Submitted by JÚLIO HEBER SILVA (julioheber@yahoo.com.br) on 2016-11-10T10:37:56Z No. of bitstreams: 2 Dissertação - Uessiley Ribeiro Barbosa - 2016.pdf: 3901383 bytes, checksum: f52c92a81d289db26f933d75a30e8788 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Jaqueline Silva (jtas29@gmail.com) on 2016-11-10T17:48:40Z (GMT) No. of bitstreams: 2 Dissertação - Uessiley Ribeiro Barbosa - 2016.pdf: 3901383 bytes, checksum: f52c92a81d289db26f933d75a30e8788 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2016-11-10T17:48:40Z (GMT). No. of bitstreams: 2 Dissertação - Uessiley Ribeiro Barbosa - 2016.pdf: 3901383 bytes, checksum: f52c92a81d289db26f933d75a30e8788 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2016-10-04 / The Paracoccidioides brasiliensis (Pb) is a thermo-dimorphic fungus described as the etiological agent of paracoccidioidomycosis (PCM), an important systemic mycosis in Latin America. The isocitrate lyase (ICL) is an enzyme involved in glyoxylate cycle, an alternative pathway to Krebs cycle, which has been described in fungi, bacteria and plants. The absence of this enzyme in mammals makes it an interesting target for design of specific antifungal compounds for PCM. In this work, we use in silico methods like homology modeling, molecular dynamics and virtual screening, aiming the development of inhibitors compounds for PbICL enzyme in the absence and presence of cofactor and positive control. From a molecular docking protocol, it was possible to select promising compounds by criteria of affinity and efficiency based on screening of natural products. Two regions were selected for molecular docking, one involving a region already known by the binding of argentilactona inhibitor and another region involving the cofactor Mg 2+, a possible catalytic site of ICL. All compounds selected by affinity criteria had more than 80% of success rate in achieving lower energy and allowed to describe common residues within the protein interaction for the target sites. The structural quality parameters are significantly improved after 100 ns of simulation. The structure of PbICL and PbICL with magnesium show all quality parameters as acceptable to define them as highresolution structures. All selected compounds show aromatic chains with the possibility of being a pharmacophore, which is essential for biological activity. Another interesting aspect is that, through the selected compounds, it was possible to describe structural patterns related to the ligand specificity, that might be promising for a basic chemical sketch for rational drug design. / O Paracoccidioides brasiliensis (Pb) é um fungo termo-dimórfico descrito como agente etiológico da paracoccidioidomicose (PCM), uma micose sistêmica importante na América Latina. A Isocitrato liase (ICL) é uma enzima envolvida no ciclo do glioxilato, uma via do ciclo Krebs já descritos em fungos, bactérias e plantas. Sua ausência em mamíferos faz dessa enzima, um alvo interessante para o desenho de compostos antifúngicos específicos para a PCM. Neste trabalho, utilizamos métodos in silico como modelagem por homologia, dinâmica molecular e triagem virtual na busca de compostos inibidores para enzima isocitrato liase de P brasilienses (PbICL) na ausência e presença do cofator e com a seleção de controle positivo. A partir de um protocolo de ancoragem molecular, foi possível selecionar compostos promissores por critérios de afinidade e eficiência a partir de triagens virtuais com produtos naturais. Duas regiões foram selecionadas para a ancoragem molecular, uma envolvendo uma região já conhecida pela ligação do inibidor argentilactona e outra envolvendo a região de ligação do cofator Mg2+ , posicionada no possível sítio catalítico da enzima. Todos os compostos selecionados pelo critério de afinidade tiveram mais do que 80% de sucesso em alcançar a mais baixa energia e permitiram descrever resíduos frequentes na interação proteína-inibidor para sítios alvo da enzima. Os parâmetros de qualidade da estrutura são sensivelmente melhorados após 100 ns de simulação. As estruturas de PbICL e PbICL com Mg2+ apresentam todos os parâmetros dentro do aceitável para defini-la como estrutura de alta resolução. Todos os compostos selecionados apresentaram cadeias aromáticas com possibilidade de ser um grupo farmacofórico, o qual é essencial para a atividade biológica. Outro aspecto interessante é que através dos compostos selecionados foi possível descrever padrões estruturais relacionados à especificidade do composto, os quais podem vir a ser promissores para um esqueleto básico no desenho racional de fármacos.

