Antibodies are soluble proteins produced by the adaptive immune system to bind and counteract invading pathogens. The binding properties of a typical human antibody are determined by the structure of its variable domain, composed of two chains – heavy and light and by the conformation of six loops located on the surface of the variable domain, known as Complementarity Determining Regions (CDRs). In the first chapter, we describe our analysis of the conformational space occupied by five out of six antibody CDRs (L1, L2, L3, H1 and H2) and the development of a novel, length-independent method for grouping these CDRs into structural clusters (canonical forms). We show that using our method we can increase coverage and precision of assigning CDR sequences into clusters. In the next chapter, we describe a method for ranking structural decoys of the CDR-H3 loop. We show that by computationally perturbing CDR-H3 decoys we can improve the performance of existing ranking methods. In the same chapter, we discuss the development of a method for high-throughput assignment of heavy-light chain orientation. The power of the method was demonstrated by assigning orientation to billions of potential Fv sequences. The third Chapter describes the analysis of a large dataset of CDR sequences with the aim of identifying sequence patterns responsible for the loops' structure. Using a neural network methodology, we found several groups of CDR sequences which might be indicative of previously-unseen conformations. In the final results Chapter, we describe how we used the structural knowledge developed throughout the rest of the thesis to create a novel pipeline for computational antibody design. We show that the binders developed using our methodology had similar features to available antibody therapeutics and low predicted propensity to cause an immunogenic response. These results demonstrate the potential for using computational methods for designing high affinity therapeutics with human properties.

O problema de predição de estruturas tridimensionais de proteínas pode ser visto computacionalmente como um problema de otimização, tal que dada a sequência de aminoácidos, deve-se encontrar a estrutura tridimensional da proteína dentre as muitas possíveis através da obtenção de mínimos de funções de energia. Vários pesquisadores têm proposto estratégias de Computação Evolutiva para a determinação de estruturas tridimensionais das proteínas, entretanto nem sempre resultados animadores têm sido alcançados visto que entre outros fatores, há um grande número de ótimos locais no espaço de busca. Geralmente as funções de fitness empregadas pelos algoritmos de otimização são baseadas em campos de força com diferentes termos de energia, sendo que os parâmetros destes termos são ajustados a priori e são mantidos estáticos ao longo do processo de otimização. Alguns pesquisadores sugerem que o uso de funções de fitness dinâmicas, ou seja, que mudam durante um processo de otimização evolutivo, pode aumentar a capacidade das populações fugirem de ótimos locais em problemas altamente multimodais. Neste trabalho, propõe-se que os parâmetros dos termos do campo de força utilizado sejam modificados durante o processo de otimização realizado por Algoritmos Genéticos (AGs) no problema de predição de estruturas de proteínas, sendo aumentados ou diminuídos, por exemplo, de acordo com a sua influência na formação de estruturas secundárias e no seu ajuste fino. Como a função de avaliação será modificada durante o processo de otimização, a predição de estruturas tridimensionais de proteínas torna-se um problema de otimização dinâmica, sendo que o uso de Algoritmos Genéticos específicos para tais problemas, como o AG com hipermutação e os AGs com imigrantes aleatórios são investigados aqui. É proposta uma nova métrica relacionada ao alinhamento da estrutura secundária da proteína, para auxiliar a análise dos dados obtidos e os resultados dos experimentos indicam que os algoritmos com função de avaliação dinâmica obtiveram resultados melhores que os algoritmos estáticos, o que é explicado pelo fato de as mudanças na função de fitness possibilitarem eventuais fugas de ótimos locais, bem como um aumento da diversidade da população. / The protein structure prediction can be seen as an optimization problem where given an amino acid sequence, the tertiary protein structure must be found amongst many possible by obtaining energy functions minima. Many researchers have been proposing Evolutionary Computation strategies to find tridimensional structures of proteins; however results are not always satisfactory since among other factors, there are always a great number of local optima in the search space. Usually, the fitness functions used by optimization algorithms are based on force fields with different energy terms with parameters from those terms being adjusted a priori, kept static through the optimization process. Some researchers suggest that the use of dynamic functions, i.e., that can be changed during the evolutionary process, can help the population to escape from local optima in highly multimodal problems. In this work we propose that the force field parameters can be changed during the optimization process of Genetic Algorithms (GAs) in the protein structure prediction problem, being increased or decreased, for instance, according with its influence on formation of secondary structures and its fine tuning. Since the cost function will be changed during the optimization process, the protein tridimensional structure prediction becomes a dynamic optimization problem and specific Genetic Algorithms for this kind of problem, like the hypermutation GA and random immigrants GA are investigated. We also propose a new metric related to the proteins secondary structure alignment to help the analysis of obtained data. Results indicate that the dynamic function algorithms obtained better results than static algorithms since changes on the fitness function allow the population to escape local optima, as well as an increase on the population diversity.

