SAnDReS 2.0: Development of machine-learning models to explore the scoring function space.

Classical scoring functions may exhibit low accuracy in determining ligand binding affinity for proteins. The availability of both protein-ligand structures and affinity data make it possible to develop machine-learning models focused on specific protein systems with superior predictive performance. Here, we report a new methodology named SAnDReS that combines AutoDock Vina 1.2 with 54 regression methods available in Scikit-Learn to calculate binding affinity based on protein-ligand structures. This approach allows exploration of the scoring function space. SAnDReS generates machine-learning models based on crystal, docked, and AlphaFold-generated structures. As a proof of concept, we examine the performance of SAnDReS-generated models in three case studies. For all three cases, our models outperformed classical scoring functions. Also, SAnDReS-generated models showed predictive performance close to or better than other machine-learning models such as KDEEP, CSM-lig, and ΔVinaRF20. SAnDReS 2.0 is available to download at https://github.com/azevedolab/sandres.

Applying the bioisosterism strategy to obtain lead compounds against SARS-CoV-2 cysteine proteases: An in-silico approach.

SARS-CoV-2 cysteine proteases are essential nonstructural proteins due to their role in the formation of the virus multiple enzyme replication-transcription complex. As a result, those functional proteins are extremely relevant targets in the development of a new drug candidate to fight COVID-19. Based on this fact and guided by the bioisosterism strategy, the present work has selected 126 out of 1050 ligands from DrugBank website. Subsequently, 831 chemical analogs containing bioisosteres, some of which became structurally simplified, were created using the MB-Isoster software, and molecular docking simulations were performed using AutoDock Vina. Finally, a study of physicochemical properties, along with pharmacokinetic profiles, was carried out through SwissADME and ADMETlab 2.0 platforms. The promising results obtained with the molecules encoded as DB00549_BI_005, DB04868_BI_003, DB11984_BI_002, DB12364_BI_006 and DB12805_BI_004 must be confirmed by molecular dynamics studies, followed by in vitro and in vivo empirical tests that ratify the advocated in-silico results.

Fragment-based Drug Discovery Successful Contributions to Current Pharmacotherapeutic Agents Arsenal against Aggressive Cancers: A Mini-Review.

After a decade of approval of the drug vemurafenib in 2011, the hopeless scenario imposed by some severe cancer types has been mitigated by the magic bullets developed through fragment-based drug discovery. Moreover, this recent approach to medicinal chemistry has been successfully practiced by academic laboratories and pharmaceutical industry workflows focused on drug design with an enhanced profile for chemotherapy of aggressive tumors. This mini-review highlights the successes achieved by these research campaigns in the fruitful field of the molecular fragment paradigm that resulted in the approval of six new anticancer drugs in the last decade (2011-2021), as well as several promising clinical candidates. It is a particularly encouraging opportunity for other researchers who want to become aware of the applicability and potency of this new paradigm applied to the design and development of powerful molecular weapons in the constant war against these merciless scourges of humanity.

Current Fragment-to-lead Approaches Starting from the 7-azaindole: The Pharmacological Versatility of a Privileged Molecular Fragment.

Fragment-based drug discovery is one of the most powerful paradigms in the recent context of medicinal chemistry and is being widely practiced by academic and industrial researchers. Currently, azaindoles are among the most exploited molecular fragments in pharmaceutical innovation projects inspired by fragment-to-lead strategies. The 7-azaindole is the most prominent representative within this remarkable family of pyrrolopyridine fragments, as it is present in the chemical structure of several approved antitumor drugs and also of numerous therapeutic candidates. In this paper, a brief overview on existing proofs of concept in the literature will be presented, as well as some recent works that corroborate 7-azaindole as a privileged and pharmacologically versatile molecular fragment.

A Study on the Anti-NF-κB, Anti-Candida, and Antioxidant Activities of Two Natural Plant Hormones: Gibberellin A4 and A7.

