Neurodegenerative Erkrankungen (NDs) wie Alzheimer (AD), Parkinson (PD), und amyotrophe lateral Sklerose (ALS) sind Hirnerkrankungen, die durch unlösliche Proteinaggregate in Neuronen oder im Extrazellularraum charakterisiert sind. In dieser Arbeit habe ich für verschiede bekannte und vorhergesagte neurodegenerative Krankheitsproteine (NDPs) Proteininteraktionsnetzwerke erstellt, um mögliche gemeinsame Krankheitsmechanismen genauer zu studieren. Mit Hilfe eines automatisierten Hefe-Zwei-Hybrid-Systems (Y2H) konnte ich 18.663 Protein-Protein-Interaktionen (PPIs) für 449 wildtyp und 22 mutierte Proteine identifizieren. Eine genaue funktionelle Analyse der Interaktionspartner von korrespondierenden wildtyp und mutierten Proteinen ergab deutliche Unterschiede zum einen im Fall von allen untersuchten Proteinen und insbesondere im Fall vom ALS Krankheitsprotein TDP-43. Die identifizierten PPIs wurden außerdem verwendet um krankheitsspezifische Netzwerke zu erstellen und um Proteine zu identifizieren, die mit mehreren NDPs verbunden sind. Ich habe auf diese Weise vier Proteine (APP, IQSEC1, ZNF179 und ZMAT2) gefunden, die mit bekannten NDPs with Huntingtin, TDP-43, Parkin und Ataxin-1 interagieren und so fünf verschiedene NDs miteinander verbinden. Die Reduktion der mRNA Expression von IQSEC1, ZNF179 oder ZMAT2 mit Hilfe von siRNA führte zu einer Verstärkung von pathogenen Mechanismen wie der Aggregation von mutiertem Huntingtin und TDP-43 sowie der Hyperphosphorylierung des Proteins Tau. Außerdem habe ich 22 Proteine entdeckt, die die Aggregation von TDP-43 deutlich verändern und außerdem Mitglieder in sieben vorhergesagten Proteinkomplexen sind. Die Proteinkomplexe habe ich durch Kombination von Interaktionsdaten und Daten eines siRNA Screenings vorhergesagt. Zusätzlich habe ich herausgefunden, dass die Proteine eines vorhergesagten Komplexes, nämlich HDAC1, pRB, HP1, BRG1 und c-MYC, die Aggregation von TDP-43 durch Veränderung von dessen Genexpression beeinflussen. / Neurodegenerative diseases (NDs) such as Alzheimer’s disease (AD), Parkinson’s disease (PD) or amyotrophic lateral sclerosis (ALS) are progressive brain disorders characterized by the accumulation of insoluble protein aggregates in neuronal cells or the extracellular space of patient brains. To elucidate potential common pathological mechanisms in different NDs, I created comprehensive interaction networks for various known and predicted neurodegenerative disease proteins (NDPs). I identified 18,663 protein-protein interactions (PPIs) for 449 bioinformatically selected wild-type target proteins and 22 mutant variants of 11 known NDPs by using an automated yeast two-hybrid (Y2H) system. The functional analysis of the interaction partners of corresponding wild-type and mutant NDPs revealed strong differences in the case of all 11 NDPs and especially for the ALS protein TDP-43. The identified PPIs were used to generate networks for individual NDs such as AD or PD and to identify proteins that are connected to multiple NDPs. For example, I found that five neurodegenerative diseases are connected by four proteins (APP, ZMAT2, ZNF179 and IQSEC1) that link known NDPs such as huntingtin, TDP-43, parkin, ataxin-1 and SOD1. Analysis of publicly available gene expression data suggested that the mRNA expression of the four proteins is abnormally altered in brains of ND patients. Moreover, the knock-down of IQSEC1, ZNF179 or ZMAT2 aggravates pathogenic disease mechanisms such as aggregation of mutant huntingtin or TDP-43 as well as hyperphosphorylation of tau. Additionally, I identified 22 modifiers of TDP-43 aggregation, which are members in 7 protein complexes. These complexes were predicted based on combined data from PPI as well as siRNA screenings. Finally, I found that the proteins HDAC1, pRB, HP1, BRG1 and c-MYC, which form one of the predicted complexes, influence TDP-43 aggregation by altering its mRNA expression.