Submitted by Erika Demachki (erikademachki@gmail.com) on 2017-01-02T16:51:48Z No. of bitstreams: 2 Tese - Bruno Junior Neves - 2016.pdf: 19331549 bytes, checksum: a7cd8b65a0a68a6cc99718604d4a6aa4 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2017-01-03T09:47:27Z (GMT) No. of bitstreams: 2 Tese - Bruno Junior Neves - 2016.pdf: 19331549 bytes, checksum: a7cd8b65a0a68a6cc99718604d4a6aa4 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-01-03T09:47:27Z (GMT). No. of bitstreams: 2 Tese - Bruno Junior Neves - 2016.pdf: 19331549 bytes, checksum: a7cd8b65a0a68a6cc99718604d4a6aa4 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2016-11-01 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / Schistosomiasis is a serious endemic disease caused by trematodes of the genus Schistosoma. Currently the control of this disease is based solely on the administration of praziquantel. However, its extensive use has raised concerns about the emergence of resistant worms and the need of discovering new anti-schistosomal drugs. Given the above, the main goal of this study was to design new S. mansoni thioredoxin glutathione reductase (SmTGR) inhibitors showing antischistosomal activity through cheminformatics tools and to repurpose new drugs for schistosomiasis employing bioinformatics tools. At the stage of drug design, binary QSAR models were constructed and validated to predict inhibitory activity of SmTGR, a validated target in schistosomes. From the individual models, consensus and consensus rigor models and were generated (CCRs = 0.87 and 0.91, respectively) and used for the virtual screening of 150,000 compounds. At the end of this process, 29 compounds were prioritized and purchased for biological evaluation. As a result, two new hits representing new molecular scaffolds showed EC 50 ≤ 3.5 µM to schistosomula and ≤ 6.0 µM for adult female worms, low cytotoxicity against WSS-1 mammalian cells (IC50 > 16 µM) and low reactivity with cysteine proteases (IC50 > 100 µM). In the second part of this work, that is drug repositioning, 2,114 proteins of S. mansoni were used in a search for orthologs in therapeutic drug target databases (DrugBank, TTD and STITCH). As a result, 215 drugs were predicted to interact with 49 schistosome proteins, of which 47 had already anti-schistosomal activity reported in the literature. Then, principal component analysis (PCA) and k-means showed that 115 drugs were in the chemical space of known anti-schistosomal agents, increasing the overall confidence of predictions. Among them, paroxetine (PAR), an antidepressant drug predicted to inhibit serotonin transporter proteins of S. mansoni (SmSERTs) presented antischistosomal activity against schistosomula (EC50 = 2.5 µM) and adult worms (EC50 = 5.1 µM and 9.9 µM for males and females respectively) of S. mansoni. Lastly, molecular docking studies with SmSERT-A and its ortholog in humans (hSERT) explored the molecular basis for antischistosomal activity of PAR and provided information for the design of more potent and selective analogs. / A esquistossomose é uma doença endêmica grave causada por trematódeos do gênero Schistosoma. Atualmente o controle desta doença é baseado exclusivamente no uso do fármaco praziquantel. Todavia, seu uso extensivo tem levantado a preocupação quanto ao surgimento de vermes resistentes e a necessidade de descobrir novos fármacos esquistossomicidas. Face ao exposto, o objetivo deste trabalho foi planejar novos inibidores da enzima tiorredoxina glutationa redutase de S. mansoni (SmTGR) com atividade esquistossomicida utilizando ferramentas de quimioinformática e reposicionar novos fármacos para esquistossomose utilizando ferramentas de bioinformática. No estudo de planejamento de fármacos, modelos de QSAR binários foram construídos e validados para predição da atividade inibitória da SmTGR, um alvo validado em esquistossomos. A partir dos modelos individuais, modelos de consenso e consenso rigoroso foram construídos (CCRs = 0,87 e 0,91, respectivamente) e utilizados na triagem virtual de 150 mil compostos. Ao final deste processo, 29 compostos foram priorizados e adquiridos para avaliação biológica. Como resultado, dois novos hits representando novos scaffolds moleculares apresentaram EC50 ≤ 3,5 µM para esquistossômulos e ≤ 6,0 µM para fêmeas de vermes adultos, baixa citotoxicidade em células WSS-1 de mamíferos (IC50 > 16 µM) e baixa reatividade com cisteíno proteases (IC50 > 100 µM). No estudo de reposicionamento de fármacos, 2.114 proteínas de S. mansoni foram utilizadas em uma busca por ortólogos em bases de dados de alvos terapêuticos de fármacos (DrugBank, TTD e STITCH). Como resultado, 215 fármacos foram preditos para interagir com 49 proteínas do esquistossomo, dos quais 47 já tinham atividade esquistossomicida reportada na literatura. Em seguida, a análise de componentes principais (PCA) e k-means demonstrou que 115 fármacos estavam dentro do espaço químico de agentes esquistossomicidas conhecidos, atribuindo maior confiabilidade as predições. Dentre eles, a paroxetina (PAR), um fármaco antidepressivo predito para inibir proteínas transportadoras de serotonina do S. mansoni (SmSERTs), apresentou atividade esquistossomicida em esquistossômulos (EC50 = 2,5 µM) e vermes adultos (EC50 = 5,1 μM e 9,9 μM para machos e fêmeas) de S. mansoni. No final, estudos de docagem molecular com a SmSERT-A e seu ortólogo em humanos (hSERT) exploraram as bases moleculares para a atividade esquistossomicida da PAR e forneceram informações para o planejamento de novos análogos mais potentes e seletivos.

