Atomistic modeling and simulation methods facilitate biomedical research from many respects, including structure-based drug design. The ability of these methods to address biologically relevant problems is largely determined by the accuracy of the treatment of complex solvation effects in target biomolecules surrounded by water. The implicit solvent model – which treats solvent as a continuum with the dielectric and non-polar properties of water – offers a good balance between accuracy and speed. Simple and efficient, generalized Born (GB) model has become a widely used implicit solvent responsible for the estimation of key electrostatic interactions. The main goal of this research is to improve the accuracy of protein-ligand binding calculations in the implicit solvent framework. To address the problem (1) GBNSR6, an accurate yet efficient flavor of GB, has been thoroughly explored in the context of protein-ligand binding, (2) a global multidimensional optimization pipeline is developed to find the optimal dielectric boundary made of atomic and water probe radii specifically for protein-ligand binding calculations using GBNSR6. The pipeline includes (3) two novel post-processing steps for optimum robustness analysis and optimization landscape visualization. In the final step of this research, (4) accuracy gain the optimal dielectric boundary can bring in practice is explored on binding benchmarks, including the SARS-CoV-2 spike receptor-binding domain and the human ACE2 receptor. / Doctor of Philosophy / Drug discovery is one of the most challenging tasks in biological sciences as it takes about 10-15 years and $1.5-2 billion on average to discover a new drug. Therefore, efforts to speed up this process or lower its costs are highly valuable. Computer-aided drug design (CADD) plays a crucial role in the early stage of drug discovery. In CADD, computational approaches are used in order to discover, develop, and analyze drugs and similar biologically active molecules, such as proteins. Proteins are an important class of biological macromolecules that perform their functionality mainly through interactions with other molecules, for example, binding to small molecules so-called ligands. Thorough understanding of protein-ligand interactions is central to comprehending biology at the molecular level. In this study, we introduce and analyze a computational model used for protein-ligand binding free energy calculations. A global multidimensional optimization pipeline is developed to find the optimal parameters of the model,˘aparticularly˘athose parameters involved in the dielectric boundary. In order to examine the robustness of the optimal model to unavoidable perturbations and uncertainties, virtually inevitable in any complex system being optimized, a novel robustness metric is introduced. Finally, the robust optimal model is tested on protein-ligand benchmarks, including a complex related to the novel coronavirus. Results demonstrate relatively higher accuracy in terms of binding free energy calculations compared to reference models.

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

Proteins and DNA are large, complex molecules that carry out biological functions essential to all life. Their successful operation relies on adopting specific structures, stabilized by intra-molecular interactions between atoms. The spatial and temporal resolution required to study the mechanics of these molecules in full detail can only be obtained using computer simulations of molecular models. In a molecular dynamics simulation, a trajectory of the system is generated, which allows mapping out the states and dynamics of the molecule. However, the time and length scales characteristic of biological events are many orders of magnitude larger than the resolution needed to accurately describe the microscopic processes of the atoms. To overcome this problem, sampling methods have been developed that enhance the occurrence of rare but important events, which improves the statistics of simulation data. This thesis summarizes my work on developing the AWH method, an algorithm that adaptively optimizes sampling toward a target function and simultaneously finds and assigns probabilities to states of the simulated system. I have adapted AWH for use in molecular dynamics simulations. In doing so, I investigated the convergence of the method as a function of its input parameters and improved the robustness of the method. I have also worked on a generally applicable approach for calculating the target function in an automatic and non-arbitrary way. Traditionally, the target is set in an ad hoc way, while now sampling can be improved by 50% or more without extra effort. I have also used AWH to improve sampling in two biologically relevant applications. In one paper, we study the opening of a DNA base pair, which due to the stability of the DNA double helix only very rarely occurs spontaneously. We show that the probability of opening depends on both nearest-neighbor and longer-range sequence effect and furthermore structurally characterize the open states. In the second application the permeability and ammonia selectivity of the membrane protein aquaporin is investigated and we show that these functions are sensitive to specific mutations. / <p>QC 20171117</p>

No description available.

Protein-ligand binding and protein allostery play a crucial role in cell signaling, cell regulation, and modern drug discovery. In recent years, experimental studies of protein structures including crystallography, NMR, and Cryo-EM are widely used to investigate the functional and inhibitory properties of a protein. On the one hand, structural classification and feature identification of the structures of protein kinases, HIV proteins, and other extensively studied proteins would have an increasingly important role in depicting the general figures of the conformational landscape of those proteins. On the other hand, free energy calculations which include the conformational and binding free energy calculation, which provides the thermodynamics basis of protein allostery and inhibitor binding, have proven its ability to guide new inhibitor discovery and protein functional studies. In this dissertation, I have used multiple different analysis and free energy methods to understand the significance of the conformational and binding free energy landscapes of protein kinases and other disease-related proteins and developed a novel alchemical-based free energy method, restrain free energy release (R-FEP-R) to overcome the difficulties in choosing appropriate collective variables and pathways in conformational free energy methods like umbrella sampling and metadynamics. / Chemistry

No description available.

No description available.

No description available.

