Protein-protein interactions are key to several biological processes that facilitate signal transduction and many other processes. These interactions are involved in pathways that are critical to many human diseases. Targeting specific protein-protein interactions is a challenging goal because protein-protein interactions are predominately through hydrophobic interactions. Antagonists of the protein-protein interactions need to be perfectly fit into the binding pockets to ensure the activity. The -helical domain of the proteins behaves as the recognition motifs for numerous protein-protein, and protein-nucleic acid interactions. Research has shown that pathways of many diseases contain protein-protein interactions involving -helical domains, e.g. neurological disorders, bacterial infections, HIV and cancer, etc. It is difficult yet very important to design small molecules to target the shallow binding areas of protein-protein interactions. So far the most successful one is Hamilton’s 1,4-terphenylene scaffold, which has been used to target the interactions between p53/MDM2, Bak/Bcl-xL etc. Inspired by this, we designed and synthesized three new scaffolds of non-pepditic -helical mimetics, mimicking the i, i + 4, i + 7 positions of an -helix. There are three basic principles that were leading our design. The side chains of our designed molecules should act as mimetics of the side chains of an -helix. Second, our molecules should possess improved water solubility. Third, the molecules should be easy to synthesize to generate a focused library. Some of our molecules, including the ones whose molecular weight are as low as 294, started to show some inhibition against p53/MDM2 interactions.

The research project, Optimization of an array of Peptidic Indicator Displacement Assays for the Discrimination of Cabernet Sauvignon Wines, describes the multiple step lab trials conducted to optimize an array of ensembles composed of synthesized peptides and PCV:Cu+2 complexes for the differentiation of seven Cabernet Sauvignon wines with different tannin levels. This report also includes the methods and analysis used. The analysis interpreted by principal component analysis. / text

No description available.

TRAIL (TNF-Related Apoptosis Inducing Ligand) et CD40 Ligand, membre de la superfamille (SF) du TNF (Tumor Necrosis Factor) apparaissent comme des cibles attractives pour la thérapie ciblée du cancer. Une caractéristique commune des membres de la SF-TNF est la structure homotrimérique du ligand qui oligomérise les récepteurs pour induire la signalisation cellulaire. Cette interaction multivalente entre ligands et récepteurs est déterminante dans l’induction de la réponse cellulaire. L’objectif de cette thèse est de i) développer des ligands multivalents de TRAIL-R2 et CD40 afin d’améliorer leur efficacité anti-tumorale puis de les associer dans des stratégies anti-tumorales innovantes et ii) de caractériser les interactions entre ces ligands et les récepteurs à la surface de la cellule pour comprendre l’impact de l’oligomerisation sur l’initiation de l’apoptose au niveau membranaire.Ainsi nous avons utilisé des peptides cycliques spécifiques du récepteur de TRAIL-R2 (TRAILmim/DR5) que nous avons multimérisés sur des plateformes chimiques innovantes. Nous nous sommes également intéressés à l’état d’oligomérisation des récepteurs avant leur contact avec le ligand dans le but de mieux comprendre les interactions multivalentes. Enfin, nous avons caractérisé à l’aide de la technique de résonance plasmonique de surface et d’un biocapteur basé sur la fluorescence les interactions entre ces ligands multivalents et le récepteur TRAIL-R2. Ces travaux nous ont donc montré la nécessité d’analyser les interactions entre ligands et récepteurs directement à la surface de la cellule pour améliorer le développement de ligands pro apoptotiques efficaces. / TRAIL (TNF- Related Apoptosis Inducing Ligand) and CD40Ligand, members of the superfamily (SF) of TNF (Tumor Necrosis Factor), appear as attractive targets for cancer therapy. A common feature of members of the TNF-SF is the homotrimeric structure that oligomerized their receptors to induce cell signaling. This multivalent interaction between ligands and receptors is crucial for apoptosis induction.The aims of this thesis is to i) develop multivalent ligands of TRAIL -R2 and CD40 to enhance their antitumor efficacy and to associate them in innovative strategies and ii) to characterize the interactions between ligands and receptors at cell surface to understand the impact of oligomerization on the initiation of apoptosis at membrane level.We used cyclic peptides specific of TRAIL-R2 (TRAILmim/DR5) that we multimerized on innovative chemical scaffolds. We also interested in the state of receptor oligomerization before their contact with the ligand in order to better understand the multivalent interactions. Finally, we characterized using the technique of surface plasmon resonance and biosensor based fluorescence interactions between multivalent ligands and TRAIL- R2.In this work we showed the need to analyze the interactions between ligands and receptors directly on the surface of the cell to improve the development of effective pro -apoptotic ligands.

