Automatic translation has seen tremendous progress in recent years, mainly thanks to statistical methods applied to large parallel corpora. Transductions represent a principled approach to modeling translation, but existing transduction classes are either not expressive enough to capture structural regularities between natural languages or too complex to support efficient statistical induction on a large scale. A common approach is to severely prune search over a relatively unrestricted space of transduction grammars. These restrictions are often applied at different stages in a pipeline, with the obvious drawback of committing to irrevocable decisions that should not have been made. In this thesis we will instead restrict the space of transduction grammars to a space that is less expressive, but can be efficiently searched. First, the class of linear transductions is defined and characterized. They are generated by linear transduction grammars, which represent the natural bilingual case of linear grammars, as well as the natural linear case of inversion transduction grammars (and higher order syntax-directed transduction grammars). They are recognized by zipper finite-state transducers, which are equivalent to finite-state automata with four tapes. By allowing this extra dimensionality, linear transductions can represent alignments that finite-state transductions cannot, and by keeping the mechanism free of auxiliary storage, they become much more efficient than inversion transductions. Secondly, we present an algorithm for parsing with linear transduction grammars that allows pruning. The pruning scheme imposes no restrictions a priori, but guides the search to potentially interesting parts of the search space in an informed and dynamic way. Being able to parse efficiently allows learning of stochastic linear transduction grammars through expectation maximization. All the above work would be for naught if linear transductions were too poor a reflection of the actual transduction between natural languages. We test this empirically by building systems based on the alignments imposed by the learned grammars. The conclusion is that stochastic linear inversion transduction grammars learned from observed data stand up well to the state of the art.

La plupart des bactéries vivent en communautés organisées, appelées biofilms, augmentant leur résistance aux traitements antibiotiques. Ainsi, la formation de biofilms sur les organes et matériels médicaux est considérée comme la cause de la majorité des infections bactériennes. Il est alors important de trouver des traitements pour empêcher ou perturber la formation de ces biofilms. Il a été proposé que les porines de P. Stuartii, Omp-Pst1 et Omp-Pst2, peuvent s’auto-assembler via un steric-zipper, étape responsable du développement initial du biofilm. Ainsi, notre objectif est de synthétiser des inhibiteurs d’interactions entre porines pour limiter ce contact intercellulaire. Nous avons développé des molécules peptidomimétiques basées sur la séquence LGNYR, active dans les deux porines. Pour cela, des réactions de chimie click sur phase solide ont été mises en oeuvre afin de synthétiser des analogues de cette séquence, comme la CuAAC pour introduire un motif triazole, dans une position variable au sein du peptide. Nous avons ainsi développé une méthode rapide et efficace afin de réaliser cette réaction par l’utilisation d’un catalyseur cuivre(I)-N-hétérocyclique carbène stable à l’air. Similairement, de nouvelles conditions ont aussi été mises au point en phase solide afin d’obtenir régiosélectivement des peptides comportant un motif isoxazole 3,4- ou 3,5-disubstitué, par la réaction entre un alcyne et un oxyde de nitrile. Ces cycloadditions 1,3-dipolaires nous ont ainsi permis d’obtenir une première librairie de peptidotriazoles et de peptidoisoxazoles. Il sera enfin possible d’étudier les biofilms grâce à la synthèse de sondes fluorescentes basées sur les inhibiteurs montrant de fortes affinités avec la cible, couplées à un fluorophore dérivé de coumarine. / In the environment, most of bacteria live as organized communities, known as biofilms, enhancing their resistance to antibiotic treatments. Thus the formation of biofilms on organ and indwelling medical devices is considered to cause the majority of bacterial infections in the human body. Thus, to enhance antibiotics efficacy, there is a high need to find treatments to prevent or disrupt biofilm formation. It has been proposed that P. stuartii porins, Omp-Pst1 and Omp-Pst2, can self-associate through a steric zipper, being responsible for the initial development of biofilms. Thus, our objective is to synthesize porin’s self-matching interactions (PSMI) inhibitors to counterfeit this intercellular contact. We developed peptidomimetic molecules based on the LGNYR sequence, that have shown to be active in both porins. Then we used click chemistry to synthesize on solid phase analogues of this sequence, as the solid phase Cu(I)-catalyzed azide-alkyne cycloaddition (CuAAC) to introduce a triazole moiety into the peptide chain at different positions. We thus developed a fast and efficient method to perform this reaction using a stable copper(I)-N-heterocyclic carbene catalyst. Similarly, new conditions were developed on solid phase to synthesize regioselectively peptides containing a 3,4- or 3,5-disubstituted isoxazole moiety, through the reaction between an alkyne and a nitrile oxide. These 1,3-dipolar cycloadditions allowed us to developed a first library of peptidotriazoles and peptidoisoxazoles. We also obtained fluorescent probes based on the inhibitors showing higher affinity for the target as tools to study biofilm formation.

