Orientadora: Profa. Dra. Káthia Maria Honório / Dissertação (mestrado) - Universidade Federal do ABC. Programa de Pós-Graduação em Ciência e Tecnologia/Química, 2015. / A proteina quinase TGFBRI/ALK5 esta envolvida em uma variedade de processos patologicos, tais como fibroses e cancer. Essa proteina propaga uma sinalizacao intracelular capaz de atingir o nucleo e modular a transcricao de genes. Neste trabalho, os metodos do holograma QSAR (HQSAR) e da analise comparativa de campos de interacao molecular (CoMFA), foram utilizados para uma serie de inibidores da proteina TGF¿ÒRI/ALK5. Os modelos obtidos apresentaram valores significativos de validacao interna (CoMFA, r2calibracao=0,99 e q2cv=0,85; HQSAR, r2calibracao=0,92 e q2cv=0,0,72) e externa (CoMFA, r2teste=0,85 e r2m=0,64, HQSAR, r2teste= 0,79 e r2m= 0,69), indicando capacidade preditiva dos modelos 2D e 3D para os compostos testados. Os modelos foram usados para analisar a capacidade preditiva de um conjunto de compostos-teste e os valores preditos a partir dos modelos de HQSAR e CoMFA apresentaram boa concordancia com os resultados experimentais. Os modelos finais juntamente com as informacoes obtidas a partir dos mapas 2D e 3D (estereoquimica e eletrostatica) para os compostos mais ativos e menos ativos da serie foram capazes de indicar caracteristicas importantes para a inibicao do alvo biologico em estudo. Com isso, os resultados obtidos neste trabalho podem guiar futuros projetos para o desenvolvimento de novos candidatos a farmacos para o tratamento de doencas como fibrose e cancer. / TGFBRI/ALK5 protein is a biological receptor involved in a variety of pathological processes such as cancer and fibrosis. TGF¿ÒRI/ALK5 receptor propagates an intracellular signaling that forms a protein complex capable of reaching the nucleus and modulating the gene transcription. In the present study, comparative molecular field analysis (CoMFA) and hologram quantitative structure-activity (HQSAR) studies were conducted on a series of potent TGF¿ÒRI/ALK5 inhibitors. Significant correlation indexes from the internal (CoMFA, r2calibration=0,99 e q2cv=0,85; HQSAR, r2calibration=0,92 e q2cv=0,0,72) and external (CoMFA, r2teste=0,85 e r2m=0,64, HQSAR, r2teste= 0,79 e r2m= 0,69) validations indicated the predictive power of the 2D and 3D models for untested compounds. The models were then used to investigate the predictive ability of a test-set, and the predicted values from the HQSAR and CoMFA models were in good agreement with the experimental results. The final QSAR models, along with the information obtained from 3D (steric and electrostatic) contour maps and 2D contribution maps, can be useful for the design of novel bioactive ligands.

Submitted by Caroline Falcao (caroline.rfalcao@ufpe.br) on 2017-04-04T19:45:56Z No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) 2012-Dissertação-MarceloRabello.pdf: 28475883 bytes, checksum: a9615b2d4a3b20d9263dba3af92c324c (MD5) / Made available in DSpace on 2017-04-04T19:45:56Z (GMT). No. of bitstreams: 2 license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) 2012-Dissertação-MarceloRabello.pdf: 28475883 bytes, checksum: a9615b2d4a3b20d9263dba3af92c324c (MD5) Previous issue date: 2012 / Este projeto pode ser resumido como uma comparação metodológica de abordagens in silico, mais precisamente de docking molecular. Foram utilizadas como alvos biológicos as fosfolipases A2 (PLA2), sendo estas um grupo de toxinas presentes abundantemente em venenos de serpentes. Foram utilizadas, como ligantes (potenciais inibidores), um total de 103 moléculas conhecidas na literatura. Um banco de dados com os ligantes e alvos foi construído, agregando-se informações sobre a atividade biológica e também suas respectivas estruturas tridimensionais. Os programas AUTODOCK, AUTODOCK VINA, GOLD, DOCK, SURFLEX e PLANTS foram utilizados para gerar os resultados de docking. O programa BINANA foi utilizado na análise das interações intermoleculares presentes nos complexos ligante-receptor, obtidos como resultados dos cálculos. Os dados gerados foram analisados com métodos multivariados, como por exemplo, a Análise de Componentes Principais (PCA) e a Análise Hierárquica de Clusters (HCA). Resíduos de aminoácidos, importantes para a estabilidade do complexo ligantereceptor, também foram identificados através de uma análise detalhada dos melhores resultados. Adicionalmente, foi reportado um artigo publicado, envolvendo PLA2s, com foco especial na molécula Quercetina como inibidor dos efeitos de veneno de serpente. Foi possível identificar os programas que apresentaram menor demanda computacional na análise comparativa de seus tempos de processamento, o que pode vir a ser muito útil em futuros estudos de docking, até mesmo com um número ainda maior de ligantes e alvos que podem ser submetidos ao procedimento de docking molecular, no intuito de aumentar o banco de dados de inibidores potenciais de PLA2s. As abordagens analíticas multivariadas utilizadas foram capazes de revelar aspectos interessantes sobre as semelhanças e diferenças entre os