3D proteomics combines chemical cross-linking with mass spectrometry to study the structure of protein assemblies and protein-protein interactions both in vitro and in vivo by providing distance constraints that indicate which residues are in close spatial proximity. I addressed the main bottleneck of this technology: the reliable identification of cross-linked peptides. Reporter ion signatures for cross-linked peptides were developed, by fragmenting model compounds containing two lysine residues joined by a cross-linker backbone or a lysine residue modified with a hydrolysed cross-linker. The reporter ion signatures showed 97% specificity at 90% sensitivity and segregated cross-linked from modified and linear peptides. They decreased the false discovery rate of the identification of cross-linked peptides from 5% to 1% in a large dataset. The signatures permit data sorting during and after mass spectrometry acquisition. The advanced 3D proteomics workflow was applied to study the protein-protein interactions in Mycoplasma pneumoniae cells. In lysates of the bacterium we identified 128 protein-protein interactions (of which 24 are novel) and obtained in vivo topological data on 208 proteins, even for cases where high-resolution structures are not yet available. We showed that our data are in excellent agreement with crystal structures of proteins and complexes where available. We defined a network of ribosomal and RNA polymerase proteins that reveals an intricate link between transcription and translation in bacteria. We demonstrated that the method is suitable for identification of homomultimeric protein complexes by exploiting peptide pairs of identical amino acid sequence. The technology has the potential to provide a complete protein interaction network map after the selective enrichment of cross-lined peptides is achieved. The method was next applied to investigate the structure of condensin and cohesin complexes, which play a crucial role in stabilization of chromosome structure during mitosis. The complexes were purified, cross-linked and their linkage map created. The condensin coiled coil cross-linked on the entire length was modeled. The information was used to direct the analysis of in situ cross-linked condensin in intact chromosomes. I found two high confidence linkages between SMC2 and SMC4 coiled coils and identified H2A as a potential condensin receptor on chromosomes.