My doctoral thesis is comprised of two bodies of research: two artworks taking the form of installations (videos, audio recordings, textiles, texts), which will be exhibited for viva. Femme Fabrications, 2016, is made from research into the American artist Joseph Cornell's (1903-1972) source materials held in the Smithsonian American Art Museum alongside research on Jean Wilkinson's 1977 book Flower Fabrications. A series of textile works encased in silk lined boxes trace my step-by-step construction of a rose from organdie. The floral emblem of the white rose (dried), 'death is preferable to a loss of innocence' , becomes an editing device, which I use to consider a number of possible recipients for the rose, such as Cornell himself. Spoken word audio recordings, which ruminate on how his sexuality pertain to the criteria of the rose are edited together with home-camcorder video footage of the house that Cornell lived in for most of his life - the house he made the entirety of his artworks within. Central to The Cesspool of Rapture, 2017, are moving-image studies of zippers, stains, rips, abrasions, openings, and closings in a series of dresses made by the American couturier Charles James (1906-1978). These videos register and move through the material research, the garments, at alternating speeds. The changing speed is registered in sound by clicks synced to each individual frame. It is at times violent and at other times tentative and gentle as the uncovering of the damage to the dresses unfolds. Audio recordings of James explicating his interests in eroticism and sexuality persistently interject the footage. This work includes the installation of a series of reconstructions of James's 1932 Taxi dress. Its black linen body is reconfigured and abstracted as the splayed design makes unfinished seams and unzipped zippers visible. In each artwork I configure viewing and consuming as a mode of authorship. I show how these diverse processes of identification become authored acts. When drawn into intimate relation with the leftover material of these historical authors, my contamination proves deviant: I gain possession of the capacity to speak for them, to expose, idolise, misrepresent, and fictionalise them. My thesis is composed of this group of methodologies, which were found and developed within the production of the artworks.

This thesis investigates the performance of concentrically braced zipper frames through complementary experimental and numerical simulation approaches and proposes a design methodology for an innovative bracing scheme labeled as the suspended zipper frame. The suspended zipper frame intends to ensure that the top-story hat truss remains elastic, resulting in very ductile behavior of the structure. In the first part of the work, a three-story prototype frame was designed based on a preliminary design method. Three tests were conducted on one-third scale models of this prototype to verify the design procedure and assess the system performance under very different load histories. Comparisons of the results between analyses and experiments validated the partial-height zipper mechanism envisioned, and led to refinements of the design procedure and establishment of appropriate design details for these frames. The design and performance of this structural system are illustrated with three-, nine-, and twenty-story buildings designed for the same masses as those used in the SAC studies for the Los Angeles area. The proposed design strategy results in suspended zipper frames having more ductile behavior and higher strength than typical zipper frames. In addition, the suspended zipper frames also appear to reduce the tendency of chevron-braced frames to form soft stories and to improve seismic performance without having to use overly stiff beams. Finally, an explanation of the design philosophy as well as code language format of the design procedure is given.

Halorhodospira halophila photoactive yellow protein (PYP) is a promising candidate to act as a photoswitching domain in engineered proteins due to the structural changes that occur during its photocycle. Absorption of a photon of wavelength 446 nm triggers trans to cis isomerization of its 4-hydroxycinnamic acid chromophore leading to large structural perturbations in the protein, particularly in the N-terminus. In the dark, a slower cis to trans reisomerization of the chromophore restores the protein’s native fold. The fusion of proteins to PYP’s N-terminus may therefore enable photomodulation of the activity of the attached protein. To test this hypothesis, this thesis descibes genetically encoded photoswitchable DNA-binding proteins that were developed by fusing the prototypical leucine-zipper type DNA-binding protein GCN4 bZIP to the N-terminus of PYP. Five different fusion constructs of full length or truncated GCN4 bZIP and full length PYP as well as fusion constructs of full length GCN4 bZIP and N-terminally truncated PYP mutants were designed in a structure-based approach to determine if the dimerization and DNA binding activities could be controlled by the PYP photocycle. Extensive biophysical characterization of the fusion constructs in the dark and under blue light irradiation using electronic absorption, circular dichroism and fluorescence spectroscopic techniques were performed. As all the fusion proteins could complete photocycles, the DNA binding abilities of the dark and light-adapted states of the proteins were characterized using spectroscopic techniques as well as by the electrophoretic mobility shift assay. All the fusion constructs maintained DNA-binding abilities, however they each differed in their affinities and the extent to which they were activated by blue light irradiation. The reasons for these differences in DNA-binding abilities and photoactivation are explored. Using the results from the characterization of these constructs, proposals are also made to develop more robust genetically encoded photoactivatable DNA-binding proteins of the same type.