Proteininteraktionen sind entscheidend für zelluläre Funktion. Interaktionen reflektieren direkte funktionale Beziehungen zwischen Proteinen. Veränderungen in spezifischen Interaktionsmustern tragen zur Entstehung von Krankheiten bei. In dieser Arbeit werden funktionale und pathologische Aspekte von Proteininteraktionen analysiert, um Funktionen für bisher nicht charakterisierte Proteine vorherzusagen und Proteine mit Krankheitsphänotypen zu assoziieren. Verschiedene Methoden wurden in den letzten Jahren entwickelt, die die funktionalen Eigenschaften von Proteinen untersuchen. Dennoch bleibt ein wesentlicher Teil der Proteine, insbesondere menschliche, uncharakterisiert. Wir haben eine Methode zur Vorhersage von Proteinfunktionen entwickelt, die auf Proteininteraktionsnetzwerken verschiedener Spezies beruht. Dieser Ansatz analysiert funktionale Module, die über evolutionär konservierte Prozesse definiert werden. In diesen Modulen werden Proteinfunktionen gemeinsam über Orthologiebeziehungen und Interaktionspartner vorhergesagt. Die Integration verschiedener funktionaler Ähnlichkeiten ermöglicht die Vorhersage neuer Proteinfunktionen mit hoher Genauigkeit und Abdeckung. Die Aufklärung von Krankheitsmechanismen ist wichtig, um ihre Entstehung zu verstehen und diagnostische und therapeutische Ansätze zu entwickeln. Wir stellen einen Ansatz für die Identifizierung krankheitsrelevanter Genprodukte vor, der auf der Kombination von Proteininteraktionen, Proteinfunktionen und Netzwerkzentralitätsanalyse basiert. Gegeben einer Krankheit, werden krankheitsspezifische Netzwerke durch die Integration von direkt und indirekt interagierender Genprodukte und funktionalen Informationen generiert. Proteine in diesen Netzwerken werden anhand ihrer Zentralität sortiert. Das Einbeziehen indirekter Interaktionen verbessert die Identifizierung von Krankheitsgenen deutlich. Die Verwendung von vorhergesagten Proteinfunktionen verbessert das Ranking von krankheitsrelevanten Proteinen. / Protein interactions are essential to many aspects of cellular function. On the one hand, they reflect direct functional relationships. On the other hand, alterations in protein interactions perturb natural cellular processes and contribute to diseases. In this thesis we analyze both the functional and the pathological aspect of protein interactions to infer novel protein function for uncharacterized proteins and to associate yet uncharacterized proteins with disease phenotypes, respectively. Different experimental and computational approaches have been developed in the past to investigate the basic characteristics of proteins systematically. Yet, a substantial fraction of proteins remains uncharacterized, particularly in human. We present a novel approach to predict protein function from protein interaction networks of multiple species. The key to our method is to study proteins within modules defined by evolutionary conserved processes, combining comparative cross-species genomics with functional linkage in interaction networks. We show that integrating different evidence of functional similarity allows to infer novel functions with high precision and a very good coverage. Elucidating the pathological mechanisms is important for understanding the onset of diseases and for developing diagnostic and therapeutic approaches. We introduce a network-based framework for identifying disease-related gene products by combining protein interaction data and protein function with network centrality analysis. Given a disease, we compile a disease-specific network by integrating directly and indirectly linked gene products using protein interaction and functional information. Proteins in this network are ranked based on their network centrality. We demonstrate that using indirect interactions significantly improves disease gene identification. Predicted functions, in turn, enhance the ranking of disease-relevant proteins.

