Cette thèse a pour but de développer le pendant in silico des étapes clés du Fragment-Based Drug Design (FBDD), et ce dans le cadre plus général du développement de l'outil S4MPLE. Le FBDD génère des ligands drug-like à partir de petites molécules (fragments). Après une étape de validation de S4MPLE et de sa fonction d’énergie, un recentrage autour du FBDD est réalisé, à travers le docking puis l'optimisation virtuelle de fragments par growing ou linking (G/L). Cette stratégie reposesur 1) la création d’une chimiothèque focalisée en connectant un ou deux fragment(s) avec des linkers pré-générés, et 2) l’échantillonnage avec S4MPLE des composés chimères dans le site avec des contraintes. Des simulations de G/L plus ou moins ambitieuses (site flexible, ajout de H2O libres) permettent de valider cette approche avec des études rétrospectives basées sur des données expérimentales. La dernière phase de la thèse a consisté à appliquer ce protocole in silico à un projet de l’entreprise. / This work aims to develop in silico methods targeting the key stages of Fragment-Based Drug Design (FBDD), participating to the development of the molecular modeling tool S4MPLE. Briefly, FBDD generates ıdrug-likeı ligands from small organic molecules called fragments. After a validation step of S4MPLE and its energy function, the work focused on FBDD: molecular docking of fragments and their subsequent virtual optimization. The latter mimics standard evolution strategies in FBDD(growing and linking). This in silico approach involves among other two key stages 1) building of a focused library by plugging in pre-generated linkers into reference fragments using rules and 2) sampling of these new compounds under atomic and binding site constraints. Validation simulations, relying on known experimental data, included ıclassicalı growing / linking and more challenging ones (site flexibility, free waters). Finally, this strategy is applied to one project of the company.

À l’heure actuelle, le paludisme reste un problème de santé majeur et demeure une des maladies parasitaires les plus menaçantes. Parmi les cinq espèces de malaria infectant l’homme, Plasmodium falciparum est la forme la plus mortelle. Lors de la phase érythrocytaire de son cycle de vie, causant tous les symptômes du paludisme, P.falciparum utilise les phospholipides pour créer les membranes nécessaires au développement de cellules filles. Chez P. falciparum, la phosphatidylcholine est principalement obtenue grâce à la voie de synthèse de novo, dite voie de Kennedy. Dans cette voie de biosynthèse, la seconde étape catalysée par la CTP:phosphocholine cytidylyltransferase [EC 2.7.7.15] est limitante et apparait essentielle pour la survie du parasite murin P. berghei lors de la phase sanguine. Les objectifs de mon travail de thèse ont été de caractériser structuralement cette enzyme et d’identifier des effecteurs, principalement grâce à des approches de « fragment-based drug design » (FBDD). Ainsi, la première structure cristalline du domaine catalytique de l’enzyme (PfCCT) a été déterminée avec une résolution de 2.2 Å. De plus, les structures de trois complexes enzyme-substrat (en présence de CMP, de phosphocholine ou de choline) et d’un complexe enzyme-produit (CDP-Choline) ont été déterminées. Ces structures cristallographiques apportent des informations détaillées sur la poche de liaison de l’enzyme et elles ont révélé des informations sur le mécanisme de la réaction catalytique à l’échelle atomique. La seconde partie de ma thèse présente les méthodes développées pour identifier des inhibiteurs potentiels de la PfCCT. Une approche de FBDD a été utilisée pour identifier et sélectionner de petites molécules (fragments, PM<300 Da) se liant à la PfCCT. Diverses techniques biophysiques (fluorescence-based thermal shift assay, différence de transfert de saturation par RMN, dénaturation chimique isotherme) ont permis la sélection de 23 fragments à partir du criblage d’une bibliothèque (~ 300 molécules). En parallèle, un criblage in silico de plus grandes bibliothèques de fragments (environ 15 000 composés) a permis d’identifier 100 fragments “hits”. Enfin, 5 composés déjà connus pour inhiber la croissance parasitaire (Malaria Box fournit par Medecines for Malaria Venture) ont été sélectionnés pour leur inhibition de l’activité de la