Halorhodospira halophila photoactive yellow protein (PYP) is a promising candidate to act as a photoswitching domain in engineered proteins due to the structural changes that occur during its photocycle. Absorption of a photon of wavelength 446 nm triggers trans to cis isomerization of its 4-hydroxycinnamic acid chromophore leading to large structural perturbations in the protein, particularly in the N-terminus. In the dark, a slower cis to trans reisomerization of the chromophore restores the protein’s native fold. The fusion of proteins to PYP’s N-terminus may therefore enable photomodulation of the activity of the attached protein. To test this hypothesis, this thesis descibes genetically encoded photoswitchable DNA-binding proteins that were developed by fusing the prototypical leucine-zipper type DNA-binding protein GCN4 bZIP to the N-terminus of PYP. Five different fusion constructs of full length or truncated GCN4 bZIP and full length PYP as well as fusion constructs of full length GCN4 bZIP and N-terminally truncated PYP mutants were designed in a structure-based approach to determine if the dimerization and DNA binding activities could be controlled by the PYP photocycle. Extensive biophysical characterization of the fusion constructs in the dark and under blue light irradiation using electronic absorption, circular dichroism and fluorescence spectroscopic techniques were performed. As all the fusion proteins could complete photocycles, the DNA binding abilities of the dark and light-adapted states of the proteins were characterized using spectroscopic techniques as well as by the electrophoretic mobility shift assay. All the fusion constructs maintained DNA-binding abilities, however they each differed in their affinities and the extent to which they were activated by blue light irradiation. The reasons for these differences in DNA-binding abilities and photoactivation are explored. Using the results from the characterization of these constructs, proposals are also made to develop more robust genetically encoded photoactivatable DNA-binding proteins of the same type.

