Expanding the NMR toolbox for nucleosome studies

Abstract

The way a cell controls which part of its long DNA to access and read, which part not, and the way it maintains the integrity of the genome, are crucial for the proper functioning of an organism. These processes depend to a large extent on the packing of the DNA in nucleosomes that, through their presence or absence and types of epigenetic modifications they carry, allow genome maintenance and regulation. Conjointly, malfunctions in these mechanisms are detrimental to the cell and have been frequently correlated with cancer. A structural understanding of how these mechanisms operate on nucleosomes to ensure the proper functioning of a cell is thus crucial. In this work, we attempted to extend the toolbox of Nuclear Magnetic Resonance (NMR)-based approaches for the study of the nucleosome, with the long-term aim of helping with the biological understanding of chromatin function and epigenetics. In the first chapter are provided an introduction of key aspects of nucleosome structure and function and an introduction of the basic principles of NMR in layman terms. In Chapter 2, the remarkable 1D spectrum of a small molecule, designed to mimic a trimethyl lysine for methyl-TROSY NMR studies, is described in detail. The protons in its symmetrical and isotope-labelled trimethylammonium group are chemically equivalent, yet magnetically inequivalent due to the presence of 3JCH couplings. This resulted in small 4JHH couplings that perturbed the spectral pattern in a complex yet surprisingly simple way due to the large 1JCH coupling maintaining a weak coupling system. Subspectral analysis allowed us to intuitively describe and build up the deceptively complicated spectrum. Additionally, we showed that the system can be returned to magnetic equivalence by selective homonuclear decoupling. In Chapters 3 to 5, we developed, characterized and used a new approach for the study of the nucleosome making use of the well-established techniques of sedimentation coupled to 1H-detected solid-state NMR. High quality spectra were obtained for histones H2A and H3, which were subsequently assigned. Their secondary structure within the nucleosome was de-novo determined and was in good agreement with previously reported crystal structures, confirming the sample integrity. Co-sedimentation and further characterization of a proof-of-concept known peptide allowed the obtention of clear perturbation of H2A signals, allowing the determination of the binding site and subsequent data-driven docking, in high agreement with the previously reported crystal structure. The new approach was then applied to the study of Dot1p, the yeast H3K79 methyltransferase. Through a divide-and-conquer approach and the combined use of biochemical and NMR methods, we found that the lysine-rich sequence directly N-terminal of the Dot1p core has a high affinity towards DNA, suggesting a nucleosome-anchoring role, in close vicinity from H3K79, for this domain. Additionally, the catalytic core of Dot1p has very low affinity for ubiquitin, in agreement with previous in-vivo assays. We then characterized the nucleosome sediment through SAXS and NMR, showing that our preparation is devoid of long-range ordering of nucleosomes. We finally attempted to co-sediment a low affinity, H3 tail-binding protein with the nucleosome, challenging the formation of a rigid, saturated protein nucleosome complex. Using solution- and solid-state NMR, we showed that the PHD2 domain of CHD4 counterintuitively combines low affinity interaction and a slow exchange interaction, while the bound state remained invisible. In Chapter 6, the results from the thesis are further discussed and put in perspective along with suggestions for future works and prospects