PfCCT recombinante. L’ensemble de ces données ouvre la voie pour l’élaboration de futurs composés ciblant la PfCCT et inhibant la biosynthèse de phosphatidylcholine chez P. falciparum. / Malaria remains a major global health problem and the most threatening parasitic disease. Among the 5 malaria species that affect humans, Plasmodium falciparum is the most deadly form. During its life cycle, in erythrocytic stage, which causes all the malaria symptoms, P. falciparum relies on phospholipids to build the membranes necessary for daughter cell development. Approximately 85% of parasite phospholipids consist of phosphatidylcholine (PC) and phosphatidylethanolamine (PE) synthesized by the parasite through the de novo Kennedy pathways. In the pathway of phosphatidylcholine biosynthesis, the second step catalyzed by CTP:phosphocholine cytidylyltransferase [EC 2.7.7.15] is rate limiting and appears essential for the parasite survival at its blood stage. In this PhD thesis I focus on the structural characterization of this enzyme and the identification of effectors mainly by fragment-based drug design approach (FBDD). The first reported crystal structure of the catalytic domain of the enzyme target (PfCCT) has been solved at resolution 2.2 Å. Four other crystal structures of PfCCT in complex with substrates (CMP, phosphocholine and choline) or product (CDP-choline) have been determined. These structural data give detailed images of the binding pocket and reveal the enzyme structures at all catalytic steps that provide crucial information on the catalytic mechanism at atomic level. The second part of the project present the methods developed to identify potential PfCCT inhibitors. A FBDD approach was used in order to identify and select small molecules (fragments, MW< 300 Da) binding to the PfCCT. A combination of biophysical techniques (fluorescence-based thermal shift assay, saturation transfer difference NMR and isothermal chemical denaturation) allowed the selection of 23 fragment hits from the screenings of fragment library (~ 300 molecules). In parallel in silico screening of larger fragment libraries (~15,000 compounds) resulted in 100 selected hits. Finally, 5 compounds already known to inhibit parasite growth (Malaria Box from Medicines for Malaria Venture) were selected for their inhibition of the recombinant PfCCT activity. The results obtained within this thesis brought important knowledge and structural insights on the catalytic mechanism of PfCCT. Taken together, these results pave the way for future structure-based drug design to target PfCCT and to inhibit the essential phosphatidylcholine biosynthesis in P. falciparum.

Arf1 est une petite protéine G (pG), essentiellement impliquée dans le trafic vésiculaire. Arf1 oscille entre deux conformations, l'une active liée au GTP et l'autre inactive associée au GDP. Arno est un des facteurs d'échange (GEF) capable d'activer Arf1 en stimulant l'échange GDP/GTP. Suractivée dans les cellules invasives du cancer du sein, Arf1 joue un rôle important dans la migration et la prolifération des cellules cancéreuses.Le but de ma thèse s'inscrit dans l'étude et la modulation de l'interaction pG-GEF, et plus spécifiquement, le couple Arf1-Arno. Mon travail a été planifié autour de deux axes: (1) L'étude fine de l'interaction entre Arf1 et Arno, et sa modulation avec un inhibiteur connu la Bréféldine A (BFA). (2) La mise en place d'une stratégie de conception d'inhibiteurs de l'interaction protéine-protéine du couple Arf1-Arno.Dans un premier temps, nous avons mis en place une méthode basée sur la résonance plasmonique de surface (SPR) permettant la détermination des paramètres cinétiques de l'interaction entre Arf1 et Arno. Nous avons précisé aussi les conséquences des partenaires allostériques (GDP, GTP, et Mg2+) et de la BFA sur les paramètres cinétiques de l'interaction. Ceci a permis une analyse fine de la régulation allostérique et du mode d'action de la BFA. Appliquée à d'autres inhibiteurs, cette méthode permettra d'examiner leur mécanisme d'inhibition.Dans la deuxième partie j'expose, la stratégie que nous avons utilisé pour la conception rationnelle d'inhibiteur de l'interaction entre Arf1 et Arno. Elle est basée sur le criblage virtuel de fragments au niveau des résidus clé « hotspots » de l'interaction, la validation des