published_or_final_version / abstract / Chemistry / Doctoral / Doctor of Philosophy

published_or_final_version / Biological Sciences / Master / Master of Philosophy

published_or_final_version / Biochemistry / Master / Master of Philosophy

published_or_final_version / Paediatrics and Adolescent Medicine / Doctoral / Doctor of Philosophy

Non-competitive antagonists of the glutamate N-methyl-D-aspartate receptor (NMDAR) induce a broad range of schizophrenia-like symptoms in humans. Consequently hypothesis has emerged suggesting that glutamate or NMDAR hypofunction may occur in schizophrenia. The NMDAR is localised at dendritic spines of neurons and is embedded in a multi-protein complex called the post-synaptic density (PSD). The biochemical composition of the postsynaptic membrane and the structure of dendritic spines are continuously modulated by glutamatergic synaptic activity. The activity-dependent interaction between glutamate receptors and proteins of the PSD stimulate intracellular signalling pathways underlying learning and memory processes. These may be disturbed in schizophrenia. In the present study we hypothesised that molecules of the PSD may be disturbed in expression in the premotor cortex of patients with schizophrenia. Postmortem premotor cortex from patients with schizophrenia, major depressive disorder, bipolar disorder and healthy controls were processed for PSD extraction and purification. Protein expression of the PSD fraction was assessed using co-immunoprecipitation (co-IP) and Western blotting (WB) methods. The expression of NMDAR subunit NR2A, PSD-95, Ca2+/calmodulin-dependent protein kinase II subunit β (CaMKIIβ) and truncated isoform of the tropomyosin receptor kinase type B (TrkB-T1) were significantly reduced in schizophrenia. A significant decrease in the expression of NR2A was also observed in patients with major depressive disorder relative to controls. A decrease in the abundance of key PSD proteins in schizophrenia provides strong evidence that PSD function and possibly synaptic plasticity may be disturbed in the premotor cortex in the disease. There may also be more subtle disturbances in PSD function in major depressive disorder.

Successful bacterial circular chromosome segregation requires that any dimeric chromosomes, which arise by crossing over during homologous recombination, are converted to monomers. Resolution of dimers to monomers requires the action of the XerCD site-specific recombinase at dif in the chromosome replication terminus region. This reaction requires the DNA translocase, FtsK(C), which activates dimer resolution by catalysing an ATP hydrolysis-dependent switch in the catalytic state of the nucleoprotein recombination complex. We show that a 62-amino-acid fragment of FtsK(C) interacts directly with the XerD C-terminus in order to stimulate the cleavage by XerD of BSN, a dif-DNA suicide substrate containing a nick in the 'bottom' strand. The resulting recombinase-DNA covalent complex can undergo strand exchange with intact duplex dif in the absence of ATP. FtsK(C)-mediated stimulation of BSN cleavage by XerD requires synaptic complex formation. Mutational impairment of the XerD-FtsK(C) interaction leads to reduction in the in vitro stimulation of BSN cleavage by XerD and a concomitant deficiency in the resolution of chromosomal dimers at dif in vivo, although other XerD functions are not affected.