A cana-de-açúcar é uma das principais fontes de energia renovável, constituindo a matéria-prima mais importante na busca por energia limpa e sustentável. Os benefícios ambientais provenientes da produção e do uso dos derivados de cana-de-açúcar fomentam o desenvolvimento de métodos e produtos que aumentem, de modo sustentável, a geração de bioenergia. Dentre os diversos fatores limitantes para o aumento da produção de cana-de-açúcar, destaca-se a ocorrência e a severidade de fitopatologias como a escaldadura das folhas. Essa doença, causada pela bactéria Xanthomonas albilineans, causa diminuição da produtividade, necessidade de reforma precoce dos canaviais e queda de qualidade do caldo extraído que determinam prejuízos econômicos significativos para os agricultores. Atualmente, não há alternativas disponíveis para o controle químico ou biológico dessa fitopatologia. Portanto, existe uma necessidade urgente de desenvolvimento de novas moléculas como defensivos agrícolas que sejam eficazes, seletivas, de baixo custo e impacto ao meio ambiente. A X. albilineans produz uma família de antibióticos e fitotoxinas conhecida como albicidinas. As enzimas envolvidas na biossíntese de albicidinas são alvos moleculares extremamente atrativos para o planejamento de novos agroquímicos. Entre as enzimas dessa via, destaca-se a fosfopanteteinil transferase (XaPPT, E.C. 2.7.8.7), uma enzima essencial para o desenvolvimento da X. albilineans. Essa dissertação está dividida em duas partes: i. estudos computacionais e ii. estudos experimentais. Na parte computacional foi desenvolvida uma nova ferramenta, denominada ViTaMIn para auxiliar o desenvolvimento de modelos tridimensionais (3D) de proteínas. A enzima XaPPT foi utilizada como estudo de caso para validar o programa. Os resultados obtidos indicaram que ViTaMIn foi capaz de auxiliar a construção de um modelo 3D robusto da XaPPT. Além disso, foi verificado que ViTaMIn é uma alternativa útil para usuários iniciante e experientes na modelagem molecular. O modelo de XaPPT construído foi utilizado na triagem virtual para a identificação de novos candidatos a inibidores. A estratégia incluiu a aplicação de filtros moleculares e estudos de docagem molecular que resultaram na seleção de 10 candidatos a inibidores da enzima-alvo. Na parte experimental, o cultivo de culturas de X. albilineans foi padronizado e estudos de biologia molecular com a XaPPT conduzidos a partir da extração de DNA genômico da bactéria. Estratégias clássicas e modernas foram empregadas para a clonagem da XaPPT. Os resultados obtidos indicaram a obtenção de proteína expressa na forma solúvel. Os trabalhos integrando estudos computacionais e experimentais apresentados nessa dissertação de mestrado significam importantes contribuições no desenvolvimento de bases científicas sólidas para o desenvolvimento de novos candidatos a agroquímicos para a cultura de cana de açúcar. / Sugarcane is a major source of renewable energy, representing the most important source for clean and sustainable energy. The environmental benefits collected from the production and use of sugarcane derivatives boost the development of new methods and products to improve, in a sustainable way, bioenergy generation. However, the occurrence and severity of plant diseases, such as leaf scald, hinder the productivity of sugarcane crops. Sugarcane leaf scald is a widespread and devastating disease caused by the bacteria Xanthomonas albilineans. The disease has a dramatic impact on crop productivity, including reduced yields and drop in quality of the juice. Currently, there is no chemical or biological treatment for disease control. Therefore, there is an urgent need for new effective and selective pesticides with low cost and environmental impact. The X. albilineans produces a family of antibiotics and phytotoxins known as albicidin. The enzymes involved in the biosynthesis of albicidin are attractive targets for the design of new agrochemicals. Among them, the phosphopantetheinyl transferase (XaPPT, EC 2.7.8.7) enzyme plays an essential role to the development of X. albilineans. This dissertation is divided into two parts: i. computational studies and ii. experimental studies. In the computational part, we developed a new tool, named ViTaMIn to assist the development of three-dimensional models (3D) of proteins. The XaPPT enzyme was employed as a case study to validate the program. The results indicated that ViTaMIn was capable of assisting the construction of a robust 3D model of XaPPT. Moreover, Vitamin is a useful alternative for beginners and advanced modelers. The XaPPT model was used in virtual screening approach to identify new candidate inhibitors. The strategy included the application of molecular filters and molecular docking studies which afforded 10 inhibitor candidates for the target enzyme. In the experimental part, X. albilineans was cultured and standardized. On the basis of that, molecular biology studies were conducted on XaPPT. Classical and modern strategies were employed to cloning XaPPT. The results indicated that protein is expressed soluble. The integration of computational and experimental studies presented in this dissertation are important contributions in the development of strong scientific basis for the design of new agrochemicals for sugarcane cultures.