Orientador: Michel Georges Albert Vincentz / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Biologia / Made available in DSpace on 2018-08-25T14:21:25Z (GMT). No. of bitstreams: 1 Viana_AmericoJoseCarvalho_D.pdf: 5804951 bytes, checksum: a3c65fa34ad298641a9b177952050af3 (MD5) Previous issue date: 2014 / Resumo: O fator de transcrição do tipo basic leucine leucine zipper (bZIP) de Arabidopsis thaliana AtbZIP63 faz parte da via de resposta a carência energética coordenada pelas quinases KIN10/11, integradoras centrais dos sinais relacionados ao estado de privação de energia. O mutante de inserção de T-DNA atbzip63-2 apresenta uma redução do crescimento e desenvolvimento das folhas assim como um atraso do florescimento em comparação ao tipo selvagem (TS, acesso Ws) quando cultivado em fotoperíodo de dia curto (10 h/14 h). Condições de fotoperíodo de dia longo ou luz contínua promoveram uma reversão parcial ou completa, respectivamente, do fenótipo mutante para o tipo selvagem, levantando a possibilidade de que este fenótipo seja o resultado de uma carência energética. Plantas silenciadas para expressão de AtbZIP63 por RNAi apresentaram características similares a do mutante atbzip63-2 confirmando o envolvimento deste fator de transcrição no crescimento. O perfil de expressão gênica e os níveis de alguns metabólitos do mutante atbzip63-2 indicaram que AtbZIP63 participa do controle da degradação do amido, pois a expressão de alguns genes centrais na degradação deste carboidrato de reserva está desregulada neste mutante. Mostramos que as oscilações no nível do transcrito AtbZIP63 são reguladas pelo relógio circadiano e a fase da oscilação do AtbZIP63 é aparentemente influenciada pela disponibilidade de carboidratos na célula. Além de estar sob o controle do relógio, AtbZIP63 também atua como um ativador direto da expressão de PRR7, que codifica um dos componentes chave do oscilador central do relógio. Portanto, evidenciamos uma interação recíproca entre o relógio e AtbZIP63 que possivelmente está impactando o processo de degradação do amido à noite. Este conjunto de evidências revela novos aspectos do ajuste do relógio circadiano pelo status de açúcar na célula que estão de acordo com trabalhos recentes mostrando que os açúcares afetam diretamente o funcionamento do relógio. Nossa hipótese é que o AtbZIP63 está agindo como um mediador entre a disponibilidade de viii açúcar e o mecanismo oscilatório do relógio circadiano de A. thaliana. Adicionalmente, verificamos que o perfil de transcritos no final do dia no mutante atbzip63-2 é diferente do observado no final da noite, sugerindo a participação do AtbZIP63 na regulação de genes envolvidos em redes regulatórias distintas em função do período do dia. Dentre os genes desregulados no atbzip63-2 no final do dia, observamos um enriquecimento para genes relacionados com metabolismo secundário e síntese de trealose, o que sugere a participação do AtbZIP63 na regulação da síntese destes compostos durante o dia, e possivelmente reflete a ocorrência de stress no mutante / Abstract: he Arabidopsis thaliana basic leucine zipper domain (bZIP) AtbZIP63 transcription factor is part of the response pathway to energy shortage coordinated by kinases KIN10/11. The T-DNA insertion mutant atbzip63-2 shows a reduction in the growth and development of leaves, as well as a delay in flowering compared to wild type (WT; ecotype Ws), when grown in short-day conditions. Long day or continuous light conditions promoted a partial or complete reversion, respectively, of the mutant to wild-type phenotype, raising the possibility that this phenotype is the result of an energy shortage. Plants silenced for AtbZIP63 showed similar characteristics to the atbzip63-2 mutant, confirming the involvement of this transcription factor in the growth. The gene expression profile and the levels of some metabolites of the atbzip63-2 indicated that AtbZIP63 takes part in the control of starch degradation, regulating the expression of some key genes in starch degradation. Diurnal AtbZIP63 mRNA level fluctuation is regulated by the circadian clock, and the phase oscillation is influenced by the availability of carbohydrates. In addition, to be controlled by the circadian clock, AtbZIP63 directly regulates the expression of PRR7 which encodes one of the key regulators of the core clock. We have therefore identified a reciprocal interaction between the clock and AtbZIP63 which is probably affecting the starch degradation process. This set of evidence reveals new aspects of the entrainment of the circadian clock by sugars, and is consistent with recent studies showing that sugars directly regulate the circadian clock. Our hypothesis is that AtbZIP63 is acting as a mediator between the energy status (availability of sugar) and the oscillatory mechanism of the A. thaliana circadian clock. Additionally, we found that the profile of transcripts at the end of the day in atbzip63-2 mutant is different from that observed in the end of the night, suggesting the involvement of AtbZIP63 in the regulation of genes involved in distinct regulatory networks according to the period of day. Among the genes deregulated in atbzip63-2 at the end of the x day, an enrichment for genes related to secondary metabolism and trehalose biosynthesis was observed. Suggesting the involvement of AtbZIP63 in regulating the synthesis of these compounds during the day, and probably reflects the occurrence of stress in the mutant / Doutorado / Genetica Vegetal e Melhoramento / Doutor em Genetica e Biologia Molecular

Herewithin, two transcription factor (TF) regulatory complexes were investigated. A bHLH–MYB–WDR (BMW) DNA-binding complex from maize was the first complex to be studied. R, a maize bHLH involved in the activation of genes in the anthocyanin pathway, had been characterized to indirectly bind DNA despite the presence of a functional DNA-binding domain. Findings presented here reveal that this is only partially correct. Direct DNA-binding by R was found to be dependent upon two distinct dimerization domains that function as a switch. This switch-like mechanism allows R to be repurposed for the activation of promoters of differing cis-element structure. The second regulatory complex studied was of the Arabidopsis thaliana MIKC-MADS TF family. For many TFs, DNA-binding site recognition is relatively straightforward and very sequence specific, while others exhibit relaxed sequence specificity. MADS-domain TFs are one family of TFs with a wider range of cis-element sequences. Though consensus cis-element sequences have been determined for various MADS-domains, correctly predicting and identifying biologically functional cis-elements has been a challenge. In order to study the influence of nucleobase associations within the cis-element, a DNA-Protein Interaction (DPI)-ELISA method was modified and optimized to screen a panel of specific probes. Screening of the SEP3 homodimer against a panel of sequential, palindromic probes revealed that nucleobases in position -1:+1 of the CArG-box influence binding strength between the MADS-domain and DNA. Additionally, the specificity of AGL15 towards CT-W6-AG forms was discovered to be determined by the functional groups present in the minor groove at position -4:+4 using inosine:cytosine (I:C) base pairs. Finally, the FLC–SVP MADS-domain heterodimer, bound to a native cis-element, was modeled and binding simulated using molecular dynamics. In conjunction with simulations of AGL15 and SEP3 homodimers, a potential binding mechanism was identified for this unique heterodimer. DNA sequence recognition by the MADS-domain was found to occur asymmetrically. In the case of the FLC–SVP heterodimer, the direction of asymmetrical DNA-binding in heterodimers was found to be fixed. Furthermore, the molecular dynamics simulations provided insight towards understanding the results generated from previous DPI-ELISA experiments, which should provide an improved means for predicting biologically significant CArG-boxes around genes.