Protein interactions as mediated by catalytic or non-catalytic protein domains contribute to cellular signal transduction processes by covalent protein modification of or non-covalent binding to interaction partners. Ena/VASP homology 1 (EVH1) domains are found in different signal transduction proteins as N-terminal non-catalytic adaptor modules of ~ 115 amino acids sharing a common fold. By targeting their host proteins to subcellular sites of action they are involved in several signalling cascades which include protein phosphorylation and cytoskeletal reorganisation. In this study, protein interactions of the two EVH1 domain containing proteins VASP and Spred2 were studied according to their involvement in different and non-overlapping signal transduction pathways of the cell. EVH1 domains were first described in the Ena/VASP protein family with the Vasodilator-stimulated phosphoprotein VASP being its founding member. As a cytoskeleton-associated protein VASP not only interacts with different proteins of the actin network but it is also a substrate for cAMP- and cGMP-dependent protein kinases. However the full complement of protein kinases targeting VASP as their substrate is still unknown. Here we used mouse cardiac fibroblast (MCFB) cells in order to study the phosphorylation status of VASP and identify new candidate protein kinases involved after serum stimulation of these cells. Using phosphosite-specific antibodies we found that serum stimulation induces a phosphorylation of VASP at Ser-157 in a time-dependent manner reaching its maximum after 90 min of stimulation. We developed an interaction graph model of possible candidate protein kinases involved. Using a pharmacological perturbation analysis with different combinations of specific protein kinase inhibitors and activators we excluded any contribution of cGMP-dependent protein kinase and Rho kinases to this process and identified a combined action of classical isoforms of PKCs and PKA in serum-stimulated VASP phosphorylation at Ser-157 positioning PKC upstream of PKA in this signalling pathway. We hypothesise that PKC receives an external stimulatory signal upon serum stimulation of MCFB cells which is passed either directly or indirectly to PKA which finally phosphorylates VASP at Ser-157. A new EVH1 domain has been described recently in the Spred proteins (Sprouty related proteins containing an EVH1 domain) which are inhibitors of the Ras/Raf/MAP kinase pathway. Our laboratory has been involved in the elucidation of the atomic structure of the human Spred2 EVH1 domain by protein NMR spectroscopy (PDB 2JP2; 2007). A positively charged binding interface of this EVH1 domain suggests an interaction with negatively charged ligands; however no interaction partners of this domain have been described so far. In the second part of this study, we used different genetic and biochemical screening methods to search for ligands of the Spred2 EVH1 domain. A bacterial two-hybrid system was established using a physically well characterized interaction of the VASP EVH1 domain with a panel of its ActA binding peptides as positive controls to screen a human brain cDNA expression library at different stringencies for candidate Spred2 EVH1 interaction partners. However none of the clones isolated could be genetically and physically validated to support Spred2 EVH1 specific interactions. An in-vitro screening of a 9-mer phage display peptide library using purified GST-Spred2 EVH1 fusion protein was performed together with a Fyn-SH3 fusion protein as a positive control. In contrast to the Fyn-SH3 domain the majority of phages isolated with the Spred2 EVH1 domain either carried no inserts or inserts with stop codons suggesting a highly non-specific interaction of the phage coat protein with the latter domain but neither the Fyn-SH3 domain nor the GST moiety. Isolation of a 13-mer proline-rich sequence was particularly surprising in this context. In order to address possible interactions of the Spred2 EVH1 domain with non-peptidergic ligands protein-lipid interaction assays were performed. Quantitative binding studies to purified Spred2 EVH1 using a liposome sedimentation assay however excluded any interaction of candidate phospholipids of the phosphatidyl inositol phosphate class with the Spred2 EVH1 domain. A natively folded and thus binding-competent conformation of the purified proteins used was assessed independently by 1H protein NMR spectroscopy. In summary the cumulative evidence of our genetic and biochemical screening experiments suggests that the still elusive Spred2 EVH1 ligand(s) may be formed of hydrophobic peptide epitopes larger than nine amino acids in size and carrying negative charge(s). A phosphorylation of Spred2 EVH1 binding epitopes by a post-translational modification should be seriously considered in future experiments. / Proteininteraktionen, wie sie durch katalytisch oder nicht-katalytisch wirksame Proteindomänen vermittelt werden können, spielen eine wesentliche Rolle in zellulären Signaltransduktionsprozessen durch die kovalente Modifikation oder nicht-kovalente Bindung von Interaktionspartnern. Ena/VASP Homologie 1 (EVH1) Domänen finden sich als N-terminale, nicht-katalytische, etwa 115 Aminosäuren große und konserviert gefaltete Adaptormodule in vielen verschiedenen Signaltransduktionsproteinen. Indem sie ihre jeweiligen Wirtsproteine an deren subzellulärem Wirkort verankern helfen, sind sie an vielen verschiedenen Signalkaskaden wie z.B. Proteinphosphorylierungen oder Umbauprozessen des Zytoskeletts beteiligt. In dieser Arbeiten wurden Proteininteraktionen der beiden EVH1 domänen-haltigen Proteine VASP and Spred2 untersucht, die in nicht überlappenden Signaltransduktionswegen der Zelle vorkommen. EVH1 Domänen wurden zuerst innerhalb der Ena/VASP-Proteinfamilie beschrieben, deren Gründungsmitglied das Vasodilator-stimulierte Phosphoprotein VASP ist. Als zytoskelett-assoziiertes Protein wechselwirkt VASP nicht nur mit verschiedenen Aktin-bindenden Proteinen, sondern ist auch ein Substrat der cAMP- und cGMP-abhängigen Proteinkinasen. Der vollständige Satz jener Proteinkinasen, die VASP als eines ihrer Substrate aufweisen, ist immer noch unbekannt. Hier haben wir kardiale Mausfibroblasten (MCFB) Zellen verwendet, um nach Serum-Stimulation dieser Zellen den Phosphorylierungsstatus von VASP zu bestimmen und daran beteiligte, neue Kandidaten-Proteinkinasen zu identifizieren. Mit Hilfe von Phosphorylierungsstellen-spezifischen Antikörpern konnten wir zeigen, dass eine Serum-Stimulation eine zeitabhängige Phosphorylierung von VASP an Serin 157 induziert, die ein Maximum 90 min nach Stimulation erreicht. Wir entwickelten ein Interaktionsgraphen-Modell möglicher Kandidaten-Proteinkinasen, die an diesem Prozess beteiligt sein könnten. Mit Hilfe pharmakologischer Perturbationsexperimente auf der Grundlage spezifischer Proteinkinase-Inhibitoren und Aktivatoren konnten wir einerseits eine Beteiligung der löslichen cGMP-abhängigen Proteinkinase und von Rho-Kinasen an diesem Prozess ausschliessen und anderseits die gemeinsame Beteiligung der klassischen Proteinkinase C Isoform(en) und der cAMP-abhängigen Proteinkinase nachweisen. In diesem Signalweg liegt dabei die Proteinkinase C stromaufwärts vor letzterer. Nach unserer Interpretation der Daten wird die PKC nach Serum-Stimulation der MCFB-Zellen aktiviert und aktiviert ihrerseits direkt oder indirekt die cAMP-abhängige Proteinkinase, die schliesslich VASP als proximales Substrat am Serin 157 phosphoryliert. Eine neue EVH1 Domäne wurde kürzlich in den Spred Proteinen (Sprouty related proteins containing an EVH1 domain) beschrieben, die neue Inhibitoren im Ras/Raf/MAP-Kinase-Signalweg darstellen. Unser Labor war an der NMR-gestützten Aufklärung der atomaren Struktur der Spred2 EVH1 Domäne beteiligt (PDB 2JP2; 2007). Die positiv geladene Bindungsfurche dieser EVH1 Domäne legt eine Interaktion mit anionischen Liganden nahe. Interaktionspartner für diese Domäne sind bisher jedoch nicht beschrieben worden. Im zweiten Teil dieser Arbeit verwendeten wir verschiedene genetische und biochemische Suchverfahren zur Identifizierung möglicher Spred2 EVH1 Liganden. Ein bakterielles Two-Hybrid-System mit der Spred2 EVH1 Domäne als Köderprotein wurde dazu etabliert unter Verwendung der physikalisch gut charakterisierten Wechselwirkung der VASP EVH1 Domäne mit ihren ActA Bindungspeptiden als eines positiven Kontroll-Interaktionspaars und zum verschieden stringenten Durchmustern einer humanen cDNA Expressionsgenbank aus Gehirn eingesetzt. Keiner der isolierten Klone ließ sich jedoch genetisch oder nach Sequenzierung in Hinblick auf eine Spred2 EVH1 spezifische Wechselwirkung validieren. Mittels gereinigtem GST-Spred2 EVH1 Protein wurde daher eine 9-mer Peptid-Genbank im Phage-Display-Verfahren durchgemustert unter Verwendung eines Fyn-SH3 Fusionsproteins als positiver Kontrolle. Im Gegensatz zu den Ergebnissen mit letzterer trugen die mit der Spred2 EVH1 Domäne isolierten Phagen überwiegend keine Inserts oder solche mit Stop-Codons, was eine unspezifische Wechselwirkung mit den Phagen-Hüllenproteinen dieser Domäne nicht jedoch der Fyn-SH3 Domäne oder des GST-Partners nahelegt. Die Isolierung einer 13-mer großen prolin-reichen Bindesequenz war in diesem Zusammenhang besonders überraschend. Um eine mögliche Wechselwirkung von Spred2 EVH1 mit nicht-peptidergen Liganden zu untersuchen, wurden Protein-Phospholipid-Interaktionsassays durchgeführt. Mittels quantitativer Bindungsstudien unter Verwendung der isolierten Domäne konnte eine Interaktion mit Kandidaten-Phospholipiden aus der Klasse der Phosphatidylinositolphosphate in einem Liposomen-Sedimentationsassay ausgeschlossen werden. Eine native Faltung und damit prinzipiell bindungskompetente Konformation(en) der gereinigten Proteine konnten mittels 1H Protein-NMR-Spektroskopie sichergestellt werden. Zusammengenommen lassen unsere Experimente vermuten, dass es sich bei den noch immer nicht dingfest gemachten Spred2 EVH1 Liganden um hydrophobe, negative geladene, mehr als neun Aminosäuren umfassende Peptidepitope handeln könnte. Bei deren Identifizierung in zukünftigen Experimenten sollte mit ihrer Phosphorylierung durch post-translationale Modifikationen gerechnet werden.