L'utilisation des oligonucléotides antisens PMO ou PNA, pour corriger les erreurs d'épissage par blocage stérique, constitue une nouvelle stratégie prometteuse pour réguler l'expression génétique. Ces ON peuvent mener au traitement de maladies comme la β-thalassémie, la dystrophie musculaire de Duchenne (DMD) ou les cancers. Cependant leur développement clinique requiert un système de délivrance efficace. Les peptides cationiques (CPPs) sont caractérisés par leur capacité à s'internaliser dans les cellules eucaryotes. Cependant leur efficacité à promouvoir la délivrance cytoplasmique et nucléaire des ON est limitée par leur séquestration dans des vésicules d'endocytose, ce qui est à l'origine de la dégradation du matériel internalisé. Nous avons contribué à l'étude du trafic intracellulaire et de l'activité dans un essai de correction d'épissage de plusieurs familles de CPPs capables de délivrer efficacement des analogues d'ON à des doses non toxiques et en absence d'agents endosomolytiques. Nos études mécanistiques indiquent que ces constructions (covalentes ou non covalentes) CPP-ON sont endocytées par la voie clathrine, que la ségrégation dans les endosomes reste une limitation et qu'il existe une bonne corrélation entre leur activité biologique et leur capacité à déstabiliser les membranes endosomales. / The use of antisense oligonucleotides PMO or PNA to correct splicing errors by steric- block represents a new promising therapeutic strategy. These ONs lead to the treatment of diseases such as β-thalassemia, Duchenne muscular dystrophy (DMD) or cancers. However their functional success requires efficient delivery. Cationic cell penetrating peptides (CPPs) are characterized by their ability to be internalized in eukaryotic cells. However their efficiency in promoting cytoplasmic and nuclear delivery of ON has been hampered by endocytic sequestration and subsequent degradation of internalized material in endocytic vesicles, which is responsible for the degradation of internalized material. We have contributed to the study of intracellular trafficking and activity (using splicing correction assay) of several families of CPPs capable of delivering effective analogs ON at nontoxic doses and in the absence of agents endosomolytic. Our mechanistic studies indicate that these constructs (covalent or noncovalent) CPP-ON are internalized through clathrin, that segregation in endosomes remains a limitation and that there is good correlation between biological activity and their ability to destabilize endosomal membranes.

A tendência da indústria em produzir alimentos mais saudáveis, com menos gordura vem aumentando sensivelmente. O desafio do melhoramento nos dias atuais é melhorar a qualidade da carcaça dos frangos juntamente com a redução nos teores de gordura, que é uma exigência do mercado consumidor. A leptina é um hormônio polipeptídico secretado principalmente pelo tecido adiposo com um importante papel na regulação da ingestão de alimentos, metabolismo de energia e reprodução. O gene da leptina e seu receptor vem sendo objeto de intensa investigação representando excelentes genes candidatos para deposição de gordura na carcaça nos programas de seleção assistida por marcadores. A função desses genes tem sido intensamente estudada em mamíferos, porém, em aves poucos estudos foram realizados. O presente trabalho teve como objetivo investigar a ocorrência de polimorfismos no gene da leptina e de seu receptor em galinhas, e avaliar os efeitos desses polimorfismos sobre características de desempenho e carcaça. Duas linhagens de aves, uma de corte (TT) e uma de postura (CC) foram utilizadas para desenvolver a população referência F2 empregada no presente estudo. A polução referência foi desenvolvida no sistema de Melhoramento Genético de Aves da Embrapa Suínos e Aves, em Concórdia/SC. Foram desenhados primers para amplificar o gene da leptina e os íntrons 8 e 19 do do gene do receptor da leptina. Entretanto, após diversas tentativas não foi possível amplificar