Regulation of Histone Chaperones by Intrinsically Disordered Regions and Acidic Stretches

Abstract

The nucleosome is the fundamental repeating unit of chromatin that is responsible for the condensation of a large amount of genetic information into the nucleus as well as regulating gene expression. Nucleosomes are composed of 147bp of DNA wrapped around an octamer of histone proteins: 2 copies of each H2A, H2B, H3, and H4. Nucleosomes do not form spontaneously, and highly basic histone proteins are prone to interact non-specifically and aggregate with DNA as well as with other nucleic acids and proteins within the cell. A specialized group of histone chaperone proteins are therefore necessary to specifically bind the basic histones, shield them from non-specific interactions, facilitate their nuclear import, and transfer them onto DNA in a stepwise fashion to form nucleosomes.

Nucleoplasmin (Npm) is a H2A/H2B-specific histone chaperone that is only present during the early stages of vertebrate development. Nucleoplasmin is composed of two major domains, an N-terminal Core domain and a C-terminal Tail domain. The Core domain has been shown to fold into an extremely stable homopentamer, whereas the Tail domain is predicted to be an intrinsically disordered region (IDR). Npm contains three distinct acidic stretches, the short Al in the Core domain, the longest A2 in the Tail domain, and a short A3 also in the Tail domain. The Tail domain also contains a bipartite nuclear localization signal (NLS) and an arginine and lysine rich C-terminus.

In the first chapter of this thesis, using bioinformatics analyses I show that IDRs and acidic stretches are common features of the otherwise structurally diverse histone chaperone family. By reviewing the literature I show that these regions often make critical contacts on histones necessary for the high affinity and specificity of histone chaperones. Additionally, many of these regions engage histones through a conserved "anchoring and capping" motif using both aromatic and acidic residues. These regions often work in conjunction with ordered domains to bind and deposit histone proteins, and provide optimal sites for post-translational modification and regulation of histone chaperone function. Finally, I discuss the published literature on the role of Npm as a histone chaperone during Xenopus laevis embryogenesis.

In the second chapter of this thesis I show that the intrinsically disordered C-terminal Tail domain of Npm autoregulates the histone binding and deposition activity of Npm through multiple competitive intramolecular interactions. I show by NMR and circular dichroism (CD) spectroscopy that the Tail domain is intrinsically disordered, as predicted, but transiently samples β-sheet-like conformations. Salt titration experiments using analytical ultracentrifugation (AUC) show that the Tail domain exists in a compact conformation that becomes extended in high salt conditions. Similar salt titration NMR experiments showed chemical shift perturbations (CSPs) in the fast-exchange regime at two regions between segments of opposing charge, indicating that these regions are likely hinges in a compact structural ensemble. NMR CSP experiments upon titrating H2A/H2B dimers shows peak disappearance for resides in A2 indicating a direct interaction, and fast exchanging CSPs at the first hinge region indicating a partial conformational opening upon binding to histones. NMR relaxation experiments on the Npm Tail domain in both the unbound and histone-bound state indicates that the protein is highly dynamic and that residues in and around the A2 acidic stretch (120-150) become less dynamic upon binding to H2A/H2B. Using paramagnetic relaxation enhancement (PRE) NMR I show that multiple intramolecular contacts are made from A2 to the NLS, and A2 to the basic C-terminus in the unbound state and that histone binding disrupts A2 to basic C-terminus contacts. Ensemble structural modeling and MD simulations showed dynamic intramolecular contacts made in the unbound state and allowed for the determination of the approximate position of A2 relative to H2A/H2B in the histone bound state. NMR and small angle X-ray scattering (SAXS) analyses of the full-length protein show that these structural states are highly relevant to the pentameric protein. Quantitative histone binding assays and chromatin assembly assays demonstrate that these competitive intramolecular interactions negatively regulate histone binding and deposition functions of Npm. Together these data demonstrated a novel mechanism of histone chaperone regulation through intramolecular competition on highly charged IDRs.

In the third chapter I explore the effects of post-translational modifications on the structural states and histone binding properties of the Npm Tail domain. Glutamylation, in particular, occurs on four conserved glutamate residues within the binding interface of A2 and H2A/H2B defined in Chapter 2, and is found to occur on acidic stretches of many other histone chaperones and acidic nuclear proteins. Using quantitative histone binding assays, I showed that glutamylation of a short peptide from A2 significantly increased its affinity toward H2A/H2B compared to the non-glutamylated peptide. Using recombinant tubulin-tyrosine ligase-like 4 (TTLL4) glutamyltransferase enzyme, I showed that I could glutamylate the Npm Core+A2 truncation (1-145), but not a GST-tagged A2 (119-145), indicating that the Core domain is important for TTLL4 modification of A2. I also showed that I can enrich for glutamylated forms of the Core+A2 truncation using anion-exchange chromatography. Chromatin assembly assays comparing unmodified and glutamylated forms of the Core+A2 truncation show a subtle decrease in the histone deposition activity of glutamylated Core+A2. Finally, extensive molecular dynamics (MD) simulations yielded insights into the effects of combinatorial phosphorylation, glutamylation and arginine methylation on intramolecular interactions and acidic stretch exposure in the Npm Tail domain. I show that oocyte and egg-mimicking phosphorylations effectively disrupt long-range intramolecular interactions, whereas glutamylation and arginine methylation may act to more subtly disrupt short-range interactions.

In the fourth chapter, I define new histone-based structural tools that will be useful for future structural studies of Npm, other H2A/H2B chaperones, and potentially many other histone-interacting proteins. First, I isotope labeled and assigned the majority of residues in HSQC spectra of the Xenopus H2A/H2B dimer. Using this new tool I was able to map the binding sites of the Npm Tail domain, and a short peptide derived from A2 ± glutamylation on the H2A/H2B dimer by NMR CSP experiments. These experiments also showed that chaperone binding likely leads to a dramatic change in the dynamics of the H2A/H2B dimer. Finally, based on structural studies of other H2A.Z/H2B chaperones, I engineered a single-chain construct of Xenopus H2B fused directly to H2A with the histone tails removed (termed scH2BH2A). I show that this construct is easily expressed in bacteria, is soluble, and forms highly reproducible crystals. A high-resolution crystal structure shows that the scH2BH2A adopts a nearly identical fold to nucleosomal H2A/H2B. Attempts were made to co-crystallize this protein with Npm peptides. These new histone-based NMR and X-ray crystallography tools can be used to structurally study many other H2A/H2B chaperones and interacting proteins.

In the final chapter, I present a brief discussion on the data presented in the prior four chapters, as well as address open questions and future areas of study for the field.