Die meisten zellulären Prozesse werden durch Interaktionen zwischen Proteinen reguliert, weswegen die Charakterisierung dieser Interaktionen zu den wichtigsten Zielen der Proteomik gehört. Allerdings sind experimentelle Verfahren zur Detektion von Proteininteraktionen mit hohen Fehlerraten assoziiert und die Bedingungen unter denen die Interaktionen gemessen werden sind zu einem gewissen Grad artifiziell. Wir implementieren eine Anwendung, die menschliche Proteininteraktionsdaten aus von Experten gepflegten Datenbanken integriert. Um die hohen Fehlerraten von experimentell detektierten Proteininteraktionen zu adressieren, entwickeln wir eine Funktion, die sowohl computergestützt als auch von Experten dahingehend optimiert wird, Menge und Qualität der Evidenz einer Proteininteraktion zu bewerten. Um das Problem der fehlenden Kontextinformationen zu beheben, entwickeln wir eine Methode, die Interaktionsannotationen von verschiedenen Attributen der interagierenden Proteine ableitet. Wir benutzen die kontextspezifischen Netzwerke, um Proteininteraktionen zu identifizieren, die vermutlich eine Rolle in Krankheiten spielen. Schliesslich verwenden wir das integrierte humane Netzwerk interagierender Proteine für die Untersuchung der Wildtyp-Funktion von Polyglutaminketten. Expansionen dieser Ketten wurden mit verschiedenen neurodegenerativen Erkrankungen (wie zum Beispiel Chorea Huntington) assoziiert. Allerdings sind Polyglutaminketten normaler Bestandteil vieler menschlicher Proteine, was suggeriert, dass diese Ketten eine wichtige zelluläre Funktion haben. Um Hinweise auf eine solche Funktion in biologischen Systemen zu sammeln, untersuchen wir die Charakteristika von Proteinen mit Polyglutaminketten in Interaktionsnetzwerken und Eigenschaften der Ketten auf Nukleotid-, Protein- und Organismen-Ebene. Zusammengenommen legen unsere Beobachtungen nahe, dass Polyglutaminketten Interaktionen zwischen Proteinen stabilisieren. / Protein-protein interactions (PPIs) regulate many cellular functions. Therefore, characterizing the entire human interactome is a key effort in current proteomics research. However, the experimental reliability of the techniques used to detect PPIs can have widely different quality with some methods being associated with high error rates. Another problem of PPI detection methods is that many interactions are measured under artificial conditions. We implement a resource that integrates human PPI data from the major expert-curated PPI databases. To address the high uncertainty associated with experimentally detected PPIs, we develop a scoring scheme that has been optimized both computationally and by human experts to reflect the amount and quality of evidence for a given PPI. To deal with the problem of missing context, we develop a method that assigns information to PPIs inferred from various attributes of the interacting proteins. We use these context-specific networks to identify PPIs that likely play a role in disease. Finally, we use the integrated human PPI network for the study of the wild type function of polyglutamine (polyQ) stretches. Expansions of these stretches have been observed in the proteins of a large number of patients with different neurodegenerative diseases such as Huntington''s. However, polyQ tracts are a normal feature of many human proteins, suggesting that they have an important cellular function. To clarify the potential function of polyQ repeats in biological systems, we study the characteristics of polyQ-containing proteins in the human PPI network. We complement the network analysis studying the repeats at nucleotide, protein and organism level. Together, our observations suggest that polyQ tracts in proteins stabilize protein interactions.

Spinal Muscular Atrophy (SMA) is a neurodegenerative, inherited disease caused by an insufficient amount of functional Survival of Motor Neurone protein (SMN), though the exact mechanism underlying this is not fully understood. The primary function of SMN is assembling a ring of Sm proteins around small nuclear RNA (snRNA) in an early, cytoplasmic stage of small nuclear ribonucleoprotein (snRNP) biogenesis, a process essential in eukaryotes. SMN, together with several mRNA binding proteins, has been linked to neural transport of mRNA towards areas of growth in Motor neurons for local translation of transcripts. Previous research in our group has found that this may involve Coatomer protein-containing vesicles transported by Dynein and requiring the Sm family protein, SmB, for maintenance. Little is known, however, about what other proteins are also present and required for correct transport and localisation of these vesicles. To further investigate this, we have produced plasmids expressing each Sm protein tagged to fluorescent proteins to help track their behaviour, in some cases for the first time, and developed a detergent-free fractionation protocol to enrich for SMN containing vesicles, providing tools that can be used to further probe behaviour and interactions in the future. Using these approaches, SmN, a neural specific Sm protein, was identified to also be present in SMN-containing vesicles similarly to SmB. Analysis of the interactomes of different Sm proteins identified a novel interactor of SMN, Neurochondrin (NCDN), that appears to be required for the correct localisation of SMN in neural cells. NCDN was found to not associate with snRNPs, indicating an snRNP-independent interaction with SMN. NCDN and SMN both independently associated and co-enriched with Rab5, indicating a potential endocytic and cell polarity role for the interaction. This interaction has the potential to be key in SMA pathology and may have therapeutic potential.