o DNA genômico e cDNA do gene da leptina. No íntron 8 do gene do receptor da leptina o SNP C352T foi genotipado por seqüenciamento em 247 animais da população F2. O alelo T foi associado ao maior rendimento de carcaça (P=0,0165), rendimento de peito (P=0,0137), gramas de proteína bruta (P=0,0112) e cinza na carcaça (P=0,0150), enquanto o alelo C foi associado somente ao rendimento de fígado (0,0102). No íntron 19 o SNP G915A foi genotipado por PCR-RFLP em 137 animais da população F2. O alelo A foi associado ao consumo de ração (P=0,0339)), o alelo G ao rendimento de pulmão (P=0,287) e os alelos A e G quando em homozigose foram associados com rendimento de coxas e sobrecoxas (0,0302). Por ocorrerem em região de íntron, provavelmente esses SNPs não estão envolvidos diretamente com as características associadas, mas podem estar ligados a outra mutação localizada na região regulatória do gene do receptor da leptina ou em outro gene próximo. No futuro outros polimorfismos deverão ser explorados, em regiões codificadoras desse gene, para serem utilizados como marcadores associados a características de desempenho e carcaça em aves. Em adição o gene do receptor da leptina foi localizado no mapa de ligação do cromossomo 8 da população de aves da Embrapa, o próximo passo será realizar a análise de QTL incluindo os SNPs estudados como marcadores. / The trend of industry in producing healthier nourishment, with less fat has raised sensibly. The challenge of animal breeding in the current days is to improve carcass quality of chickens with concomitant reduction in the fat content, witch is a consumer\'s demand. Leptin is a polypeptide hormone secreted mainly by adipocytes with an important role in the regulation of food ingestion, energy metabolism energy and reproduction. The leptin (LEP) and its receptor (LEPR) genes are being subject of intense investigation representing excellent candidate genes for fat deposition in the carcass in marker assisted selection programs. The function of those genes has been intensely studied in mammals, however, in birds few studies were accomplished. The present work aimed to investigate the occurrence of polymorphisms in the LEP and LEPR genes in Chickens, and also evaluate the effects of those polymorphisms on performance and carcass traits. Two chicken lines, a broiler (TT) and a layer line (CC) were used to develop an F2 reference population used in the present study. The reference population was developed in the Poultry Genetic Breeding System at Embrapa Suínos e Aves, in Concórdia/SC. Primers were designed for the PCR amplification of the LEP gene leptina and introns 8 and 19 of the LEPR gene. However, after several attempts it was not possible to amplify the genomic DNA or cDNA of the leptin gene. In the intron 8 of the LEPR gene, SNP C352T was genotyped by DNA sequencing of 247 animals of the F2 population. The T allele was associated with greater carcass yield (P=0,0165), breast yield (P=0,0137), grams of crude protein (P=0,0112) and ashes in the carcass (P=0,0150), while the C allele was associated only to liver yield (0,0102). The SNP G915A in the intron 19 was genotyped by PCR-RFLP of 137 animals from the F2 population, the allele A was associated to the feed intake (P=0,0339), the allele G to the lung yield (P=0,287), and the alleles A and G, when in homozigose, were associated with drumstick and thigh yield (0,0302). Because these SNP are located within an intron area, they are not likely directly involved with the associated characteristics, but they can be linked the other mutation located in the regulatory area of the LEPR gene. In the future other polymorphisms should be explored, in the coding regions of this gene, so they can be used as markers associated to performance and carcass characteristics in poultry. In addition, the LEPR gene was located in the ligation map of the chromosome 8 for the Embrapa population. The next step will be to accomplish the analysis of QTL including the SNPs studied as markers.

