Olsson, T.; (2009) An informatics approach to investigating structure-thermodynamic correlations in protein-ligand interactions. Doctoral thesis, UCL (University College London).
Full text not available from this repository.
Protein-ligand interactions are central to many biological processes. Developing a detailed understanding of protein-ligard interactions is therefore fundamental to the molecular life sciences. This study aims to further our understanding of how structure relates to the thermodynamics of binding. By constructing a database (SCORPIO) of calorimetric and structural data on the interactions of proteins with both biological ligands and synthetic inhibitors, we have created a resource which enables analysis and hypothesis testing of structure-thermodynamic correlations. The data in SCORPIO have been used to identity differences in the thermodynamics of binding of native ligands and synthetic inhibitors. The SCORPIO data have also been used to show that enthalpy-entropy compensation is a real phenomenon and not just an artefact of a limited affinity window. By-analysing the changes in surface area upon protein-ligand complexation, buried apolar surface is found to correlate well with ΔG (change in free energy) (R^2=0.65). Contrary-to common belief, this relationship cannot be simply explained by an underlying correlation with ΔS (change in entropy). The surface area analysis has some inherent weaknesses in that complementarity is assumed and water-mediated interactions are ignored. An atom-atom contact approach that explicitly deals with waters has therefore been developed and tested. Using the atom-atom contact approach a multiple linear regression (MLR) model describing ΔG has been created. The descriptors selected reflect the changes in polar- and apolar- hydration upon forming the complex. Interestingly direct contacts between the protein and the ligand do not appear to be important for describing the variance in ΔG, although some results indicate that the polar atom-atom contact descriptor is not sensitive enough to be used as a predictor of ΔG in the MLR model.
|Title:||An informatics approach to investigating structure-thermodynamic correlations in protein-ligand interactions|
|Additional information:||Authorisation for digitisation not received. Abstract contains LATEX text. Please see thesis for rendered formulae and equations|
|UCL classification:||UCL > School of Life and Medical Sciences > Faculty of Life Sciences > Biosciences (Division of) > Structural and Molecular Biology|
Archive Staff Only: edit this record