As concepções de moduladores da dinâmica dos microtúbulos, que levam ao bloqueio do ciclo celular, e de bloqueadores de canais de cálcio tipo L (Cav), tais como o 1,4-di-hidropiridinas e análogos, que diminuem a resistência do organismo humano aos tratamentos quimioterápicos através da inibição da proteína de transmembrana P-gp, são estratégias importantes tanto para terapias antitumorais quanto para leishmanicida. Esta abordagem tem mostrado resultados interessantes na diminuição da resistência à quimioterapia em câncer chamada de MDR (do inglês Multi Drug Resistence), além de também serem uma estratégia importante para controlar a fase inicial da leishmaniose. Diante desse contexto, e baseado no estudo de Ueki 2013 e colaboradores que, a partir de estudos anteriores, os quais relatam a superexpressão das enzimas estona deacetilase (HDAC) e catepsina L (CTSL) em células tumorais, propuseram um pró-fármaco seletivo, planejado a partir de um espaçador de lisina acetilada, que garante a liberação do fármaco seletivamente nas células tumorais, trabalhamos no desenvolvimento de uma nova proposta de pró-fármaco trifuncional. Nossa proposta foi desenvolvida a partir de estudos de triagem virtual, baseados em ligantes e em estrutura, predição das propriedades farmacocinéticas e toxicológicas (ADME/Tox) e também técnicas de bioinformática para a construção de um modelo de canal de cálcio, devido à inexistência de estruturas, do mesmo, que estivessem depositadas no banco de dados de proteínas PDB (Protein Data Bank). Paralelamente, nosso grupo de síntese colaborador sintetizou, através de técnicas de \"Click Chemistry\" e reações de Mitsunobu multicomponentes, uma biblioteca de novos híbridos trifuncionais, os quais, após estudos de atividade biológica, foram avaliados (in silico) frente à estrutura da tubulina, e os compostos mais promissores desta biblioteca serviram de base para novos estudos de triagem virtual. Para a obtenção dos nossos hits, executamos 4 estratégias de triagem virtual, separadas em 2 tarefas. Ao final, selecionarmos um total 59 hits, dos quais, 9 hits apresentam promissoras atividades bloqueadoras do canal de cálcio e 65 hits apresentam promissoras atividades moduladoras da tubulina. Estes hits seguem em estágio de compra e ensaios in vitro e após comprovada a eficácia dos mesmos, estes futuramente farão parte de uma nova proposta de pró-farmaco trifuncional. / The concepts of modulating microtubule dynamics, and calcium channel L-types (CAV) blockers are important strategies for anticancer and antileishmanial therapies. Microtubule modulators that blocks the cell cycle and the calcium channel blockers, such as, 1,4-dihydropyridines and analogues, reduce the resistance of the human body to chemotherapeutic treatments by inhibiting transmembrane P-gp protein. This approach has shown interesting results in reduced resistance to chemotherapy in cancer called MDR (Multi Drug Resistance), and an important strategy for controlling the early stage of leishmaniasis. In this context, we work to develop a new proposal for trifunctional prodrug. We have based on the study of Ueki 2013 and collaborators, which, from earlier studies with reported overexpression of both deacetylase estona enzymes (HDACs) and cathepsin L (CTSL) in tumor cells, proposed a selective prodrug. We considering an acetylated lysine link/spacer, which ensures the release of the drug selectively in tumor cells, have now designed this selective prodrug. Our proposal was developed from virtual screening studies, based on ligands and structure, prediction of pharmacokinetic and toxicological properties (ADME / Tox) and also bioinformatics techniques for the construction of a calcium channel model, due to the inexistence of Structures of the same that were deposited in the database of proteins PDB (Protein Data Bank). At the same time, our collaborating synthesis group synthesized, through Click Chemistry techniques and Mitsunobu multicomponent reactions, a library of new trifunctional hybrids, which, after studies of biological activity, were evaluated (in silico) against the structure of tubulin , And the most promising compounds from this library served as the basis for further virtual screening studies. To obtain our hits, we performed 4 virtual screening strategies, separated into 2 tasks. In the end, we selected 59 hits, of which 9 hits show promising calcium channel blocking activities and 65 hits show promising tubulin modulating activities. These hits follow in vitro purchase and testing, and after proven effectiveness, they will be part of a new tri prodrug proposal.

Computational methods have been extensively developed at various levels of approximation in recent years to model biomolecular interactions and for rational drug design. This research work aims to explore the feasibility of using quantum mechanical (QM) methods within the two broad categories of in silico ligand-based and structure-based drug design. First, density functional theory at the M06L level of theory was employed to examine structure-activity relationships of boron-based heterocyclic compounds, anti-inflammatory inhibitors targetting the interleukin-1β (IL-1β) cytokine. Our findings from computed energies and shapes of the molecular orbitals provide understanding of electronic effects associated with the inhibitory activity. We also found that the boron atom, specifically its electrostatic polarity, appears to be essential for the anti-IL-1β activity as evidenced by the biological assay of non-boron analogues selected from the ligand-based virtual screening results. Secondly, we aimed to dock ligands at the active sites of zinc-containing metalloproteins with reasonable computational cost and with accuracy. Therefore, an in-house docking scheme based on a Monte Carlo sampling algorithm using the semiempirical PM6/AMBER force field scoring function was compiled for the first time within the Gaussian 09 program. It was applied to four test cases, docking to cytidine deaminase and human carbonic anhydrase II proteins. The docking results show the method’s promise in resolving false-positive docking poses and improving the predicted binding modes over a conventional docking scheme. Finally, semiempirical QM methods which include dispersion and hydrogen-bond corrections were assessed for modelling conformations of β-cyclodextrin (βCD) and their adsorption on graphene. The closed in vacuo βCD cccw conformer was found to be in the lowest energy, in good agreement with previous ab initio QM studies. DFTB3, PM6-DH2 and PM7 methods were applied to model the intermolecular interactions of large βCD/graphene complexes, over a thousand atoms in size. We found that the binding preference of βCD on graphene is in a closed conformation via its C2C3 rim, agreeing with reported experimental and computational findings.