O problema de predição de estruturas tridimensionais de proteínas pode ser visto computacionalmente como um problema de otimização, tal que dada a sequência de aminoácidos, deve-se encontrar a estrutura tridimensional da proteína dentre as muitas possíveis através da obtenção de mínimos de funções de energia. Vários pesquisadores têm proposto estratégias de Computação Evolutiva para a determinação de estruturas tridimensionais das proteínas, entretanto nem sempre resultados animadores têm sido alcançados visto que entre outros fatores, há um grande número de ótimos locais no espaço de busca. Geralmente as funções de fitness empregadas pelos algoritmos de otimização são baseadas em campos de força com diferentes termos de energia, sendo que os parâmetros destes termos são ajustados a priori e são mantidos estáticos ao longo do processo de otimização. Alguns pesquisadores sugerem que o uso de funções de fitness dinâmicas, ou seja, que mudam durante um processo de otimização evolutivo, pode aumentar a capacidade das populações fugirem de ótimos locais em problemas altamente multimodais. Neste trabalho, propõe-se que os parâmetros dos termos do campo de força utilizado sejam modificados durante o processo de otimização realizado por Algoritmos Genéticos (AGs) no problema de predição de estruturas de proteínas, sendo aumentados ou diminuídos, por exemplo, de acordo com a sua influência na formação de estruturas secundárias e no seu ajuste fino. Como a função de avaliação será modificada durante o processo de otimização, a predição de estruturas tridimensionais de proteínas torna-se um problema de otimização dinâmica, sendo que o uso de Algoritmos Genéticos específicos para tais problemas, como o AG com hipermutação e os AGs com imigrantes aleatórios são investigados aqui. É proposta uma nova métrica relacionada ao alinhamento da estrutura secundária da proteína, para auxiliar a análise dos dados obtidos e os resultados dos experimentos indicam que os algoritmos com função de avaliação dinâmica obtiveram resultados melhores que os algoritmos estáticos, o que é explicado pelo fato de as mudanças na função de fitness possibilitarem eventuais fugas de ótimos locais, bem como um aumento da diversidade da população. / The protein structure prediction can be seen as an optimization problem where given an amino acid sequence, the tertiary protein structure must be found amongst many possible by obtaining energy functions minima. Many researchers have been proposing Evolutionary Computation strategies to find tridimensional structures of proteins; however results are not always satisfactory since among other factors, there are always a great number of local optima in the search space. Usually, the fitness functions used by optimization algorithms are based on force fields with different energy terms with parameters from those terms being adjusted a priori, kept static through the optimization process. Some researchers suggest that the use of dynamic functions, i.e., that can be changed during the evolutionary process, can help the population to escape from local optima in highly multimodal problems. In this work we propose that the force field parameters can be changed during the optimization process of Genetic Algorithms (GAs) in the protein structure prediction problem, being increased or decreased, for instance, according with its influence on formation of secondary structures and its fine tuning. Since the cost function will be changed during the optimization process, the protein tridimensional structure prediction becomes a dynamic optimization problem and specific Genetic Algorithms for this kind of problem, like the hypermutation GA and random immigrants GA are investigated. We also propose a new metric related to the proteins secondary structure alignment to help the analysis of obtained data. Results indicate that the dynamic function algorithms obtained better results than static algorithms since changes on the fitness function allow the population to escape local optima, as well as an increase on the population diversity.

Afin d’évaluer la qualité alimentaire et l’efficacité métabolique des aliments mixtes combinant différentes sources protéiques végétales ou des sources protéiques végétales/animales, deux aliments de base, les pâtes alimentaires et les gels laitiers, ont été choisis comme vecteurs et ont été enrichis par des farines ou des protéines de légumineuses. La structure de la fraction protéique des aliments mixtes a été étudiée à l’échelle moléculaire. La relation entre cette structure et la digestibilité in vitro et in vivo des protéines a été évaluée. L’effet de la formulation et/ou du procédé de fabrication de ces aliments mixtes sur le métabolisme protéique in vivo a été étudié chez des rats jeunes en croissance et des rats âgés. Le changement de la formulation des pâtes alimentaires, c'est à dire l’incorporation de trois farines de légumineuses différentes (féverole, lentille ou pois