resultados de docking obtidos. Com a aplicação destas análises, pôde-se estabelecer uma metodologia de análise para a interpretação dos resultados nas escalas visual, numérica e gráfica, através, respectivamente, da geração das imagens para visualização molecular, das análises estatísticas (multivariadas) e da elaboração dos gráficos provenientes destas análises. Este estudo foi importante para aumentar a base de conhecimento do nosso grupo de pesquisa, e da comunidade científica como um todo, a respeito dos cálculos de modelagem molecular que envolvem os métodos de docking, e certamente serão úteis em estudos futuros com PLA2s ou outros alvos farmacológicos. / This project can be summarized as a methodological comparison of in silico approaches, more precisely between docking methods. Several phospholipase A2 (PLA2), a group of toxins abundantly present in snake venoms, were used as targets. A total number of 103 molecules were used for the definition of the ligands. It was built a database with all these ligands, adding information about their activity and also their three-dimensional structures. The programs AUTODOCK, AUTODOCK VINA, GOLD, DOCK, SURFLEX and PLANTS were used in docking calculations. The program BINANA was used to analyze the intermolecular interactions present in the ligand-receptor complexes obtained as poses from docking results. The data generated by BINANA was statistically analyzed by applying multivariate analysis methods, such as principal component analysis (PCA) and hierarchical cluster analysis (HCA). Important aminoacids residues for the stability of the complex ligandreceptor were indetified through a detailed analysis of the best results. Additionally, an article was reported, involving PLA2s, with special focus on the Quercetin molecule as an inhibitor of snake venom. In a comparative analysis of the docking calculation’s time, it was possible to identify the programs that present low computational demands. This performance analysis can be very useful in future studies of docking, even with a much larger number of ligands and targets, increasing the database of potential PLA2s inhibitors. By applying these tests, we could establish a methodological analysis for the results interpretation on visual, numerical and graphic scales, means, respectively by the generation of images for molecular visualization, statistical analysis (multivariates) and graphing from these tests. This study was important to increase the knowledge of our research group and also of the scientific community, about the molecular modeling calculations that involves docking methods, and will certainly be useful in future studies with PLA2 or other pharmacological targets.

A conhecida regra do acoplamento em W, que estabelece que núcleos na conformação em W planar exibem valores significativos de constantes de acoplamento através de quatro ligações, tem sido uma ferramenta útil para a determinação estrutural desde o início do uso da RMN para essa finalidade. Muitas configurações e conformações foram decididas com base nessa regra. A contínua evolução do equipamento de RMN, porém, resulta em modificações na qualidade e no número dos dados experimentais obtidos, obrigando os químicos a freqüentes revisões de seus pontos de vista sobre a importância relativa dos dados que podem ser obtidos dos espectros de RMN. O equipamento mais recente tem uma resolução maior e várias características adicionais que obscurecem um pouco os conceitos mais antigos: por um lado, o alto valor informativo de algumas técnicas modernas tais como NOE DIFF e outros métodos multi-dimensionais reduzem a importância das constantes de acoplamento; por outro lado, agora é possível determinar um número muito maior de constantes de acoplamento, devido principalmente à maior resolução. Uma conseqüência natural é que o químico agora pode explorar o uso de desdobramentos sutis em análise conformacional. Como parte de nosso trabalho de pesquisa na síntese de produtos naturais, preparamos há algum tempo um número apreciável de derivados de ciclopentanos. Nossa atenção foi fortemente atraída para certas constantes de acoplamento a longa distância (4JHH) que ocorriam em alguns destes compostos, já que uma conformação W planar não parece ser possível em ciclopentanos. Decidimos então iniciar um estudo mais detalhado dos espectros de RMN desses compostos, com vistas a uma interpretação mais clara dos dados. Inicialmente fizemos as atribuições para todos os hidrogênios, incluindo a estéreo-química de cada um, e medimos todos os valores de J para os compostos; para tanto fizemos extenso uso dos espectros de RMN tanto de 1H (300 MHz)como de 13C (75 MHz), medidas de efeito NOE, etc. Os ângulos entre as ligações e os ângulos diedros (de torção) foram calculados com programas de modelagem molecular; vários programas e métodos diferentes foram usados, para aumentar a confiabilidade. O primeiro resultado obtido é a confirmação de que 4JHH 'diferente' 0 pode ocorrer mesmo em casos em que uma conformação W planar não é possível. Mais importante, porém, é a