Thesis advisor: Udayan Mohanty / Computational research can play a crucial component in the discovery of unique biochemical phenomena, from answering fundamental questions about molecular function and structure to the modeling of designed pharmaceuticals to cure many debilitating illnesses. Here computational methods are employed to examine the exquisite role that magnesium ions play in stabilizing ribosomal subunits responsible for protein translation and to analyze the potential of a proposed anticancer drug for a pathway that is impaired in the majority of human cancer cases. / Thesis (BS) — Boston College, 2011. / Submitted to: Boston College. College of Arts and Sciences. / Discipline: College Honors Program. / Discipline: Chemistry.

Accurate prediction of binding free energies of protein-ligand system has long been a focus area for theoretical and computational studies; with important implications in fields like pharmaceuticals, enzyme-redesign, etc. The aim of this project was to develop such a predictive model for calculating binding free energies of protein-ligand systems based on the LIE-SASA methods. Many models have been successfully fit to experimental data, but a general predictive model, not reliant on experimental values, would make LIE-SASA a more powerful and widely applicable method. The model was developed such that There is no significant increase in computational time No increase in complexity of system setup No increase in the number of empirical parameters. The method was tested on a small number of protein-ligand systems, selected with certain constraints. This was our training set, from which we obtain the complete expression for binding free energy. Expectedly, there was good agreement with experimental values for the training set On applying our model to a similar sized validation set, with the same selection constraints as for the training set, we achieved even better agreement with experimental results, with lower standard errors. Finally, the model was tested by applying it to a set of systems without such selection constraints, and again found good agreement with experimental values. In terms of accuracy, the model was comparable to a system specific empirical fit that was performed on this set. These encouraging results could be an indicator of generality.

Os sistemas biológicos macromoleculares são conhecidos por serem sistemas interagentes. Essas interações são fundamentais para processos como comunicação celular, especificidade de reações enzimáticas e regulação da expressão gênica. Os métodos disponíveis atualmente para estimar a afinidade das interações biomoleculares podem ser divididos, basicamente, em dois grupos: métodos rápidos que estimam a energia livre de ligação através de aproximações de campo de força (por exemplo, docking); e os métodos que são baseados em ensembles de Dinâmica Molecular (DM) para calcular as energias livres de ligação de maneira mais rigorosa, porém, com custo computacional mais elevado. O objetivo deste trabalho é aprimorar e validar um método menos custoso para o cálculo da energia livre de ligação. Para isso, simulações atomísticas de Monte Carlo (MC) dos ligantes no sítio de ligação são usadas para gerar ensembles termodinâmicos. Depois disso, as energias livres de ligação são calculadas usando uma combinação de energias e entropias estimadas através de uma estratégia de aproximação de primeira ordem. Dois algoritmos de amostragem foram avaliados no cálculo de energia de ligação. O primeiro algoritmo amostra graus de liberdade de translação e rotação randômicas do centro de massa do ligante no sítio de ligação, além de variações randômicas nos ângulos de torção envolvendo átomos pesados (não hidrogênio). O segundo amostra graus de liberdade rotacional e translacional do centro de massa, além de deslocamentos atômicos individuais para cada átomo do ligante. Além disso, diferentes modelos para calcular as contribuições polares para interação intermolecular foram utilizados. Comparações entre as energias livres de ligação calculadas com baixo custo computacional e as experimentais disponíveis na literatura para o sistema modelo utilizado, lisozima do vírus T4, mostraram uma correlação considerável (r=0,64 para N=27). Esses dados também apresentaram resultados interessantes quando comparados com outras metodologias, tais como LIE, MM-PBSA e MM-GBSA. Assim, a abordagem utilizada para a determinação das energias de interação mostrou-se eficiente em termos de tempo computacional e para comparação com dados de energia livre de ligação determinados experimentalmente. / Macromolecular biological systems are widely known by its interaction properties. Those interactions play fundamental roles in processes such as cellular communication, specificity of enzymatic reactions and regulation of gene expression. The methods currently available to estimate the affinity of biomolecular interactions can be divided basically into two groups: fast methods that estimate the free energy of binding through force field approximations (e.g., docking); and methods that are based on Molecular Dynamics (DM) ensembles to calculate binding free energies more rigorously, however, with higher computational cost. The objective of this work is to improve and validate a less costly method for calculating binding free energy. For this, atomistic Monte Carlo (MC) simulations of ligands at the binding site are used to generate thermodynamic ensembles. Thereafter, the binding free energies are calculated using a combination of energies and entropy estimated through a first-order approximation strategy. Two sampling algorithms were evaluated in the calculation of the binding energy. The first one samples the degrees of freedom from translation and rotation of the center of mass of the binder at the binding site, as well as random variations in the torsion angles involving heavy atoms (non-hydrogen). The second one samples the rotational and translational degrees of freedom of the ligand center of mass, as well as individual atomic displacements for each atom of the ligand. In addition, different models to calculate the polar contributions for intermolecular interaction were used. Comparisons between the binding free energies calculated with low computational cost and the experimental ones available in the literature for the system used, T4 virus lysozyme, resulted in acceptable correlation values (r=0.64 for N=27). Those data also showed interesting results compared to different methodologies such as LIE, MM-PBSA and MM-GBSA. Therefore, the used approach for determining the binding energies was efficient in terms of computational time and for comparison with free energy data determined experimentally.