cassé), génère des modifications de structure du réseau protéique influençant la digestibilité des protéines. Les études animales montrent que la qualité alimentaire des pâtes enrichies en légumineuses est comparable à celle d’une protéine animale comme la caséine et ce, quel que soit le type de légumineuses utilisé. La rétention protéique corporelle et la synthèse protéique musculaire des rats âgés, consommant des régimes iso- protéiques à base de pâtes alimentaires enrichies en légumineuses ou de caséine, sont comparables. Elles restent cependant inférieures à celles induites par les protéines solubles du lait. L’utilisation de gels laitiers enrichis en protéines de féverole chez le rat a révélé un effet de la formulation et du procédé de gélification sur la digestion et la rétention protéiques. La digestibilité in vivo des protéines est plus élevée chez les rats consommant le régime contenant le gel fermenté mixte composé de protéines de caséine et de féverole comparativement à son homologue de même composition mais acidifié par voie chimique. La rétention protéique est encore améliorée chez les rats ayant consommé le régime contenant le gel fermenté composé de protéines de caséine, de féverole et de lactosérum. Ces aliments enrichis en légumineuses, riches en protéines, équilibrés en acides aminés indispensables commencent à être disponibles sur le marché. Ils pourraient être proposés à la population âgée notamment dans des situations physiopathologiques impliquant une perte de protéines corporelles. / In order to assess food quality and metabolic efficiency of mixed foods combining different vegetable protein sources or vegetable-animal protein sources, two staple foods, pasta and dairy gels, were selected as vectors and enriched with flour or protein extracted from legumes. The protein structure of the mixed feeds was studied at a molecular level. The relationship between this structure and the in vitro and in vivo digestibility of proteins was evaluated. The effect of the formulation and / or manufacturing process of these mixed foods on protein metabolism in vivo has been studied in growing young rats and aged rats. The change of the pasta formulation, ie the incorporation of three different leguminous flours (faba beans, lentils or split peas), generates structural modifications of the protein network influencing the digestibility of the proteins. Animal studies show that the nutritional quality of pasta enriched with legumes is comparable to that of an animal protein such as casein, regardless of the type of legume used. Body protein retention and muscle protein synthesis in aged rats, consuming iso-protein diets based on legume-enriched pasta or casein, are comparable. However, they remain lower than those induced by soluble milk proteins. The use of faba bean protein-enriched milk gels has shown an effect of the formulation and gelation process on protein digestion and retention. In vivo digestibility of proteins is higher in rats consuming the mixed fermented gel diet composed of casein and faba bean proteins compared to its counterpart of the same composition but chemically acidified. Protein retention is further improved in rats fed with a diet containing fermented gel composed of casein proteins, faba beans and whey. These legume-enriched foods, rich in protein, balanced in essential amino acids are beginning to be available on the market. They could be offered to the elderly population, especially in physiopathological situations involving a loss of body proteins.

<p>Experimentally derived biomolecular structures were determined by Nuclear Magnetic Resonance (NMR). The properties of selected peptides and proteins in solution and in membrane mimicking micelles were observed by circular Dichroism (CD), mass spectrometry (MS), and other spectroscopic techniques.</p><p>The mDpl(1-30) peptide (30 residues) of the mouse Doppel protein was found to be positioned as an α-helix in a DHPC micelle. The same peptide can disrupt and cause leakage in small unilamellar vesicles.</p><p>Single D-amino acid isomers of Trp-cage (20 residues), the smallest peptide with a protein-like fold, were analyzed by CD spectroscopy and were found to have different secondary structures and melting temperatures. They were compared against MS measurements specially designed to reveal the secondary structure of proteins.</p><p>We studied a novel protein in E. coli of unknown structure that is encoded by the putative transcription factor ORF: ygiT (131 residues). This protein comprises a helix-turn-helix (HTH) domain in the C-terminus and contains two CxxC motives in the N-terminal domain, which binds Zn. This protein was named 2CxxC. We succeeded in overexpressing and purifying 2CxxC in E. coli with enough yield for a 13C, 15N uniformly labeled NMR sample. The chemical shift assignment was completed and the NMR structure was calculated in reducing, slightly acidic conditions (1mM DTT, pH 5.5). The determined HTH domain shows good similarity with structures predicted by a homology search, while the N-terminal domain has no other homologous structure in the Protein Data Bank (PDB).</p><p>The structure of the paddle region (27 residues) of the HsapBK(233-260) voltage and Ca+2 activated potassium channel, in DPC micelles, was determined by NMR. It shows a helix-turn-helix loop, which agrees well with the expected structure and could help to verify the proposed models of the voltage gating mechanism.</p><p>The C-repressor (dimer of 99 residues) of bacteriophage P2 was analyzed by NMR. We assigned the chemical shifts and NMR structure determination is under way.</p>

