Refine
Year of publication
- 2006 (2) (remove)
Document Type
- Doctoral Thesis (2)
Has Fulltext
- yes (2)
Is part of the Bibliography
- no (2)
Keywords
- J coupling constants (1)
- Konformation (1)
- Kopplungskonstante (1)
- MD simulations (1)
- MD-Simulation (1)
- Magnetische Kernresonanz (1)
- Molekulardynamik (1)
- NMR spectroscopy (1)
- NMR-Spektroskopie (1)
- Peptide (1)
Institute
One of the most important tasks in chemistry and especially in structural biology has always been the elucidation of three-dimensional molecular structures - either of small molecules or large biopolymers. Among the (bio)physical methods to acquire structural data at atomic resolution electron paramagnetic resonance (EPR) spectroscopy is the most valuable technique for obtaining structural information about many different kinds of paramagnetic species. In biological systems, either paramagnetic metal ions/clusters, transient paramagnetic intermediates in electron transfer processes or artificially attached stable spin labels can be found. The usual approach to interpret EPR spectra is to perform simulations based on the so-called spin Hamiltonian (SH). This means that the well-defined numerical parameters (tensors) in the SH representing different types of interaction are obtained by fitting the experimental data. The SH parameters include electronic g-values, hyperfine coupling (HFC) and quadrupole coupling (&C) constants, zero-field splittings and constants to describe exchange and dipolar interactions between electron spin systems. However, since the SH only contains spin degrees of freedom, a direct translation of the SH EPR parameters into structural information is not straightforward. Therefore, methods to predict such SH interaction parameters starting from molecular structures are required. In this thesis it was investigated whether quantum chemical calculations of EPR parameters based on density functional theory (DFT) methods may be employed to overcome these problems thus enabling a correlation of experimental EPR data with molecular structure. It was the central goal of this work to point out the potential of a fruitful interplay between quantum chemistry and experiment and to study how both can benefit from each other. For this purpose DFT methods were applied to a variety of organic radical or transition metal systems to calculate different EPR parameters. Using the 'broken symmetry' formalism it was possible to compute the exchange coupling constant for a nitroxide biradical and furthermore decompose the exchange mechanism in different through-bond and through-space interactions. Spin density distributions, 14N and 1H HFC constants as well as dipole moments and polarizabilities were computed for a number of aromatic nitroxides to examine their properties and select promising candidates which may serve as DNA-intercalating spin labels. Systematic investigations of the influence of hydrogen bond geometry on the 14N QC parameters for imidazole-water and methylimidazole-benzosemiquinone complexes lead to the conclusion that especially the imidazole amino nitrogen &C parameters are very sensitive probes of the bond geometry, in particular of the hydrogen bond length. The results of this study may be applied to biological systems, e.g. to gain structural information about quinone binding sites. Moreover, quantum chemical methods were applied to elucidate the structure of a nitrogen-centered radical intermediate in the inhibition process of ribonucleotide reductase (RNR). It was possible to find a molecular structure in accordance with all experimentally available data, thus revealing the longsought structure of the No radical and providing evidence for the trapping of a 3'-ketonucleotide in the reduction process catalyzed by RNR. To test the capability of modern DFT methods to predict g- and molybdenum HFC tensors for MoV complexes, validation studies were carried out. Comparison of computed EPR parameters of a number of MoV compounds with corresponding experimental values showed that g- and HFC tensors could be predicted in good accuracy, although some systematic errors of the computational methods have to be considered for such heavy 4d1 transition meta1 systems. Furthermore, DFT calculations on a Mn2+ binding site model of the hammerhead ribozyme allowed to conclude that the structure of the binding site as studied by EPR spectroscopy in frozen solution is very likely to be identical to the site found occupied by Mn2+ in crystals. Finally, computational methods were employed to aid in the structural characterization of the Mn2+ binding site in Ras (rat sarcoma protein) by providing accurate starting parameters for spectral simulations and furthermore helping to interpret the experimental data. In conclusion, it was demonstrated in this thesis that the combination of sophisticated experimental and quantum chemical methods represents a powerful approach in the field of EPR spectroscopy and that it may be essential to employ EPR parameter computations to extract the full information content from EPR spectra. Therefore, great potential lies in future applications of DFT methods to the large number of systems where detailed and reliable experimental data is available but where an unequivocal correlation of these data with structural information is still lacking.
Für das Verständnis der Proteinfaltung ist es von Interesse, die phi,psi-Torsionswinkelverteilung und deren Abhängigkeiten innerhalb einer Polypeptidkette zu kennen. Mit der in dieser Arbeit verwendeten Kombination aus MD-Simulation und NMR-Spektroskopie wird die Abhängigkeit der Konformationsverteilung kurzer alaninbasierter Modellpeptide mit einer Genauigkeit von 5 % bestimmt. Die Berechnung der thermischen Populationen der einzelnen Konformationen beruht auf einer Minimierung der Differenz aus experimentellen und berechneten skalaren Kopplungskonstanten. Trialanin populiert überwiegend den Bereich der Polyprolin Typ II Helix (~ 90 %) und daneben den beta-Faltblattbereich mit ca. 10%, jedoch nicht den alphaR-helicalen Bereich. Diese Konformationsverteilung ändert sich nicht signifikant mit zunehmender Kettenlänge in der Peptidreihe Ala3 bis Ala7. Das in der Seitenkette verzweigte Trivalin populiert dagegen alle drei Konformationsbereiche signifikant. Aufgrund der Periodizität der Torsionswinkel populiert Triglycin einen zusammenhängenden Bereich, der sich an den vier Ecken des Ramachandran-Diagramms befindet. Zudem befindet es sich in einem langsamen konformationellen Gleichgewicht zwischen der cis- und trans-Konformation der Peptidbindung. Die Temperaturabhängigkeit der Konformationsverteilung wird am Beispiel von Trialanin untersucht. Die 3J(HN,Ha) Kopplungskonstanten nehmen linear mit der Temperatur zu. Dies ist auf eine Zunahme des beta-Faltblattanteils zurückzuführen und kann theoretisch beschrieben werden. Die Konformationsverteilung der Trialaninsequenz innerhalb einer heteropolymeren Aminosäuresequenz ist von der Kettenlänge der an dem N- und C-Terminus angefügten heteropolymeren Aminosäuresequenz abhängig. Dies wird an zwei Peptiden, abgeleitet von der Sequenz des Proteins Lysozym aus Hühnereiweiß, gezeigt. Das kürzere Peptid hat an beiden Enden jeweils drei Aminosäurereste angefügt, das längere jeweils acht Aminosäurereste. Die Konformationsverteilung der Trialanisequenz des kürzeren Peptids entspricht nahezu der in der Peptidreihe Ala3 bis Ala7. Die Verteilung des längeren Peptids ist dagegen deutlich verschieden (~ 35% alphaR-helicaler Anteil). Die 1HN und 15N chemischen Verschiebungen der Trialaninsequenz des längeren Peptids sind mit denen des entfalteten Lysozym-Proteins identisch und demzufolge aller wahrscheinlichkeit nach auch die Konformationsverteilung. Kurze homopolymere Peptide eignen sich deshalb nicht als Modell für Aminosäuresequenzen in längeren heteropolymeren Peptiden.