L'ADN chromosomique des cellules eucaryotes est fortement condensé au sein d'un complexe nucléoprotéïque, la chromatine. Aussi bien l'organisation spatiale que la composition biochimique (état “épigénomique”) de la chromatine jouent un rôle fondamental dans la régulation des gènes. Grâce aux récents développements des techniques de séquençage à haut-débit, il est possible de déterminer l'état épigénomique local de la chromatine ainsi que la probabilité de contact entre deux sites génomiques (technique dite de “Hi-C”). Ces deux techniques ont permis de mettre en évidence l’existence de domaines d’interaction dont les positions corrèlent fortement avec la segmentation épigénomique de la chromatine. Cependant, les mécanismes responsables de ce couplage sont encore mal compris. L’objectif de cette thèse est de bâtir des modèles physiques permettant de valider l’hypothèse que l’épigénome est un acteur majeur dans le repliement 3D de la chromatine. Pour cela, nous avons tout d’abord développé “IC-Finder”, un algorithme permettant de segmenter les cartes Hi-C en domaines d’interaction. Nous avons alors pu quantifier précisément l’association entre épigénome et organisation de la chromatine. Les corrélations trouvées justifient l’idée de modéliser la chromatine par un copolymère par bloc dont les monomères ont chacun un état épigénomique. Dans ce cadre, nous avons développé une méthode d’inférence des potentiels d'interaction entre sites génomiques à partir des cartes Hi-C expérimentales. Ce travail permettra à plus long terme de prévoir l’organisation de la chromatine sous différentes conditions, ce qui permettra d’étudier en particulier les changements de structure résultant de l’altération de l’épigénome. / DNA of eukaryotes is highly condensed in a nucleoprotein complex called chromatin. Both the spatial organization and the biochemical composition (“epigenomic” state) of the chromatin are fundamental for gene regulation. Remarkably, recent studies indicate that1D epigenomic domains tend to fold into 3D topologically associated domains (TADs) forming specialized nuclear chromatin compartments. In this thesis, we address the question of the coupling between chromatin folding and epigenome. We first built a software called IC-finder to segment HiC maps into interacting domains. We next used it to quantify correlations between the TADs and epigenomic partitions of the genome. This led us to develop a physical model of the chromatin with the working hypothesis that chromatin organization is driven by physical interactions between epigenomic loci. We modeled chromatin as a block copolymer where each block corresponds to an epigenomic domain. With this framework, we developed a method to infer interaction parameters between chromatin loci from experimental Hi-C map. An outcome of such inference process would be a powerful tool to predict chromatin organization in various conditions, allowing investigating in silico changes in TAD formations and long-range contacts when altering the epigenome.

De nombreuses études théoriques et expérimentales ont été menées au cours des 30 dernières années afin d'établir le lien entre les propriétés d'interaction des protéines, leurs transitions de phase et leur auto-assemblage. Des avancées significatives ont ainsi été permises grâce à l'application de concepts et méthodes de la physique des polymères et des colloïdes. Ces études ont, pour la majeure partie d'entre elles, été limitées à des protéines d'intérêt médical et à des protéines animales. Ce travail de thèse vise à appliquer ce type d'approche aux protéines végétales afin de mieux comprendre leurs propriétés d'interaction à l'origine de leurs propriétés fonctionnelles au sein des grains et dans les matrices alimentaires. Ce travail a été mené sur un isolat de protéines de réserve du blé composé principalement de la fraction monomérique: les gliadines. Nous avons étudié les transitions de phase des gliadines afin de mieux comprendre leurs propriétés d'interaction d'une part et les structures associées d'autre part. Dans un premier temps, une procédure d'extraction a été développée afin de travailler sur un isolat de composition contrôlée dont les masses moléculaires sont comprises entre 20 kDa et 300 kDa. Le comportement de phase de cet isolat a ensuite été étudié en diminuant la qualité du solvant. Nous avons ainsi déterminé le diagramme de phases (T-Φ), où T est la température et Φv la fraction volumique des gliadines. Cette étude a mis en évidence une séparation de phase de type liquide-liquide dans le système par diminution de la température. Une analyse détaillée de la répartition des protéines au sein des deux phases en fonction de leur masse moléculaire a permis d'identifier une masse moléculaire critique séparant des protéines de comportement de type colloïdal et des protéines de comportement de type polymérique. A partir du diagramme de phase, deux études structurales ont été effectuées. La première a étudié les cinétiques de séparation de phase lors de la diminution de la température pour caractériser la dynamique locale de séparation de phase et identifier les mécanismes qui génèrent les systèmes concentrés. Deux grands types de mécanismes de séparation de phase ont été identifiés : nucléation-croissance et décomposition spinodale. La seconde étude structurale a consisté à établir l'équation d'état pression osmotique vs concentration dans des conditions de bon solvant et à caractériser la structure des dispersions de protéines associée. La relation pression osmotique vs fraction volumique a permis de mettre en évidence l'existence de plusieurs régimes de structuration, associés à des changements de structure secondaire et de propriété rhéologique. La discussion générale permet de mettre en relation les propriétés thermodynamiques déduites de cette approche expérimentale et les changements structuraux observés à différentes échelles. / A substantial body of theoretical and experimental studies has been conducted over the last 30 years to establish the link between protein interaction properties, phase transitions and