Extreme conditions allow us to reveal unusual material properties. At the same time an experimental approach is di-cult under such conditions. Capabilities of a theoretical approach based on simplied models are limited. This explainsa wide application of computer simulations at extreme conditions. My thesis is concerned with computer simulations undersuch a conditions. I address such problems as melting, solidsolid phase transitions, shockwave impact on material properties and chemical reactions under extreme conditions. We addressed these problems to facilitate simulations of phase transitions to provide some interpretation of experimental data andexplain enigmatic phenomena in interior of the Earth. / <p>QC 20160615</p>

Yes / The adsorption of three homo-tripeptides, HHH, YYY, and SSS, at the aqueous Au interface is investigated, using molecular dynamics simulations. We find that consideration of surface facet effects, relevant to experimental conditions, opens up new questions regarding interpretations of current experimental findings. Our well-tempered metadynamics simulations predict the rank ordering of the tripeptide binding affinities at aqueous Au(111) to be YYY > HHH > SSS. This ranking differs with that obtained from existing experimental data which used surface-immobilized Au nanoparticles as the target substrate. The influence of Au facet on these experimental findings is then considered, via our binding strength predictions of the relevant amino acids at aqueous Au(111) and Au(100)(1 × 1). The Au(111) interface supports an amino acid ranking of Tyr > HisA ≃ HisH > Ser, matching that of the tripeptides on Au(111), while the ranking on Au(100) is HisA > Ser ≃ Tyr ≃ HisH, with only HisA showing non-negligible binding. The substantial reduction in Tyr amino acid affinity for Au(100) vs Au(111) offers one possible explanation for the experimentally observed weaker adsorption of YYY on the nanoparticle-immobilized substrate compared with HHH. In a separate set of simulations, we predict the structures of the adsorbed tripeptides at the two aqueous Au facets, revealing facet-dependent differences in the adsorbed conformations. Our findings suggest that Au facet effects, where relevant, may influence the adsorption structures and energetics of biomolecules, highlighting the possible influence of the structural model used to interpret experimental binding data. / Air Office of Scientific Research, Grant No. FA9550-12-1-0226

It has been 10 years since the first hybrid halide perovskite photovoltaics was fabricated. Power conversion efficiency increases from the initial 3.8% to the current 25.2%. Fabrication method envolves from spin-coating to printable technology, and we deeply experience the drastic development of hybrid halide perovskite photovoltaics. Although hybrid halide photovoltaics render a variety of advantages over traditional photovoltaics, we still cannot find any practical application of these hybrid halide photovoltaics. There exist a few issues which hinder the commercialization of this type of solar cell. Among these issues, the long-term instability of hybrid halide perovskite is the main concern for the next development. This thesis expands on investigating the instability of hybrid halide perovskites from first principles. In Chapter 1, two computational methods employed in the thesis: density functional theory and Ab initio molecular dynamics are introduced. Theoretical investigations of the instability of CH3NH3PbI3 using density functional theory method are mainly conducted at 0 K. The finite temperature effect on this instability of CH3NH3PbI3 is usually neglected. In Chapter 2 of this thesis, we combined density functional calculations and additional thermodynamic data to explain the intrinsic instability of CH3NH3PbI3 under finite temperature conditions. We also analyzed the stability under humid conditions. It is shown that the aqueous solubilities of reactants play an important role in the products’ stabilities. The Born–Haber cycle of NaCl splits the enthalpy change into several components which will give a better understanding of the origin of the corresponding enthalpy change. In Chapter 3, with the extension of the Born–Haber cycle to the hybrid halide perovskites, the reaction enthalpies which govern the intrinsic instability of ionic compounds were analyzed. We proposed a criterion that helps to filter the hybrid halide perovskites with improved stability aimed for photovoltaics. Since the instability of CH3NH3PbI3 is intrinsic. The long-term instability can be settled by discovering alternative perovskite absorber. In Chapter 4, based on literature research, we propose a three-membered ring cation which has a suitable size to fit into the Pb-I framework, leading to optimal band gap for photovoltaics. Besides, the cation has a good ionization energy which will potentially render better stability. Whereas, a comprehensive study of this cyclic ring based perovskite indicates that the instability of the three-membered ring cation will make it impossible to synthesize this theoretical structure. Moisture degradation mechanisms of CH3NH3PbI3 are investigated intensively. More importantly, for practical photovoltaics, we have to imagine different situations the modules will encounter, e.g. after a couple of years, cracks appearing on the modules are inevitable, at this stage, understanding of the degradation mechanism of CH3NH3PbI3 according to liquid water becomes important. Chapter 6 elaborately describes a comprehensive degradation mechanism of CH3NH3PbI3 under liquid water. We investigate the energy barrier for the first dissolution event of CH3NH3PbI3 in water. Furthermore, thermodynamic analyses of CH3NH3PbI3 dissolution in water clearly explain the spontaneity of CH3NH3PbI3 degradation in water. Besides, different mechanisms of CH3NH3PbI3 and CsPbI3 dissolution in water are discussed. / Dissertation / Doctor of Philosophy (PhD)