conclusão de que há uma relação entre os valores de 4JHH e os ângulos diedros envolvidos. O acoplamento entre H1 e H2 ocorre através das quatro ligações 'sigma' definidas pelo caminho H1-C1-C2-C3-H2 e envolve dois ângulos diedros 'teta'1 e 'teta'2. Os valores de 4JHH foram plotados contra (cos2'teta'1 x cos2'teta'2) (uma simples extensão da equação de Karplus); os pontos resultantes não se alinham com perfeição sobre uma curva contínua, mas mostram clara tendência de aumento do valor de 4JHH conforme os ângulos 'teta'1 e 'teta'2 se afastam de 90º e se aproximam de 180º. / The well known W rule, which establishes that nuclei in a planar W arrangement exhibit significant four bond coupling constants, has been a useful tool in molecular structure determination since early times of the use of nmr spectra for this purpose. Many configurations and conformations have been decided based on this rule. The continuous evolution of the nmr equipment, however, produces modifications in quality and number of available experimental data, thus forcing the chemists to frequent revisions of their points of view about the relative importance of the data that can be obtained from nmr spectra. The more recent equipment has a higher resolution and several additional features that throw some shadow over early concepts: on one hand, the high power of modern techniques such as NOE DIFF and other multi-dimensional methods reduce the importance of coupling constants; on the other hand, it is now possible to determine many more coupling constants, due mainly to the higher resolution. A natural consequence is that the chemist can now exploit the use of subtle splittings in conformational analysis. As part of our research work on the synthesis of natural products, we have prepared a number of cyclopentane derivatives. Our attention was strongly attracted to the long-range (4JHH) coupling constants that occurred in some of these compounds, as no planar W conformation seems to be possible in cyclopentanes. We have thus decided to start a more detailed study of the nmr spectra of these compounds, seeking for more clear interpretation of the data. We have first assigned all hydrogens, including the stereochemistry of each hydrogen, and measured all J values for the compounds; for this task we have used nmr spectra both of 1H (300 MHz) and 13C (75 MHz), NOE measurements, etc. Bond angles and dihedral (torsion) angles were calculated with molecular modeling programs; several different programs and methods were used to improve reliability. The first result obtained is a confirmation that a 4JHH 'different' 0 occurs even in cases where a planar W conformation is not possible. More important, however, is the conclusion that there is a correlation between the 4JHH values and the involved dihedral angles. There are two dihedral angles in the path through the bonds between two hydrogens which show 4JHH coupling. When 4JHH values are plotted against(cos2'teta'1 x cos2'teta'2) (a simple extension of Karplus equation) the points are not aligned over a continuous curve, but they show a clear tendency: 4JHH values become higher as the angles 'teta'1 and 'teta'2 vary from 90 to 180º.

Adsorption in porous materials is a promising technology for CO2 capture and storage. Particularly important applications are adsorption separation of streams associated with the fossil fuel power plants operation, as well as natural gas sweetening. High surface area activated carbons are a promising family of materials for these applications, especially in the high pressure regimes. As the streams under consideration are generally multi-component mixtures, development and optimization of adsorption processes for their separation would substantially benefit from predictive simulation models. In this project we combine experimental data and molecular simulations to systematically develop a model for a high surface area carbon material, taking activated carbon Maxsorb MSC-30 as a reference. Our study starts from the application of the well-established slit pore model, and then evolves through the development of a more realistic model, based on a random packing of small graphitic fragments. In the construction of the model, we introduce a number of constraints, such as the value of the accessible surface area, concentration of the surface groups and pore volume, to bring the properties of the model structure close to the reference porous material. Once a plausible model is developed, its properties are further tuned through comparison between simulated and experimental results for carbon dioxide and methane. The model is then validated by predictions for the same species at different conditions and by prediction of other species involved in the carbon capture processes. The model is applied to simulate the separations involved in pre and post combustion capture processes and sweetening of sour natural gas, using realistic conditions and compositions for the multicomponent mixtures. Finally, it is used to explore the effect of water in pre and post combustion separations.