self-assembly. Both colloidal and polymer physics provide a new framework for understanding the driving force for proteins phase behaviour. Such studies have been limited to health-related proteins and to a few food proteins, mainly animal proteins such as casein, whey proteins. This thesis aims to apply this approach to plant proteins to better understand their interactions properties, at the basis of their functional properties within grains and food matrices. This work was carried out on a wheat storage protein isolate mainly composed of the monomeric fraction: gliadins.The objective of this PhD thesis is to investigate the phase transitions of wheat proteins to develop our knowledge on their interaction properties and the associated structures. We organized our experimental approach in five steps. First, we developed an extraction procedure to work on a protein isolate of controlled composition with molecular weight ranging from 20 to 300 kg mol-1. Then, we investigated the phase behaviour of the protein isolate by decreasing the solvent quality, here the temperature. We determined the T-Φ phase diagram, where T is the temperature and Φv the protein volume fraction, that maps the phase and structural transitions of the proteins. This study showed the existence of a liquid-liquid phase separation in the system upon a temperature decrease. We evidenced two different behaviours among proteins as a function of their MWs and highlighted a critical protein size above which the molecular weight is the key determinant of the protein properties. From the phase diagram, two structural studies were conducted. The first one studied the kinetics of phase separation upon temperature decrease to characterize the local dynamics of phase separation and to identify the mechanisms that generate concentrated systems. Two main mechanisms of phase separation have been identified: nucleation-growth and spinodal decomposition. The second one studied the effect of protein concentration on the multi-scale structure of wheat gliadins in good solvent. The integration of all these results allowed us to build the phase diagram of wheat gliadins, integrating thermodynamic and structural data.

<div>Despite the abundant current applications for ionic hydrogels, much about the stimuli-responsive behavior of these materials remains poorly understood. Due to the soft nature of these materials, the number of traditional characterization methods which can be applied to these systems is limited. Many studies have been conducted to characterize bulk property responses of these materials, and experimental studies have been produced examining the distribution of free ions around single polyelectrolyte chains. However, little experimental work has been published in which molecular-scale interactions are elucidated in confined polyelectrolyte networks. Furthermore, the way in which responsive properties, other than bulk swelling capacity, scale with ionic fraction in mixed polyelectrolyte-non-polyelectrolyte hydrogel systems has not been thoroughly investigated.</div><div>The distribution and strength of polymer-counter-ion bonds has a remarkable effect on hydrogel properties such as absorption capacity, mechanical strength, and size and chemical selectivity. In order to tailor these properties for targeted applications in ionic environments, it is imperative that we thoroughly understand the character of these polymer-ion interactions and their arrangement within the bulk hydrogel. In order to do so, however, non-traditional methods of analysis must be employed.</div><div>This dissertation focuses on a model part-ionic hydrogel system, poly(sodium acrylate-co-acrylamide), in order to assess not only the polymer-counter-ion interactions but also the impact of gel ionic fraction on these interactions and the responses which they induce in gel performance properties. A model alkali (NaCl), alkaline earth (CaCl2), and transition (CuSO4) metal salt are employed to investigate changes in polymer properties from the macroscale to the nanoscale. The aim of this dissertation is to lay the foundation for the development of fundamental structure-property relationships by which we may fully understand the ionically-induced performance properties of polyelectrolyte networks.</div>

Density functional theory (DFT) calculations of polyethylene (PE) HVDC cable insulation have been performed for systems containing four different chemical impurities: acetophenone, cumene, $\alpha$-methyl styrene and $\alpha$-cumyl alcohol. Systems were generated by molecular dynamics (MD) equilibration at four different temperatures relevant for cable insulation applications: 277 K, 293 K, 343 K and 363 K. With the goal of gaining better measure of variations in hole and electron traps energies, four initial configurations were also stochastically generated at each temperature, which yielded four different final configurations after equilibration. The counterpoise correction scheme was implemented for DFT calculations, by distributing ghost atoms thought any empty pockets of space in between the PE chains. The PBE functional was selected for DFT simulations. The resulting band gaps were in agreement with those of earlier GGA-based studies, and thus lower by 3 eV than empirical band gaps. For all impurities, the first HOMO state and the first two LUMO states were generally located on the impurity molecule, forming one hole trap and two electron traps, but certain configurations generated increased electron trap numbers, or eliminated hole traps. No dependence could be derived between temperature and trap depth for either electron or hole traps. Mean electron trap energies were largely in agreement with results from earlier studies, they were deepest for acetophenone, and they varied by as much as 0.6 eV between different configurations. Hole traps are universally shallow and vary by up to 0.7 eV between configurations, and are similar in depth for all impurities. Results suggest that electron trap depths correlate with the presence of molecular features such as oxygen atoms and