A presente tese de doutorado propõe investigar mapas peptídicos urinários via eletroforese capilar acoplada ao espectrômetro de massas e avaliar se existe correlação entre estes e a presença de refluxo vésico-ureteral (RVU), visando desenvolver um novo método não invasivo para diagnosticar RVU em crianças, e ainda averiguar a existência de possíveis biomarcadores para a doença. Vinte e quatro amostras de urina de crianças positivas para RVU e quinze saudáveis, anteriormente submetidas à uretrocistografia miccional, foram disponibilizadas ao nosso grupo. Estas foram filtradas para eliminar proteínas de alta massa molar e pré-concentradas em coluna de fase reversa C2. Os mapas peptídicos foram obtidos via CE-MS em um eletrólito composto por 0,8% de ácido metanóico e 20% de metanol; já o líquido auxiliar consistia em 0,8% de ácido metanóico e 60% de metanol. Diversos métodos de classificação de aglomerados foram experimentados com os mapas peptídicos. Todos indicaram que as variáveis mais importantes para este discernimento eram os picos mais intensos ordenados em blocos de tempo. Entre os métodos de classificação não supervisionada, PCA (Principal Component Analysis) foi o mais perceptível na tarefa de distribuir os conjuntos de amostras. Para este, a porcentagem de acertos entre as amostras positivas foi de 75% e, entre as amostras negativas foi de 86,7%. No entanto, a taxa de erro encontrada por este método foi de 20,5%, não cumprindo o objetivo proposto. Entre os métodos de classificação supervisionada, SVM (Support Vector Machines) foi o selecionado, exprimindo melhor habilidade em prever que os modelos lineares e boa capacidade de generalização. O método otimizado apontou para o uso da função de base radial, o que está de acordo com outros trabalhos na área. Outras três variáveis ajustadas foram o parâmetro capacidade, o responsável por controlar a amplitude da função gaussiana e o ε-insensitive loss function. A validação cruzada LOO (leave-one-out) realizada para a rede treinada pelo SVM exibiu 0,00454 para a raiz quadrada da média dos erros e coeficiente de correlação próximo de um, indicando que o ajuste dos dados foi bem realizado. Empregando um mínimo de 80% das amostras para treinamento e 10% para verificação, o modelo foi capaz de classificar corretamente as amostras restantes. Três grupos de amostras para testes foram separados e a rede foi capaz de classificá-los corretamente. Biomarcadores específicos para RVU foram pesquisados durante o trabalho. Peptídeos que apareceram em mais de 70% das amostras de urina positivas para RVU e em menos de 15% das amostras de urina saudáveis foram selecionados. Espectros MS2 foram obtidos para estes peptídeos, e a pesquisa em bancos de espectros indicou um fragmento da imunoglobulina G como um possível candidato. No entanto, a existência de um biomarcador específico para RVU é ainda incerta, carecendo de uma investigação mais aprofundada para se concretizar este objetivo. Nos mapas peptídicos existem disparidades entre os dois grupos de amostras, que foram correlacionadas nesta tese com a presença de refluxo vésico-ureteral, sendo o nosso espaço amostral classificado corretamente / The present doctorate thesis intends to investigate urinary peptides maps by capillary electrophoresis coupled with mass spectrometry (CE-MS) to evaluate whether there is a correlation between these maps and vesicoureteral reflux (VUR) pathology, with the purpose of developing a new non invasive method for RVU diagnosis in children, and to investigate possible biomarkers for the illness. Twenty four urine samples of positive children for RVU and fifteen healthy children samples, previously submitted to the miccional cystourethrography, were available. Samples were filtered to eliminate proteins of high molar mass and preconcentrated in reversed-phase C2 columns. Peptide enriched samples were submitted to analysis by CE-MS in an electrolyte consisted of 0.8% formic acid and 20% methanol; sheath liquid was composed of 0.8% formic acid and 60% methanol. Diverse cluster analysis techniques were attempted to classify the peptide maps. The most important variables for screening were the most intense peaks organized by blocks of time. Among the non supervised methods, PCA (Principal Component Analysis) performed best in the task of discriminating the sample sets. For PCA, 75% of positive samples and 86.7% of negative samples were correctly assigned. However, the error found for this method was 20.5%, not fulfilling the purpose. Among the supervised methods, SVM (Support Vector Machines) performed best, exhibiting better prediction ability than the linear models and good generalization. The optimized method used a radial basis function, which is in agreement with literature. Three other variables were adjusted: the capacity parameter, responsible for controlling the Gaussian function amplitude and the e- insensitive loss function. The LOO (leave-one-out) cross-validation for the training set showed 0.005011 for the root-mean-square error (RMS) and a coefficient of correlation close to one, indicating good fitting and consistency. With a minimum