The highly competitive market in the oil refining industry forces refiners look for more detailed information of both feedstocks and products to achieve the optimal economic performance. Due to stricter environmental legislations, the molecular level characterisation has been investigated by various researchers and shows promising advantages in modern refinery design and operation. Although various molecular characterisation methods have been developed, there is an unavoidable trade-off between keeping astronomical molecule details and practicality in industrial applications. In the meantime, many of these methodologies have different characteristics and different focuses according to a particular application purpose. Our aim is hence to tackle the problems of developing manageable and practical technical solutions for molecular characterisation of petroleum fractions for vary refinery processes. A pseudo-component based approach is developed within a modified MTHS (Molecular Type Homologous Series) matrix framework (Peng, 1999) to represent the molecular information of a particular refining stream. This proposed methodology incorporates both molecular type and pseudo-component information by the conjunction of homologous series and boiling points in the matrix framework. To increase the usability of this method, a 3-parameter gamma distribution function is introduced to describe the composition of each structural molecular type. Typical PIONA (paraffin, iso-paraffin, olefin, naphthene, aromatic) analysis, ratios between each homologous types and the percentage of particular carbon type are considered as well as the distillation curve and the density of a stream. More strict product specifications and environmental legislations make strong restriction to the benzene and aromatics content in gasoline products, which motivate refiners to understand, characterise and simulate gasoline catalytic reforming on molecular-level. In this work, kinetic and reactor model of naphtha catalytic reforming is developed based on the proposed MTHS method. The naphtha feedstock composition is represented by the MTHS matrix, and a kinetic network is constructed according to conversions among matrix elements. A process model proposed by Wu (2010) is employed for reforming modelling. The proposed model is then applied to a bench-scale semi-regenerative catalytic reforming unit, which contains 3 fixed-bed reactors, for validation. The influences of essential operating conditions, such as reactor inlet temperature, pressure and weight hourly space velocity (WHSV), on the product distribution and quality are explored. The developed characterisation is also applied in gasoline blending modelling. A molecular-level nonlinear gasoline blending model is developed based on proposed MTHS method with validation. Key properties such as Octane Numbers (ONs) and RVP are blended by molecular matrix elements, and the influence of molecular composition on bulk properties is obvious. A case of recipe optimisation is studied to show the applicability of the proposed method. The implementation of the developed MTHS method for catalytic reforming and gasoline blending demonstrates the compatibility when characterising different petroleum streams, and provides a common platform to simulate and optimise refining operations on the same molecular basis.

Accumulation and aggregation of the 42-residue amyloid-[beta] (A[beta]) protein fragment, which originates from the cleavage of amyloid precursor protein by beta and gamma secretase, correlates with the pathology of Alzheimer's disease (AD). Possible therapies for AD include peptides based on the A[beta] sequence, and recently identified small molecular weight compounds designed to mimic these, that interfere with the aggregation of A[beta] and prevent its toxic effects on neuronal cells in culture. Here, we use molecular dynamics simulations to compare the mode of interaction of an active (LPFFD) and inactive (LHFFD) [beta]-sheet breaker peptide with an A[beta] fibril structure from solid state NMR studies. We found that LHFFD had a weaker interaction with the fibril than the active peptide, LPFFD, from geometric and energetic considerations, as estimated by the MM/PBSA approach. Cluster analysis and computational alanine scanning identified important ligand-fibril contacts, including a possible difference in the effect of histidine on ligand-fibril [pi]-stacking interactions, and the role of the proline residue establishing contacts that compete with those essential for maintenance of the inter-monomer [beta]-sheet structure of the fibril. Our results show that molecular dynamics simulations can be a useful way to classify the stability of docking sites. These mechanistic insights into the ability of LPFFD to reverse aggregation of toxic A[beta] will guide the redesign of lead compounds, and aid in developing realistic therapies for AD and other diseases of protein aggregation. We have also performed long explicit solvent MD simulations of unliganded amyloid fibril in three putative protonation states, in order to better understand the energetic and mechanical features of the fibril receptor. Over 100 ns MD simulations, the trajectories where fibril has Glu11 and Glu22 side-chains protonated exhibit the least deviation from the initial solid state NMR structures. Free energy calculations on these rajectories suggest that the weakest fibril interface lies in the lateral rather than transverse direction and that there is little dependence on whether the lateral interface is situated at the edge or middle of the fibril. This agrees with recent reported steered molecular dynamics calculations. Secondly, in an effort to improve the ability of atomistic simulation techniques to directly resolve protein tertiary structure from primary amino acid sequence, we explore the use of a molecular dynamics technique based on swarm intelligence, called SWARM-MD, to identify the native states of two peptides, polyalanine and AEK17, as well as Trp-cage miniprotein. We find that the presence of cooperative swarm interactions significantly enhanced the efficiency of molecular dynamics simulations in predicting native conformation. However, it also is evident that the presence of outlying simulation replicas can adversely impact correctly folded replica structures. By slowly removing the swarm potential after folding simulations, the negative effect of the swarm potential can be alleviated and better agreement with experiment obtained.