A defining characteristic of neurodegenerative disease is the accumulation of mutant or misfolded proteins within neurons. Selective macroautophagy of aggregates, or aggrephagy, is a lysosome-mediated protein degradation pathway implicated in the turnover of disease-relevant accumulated proteins, but its specific function in vivo in the mammalian nervous system is poorly understood. The large PI3P-binding protein Alfy (Autophagy-linked FYVE protein) is an adaptor required for selective macroautophagy of aggregated proteins in cellular model systems. We sought to address Alfy-mediated aggrephagy in the mammalian brain in mouse models of Huntington’s disease (HD). HD is a neurodegenerative disorder caused by autosomal dominant inheritance of an expanded CAG repeat within the IT15, or huntingtin (htt) gene. The mutation causes an expansion of a polyglutamine (polyQ) tract in the protein Huntingtin (Htt), which results in psychiatric, cognitive, and motor symptomology. A pathological hallmark of HD is the accumulation of intracellular deposits of mutant Htt and ubiquitin. The exact relevance of these deposits remains unclear, but their elimination, hypothesized to occur via macroautophagy, correlates with behavioral improvements in mouse models of HD. The selective mechanisms of this phenomenon are largely unexplored in vivo. We have created two mouse models to address the role of Alfy-mediated selective macroautophagy in mammalian HD brain. First, we created tamoxifen-inducible Alfy knockout mice (Alfy iKO) and crossed them with a redesigned inducible HD mouse (HD103Q) that uses a tetracycline-regulated system to control reversible expression of mutant exon-1 Htt. Western blot, in situ, and PCR analysis confirm that Alfy can be eliminated from brain in adult Alfy iKO mice. A timecourse of Htt aggregation and clearance reveals that HD103Q mice accumulate huntingtin deposits, which clear in a linear manner upon transgene suppression over the course of four months. The loss of Alfy significantly impedes the removal of these deposits. Second, an Alfy knockout mouse was created using gene-trap technology, and mice hemizygous for Alfy knockout were crossed with BACHD mice expressing full-length human mutant Htt. We find that 50% Alfy depletion in the BACHD leads to increased insoluble Htt aggregate deposition along with accelerated decline in motor behavioral performance. Furthermore, inducible knockout of Alfy alone has a severe and age-dependent motor behavioral phenotype. This work reveals an in vivo role for Alfy in turnover of mutant Htt deposits, suggests that the accumulation of detergent-insoluble mutant Htt species contributes to behavioral pathogenesis, and supports an important function for Alfy at the intersection of HD and aging.