Les interactions protéine-protéine sont souvent contrôlées par de petits domaines protéiques qui régulent les chemins de signalisation au sein des cellules eucaryotes. Les domaines PDZ sont parmi les domaines les plus répandus et les plus étudiés. Ils reconnaissent spécifiquement les 4 à 10 acides aminés C-terminaux de leurs partenaires. Tiam1 est un facteur d'échange de GTP de la protéine Rac1 qui contrôle la migration et la prolifération cellulaire et dont le domaine PDZ lie les protéines Syndecan-1 (Sdc1), Caspr4 et Neurexine. Des petits peptides ou des molécules peptidomimétiques peuvent potentiellement inhiber ou moduler son activité et être utilisés à des fins thérapeutiques. Nous avons appliqué des approches de dessin computationnel de protéine (CPD) et de calcul d'énergie libre par simulations dynamique moléculaire (DM) pour comprendre et modifier sa spécificité. Le CPD utilise un modèle structural et une fonction d'énergie pour explorer l'espace des séquences et des structures et identifier des variants protéiques ou peptidiques stables et fonctionnels. Nous avons utilisé le programme de CPD Proteus, développé au laboratoire, pour redessiner entièrement le domaine PDZ de Tiam1. Les séquences générées sont similaires à celles des domaines PDZ naturels, avec des scores de similarité et de reconnaissance de pli comparables au programme Rosetta, un outil de CPD très utilisé. Des séquences contenant environ 60 positions mutées sur 90, ont été testées par simulations de DM et des mesures biophysiques. Quatre des cinq séquences testées expérimentalement (par nos collaborateurs) montrent un dépliement réversible autour de 50°C. Proteus a également déterminer correctement la spécificité de la liaison de quelques variants protéiques et peptidiques. Pour étudier plus finement la spécificité, nous avons paramétré un modèle d'énergie libre semi-empirique de Poisson-Boltzmann ayant la forme d'une énergie linéaire d'interaction, ou PB/LIE, appliqué à des conformations issues de simulations de DM en solvant explicite de complexes PDZ:peptide. Avec trois paramètres ajustables, le modèle reproduit correctement les affinités expérimentales de 41 variants, avec une erreur moyenne absolue de 0,4~kcal/mol, et donne des prédictions pour 10 nouveaux variants. Le modèle PB/LIE a ensuite comparé à la méthode non-empirique de calcul d'énergie libre par simulations alchimiques, qui n'a pas de paramètre ajustable et qui prédit correctement l'affinité de 12 complexes Tiam1:peptide. Ces outils et les résultats obtenus devraient nous permettre d'identifier des peptides inhibiteurs et auront d'importantes retombées pour l'ingénierie des interactions PDZ:peptide. / Small protein domains often direct protein-protein interactions and regulate eukaryotic signalling pathways. PDZ domains are among the most widespread and best-studied. They specifically recognize the 4-10 C-terminal amino acids of target proteins. Tiam1 is a Rac GTP exchange factor that helps control cellmigration and proliferation and whose PDZ domain binds the proteins syndecan-1 (Sdc1), Caspr4, and Neurexin. Short peptides and peptidomimetics can potentially inhibit or modulate its action and act as bioreagents or therapeutics. We used computational protein design (CPD) and molecular dynamics (MD) free energy simulations to understand and engineer its peptide specificity. CPD uses a structural model and an energy function to explore the space of sequences and structures and identify stable and functional protein or peptide variants. We used our in-house Proteus CPD package to completely redesign the Tiam1 PDZ domain. The designed sequences were similar to natural PDZ domains, with similarity and fold recognition scores comarable to the widely-used Rosetta CPD package. Selected sequences, containing around 60 mutated positions out of 90, were tested by microsecond MD simulations and biophysical experiments. Four of five sequences tested experimentally (by our collaborators) displayed reversible unfolding around 50°C. Proteus also accurately scored the binding specificity of several protein and peptide variants. As a more refined model for specificity, we parameterized a semi-empirical free energy model of the Poisson-Boltzmann Linear Interaction Energy or PB/LIE form, which scores conformations extracted from explicit solvent MD simulations of PDZ:peptide complexes. With three adjustable parameters, the model accurately reproduced the experimental binding affinities of 41 variants, with a mean unsigned error of just 0.4 kcal/mol, andgave predictions for 10 new variants. The PB/LIE model was tested further by comparing to non-empirical, alchemical, MD free energy simulations, which have no adjustable parameters and were found to give chemical accuracy for 12 Tiam1:peptide complexes. The tools and insights obtained should help discover new tight binding peptides or peptidomimetics and have broad implications for engineering PDZ:peptide interactions.

