Thesis advisor: Mary F. Roberts / Amphitropic proteins, such as the virulence factor phosphatidylinositol-specific phospholipase C (PI-PLC) from Bacillus thuringiensis, often depend on lipid-specific recognition of target membranes. However, the recognition mechanisms for zwitterionic lipids such as phosphatidylcholine (PC), which is enriched in the outer leaflet of eukaryotic cell membranes, are not well understood. Molecular dynamics (MD) simulation and mutagenesis results strongly indicate that PI-PLC interacts with PC head groups via cation-π interactions with aromatic tyrosine residues, and suggest that cation-π interactions at the interface may be a mechanism for specific lipid recognition by amphitropic and membrane proteins. Aromatic amino acids can not only form cation-π interactions at the interface but also insert into membranes and have hydrophobic interactions with lipid tails. Heretofore there has been no facile way to differentiate these two types of interactions. We show that specific incorporation of fluorinated amino acids into proteins can experimentally distinguish cation-π interactions from membrane insertion of the aromatic side-chains. Fluorinated aromatic amino acids destabilize the cation-π interactions by altering electrostatics of the aromatic ring while their enhanced hydrophobicity enhances membrane insertion. Incorporation of pentafluorophenylalanine or difluorotyrosine into a Staphylococcus aureus phosphatidylinositol-specific phospholipase C (PI-PLC) variant engineered to contain a specific PC-binding site demonstrates the effectiveness of this methodology. Applying this methodology to the plethora of tyrosine residues in Bacillus thuringiensis PI-PLC identifies those involved in cation-π interactions with PC. Cation-π interactions provide a likely molecular mechanism for BtPI-PLC PC specificity but do not account for its preference for bilayers containing a small fraction of anionic lipids. MD simulations and fluorescence correlation spectroscopy (FCS) vesicle binding measurements of positively charged amino acids as well as surface tyrosine residues are used to formulate a complete model of BtPI-PLC specific binding to mixed anionic phospholipid/PC membrane. DNF2, a new plant protein with homology to bacterial PI-PLC, is confirmed to be the first plant small PI-PLC enzyme that can cleave both PI and glycosylphosphatidylinositol (GPI) anchored proteins. GPI-anchored protein cleavage also confirms that DNF2 plays an important role in symbiosome, the intracellular compartment formed by the plant that contains nitrogen fixing bacteria. / Thesis (PhD) — Boston College, 2015. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Chemistry.

Interactions between proteins and membranes are central to many signalling pathways and other cellular processes. Phosphatidylinositol phosphates (PIPs) are a family of lipids often acting as second messengers and targeted by peripheral proteins in these processes. A pipeline was developed combining the molecular dynamics (MD) approaches of umbrella sampling and coarse-grain modelling, and used to quantify and compare the interactions with PIP-containing model membranes of 13 pleckstrin homology (PH) domains, a common lipid-binding domain found in many proteins showing varied affinities and specificities for PIPs. Lipid selectivity generally agreed with previous observations. Several membrane-binding modes were identified, revealing PIP interactions through a secondary site are more common than suggested experimentally and appear to be related to overall affinity. Results suggest that simultaneous binding of multiple PIP lipids is required to achieve the high affinities characteristic of PH domains. Multiscale MD, combining coarse-grain binding simulations and atomistic refinement, was used to investigate PTEN, a tumour suppressor catalysing interconversion of PIPs and associated with many cancers and other disorders. Regions often ignored in previous studies were revealed to favour productive binding, largely via electrostatics. PIP clustering by bound PTEN and membrane insertion in the productive mode were demonstrated. Existence of an N-terminal PIP-binding site was supported, with this region appearing disordered, rather than helical as previously suggested. Changes in interdomain orientation when bound and with the clinically-relevant R173C mutation further suggest the importance of the interdomain interface for productive binding. Together, this work demonstrates the important contributions MD can make towards understanding protein/membrane interactions, particularly in the context of managing the diseases caused by their disruption.

A combined attenuated total reflection-Fourier-transform infrared spectroscopy (ATR-FTIR)-atomic force microscopy (AFM) platform was used to visualize and characterize membranolytic protein- and peptide-membrane interactions, allowing spectroscopic details to be correlated with structural features. Modifications to a previous combined platform permitted IR results for physiologically-relevant protein or peptide concentrations as well as provided nanometer-resolution height data for AFM. This combination provides greater insight than individual techniques alone. The interactions of hemolytic sticholysin proteins on a model red blood cell membrane showed evidence of conformational changes associated with a membrane-induced organization. In addition, the examination of a de novo cationic antimicrobial peptide on a model bacterial membrane showed that the peptide adopted a helical structure upon interaction with the membrane, and also provided evidence of membrane disruption and peptide aggregation. These results demonstrate that ATR-FTIR-AFM can be a powerful tool for understanding protein- and peptide-membrane interactions.

A combined attenuated total reflection-Fourier-transform infrared spectroscopy (ATR-FTIR)-atomic force microscopy (AFM) platform was used to visualize and characterize membranolytic protein- and peptide-membrane interactions, allowing spectroscopic details to be correlated with structural features. Modifications to a previous combined platform permitted IR results for physiologically-relevant protein or peptide concentrations as well as provided nanometer-resolution height data for AFM. This combination provides greater insight than individual techniques alone. The interactions of hemolytic sticholysin proteins on a model red blood cell membrane showed evidence of conformational changes associated with a membrane-induced organization. In addition, the examination of a de novo cationic antimicrobial peptide on a model bacterial membrane showed that the peptide adopted a helical structure upon interaction with the membrane, and also provided evidence of membrane disruption and peptide aggregation. These results demonstrate that ATR-FTIR-AFM can be a powerful tool for understanding protein- and peptide-membrane interactions.