INTRODUCTION: Gibberellins (GA) are terpenoids that serve as important plant hormones by acting as growth and response modulators against injuries and parasitism. In this study, we investigated the in vitro anti-NF-κB, anti-Candida, and antioxidant activity of gibberellin A4 (GA4) and A7 (GA7) compounds, and further determined their toxicity in vivo. METHODS: GA4 and GA7 in vitro toxicity was determined by MTT method, and nontoxic concentrations were then tested to evaluate the GA4 and GA7 anti-NF-κB activity in LPS-activated RAW-luc macrophage cell culture (luminescence assay). GA4 in silico anti-NF-κB activity was evaluated by molecular docking with the software "AutoDock Vina", "MGLTools", "Pymol", and "LigPlot+", based on data obtained from "The Uniprot database", "Protein Data Bank", and "PubChem database". The GA4 and GA7 in vitro anti-Candida effects against Candida albicans (MYA 2876) were determined (MIC and MFC). GA7 was also evaluated regarding the viability of C. albicans preformed biofilm (microplate assay). In vitro antioxidant activity of GA4 and GA7 was evaluated against peroxyl radicals, superoxide anions, hypochlorous acid, and reactive nitrogen species. GA4 and GA7 in vivo toxicity was determined on the invertebrate Galleria mellonella larvae model. RESULTS: Our data show that GA4 at 30 µM is nontoxic and capable of reducing 32% of the NF-κB activation on RAW-luc macrophages in vitro. In vitro results were confirmed via molecular docking assay (in silico), since GA4 presented binding affinity to NF-κB p65 and p50 subunits. GA7 did not present anti-NF-κB effects, but exhibited anti-Candida activity with low MIC (94 mM) and MFC (188 mM) values. GA7 also presented antibiofilm properties at 940 mM concentration. GA4 did not present anti-Candida effects. Moreover, GA4 and GA7 showed antioxidant activity against peroxyl radicals, but did not show scavenging activity against the other tested radicals. Both compounds did not affect the survival of G. mellonella larvae, even at extremely high doses (10 g/Kg). CONCLUSION: Our study provides preclinical evidence indicating that GA4 and GA7 have a favorable low toxicity profile. The study also points to GA4 and GA7 interference with the NF-κB via, anti-Candida activity, and a peroxyl radical scavenger, which we argue are relevant biological effects.

MB-Isoster: A software for bioisosterism simulation.

Bioisosterism is a technique used in medicinal chemistry to optimize lead compounds in drug research. One can replace a substituent group in original molecule by another with similar physical chemistry properties and then test how this replacement affects biological activity. To help researchers in their bioisosteric replacement choose, computational efforts such as programs and databases was developed. In this article, it is presented MB-Isoster, a software that draws bioisosteric molecules. Starting from an input molecule, user selects a molecular subregion formed by connected atoms to be replaced and MB-Isoster queries an internal library to find bioisosteric substituents for selected subregion, and makes the bioisosteres. Another functionality is receptor-ligand pdb complex reading, in which nonbonded interactions are computed between receptor and ligand in a pdb file, helping in atom/subregion selection to bioisosteric replacement. Physical-chemical properties computing, and virtual screening evaluation is also available. MB-Isoster is freely available at http://molmod-cs.unifal-mg.edu.br/tools.html. © 2018 Wiley Periodicals, Inc.

Phylogenetic analysis of Dengue virus 1 isolated from South Minas Gerais, Brazil.

Dengue is a major worldwide public health problem, especially in the tropical and subtropical regions of the world. Primary infection with a single Dengue virus serotype causes a mild, self-limiting febrile illness called dengue fever. However, a subset of patients who experience secondary infection with a different serotype can progress to a more severe form of the disease, called dengue hemorrhagic fever. The four Dengue virus serotypes (1-4) are antigenically and genetically distinct and each serotype is composed of multiple genotypes. In this study we isolated one Dengue virus 1 serotype, named BR/Alfenas/2012, from a patient with dengue hemorrhagic fever in Alfenas, South Minas Gerais, Brazil and molecular identification was performed based on the analysis of NS5 gene. Swiss mice were infected with this isolate to verify its potential to induce histopathological alterations characteristic of dengue. Liver histopathological analysis of infected animals showed the presence of inflammatory infiltrates, hepatic steatosis, as well as edema, hemorrhage and necrosis focal points. Phylogenetic and evolutionary analyses based on the envelope gene provided evidence that the isolate BR/Alfenas/2012 belongs to genotype V, lineage I and it is probably derived from isolates of Rio de Janeiro, Brazil. The isolate BR/Alfenas/2012 showed two unique amino acids substitutions (SER222THRE and PHE306SER) when compared to other Brazilian isolates from the same genotype/lineage. Molecular models were generated for the envelope protein indicating that the amino acid alteration PHE 306 SER could contribute to a different folding in this region located within the domain III. Further genetic and animal model studies using BR/Alfenas/2012 and other isolates belonging to the same lineage/genotype could help determine the relation of these genetic alterations and dengue hemorrhagic fever in a susceptible population.

DBMODELING: a database applied to the study of protein targets from genome projects.

Genome sequencing efforts are providing us with complete genetic blueprints for hundreds of organisms. We are now faced with assigning, understanding, and modifying the functions of proteins encoded by these genomes. DBMODELING is a relational database of annotated comparative protein structure models and their metabolic pathway characterization, when identified. This procedure was applied to complete genomes such as Mycobacterium tuberculosis and Xylella fastidiosa. The main interest in the study of metabolic pathways is that some of these pathways are not present in humans, which makes them selective targets for drug design, decreasing the impact of drugs in humans. In the database, there are currently 1,116 proteins from two genomes. It can be accessed by any researcher at http://www.biocristalografia.df.ibilce.unesp.br/tools/. This project confirms that homology modeling is a useful tool in structural bioinformatics and that it can be very valuable in annotating genome sequence information, contributing to structural and functional genomics, and analyzing protein-ligand docking.