SHAPE-DEPENDENT MOLECULAR RECOGNITION OF SPECIFIC SEQUENCES OF DNA BY HETEROCYCLIC CATIONS by YI MIAO Under the Direction of Dr. W. David Wilson ABSTRACT DB921 and DB911 are biphenyl-benzimidazole-diamidine isomers with a central para- and meta-substituted phenyl group, respectively. Unexpectedly, linear DB921 has much stronger binding affinity with DNA than its curved isomer, DB911. This is quite surprising and intriguing since DB911 has the classical curved shape generally required for strong minor groove binding while DB921 clearly does not match the groove shape. Several biophysical techniques including thermal melting (Tm), circular dichroism (CD), biosensor-surface plasmon resonance (SPR), and isothermal titration calorimetry (ITC) have been utilized to investigate the interactions between these compounds and DNA. The structure of the DB921-DNA complex reveals that DB921 binds to DNA with a reduced twist of the biphenyl for better fit of DB921 into the minor groove. A bound water molecule complements the curvature of DB921 and contributes for tight binding by forming H-bonds with both DNA and DB921. Structure-affinity relationship studies of a series of DB921 analogs show that the benzimidazole group is one of the key groups of DB921 for its strong binding to the minor groove. Thermodynamic studies show that the stronger binding of DB921 is due to a more favorable binding enthalpy compared to DB911 even though the complex formation with DNA for these compounds are all predominantly entropically driven. DB921 also has more negative heat capacity change than DB911. The initial studies of inhibition of the interaction between an AT hook peptide of HMGA proteins and its target DNA by a set of diamidine AT-minor groove binders using biosensor-SPR technique show that the inhibitory ranking order is consistent with that of binding affinity and linear-shaped DB921 still has excellent inhibitory effects. These heterocyclic cations rapidly inhibit the binding of DBD2 peptide to the DNA and may only block the specific AT binding of the peptide without hindering the non-specific binding interaction. The results of this project have shown that DB921 represents a new novel effective minor groove binder that does not fit the traditional model and is a potential inhibitor for DNA/protein complexes. INDEX WORDS: Molecular recognition, DNA binding, Minor groove binding, Linear shape, Compound curvature, Binding affinity, Binding kinetics, Thermodynamics, Surface plasmon resonance, Isothermal titration calorimetry, Inhibition

A central goal of computational biophysics and biochemistry is to understand the behavior, interactions, and reactions of molecules, and to interpret and facilitate experimental design. The objective of this thesis research is to use the molecular modeling and simulation techniques to advance our understanding of principles in molecular structure properties, recognition and interaction at the atomic level. First, a physical molecular mechanics model is built to study the conformational properties of depsipeptide, which shows potential for engineered protein mimetics with controllable structure and function. We explore the possible kinase-substrate binding modes and the likelihood of an [alpha]-helix docking interaction within a kinase active site. Finally, efficient physical models based on a polarizable potential function are developed to describe the structural properties and calculate protein-ligand binding affinities accurately for both trypsin and matrix metalloproteinase. / text

