Implementation of a methodology for the structural investigation of proteins based on EPR spectroscopy with porphyrin probes

This Thesis focuses on the analysis of spectroscopic results from Pulsed Dipolar Spectroscopy (PDS) by means of Atomistic Molecular Dynamics (MD) simulations. PDS is a pulsed EPR technique that has become an important biophysical tool in the study of macromolecular systems since it allows the measurement of nanometric distances and distance distributions that can be used to unravel structural details of biomacromolecules. It relies on the measurement of the magnetic dipolar interaction between two paramagnetic centres, typically a pair of nitroxide spin labels, artificially introduced by site-directed spin labelling. Furthermore, orientation-selective PDS experiments do not provide only the distance distribution but also the relative orientation of the spin labels, which is a valuable information on the mutual orientation of interacting proteins or proteins subunits with respect to each other. A problem in the analysis and interpretation of experimental data is related to the flexibility of most spin labels, which contributes to the width of the measured distance distributions. To take into account this effect, a common practice is to use rotamer libraries, which include the stable conformers of the spin labels. An alternative is represented by atomistic MD simulations, which allow sampling of the spin label conformations within their environment. Even more important, these can account for the intrinsic flexibility of the whole spin labeled system; thus, the combination of PDS experiments and MD can provide unique insights into the conformational landscape of biomacromolecules. Despite these valuable aspects, the application of MD methodology to simulate pulsed dipolar EPR experiments combined with spin labeling has remained rather limited. The use of photogenerated triplet states as spin labels in light-induced PDS offers an effective alternative to conventional nitroxides for investigations on biological structures. The characteristic spin polarization of photoexcited triplet states, which is due to the non-Boltzmann populations of the triplet sublevels as a result of the inter-system crossing mechanisms, can lead to an enhanced spectroscopic sensitivity of the technique. The performance of light-induced PDS were tested on model systems, containing porphyrin-nitroxide probes or two porphyrin chromophores, separated by rigid alpha-helical bridges in previous investigations performed by the EPR group in collaboration with the University of Oxford and Manchester. In this Thesis, a conformational analysis on the above-mentioned model systems was carried out by MD simulations in order to complement the experimental data obtained from light-induced PDS and assess the validity of the PDS-MD combined methodology for the structural characterization of biomolecules. Since the reliability of MD results depends on the quality of the model used, firstly, an accurate parameterization of the potential energy for the systems under investigation was derived. MD was employed to sample the conformational distribution of the two peptides. The conformational analysis, performed at room temperature and at 200 K, have shown a different behaviour for the two model systems, where different probes and different peptide linkers are present. The following step, was to compare the sampled conformational distributions with the corresponding light-induced PDS results, allowing the selection of the conformational ensemble contributing to the experimental distance distribution. For this purpose, a procedure was implemented to calculate the distance distributions corresponding to the set of conformations obtained by MD. Overall, this work demonstrates that the integration of the distance and orientation analysis of PDS results with MD simulations is very important in order to provide a detailed information on the conformational landscape of the model systems under investigation.