TY - THES A1 - Calzadiaz Ramirez, Liliana T1 - Engineering highly efficient NADP-dependent formate dehydrogenases using a NADPH biosensor Escherichia coli strain N2 - NADPH is an essential cofactor that drives biosynthetic reactions in all living organisms. It is a reducing agent and thus electron donor of anabolic reactions that produce major cellular components as well as many products in biotechnology. Indeed, the engineering of metabolic pathways for the production of many products is often limited by the availability of NADPH. One common strategy to address this issue is to swap cofactor specificity from NADH to NADPH of enzymes. However, this process is time consuming and challenging because multiple parameters need to be engineered in parallel. Therefore, the first aim of this project is to establish an efficient metabolic biosensor to select enzymes that can reduce NADP+. An NADPH auxotroph strain was constructed by deleting major reactions involved in NADPH biosynthesis in E. coli’s central carbon metabolism with the exception of 6-phosphogluconate dehydrogenase. To validate this strain, two enzymes were tested in the presence of several carbon sources: a dihydrolipoamide dehydrogenase variant of E. coli harboring seven mutations and a formate dehydrogenase (FDH) from Mycobacterium vaccae N10 harboring four mutations were found to support NADPH biosynthesis and growth. The strain was subjected to adaptive laboratory evolution with the goal of testing its robustness under different carbon sources. Our evolution experiment resulted in the random mutagenesis of the malic enzyme (maeA), enabling it to produce NADPH. The additional deletion of maeA rendered a more robust second-generation biosensor strain for NADP+ reduction. We devised a structure-guided directed evolution approach to change cofactor specificity in Pseudomonas sp. 101 FDH. To this end, a library of >106 variants was tested using in vivo selection. Compared to the best engineered enzymes reported, our best variant carrying five mutations shows 5-fold higher catalytic efficiency and 13-fold higher specificity towards NADP+, as well as 2-fold higher affinity towards formate. In conclusion, we demonstrate the potential of in vivo selection and evolution-guided approaches to develop better NADPH biosensors and to engineer cofactor specificity by the simultaneous improvement of multiple parameters (kinetic efficiency with NADP+, specificity towards NADP+, and affinity towards formate), which is a major challenge in protein engineering due to the existence of tradeoffs and epistasis. N2 - NADPH ist ein essentieller Kofaktor, der biosynthetische Reaktionen in allen lebenden Organismen antreibt. Es ist ein Reduktionsmittel und damit Elektronenspender für anabole Reaktionen, die wichtige Zellkomponenten sowie viele Produkte in der Biotechnologie erzeugen. In der Tat ist das Engineering von Stoffwechselwegen in Mikroben für die Herstellung vieler Produkte oft durch die Verfügbarkeit von NADPH begrenzt. Eine gängige Strategie zur Lösung dieses Problems ist der Austausch der Kofaktor-Spezifität von NADH gegen NADPH in Enzymen von Stoffwechselwegen, da der erstgenannte Kofaktor reichlicher vorhanden ist als der letztere. Dieser Prozess ist jedoch zeitaufwendig und schwierig, da mehrere Parameter parallel entwickelt werden müssen. Daher ist das erste Ziel dieses Projekts die Etablierung eines effizienten metabolischen Biosensors zur Auswahl von Enzymen, die NADP+ reduzieren können. Ein auxotropher NADPH-Stamm wurde durch die Entfernung der wichtigsten Reaktionen, die an der NADPH-Biosynthese im zentralen Kohlenstoffmetabolismus von E. coli beteiligt sind, mit Ausnahme der 6-Phosphogluconat-Dehydrogenase, konstruiert. Um diesen Stamm zu validieren, wurden zwei Enzyme in Gegenwart mehrerer Kohlenstoffquellen getestet: eine Dihydrolipoamid-Dehydrogenase-Variante von E. coli mit sieben Mutationen und