Apart from biological functions, peptides are of uttermost importance as models for un- folded, denatured or disordered state of the proteins. Similarly, miniproteins such as Trp-cage have proven their role as simple models of both experimental and theoretical studies of protein folding. Molecular dynamics and computer simulations can provide an unique insight on processes at atomic level. However, simulations of peptides and minipro- teins face two cardinal problems-inaccuracy of force fields and inadequate conformation sampling. Both principal issues were tackled in this theses. Firstly, the differences in several force field for peptides and proteins were questioned. We demonstrated the inability of the used force fields to predict consistently intrinsic conformational preferences of individual amino acids in the form of dipeptides and the source of the discrepancies was traced. In order to shed light on the nature of conformational ensembles under various denatur- ing conditions, we studied host-guest AAXAA peptides. The simulations revealed that thermal and chemical denaturation by urea produces qualitatively different ensembles and shift propensities of individual amino acids to particular conformers. The problem of insufficient conformation sampling was dealt by introducing gyration- and...

Genetic sequences are being collected at an ever increasing rate due to rapid cost reductions; however, experimental approaches to determine the structure and function of the protein(s) each gene codes are not keeping pace. Therefore, computational methods to augment experimental structures with comparative (i.e. homology) models using physics-based methods for building residues, loops and domains are needed to thread new sequences onto homologous structures. In addition, even experimental structure determination relies on analogous first principles structure refinement and prediction algorithms to place structural elements that are not defined by the data alone. Computational methods developed to find the global free energy minimum of an amino acid sequence (i.e. the protein folding problem) are increasingly successful, but limitations in accuracy and efficiency remain. Optimization efforts have focused on subsets of systems and environments by utilizing potential energy functions ranging from fixed charged force fields (Fiser, Do, & Sali, 2000; Jacobson et al., 2004), statistical or knowledge based potentials (Das & Baker, 2008) and/or potentials incorporating experimental data (Brunger, 2007; Trabuco, Villa, Mitra, Frank, & Schulten, 2008). Although these methods are widely used, limitations include 1) a target function global minimum that does not correspond to the actual free energy minimum and/or 2) search protocols that are inefficient or not deterministic due to rough energy landscapes characterized by large energy barriers between multiple minima. Our Global Optimization Using Metadynamics and a Polarizable Force Field (GONDOLA) approach tackles the first limitation by incorporating experimental data (i.e. from X-ray crystallography, CryoEM or NMR experiments) into a hybrid target function that also includes information from a polarizable molecular mechanics force field (Lopes, Roux, & MacKerell, 2009; Ponder & Case, 2003). The second limitation is overcome by driving the sampling of conformational space by adding a time-dependent bias to the objective function, which pushes the search toward unexplored regions (Alessandro Barducci, Bonomi, & Parrinello, 2011; Zheng, Chen, & Yang, 2008). The GONDOLA approach incorporates additional efficiency constructs for search space exploration that include Monte Carlo moves and fine grained minimization. Furthermore, the dimensionality of the search is reduced by fixing atomic coordinates of known structural regions while atoms of interest explore new coordinate positions. The overall approach can be used for optimization of side-chains (i.e. set side-chain atoms active while constraining backbone atoms), residues (i.e. side-chain atoms and backbone atoms active), ligand binding pose (i.e. set atoms along binding interface active), protein loops (i.e. set atoms connecting two terminating residues active) or even entire protein domains or complexes. Here we focus on using the GONDOLA general free energy driven optimization strategy to elucidate the structural details of missing protein loops, which are often missing from experimental structures due to conformational heterogeneity and/or limitations in the resolution of the data. We first show that the correlation between experimental data and AMOEBA (i.e. a polarizable force field) structural minima is stronger than that for OPLS-AA (i.e. a fixed charge force field). This suggests that the higher order multipoles and polarization of the AMOEBA force field more accurately represented the true crystalline environment than the simpler OPLS-AA model. Thus, scoring and optimization of loops with AMOEBA is more accurate than with OPLS-AA, albeit at a slightly increased computational cost. Next, missing PDZ domain protein loops and protein loops from a loop decoy data set were optimized for 5 ns using the GONDOLA approach (i.e. under the AMOEBA polarizable force field) as well as a commonly used global optimization procedure (i.e. simulated annealing under the OPLS-AA fixed charge force field). The GONDOLA procedure was shown to provide more accurate structures in terms of both experimental metrics (i.e. lower Rfree values) and structural metrics (i.e. using the MolProbity structure validation tool). In terms of Rfree, only one out of seven simulated annealing results was better than the Gondola global optimization. Similarly, one simulated anneal loop had a better MolProbity score, but none of the simulated annealing loops were better in both categories. On average, GONDOLA achieved an Rfree value 19.48 and simulated annealing saw an average Rfree value of 19.63, and the average MolProbity scores were 1.56 for GONDOLA and 1.75 for simulated annealing. In addition to providing more accurate predictions, GONDOLA was shown to converge much faster than the simulated annealing protocol. Ten separate 5 ns optimizations of the 4 residue loop missing from one of the PDZ domains were conducted. Five were done using GONDOLA and five with the simulated annealing protocol. The fastest four converging results belonged to the GONDOLA approach. Thus, this work demonstrates that GONDOLA is well-suited to refine or predict the coordinates of missing residues and loops because it is both more accurate and converges more rapidly.

