Dissecting the Histone Code Involving Bromodomains and the Histone Variant, H2A.Z

Abstract

Modifications to DNA-packaging histone proteins within chromatin act as a complex chemical “language” to recruit epigenetic “reader” proteins to mediate the transcriptional on and off states of specific genomic regions. One such protein, BPTF, associates with chromatin through the recognition of N-ε-acetylated lysine residues on histone N-termini by a bromodomain structural motif. While bromodomain-containing proteins play an essential role in regulating gene expression for immune homeostasis and tissue development, aberrant function of BPTF has been associated with various human cancers. Therefore, learning the specific types and patterns of modifications BPTF “reads” as part of the Histone Code and the resulting downstream effects can assist therapeutic efforts. Additionally, while bromodomain recognition of canonical histones (H2A, H2B, H3, and H4) has been well studied, little is known about possible interactions with two oncogenic histone variants, H2A.Z I and H2A.Z II.In the first part of this thesis, I will describe my work using NMR techniques to study H2A.Z interactions with the bromodomain of BPTF. In Chapter 2, I used a protein-observed fluorine NMR (PrOF NMR) binding assay to characterize the affinities of N-terminal H2A.Z peptides with various acetylation motifs for the bromodomain of BPTF to assess functional differences between the two highly homologous H2A.Z isoforms. Here we found the diacetylated mark of lysine 7 and 13 on H2A.Z II to have the strongest interaction with BPTF. Additional structural analyses with ligand-observed 1H CPMG NMR and X-ray crystallography demonstrated that a single lysine residue, K7ac, preferentially engaged the bromodomain binding site. A further panel of H2A.Z isoform iii selectivity panel against a variety of bromodomains revealed the bromodomain of CECR2 binds with the highest affinity and specificity for acetylated H2A.Z I over isoform II. To further validate these H2A.Z-bromodomain interactions in a physiologically relevant context, Chapter 3 of this thesis describes my approaches to develop both an affinity-based proteomics assay using bead-immobilized H2A.Z peptides and a coimmunoprecipitation assay using full-length H2A.Z protein constructs. While our preliminary proteomics data has identified bromodomain-containing proteins associated with acetylated H2A.Z peptides, and affinity purification of exogenously added BPTF bromodomain, we have yet to identify endogenous BPTF in pull-down experiments, likely due to the limited sensitivity of the assay for detecting weak interactions. The introduction of a photo-activatable diazirine moiety to the N-terminus of the peptide baits is an alternative approach that we are pursuing and is showing promising initial results. In a tandem effort to develop a complimentary co-immunoprecipitation assay, we have evaluated many magnetic bead-based pull-down systems using both affinity-tagged, fulllength H2A.Z and BPTF(BRD) constructs. While our results using GST-tagged BPTF support a BPTF-H2A.Z interaction, we are still investigating the best conditions for a minimizing background signal from nonspecific binding. We hope that with continued optimization, both assays will provide a basis for understanding the elusive mechanism of H2A.Z in gene regulation. In Chapter 4, I describe an effort to develop a useful protein-labeling system as a cost-effective approach to the traditional methods used in the Pomerantz lab for proteinobserved 19F NMR. Current experiments that rely on biosynthetic, metabolic protein labeling with 19F often require fluorinated amino acids, which in the case of 2- and 3- iv fluorotyrosine can be expensive. However, using these amino acids has provided valuable insight into protein dynamics, structure, and function. We developed a new in-cell method for fluorinated tyrosine generation from readily available substituted phenols and subsequent metabolic labeling of proteins in a single bacterial expression culture. This approach uses a dual-gene plasmid encoding for a model protein BRD4(D1) and a tyrosine phenol lyase from Citrobacter freundii, which catalyzes the formation of tyrosine from phenol, pyruvate, and ammonium. Our system demonstrated both enzymatic fluorotyrosine production and expression of 19F-labeled proteins as analyzed by 19F NMR and LC-MS methods. Further optimization of our system should provide a cost-effective alternative to a variety of traditional protein-labeling strategies. Preliminary studies for studying the function of the fluorinated KIX domain, and additional collaborative studies are outlined in Appendices A-E.