<p>The work in this thesis has demonstrated the advantages and limitations of using MS based technologies in protein and peptide structural studies. </p><p>Tandem MS, specifically electron capture dissociation (ECD) have shown the ability to provide structural insights in molecules containing the slightest of all modifications (D-AA substitution). Additionally, it can be concluded that charge localization in molecular ions is best identified with ECD and to a lesser degree using CAD. </p><p>Fragment ion abundances are a quantifiable tool providing chiral recognition (R_Chiral). An analytical model demonstrating the detection and quantification of D-AAs within proteins and peptides has been achieved. ECD has demonstrated the ability to quantify stereoisomeric mixtures to as little as 1%. Chirality elucidation on a nano LC-MS/MS time scale has been shown. </p><p>The structures of various stereoisomers of the mini protein Trp Cage were explored, each providing unique ECD fragment ion abundances suggestive of gas phase structural differences. The uniqueness of these abundances combined with MDS data have been used in proposing a new mechanism in c and z fragment ion formation in ECD. This mechanism suggests initial electron capture on a backbone amide involved in (neutral) hydrogen bonding.</p><p>The wealth of solution phase (circular dichroism), transitition phase (charge state distribution, CSD) and gas phase (ECD) data for Trp Cage suggest that at low charge states (2+) the molecule has a high degree of structural similarity in solution- and gas- phases. Furthermore, quantitative information from CSD studies is garnered when using a “native” deuteriated form as part of the stereoisomeric mixture. It has also been shown that the stability of the reduced species after electron capture is indicative of the recombination energy release, which in turn is linked to the coulombic repulsion- a structural constraint that can be used for approximation of the inter-charge distance for various stereoisomers.</p>

Phosphoinositide-dependent protein kinase-1 (PDK1) plays an integral role in signaling cellular growth and proliferation, one that's dependent on its ability to autophosphorylate Ser-241 in its T-loop. This process appears to have a strict requirement for its C-terminal pleckstrin homology (PH) domain. Thus, the overall objective of this work was to determine the mechanism by which the PH domain induces an active kinase conformation in unphosphorylated PDK1, capable of Ser-241 autophosphorylation. First, computational modeling and protein cross linking studies were combined with site-directed mutagenesis and kinetic assays in order to provide initial assessment of how the PH domain scaffolds Ser-241 autophosphorylation. A significant number of contacts were identified between the enigmatic "N-bud" region of the PH domain and the kinase domain. Specifically, these studies implicated Glu-432 and Glu-453 of the N-bud region of the PH domain that bind and serve as mimics of the phosphorylated Ser-241 in the T-loop and the phosphorylated C-terminal tail of PDK1 substrates, respectively. Next, a novel method for protein trans-splicing of the regulatory and catalytic kinase domains of PDK1 was developed. The method utilizes the N- and C-terminal split inteins of the gene dnaE from Nostoc punctiforme [(N)NpuDnaE] and Synechocystis sp. strain PCC6803 [(C)SspDnaE], respectively. The cross-reacting KINASE(AEY)-(N)NpuDnaE-His6 and GST-His6-(C)SspDnaE-(CMN)PH fusion constructs generated full length spliced-PDK1 with kobs = (2.8 +- 0.3) x 10-5 s-1. Finally, NMR was used to further characterize the structural and dynamical properties of the PH domain in both its isolated form and in full length PDK1. Whereas, it was not possible to obtain chemical shift assignments of any backbone or side chain nuclear resonances, methods were optimized for 2H,13C,15N-isotopic labeling of the recombinant PH domain. Furthermore, the protein trans-splicing method was significantly improved and utilized for segmental isotopic labeling of the PH domain in full length PDK1. These new findings and developments may provide specific insight and technological improvements towards future studies aimed to better understand and target autoinhibited conformations of PDK1 for translational purposes.