molécules-touches par des techniques biophysiques, et l'élimination de molécules artefacts. Les structures des complexes fragments-Arno ont été résolues, ce qui confirme la validité de cette stratégie ouvrant la voie vers l'optimisation moléculaire pour obtenir des inhibiteurs plus efficaces. / Arf1 is a small GTPases, essentially involved in the vesicular traffic. Arf1 switch between two conformations, an active form bound to GTP and an inactive form bound to GDP. Arno is one of the exchange factors (GEF) that can activate Arf1, through its catalytic Sec7 domain, promoting the exchange of GDP by GTP. Activated in breast cancer cells, Arf1 plays an important role in the migration and proliferation of cancer cells.The aim of my thesis was the study and the modulation of the interaction between small G proteins and their GEFs, more precisely the Arf1-Arno interaction. My work has been planned around two axes: (1) the study of the interaction between Arf1 and Arno, and its modulation with a known inhibitor Brefeldin A (BFA). (2) The development of a rational strategy for designing inhibitors of protein-protein interaction for the Arf1-Arno complex.In the first part of my PhD work, we set up a Surface Plasmon Resonance (SPR) method allowing to determine the kinetic parameters of the interaction between Arf1 and Arno. We also studied the effects of allosteric partners such as GDP, GTP and Mg2+ as well as the known uncompetitive inhibitor (Brefeldin A). This SPR approach allowed a very informative analysis at qualitative and quantitative levels of the various complexes taking place during the exchange reaction that should help to solve the inhibitory mechanism for the known inhibitors reported in the literature. In the second part of my thesis, we propose a strategy for targeting the interaction between Arf1and Arno. This approach is based on virtual screening of fragments at hotspot regions. Using biophysical techniques such fluorescence techniques, SPR, NMR and X-Ray crystallography, we identified and validated Hits, showing by crystallographic structural data their modes of interaction with the target protein Arno. A fluorescence polarization test was also developed to identify false positive fragments to eliminate promiscuous aggregators. Taken together, our work proposes a method based on SPR allowing the study of known inhibitors of GEFs, understanding at molecular level their mode of action. We also propose a general strategy for finding Hit fragments that designing competitive inhibitor of the interaction small G protein with its GEFs, that can be the scaffold for designing more powerful inhibitors.

A Doença de Chagas, endêmica na América Latina, é causada pelo parasito tripanossomatídeo Trypanosoma cruzi e atualmente já se espalha para o restante do mundo devido à migração humana. Os dois medicamentos disponíveis para o tratamento dessa doença, o Nifurtimox, (banido do Brasil), e o Benzonidazol, são eficazes somente na etapa aguda da doença e possuem efeitos colaterais severos. Recentemente, três novas substâncias para o tratamento chegaram à fase clínica de testes contra essa doença, mas ainda é necessária a pesquisa de novas moléculas contra esse parasito. A enzima Gliceraldeído-3-fosfato Desidrogenase (GAPDH) foi selecionada como alvo para busca de moléculas potencialmente tripanossomicidas. De forma a inibir essa enzima, a busca de moléculas baseou-se na abordagem de fragmentos moleculares para encontrar substâncias com elevada eficiência de ligante. A partir de um banco de dados comercial de 500 mil moléculas, filtros moleculares de solubilidade e da Regra dos Três foram aplicados para montar uma biblioteca focalizada de moléculas. Essa biblioteca foi submetida a estudos integrados baseados na estrutura do alvo macromolecular (docagem) e do substrato da enzima (similaridade química e eletrostática) e a inspeção visual dos fragmentos bem classificados em ambas técnicas foi realizada de modo a selecionar cinco compostos de classes químicas diversas. Essas substâncias foram submetidas a ensaios enzimáticos in vitro por meio de espectroscopia de fluorescência, sendo encontrado então um fragmento com Ki de (425 &plusmn; 53) &mu;M e eficiência de ligante de 0,33, valor bastante promissor para abordagem baseada em fragmentos moleculares. Estudos de simulação por