Abstract by João Carlos Moreno Ramos

Protein biological function is intimately related to structure as well as dynamics at the molecular and atomic scales. X-ray diffraction has contributed immensely to the understanding of protein function mainly through the description of structural properties. However, the study of atomic motion has been somewhat overlooked due to the limitations inherent to X-ray models and to the challenges faced in obtaining neutron diffraction data. For many years, in small-molecule crystallography, the use of multipole models to describe X-ray data has been employed to accurately model the electron density and characterize the nature of chemical bonding in molecules. At present, the accuracy of X-ray diffraction data from protein crystals is not at the level required to enable this type of refinements. In contrast, small-molecule studies have shown the capabilities of neutron diffraction in terms of providing accurate information on atomic and molecular motion. It is the aim of the research presented in this thesis to extend the potential of neutron diffraction to the study of proteins to get information of their structure and dynamics not biased by the limitations of the X-ray models.

The main aim of the present work is to push the boundaries of neutron macromolecular crystallography to obtain anisotropic atomic displacement parameters (ADPs) for perdeuterated hen egg-white lysozyme (D-HEWL), providing a basis for comparisons with X-ray derived ADPs.

For the success of this study, a strategy for the production of significant quantities of D-HEWL was developed, consisting in its overexpression in E. coliinclusion bodies, followed by protein purification and in vitrorefolding.  The refolded D-HEWL, and its hydrogenated variant, were characterized in terms of biophysical and structural properties. Both variants were shown to be thermally stable and active, and with an identical overall structural fold compared to the native unlabelled protein, from Gallus gallus. Minor variations in protein structure, were however observed, mainly in the Lys97-Gly104 region, related to the Asn103 peptide-plane flip. These changes were linked to slight decreases in protein thermal stability and enzymatic function.

Complete atomic resolution neutron diffraction data was successfully measured at room temperature (RT) and 100 K, for D-HEWL crystals grown at the pH where the enzyme is fully active. The neutron data were complemented by equivalent X-ray datasets obtained in the same conditions. The RT D-HEWL neutron structure elucidated the detailed configuration and dynamics of protein residues and water molecules in the enzyme’s active site, in an active conformation. This dataset also provided insight into the potential of complete atomic resolution neutron data for a protein crystal, which has not been reported earlier. The D-HEWL X-ray and neutron datasets at RT and 100 K are the foundation for the present study of atomic motion. Preliminary results, from the D-HEWL RT data, suggest differences in the ADPs derived from the X-ray and neutron models. The X-ray derived ADPs appear to be larger and more isotropic that those obtained from neutrons. Additionally, the neutron model seems to describe more accurately structural disorder, affecting less the modelling of the ADPs. The neutron ADPs seem to contain biologically relevant information that is more physically realistic than can be derived from the X-ray model. Finally, an outline is given of additional work in order to extract reliable ADPs from the 100 K neutron dataset, which could corroborate the observations made at RT.