Targeting HIF/Coregulator Complexes in Cancer: Characterizing Fragile Fold and Inhibitor Binding Sites of ARNT

Author

Xingjian Xu, The Graduate Center, City University of New YorkFollow

Date of Degree

6-2023

Document Type

Dissertation

Degree Name

Ph.D.

Program

Biochemistry

Advisor

Kevin H. Gardner

Committee Members

Ranajeet Ghose

Amédée des Georges

Emilio Gallicchio

Catherine A. Royer

Subject Categories

Biochemistry, Biophysics, and Structural Biology

Keywords

ARNT, HIF, transcription factor, protein-protein/protein-ligand interactions, NMR, integrative structural biology

Abstract

Hypoxia-inducible factors (HIFs) are transcription factors that play crucial roles in regulating adaptive cellular responses to low oxygen concentrations (hypoxia). Structurally, HIFs are heterodimers composed of one of three α-subunits (HIF-1α, -2α, and -3α) and a β subunit (HIF-β, also known as ARNT), with the heterodimerization mediated by the tandem PAS domains (PAS-A and PAS-B) found within both proteins. Dysregulation of the HIF pathway has been implicated in various types of cancers, contributing to cancer cell survival, tumor growth and progression, and resistance to therapies, resulting in poor treatment outcomes. As such, HIF-specific inhibitors have been extensively investigated as novel therapeutic strategies. Recently, a group of transcriptional coactivators known as the coiled-coil coactivators (CCCs) has been shown to form complexes with HIFs via the PAS-B domain of ARNT, establishing ARNT/CCC interactions as potential targets for regulation. Here, several studies were conducted to better understand the structural characteristics of ARNT PAS-B and its interactions with small molecules and CCCs, with an eye toward developing more potent inhibitors.

ARNT PAS-B has a ‘fragile’ fold, making it susceptible to structural destabilization when point mutations are made to its functionally important β-sheet surface. Specifically, a single point mutation, Y456T, splits the protein between two slowly exchanging states. In the first project, we used high-pressure NMR to characterize this interconversion process and derived biophysical parameters such as volume and compressibility differences between the two conformations. We also established that the interconversion proceeds through a chiefly unfolded transition intermediate. Furthermore, we performed a series of ligand titration experiments and showed that small molecule interactions could also modulate the protein’s conformational equilibria. In summary, this work advanced our understanding of the ARNT PAS-B conformational dynamics and demonstrated an array of NMR-based approaches that could be readily implemented for characterizing other proteins that exhibit metamorphic behavior.

In the second project, we coupled mutagenesis studies with advanced NMR techniques and X-ray crystallography to explore the binding sites of previously-identified small molecules for ARNT PAS-B. We characterized the binding of several ligands from our in-house fragment library, including KG-548 and KG-655, which antagonize ARNT PAS-B/CCC interactions. Our data indicate that KG-548 binds exclusively to the external β-sheet surface, whereas KG-655 binds to the same solvent-exposed site but can also enter a water-accessible internal cavity. Furthermore, we showed that another ligand, KG-279, binds preferentially to the internal cavity and only weakly interacts with the surface site. While the internal binding mode resembles how other PAS domains interact with their ligands, the surface binding site is of particular interest, as it is utilized by ARNT PAS-B to simultaneously mediate HIF-α/ARNT heterodimerization and recruit other binding partners, including CCCs. This work provides the structural basis for designing more potent ARNT-specific small molecule regulators.

To date, all the structural information on ARNT/CCC interactions has come from divide-and-conquer studies on individual domains, leading to several proposed models which conflict with each other. For our final project, we expanded our research beyond isolated ARNT PAS-B and investigated the assembly of a more complete HIF-2 heterodimeric complex to explore the inter-molecular interactions between HIF-2α, ARNT, and CCCs. To accurately map the binding interfaces, we utilized an integrative approach that combined data from cryo-EM, HDX-MS, and biophysical solution binding assays. At first, we used cryo-EM to examine the HIF-2 heterodimer in the absence of CCCs and discovered that ARNT PAS-B was highly flexible, which allowed the domain to sample an "open" conformation detached from the other domains. Upon adding CCCs of different boundaries, the resulting HIF-2/CCC complexes exhibited significant conformational heterogeneity, which prevented the reconstruction of high-resolution models. Nonetheless, evidence was clear that CCCs interact with the PAS domain regions of HIF-2 via their C-terminal ends. In addition, we quantitatively characterized the binding of CCCs using MST, and validated that CCCs directly interact with ARNT PAS-B. Taken together, we proposed that CCCs are recruited to HIFs by leveraging the structural motions within the PAS-B domain and flexible linkers of ARNT. This hypothesis was supported by the identification of a novel coactivator binding site on the β-sheet surface of ARNT PAS-B. Overall, this work further exemplifies the importance of ARNT in mediating HIF function.

In conclusion, the studies presented in this thesis provided a comprehensive analysis of the ARNT PAS-B dynamics and its interactions with various ligands and CCCs. These findings shed light on the molecular mechanisms underlying HIF function and regulation and laid the groundwork for developing more potent HIF inhibitors exploiting ARNT/coactivator interactions.

Recommended Citation

Xu, Xingjian, "Targeting HIF/Coregulator Complexes in Cancer: Characterizing Fragile Fold and Inhibitor Binding Sites of ARNT" (2023). CUNY Academic Works.
https://academicworks.cuny.edu/gc_etds/5367