L’appréhension des mécanismes d’action biologiques des protéines membranaires nécessite de comprendre les interactions des xénobiotiques avec ces protéines et avec les membranes lipidiques. Les méthodes expérimentales sont parfois coûteuses et ne permettent d’obtenir que des informations partielles sur les interactions xénobiotiques-membrane-protéine. La modélisation moléculaire est une sérieuse alternative. Les simulations de dynamique moléculaire et de dynamique biaisées ont ouvert de nombreuses perspectives en permettant de décrire ces interactions moléculaires à l’échelle atomique. Grâce à des simulations de dynamique moléculaire, nous avons été capables de construire un modèle de transporteur humain ABC : ABCC4/MRP4. Cette protéine a été choisie pour sa présence dans le rein, notamment, et son importance clinique. Nous avons évalué l’influence du cholestérol sur cette protéine. L’étude de domaines spécifiques et l’impact d’un polymorphisme a été reliée à l’activité de transport de cette protéine. Nous avons également étudié l’interaction de xénobiotiques avec ce transporteur humain. Le cycle de transport des transporteurs ABC a été examiné afin de comprendre leur fonctionnement. L’incorporation de cholestérol a montré un impact significatif sur la protéine humaine ABCC4/MRP4 et sur les xénobiotiques étudiés. L’importance de domaines constituant la protéine ABCC4/MRP4 ainsi que l’importance de résidus individuels a clairement été prouvée. Nous avons également pu observer des intermédiaires du cycle de transport d’un transporteur ABC conjointement avec des changements structuraux. / Understanding the biological mechanisms of action of membrane proteins requires the comprehension of the interactions of xenobiotics with these proteins and with lipid membranes. Experimental methods are often demanding and only partially respond to xenobiotic-membrane-protein interactions. In silico molecular modeling is a serious alternative to tackle these issues. Molecular dynamics (MD) and biased dynamics simulations have opened many perspectives by providing an atomistic description of these intermolecular interactions. Using MD simulations, we built a model of the human ABC ABCC4/MRP4 transporter. We explored the influence of cholesterol on this protein as well as the impact of a polymorphism known to shut down the transport activity of this protein. We also studied the interaction of xenobiotics with this human transporter. The transport cycle of the ABC transporters was investigated in an attempt to better understand how it works.Interactions between lipid membranes and xenobiotics were explored by examining their ability to incorporate lipid membranes. Lipid mixtures with cholesterol showed a significant impact on the human protein ABCC4/MRP4 and on the xenobiotics studied. The importance of regions, domains constituting the ABCC4/MRP4 protein as well as the importance of specific residues has been clearly demonstrated. We also observed intermediates in the transport cycle of an ABC transporter in conjunction with structural changes occurring during this cycle.