Molecular models of tryptophan synthase from mycobacterium tuberculosis complexed with inhibitors.

The development of new therapies against infectious diseases is vital in developing countries. Among infectious diseases, tuberculosis is considered the leading cause of death. A target for development of new drugs is the tryptophan pathway. The last enzyme of this pathway, tryptophan synthase (TRPS), is responsible for conversion of the indole 3-glycerol phosphate into indol and the condensation of this molecule with serine-producing tryptophan. The present work describes the molecular models of TRPS from Mycobacterium tuberculosis (MtTRPS) complexed with six inhibitors, the indole 3-propanol phosphate and five arylthioalkyl-phosphonated analogs of substrate of the alpha-subunit. The molecular models of MtTRPS present good stereochemistry, and the binding of the inhibitors is favorable. Thus, the generated models can be used in the design of more specific drugs against tuberculosis and other infectious diseases.

Molecular models of protein kinase 6 from Plasmodium falciparum.

Cyclin-dependent kinases (CDKs) have been identified as potential targets for development of drugs, mainly against cancer. These studies generated a vast library of chemical inhibitors of CDKs, and some of these molecules can also inhibit kinases identified in the Plasmodium falciparum genome. Here we describe structural models for Protein Kinase 6 from P. falciparum (PfPK6) complexed with Roscovitine and Olomoucine. These models show clear structural evidence for differences observed in the inhibition, and may help designing inhibitors for PfPK6 generating new potential drugs against malaria.

Structural bioinformatics study of PNP from Schistosoma mansoni.

The parasite Schistosoma mansoni lacks the de novo pathway for purine biosynthesis and depends on salvage pathways for its purine requirements. Schistosomiasis is endemic in 76 countries and territories and amongst the parasitic diseases ranks second after malaria in terms of social and economic impact and public health importance. The PNP is an attractive target for drug design and it has been submitted to extensive structure-based design. The atomic coordinates of the complex of human PNP with inosine were used as template for starting the modeling of PNP from S. mansoni complexed with inosine. Here we describe the model for the complex SmPNP-inosine and correlate the structure with differences in the affinity for inosine presented by human and S. mansoni PNPs.

Molecular models for shikimate pathway enzymes of Xylella fastidiosa.

The Xylella fastidiosa is a bacterium that is the cause of citrus variegated chlorosis (CVC). The shikimate pathway is of pivotal importance for production of a plethora of aromatic compounds in plants, bacteria, and fungi. Putative structural differences in the enzymes from the shikimate pathway, between the proteins of bacterial origin and those of plants, could be used for the development of a drug for the control of CVC. However, inhibitors for shikimate pathway enzymes should have high specificity for X. fastidiosa enzymes, since they are also present in plants. In order to pave the way for structural and functional efforts towards antimicrobial agent development, here we describe the molecular modeling of seven enzymes of the shikimate pathway of X. fastidiosa. The structural models of shikimate pathway enzymes, complexed with inhibitors, strongly indicate that the previously identified inhibitors may also inhibit the X. fastidiosa enzymes.

Structural bioinformatics study of EPSP synthase from Mycobacterium tuberculosis.

The shikimate pathway is an attractive target for herbicides and antimicrobial agent development because it is essential in algae, higher plants, bacteria, and fungi, but absent from mammals. Homologues to enzymes in the shikimate pathway have been identified in the genome sequence of Mycobacterium tuberculosis. Among them, the EPSP synthase was proposed to be present by sequence homology. Accordingly, in order to pave the way for structural and functional efforts towards anti-mycobacterial agent development, here we describe the molecular modeling of 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase isolated from M. tuberculosis that should provide a structural framework on which the design of specific inhibitors may be based on. Significant differences in the relative orientation of the domains in the two models result in "open" and "closed" conformations. The possible relevance of this structural transition in the ligand biding is discussed.

Structural basis for inhibition of cyclin-dependent kinase 9 by flavopiridol.

Flavopiridol has been shown to potently inhibit CDK1 and 2 (cyclin-dependent kinases 1 and 2) and most recently it has been found that it also inhibits CDK9. The complex CDK9-cyclin T1 controls the elongation phase of transcription by RNA polymerase II. The present work describes a molecular model for the binary complex CDK9-flavopiridol. This structural model indicates that the inhibitor strongly binds to the ATP-binding pocket of CDK9 and the structural comparison of the complex CDK2-flavopiridol correlates the structural differences with differences in inhibition of these CDKs by flavopiridol. This structure opens the possibility of testing new inhibitor families, in addition to new substituents for the already known leading structures such as flavones and adenine derivatives.