<p>Drug-drug interactions (DDIs) occur when multiple drugs are used concurrently. Caused by one drug inhibiting or inducing the metabolism of a second drug, DDIs often alter plasma concentrations and could seriously impact efficacy and safety of co-administered medications. Cytochrome P450 (CYP), a superfamily of enzymes, plays an important role in metabolizing a majority of FDA approved drugs currently on the market. 70% of predicable DDIs are associated with CYP enzymes inhibition. In-silico methods are increasingly adopted as a cost-effective complement to guide and prioritize efforts in drug discovery. Recent emerging applications of artificial intelligence algorithms have demonstrated promising results capable of prioritizing the selection of large chemical libraries, thereby outlining the future of in-silico methods assisting in drug discovery. Nevertheless, current methods rely on molecular descriptors that almost exclusively focus on chemical properties and atomic structures that fail to capture critical conformation and biological interaction related properties. There is also a lack of trainable molecular descriptors with feature specificity that reflect detailed protein-ligand binding energy and enable biological activity prediction. The overall objective of this dissertation is to understand molecular biological binding activity through electronic structure-based local descriptors derived from quantum based conceptual density functional theory (CDFT). This method will be used to assess the correlation of intermolecular interaction energy with ligand-protein binding with 2D feature maps reduced from the 4D molecular surfaces of the binding site and ligand (3D molecular surface with 1D electronic property). Additionally, it will be used to explore the possibility of predicting CYP related DDIs using descriptors generated using first principles including protein-ligand binding with specificity and strength and deep learning algorithms. Using quantum chemistry to interpret topological molecular information residing on 3D molecular surface permits the extraction of interacting features directly from the ligand structure. To achieve that, a set of curatable data containing consistent measurements was accessed through publicly accessible libraries. A series of novel Manifold Embedding of Molecular Surface (MEMS) descriptors were generated containing local electronic properties residing on the 3D molecule structure surface of each ligand using manifold learning. Major information were captured featuring electronic characteristics on the molecular 3D surface. Shape context was employed to derive transnational invariance feature vectors from MEMS with high granularity, thus preserving molecular information with specificity. DeepSet was utilized to perform permutation equivariance model training and validation. Powerful model learning is observed with an F-measure for all targets above 75% with the highest of 87% from external testing. Despite their promising prediction performance, molecular conformation changes and analytical featurization methods need to be implemented to expand model applicability and improve model reliability.</p>

Le récepteur tyrosine kinase Axl a été découvert en 1988. Depuis, son implication dans les phénomènes de cancérisation a été mis en lumière. Ce récepteur est surexprimé, entre autres, dans les lignées cellulaires du cancer du pancréas et du cancer du sein triple négatif. Le succès des inhibiteurs de kinase contre les cancers (imatinib, erlotinib …) nous a poussés à nous focaliser sur la conception d'inhibiteurs du domaine kinase de la protéine Axl afin d'élaborer de nouveaux anticancéreux. Pour ce faire, nous avons décidé de modéliser le domaine kinase de la protéine Axl en conformations dites ‘active' et ‘inactive'. Les modèles ont ensuite été validés par différentes méthodes : des méthodes de bioinformatique structurale mais aussi par amarrage comparatif et par criblage virtuel focalisé. Sur la base de ces modèles, une chimiothèque virtuelle focalisée a été construite et amarrée dans les modèles d'Axl. J'ai ensuite effectué la synthèse chimique de 15 des ligands conçus à l'étape précédente et ciblant la conformation ‘inactive' du domaine kinase d'Axl. Aucun de ces ligands n'est apparu actif dans les tests in vitro. Dans un second temps, nous nous sommes intéressés à la chimie des noyaux 4- et 7-azaindoles. Ces travaux ont permis la synthèse de 12 ligands dirigés contre les conformations ‘inactives' des domaines kinases d'Abl et de Src dont certains ont montré une activité prometteuse. Parallèlement, un criblage a très large échelle a été publié et nous avons utilisé ces nouveaux résultats pour réévaluer l'existence d'une conformation -inactive' – de type « DFG-out » - du domaine kinase d'Axl. Ces travaux permettront la conception de nouveaux ligands ciblant efficacement Axl. / The receptor tyrosine kinase Axl was discovered in 1988. Latter on, its involvement in the cancer development was highlighted. Axl is overexpressed in pancreatic cancer and triple-negative breast cancer cell lines. The success of kinase inhibitors (imatinib, erlotinib ...) led us to focus on the design of inhibitors targetting the kinase domain of Axl. As a guide, we modeled the protein-kinase domain in its active and inactive conformations to perform structure-based drug design. The models were then validated by different methods: structural bioinformatics, comparative docking and focused virtual screening. A virtual chemical library was built and docked into Axl models.Then, I synthetized 15 chemical compounds targetting the ‘inactive' conformation of the kinase domain of Axl. However, none were active in an in vitro assay. Then we were interested in the chemistry of 4 and 7-azaindole cores. This work led to the synthesis of 12 ligands among which several showed promising activity against the ‘inactive' conformation of the kinase domains of Abl and Src.Meanwhile, a large-scale screening was published and we used that new data to re-evaluate the modeling of a "DFG-out" inactive conformation of Axl.