Proteins are synthesised as linear polymeric chains. The subtle energetic interplay of interatomic interactions results in chain folding, through which proteins may acquire defined structures. This spatial organisation is encoded by the protein sequence itself; the so-called thermodynamic hypothesis formulated by Anfinsen in 1961. A defined structure is often considered a pre-requisite to protein function, but widespread existence of intrinsically disordered proteins (IDPs) has prompted a re- evaluation of the ways biological function may be encoded into polypeptide chains. Furthermore, proteins often exist as part of multi-component entities, where regulation of assembly is integral to their properties. The interplay between disorder, oligomerisation and function is the focus of this thesis. Some IDPs fold conditionally upon interacting with a partner protein; a process known as coupled folding and binding. What are the biophysical advantages and consequences of disorder in the context of these interactions? A common feature of IDPs is their sequence composition bias, with charged residues being often over-represented. It is therefore tempting to speculate that electrostatic interactions may play a major role in coupled folding and binding reactions. Surprisingly, the opposite was found to be true. Charge-charge interactions only contributed about an order of magnitude to the association rate constants of two contrasting model systems. The lack of pre-formed binding interfaces-a consequence of disorder-might preclude electrostatic acceleration from complementary patches. By looking at the role of the sequence, many studies have taken a protein-centric approach to understanding disorder. Yet there is paucity of data about the effect of extrinsic factors on interactions involving disordered partners. Investigating the role of co-solutes, it was discovered that the kinetic and thermodynamic profiles of coupled folding and binding reactions were sensitive to ion-types. This effect followed the Hofmeister series, and occurred at physiological concentrations of salt. The sensitivity of coupled folding and binding reactions-a consequence of the lack of stability of IDPs-might be advantageous. Given the role of ions in biology, this 'biophysical sensing' could be a mechanism of physiological relevance, allowing modulation of protein-protein interactions involving disordered partners in response to changes in their environments. In cells, signalling networks are often multi-layered, and involve competing protein-protein interactions. The interplay between the biophysical characteristics of the components, and the behaviour of the network were investigated in a model tripartite system composed of folded and disordered proteins. The BCL-2 family regulates the intrinsic pathway of apoptosis through control of mitochondrial outer-membrane permeabilisation; a result of BAK and BAX oligomerisation. Through a shared homology motif (termed BH3), the subtle balance of their interactions determines cellular fate at the molecular level. Characterisation of the model under simple biochemical conditions revealed large differences in affinities among binary interactions; the consequence of the lifetime of the complexes, not their speed of association. A membrane-like environment, re-created using detergents, allows the oligomerisation of BAK and BAX in vitro. Furthermore, investigation of the tripartite system under detergent conditions showed that regulation of the network was the result of competing hetero- and homo-oligomerisation events. Relationships to their biophysical properties were gained by probing their energy landscapes using protein folding techniques. The connection between the biophysical properties of the components of the network and their interactions provides a molecular explanation for the regulation of apoptosis. This thesis offers insights into the ways structured assemblies and environmentally responsive disorder elements may encode functions into proteins.

Au cours des dernières années, le développement de nouvelles thérapies contre les infections bactériennes a suscité un grand intérêt face à l’émergence des nombreuses souches résistantes aux antibiotiques. Le point de départ de cette recherche de nouveaux antibiotiques, pour lesquels les bactéries n’ont pas encore acquis de mécanismes de résistance, est l’identification de nouvelles cibles cellulaires. En 2000, des études génétiques ont identifié engA, un gène bactérien dont le produit est une GTPase, comme une cible pharmacologique pertinente: elle est essentielle à la survie cellulaire, conservée au sein des bactéries et absente chez les eucaryotes.Puisque EngA agit comme un facteur d’assemblage pour le ribosome bactérien, un de nos objectifs a été de développer un test de criblage pour identifier des inhibiteurs des interactions EngA-ribosome. Ces interactions sont modulées par des changements conformationnels d'EngA, qui sont eux-mêmes déclenchés par la fixation de différents nucléotides dans le domaine catalytique. Cependant, les liens entre ces différents changements restent encore méconnus. Nous avons utilisé une approche multi-technique pour étudier ces questions et obtenir des informations utiles pour l’optimisation de notre test de