La Programmation Orientée Langage est un paradigme de programmation qui tente, par la technique de méta-programmation, de changer les habitudes des développeurs de systèmes informatiques en leur permettant de "travailler en termes de concepts et notions du problème à résoudre, au lieu d'être toujours obligés de traduire leurs idées aux notions qu'un langage généraliste est capable de comprendre" (Ward et Sergey). Le développement de logiciels passe de ce fait par la conception de langages dédiés : on définit un ou plusieurs langages qui capturent les caractéristiques du domaine étudié, puis on écrit les applications visées en utilisant ces langages. Dans cette thèse, nous proposons une démarche méthodologique de développement logiciel reposant sur ce concept. Il s'agit de conduire la même démarche méthodologique au niveau des langages que ce qui est classiquement fait au niveau des composants logiciels. En l'occurrence, nous utilisons le formalisme des grammaires attribuées pour tenter de répondre à la question suivante : comment peut-on créer de nouveaux langages par composition de langages réutilisables existants ? Nous tirons profit de leur traduction en algèbres de combinateurs fonctionnels pour définir des spécifications exécutables de langages dédiés (vus comme composants logiciels), plongés dans le langage fonctionnel pur Haskell. \' partir d'exemples significatifs d'extension et de réutilisation de langages dédiés (par stratification de ceux-ci, ou par changement de monade), nous proposons un typage de langages dédiés en vue de leur assemblage et leur réutilisation. Pour illustrer cette démarche, nous décrivons un langage dédié (bibliothèque de combinateurs) pour l'édition de documents structurés. Un document y est représenté par un zipper attribué, une structure arborescente localisable, représentant un arbre et son contexte, et caractérisée par une grammaire attribuée. L'édition consiste alors à la modification interactive de cette structure ; ce qui entraîne une réévaluation totale ou partielle des attributs. L'édition peut aussi être réalisée à travers une vue abstraite obtenue par projection de la structure concrète. Ce qui pose le problème de \textit, un problème familier de la communauté des bases de données, auquel nous donnons une solution grâce à nos combinateurs d'éditeurs.

Transcription factors are proteins that regulate transcription of genes by binding to specific DNA sequences proximal to the gene. The specificity and affinity of protein-DNA recognition is critical for proper gene regulation. This thesis explores the mechanisms of binding to the sequence 5’CACGTG, a common recognition sequence both in plants where it is known as the G-box and in mammalian cells where it is termed the E-box. This sequence is of clinical interest because it is the target of the transcription factor Myc, an oncogene linked to many cancers. A number of alpha-helical proteins with different dimerization elements, from the basic region-leucine zipper (bZIP), basic region helix-loop-helix leucine zipper (bHLHZ) and basic region helix-loop-helix-PAS (bHLH-PAS) protein families, are capable of binding to this sequence. The basic regions of all these protein families contain residues that contact DNA and determine DNA sequence specificity while the other subdomains are responsible for dimerization specificity. First, the influence of protein-DNA contacts on sequence specificity of the plant bZIP protein EmBP-1 was probed by point mutations in the basic region. Residues that contact the DNA outside the core G-box sequence and residues that contact the phosphate backbone were found to be important for sequence specificity. Second, the impact of the dimerization subdomains of bHLHZ protein Max, the required heterodimerization partner of the Myc protein, and bHLH-PAS protein Arnt was probed by mutation, deletion and inter-family subdomain swapping studies. All studied protein families are intrinsically disordered, forming structure upon dimerization and DNA binding. The dimerization domains were found to indirectly influence DNA binding by affecting folding, dimerization ability or proper orientation of the basic regions relative to DNA. Lastly, a new strategy for selection of G-box binding proteins in the Yeast One-hybrid system is explored. Together, these studies broaden our understanding of the structure-function relationship of the DNA-binding activities of these closely related families of transcription factors. The creation and characterization of mutants with altered specificity, affinity and dimerization specificity may also be useful for biotechnology applications.