Oxidative protein surface mapping has gained popularity over recent years within the mass spectrometry (MS) community for gleaning information about the solvent accessibility of folded protein structures. The hydroxyl radical targets a wide breadth of reactive amino acids with a stable mass tag that withstands subsequent MS analysis. A variety of techniques exist for generating hydroxyl radicals, with most requiring sources of radiation or caustic oxidizing reagents. The purpose of this research was to evaluate the novel use of electrochemistry for accomplishing a comparable probe of protein structure with a more accessible tool. Two different working electrode types were tested across a range of experimental parameters, including voltage, flow rate, and solution electrolyte composition, to affect the extent of oxidation on intact proteins. Results indicated that the boron-doped diamond electrode was most valuable for protein research due to its capacity to produce hydroxyl radicals and its relatively low adsorption profile. Oxidized proteins were collected from the electrochemical cell for intact protein and peptide level MS analysis. Peptide mass spectral data were searched by two different “hybrid” software packages that incorporate de novo elements into a database search to accommodate the challenge of searching for more than forty possible oxidative mass shifts. Preliminary data showed reasonable agreement between amino acid solvent accessibility and the resulting oxidation status of these residues in aqueous solution, while more buried residues were found to be oxidized in “non-native” solution. Later experiments utilized higher flow rates to reduce protein residence time inside the electrochemical flow chamber, along with a different cell activation approach to improve controllability of the intact protein oxidation yield. A multidimensional chromatographic strategy was employed to improve dynamic range for detecting oxidation of lower reactivity residues. Along with increased levels of oxidation around “reactive hotspot” sites, the enhanced sensitivity of these measurements uncovered a significant level of background oxidation in control proteins. While further work is needed to determine the full utility that BDD electrochemistry can lend protein structural studies, the experimental refinements reported here pave the way for improvements that could lead to a high-throughput structural pipeline complementary to predictive modeling efforts.

Circadian oscillators or clocks are a widespread, endogenous class of oscillatory mechanisms that control the ~24h temporal pattern of diverse organism functions. In cyanobacteria this mechanism is formed by three proteins, KaiA, KaiB and KaiC. KaiA is shown here to be a two domain protein that directly interacts with KaiC and enhances the KaiC autokinase activity. The amino-terminal domain of KaiA can be structurally categorized as a pseudo-receiver, a class of proteins used in signaling cascades and activated by direct protein??protein interactions. The carboxy-terminal domain interacts directly with KaiC, is sufficient to enhance the KaiC autokinase activity in a manner similar to full-length KaiA, and adopts a unique, all α-helical dimeric fold. The structure of this domain raises interesting probabilities regarding the mode of KaiA??KaiC interaction. The two KaiA domains are shown to directly interact with each other, which suggests a possible mechanism of signal transfer from the amino to carboxy-terminal domain. Hydrogen bonds are of paramount importance in nucleic acid structure and function. Here we show that changes in the width and anharmonicity of vibrational potential energy wells of hydrogen bonded groups can be measured in nucleic acids and can possibly be correlated to structural properties, such as length. Deuterium/protium fractionation factors, which are sensitive to the vibrational potential well width, were measured for the imino sites of thymidine residues involved in A:T base pairs or free in solution, and a correlation was established between decreasing fractionation factors and increasing imino proton chemical shift, δH3. Similarly, a correlation was observed between δH3and deuterium isotope effects (DIE) on chemical shift of thymidine carbon atoms. Combined these results indicate that as hydrogen-bond strength increases the vibrational potential wells of imino protons widen with a corresponding increase in anharmonicity. However, trans-hydrogen bond DIE on carbon chemical shifts of A:T base-paired adenosine residues do not correlate with those measured on thymidine residues. We propose that this lack of correlation is due to DIE dependence on base-pair geometry, which is not easily measured by traditional NMR experiments.