@phdthesis{Shkumatov2011,
  author    = {Shkumatov, Alexander V.},
  title     = {Methods for hybrid modeling of solution scattering data and their application},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-65044},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2011},
  abstract  = {Small-angle X-ray scattering (SAXS) is a universal low-resolution method to study proteins in solution and to analyze structural changes in response to variations of conditions (pH, temperature, ionic strength etc). SAXS is hardly limited by the particle size, being applicable to the smallest proteins and to huge macromolecular machines like ribosomes and viruses. SAXS experiments are usually fast and require a moderate amount of purified material. Traditionally, SAXS is employed to study the size and shape of globular proteins, but recent developments have made it possible to quantitatively characterize the structure and structural transitions of metastable systems, e.g. partially or completely unfolded proteins. In the absence of complementary information, low-resolution macromolecular shapes can be reconstructed ab initio and overall characteristics of the systems can be extracted. If a high or low-resolution structure or a predicted model is available, it can be validated against the experimental SAXS data. If the measured sample is polydisperse, the oligomeric state and/or oligomeric composition in solution can be determined. One of the most important approaches for macromolecular complexes is a combined ab initio/rigid body modeling, when the structures (either complete or partial) of individual subunits are available and SAXS data is employed to build the entire complex. Moreover, this method can be effectively combined with information from other structural, computational and biochemical methods. All the above approaches are covered in a comprehensive program suite ATSAS for SAXS data analysis, which has been developed at the EMBL-Hamburg. In order to meet the growing demands of the structural biology community, methods for SAXS data analysis must be further developed. This thesis describes the development of two new modules, RANLOGS and EM2DAM, which became part of ATSAS suite. The former program can be employed for constructing libraries of linkers and loops de novo and became a part of a combined ab initio/rigid body modeling program CORAL. EM2DAM can be employed to convert electron microscopy maps to bead models, which can be used for modeling or structure validation. Moreover, the programs CRYSOL and CRYSON, for computing X-ray and neutron scattering patterns from atomic models, respectively, were refurbished to work faster and new options were added to them. Two programs, to be contributed to future releases of the ATSAS package, were also developed. The first program generates a large pool of possible models using rigid body modeling program SASREF, selects and refines models with lowest discrepancy to experimental SAXS data using a docking program HADDOCK. The second program refines binary protein-protein complexes using the SAXS data and the high-resolution models of unbound subunits. Some results and conclusions from this work are presented here. The developed approaches detailed in this thesis, together with existing ATSAS modules were additionally employed in a number of collaborative projects. New insights into the "structural memory" of natively unfolded tau protein were gained and supramodular structure of RhoA-specific guanidine nucleotide exchange factor was reconstructed. Moreover, high resolution structures of several hematopoietic cytokine-receptor complexes were validated and re-modeled using the SAXS data. Important information about the oligomeric state of yeast frataxin in solution was derived from the scattering patterns recorded under different conditions and its flexibility was quantitatively characterized using the Ensemble Optimization Method (EOM).},
  subject      = {R{\"o}ntgen-Kleinwinkelstreuung},
  language  = {en}
}