurs in many types of cancer. ATP competitive inhibitors have failed to be approved for clinical use due to toxicity issues raised by a lack of selectivity. However, in the last few years type III allosteric inhibitors have emerged as an alternative strategy to selectively modulate CDK2 activity. In this study we have investigated the conformational variability of CDK2. A low dimensional conformational landscape of CDK2 was modeled using classical multidimensional scaling on a set of 255 crystal structures. Microsecond-scale plain and accelerated MD simulations were used to populate this landscape by using an out-of-sample extension of multidimensional scaling. CDK2 was simulated in the apo-form and in complex with the allosteric inhibitor 8-anilino-1-napthalenesulfonic acid (ANS). The apo-CDK2 landscape analysis showed a conformational equilibrium between an Src-like inactive conformation and an active-like form. These two states are separated by different metastable states that share hybrid structural features with both forms of the kinase. In contrast, the CDK2/ANS complex landscape is compatible with a conformational selection picture where the binding of ANS in proximity of the αC helix causes a population shift toward the inactive conformation. Interestingly, the new metastable states could enlarge the pool of candidate structures for the development of selective allosteric CDK2 inhibitors. The method here presented should not be limited to the CDK2 case but could be used to systematically unmask similar mechanisms throughout the human kinome.

A dynamic target-based pharmacophoric model mapping the CD4 binding site on HIV-1 gp120 to identify new inhibitors of gp120-CD4 protein-protein interactions.

A dynamic target-based pharmacophoric model mapping the CD4 binding site on HIV-1 gp120 was built and used to identify new hits able to inhibit gp120-CD4 protein-protein interactions. Two compounds showed micromolar inhibition of HIV-1 replication in cells attributable to an interference with the entry step of infection, by direct interaction with gp120. Inactivity of compounds toward a M475I strain suggested specific contacts with the Phe43 cavity of gp120.

Use of large multiconformational databases with structure-based pharmacophore models for fast screening of commercial compound collections.

In the last years high-throughput pharmacophore screenings have been rediscovered as an effective and rapid tool for guiding the selection of new hit compounds with predefined biological activity. This renaissance has also been fostered by the current possibility to generate pharmacophore hypotheses directly from crystallographic, NMR or computational models of protein-ligand complexes [1,2]. Indeed the pharmacophore notion provides a powerful way to identify and compare structural features across a large set of molecules and the possibility to screen virtually millions of compounds. At the present time pharmacophore screenings are further stimulated by the increasing number of chemical vendors that offer their catalogs of chemical compounds also for this purpose. Unfortunately, such an advantage is in many cases blunted by the fact that an high-throughput pharmacophore screening campaign takes time during the preparation steps of compound libraries, e.g. preparing ligand structures with full hydrogen, tautomers, stereoisomers and, most importantly, conformers. The latter point is extremely important in the context of pharmacophore screenings since the matching of a ligand molecule to a pharmacophore hypothesis is dependent upon the molecular conformation of the ligand. Furthermore, there is no specific way to predict what conformations are the biologically active ones for a given biological target. To address these issues we have recently introduced CoCoCo, a suite of free muticonformational databases that can be used for such pharmacophore screenings [3,4]. Here we will present how these ready-to-use chemical databases, that bear multiconformational information for each ligand, may provide a straightforward and time-effective way to select candidate active compounds. Different cases will be analyzed to highlight how different factors, e.g. pharmacophore hypotheses and conformational states, may influence the outcome of high-throughput screenings. Finally, a proof-of-concept experimental study that strongly supports these approaches will be presented.

Probing an allosteric pocket of CDK2 with small-molecules.

The availability of well‐characterized allosteric modulators is crucial for investigating the allosteric regulation of protein function. In a recently identified inactive conformation of cyclin‐dependent kinase 2 (CDK2), an open allosteric pocket was detected and proposed as a site to accommodate allosteric inhibitors. Previous structure‐based approaches allowed the identification of a hit compound expected to bind to this pocket. Herein we report the characterization of this compound by X‐ray crystallography, which surprisingly provided a chemical structure different from that previously reported. Therefore, the compound was synthesized and completely characterized. X‐ray structures of the synthesized and purchased compounds were found to be superimposable. A reaction mechanism was proposed to explain the formation of the structure indicated by crystallography. Moreover, a stereoselective synthesis was developed to evaluate the biological activity of the pure stereoisomers. Modeling studies were performed to unveil the details of the interaction with CDK2. The activity of the obtained compounds was evaluated with various biological assays. Mutagenesis experiments confirmed binding to the allosteric pocket. Finally, the allosteric ligands were shown to inhibit the growth of lung (A549) and ovarian (SKOV3) cancer cell lines. Therefore, this report presents a thorough chemical and biological characterization of the first small‐molecule ligands to be used as probes to study the allosteric modulation of CDK2 activity.

Contribution of the residue at position 4 within classical nuclear localization signals to modulating interaction with importins and nuclear targeting

Nuclear import involves the recognition by importin (IMP) superfamily members of nuclear localization signals (NLSs) within protein cargoes destined for the nucleus, the best understood being recognition of classical NLSs (cNLSs) by the IMPα/β1 heterodimer. Although the cNLS consensus [K-(K/R)-X-(K/R) for positions P2-P5] is generally accepted, recent studies indicated that the contribution made by different residues at the P4 position can vary. Here, we apply a combination of microscopy, molecular dynamics, crystallography, in vitro binding, and bioinformatics approaches to show that the nature of residues at P4 indeed modulates cNLS function in the context of a prototypical Simian Virus 40 large tumor antigen-derived cNLS (KKRK, P2-5). Indeed, all hydrophobic substitutions in place of R impaired binding to IMPα and nuclear targeting, with the largest effect exerted by a G residue at P4. Substitution of R with neutral hydrophobic residues caused the loss of electrostatic and van der Waals interactions between the P4 residue side chains and IMPα. Detailed bioinformatics analysis confirmed the importance of the P4 residue for cNLS function across the human proteome, with specific residues such as G being associated with low activity. Furthermore, we validate our findings for two additional cNLSs from human cytomegalovirus (HCMV) DNA polymerase catalytic subunit UL54 and processivity factor UL44, where a G residue at P4 results in a 2-3-fold decrease in NLS activity. Our results thus showed that the P4 residue makes a hitherto poorly appreciated contribution to nuclear import efficiency, which is essential to determining the precise nuclear levels of cargoes.