Les protéines sont au coeur de la vie. Ce sont d'incroyables nanomachines moléculaires spécialisées et améliorées par des millions d'années d'évolution pour des fonctions bien définies dans la cellule. La structure des protéines, c'est-à-dire l'arrangement tridimensionnel de leurs atomes, est intimement liée à leurs fonctions. L'absence apparente de structure pour certaines protéines est aussi de plus en plus reconnue comme étant tout aussi cruciale. Les protéines amyloïdes en sont un exemple marquant : elles adoptent un ensemble de structures variées difficilement observables expérimentalement qui sont associées à des maladies neurodégénératives. Cette thèse, dans un premier temps, porte sur l'étude structurelle des protéines amyloïdes bêta-amyloïde (Alzheimer) et huntingtine (Huntington) lors de leur processus de repliement et d'auto-assemblage. Les résultats obtenus permettent de décrire avec une résolution atomique les interactions des ensembles structurels de ces deux protéines. Concernant la protéine bêta-amyloïde (AB), nos résultats identifient des différences structurelles significatives entre trois de ses formes physiologiques durant ses premières étapes d'auto-assemblage en environnement aqueux. Nous avons ensuite comparé ces résultats avec ceux obtenus au cours des dernières années par d'autres groupes de recherche avec des protocoles expérimentaux et de simulations variés. Des tendances claires émergent de notre comparaison quant à l'influence de la forme physiologique de AB sur son ensemble structurel durant ses premières étapes d'auto-assemblage. L'identification des propriétés structurelles différentes rationalise l'origine de leurs propriétés d'agrégation distinctes. Par ailleurs, l'identification des propriétés structurelles communes offrent des cibles potentielles pour des agents thérapeutiques empêchant la formation des oligomères responsables de la neurotoxicité. Concernant la protéine huntingtine, nous avons élucidé l'ensemble structurel de sa région fonctionnelle située à son N-terminal en environnement aqueux et membranaire. En accord avec les données expérimentales disponibles, nos résultats sur son repliement en environnement aqueux révèlent les interactions dominantes ainsi que l'influence sur celles-ci des régions adjacentes à la région fonctionnelle. Nous avons aussi caractérisé la stabilité et la croissance de structures nanotubulaires qui sont des candidats potentiels aux chemins d'auto-assemblage de la région amyloïde de huntingtine. Par ailleurs, nous avons également élaboré, avec un groupe d'expérimentateurs, un modèle détaillé illustrant les principales interactions responsables du rôle d'ancre membranaire de la région N-terminal, qui sert à contrôler la localisation de huntingtine dans la cellule. Dans un deuxième temps, cette thèse porte sur le raffinement d'un modèle gros-grain (sOPEP) et sur le développement d'un nouveau modèle tout-atome (aaOPEP) qui sont tous deux basés sur le champ de force gros-grain OPEP, couramment utilisé pour l'étude du repliement des protéines et de l'agrégation des protéines amyloïdes. L'optimisation de ces modèles a été effectuée dans le but d'améliorer les prédictions de novo de la structure de peptides par la méthode PEP-FOLD. Par ailleurs, les modèles OPEP, sOPEP et aaOPEP ont été inclus dans un nouveau code de dynamique moléculaire très flexible afin de grandement simplifier leurs développements futurs. / Proteins are at the center of life. They are formidable molecular nanomachines specialized and optimized during million years of evolution for well-defined functions in the cell. The structure of proteins, meaning the tridimensional setting of their atoms, is closely related to their function. Absence of structure for a subset of proteins is also recognized to be as crucial. Amyloid proteins is a striking example : they fold into an ensemble of various structures hardly observable experimentally that are associated with neurodegenerative diseases. This thesis, firstly, is on the study of the structural ensemble of the amyloid proteins amyloid-beta (Alzheimer) and huntingtin (Huntington) during their folding and aggregation. Our results describe in details, with an atomic resolution, the characteristic interactions present in the structural ensemble of these two proteins. Concerning the amyloid-beta protein (AB), our results show the structural differences between three of its physiological forms during its first aggregation steps in an aqueous environment. We have then compared these results with those obtained during the past few years by several other research groups using various experimental and simulation protocols. Clear trends come out of this comparison regarding the influence of AB physiological form on its structural ensemble during its first aggregation steps. Their distinct aggregation pathways are rationalized by the identified differences. For their part, the identified similarities offer targets for therapeutical compounds disrupting the aggregation of the neurotoxic oligomers. Concerning the huntingtin protein, we identify the structural ensemble of its functional region at its N-terminal in an aqueous environment and in a phospholipid membrane. In agreement with the available experimental results on the global structure of this region in aqueous solution, our results reveal the dominant interactions, at an atomic precision, in its structural ensemble as well as the influence of its neighboring regions. We have also characterized the stability and the growth of nanotube-like structures that could occur during the aggregation of the amyloid region of huntingtin. Moreover, we have developed, in collaboration with a group of experimentalists, a precise model describing the main membrane interactions of huntingtin N-terminal, which serves as a membrane anchor that controls the localization of huntingtin in the cell. Secondly, this thesis is on the refinement of a coarse-grained model (sOPEP) and on the development of a new all-atom model (aaOPEP) that are both based on the coarse-grained OPEP force field, commonly used to study protein folding and amyloid protein aggregation. The goal behind the optimization of these models is to improve the de novo structure prediction of the PEP-FOLD method. These three models -- OPEP, sOPEP and aaOPEP -- are now also implemented in a new molecular dynamics software that we have developed specifically to greatly ease their future developments.