Formiat-Dehydrogenase (FDH) aus Mycobacterium vaccae N10 mit vier Mutationen wurden gefunden, die die NADPH-Biosynthese und das Wachstum unterstützen. Der Stamm wurde dann einer adaptiven Laborentwicklung unterzogen mit dem Ziel, seine Robustheit unter verschiedenen Kohlenstoffquellen zu testen. Unser Evolutionsexperiment führte zu einer zufälligen Mutagenese des Apfelsäure-Enzyms (maeA), die es ihm ermöglicht, NADPH zu produzieren. Die zusätzliche Entfernung von maeA machte einen robusteren Biosensor-Stamm der zweiten Generation für die NADP+-Reduktion möglich. Wir entwickelten einen strukturgesteuerten Evolutionsansatz zur Änderung der Kofaktorspezifität von Pseudomonas sp. 101 FDH. Zu diesem Zweck wurde eine Bibliothek von >106 Varianten mit Hilfe der in vivo-Selektion getestet. Im Vergleich zu den am besten entwickelten Enzymen über die berichtet wurde, zeigt unsere beste Variante mit fünf Mutationen eine 5-fach höhere katalytische Effizienz und eine 13-fach höhere Spezifität gegenüber NADP+ sowie eine 2-fach höhere Affinität gegenüber Formiat. Zusammenfassend zeigen wir das Potenzial der in vivo-Selektion und evolutionsgesteuerten Ansätze zur Entwicklung 14 besserer NADPH-Biosensoren und zur Entwicklung der Kofaktor-Spezifität durch die gleichzeitige Verbesserung mehrerer Parameter (kinetische Effizienz mit NADP+, Spezifität gegenüber NADP+ und Affinität zu Formiat), was aufgrund der Existenz von Zielkonflikten und Epistase eine große Herausforderung im Protein-Engineering darstellt. KW - formate dehydrogenases Y1 - 2020 ER - TY - THES A1 - Michaelis, Marcus T1 - Molekulare Erkennung von Cellulose und Cellulose-Fragmenten durch Cellulose-Bindemodule & Interaktionsstudien zwischen den zytoplasmatischen Domänen von Integrin-β1/β3 und dem fokalen Adhäsionsprotein Paxillin T1 - Molecular recognition of cellulose and cellulose fragments by cellulose binding modules & Interaction studies between the cytoplasmic domains of integrin-β1/β3 and the focal adhesion protein paxillin N2 - Proteine erfüllen bei einer Vielzahl von Prozessen eine essenzielle Rolle. Um diese Funktionsweisen zu verstehen, bedarf es der Aufklärung derer Struktur und deren Bindungsverhaltens mit anderen Molekülen wie Proteinen, Peptiden, Kohlenhydraten oder kleinen Molekülen. Im ersten Teil dieser Arbeit wurden der Wildtyp und die Punktmutante N126W eines Kohlenhydrat-bindenden Proteins aus dem hitzestabilen Bakterium C. thermocellum untersucht, welches Teil eines Komplexes ist, der Kohlenhydrate wie Cellulose erkennen, binden und abbauen kann. Dazu wurde dieses Protein mit E.coli Bakterien hergestellt und durch Metallchelat- und Größenausschlusschromatographie gereinigt. Die Proteine konnten isotopenmarkiert mittels Kernspinresonanz-Spektroskopie (NMR) untersucht werden. H/D-Austauschexperimente zeigten leicht und schwer zugängliche Stellen im Protein für eine mögliche Ligandenwechselwirkung. Anschließend konnte eine Interaktion beider Proteine mit Cellulosefragmenten festgestellt werden. Diese interagieren über zwischenmolekulare Kräfte mit den Seitenketten von aromatischen Aminosäuren und über Wasserstoffbrückenbindungen mit anderen Resten. Weiterhin wurde die Calcium-Bindestelle analysiert und es konnte gezeigt werden, das diese nach der Proteinherstellung mit einem Calcium-Ion besetzt ist und dieses mit dem Komplexbildner EDTA entfernbar ist, jedoch wieder reversibel besetzt werden kann. Zum Schluss wurde mittels zweier Methoden versucht (grafting from und grafting to), das Protein mit einem temperatursensorischen Polymer (Poly-N-Isopropylacrylamid) zu koppeln, um so Eigenschaften wie Löslichkeit oder Stabilität zu beeinflussen. Es zeigte sich, das während die grafting from Methode (Polymer wächst direkt vom Protein) zu einer teilweisen Entfaltung und Destabilisierung des Proteins führte, bei der grafting to Methode (Polymer wird separat hergestellt und dann an das Protein gekoppelt) das