of 80% samples for training and 10% for verification, the model was capable to classify correctly the remaining samples. Three sample groups for tests had been separated and the net was capable to classify them correctly. Specific biomarkers for VUR were searched. Peptides that appeared in more than 70% of positive urine samples for VUR and less than 15% of negative samples were selected. MS2 spectra were acquired for these peptides and database search pointed to an IgG fragment as a possible candidate biomarker. However, the existence of a specific biomarker for VUR is not conclusive with the present data and a more thorough investigation must be pursued. Nevertheless, the peptidic maps inspected in this work carried enough information that allowed discrimination of two sample sets, one of them correctly associated with VUR

A presente tese de doutorado propõe investigar mapas peptídicos urinários via eletroforese capilar acoplada ao espectrômetro de massas e avaliar se existe correlação entre estes e a presença de refluxo vésico-ureteral (RVU), visando desenvolver um novo método não invasivo para diagnosticar RVU em crianças, e ainda averiguar a existência de possíveis biomarcadores para a doença. Vinte e quatro amostras de urina de crianças positivas para RVU e quinze saudáveis, anteriormente submetidas à uretrocistografia miccional, foram disponibilizadas ao nosso grupo. Estas foram filtradas para eliminar proteínas de alta massa molar e pré-concentradas em coluna de fase reversa C2. Os mapas peptídicos foram obtidos via CE-MS em um eletrólito composto por 0,8% de ácido metanóico e 20% de metanol; já o líquido auxiliar consistia em 0,8% de ácido metanóico e 60% de metanol. Diversos métodos de classificação de aglomerados foram experimentados com os mapas peptídicos. Todos indicaram que as variáveis mais importantes para este discernimento eram os picos mais intensos ordenados em blocos de tempo. Entre os métodos de classificação não supervisionada, PCA (Principal Component Analysis) foi o mais perceptível na tarefa de distribuir os conjuntos de amostras. Para este, a porcentagem de acertos entre as amostras positivas foi de 75% e, entre as amostras negativas foi de 86,7%. No entanto, a taxa de erro encontrada por este método foi de 20,5%, não cumprindo o objetivo proposto. Entre os métodos de classificação supervisionada, SVM (Support Vector Machines) foi o selecionado, exprimindo melhor habilidade em prever que os modelos lineares e boa capacidade de generalização. O método otimizado apontou para o uso da função de base radial, o que está de acordo com outros trabalhos na área. Outras três variáveis ajustadas foram o parâmetro capacidade, o responsável por controlar a amplitude da função gaussiana e o ε-insensitive loss function. A validação cruzada LOO (leave-one-out) realizada para a rede treinada pelo SVM exibiu 0,00454 para a raiz quadrada da média dos erros e coeficiente de correlação próximo de um, indicando que o ajuste dos dados foi bem realizado. Empregando um mínimo de 80% das amostras para treinamento e 10% para verificação, o modelo foi capaz de classificar corretamente as amostras restantes. Três grupos de amostras para testes foram separados e a rede foi capaz de classificá-los corretamente. Biomarcadores específicos para RVU foram pesquisados durante o trabalho. Peptídeos que apareceram em mais de 70% das amostras de urina positivas para RVU e em menos de 15% das amostras de urina saudáveis foram selecionados. Espectros MS2 foram obtidos para estes peptídeos, e a pesquisa em bancos de espectros indicou um fragmento da imunoglobulina G como um possível candidato. No entanto, a existência de um biomarcador específico para RVU é ainda incerta, carecendo de uma investigação mais aprofundada para se concretizar este objetivo. Nos mapas peptídicos existem disparidades entre os dois grupos de amostras, que foram correlacionadas nesta tese com a presença de refluxo vésico-ureteral, sendo o nosso espaço amostral classificado corretamente / The present doctorate thesis intends to investigate urinary peptides maps by capillary electrophoresis coupled with mass spectrometry (CE-MS) to evaluate whether there is a correlation between these maps and vesicoureteral reflux (VUR) pathology, with the purpose of developing a new non invasive method for RVU diagnosis in children, and to investigate possible biomarkers for the illness. Twenty four urine samples of positive children for RVU and fifteen healthy children samples, previously submitted to the miccional cystourethrography, were available. Samples were filtered to eliminate proteins of high molar mass and preconcentrated in reversed-phase C2 columns. Peptide enriched samples were submitted to analysis by CE-MS in an electrolyte consisted of 0.8% formic acid and 20% methanol; sheath liquid was composed of 0.8% formic acid and 60% methanol. Diverse cluster analysis techniques were attempted to classify the peptide maps. The most important variables for screening were the most intense peaks organized by blocks of time. Among the non supervised methods, PCA (Principal Component Analysis) performed best in the