The application of commercial herbicide restricts weed growth and significantly improves control over crop vitality and yield. Despite their utility in the agriculture sector, herbicides have the potential to contaminate local water sources. To minimize environmental impacts, the development of efficient separation processes to clean-up contaminated water bodies is necessary. However, complex speciation and conformational flexibility in the condensed phase poses a significant challenge. In this work, we investigate structure and speciation of three common organic herbicides (glyphosate, atrazine, and metolachlor) in aqueous solution. We employ the PBE-D3 density functional to perform ab initio molecular dynamics (MD) simulations in the canonical and isothermal-isobaric ensembles. We analyze MD trajectories to understand hydrogen bonding dynamics and lifetime as well as diffusional and vibrational characteristics. To enhance configurational sampling, we conduct metadynamics simulations to obtain the free energies of dissociation and intramolecular proton transfer of glyphosate.

La description des interactions protéine-ligand est d'une importance capitale pour la compréhension de phénomènes biologiques de toutes sortes, et dans le procédé de conception de nouvelles molécules bioactives. Avec l'avènement de nouveaux moyens de calculs numériques, la modélisation des équilibres d'interaction utilisant la dynamique moléculaire est de plus en plus utilisée pour l'étude au niveau microscopique permettant ainsi de guider la conception et l'optimisation de nouveaux inhibiteurs biologiques. Cette thèse est centrée sur l'utilisation de la funnel-métadynamique pour la prédiction d'affinités protéine-ligand et la description des interactions, en corrélation avec des données expérimentales de RMN et de cinétique enzymatique. Sur le modèle biologique des peroxyrédoxines humaines impliquées dans la cascade inflammatoire survenant après un choc ischémique, l'interaction et l'inhibition de composés catéchols ont été étudiées. Les données obtenues par dynamique sur l'interaction de dérivés catéchols ont été utilisées pour guider l'optimisation de cette molécule fragment pour une meilleure affinité et inhibition des peroxyrédoxines humaines. Nos résultats ont montré que la funnel-métadynamique, en plus de permettre la prédiction des affinités protéine-ligand, donnait une description réaliste de l'interaction pouvant mener à l'identification et l'optimisation de nouvelles molécules bioactives dont le potentiel inhibiteur a pu être examiné par cinétique enzymatique. Cette recherche fournit un aperçu des possibilités offertes par les nouvelles méthodes numériques, et leur application en chimie médicinale par exemple / Description of protein-ligand interactions is crucial for the understanding of all biological phenomena, and the drug design process. Thanks to new developments in computational devices, simulation of interaction equilibria using molecular dynamics are becoming state-of-the-art approach for the microscopic study of these molecular interactions. These new methods guide the conception and the optimization of new biological inhibitors. This project is focused on the use of funnel metadynamics for the prediction of protein-ligand affinities, and the description of interactions. Data obtained by metadynamics are correlated with experimental data from NMR and enzymatic assays.On the biological model of human peroxiredoxins, involved in the post ischemic inflammation cascade, interaction an inhibition of catechols derivatives have been studied. Data obtained from simulations have been used in the optimization process of the catechol fragment, to reach a better affinity and inhibition of human peroxiredoxins. Our results have shown that funnel-metadynamics allows the prediction of protein-ligand affinities, and the realistic description of the interaction, that lead to identify and optimize new bioactive molecules. Their inhibitory strengths and mechanisms have been examinated using enzymatic assays.Our research gives an overview of the possibilities brought by new numerical approaches, and their application in medicinal chemistry for example