Knowledge of protein-protein interactions (PPIs) is essential to understanding regulatory processes in a cell. High-throughput experimental methods have made significant contributions to PPI determination, but they are known to have many false positives and fail to identify a signification portion of bona fide interactions. The same is true for the many computational tools that have been developed. Significantly, although protein structures provide atomic details of PPIs, they have had relatively little impact in large-scale PPI predictions and there has been only limited overlap between structural and systems biology. Here in this thesis, I present our progress in combining structural biology and systems biology in the context of studies analyzing, coarse-grained modeling and prediction of protein-protein interactions. I first report a comprehensive analysis of the degree to which the location of a protein interface is conserved in sets of proteins that share different levels of similarities. Our results show that while, in general, the interface conservation is most significant among close neighbors, it is still significant even for remote structural neighbors. Based on this finding, we designed PredUs, a method to predict protein interface simply by "mapping" the interface information from its structural neighbors (i.e., "templates") to the target structure. We developed the PredUs web server to predict protein interfaces using this "template-based" method and a support vector machine (SVM) to further improve predictions. The PredUs webserver outperforms other state-of-the-art methods that are typically based on amino acid properties in terms of both prediction precision and recall. Meanwhile, PredUs runs very fast and can be used to study protein interfaces in a high throughput fashion. Maybe more importantly, it is not sensitive to local conformational changes and small errors in structures and thus can be applied to predict interface of protein homology models, when experimental structures are not available. I then describe a novel structural modeling method that uses geometric relationships between protein structures, including both PDB structures and homology models, to accurately predict PPIs on a genome-wide scale. We applied the method with considerable success to both the yeast and the human genomes. We found that the accuracy and the coverage of our structure-based prediction compare favorably with the methods derived from sequence and functional clues, e.g. sequence similarity, co-expression, phylogenetic similarity, etc. Results further improve when using a naive Bayesian classifier to combine structural information with non-structural clues (PREPPI), yielding predictions of comparable quality to high-throughput experiments. Our data further suggests that PREPPI predictions are substantially complementary to those by experimental methods thus providing a way to dissect interactions that would be hard to identify on a purely high-throughput experimental basis. We have for the first time designed a "template-based" method that predicts protein interface with high precision and recall. We have also for the first time used 3D structure as part of the repertoire of experimental and computational information and find a way to accurately infer PPIs on a large scale. The success of PredUs and PREPPI can be attributed to the exploitation of both the information contained in imperfect models and the remote structure-function relationships between proteins that have been usually considered to be unrelated. Our results constitute a significant paradigm shift in both structural and systems biology and suggest that they can be integrated to an extent that has not been possible in the past.