criblage.Des analyses de SAXS et protéolyse limitée ont démontré un changement conformationnel en solution après adition de nucléotides di- ou tri-phosphate. La comparaison des données avec des modèles cristallographiques d'EngA a confirmé la conformation de la protéine liée au GDP. Cependant, la conformation de la protéine liée au GTP ne correspond à aucune structure connue. Des essais d’interaction ont démontré que la fixation de différents nucléotides au niveau des domaines catalytiques régule l’interaction d'EngA avec le ribosome. En outre, les effets des nucléotides se produisent en utilisant des fortes concentrations, ce qui suggère que le rôle d'EngA dans la biogenèse du ribosome peut être contrôlé par la concentration intracellulaire de nucléotides. Les travaux visant la détermination de la structure d'EngA dans sa conformation liée au GTP par cristallographie nous ont permis d’obtenir la structure d’EngA dans différentes formes cristallines. Cependant, ces structures représentent la conformation liée au GDP. L’analyse de l’empilement des cristaux a montré des contacts intermoléculaires conservés qui peuvent stabiliser cette conformation pendant la nucléation. Des mutations spécifiques permettant la rupture de ces contacts peuvent éventuellement aider à promouvoir la cristallisation de conformations alternatives. Des analyses de cryo-microscopie électronique ont débuté afin d’obtenir la structure du complexe EngA:50S de chez B. subtilis. Des résultats préliminaires montrent une carte de densité électronique à 6.4 Å de résolution. L’interprétation de ces résultats est en cours. / The development of new therapeutics against bacterial infections has aroused great interest over the last years in the context of drug resistance. The starting-point in the pursuit of new antibiotics for which bacterial resistance mechanisms do not exist is the identification of novel cellular targets. Genetics studies in the early 2000s have identified engA as a conserved bacterial gene whose product is a GTPase that could represent a potential drug target: it is conserved among bacteria, essential for cell survival, and absent in humans.Since EngA acts as an assembly factor for the bacterial ribosome, one of our aims was to develop an assay to screen inhibitors of the EngA-ribosome interactions. These interactions are modulated by EngA conformational changes that are in turn triggered by the binding of different nucleotides to the catalytic G-domain. As the interplay between all these events in bacteria is still not resolved, we have used a multi-technique approach to explore these questions in order to obtain useful information for the setting up of a robust screening assay.SAXS and limited proteolysis showed a conformational change occurring in solution upon addition of either di- or tri-phosphate nucleotides. While model validation analysis confirmed the GDP-bound conformation, the GTP-bound state does not match any known EngA structure. Binding studies have revealed modulation of interactions by different nucleotide-bound states. Furthermore, response to nucleotides occurs at high concentrations, suggesting that the role of EngA in promoting ribosome assembly could be monitored by the intracellular nucleotide concentration. Efforts on identifying the GTP-bound state 3D structure by crystallography have resulted in EngA structures in different crystal forms. Although all the obtained structures represent the GDP-bound state, packing analysis has revealed conserved crystal contacts that can potentially stabilise this conformation during nucleation. Specific mutations aiming at disrupting these contacts may help to promote crystallisation of alternative conformations. Cryo-EM investigation has been initiated in order to obtain the structure of the B. subtilis EngA:50S complex. So far, an electron density map at 6.4 Å resolution has been obtained and its interpretation is underway.

With the advent of large scale genome sequencing efforts along with more sophisticated methods of genetic mapping, a number of loci have been identified that are associated with human diseases. Intriguingly, many genes identified in these loci remain uncharacterized. Although current annotation can provide a prediction of putative function of some of these proteins at a biochemical level, understanding their cellular roles require analysis at a single gene level. Bioinformatic analysis carried out in the laboratory during studies on cyclic nucleotide metabolism in mycobacteria identified putative Class III cyclic nucleotide phosphodiesterases (Class III cNMP PDEs) from the non-redundant database of proteins. One of the proteins identified was the Rv0805 gene product from Mycobacterium tuberculosis. Detailed biochemical characterization of this protein revealed that Rv0805 is indeed a phosphodiesterase (PDE) and could hydrolyze 3’, 5’-cyclic adenosine monophosphate (cAMP) as well as 3’, 5’-cyclic guanosine monophosphate (cGMP). Structural analysis of Rv0805 revealed a metallophosphoesterase (MPE) like fold and presence of two metal atoms at the binuclear metal centre of the protein. Moreover, overexpression of Rv0805 in E. coli and M. smegmatis reduced intracellular cAMP levels indicating that it possesses cAMP PDE activity in vivo. The