Protein seine Stabilität behielt und nur wenige Polymerketten angebaut waren. Der zweite Teil dieser Arbeit beschäftigte sich mit der Interaktion von zwei LIM-Domänen des Proteins Paxillin und der zytoplasmatischen Domäne der Peptide Integrin-β1 und Integrin-β3. Diese spielen eine wichtige Rolle bei der Bewegung von Zellen. Dabei interagieren sie mit einer Vielzahl an anderen Proteinen, um fokale Adhäsionen (Multiproteinkomplexe) zu bilden. Bei der Herstellung des Peptids Integrin-β3 zeigte sich durch Größenausschlusschromatographie und Massenspektrometrie ein Abbau, bei dem verschiedene Aminosäuregruppen abgespalten werden. Dieser konnte durch eine Zugabe des Serinprotease-Inhibitors AEBSF verhindert werden. Anschließend wurde die direkte Interaktion der Proteine untereinander mittels NMR untersucht. Dabei zeigte sich, das Integrin-β1 und Integrin-β3 an die gleiche Position binden, nämlich an den flexiblen Loop der LIM3-Domäne von Paxillin. Die Dissoziationskonstanten zeigten, dass Integrin-β1 mit einer zirka zehnfach höheren Affinität im Vergleich zu Integrin-β3 an Paxillin bindet. Während Paxillins Bindestelle an Integrin-β1 in der Mitte des Peptids liegt, ist bei Integrin-β3 der C-Terminus essenziell. Daher wurden die drei C-terminalen Aminosäuren entfernt und erneut Bindungsstudien durchgeführt, welche gezeigt haben, das die Affinität dadurch fast vollständig unterbunden wurde. Final wurde der flexible Loop der LIM3-Domäne in zwei andere Aminosäuresequenzen mutiert, um die Bindung auf der Paxillin-Seite auszulöschen. Jedoch zeigten sowohl Zirkulardichroismus-Spektroskopie als auch NMR-Spektroskopie, dass die Mutationen zu einer teilweisen Entfaltung der Domäne geführt haben und somit nicht als geeignete Kandidaten für diese Studien identifiziert werden konnten. N2 - Proteins play an essential role in a variety of processes. Understanding these functions requires elucidation of their structure and their binding behavior with other molecules such as proteins, peptides, carbohydrates, or small molecules. In the first part of this work, the wild type and the point mutant N126W of a carbohydrate-binding protein from the heat-stable bacterium C. thermocellum were studied, which is part of a complex that can recognize, bind and degrade carbohydrates such as cellulose. For this purpose, this protein was produced with E. coli bacteria and purified by metal chelation and size exclusion chromatography. The proteins could be isotopically labeled by nuclear magnetic resonance (NMR) spectroscopy. H/D exchange experiments revealed easy and difficult sites in the protein for possible ligand interaction. Subsequently, interaction of both proteins with cellulose fragments was detected. These interact with the side chains of aromatic amino acids via intermolecular forces and with other residues via hydrogen bonds. Furthermore, the calcium binding site was analyzed and it could be shown that it is occupied by a calcium ion after protein production and that this can be removed with the complexing agent EDTA, but that it can be reversibly occupied again. Finally, two methods (grafting from and grafting to) were used to couple the protein with a temperature-sensitive polymer (poly-N-isopropylacrylamide) in order to influence properties such as solubility or stability. It was found that while the grafting from method (polymer grows directly from the protein) resulted in partial unfolding and destabilization of the protein, in the grafting to method (polymer is prepared separately and then coupled to the protein) the protein retained its stability and only a few polymer chains were attached. The second part of this work dealt with the interaction of two LIM domains of the protein paxillin and the cytoplasmic domain of the peptides integrin-β1 and integrin-β3, which play an important role in cell movement. In doing so, they interact with a variety of other proteins to form