conjugated double bonds. The dependence of trap depth on the spatial configuration of the impurity molecule suggests that results could be improved by more precise quantum mechanical treatment of the dynamics of the impurity. / Täthetsfunktionalteori (DFT) har använts för beräkningar av isolering till HVDC kablar som består av polyeten innehållande fyra olika kemiska orenheter: acetofenon, kumen, alfa-metylstyren och alfa-kumylalkohol. System att studera genererades genom molekylärdynamisk ekvilibrering vid fyra olika temperaturer relevanta för tillämning till kabelisolering: 277 K, 293 K, 343 K och 363 K. För att få ett mått på de variationer som existerar i energierna på hål- och elektronfällor genererades stokastiskt fyra initialkonfigurationer vid varje temperatur, vilket fyra olika konfigurationer efter relaxering. Motviktskorrigering implementerades för DFT-beräkningar, genom att fördela "spökatomer" i de tomrum som bildas mellan PE-kedjorna i den amorfa fasen. PBE-funktionalen användes för DFT-simuleringar. De resulterande bandgapen stämde överens med tidigare GGA-baserade studier, och var därmed runt 3 eV smalare än empiriskt uppmätta bandgap. För alla orenheter var det första HOMO-tillståndet och de två första LUMO-tillstånden i allmänhet placerade på orenheten, vilket resulterade i en hålfälla och två elektronfällor, men vissa konfigurationer gav upphov till fler elektronfällor, eller eliminerade hålfällorna. Inget samband kunde härledas mellan temperaturen och djupet på fällorna för vare sig elektron- eller hålfällor. Medelvärdet på elektronfällornas energier överensstämde till stor del med resultat från tidigare studier, energierna var högst för acetofenon, och de varierade med så mycket som 0,6 eV mellan olika konfigurationer. Hålfällorna var genomgående grunda, varierade med upp till 0,7 eV mellan olika konfigurationer, och hade likartat djup för alla orenheter. Resultaten indikerar att variationerna elektronfällornas medeldjup uppstår på grund av orenheternas olika molekylära uppbyggnad: förekomst av syreatomer och konjugerade dubbelbindningar i orenheterna leder till djupare elektronfällor. Det faktum att djupet på elektron- och hålfällor varierar mellan olika rumsliga konfigurationer av av orenheten och polyetenstrukturen ger en antydan om att resultaten kan komma att förbättras om dynamiken hos orenheten simuleras med mer exakta kvantmekanisk metoder.

Binary mixtures of polymers and soft nanocolloids, also called as polymer nanocomposites are well known and studied for their enormous potentials on various technological fronts. In this thesis blends of polystyrene grafted gold nanoparticles (PGNPs) and polystyrene (PS) are studied experimentally, both in bulk and in thin films. This thesis comprises three parts; 1) evolution of microscopic dynamics in the bulk(chapter-3),2) dispersion behavior of PGNPs in thin and ultra thin polymer matrices (chapter-4) 3) effect of dispersion on the glass transition behavior (chapter-5). In first part, the state of art technique, x-ray photon correlation spectroscopy is used to study the temperature and wave vector dependent microscopic dy¬namics of PGNPs and PGNP-PS mixtures. Structural similarities between PGNPs and star polymers (SPs) are shown using small angle x-ray scatter¬ing and scaling relations. We find unexpected (when compared with SPs) non-monotonic dependence of the structural relaxation time of the nanoparticles with functionality (number of arms attached to the surface). Role of core-core attractions in PGNPs is shown and discussed to be the cause of anomalous behavior in dynamics. In PGNP-PS mixtures, we find evidence of melting of the dynamically arrested state of the PGNPs with addition of PS followed by a reentrant slowing down of the dynamics with further increase in polymer frac¬tion, depending on the size ratio(δ)of PS and PGNPs. For higher δ the reen¬trant behavior is not observed with polymer densities explored here. Possible explanation of the observed dynamics in terms of the presence of double-glass phase is provided. The correlation between structure and reentrant vitrifica¬tion in both pristine PGNPs and blends are derived rather qualitatively. In the second part, the focus is shifted to miscibility between PGNPs and polymers under confinement i.e., in thin films. This chapter provide a compre¬hensive study on the different parameters affecting dispersion viz., annealing conditions, fraction of the added particles, polymer-particle interface and more importantly the thickness of the films. Changes in the dispersion behavior with annealing is shown and the need for annealing the films at temperatures higher than the glass transition temperature of the matrix polymers is clearly elucidated. Irrespective of the thickness of the films( 20 and 65 nm) studied, immiscible particle-polymer blends unequivocally prove the presence of gradi¬ent in dynamics along the depth of the films. To our knowledge for the first time, we report results on confinement induced enhancement in the dispersion of the nanoparticles in thin polymer films. The enhanced dispersion is argued to be facilitated by the increased free volume in the polymer due to confinement as shown by others. Based on these results we have proposed a phase diagram for dispersibility of the nanoparticles in polymer films. The phase diagram for ultra thin films highlights an important point: In ultra thin films the particles are dispersed even with grafting molecular weight less than matrix molecular weight. In the third part, we have studied the glass transition of the thin films whose structure has been studied earlier in the earlier part. Non-monotonic variation in glass transition with the fraction of particles in thin films has increased our belief on the gradient in the dynamics of thin polymer films. En¬hanced dispersion with confinement is captured with the enhanced deviation in glass transition temperature of ultra thin films. Effect of miscibility param¬eter on Tgis studied and the results are explained with the subtle interplay of polymer-particle interface and confinement.