majority of proteins identified in this bioinformatic analysis were of bacterial or archaebacterial in origin but it was interesting to find some mammalian proteins, since, till date, no Class III cNMP PDE has been found in higher eukaryotes. Interestingly, two genes were identified in the human genome. These genes, 239FB and 239AB, are expressed in the fetal brain and adult brain, respectively and have been annotated as metallophosphoesterases but there has been no biochemical or functional characterization of these proteins. The 239FB gene is present between the FSHB and PAX6 genes on chromosome 11. This gene locus is present within a deletion interval (11p13-14) that is associated with the mental retardation phenotype of WAGR syndrome (Wilms’ tumor, aniridia, genitourinary anomalies, mental retardation). Inspection of available sequenced mammalian genomes indicated a shared synteny of the genes in the WAGR locus, highlighting it’s evolutionary conservation. Most interestingly, nucleotide sequences within the WAGR locus (which include the 5 genes WT1, PAX6, RCN1, ELP4 and 239FB) are amongst the 481 ultra conserved regions of the human genome. Moreover, 239FB is one of only 24 instances where an ortholog of an ultra-conserved element could be partially traced back by sequence similarity in lower eukaryotes such as Ciona intestinalis, Drosophila melanogaster, or Caenorhabditis elegans. Although the function of the 239FB protein is unknown so far, the distinctive expression of the gene in the fetal brain and the presence of an “ancient conserved region” in this gene suggest that this gene may be vital for the development of the nervous system. The work carried out in this thesis has attempted to understand the physiological functions of the 239FB/AB gene family. Amino acid sequence comparison revealed two amino acids changes between the human and rat proteins indicating the extra-ordinary sequence conservation of these proteins. Therefore, to characterize the biochemical properties of 239FB and 239AB proteins, rat proteins were used as model enzymes. Reverse transcription-PCR analysis of RNA prepared from the fetal and adult rat brains as well as Western blot analysis on cytosolic fractions of rat brains from various developmental stages indicated that 239FB is predominantly expressed in fetal brain. Detailed biochemical analyses of the rat 239FB and 239AB proteins were performed which showed that they possess metallophosphodiesterase activity. 239FB showed activity only in the presence of Mn2+ and Co2+ as the added metal cofactors. Surprisingly, the Km for Mn2+ of 239FB was found to be 1.5 mM, which is nearly 60-fold higher than that of its mycobacterial ortholog, Rv0805. A systematic mutational analysis was performed to characterize the residues that are involved in binding either one or both the metals found in the catalytic site of 239FB. Although 239FB shares some of the residues that have been shown to be essential for metal binding and catalytic activity with other MPEs including Rv0805, there are some differences as well. One histidine residue that has been conserved in other MPEs and has been shown to be important for metal binding is replaced by glycine (Gly-252) in 239FB. To study the consequence of replacing the glycine with a histidine in 239FB, a 239FBGly252His mutant protein was generated and characterized. Interestingly, the single mutation of Gly-252 to a histidine residue not only increased the affinity of the protein for metals but increased catalytic activity as well with various phosphodiesters. Moreover, 239FBGly252His mutant protein showed significant activity with cAMP and cGMP which were not hydrolysed by wild type 239FB. Interestingly, in the 239AB protein, histidine 284 is present at a position equivalent to Gly-252 in the 239FB protein. Biochemical characterization of 239AB showed 2’, 3’-cAMP hydrolyzing activity similar to 239FBGly252His mutant protein. A rat 239FB protein with a mutation (His67Arg) corresponding to a single nucleotide polymorphism seen in human 239FB, led to complete inactivation of the protein. The occurrence of this SNP at a very low frequency and only as a heterozygous condition suggests that a complete loss-of-function mutation of 239FB in human populations cannot be tolerated. To gain insights into the function of 239FB in its physiological milieu, yeast two-hybrid screening was performed with 239FB using human fetal brain cDNA library. Dipeptidyl peptidase III, a zinc dependent metallopeptidase, was found as an interacting partner of 239FB in this analysis and the functional consequences of this interaction would be an interesting area of study in future. While a number of metallophosphoesterases have been characterized biochemically and structurally, their biological role(s) and in vivo substrate(s) remain elusive. In order to elucidate the physiological role of 239FB/AB family, the ortholog of 239FB/AB in D. melanogaster was characterized. Sequence comparison of Drosophila ortholog with both the mammalian proteins indicated that it may be an ortholog of 239AB and hence, it was named as d239AB. Enhancer-promoter analysis with a putative promoter region of the d239AB indicated the expression of d239AB in the mushroom bodies in brain and in enterocytes