This thesis focuses on an important aspect of protein functionality – protein-protein interactions (PPI). Three physical chemistry techniques for or derived from protein-protein interaction investigation are discussed. First, in Chapter 2, we demonstrate a new fluorescent imaging technique that creates high-order nonlinear signals by harnessing the frustrated fluorescence resonance energy transfer (FRET) – energy transfer between certain proteins close in proximity which is commonly used in PPI studies. In Chapter 3, we combine fluorescence resonance energy transfer (FRET) and bioluminescence resonance energy transfer (BRET), two most commonly used approaches to monitor protein-protein interactions in vivo, to create a novel hybrid strategy, bioluminescence assisted switching and fluorescence imaging (BASFI), which integrates the advantages of FRET and BRET. We demonstrate BASFI with Dronpa-RLuc8 fusion constructs and drug-inducible intermolecular FKBP-FRB protein-protein interactions in live cells with high sensitivity, resolution, and specificity. Finally, in Chapter 4, we propose a systematic free energy perturbation (FEP) protocol to computationally calculate the binding affinities between proteins. We demonstrate our protocol with the gp120 envelope glycoprotein of HIV-1 and three broadly neutralizing antibodies (bNAbs) of the VRC01 class and analyze antibody residues’ contributions to the binding which further provides insights for antibody design.