task of discriminating the sample sets. For PCA, 75% of positive samples and 86.7% of negative samples were correctly assigned. However, the error found for this method was 20.5%, not fulfilling the purpose. Among the supervised methods, SVM (Support Vector Machines) performed best, exhibiting better prediction ability than the linear models and good generalization. The optimized method used a radial basis function, which is in agreement with literature. Three other variables were adjusted: the capacity parameter, responsible for controlling the Gaussian function amplitude and the e- insensitive loss function. The LOO (leave-one-out) cross-validation for the training set showed 0.005011 for the root-mean-square error (RMS) and a coefficient of correlation close to one, indicating good fitting and consistency. With a minimum of 80% samples for training and 10% for verification, the model was capable to classify correctly the remaining samples. Three sample groups for tests had been separated and the net was capable to classify them correctly. Specific biomarkers for VUR were searched. Peptides that appeared in more than 70% of positive urine samples for VUR and less than 15% of negative samples were selected. MS2 spectra were acquired for these peptides and database search pointed to an IgG fragment as a possible candidate biomarker. However, the existence of a specific biomarker for VUR is not conclusive with the present data and a more thorough investigation must be pursued. Nevertheless, the peptidic maps inspected in this work carried enough information that allowed discrimination of two sample sets, one of them correctly associated with VUR

Les nanocristaux de cellulose (NCC), obtenus par hydrolyse acide des microfibrilles de cellulose native sous forme de suspensions colloïdales aqueuses, sont des nanoparticules biosourcées ayant des propriétés mécaniques et optiques particulièrement séduisantes pour la conception de nanomatériaux à haute performance. Pour éliminer certaines de leurs limitations comme la sensibilité au sel et l'absence de contrôle de leurs interactions, nous nous sommes attachés au cours de ce travail à modifier chimiquement ces NCC par greffage de chaînes de polymère thermosensible de la famille de polyétheramines Jeffamine® sur leur surface. La première méthode de greffage utilisée faisait appel à une oxydation TEMPO suivie d'un couplage peptidique. Après l'optimisation de la réaction de greffage, les nouveaux systèmes (NCC-g-Jeffamine) ont été caractérisés et de nouvelles propriétés ont été identifiées notamment leur insensibilité au sel, leur capacité de redispersion dans des solvants organiques, leur caractère tensioactif, leur stabilité thermique améliorée ainsi que leur thermoagrégation réversible. L'étude structurale par diffusion des neutrons aux petits angles a permis d'avoir des informations sur les épaisseurs des couches de polymère greffé ainsi que sur la conformation des chaînes. De nouvelles caractéristiques de l'auto-organisation en phase chirale-nématique de ces nouveaux systèmes ont été notées concernant leur diagramme de phase et leur pas cholestériques. La deuxième méthode de greffage explorée consistait en une oxydation aux ions métapériodate suivie d'une amination réductrice. Les résultats ont montré un détachement partiel des chaînes de cellulose suite à l'oxydation puis leur détachement complet après le greffage du polymère pour obtenir des copolymères cellulose-Jeffamine ayant éventuellement des propriétés intéressantes à étudier. / Cellulose nanocrystals (CNC), obtained by acid hydrolysis of native cellulose microfibrils as colloidal aquous suspensions, are bioresourced nanoparticles that have great mechanical and optical properties well adapted for the conception of new nanomaterials with high performance. In order to eliminate some of their limitations like sensitivity to salt and absence of interactions control, we studied in this work the chemical modification of these CNC by grafting thermoresponsive polymer chains (polyetheramines Jeffamine®) on the their surface. A first grafting strategy used was a TEMPO oxidation followed by peptidic coupling. After the optimization of the grafting reaction, the new systems (CNC-g-Jeffamine) were characterized and new properties were identified particularly their insensitivity to salts, their ability to be redispersed in organic solvents, their surfactant character, their enhanced thermal stability and their reversible thermoagregation. The structural study using small angles neutrons scattering gave us information about thicknesses of the polymer corona and the chains conformation. New characteristics of their self-assembly into chiral-nematic phases were noticed more particularly concerning phase diagrams and cholesteric pitches. The second grafting strategy consisted in an oxidation using periodate ion followed by reductive amination. Results showed a partial detachment of cellulose chains after oxidation. A complete detachment was observed after grafting the polymer chains to finally obtain cellulose-Jeffamine copolymers which could have interesting properties to study.