in mid gut. Characterization of a Drosophila line, BS#16242, with a piggybac element inserted in the intron of d239AB showed disruption of d239AB expression. This suggested that BS#16242 line can serve as a d239AB knockout line and hence, was selected for further phenotypic characterization to unravel the physiological roles of d239AB. Though, BS#16242 flies did not show any developmental defects, a severe reduction in the fecundity of these files was observed. Further analysis revealed defective ovulation as a probable reason for reduced fecundity of these flies. In addition to compromised fecundity, BS#16242 flies showed a significant reduction in the life span of male as well as female flies. Moreover, these flies showed less resistance to thermal stress and desiccation. Most interestingly, all these phenotypes were rescued upon neuronal expression of the d239AB transgene in BS#16242 flies indicating that neuronal function of d239AB is important for diverse physiological processes. The phenotypes observed in BS#16242 flies mimic the physiological state under increased insulin signaling, such as decrease in life span, and susceptibility to various stress conditions suggesting that d239AB could play a role in the insulin signaling pathway. Interestingly, overexpression of d239AB transgene in neurons reduced cAMP levels in the brains of Drosophila, indicating that the protein may have cAMP phosphodiesterase activity in vivo. This is the first analysis of the presence of a Class III phosphodiesterase in eukaryotes. Thus, d239AB mediated regulation of cAMP levels in a particular subsets of cells, such as neurons, could also be one of the molecular mechanisms responsible for reduced fecundity and longevity of BS#16242 flies. Interacting partners of d239AB were inspected in the Drosophila interactome (built on protein-protein interactions identified using a yeast two-hybrid approach). Strikingly, most of the d239AB interacting proteins were involved either in transcriptional or translational regulation indicating that d239AB could be involved in the regulation of expression of genes involved in diverse physiological processes. This could explain why disruption of d239AB led to various physiological defects such as reduced fecundity, decreased life span and compromised fitness. In summary, studies described in this thesis suggest that 239FB and 239AB proteins are the first Class III cyclic nucleotide phosphodiesterases reported in eukaryotes. Results shown here suggest the critical role of their ortholog in the physiology of Drosophila. Further genetic manipulation in D. melanogaster and other organisms which harbor orthologs of the 239FB/AB gene could throw light on the diverse biological roles of these enzymes in humans.

Protein-protein interactions (PPIs) are attractive targets because of their therapeutic potential. One approach to design small molecules that can disrupt the PPIs is to use structural information of proteins. With this approach, triazole-based peptidomimetics that mimic beta-turn hot-spot regions in neurotrophins were synthesized. The monovalent mimics were assembled into bivalent mimics via a combinatorial method. Three different bivalent mimics were prepared for different studies. Bivalent mimics with long-linkers bound to TrkA or TrkC receptor and showed partial antagonism for the receptors. Other mimics were conjugated with cytotoxic compounds and they were used for TrkC targeted drug delivery. The last group of bivalent mimics previously showed targeted delivery effects for pancreatic cancer cells. In this study, we synthesized Eu-chelated bivalent mimics to perform a competitive binding assay for pancreatic cancer cells. Previous research in our group focused on design of secondary structures' mimics on rigid scaffolds as "minimalist mimics." We sought to establish structural design criteria for the minimalist mimics, and we wanted to propose that sets of such compounds could mimic local pairs of amino acids in any secondary structures as "universal peptidomimetics." Thus, we designed five compounds, such as oxazoline-, pyrrole-, dyine- "kinked" and "linear" bistrizole-based peptidomimetics, and performed molecular modelings, DFT calculations, and QMD for them to validate our hypothesis. On the concepts of "minimalist mimics" and "universal peptidomimetics," we developed the C alpha ? C beta vector matching program to evaluate preferred orientations of C alpha - C beta coordinates for secondary structures. We applied the program to omegatides and pyrrolinone-pyrrolidine oligomers. The compounds matched better with strands than for helices. We expanded the C alpha ? C beta vector matching idea to a method that ranks preferred conformations of small molecules on any combination of three interface side-chains in all structurally characterized PPIs. We developed a PDB mining program (explores key orientation, EKO) to do this, and EKO applied to pyrrolinone-pyrrolidine oligomers to find targets. EKO found several interesting targets, such as AICAR Tfase, GAPDH, and HIV-1 protease. HIV-1 dimerization inhibition and Zhang-Poorman kinetic assays were performed to validate our hypothesis, and the results showed that pyrrolinone-pyrrolidine derivatives inhibited HIV-1 dimerization.