A cruza&iacute;na, principal ciste&iacute;no protease do Trypanosoma cruzi, &eacute; um alvo terap&ecirc;utico validado para o tratamento da doen&ccedil;a de Chagas. Grande parte dos inibidores desta enzima &eacute; constitu&iacute;da de pept&iacute;deo-mim&eacute;ticos do tipo covalente irrevers&iacute;vel cujo desenvolvimento, entretanto, tem sido evitado devido ao potencial efeito off target e ao perfil farmacocin&eacute;tico indesej&aacute;vel. O grupo de qu&iacute;mica medicinal NEQUIMED/ IQSC/USP v&ecirc;m utilizando m&eacute;todos avan&ccedil;ados em quimioinform&aacute;tica para a identifica&ccedil;&atilde;o de inibidores revers&iacute;veis da cruza&iacute;na e de outras ciste&iacute;no proteases. Contudo, para que estes inibidores se desenvolvam a compostos matrizes mais eficientes, informa&ccedil;&otilde;es computacionais e estruturais de seus Modos de Intera&ccedil;&atilde;o (MOB) com a prote&iacute;na alvo tornam-se fundamentais. Neste trabalho, a cristalografia de raios X foi utilizada para descrever em detalhes o MOB de tr&ecirc;s importantes inibidores revers&iacute;veis da cruza&iacute;na identificados pelo NEQUIMED: a dipeptidil-nitrila Neq0409 e os dois fragmentos moleculares Neq0147 e Neq0176. De modo a evitar a auto-prote&oacute;lise experimentada pela cruza&iacute;na nativa que torna desafiadora a cristaliza&ccedil;&atilde;o com inibidores de baixa afinidade como os fragmentos revers&iacute;veis, duas novas constru&ccedil;&otilde;es mutantes e inativas da cruza&iacute;na (a C25A e a C25S) foram desenvolvidas. A C25S foi validada como modelo cristalogr&aacute;fico representativo dado a coer&ecirc;ncia do MOB elucidado nesta enzima mutante e na cruza&iacute;na nativa para o inibidor mais estudado da cruza&iacute;na, o K777. A descri&ccedil;&atilde;o da presen&ccedil;a, orienta&ccedil;&atilde;o, conforma&ccedil;&atilde;o e modo de a&ccedil;&atilde;o no s&iacute;tio, al&eacute;m do completo padr&atilde;o de intera&ccedil;&otilde;es fornecidos por estas estruturas cristalinas, validaram ortogonalmente os MOB preditos e os m&eacute;todos de planejamento in silico usados no NEQUIMED, permitindo a identifica&ccedil;&atilde;o de inibidores muito mais potentes an&aacute;logos ao Neq0147 e ao Neq0409. O grupo 2-acetamidotiofeno-3-carboxamida do Neq0176 e o anel heteroc&iacute;clico de cinco membros baseado no 1,3,4-oxadiazol do Neq0147 foram identificados como alternativos aos tradicionais esqueletos pept&iacute;deo-mim&eacute;ticos. O impacto do efeito proteol&iacute;tico na qualidade e na resolu&ccedil;&atilde;o de estruturas cristalogr&aacute;ficas na cruza&iacute;na foi melhor compreendido a partir de duas estruturas de alta resolu&ccedil;&atilde;o obtidas para a cruza&iacute;na nativa em complexo com o metanotiossulfonato de metila (MMTS) e a iodoacetamida. Estes resultados permitiram compreender as bases experimentais para a cristaliza&ccedil;&atilde;o de inibidores revers&iacute;veis de baixa afinidade. Todos os resultados estruturais obtidos via cristalografia de raios X neste trabalho s&atilde;o &uacute;teis para mapear as bases estruturais essenciais para o planejamento de futuros inibidores mais potentes e seletivos da cruza&iacute;na. / Cruzain, the major Trypanosoma cruzi cysteine protease, is a validated therapeutic target for the search of new medicines for the treatment of Chagas disease. A myriad of inhibitors of this enzyme consists of covalent irreversible peptidomimetics whose development has been impaired due to potential off target effect and undesirable pharmacokinetic profile. Modern cheminformatic methods employed by The Medicinal Chemistry Group (NEQUIMED/IQSC/USP) were used to identify cruzain reversible inhibitors. Their optimization for more efficient compounds can be accomplished by the use of data and information gathered from computational and structural modes of interaction (MOB). In this doctoral thesis, the X-ray crystallography was used to describe in detail the MOB of three important new cruzain reversible inhibitors: the dipeptidil-nitrile Neq0409 and the two molecular fragmentos Neq0147 and Neq0176. In order to avoid cruzain self-proteolysis during crystallization with low affinity reversible inhibitors such as the identified fragments, two new mutant and inactive constructs of cruzain (the C25S and C25A) were designed to upholding the same properties of the wild type catalytic site. The C25S was validated as representative crystallographic model given the coherence of the MOB elucidated for the best-known and studied cruzain inhibitor, the K777. The description of the presence, orientation, conformation and mode of action at the site, besides the complete pattern of bimolecular interactions, provided by these crystalline structures, orthogonally validated the predicted in silico MOB and allowed the identification of other potent inhibitors analogous to the Neq0147 and the Neq0409. The 2-acetamidothiophene-3-carboxamide moiety (Neq0176) and heterocyclic five-membered ring based on the 1,3,4-oxadiazole (Neq0147) were thereby identified as attractive alternatives to traditional peptidomimetics. The impact of proteolytic effect on the quality and resolution of crystallographic structures in cruzain was best understood from two high-resolution structures obtained for the native cruzain in complex with methyl methanethiolsulfonate (MMTS) and iodoacetamide. These results allow the understanding of experimental basis for the crystallization with reversible inhibitors of low affinity such as fragments. All structural results obtained by X-ray crystallography together with the catalytic site depiction using GRID and SuperStar methods are useful for mapping the essential structural basis for the design of future more potent and selective cruzain inhibitors.

