Insight into the binding mode of HIF-2 agonists through molecular dynamic simulations and biological validation

Hypoxia-inducible factor-2 (HIF-2), a heterodimeric transcriptional protein consisting of HIF-2α and aryl hydrocarbon receptor nuclear translocator (ARNT) subunits, has a broad transcriptional profile that plays a vital role in human oxygen metabolism. M1001, a HIF-2 agonist identified by high-throughput screening (HTS), is capable of altering the conformation of Tyr281 of the HIF-2α PAS-B domain and enhancing the affinity of HIF-2α and ARNT for transcriptional activation. M1002, an analog of M1001, shows improved efficacy than M1001. However, the cocrystal structure of M1001 and HIF-2 has some defects in revealing the agonist binding mode due to the relatively low resolution, while the binding mode of M1002 remained unexplored. To in-depth understand agonist binding profiles, herein, the molecular dynamic (MD) simulations was applied to construct a stable agonist-protein model, and a possible binding mode was proposed through the analysis of the binding free energy and hydrogen bonding of the simulation results. Nine compounds were then synthesized and evaluated to verify the proposed binding mode. Among them, compound 10 manifested improved agonistic activity and reduced toxicity compared to M1002. This study provides deep insight into the binding mode of such HIF-2 agonists, which would be useful for designing novel agonists for HIF-2.