Dinâmica Molecular (MD) foram feitos para as cinco moléculas adquiridas, com energia de interação ligante-enzima calculada usando o método MM-GB/SA. A classificação das moléculas por essa energia foi idêntica à obtida experimentalmente. Além disso, a MD possibilitou a predição de modo de interação dos fragmentos no sítio ativo da TcGAPDH e a identificação de uma nova cavidade passível de modulação de sua atividade. Uma segunda série de fragmentos foi selecionada baseada no fragmento ativo de modo a construir uma relação estrutura-atividade (SAR) teórica por MD. A SAR sugere que a presença de um átomo nitrogênio capaz de doar ligações de hidrogênio é importante para a interação, com o resíduo de aminoácido Asp210. Este resíduo de aminoácido desponta como um possível ponto de seletividade para a enzima humana, que possui uma leucina nessa posição. Além disso, a posição na substituição do anel central também está diretamente relacionada à interação da molécula com a enzima, com uma substituição 2,3 em um anel de cinco membros a mais favorável. Este trabalho identificou pela primeira vez um fragmento molecular com alta eficiência de ligante para a enzima TcGAPDH, com o auxílio do uso conjunto de técnicas baseadas na estrutura do ligante e do alvo, para seleção de moléculas, e espectroscopia de fluorescência para identificação de atividade inibitória frente à enzima. Método de simulação por MD conseguiu reproduzir os resultados experimentais e prover informações teóricas de SAR para o composto ativo. / Chagas disease is a parasitic illness endemic in Latin America caused by the trypanosomatid parasite Trypanosoma cruzi that spreads around the world due to people migration. Nowadays, Benznidazole and Nifurtimox (banned in Brazil), are used for the treatment of this disease but causes severe side effects to patients. Recently, three new molecules have reached clínical trials phase in the development of drugs against Chagas disease but it is still necessary to develop new drugs. In this studies, the enzyme Glyceraldehyde-3-phosphate Dehydrogenase (GAPDH) was used as a target for the search of new antitrypanosomatid molecules. It belongs to the glycolytic pathway, the major one for parasite\'s energy generation. With the aim searching a new molecule that inhibits this enzyme, the fragment-based approach guided the search of molecules with high ligand efficiency. A focused compound library was assembled from a database of 500,000 molecules using molecular and solubility filters and the Rule of Three. The integrated use of ligand (chemical and electrostatic similarity) and target (molecular docking) based drug design was carried out to rank the molecular fragments by a consensual score. Through visual inspection of the top 500 molecules five diverse fragments were selected for the in vitro enzymatic assays using fluorescence spectroscopy. One of these molecules shows a Ki equals to (425 &plusmn; 53) &mu;M and ligand efficiency equals to 0,33, a promising value for the fragment-based approach. Additionally, Molecular Dynamics simulations (MD) were carried out with these fragments and the predicted energy of interaction for fragment-enzyme complex was able to rank the molecules as using the experimental results. Furthermore, the MD was useful to predict the mode of interaction of the fragments in the active site of enzyme and to reveal a new cavity close to the substrate binding site. A second generation of compounds was selected based on the structure of the active fragment to construct theoretical structure-activity relationship (SAR) using MD. SAR indicates that the presence of a nitrogen with hydrogen-bond donor property is important to the interaction, making hydrogen-bonding with the amino acid residue Asp210. In addition, MD shows the influence of different substituent posítion in the central ring in the energy of the interaction, with a 2,3 substitution at five-atom ring the most favorable. This study identifies the first molecular fragment with high ligand efficiency for the enzyme TcGAPDH, with the combined use of ligand and target-based tools and fluorescence spectroscopy, for selection and identification of active compounds against the enzyme. MD was able to reproduce experimental results and generate theoretical information of SAR to the active molecular fragment.