focal adhesions (multiprotein complexes). In the preparation of the peptide integrin-β3, size exclusion chromatography and mass spectrometry revealed a degradation in which various amino acid groups are cleaved. This could be prevented by addition of the serine protease inhibitor AEBSF. Subsequently, the direct interaction of the proteins with each other was investigated by NMR. This showed that integrin-β1 and integrin-β3 bind to the same position, namely to the flexible loop of the LIM3 domain of paxillin. The dissociation constants showed that integrin-β1 binds to paxillin with an approximately tenfold higher affinity compared to integrin-β3. While Paxillin's binding site to integrin-β1 is in the middle of the peptide, the C-terminus is essential for integrin-β3. Therefore, the three C-terminal amino acids were removed and binding studies were performed again, which showed that this almost completely prevented affinity. Finally, the flexible loop of the LIM3 domain was mutated into two other amino acid sequences to extinguish binding on the paxillin side. However, both circular dichroism spectroscopy and NMR spectroscopy showed that the mutations resulted in partial unfolding of the domain and thus could not be identified as suitable candidates for these studies. KW - CBM KW - cellulose-binding KW - Cellulose-Bindung KW - protein polymer conjugate KW - Protein-Polymer-Konjugat KW - focal adhesion KW - fokale Adhäsionen KW - Integrin KW - Paxillin KW - cell migration KW - Zellmigration Y1 - 2022 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:517-opus4-555162 ER - TY - THES A1 - Küçükgöze, Gökhan T1 - Purification and characterization of mouse aldehyde oxidases T1 - Aufreinigung und Charakterisierung von Maus-Aldehyd-Oxidasen N2 - Mouse aldehyde oxidases (mAOXs) have a homodimeric structure and belong to xanthine oxidase family of molybdo-flavoenzymes. In general, each dimer is characterized by three subdomains: a 20 kDa N-terminal 2x[2Fe2S] cluster containing domain, a 40 kDa central FAD-containing domain and an 85 kDa C-terminal molybdenum cofactor (Moco) containing domain. Aldehyde oxidases have a broad substrate specificity including the oxidation of different aldehydes and N-heterocyclic compounds. AOX enzymes are present in mainly all eukaryotes. Four different homologs of AOX were identified to be present with varying numbers among species and rodents like mice and rats contain the highest number of AOX isoenzymes. There are four identified homologs in mouse named mAOX1, mAOX3, mAOX2, and mAOX4. The AOX homologs in mice are expressed in a tissue-specific manner. Expression of mAOX1 and mAOX3 are almost superimposable and predominantly synthesized in liver, lung, and testis. The richest source of mAOX4 is the Harderian gland, which is found within the eye's orbit in tetrapods. Expression of mAOX2 is strictly restricted to the Bowman’s gland, the main secretory organ of the nasal mucosa. In this study, the four catalytically active mAOX enzymes were expressed in a heterologous expression system in Escherichia coli and purified in a catalytically active form. Thirty different structurally related aromatic, aliphatic and N-heterocyclic compounds were used as substrates, and the kinetic parameters of all four mAOX enzymes were directly compared. The results showed that all enzymes can catalyze a broad range of substrates. Generally, no major differences between mAOX1, mAOX3 and mAOX2 were identified and the substrate specificity of mAOX1, mAOX3, and mAOX2 was broader compared to that of mAOX4 since mAOX4 showed no activity with substrates like methoxy-benzaldehydes, phenanthridine, N1-methyl-nicotinamide, and cinnamaldehyde and 4-(dimethylamino)cinnamaldehyde. We investigated differences at the flavin site of the mAOX enzymes by measuring the ability of the four mAOX enzymes to oxidize NADH in the absence of oxygen. NADH was able to reduce only