Plusieurs indices suggèrent que le repliement du chromosome et la régulation de l’expression génétique sont étroitement liés. Par exemple, la co-expression d’un grand nombre de gènes est favorisée par leur rapprochement dans l’espace cellulaire. En outre, le repliement du chromosome permet de faire émerger des structures fonctionnelles. Celles-ci peuvent être des amas condensés et fibrillaires, interdisant l’accès à l’ADN, ou au contraire des configurations plus ouvertes de l’ADN avec quelques amas globulaires, comme c’est le cas avec les usines de transcription. Bien que dissemblables au premier abord, de telles structures sont rendues possibles par l’existence de protéines bivalentes, capable d’apparier des régions parfois très éloignées sur la séquence d’ADN. Le système physique ainsi constitué du chromosome et de protéines bivalentes peut être très complexe. C’est pourquoi les mécanismes régissant le repliement du chromosome sont restés majoritairement incompris.Nous avons étudié des modèles d’architecture du chromosome en utilisant le formalisme de la physique statistique. Notre point de départ est la représentation du chromosome sous la forme d’un polymère rigide, pouvant interagir avec une solution de protéines liantes. Les structures résultant de ces interactions ont été caractérisées à l’équilibre thermodynamique. De plus, nous avons utilisé des simulations de dynamique Brownienne en complément des méthodes théoriques, car elles permettent de prendre en considération une plus grande complexité dans les phénomènes biologiques étudiés.Les principaux aboutissements de cette thèse ont été : (i) de fournir un modèle pour l’existence des usines de transcriptions caractérisées in vivo à l’aide de microscopie par fluorescence ; (ii) de proposer une explication physique pour une conjecture portant sur un mécanisme de régulation de la transcription impliquant la formation de boucles d’ADN en tête d’épingle sous l’effet de la protéine H-NS, qui a été émise suite à l’observation de ces boucles au microscope à force atomique ; (iii) de proposer un modèle du chromosome qui reproduise les contacts mesurés à l’aide des techniques Hi-C. Les conséquences de ces mécanismes sur la régulation de la transcription ont été systématiquement discutées. / Increasing evidences suggest that chromosome folding and genetic expression are intimately connected. For example, the co-expression of a large number of genes can benefit from their spatial co-localization in the cellular space. Furthermore, functional structures can result from the particular folding of the chromosome. These can be rather compact bundle-like aggregates that prevent the access to DNA, or in contrast, open coil configurations with several (presumably) globular clusters like transcription factories. Such phenomena have in common to result from the binding of divalent proteins that can bridge regions sometimes far away on the DNA sequence. The physical system consisting of the chromosome interacting with divalent proteins can be very complex. As such, most of the mechanisms responsible for chromosome folding and for the formation of functional structures have remained elusive.Using methods from statistical physics, we investigated models of chromosome architecture. A common denominator of our approach has been to represent the chromosome as a polymer with bending rigidity and consider its interaction with a solution of DNA-binding proteins. Structures entailed by the binding of such proteins were then characterized at the thermodynamical equilibrium. Furthermore, we complemented theoretical results with Brownian dynamics simulations, allowing to reproduce more of the biological complexity.The main contributions of this thesis have been: (i) to provide a model for the existence of transcrip- tion factories characterized in vivo with fluorescence microscopy; (ii) to propose a physical basis for a conjectured regulatory mechanism of the transcription involving the formation of DNA hairpin loops by the H-NS protein as characterized with atomic-force microscopy experiments; (iii) to propose a physical model of the chromosome that reproduces contacts measured in chromosome conformation capture (CCC) experiments. Consequences on the regulation of transcription are discussed in each of these studies.