A Doença de Chagas, endêmica na América Latina, é causada pelo parasito tripanossomatídeo Trypanosoma cruzi e atualmente já se espalha para o restante do mundo devido à migração humana. Os dois medicamentos disponíveis para o tratamento dessa doença, o Nifurtimox, (banido do Brasil), e o Benzonidazol, são eficazes somente na etapa aguda da doença e possuem efeitos colaterais severos. Recentemente, três novas substâncias para o tratamento chegaram à fase clínica de testes contra essa doença, mas ainda é necessária a pesquisa de novas moléculas contra esse parasito. A enzima Gliceraldeído-3-fosfato Desidrogenase (GAPDH) foi selecionada como alvo para busca de moléculas potencialmente tripanossomicidas. De forma a inibir essa enzima, a busca de moléculas baseou-se na abordagem de fragmentos moleculares para encontrar substâncias com elevada eficiência de ligante. A partir de um banco de dados comercial de 500 mil moléculas, filtros moleculares de solubilidade e da Regra dos Três foram aplicados para montar uma biblioteca focalizada de moléculas. Essa biblioteca foi submetida a estudos integrados baseados na estrutura do alvo macromolecular (docagem) e do substrato da enzima (similaridade química e eletrostática) e a inspeção visual dos fragmentos bem classificados em ambas técnicas foi realizada de modo a selecionar cinco compostos de classes químicas diversas. Essas substâncias foram submetidas a ensaios enzimáticos in vitro por meio de espectroscopia de fluorescência, sendo encontrado então um fragmento com Ki de (425 &plusmn; 53) &mu;M e eficiência de ligante de 0,33, valor bastante promissor para abordagem baseada em fragmentos moleculares. Estudos de simulação por Dinâmica Molecular (MD) foram feitos para as cinco moléculas adquiridas, com energia de interação ligante-enzima calculada usando o método MM-GB/SA. A classificação das moléculas por essa energia foi idêntica à obtida experimentalmente. Além disso, a MD possibilitou a predição de modo de interação dos fragmentos no sítio ativo da TcGAPDH e a identificação de uma nova cavidade passível de modulação de sua atividade. Uma segunda série de fragmentos foi selecionada baseada no fragmento ativo de modo a construir uma relação estrutura-atividade (SAR) teórica por MD. A SAR sugere que a presença de um átomo nitrogênio capaz de doar ligações de hidrogênio é importante para a interação, com o resíduo de aminoácido Asp210. Este resíduo de aminoácido desponta como um possível ponto de seletividade para a enzima humana, que possui uma leucina nessa posição. Além disso, a posição na substituição do anel central também está diretamente relacionada à interação da molécula com a enzima, com uma substituição 2,3 em um anel de cinco membros a mais favorável. Este trabalho identificou pela primeira vez um fragmento molecular com alta eficiência de ligante para a enzima TcGAPDH, com o auxílio do uso conjunto de técnicas baseadas na estrutura do ligante e do alvo, para seleção de moléculas, e espectroscopia de fluorescência para identificação de atividade inibitória frente à enzima. Método de simulação por MD conseguiu reproduzir os resultados experimentais e prover informações teóricas de SAR para o composto ativo. / Chagas disease is a parasitic illness endemic in Latin America caused by the trypanosomatid parasite Trypanosoma cruzi that spreads around the world due to people migration. Nowadays, Benznidazole and Nifurtimox (banned in Brazil), are used for the treatment of this disease but causes severe side effects to patients. Recently, three new molecules have reached clínical trials phase in the development of drugs against Chagas disease but it is still necessary to develop new drugs. In this studies, the enzyme Glyceraldehyde-3-phosphate Dehydrogenase (GAPDH) was used as a target for the search of new antitrypanosomatid molecules. It belongs to the glycolytic pathway, the major one for parasite\'s energy generation. With the aim searching a new molecule that inhibits this enzyme, the fragment-based approach guided the search of molecules with high ligand efficiency. A focused compound library was assembled from a database of 500,000 molecules using molecular and solubility filters and the Rule of Three. The integrated use of ligand (chemical and electrostatic similarity) and target (molecular docking) based drug design was carried out to rank the molecular fragments by a consensual score. Through visual inspection of the top 500 molecules five diverse