A Dengue e a Febre Aftosa são infecções virais que ocorrem tanto no Brasil como no mundo, apresentando enorme impacto socioeconômico. As proteases virais são reconhecidas como alvos de grande interesse para o planejamento de antivirais, devido ao seu papel fundamental no ciclo de vida de muitos vírus, incluindo-se os flavivirus (vírus da Dengue DENV) e os picornavirus (vírus da Febre Aftosa FA). Com o objetivo de buscar e identificar novos inibidores de proteases virais (da NS2B/NS3pro do DENV ou da Lbpro do vírus da FA) propusemos modelos de busca virtual, incluindo diferentes filtros de seleção (ex. farmacofórico, drug-like, similaridade e ancoramento) aplicados, sequencialmente, a bancos de dados de compostos comerciais (280x103 a 23x106 compostos). A seleção final dos compostos foi sempre feita por inspeção visual. Para NS2B/NS3pro do DENV, construíram-se quatro modelos de busca virtual (Modelo I-DENV a Modelo IV-DENV). O primeiro foi construído, baseando-se na estrutura cristalográfica ligada a um inibidor peptidomimético, e aplicado ao banco ZINC. Ao final, dez compostos foram comprados e submetidos a testes de inibição enzimática, frente à NS2B/NS3pro, para validação experimental deste modelo. Dois compostos mostraram alguma atividade inibitória (IC50 150 - 300 &#181;M). Visando-se melhorar estes resultados, a flexibilidade da NS2B/NS3pro foi incluída, usando simulações de dinâmica molecular (DM), e um novo banco de dados construído (ZINC-Curated). Através de uma análise extensiva do banco ZINC-Curated, usando ferramentas estatísticas/quimiométricas, confirmou-se que este foi enriquecido com compostos com características de fármacos. Outros três modelos de busca virtual foram construídos incluindo-se diferentes informações obtidas nas simulações de DM. O modelo II-DENV foi feito usando ancoramento e aplicado ao banco ZINC-Curated, selecionando dezesseis compostos. Nenhum deles apresentou atividade inibitória significativa frente à NS2B/NS3pro do DENV. Os modelos III-DENV e IV-DENV utilizaram modelos farmacofóricos, que tiveram seus desempenhos previamente avaliados usando dados de literatura, e foram aplicados aos bancos NCI e ZINC-Curated, respectivamente. O modelo III-DENV selecionou quinze compostos, tendo quatro deles apresentado atividade inibitória (IC50 30 - 100 &#181;M). O modelo IV-DENV selecionou dezoito compostos, sendo quatro ativos frente a esta protease (IC50 4 - 90 &#181;M), representando uma taxa de acerto de ~22 %. Ainda, uma série de treze análogos estruturais do composto mais ativo foi construída, sendo três deles também ativos. Portanto, as modificações incluídas na busca virtual permitiram melhorar, significativamente, os resultados obtidos. Para Lbpro do vírus da FA, construíram-se dois modelos de busca virtual (Modelos I-FA e II-FA). O primeiro foi construído usando sua estrutura cristalográfica, sem ligantes, e uma série in house de potenciais inibidores covalentes. Seis compostos foram selecionados e testados frente à Lbpro, tendo dois deles baixa atividade inibitória (IC50 300 - 600 &#181;M). A partir da disponibilidade da estrutura da Lbpro com ligante, o modelo IIFA foi construído e aplicado ao banco ZINC-Curated, selecionando quinze compostos. Estes foram adquiridos e testados frente à Lbpro, não apresentando atividade inibitória significativa. Assim, as modificações incluídas ainda não foram suficientes para selecionar inibidores mais potentes. No entanto, estes modelos/resultados contribuíram para o entendimento da(s) interação(ões) no sítio ativo da Lbpro. / Dengue and Food-and-mouth disease are viral infections that occur in Brazil and in the world, causing a huge socioeconomic impact. Viral proteases are recognized as targets for antiviral design, because they are crucial for the life cycle of many viruses, such as flavivirus (Dengue virus DENV) and picornavirus (Food-and-mouth disease virus FMDV). In order to discovery novel inhibitors of viral proteases (of NS2B/NS3pro of DENV or of Lbpro of FMDV) virtual screening models were proposed comprising a sequence of different filters (e.g. pharmacophore, drug-like, similarity and docking) applied to databases of commercial compounds (280x103 to 23x106 compounds). In all models, the final selection of compounds was always done by visual inspection. For DENV NS2B/NS3pro, four virtual screening models were proposed (Model I-DENV to Model IV-DENV). Model I-DENV was built, based on the crystal structure bound to a peptidemimetic inhibitor, and applied to ZINC database. Finally, ten compounds were purchased and submitted to enzymatic assays against this protease to the experimental validation of this model. Two compounds showed some inhibitory activity (IC50 150 - 300 &#181;M). In order to improve these results, NS2B/NS3pro flexibility was included, using