BACKGROUND: A long-standing challenge in the post-genomic era of Bioinformatics is the prediction of protein-protein interactions, and ultimately the prediction of protein functions. The problem is intrinsically harder, when only amino acid sequences are available, but a solution is more universally applicable. So far, the problem of uncovering protein-protein interactions has been addressed in a variety of ways, both experimentally and computationally. MOTIVATION: The central problem is: How can protein complexes with solved threedimensional structure be utilized to identify and classify protein binding sites and how can knowledge be inferred from this classification such that protein interactions can be predicted for proteins without solved structure? The underlying hypothesis is that protein binding sites are often restricted to a small number of residues, which additionally often are well-conserved in order to maintain an interaction. Therefore, the signal-to-noise ratio in binding sites is expected to be higher than in other parts of the surface. This enables binding site detection in unknown proteins, when homology based annotation transfer fails. APPROACH: The problem is addressed by first investigating how geometrical aspects of domain-domain associations can lead to a rigorous structural classification of the multitude of protein interface types. The interface types are explored with respect to two aspects: First, how do interface types with one-sided homology reveal convergently evolved motifs? Second, how can sequential descriptors for local structural features be derived from the interface type classification? Then, the use of sequential representations for binding sites in order to predict protein interactions is investigated. The underlying algorithms are based on machine learning techniques, in particular Hidden Markov Models. RESULTS: This work includes a novel approach to a comprehensive geometrical classification of domain interfaces. Alternative structural domain associations are found for 40% of all family-family interactions. Evaluation of the classification algorithm on a hand-curated set of interfaces yielded a precision of 83% and a recall of 95%. For the first time, a systematic screen of convergently evolved motifs in 102.000 protein-protein interactions with structural information is derived. With respect to this dataset, all cases related to viral mimicry of human interface bindings are identified. Finally, a library of 740 motif descriptors for binding site recognition - encoded as Hidden Markov Models - is generated and cross-validated. Tests for the significance of motifs are provided. The usefulness of descriptors for protein-ligand binding sites is demonstrated for the case of "ATP-binding", where a precision of 89% is achieved, thus outperforming comparable motifs from PROSITE. In particular, a novel descriptor for a P-loop variant has been used to identify ATP-binding sites in 60 protein sequences that have not been annotated before by existing motif databases.

<p>Proteins with highly specific binding properties constitute the basis for many important applications in biotechnology and medicine. Immunoglobulins have so far been the obvious choice but recent advances in protein engineering have provided several novel constructs that indeed challenge antibodies. One class of such binding proteins is based on the 58 residues three-helix bundle Z domain from staphylococcal protein A (SPA). These so-called affibodies are selected from libraries containing Z domain variants with 13 randomised positions at the immunoglobulin Fc-binding surface. This thesis aims to describe the principles for molecular recognition in two protein-protein complexes involving affibody proteins. The first complex is formed by the Z_SPA-1 affibody binding to its own ancestor, the Z domain (Kd ~1 μM). The second complex consists of two affibodies: Z_Taq, originally selected to bind Taq DNA polymerase, and anti-Z_Taq, an anti-idiotypic binder to Z_Taq with a Kd ~0.1 μM. The basis for the study is the determination of the three-dimensional structures using NMR spectroscopy supported by biophysical characterization of the uncomplexed proteins and investigation of binding thermodynamics using isothermal titration calorimetry. The free Z_SPA-1 affibody is a molten globule-like protein with reduced stability compared to the original scaffold. However, upon target binding it folds into a well-defined structure with an interface topology resembling that displayed by the immunoglobulin Fc fragment when bound to the Z domain. At the same time, structural rearrangements occur in the Z domain in a similar way as in the Fc-binding process. The complex interface buries 1632 Ų total surface area and 10 out of 13 varied residues in Z_SPA-1 are directly involved in inter-molecular contacts. Further characterization of the molten globule state of Z_SPA-1 revealed a native-like overall structure with increased dynamics in the randomised regions (helices 1 and 2). These features were reduced when replacing some of the mutated residues with the corresponding wild-type Z domain residues. The nature of the free Z_SPA-1 affects the thermodynamics of the complex formation. The contribution from the unfolding equilibrium of the molten globule was successfully separated from the binding thermodynamics. Further decomposition of the binding entropy suggests that the conformational entropy penalty associated with stabilizing the molten globule state of Z_SPA-1 upon binding seriously reduces the binding affinity. The Z_Taq:anti-Z_Taq complex buries in total 1672 Ų surface area and all varied positions in anti-Z_Taq are directly involved in binding. The main differences between the Z:Z_SPA-1 and the Z_Taq:anti-Z_Taq complexes are the relative subunit orientation and certain specific interactions. However, there are also similarities, such as the hydrophobic interface character and the role of certain key residues, which are also found in the SPA:Fc interaction. Structural rearrangements upon binding are also common features of these complexes. Even though neither Z_Taq nor anti-Z_Taq shows the molten globule behaviour seen for Z_SPA-1, there are indications of dynamic events that might affect the binding affinity. This study provides not only a molecular basis for affibody-target recognition, but also contributions to the understanding of the mechanisms regulating protein stability and protein-protein interactions in general.</p>