A dehydrogen is an ‘under-wrapped’ hydrogen bond in a protein that is purported to be a hot spot for binding due to the favorable replacement of water with hydrocarbon upon binding of another protein. A model at the level of dielectric constants is used to test the validity of the claim that moving a hydrogen bond from high dielectric (i.e. a dehydron) to low dielectric (i.e. after binding of another protein) is actually a thermodynamically favorable process. In simulation, several proteins have been shown to undergo a dewetting transition when fixed components are separated a small distance. A new atomic-level hydrophobicity scale is combined with topographical information to characterize protein interfaces. The relationship between hydrophobicity and topography for protein surfaces known to be involved in binding is examined. This framework is then applied to identify surface characteristics likely to have an affect on the occurrence of a dewetting transition. Cadmium selenide (CdSe) nanoparticles form nanospheres or nanorods when grown in solutions of varying concentrations of the surfactants hexylphosphonic acid (HPA) and trioctylphosphine oxide (TOPO). Relative binding free energies are calculated for HPA and TOPO to the solvent-accessible faces of CdSe crystals. Binding free energies calculated with a Molecular Mechanics-Generalized Born model are used to identify a set of low free energy structures for which the solvation free energy is refined with the solution to the Poisson equation. These relative binding free energies provide information about the relative growth rates of these crystal faces in the presence of surfactants. Relative growth rates are then used to help understand why nanoparticles form certain shapes in the presence of specific surfactants. / text

The significance of endocrine-disrupting compounds (EDCs) in the environment has only recently come to the forefront of scientific research, policy debates, water utilities management and public awareness. EDCs have the ability to interfere with the normal functioning of the endocrine system of humans and other animals. Numerous chemicals are included in the class of compounds known as EDCs, and exposure is widespread. These compounds are found in a variety of environmental matrices (e.g., marine and freshwater systems, sediment, soil), transported there primarily through sewage effluent discharge and recycling of sewage sludge for topical fertilizer use. This transport to the environment serves as the primary route of exposure for aquatic and terrestrial organisms living there. Furthermore, these compounds are also found in consumer products, food and drinking water--which serve as the exposure source for human beings. Multiple examples of endocrine disruption have been documented in humans and animals, and certain EDCs have been implicated in each case. The future of public and environmental health will depend upon mitigating the effects of these chemicals. This purpose of this dissertation is to provide an initial understanding of EDC occurrence in the Tampa Bay region of south Florida, and to complement the existing body of EDC research with regards to marine systems. It focuses on estrogenic EDCs, specific compounds which target the estrogen axis of the endocrine system. Six estrogenic EDCs were chosen based on their documented prevalence in the environment, prevalence in sewage, and for their suspected endocrine-disrupting effects: estrone, 17β-estradiol, estriol, 17α-ethinylestradiol, bisphenol-A and nonylphenol. These compounds were verified to be amenable to and detectable by gas chromatography-mass spectrometry analysis. Since the occurrence of EDCs in aquatic environments of the Tampa Bay region had not been previously characterized, the initial phase of the research focused on quantification of the six estrogenic EDCs in Tampa Bay area water, sediment, and sewage influent and effluent. All targeted EDCs were present in 89% of sewage samples, while 100% of the samples contained at least one or more EDCs. The concentrations of EDCs in marine aqueous and sediment samples tended to decrease with increasing distance from the wastewater treatment plant discharge site. The ubiquitous presence of these estrogenic EDCs in the Tampa Bay area is cause for concern with respect to endocrine disruption in local terrestrial and aquatic wildlife. Since the Tampa Bay region is home to a wide variety of marine organisms, constant exposure to EDCs could result in ecosystem-level effects, as these compounds can impair reproductive fitness and lead to other adverse health effects. This research also served to enlarge the existing scientific literature on EDC occurrence, as many marine and freshwater systems continue to be characterized globally. The very basis for expecting to find EDCs in the Tampa Bay area had come from the fact that the main source of environmental contamination is typically the effluent discharge from area wastewater treatment plants. Conventional wastewater treatment plant processes are designed to reduce the amount of organic matter, pathogens and nutrients from the incoming influent. However, the processes are not as effective in removing micropollutants, including EDCs. These compounds notoriously evade traditional wastewater treatment technologies and are found even in tertiary-treated effluent. For this reason, the second phase of the research assessed an electro-chemical technique for the removal of the same six EDCs. The removal technique was tested on a laboratory scale and has a commercial-sized counterpart which can be integrated at the level of the wastewater treatment plant. In order to test the removal efficiency, samples of influent and tertiary-treated effluent were spiked with the six EDCs. The mean concentration of each EDC component was statistically lower after treatment (removal range = 42% - 98.2%), demonstrating the effectiveness of this electro-chemical process for EDC removal from both raw and treated sewage. The significance of the results lies in the fact that if this method is implemented, then future wastewater treatment plant effluent discharge (similar to that of the Tampa Bay region) could be less impacted by EDCs and therefore cleaner for the environment into which it is being discharged. For the final phase of the research, the use of computational techniques to simulate human endogenous estrogen binding to its receptor was started as a foundation for future models to eventually predict endocrine-disrupting potential of different chemical compounds. We built an estradiol-human estrogen receptor model, and used molecular dynamic simulations to determine the binding free energy. The calculated total binding free energy of estradiol bound to the ligand binding domain of the human estrogen receptor was found to be -16.85 kcal/mol, which is in range of the experimental value of -12.40 kcal/mol. Humans are chronically exposed to low doses of EDCs every day, which makes endocrine disruption a considerable public health issue. Human exposure to EDCs is completely different from marine organism exposure, but the adverse effects are no less significant. The successful completion of this model serves as a platform for 1. Testing the human model against endocrine-disrupting compounds, 2. Subsequent models that will be developed for different species, including marine species important to Tampa Bay. Substantial data exist regarding the exposures and health risks associated with EDCs in humans and wildlife on a global scale. As the pressing issues of climate change and carbon emissions are at the top of the list of environmental concerns, it is important to note that mitigating the effects of EDCs should not be overlooked and will be an important responsibility of regulatory agencies in the near future.