fragments were selected for the in vitro enzymatic assays using fluorescence spectroscopy. One of these molecules shows a Ki equals to (425 &plusmn; 53) &mu;M and ligand efficiency equals to 0,33, a promising value for the fragment-based approach. Additionally, Molecular Dynamics simulations (MD) were carried out with these fragments and the predicted energy of interaction for fragment-enzyme complex was able to rank the molecules as using the experimental results. Furthermore, the MD was useful to predict the mode of interaction of the fragments in the active site of enzyme and to reveal a new cavity close to the substrate binding site. A second generation of compounds was selected based on the structure of the active fragment to construct theoretical structure-activity relationship (SAR) using MD. SAR indicates that the presence of a nitrogen with hydrogen-bond donor property is important to the interaction, making hydrogen-bonding with the amino acid residue Asp210. In addition, MD shows the influence of different substituent posítion in the central ring in the energy of the interaction, with a 2,3 substitution at five-atom ring the most favorable. This study identifies the first molecular fragment with high ligand efficiency for the enzyme TcGAPDH, with the combined use of ligand and target-based tools and fluorescence spectroscopy, for selection and identification of active compounds against the enzyme. MD was able to reproduce experimental results and generate theoretical information of SAR to the active molecular fragment.

Tuberculosis (TB) remains the leading cause of death due to a single infectious agent with more than 1.5 million people killed each year. In 2018, the World Health Organization (WHO) estimated that one third of the world’s population was infected with Mycobacterium tuberculosis (Mtb), the pathogen responsible for the disease.In 2000, EthR, a mycobacterial transcriptional repressor, was identified as a key modulator of ethionamide (ETH) bioactivation. ETH is one of the main second-line drugs used to treat drug-resistant strains and it is a prodrug that is activated in Mtb by the mono-oxygenase EthA and then inhibits InhA, an enzyme involved in the mycolic acid biosynthesis. In 2009, it was demonstrated that co-administration of ETH with the drug-like inhibitors of EthR was able to boost ETH activity by a factor three in a mouse-model of TB-infection, thus validating EthR protein as a target for a new therapeutic strategy. The first part of this thesis deals with the validation and deep characterization of the solved EthR-ligand structures based on all analysis of how each ligand bind to the EthR. In this section, based on the study of both co-crystal structures and the physicochemical properties of the ligands, we have rationalized the information currently available and understood the interaction of all EthR inhibitors in order to lead to more effective inhibitor design.More recently, another mycobaterial repressor, denoted EthR2, was identified as a putative target that appears to be functionally comparable to EthR (then the locus has been termed EthA2/EthR2, due to its similarity to the EthA/EthR locus). Furthermore, a spiroisoxazoline family of small-molecules, generically denoted as SMARt, has been identified as effective ligand of EthR2. However, according to the data present in the literature, this spiroisoxazoline family can also bind to the former EthR. In order to investigate this proposition, I have solved these small molecules in complex with EthR and compared their binding interactions to the EthR2 protein as well. The opportunity for the design small-molecules is capable of targeting both repressors, thereby opening the way to a dual-target approach.Finally, the third part of this thesis is devoted to the mycobacterial transcriptional factor MabR (Rv2242). Several studies identified this protein as a regulatory transcription factor of the fatty acid synthase II operon, which is mainly responsible for the mycolic acid biosynthesis in Mtb. I therefore purified to homogeneity and characterized the MabR protein as well as I determined the crystal structure of its C-terminal part. Finally, the functional role of MabR is largely discussed, and the way on how to interfere with its DNA binding ability is commented with respect to our results. / Doctorat en Sciences biomédicales et pharmaceutiques (Pharmacie) / info:eu-repo/semantics/nonPublished

No description available.