Les interactions protéine-protéine (IPPs) médient la signalisation cellulaire. Leur ingénierie peut fournir des informations et conduire au développement de molécules thérapeutiques. Les domaines PDZ sont des médiateurs importants de IPPs. Elles lient les 4--10 résidus C-terminaux de protéines cibles. Elles lient aussi les peptides correspondants, qui peuvent servir de systèmes modèles ou d'inhibiteurs. Nous avons développé deux approches computationnelles et les avons appliquées au domaine PDZ de la protéine Tiam1, un facteur d'échange pour la protéine Rac, impliqué dans la protrusion neuronale. Sa cible est la protéine Syndecan1. Des affinités expérimentales sont connues pour le peptide C-terminal, noté Sdc1, et plusieurs mutants; elles ont servi pour tester les calculs. Nous avons d'abord développé une méthode de dessin computationnel haut débit. Une simulation Monte Carlo est faite où les chaines latérales de la protéine et du peptide peuvent changer de conformères et certaines positions peuvent muter. Le solvant est implicite. Le paysage énergétique est aplati par la méthode adaptative de Wang-Landau, de sorte qu'un vaste ensemble de variantes est échantillonné. Effectuant des simulations distinctes du complexe et du peptide seul nous avons obtenu les énergies libres relatives d'association de 75,000 variantes en heure CPU sur une machine de bureau. Les valeurs sont compatibles avec les quelques données expérimentales disponibles. Ensuite, nous avons développé une approche beaucoup plus détaillée et réaliste. Soluté et solvant sont décrits par un champ de force atomique, qui représente explicitement la polarisation électronique: le champ de force Drude de Charmm. La polarisabilité peut être importante car les résidus de l'interface PDZ:peptide passent, lors de l'association, d'un environnement riche en solvant à un autre pauvre en solvant. Nous avons fait des simulations alchimiques d'énergie libre pour comparer quatre variantes du peptide qui diffèrent par une ou deux chaines latérales ioniques. Les résultats sont en bon accord avec l'expérience. Les champs de force additifs Charmm et Amber, qui représentent la polarisabilité implicitement, donnent un moins bon accord. Ces calculs sont le premier exemple de simulations alchimiques d'énergies libre d'association relatives protéine: ligand avec un champ de force polarisable. Enfin, pour une modélisation future de peptides phosphorylés, nous avons étendu le champ de force Drude pour inclure le méthyl phosphate et la phospho tyrosine. Il en résulte un excellent accord avec les affinités expérimentales phosphate: magnésium. / Protein-protein interactions (PPIs) regulate complex signaling networks in eukaryotic cells. Many binding events between several protein domains transfer information through communication pathways. Disrupting or altering the equilibrium between PPIs plays an important role inseveral diseases and the inibition of targeted PPIs is a recognized strategy for computational drug design. In the present thesis we focused on PDZ domains, which are among the most widespread signaling domains. PDZs recognize the 4-10 C-terminal amino acids of their target proteins as well as the corresponding peptides in isolation. We studied PDZ:peptide binding for the Tiam1 protein, which is a Rac GTP exchange factor involved in neuronal protrusion and axon guidance. Tiam1 activity modulates signaling for cell proliferation and migration, whose dysregulation increases growth of metastatic cancers. Its natural binder peptide is Syndecan1 (Sdc1), composed of 8 amino acids. Its last 5 Cter residues drive interactions in the binding pocket. Experimental affinities for several mutants of Sdc1 and in the protein domain constitute a complete dataset to study many ionic interactions with molecular simulations. These calculations are still challenging, despite the dramatic improvement of biomolecular modelling in the 1990's and 2000's. Upon binding, residues are transferred from a solvent-exposed environment to a solvent-poor one. This is expected to change the electron distribution within residues and nearby solvent molecules. Comparing ligands that differ by one or more ionic side-chain mutations, more sophisticated force fields where electronic polarizability is treated explicitly may be required. We developed and tested both Computational Protein Design (CPD) models and more precise free energy calculation methods based on polarizable molecular dynamics. We developed a general, high-througtput CPD protocol to optimize protein:peptide binding. The model has been implemented in on our in-house CPD package Proteus ( Simonson et al, 2014) and has been tested computing relative binding affinities for many variants of the Tiam1:Sdc1 complex. Monte Carlo sampling of equilibrium distributions of protein sequences is performed using an adaptive bias potential which flattens the energy landscape in sequence space and allows to estimate binding affinities for thousands of protein variants in limited CPU time (~1hour). We also improved our CPD implicit solvent model, implementing a more realistic description of the solute-solvent dielectric boundary. The new method, called Fluctuating Dielectric Boundary (FDB) showed a systematic improvement in the prediction of acid:base constants of several proteins. Promising results were also obtained for the complete sequence redesign of three PDZ domains. In the second part of this work we studied Tiam1:peptide affinities with more sophisticated models, based on free energy simulations with the Drude Polarizable Force field (DrudeFF). We first computed relative binding free energies for charge mutations in the Tiam1:Sdc1 complex, obtaining a clear improvement respect to equivalent calculations performed using two additive force fields. We applied the well-enstablished Dual Topology Approach: to our knowledge, this was the first example of such a calculation for a protein:peptide complex with uses the DrudeFF. Then we went on, developing the Drude polarizable models for methyl phosphate (MP) and phospho tyrosine (pTyr). We were interested in the change in binding affinity associated with phosphorylation of a Tyrosine residue of Sdc1, but Drude pTyr parameters were not yet developed. We tested our new phosphate parameters studying standard binding free energies between MP and magnesium (Mg2+) in water solution. Results showed a good agreement with experiment, improving previous calculations performed using additive force field