Synthèse et utilisation de mimes de quadruplexes contraints pour l'évaluation de ligands II est maintenant bien connu que l'ADN simple brin peut s'associer sous différentes conformations telles que la double hélice, les triplexes, les i-motifs ou bien encore les G-quadruplexes. Ces dernières années les structures de type quadruplexes (G-quadruplexes et i-motif) ont suscité un certain intérêt notamment pour leur implications au niveau cellulaire (maintenance des télomères, activation de gènes…). C'est pourquoi de nombreuses équipes travaillent sur le développement de ligands affins pour ces structures qui pourraient agir en tant qu'anticancéreux. Cependant, le fait que les quadruplexes présentent un polymorphisme important(variabilité du nombre de brins dans la structure, des différents types de boucles et de l'orientation des brins) rend la compréhension des interactions entre un ligand et le quadruplexe plus difficile. Dans ce contexte, l'équipe développe un nouveau concept, « Template Assisted Synthesis of Quadruplexes » (TASQ) dont le but est d'obtenir un quadruplexe ne présentant qu'une topologie de façon contrôlée afin de permettre des études plus précises sur la façon dont un ligand pourrait interagir avec les quadruplexes. La première partie de ce manuscrit reporte l'évaluation par résonnance plasmonique de surface de complexes métalliques en tant que ligands de G-quadruplexe. Ces études reposent sur l'utilisation d'un premier mime de G-quadruplexe parallèle sur lequel deux séries de complexes sont testées : des métalloporphyrines et des ligands de type salphen. La seconde partie du manuscrit décrit la synthèse de mimes de G-quadruplexes antiparallèle. Elle repose sur l'utilisation du gabarit peptidique qui relié aux séquences spécifiques d'oligonucléotides de façon adéquat contraint la structure. Pour se faire, deux réactions chimiosélectives ont été utilisées : la cycloaddition 1,3 dipolaire de Huisgen et la ligation oxime. Les travaux reportés concernent trois types de structure mimant un i-motifs, des G-quadruplexes tétramoléculaires ou bimoléculaires. / Synthesis and use of constrained quadruplexe mimic for the ligand evaluation It has now been shown that single stranded DNA can fold into hairpin, triplex, i-motif and G-quadruplexe structures containing non canonical base pairing. In particular quadruplex structures (G-quadruplexes and i-motif) have generated interest because the formation is considered to have important consequences at the cellular level (telomere maintenance, oncogene activation…). Therefore, the design of small molecules (ligands) targeting quadruplexes is under development, in particular to obtain anticancerous molecules. Nevertheless G-quadruplexes exhibit a wide structural polymorphism (different kind of loop, strand orientation, number of strands) making investigation of ligand tricky. In this context, we have developed a new concept called 'Template Assisted Synthesis of Quadruplex' (TASQ) with aim to constrain the G-quadruplex in a single conformation which enables the study of the interactions of ligand more specifically for a single topology. In the first part of this manuscript, we describe studies by using surface plasmon resonance (SPR) of metal complexes as G-quadruplex ligands. The investigation was performed using a first parallel G-quadruplex as model to interact with two kinds of metal complexes: metalloporphyrins and salphen. The second part of the manuscript reports the work concerning the synthesis of antiparallel quadruplexes. The strategy is based on the use of a peptide scaffold in which the suitable oligonucleotide sequences are anchored with specific orientation. For this purpose two compatible chemistries are used (1,3 dipolar Huisgen cycloaddition and oxime ligation). Three kinds of antiparallel quadruplexe mimics have been investigated: i-motif, tetramolecular G-quadruplexes and loop-constrained G-quadruplexes.