Dissertations, Theses, and Capstone Projects

Date of Degree

9-2026

Document Type

Doctoral Dissertation

Degree Name

Doctor of Philosophy

Program

Biochemistry

Advisor

Kevin Gardner

Committee Members

Amédée des Georges

Akira Kawamura

Jill Bargonetti

Carrie Partch

Subject Categories

Biochemistry | Cell Biology | Integrative Biology | Structural Biology

Keywords

Hypoxia inducible factors, Aryl hydrocarbon receptor nuclear translocator, cryo-EM, small molecule therapeutics, hypoxia

Abstract

Hypoxia inducible factors (HIFs) are a class of basic helix-loop-helix PER-ARNT-SIM (bHLH-PAS) transcription factors that regulate oxygen-dependent cellular processes in response to low oxygen levels. HIFs are heterodimeric complexes comprised one of three isoforms of an oxygen-sensitive alpha subunit (HIF-1α, -2α, -3α) and a constitutively expressed beta subunit (HIF-1β, or aryl hydrocarbon nuclear translocator [ARNT]) that complex in the nucleus under hypoxic conditions. Dysregulation of these complexes can lead to several clinical complications, such as tumorigenesis, anemia and certain types of cancers such as renal cell carcinoma, making HIFs intriguing therapeutic targets in the discovery and development of anticancer therapeutics. Here, we structurally and biochemically characterize interactions between HIFs of varying degrees of complexity and small molecule, protein and DNA binding partners with a focus on discovering novel HIF assemblies and potentially developing new modulators of HIF activity.

First, we analyzed interactions between the isolated ARNT PAS-B domain and three small molecule binders, KG-548, KG-655 and KG-279, using a mix of NMR binding assays and molecular dynamics (MD) simulations to structurally and biophysically assess binding. Using NMR-guided MD simulations, we determined binding sites of KG-655 and KG-279 to be within the small cavity inside the domain, causing a significant shift of the Fα-helix to accommodate binding. In contrast, KG-548 and KG-655 bound on the external face of the ARNT PAS-B beta sheet. Additionally, we determined binding affinities of KG-548 and KG-655 and found that all three ligands could induce homodimerization of the ARNT PAS-B domain at low protein concentrations, suggesting their potential to act as molecular glues or tool compounds.

Second, we probed HIF complex assembly in more functional contexts than previously appreciated, building on original identifications of HIF binding sites on DNA which suggested the involvement of regions outside the minimal 6 bp hypoxia response element (HRE) sites used in many subsequent functional and structural studies. We solved a structure of HIFs bound to human EPO DNA sequences including both an HRE and a secondary downstream HRE adjacent sequence (HAS) important to complete HIF function, finding that two individually assembled HIF heterodimers bound symmetrically to the same DNA sequence at the HRE and HAS-boxes forming a complex we termed a “dimer-of-heterodimers.” We found that this dimer-of-heterodimers complex largely interacts through the ARNT PAS-B domains and thus can assemble in HIF-1 and HIF-2 and as a mixed HIF-1/HIF-2 complex. We also probed HIF recruitment of coiled-coil coactivators, finding that CCCs largely associate with disordered loops along the HIF-α and ARNT subunits, though this was not seen in the larger dimer-of-heterodimers complex. With this work, we suggest that HIFs form alternative assemblies to regulate gene expression under different cellular conditions, controlled by HIF concentrations and the presence of coactivators.

Finally, we aimed to expand upon the ARNT PAS-B binding small molecules discussed in the first project to develop them into inhibitors of all three HIF complexes. We developed and analyzed KG-548 derivatives KP-004, KP-107 and KP-143, finding all three capable of disrupting HIF-α/ARNT dimerization in both HIF-1 and HIF-2 complexes with varying potencies as high as 2 µM. We solved a crystal structure of KP-004 bound to the β-sheet surface of the ARNT PAS-B domain, suggesting these compounds act as competitive inhibitors of dimerization. Additionally, we determined that all three inhibitors were minimally impacted by ARNT mutant F446L, a mutant that suppresses the effects of clinically-utilized HIF-2α inhibitors such as belzutifan. This work details progress in the development of pan-HIF inhibitors, as well as providing insights into the mechanistic role of ARNT residue F446 in HIF inhibition.

This work is embargoed and will be available for download on Saturday, September 30, 2028

Share

COinS