The task of medicinal chemists in a drug discoveryproject is to synthesize/design analogues to the screening hits, simultaneouslyincreasing target potency and optimizing the pharmacological properties.  This requires a wide selection of moleculesto be synthesized, where both synthetic feasibility and price of startingmaterials are of great importance. In this work, a synthetic pathway from cheapand readily available starting materials to highly modifiable 2,4-disubstitutedpyrrolidines is demonstrated. Previously reported procedures to similarpyrrolidines use expensive catalysts, requires harsh conditions and requiresnon-commercially available starting materials. The suggested pathway herein has demonstrated great possibility forvariation in the 4-position, including fluoro, difluoro, nitrile and alcoholfunctional groups. There are several areas in which the synthesis can beimproved and expanded upon. Improvements can be made by optimizing thedescribed reaction conditions and further expansion of possible modificationsin both 2- and 4-position could be explored.

Locating a ligand-binding site is an important first step in structure-guided drug discovery, but current methods typically assess the pocket as a whole, doing little to suggest which regions and interactions are the most important for binding. This thesis introduces Fragment Hotspot Maps, a grid-based method that samples atomic propensities derived from interactions in the Cambridge Structural Database (CSD) with simple molecular probes. These maps specifically highlight fragment-binding sites and their corresponding pharmacophores, offering more precision over other binding site prediction methods. The method is validated by scoring the positions of 21 fragment and lead pairs. Fragment atoms are found in the highest scoring parts of the map corresponding to their atom type, with a median percentage rank of 98%. This is reduced to 72% for lead atoms, showing that the method can differentiate between the hotspots, and the warm spots later used during fragment elaboration. For ligand-bound structures, they provide an intuitive visual guide within the binding site, directing medicinal chemists where to grow the molecule and alerting them to suboptimal interactions within the original hit. These calculations are easily accessible through a simple to use web application, which only requires an input PDB structure or code. High scoring specific interactions predicted by the Fragment Hotspot Maps can be used to guide existing computer aided drug discovery methods. The Hotspots Python API has been created to allow these work flows to be executed programmatically through a single Python script. Two of the functions use scores from the Fragment Hotspot Maps to guide virtual screening methods, docking and field-based ligand screening. Docking virtual screening performance is improved by using a constraint selected from the highest scoring polar interaction. The field-based ligand screener uses modified versions of the Fragment Hotspot Maps directly to predict and score the binding pose. This workflow gave comparable results to docking, and for one target, Glucocorticoid receptor (GCR), showed much better results, highlighting its potential as an orthogonal approach. Fragment Hotspot Maps can be used at multiple stages of the drug discovery process, and research into these applications is ongoing. Their utility in the following areas are currently being explored: to assess ligandability for both individual structures and across proteomes, to aid in library design, to assess pockets throughout a molecular dynamics trajectory, to prioritise crystallographic fragment hits and to guide hit-to-lead development.

