For understanding the function of a protein, the protein structure plays an important role. The prediction of protein structure from its primary sequence has significant assistance in bioinformatics. Generally, the real protein structures can be reconstructed by some costly techniques, but predicting the protein structures helps us guess the functional expression of a protein in advance. In this thesis, we develop three terms as the materials of the fitness function that can be successfully used in protein backbone structure prediction. In the result of this thesis, it shows that over 80% of good values calculated from our fitness function, which are generated by the genetic programming, are better than the average in the CASP8.

Structured RNAs are pervasive in biology, spanning a functional repertoire that includes messengers, regulators of gene expression and catalysts of translation and splicing. From the relatively simple tRNAs and riboswitches to the highly structured ribosomal RNAs, the ability of RNAs to function is dependent on well-defined secondary and tertiary structures. However, studies of RNA folding in vitro have revealed an extreme propensity to form alternative structures, which can be long-lived and interfere with function. In the cell, a diverse array of RNA binding proteins and RNA chaperones guide RNAs towards the correct structure and disrupt misfolded intermediates. Among these proteins, DEAD-box protein family stands out as one of the largest groups, with its members ubiquitously involved in RNA metabolism across all domains of life. DEAD-box proteins can function as both specific and general RNA chaperones by disrupting RNA structures in an ATP-dependent manner. Here I describe my work studying the general RNA chaperone mechanism of the Neurospora crassa protein CYT-19, a model DEAD-box protein and a biological RNA chaperone that is required for efficient folding of self-splicing group I intron RNAs in vivo. After an introduction to DEAD-box proteins and their mechanisms as RNA remodelers (Chapter 1), I will first describe studies of group I intron unfolding by CYT-19, focusing on the effects of RNA tertiary structure stability on CYT-19 activity and targeting to RNA substrates (Chapter 2). I will then describe the characterization of ATP-dependent mechanisms during CYT-19-mediated refolding of the misfolded group I intron (Chapter 3). In Chapter 4, I will present small-angle X-ray scattering (SAXS) studies of structural features of DEAD-box proteins that allow them to efficiently interact with large structured RNA substrates. Finally, I will turn to studies of DEAD-box protein involvement during early steps of RNA compaction and folding, using SAXS and activity-based approaches (Chapter 5). I will conclude with a general discussion of superfamily 2 RNA helicases, which include DEAD-box and related proteins, and their functions and mechanisms as remodelers of structured RNAs and RNPs. / text

Multiple lines of evidence suggest that RNA performed all of the biological functions in the first life forms on earth. These functions included cleavage, ligation, polymerization, recognition, binding, and replication. In order to perform these functions, populations of RNA molecules with unevolved sequences must have been able to fold into compact three dimensional shapes, in unregulated environments, and without the help of proteins. Folding into compact tertiary structures is difficult because of the high charge density of RNA. Consequently, the ranges of temperature, salinity, pH, and pressure that allow RNA to fold into functional shapes is very restricted. We use thermodynamic arguments and Brownian dynamics simulations to compute the range of these environmental parameters that will allow RNA to fold. This is a non-trivial calculation due to the formation of an ion atmosphere around RNA that reduces its electric field. The results can be used to clarify the environments in which the transition to life is possible. Our preliminary calculations suggest that environments with low temperatures ($0-50^\circ C$) and high salt concentrations (greater than 100mM) are the most favorable for unassisted RNA folding and thus the transition to RNA-based life. Applications of our results include determining the environments on early earth where life formed, assesing the habitability of Europa, Titan, and (using modeled parameters) extrasolar planets. / text

The interaction of proteins with surfaces is a major problem involved in protein microarrays. Understanding protein/surface interactions is key to improving the performance of protein microarrays, but current understanding of the behavior of proteins on surfaces is lacking. Prevailing theories on the subject, which suggest that proteins should be stabilized when tethered to surfaces, do not explain the experimentally observed fact that proteins are often denatured on surfaces. In an attempt to develop some predictive capabilities with respect to protein/surface interactions, it was asked in previous works if the stabilization/destabilization of proteins on surfaces could be correlated to secondary structure and found that no link existed. However, further investigation has revealed that proteins with similar tertiary structure show predictable stabilization patterns. In this research, it is reported how five, alpha-helical, orthogonal-bundle proteins behave on the surface compared to the bulk. By measuring stabilization using melting temperatures and the Gibbs energies of folding, it is shown that the stability of proteins tethered to surfaces can be correlated to the shape of the loop region where the tether is placed and the free rotation ability of the part of proteins near surfaces. It is also shown that any destabilization that occurs because of the surface is an enthalpic effect and that surfaces always stabilize proteins entropically. Furthermore, the entropical stabilization effect comes from unfolded states of the tethered protein, while the enthalpical destabilization effect is from the folded states of protein. A further analysis of surface induced change of folding mechanism is also studied with a multi-state protein 7LZM in this research. The result showed that by tethering a protein on a surface, the melting temperature of part of the protein changed, which leads to a miss of state.

