<p>The posttranslation modification of histone proteins within the nucleosomes of chromatin plays important roles in the regulation of gene expression in both normal biological and pathobiological processes. These modifications alter local chromatin structure and subsequently alter the expression profile of associated genes. Histone methylation, which was long thought immutable, is one such modification that plays a dual functionality in both activation and repression of gene expression and can be thought of as an information storage mark. With the initial discovery of lysine-specific demethylase 1A (LSD1/KDM1A), an FAD-dependent enzyme that catalyzes the oxidative demethylation of histone H3K4me1/2 and H3K9me1/2 repressing and activating transcription, respectively, the missing counterbalance to dynamic histone methylation was cemented. This discovery further strengthened the link between histone demethylation and transcriptional regulation and the enzyme has since been identified as a target with therapeutic potential.</p><p>Given the significance of KDM1A enzymatic activity, herein we report our efforts to characterize novel binding interactions that dictate the enzymes biological and pathobiological functions. As KDM1A falls into the greater class of flavin-dependent amine oxidases, it contains features that are recurrent within the class, but due to its unique ability to work on histone and non-histone substrates has unprecedented structural elements. Although the active site is expanded compared to the greater amine oxidase superfamily, it is too sterically restricted to encompass the minimal 21-mer peptide substrate footprint of the histone H3 tail. The remainder of the substrate/product is therefore expected to extend along the surface of KDM1A. Using steady-state kinetic analyses, we now show that unmodified histone H3 is a tight-binding, competitive inhibitor of KDM1A demethylation activity with a Ki of 18.9 ± 1.2 nM that is approximately 100-fold higher than the 21-mer peptide product. The relative affinity of dose-response curves is independent of preincubation time suggesting that H3 rapidly reaches equilibrium with KDM1A. Rapid dilution experiments confirmed the increased binding affinity of full-length H3 toward KDM1A was at least partially caused by a slow off-rate with a koff of 0.072 min-1, a half-life (t1/2) of 9.63 min, and residence time (τ) of 13.9 min. Independent affinity capture surface plasmon resonance experiments confirmed the tight-binding nature of the H3/KDM1A interaction revealing a Kd of 9.02 ± 2.27 nM, a kon of 9.26 x 104 ± 1.5 x 104 M-1s-1 and koff of 8.35 x 10-4 ± 3.4 x 105 s-1. Additionally, consistent with H3 being the only histone substrate of KDM1A, no other core histones are inhibitors of demethylation activity. Our data suggests that KDM1A contains a histone H3 secondary specificity recognition element on the enzyme surface and required further characterization.</p><p>In order to characterize this secondary H3 binding site, we turned to the use of cysteine labeling, chemical cross-linking coupled to proteolysis and LC-MS/MS, HDX-MS, and the design of an active, tower domain deletion KDM1A mutant. We now show that the tower domain contributes to the extended binding interface of the KDM1A/H3 interaction. Additionally, we show that the KDM1A tower domain is not required for demethylation activity and that one can functionally uncouple catalytic activity from protein-protein interactions that occur along the KDM1A tower domain interface, a domain unprecedented in the greater amine oxidase family. Furthermore, this towerless mutant will be useful for dissecting molecular contributions to KDM1A function along the tower domain. Our discovery of this secondary binding site within the aforementioned domain points to how pivotal this region is to the control and localization of KDM1A enzymatic activity as it also serves a pivotal role as a protein-protein interaction motif for the nucleation of a multitude of multimeric protein complexes.</p><p>With this in mind, we set out to design a strategy to isolate the core histone demethylase complex from E. coli cellular lysates. With the use of polycistronic vectors that encode both KDM1A and CoREST for coexpression we were able to produce appreciable amounts of chromatographically pure complex. As our CoREST construct in this strategy contains both the ELM2 and SANT2 domain needed for interaction with the HDACs, this core complex will serve as a starting point for future work that will tease apart additional influences on substrate binding and recognition imparted on KDM1A from binding partners. This preparation can therefore be used in a multitude of downstream studies including reconstitution of the core histone demethylase/deactylase complex and in depth kinetic and biophysical analyses and provides an invaluable starting point</p><p>This work provides a foundational understanding of this unprecedented secondary binding site on the surface of the KDM1A tower domain and how it may play an important role in substrate and product recognition. We suspect that this extended interaction interface may control KDM1A localization within specific chromatin loci and allow the enzyme to serve as a docking element for the nucleation of protein complexes or transcriptional machinery. On the other hand, disruption of this point of contact between the KDM1A/H3 binary complex may also facilitate enzyme/product dissociation, thereby tuning the catalytic activity of the demethylase. Additionally, the ability to produce substantial quantities of the core histone demethylase complex is a necessary step in the decoding of the ‘histone code’ hypothesis of KDM1A and its associated complexes. We suspect that the body of this work will prove to be invaluable for future characterization of the enzyme and its role in biology and pathobiology.</p> / Dissertation