L’abandon progressif des ressources fossiles s’accompagne de l’exploitation croissante de la biomasse. Cette transition nécessite de développer de nouveaux procédés notamment en catalyse hétérogène. Les chimistes se heurtent alors à deux défis majeurs : (i) désoxygéner la biomasse (cellulose/lignine) pour revenir à la chimie maîtrisée des grands intermédiaires (ii) rendre les catalyseurs résistants à l’eau, omniprésente en biomasse. En collaboration avec des expérimentateurs de l’Université d’Ottawa, nous nous sommes d’abord intéressés à la désoxygénation d’aromatiques de type lignine. Les calculs ab initio (DFT) nous ont permis de dresser les caractéristiques d’adsorption de ces composés sur Pt(111) en termes de descripteurs moléculaires simples. Nous avons ensuite étudié le mécanisme de décomposition de l’anisole et du 2-phénoxyéthanol, molécules modèles. Nos études ont montré l’importance de l’hydrogène et des fragments carbonés sur la réaction de désoxygéna6on de ces composés. En parallèle nous nous sommes intéressés à la stabilité, dans l’eau, d’un des supports catalytiques majeurs : l’alumine-γ. Ce sujet clé pose des défis considérables en modélisation, puisqu’il nécessite d’utiliser des méthodes de dynamiques moléculaires ab initio. Celles-ci nous ont permis de caractériser la structuration de l’eau au contact de l’alumine et l’importance de la solvatation sur les aluminols de surface. À l’aide de méthodes d’événements rares (dynamique contrainte, métadynamique) nous avons enfin abordé la réactivité d’alcools et de l’eau avec l’alumine hydratée. Ces simulations ont permis d’identifier les premières étapes d’hydratation et de mieux comprendre comment les limiter. / Moving away from fossil ressources is currently being accompanied by the increasing exploitation of biomass.This shift requires the development of new processes, in particular in heterogeneous catalysis. Chemists are nowfacing two major challenges: (i) deoxygenate biomass (cellulose/lignin) to produce platform intermediates with aeel-known chemistry (ii) make catalysts resistant to water, ubiquitous within the context of biomass.Within a collaboration with experimentalists at the University of Ottawa, we have first studied the deoxygenationof lignin-like aromatics. From an ab initio (DFT) inspection, we have characterized and described the adsorptionof such aromatic oxygenates on Pt(111) with simple molecular descriptors. We have then investigated thedecomposition mechanism of anisole and 2-phenoxyethanol. For these two model compounds, we have showedthe significance of hydrogen and carbonaceous species to have the deoxygenation reaction proceed properly.Meanwhile, we have examined the stability, in water, of γ-alumina, a major support in heterogeneous catalysis.The necessity to perform ab initio molecular dynamics simulations makes the modeling of such a systemparticularly challenging computationally. The simulations have nevertheless enabled us to characterize thestructuration of liquid water in contact with alumina and the significance of solvation on surface aluminol groups.Using rare-event methods (constrained dynamics, metadynamics) we have eventually been able to probe thereactivity of alcohols and water with hydrated alumina. We have then identified the first steps of hydration andgained insights on how to limit them.

Les structures cristallographiques de macromolécules comme les protéines, obtenues par la biologie structurale sont des modèles statiques. Or, c'est la flexibilité et la dynamique de ces macromolécules qui sont généralement responsables de leurs fonctions. La simulation permet d'explorer cette flexibilité lors de différents phénomènes qui ont lieu dans ces systèmes : une réaction chimique, des interactions avec une petite molécule… Simuler de tels phénomènes est un défi car la dynamique moléculaire classique ne permet pas de les observer. Des algorithmes permettent d'accélérer l'échantillonnage des dynamiques pour lever cette limitation et de calculer les barrières d'activation pour de tels phénomènes. Simultanément, le choix du niveau de calcul est crucial car il faut concilier la taille importante des systèmes, la nature des interactions et les phénomènes électroniques impliqués. Dans ce travail, différentes méthodes, dont principalement la métadynamique soit au niveau classique ou quantique, ou encore en combinant les deux niveaux quantique/classique, seront utilisées pour modéliser quatre processus complexes : des changements de conformations d'une protéine, des interactions entre métalloprotéine et inhibiteur, des réactions en solution et dans une enzyme. / Crystallographic structures of macromolecules, such as proteins, obtained by structural biology are static models. However, flexibility and dynamics of macromolecules are generally responsible for their functions. Modeling allows us to explore this flexibility in different phenomena that take place in these systems: a chemical reaction, interaction with a small molecule... Modeling such phenomena is a challenge because they cannot be observed by classical molecular dynamics. Algorithms can accelerate sampling of dynamics to simulate these events and calculate their activation barriers. Simultaneously, the choice of the level of calculation is crucial because it must merge with the size of the systems, the nature of interactions and the electronic phenomena involved.In this thesis, some methods, mainly metadynamics at classical level, quantum or the hybrid quantum/classical level, will be used to model four complex processes: conformational changes of proteins, metalloprotein/inhibitor interactions, reactivity in solution and enzymatic reactivity.