The interaction of proteins with surfaces is a major process involved in protein microarrays. Understanding protein-surface interactions is key to improving the performance of protein microarrays, but current understanding of the behavior of proteins on surfaces is lacking. Prevailing theories on the subject, which suggest that proteins should be stabilized when tethered to surfaces, do not explain the experimentally observed fact that proteins are often denatured on surfaces. This document outlines several studies done to develop a model which is capable of predicting the stabilization and destabilization of proteins tethered to surfaces. As the start point of the research, part of this research showed that the stability of five mainly-alpha, orthogonal-bundle proteins tethered to surfaces can be correlated to the shape of the loop region where the tether is placed and the free rotation ability of the part of proteins near surfaces. To test the expandability of the protein stability prediction pattern derived for mainly-alpha, orthogonal-bundle proteins, same analysis is performed for proteins from other structure motifs. Besides the study in these small two-state proteins, a further analysis of surface-induced change of folding mechanism is also studied with a multi-state lysozyme protein 7LZM. The result showed that by tethering a protein on a surface, the melting temperature of a part of the protein changed, which leads to an avoidance of the meta-stable state. Besides the change of folding mechanism, by tethering the lysozyme protein to a certain site, the protein could both keep a stable structure and a good orientation, allowing active sites to be available to other proteins in bulk solution. All the work described above are done with a purely repulsive surface model which was widely used to roughly simulate solid surfaces in protein microarrays. For a next-level understanding of protein-surface interactions, a novel coarse-grain surface model was developed, parameterized, and validated according to experimental results from different groups. A case study of interaction between lysozyme protein 7LZM and three types of surfaces with the novel model has been performed. The results showed that protein stabilities and structures are dependent on the types of surfaces and their different hydrophobicities. This result is consistent with previously published experimental work.

No description available.

This thesis investigates the correlations between short protein peptide sequences and local tertiary structures. In particular, it introduces a novel algorithm for partitioning short protein segments into clusters of local sequence-structure motifs, and demonstrates that these motif clusters contain useful structural information via two applications to structural prediction. The first application utilizes motif clusters to predict local protein tertiary structures. A novel dynamic programming algorithm that performs comparably with some of the best existing algorithms is described. The second application exploits the capability of motif clusters in recognizing regular secondary structures to improve the performance of secondary structure prediction based on Support Vector Machines. Empirical results show significant improvement in overall prediction accuracy with no performance degradation in any specific aspect being measured. The encouraging results obtained illustrate the great potential of using local sequence-structure motifs to tackle protein structure predictions and possibly other important problems in computational biology.

This thesis investigates the correlations between short protein peptide sequences and local tertiary structures. In particular, it introduces a novel algorithm for partitioning short protein segments into clusters of local sequence-structure motifs, and demonstrates that these motif clusters contain useful structural information via two applications to structural prediction. The first application utilizes motif clusters to predict local protein tertiary structures. A novel dynamic programming algorithm that performs comparably with some of the best existing algorithms is described. The second application exploits the capability of motif clusters in recognizing regular secondary structures to improve the performance of secondary structure prediction based on Support Vector Machines. Empirical results show significant improvement in overall prediction accuracy with no performance degradation in any specific aspect being measured. The encouraging results obtained illustrate the great potential of using local sequence-structure motifs to tackle protein structure predictions and possibly other important problems in computational biology.