Domínio proteico é uma sequência de aminoácidos evolutivamente conservada e funcionalmente independente. Um dos aspectos mais importantes do estudo de uma proteína que contem múltiplos domínios é o entendimento da comunicação, entre os diferentes domínios, e seu papel biológico. Essa comunicação em maior parte é feita pela interação direta entre domínios. A interação poderia ser tratada como uma clássica interação proteína-proteína. Entretanto, proteínas multidomínio possuem restrições determinadas por suas regiões conectoras. Os conectores interdomínio impõem restrições e limitam espaço conformacional dos domínios. Apresentamos aqui o MAD, uma rotina capaz de obter modelos tridimensionais de alta resolução para proteínas, contendo qualquer número de domínios, a partir de sua sequencia primária. Os domínios conservados são identificados utilizando a base de domínios conservados (CDD) e seus limites são utilizados para definir as regiões conectoras. É criado um ensamble de possíveis dobramentos dos conectores e sua distribuição de distâncias C/N-terminais são utilizadas como restrição espacial na busca pela interação entre os domínios.Os modelos dos domínios são obtidos por uma modelagem comparativa. Foi implementada uma heurística, capaz de lidar com a natureza combinatorial dos múltiplos domínios e com a necessidade imposta pela limitação computacional de realizar o docking dos domínios em forma de pares. Todas combinações de domínios são submetidas as rotinas de docking. Aplica-se filtro de distância e energético, excluindo as conformações que apresentam distância C/N-terminal entre domínios maior do que o valor máximo observado no ensamble de conectores e seleciona as conformações energeticamente mais favoráveis. As conformações são submetidas a uma rotina de agrupamento hierárquico baseada em sua similaridade estrutural. Para a segunda fase as conformações selecionadas são pareadas com seu domínio complementar e ressubmetidas a rotina de docking até que todas as fases tenham sido completadas. Foi criado um conjunto de testes a partir do Protein Data Bank contendo 54 proteínas multidomínio para que a rotina de docking do MAD fosse comparada com outros softwares utilizados pela comunidade cientifica, mostrou-se superior ou equivalente aos métodos testados. A capacidade de utilizar dados experimentais foi demostrada através da proposição de um modelo da forma ativa da enzima tirosina fosfatase 2, nunca observado experimentalmente. A rotina de docking foi expandida paralelamente em uma aplicação standalone e utilizada na resolução de diversos problemas biológicos. Concluímos que a inovação metodológica proposta pelo MAD é de grande valia para a modelagem molecular e tem potencial de gerar uma nova perspectiva a respeito da interação de proteína multidomínio, visto que é possível analisar essas proteínas em sua plenitude e não como domínios separados. / Protein domain is an evolutionary conserved and functionally independent amino acid sequence. One of the most important aspects of the study of a protein that contains multiple domains is the understanding of communication between the different areas, and their biological role. This communication is made mostly by direct interaction between domains. The interaction could be treated as a classical protein-protein interaction. However, multidomain proteins have certain restrictions for its connector regions. The intra connectors impose restrictions and limit conformational space of the domains. We present the MAD, a routine able to get three-dimensional models of high-resolution protein, containing any number of domains, from its primary sequence. The conserved domains are identified using the basic conserved domains database (CDD) and its boundaries are used to define the connector regions. This creates a ensemble of possible folding of the connectors and distribution of distances C/N-terminals are used as spatial restriction in the search for interaction between domains.Os models of the domains are obtained by comparative modelling. A heuristic able to handle the combinatorial nature of the multiple areas and the need imposed by the computer to perform the limitation of the docking areas as pairs was implemented. All combinations of domains are referred to the docking routines. Distance and energy filters are applied, excluding conformations that have C/N-terminal domains distances larger than the maximum value observed in the connectors ensemble and selects the most favourable energy conformations. Conformations are subjected to hierarchical clustering routine based on their structural similarity. For the second phase, the selected conformations are paired with its complementary domain and resubmitted to the docking routine until all phases have been completed. A test set has been created from the Protein Data Bank containing 54 multidomain proteins so that the docking routine of MAD could be compared with other software used by the scientific community, it has been shown to be superior or equivalent to the tested methods. The ability to use experimental data was demonstrated by proposing a model of the active form of tyrosine phosphatase enzyme 2, never observed experimentally. The docking routine was expanded in a standalone application and used in solving various biological problems. We conclude that the methodological innovation proposed by the MAD is very useful for molecular modelling and has the potential to generate a new perspective on multidomain protein interaction as you can analyse these proteins in its entirety and not as separate domains.