molecular dynamics (MD) simulations, and a novel database was built (ZINC-Curated). Throughout an exhaustively analysis of ZINC-Curated, using statistical/chemometrics tools, we confirmed that this new database was enriched with drug like compounds. Other three virtual screening models were built including different information from MD simulations. Model II-DENV was built using docking and applied to the ZINC-Curated database, selecting sixteen compounds. None of them showed a significant inhibitory activity against DENV NS2B/NS3pro. Models III-DENV and IV-DENV were built using pharmacophore models, which have their performance previously evaluated using literature data, and applied to NCI and ZINC-Curated databases, respectively. Model III-DENV selected fifteen compounds, showing four of them inhibitory activity (IC50 30 - 100 &#181;M). Model IV-DENV selected eighteen compounds. Four of them were active against this protease (IC50 4 - 90 &#181;M), representing a hit rate of ~22 %. Moreover, a set of thirteen structural analogues of the most active compound were built, being three of them also active. Thus, the modifications done in the virtual screening procedure really improved our results. For FMDV Lbpro, two virtual screening models were built (Models I-FMDV and II-FMDV). The first one was based on the crystal structure, without ligands, and used a set of in house potential covalent inhibitors. Six of the in house compounds were selected and tested against this protease. Two of them showed a weak inhibitory activity (IC50 300 - 600 &#181;M). Later on, the Lbpro bound with ligands was available being therefore used to build another model. Model II FMDV was applied to the ZINC-Curated database, selecting fifteen compounds that were purchased and also tested against the target protease. But none of them showed a significant inhibition. Thus, the incorporated changes were not enough to retrieve active compounds. However, these models/results contributed to better understand Lbpro binding site interactions

A Doença de Alzheimer (DA) é um transtorno progressivo que acomete o Sistema Nervoso Central, causando demência em idosos. A doença leva a uma diminuição da memória, dificuldade no raciocínio e pensamento e alterações comportamentais. A fisiopatologia da doença corresponde ao aumento na concentração do peptídeo -amilóide com consequente deposição e formação da placa amiloide; e também ao aparecimento dos emaranhados neurofibrilares, que são agregados de proteína tau hiperfosforilada. A enzima glicogênio sintase cinase 3 beta (GSK-3) está diretamente envolvida nos dois processos e, por isso, a busca por novos inibidores dessa enzima é uma importante estratégia terapêutica para o tratamento da doença. Neste trabalho utilizou-se a triagem virtual em 7 bancos de dados de moléculas aplicando cinco diferentes estratégias in silico, através de planejamento de fármacos baseado em ligantes e baseado em estrutura, combinada com estudos in silico de farmacocinética, toxicidade e atividade biológica, seguido de posteriores ensaios de inibição enzimática in vitro. Obteve-se três compostos pela metodologia de farmacóforo, (Estratégia 3) dos quais dois deles demonstraram atividade inibitória interessante para GSK-3, na faixa de micromolar. A partir das outras quatro estratégias foram selecionados 16 compostos que futuramente serão também testados utilizando o mesmo protocolo de ensaio in vitro aqui utilizado. / Alzheimer\'s disease (AD) is a progressive disorder that affects the Central Nervous System, causing dementia in the elderly. The disease leads to decreased memory, difficulty in reasoning and thinking, and behavioral changes. The pathophysiology of the disease corresponds to the increase in -amyloid peptide concentration with consequent deposition and formation of the amyloid plaque and to the appearance of neurofibrillary tangles, which are aggregates of hyperphosphorylated tau protein. The enzyme glycogen synthase kinase-3 beta (GSK-3) is directly involved in both processes and, therefore, the search for new inhibitors of this enzyme is an important therapeutic strategy for the treatment of the disease. In this work, we used virtual screening in 7 molecule databases applying five different in silico strategies, using the ligand-based and structure-based drug design methodologies, combined with in silico studies of pharmacokinetics, toxicity and biological activity, followed by subsequent assays enzymatic inhibition in vitro. We obtained three compounds by the pharmacophore methodology (Strategy 3) of which two of them demonstrated interesting inhibitory activity for GSK-3 in the micromolar range. From the other four strategies, 16 compounds were selected which in future will also be tested using the same in vitro assay protocol used here.