Les acides ribonucléiques (ARN) forment des structures tri-dimensionnelles complexes stabilisées par la formation de la structure secondaire (2D), elle-même formée de paires de bases. Plusieurs méthodes computationnelles ont été créées dans les dernières années afin de prédire la structure 2D d’ARNs, en partant de la séquence. Afin de simplifier le calcul, ces méthodes appliquent généralement des restrictions sur le type de paire de bases et la topologie des structures 2D prédites. Ces restrictions font en sorte qu’il est parfois difficile de savoir à quel point la totalité des paires de bases peut être représentée par ces structures 2D restreintes. MC-Unfold fut créé afin de trouver les structures 2D restreintes qui pourraient être associées à une structure secondaire complète, en fonction des restrictions communément utilisées par les méthodes de prédiction de structure secondaire. Un ensemble de 321 monomères d’ARN totalisant plus de 4223 structures fut assemblé afin d’évaluer les méthodes de prédiction de structure 2D. La majorité de ces structures ont été déterminées par résonance magnétique nucléaire et crystallographie aux rayons X. Ces structures ont été dépliés par MC-Unfold et les structures résultantes ont été comparées à celles prédites par les méthodes de prédiction. La performance de MC-Unfold sur un ensemble de structures expérimentales est encourageante. En moins de 5 minutes, 96% des 227 structures ont été complètement dépliées, le reste des structures étant trop complexes pour être déplié rapidement. Pour ce qui est des méthodes de prédiction de structure 2D, les résultats indiquent qu’elles sont capable de prédire avec un certain succès les structures expérimentales, particulièrement les petites molécules. Toutefois, si on considère les structures larges ou contenant des pseudo-noeuds, les résultats sont généralement défavorables. Les résultats obtenus indiquent que les méthodes de prédiction de structure 2D devraient être utilisées avec prudence, particulièrement pour de larges molécules. / Ribonucleic acids (RNA) adopt complex three dimensional structures which are stabilized by the formation of base pairs, also known as the secondary (2D) structure. Predicting where and how many of these interactions occur has been the focus of many computational methods called 2D structure prediction algorithms. These methods disregard some interactions, which makes it difficult to know how well a 2D structure represents an RNA structure, especially when large amounts of base pairs are ignored. MC-Unfold was created to remove interactions violating the assumptions used by prediction methods. This process, named unfolding, extends previous planarization and pseudoknot removal methods. To evaluate how well computational methods can predict experimental structures, a set of 321 RNA monomers corresponding to more than 4223 experimental structures was acquired. These structures were mostly determined using nuclear magnetic resonance and X-ray crystallography. MC-Unfold was used to remove interactions the prediction algorithms were not expected to predict. These structures were then compared with the structured predicted. MC-Unfold performed very well on the test set it was given. In less than five minutes, 96% of the 227 structure could be exhaustively unfolded. The few remaining structures are very large and could not be unfolded in reasonable time. MC-Unfold is therefore a practical alternative to the current methods. As for the evaluation of prediction methods, MC-Unfold demonstrated that the computational methods do find experimental structures, especially for small molecules. However, when considering large or pseudoknotted molecules, the results are not so encouraging. As a consequence, 2D structure prediction methods should be used with caution, especially for large structures.

A Determinação da Estrutura tridimensional de Proteínas (DEP) a partir da sua seqüência de aminoácidos é importante para a engenharia de proteínas e o desenvolvimento de novos fármacos. Uma alternativa para este problema tem sido a aplicação de técnicas de computação evolutiva. As abordagens utilizando Algoritmos Evolutivos (AEs) tem obtido resultados relevantes, porém estão restritas a pequenas proteínas, com dezenas de aminoácidos e a algumas classes de proteínas. Este trabalho propõe a investigação de uma abordagem utilizando AEs para a predição da estrutura terciária de proteínas independentemente do seu tamanho e classe. Os resultados obtidos demonstram que apesar das dificuldades encontradas a abordagem investigada constitue-se em uma alternativa em relação aos métodos clássicos de determinação da estrutura terciária das proteínas. / Protein structure determination (DEP) from aminoacid sequences is very importante to protein engineering and development of new drugs. Evolutionary computation has been aplied to this problem with relevant results. Nevertheless, Evolutionary Algorithms (EAs) can work with only proteins with few aminoacids and some protein classes. This work proposes an approach using AEs to predict protein tertiary structure independly from their size and class. The obtained results show that, despite of the difficulties that have been found, the investigate approach is a relevant alternative to classical methods to protein structure determination.