No description available.

Molecular association is part of many important processes in living cells. Computational methods for calculating binding free energies allows for a quantitative examination of biomolecular structures and hypotheses drawn from biochemical experiments. Here, binding free energy calculations for tRNAs and release factors binding to mRNA codons on the ribosome, sugars binding to lectins and purine analogs binding to the purine riboswitch are presented. The relative affinities between cognate and non-cognate tRNAs for different states involved in codon reading on the ribosome were determined. The calculations show that tRNA discrimination varies between different conformations of the 30S subunit, where the existence of both low and high selectivity states provides an efficient common mechanism for initial selection and proofreading. The simulations reveal a desolvation mechanism for the 30S conformational switch with which the accuracy of peptide bond formation can be amplified. When an mRNA stop codon (UAA, UAG or UGA) is located in the ribosomal A-site release factors bind to the ribosome and the synthesized protein is released. RF1 is specific for UAA and UAG whereas RF2 is specific for UAA and UGA. The free energy calculations and an analysis of the performed simulations show the mechanisms for how RF1 and RF2 are able to read the stop codons with different specificities. Also mitochondrial release factors were investigated. Vertebrate mitochondria have four stop codons, UAA, UAG, AGA and AGG and two release factors mtRF1 and mtRF1a. The calculations show how the specificities of both mtRF1 and mtRF1a agree with RF1 and that none of them are likely to read the non-standard stop codons AGA and AGG. The linear interaction energy method has also been examined for the RSL and PA-IIL lectins and for the purine riboswitch. The standard parameterization of the method works well for RSL, but fails for PA-IIL and the purine riboswitch due to compositions of the active sites in these systems. The development of new parameterizations to overcome these problems leads to a better understanding of both the method and the binding mechanisms in these systems.

La reconnaissance moléculaire est un problème très intéressant et surtout un défi actuel pour la chimie biophysique. Avoir des prévisions fiables sur la reconnaissance spécifique entre les molécules est hautement prioritaire, car il fournira un aperçu des problèmes fondamentaux et suscitera des applications technologiques pertinentes. La thèse présentée ici est centrée sur une analyse quantitatif de la reconnaissance moléculaire en solution pour la liaison l'hôte-invité, la liaison protéine-ligand et la catalyse. Le cadre de la mécanique statistique utilisé pour décrire l'état de la technique de liaison récepteur-ligand est un point d'inflexion pour le développement de nouvelles méthodes améliorées. En fait, un modèle très performant et précis a été obtenu pour l'analyse de la liaison hôte-invité. Enfin, les modèles présentés ont été utilisés comme outils prédictifs fiables pour la découverte de nouvelles entités chimiques destinées à améliorer la catalyse en solution. / Molecular recognition is a very interesting problem, and foremost, a current challenge for biophysical chemistry. Having reliable predictions on the specific recognition between molecules is highly priority as it will provide an insight of fundamental problems and will raise relevant technological applications. The dissertation presented here is centered on a quantitative analysis of molecular recognition in solution for host-guest, protein-ligand binding and catalysis. The statistical mechanics framework used to describe the state-of-the-art for receptor-ligand binding is an inflection point for the developing of new improved and methods. In fact, a highly performanced and accurate model was obtained for the analysis of host-guest binding. Finally, the presented models were used as a reliable predictive tools for discovering new chemical entities for enhance catalysis in solution.