Papilomavírus (PVs) são vírus amplamente estudados, sendo enfatizada sua capacidade de infectar os tecidos epitelial e mucoso em diversos animais, incluindo humanos, causando lesões benignas que podem, ocasionalmente, resultar em câncer. Dentre os gêneros que infectam animais, os Deltapapillomavirus têm uma importância veterinária, ecológica e histórica, pois são capazes de infectar seu hospedeiro natural e outros animais. Por isso, esse trabalho contempla todos os tipos virais pertencentes aos Delta-PVs, incluindo sua história. Dentre as proteínas traduzidas pelos PVs, três são consideradas proteínas oncogênicas: E5, E6 e E7. Determinar a estrutura de uma proteína é crucial para a elucidação da sua função, possibilitando aplicações nas áreas de engenharia de proteínas, anotação genômica e desenho racional de fármacos. A estrutura tridimensional da proteína E6 de cada tipo viral pertencente ao gênero Deltapapillomavirus foi determinada por modelagem molecular por homologia. A história evolutiva dessas proteínas foi avaliada com base na geração de árvores filogenéticas e suas propriedades físico-químicas foram analisadas. Além disso, devido ao seu alto grau de conservação, a E6 demonstrou ser útil como um marcador molecular. Apesar de serem consideradas raras, foram observadas lesões papilomatosas em carneiros em uma fazenda do estado de São Paulo. Foi realizado o diagnóstico molecular dessas lesões. Os resultados mostraram pela primeira vez no mundo que, apesar de serem ovinos, o agente causador da papilomatose era um papilomavírus bovino, o BPV2, um Delta-PV. Logo, além de discutir os Delta-PVs, esta tese demonstra na prática a habilidade desse gênero em romper a barreira espécie-específica. / Papillomaviruses (PVs) are widely studied viruses, emphasizing their ability to infect the epithelial and mucosal tissues in several animals, including humans, causing benign lesions that may occasionally result in cancer. Among the genera that infect animals, Deltapapillomaviruses have a veterinary, ecological and historical importance because they are capable of infecting their natural host and other animals. Therefore, this work contemplates all viral types belonging to the Delta-PVs, including their history. Among the proteins translated by the PVs, three of them are considered oncogenic proteins: E5, E6 and E7. Determining the structure of a protein is crucial to the elucidation of its function, allowing applications in the areas of protein engineering, genomic annotation and rational design of drugs. The three-dimensional structure of the E6 protein of each viral type belonging to the genus Deltapapillomavirus was determined by molecular modeling by homology. The evolutionary history of these proteins was evaluated based on the generation of phylogenetic trees and their physicochemical properties were analyzed. In addition, due to its high degree of conservation, E6 has been shown to be useful as a molecular marker. Despite being considered rare, papillomas lesions were observed in sheep on a farm in the state of São Paulo. The molecular diagnosis of these lesions was performed. The results showed for the first time in the world that, despite being ovines, the causative agent of papillomatosis was a bovine papillomavirus, BPV2, a Delta-PV. Thus, in addition to discussing Delta-PVs, this thesis demonstrates in practice the ability of this genre to break the species-specific barrier.