La méthode de conception de médicaments à partir de molécules « fragments » (connue sous le nom de « Fragment-Based Drug Design ») a été proposée au milieu des années 90, et a depuis été reconnue comme une alternative tangible aux techniques plus classiques de recherche de médicaments telles que le criblage à haut débit par exemple. La méthode des fragments consiste à cribler un petit nombre (< 10000) de composés organiques de faible poids moléculaire (< 300 Da) afin de détecter ceux qui se lient à la cible (protéine ou acides nucléiques). Du fait de leur faible complexité, les fragments présentent une affinité faible pour la cible, et la détection s'effectue généralement grâce à une technique biophysique (en particulier, résonance magnétique nucléaire (RMN), cristallographie aux rayons X, résonance plasmonique de surface). Les fragments « hits » sont ensuite modifiés par addition de nouvelles fonctions chimiques, ou par liaison de deux fragments, afin d'élaborer, étape par étape, une molécule capable d'établir des interactions plus nombreuses avec la cible, et d'améliorer ainsi l'affinité. Comparée aux méthodes classiques de criblage haut débit, la méthode des fragments offre divers avantages, notamment une meilleure exploration de l'espace chimique, une meilleure efficacité de liaison des molécules « hits », et une plus grande facilité d'optimisation des hits en molécules plus affines. Dans le cadre de ce projet de thèse, plusieurs aspects de la méthode des fragments ont été abordés : dans une première partie, nous étudions un cas concret d'application de la méthode des fragments à la recherche d'un inhibiteur de la peroxiredoxine 5 humaine, en utilisant la RMN comme outil de criblage des fragments ainsi que comme outil d'étude des interactions protéine-fragment. La découverte d'un inhibiteur de cette enzyme représente une avancée importante, qui devrait permettre de mieux comprendre son fonctionnement. Les autres parties de ce projet de thèse abordent des aspects plus méthodologiques de la méthode des fragments : les fragments conservent-ils leur site de liaison, leur efficacité de liaison et leur mode d'interaction au cours de leur élaboration en inhibiteur ? Les fragments peuvent-ils être spécifiques d'une protéine ? D'un site de liaison particulier ? Ces questions, rarement traitées, sont pourtant essentielles à la compréhension du comportement des molécules fragments, et sont abordées d'une part en défragmentant plusieurs inhibiteurs de la protéine Bcl-xL et en étudiant par RMN le comportement de ces fragments vis-à-vis de la protéine en termes d'affinité et de site de liaison, d'autre part en réalisant le criblage par RMN d'une série de fragments sur cinq protéines différentes (peroxiredoxine 5 humaine, sérum albumine humaine et trois protéines homologues de la famille Bcl-2). De manière générale, ce projet de thèse vise à étudier des aspects peu abordés de la méthode des fragments et à proposer des pistes permettant de mieux comprendre le comportement des fragments vis-à-vis de leur cible, au cours du criblage initial comme lors de leur optimisation / Fragment-Based Drug Design (FBDD) has been proposed in 1996 and has since been recognized as a tangible alternative to the more classical drug discovery methods such as High-Throuput Screening. FBDD consists of screening a small number (< 10 000) of low-molecular weight (< 300 Da) compounds and detect those that bind to the target (protein or nucleic acids). Because of their low complexity, fragment molecules usually display low affinities for their target, hence detecting fragment-protein interactions is mostly achieved using a biophysical technique (mostly Nuclear Magnetic Resonance (NMR), X-ray crystallography or Surface Plasmon Resonance). “Hit” fragments are then modified by addition of chemical substituents, or linked together, so as to elaborate a more complex molecule, forming more interactions with the target and hence displaying an improved affinity. As compared to the more classical High Throughput Screening method, fragment screening provides several advantages, including a better exploration of chemical space, highly ligand-efficient hits and easier optimization of the hits into more affine molecules. In this PhD project, several aspects of FDBB have been addressed : first, FBDD approaches were applied to the research of an inhibitor of the human peroxiredoxin 5 protein, using NMR not only as a screening method but also for the characterization of the protein-fragment interactions. The discovery of an inhibitor against this enzyme would allow to better understand its function. Next, methodological aspects of the FBDD method were addressed : Do fragments conserve their binding site, binding efficiency and mode of interaction upon optimization? Can the fragments display specificity towards a given target? Towards a given binding site? These issues, rarely studied, are yet essential to the understanding of the behavior of fragment molecules, and will be addressed firstly by defragmentating several Bcl-xL inhibitors into fragments and studying their behavior towards the protein in terms of a_nity and binding mode, secondly by screening a set of fragments against five different proteins (human peroxiredoxin 5, human serum albumin and three homologous proteins of the Bcl-2 family of proteins). More generally, this PhD project aims at studying less characterized aspects of the fragment methodology and proposing answers to better understand the behavior of fragment molecules towards their targets, both in the initial screening step and then during their optimization