mAOX3. The four mouse AOXs are also characterized by quantitative differences in their ability to produce superoxide radicals. mAOX2 is the enzyme generating the largest rate of superoxide radicals of around 40% in relation to moles of substrate converted and it is followed by mAOX1 with a ratio of 30%. To understand the factors that contribute to the substrate specificity of mAOX4, site-directed mutagenesis was applied to substitute amino acids in the substrate-binding funnel by the ones present in mAOX1, mAOX3, and mAOX2. The amino acids Val1016, Ile1018 and Met1088 were selected as targets. An increase in activity was obtained by the amino acid exchange M1088V in the active site identified to be specific for mAOX4, to the amino acid identified in mAOX3. N2 - Mouse aldehyde oxidases (mAOXs) have a homodimeric structure and belong to xanthine oxidase family of molybdo-flavoenzymes. In general, each dimer is characterized by three subdomains: a 20 kDa N-terminal 2x[2Fe2S] cluster containing domain, a 40 kDa central FAD-containing domain and an 85 kDa C-terminal molybdenum cofactor (Moco) containing domain. Aldehyde oxidases have a broad substrate specificity including the oxidation of different aldehydes and N-heterocyclic compounds. AOX enzymes are present in mainly all eukaryotes. Four different homologs of AOX were identified to be present with varying numbers among species and rodents like mice and rats contain the highest number of AOX isoenzymes. There are four identified homologs in mouse named mAOX1, mAOX3, mAOX2, and mAOX4. The AOX homologs in mice are expressed in a tissue-specific manner. Expression of mAOX1 and mAOX3 are almost superimposable and predominantly synthesized in liver, lung, and testis. The richest source of mAOX4 is the Harderian gland, which is found within the eye's orbit in tetrapods. Expression of mAOX2 is strictly restricted to the Bowman’s gland, the main secretory organ of the nasal mucosa. In this study, the four catalytically active mAOX enzymes were expressed in a heterologous expression system in Escherichia coli and purified in a catalytically active form. Thirty different structurally related aromatic, aliphatic and N-heterocyclic compounds were used as substrates, and the kinetic parameters of all four mAOX enzymes were directly compared. The results showed that all enzymes can catalyze a broad range of substrates. Generally, no major differences between mAOX1, mAOX3 and mAOX2 were identified and the substrate specificity of mAOX1, mAOX3, and mAOX2 was broader compared to that of mAOX4 since mAOX4 showed no activity with substrates like methoxy-benzaldehydes, phenanthridine, N1-methyl-nicotinamide, and cinnamaldehyde and 4-(dimethylamino)cinnamaldehyde. We investigated differences at the flavin site of the mAOX enzymes by measuring the ability of the four mAOX enzymes to oxidize NADH in the absence of oxygen. NADH was able to reduce only mAOX3. The four mouse AOXs are also characterized by quantitative differences in their ability to produce superoxide radicals. mAOX2 is the enzyme generating the largest rate of superoxide radicals of around 40% in relation to moles of substrate converted and it is followed by mAOX1 with a ratio of 30%. To understand the factors that contribute to the substrate specificity of mAOX4, site-directed mutagenesis was applied to substitute amino acids in the substrate-binding funnel by the ones present in mAOX1, mAOX3, and mAOX2. The amino acids Val1016, Ile1018 and Met1088 were selected as targets. An increase in activity was obtained by the amino acid exchange M1088V in the active site identified to be specific for mAOX4, to the amino acid identified in mAOX3. KW - aldehyde oxidase KW - drug metabolism KW - molybdenum cofactor KW - enzyme isoforms KW - enzyme kinetics KW - Aldehyd-oxidase KW - Metabolismus von Medikamenten KW - Molybdänkofaktor KW - Isoenzyme KW - Enzymkinetik Y1 - 2019 ER -