Molecular modelling strategies for evaluation of enzyme selectivity wereinvestigated with a focus on principles of how molecular interactionscould be evaluated to provide information about selectivity. Althoughmolecular modelling provides tools for evaluation of geometrical andenergy features of molecular systems, no general strategies for evaluationof enzyme selectivity exist. Geometrical analyses can be based uponinspection and reasoning about molecular interactions, which provide aneasily accessible way to gain information, but suffer from the risk of biasput in by the modeller. They can also be based on geometrical features ofmolecular interactions such as bond lengths and hydrogen-bond formation.Energy analyses are appealing for their modeller independenceand for the possibility to predict not only stereopreference, but also itsmagnitude.In this thesis, four examples of enantio- or regioselective serinehydrolase-catalysed reaction systems are presented together with developedmodelling protocols for explanation, prediction or enhancement ofselectivity. Geometrical as well as energy-based methodology were used,and provided an understanding of the structural basis of enzymeselectivity. In total, the protocols were successful in making qualitative explanationsand predictions of stereoselectivity, although quantitative determinationswere not achieved.

The Diels-Alder is an important reaction that is one of the primary tools for synthesizing cyclic carbon structures, while simultaneously introducing up to four stereocenters in the resulting product. Not only is it a widely explored reaction in organic chemistry, but a vital tool in industry to construct novel compounds for pharmacological applications. Still, a remaining concern is the fact that upon the introduction of stereogenic carbons, the possibility of stereoselective control is greatly diminished. A common solution to the problem of undesirable stereoisomers is to employ chiral auxiliaries and ligands as means to increase the yield of a certain stereoisomer. However, incorporating these types of compounds in order to obtain an enantiomerically pure product increases the amount of synthetic steps to be regulated, implying that one or more purification steps are necessary to obtain the desired result. An accompanying thought leans toward the environmental aspect, as the principles of green chemistry are of great importance. This thesis presents the attempts to explore the possibility of engineering an enzyme that can catalyze an asymmetric Diels-Alder reaction through the use of molecular modeling. Based on previous work, the catalytically proficient enzyme ketosteroid isomerase had been deemed a probable candidate as a Diels-Alderase. To evaluate the enzyme thoroughly, a set of compounds was scored against the active binding site where the best hits against the wild type were saved and evaluated repeatedly after the introduction of rational mutations. Although no conclusive indication of an optimal design could be obtained at the end of this work, valuable insight was retrieved on plausible design strategies, which eventually could help lead to the first catalytically proficient Diels-Alderase. / Diels-Alder är en viktig reaktion då den är ett redskap för att syntetisera cykliska kolstrukturer, samtidigt som uppemot fyra stereocentra introduceras i den resulterande produkten. Reaktionen används inte enbart inom organisk kemi, utan är även ett viktigt redskap inom industriella sammanhang för att ta fram nya preparat som direkt kan tillämpas inom farmakologi. En återstående problematik är faktumet att introduktionen av nya stereogena kol bidrar till att drastiskt minska möjligheten att bibehålla en stereoselektiv kontroll. En vanlig lösning för att undvika oönskade stereoisomerer är att nyttja kirala hjälpmolekyler och ligander för att öka utbytet av en specifik stereoisomer. Dock innebär införandet av dessa hjälpmolekyler i strävan att erhålla en enantiomeriskt ren produkt ett ökat antal syntes-steg att hantera, vilket antyder att ett eller flera reningssteg är nödvändiga för att uppnå önskat resultat. Ur en miljösynpunkt är detta värt att ha i åtanke, då principerna för grön kemi är viktiga. Detta arbete utforskar möjligheterna att konstruera ett enzym som kan katalysera en asymmetrisk Diels-Alder-reaktion, med hjälp av molekylär modellering. Baserat på tidigare arbeten har enzymet ketosteroid isomeras valts ut som en potential kandidat till ett Diels-Alderase. För att noggrant evaluera enzymet så screenades ett set av substrat mot dess aktiva säte, där de bästa träffarna gentemot vildtypen sparades och återevaluerades allteftersom rationella mutationer kontinuerligt introducerades. Trots avsaknaden av klara indikationer på att en optimal design har kunnat tas fram vid slutet av detta arbete, så erhölls värdefull insikt på möjliga design-strategier, vilket skulle kunna bistå sökandet av det första katalytiskt effektiva Diels-Alderase.