Abstract by Samrina Hussain

Aldo keto reductase 1C3 (AKR1C3) is a key enzymatic driver of castration resistant prostate cancer through its dual role in intratumoral androgen biosynthesis and prostaglandin metabolism. Despite extensive medicinal chemistry efforts, the development of clinically viable AKR1C3 inhibitors has been hampered by high sequence and structural similarity among AKR1C isoforms and by the intrinsic conformational adaptability of the enzyme active site. This thesis addresses these challenges by establishing a robust structural biology pipeline to systematically interrogate the molecular determinants of AKR1C3 inhibition and selectivity. Human AKR1C3 was recombinantly expressed, purified, and crystallized in complex with NADP⁺ and a chemically diverse panel of inhibitors, including synthetic scaffolds derived from structure guided design and natural product flavonoids. High resolution X ray crystal structures were determined, enabling direct visualization of ligand binding modes, subpocket engagement, and protein conformational responses. Across all productive complexes, inhibitors were anchored via conserved interactions at the catalytic oxyanion site involving Tyr55, His117, and the NADP⁺ cofactor, while selectivity and affinity were governed by ligand dependent exploitation of distal subpockets SP1, SP2, and the steroid channel. Distinct inhibitor classes preferentially engaged either SP1 or SP2, demonstrating that no single subpocket is uniquely responsible for selective inhibition. Central to this plasticity is Trp227, which functions as a dynamic gatekeeper residue controlling access to the steroid channel and adjacent subpockets. Ligand dependent reorientation of Trp227, often coupled with movements of Phe306 and Phe311, modulates pocket volume and shape, defining both permissive and restrictive binding modes. Structural analysis indicates that the oxyanion site can be productively engaged by diverse ligand chemotypes, whereas bulky glycosylated natural products exceed the steric limits of the active site and do not form stable, ordered complexes. Together, this work provides an integrated structural framework for understanding AKR1C3 ligand recognition, pocket plasticity, and isoform selectivity. The results establish generalizable design principles for next generation AKR1C3 inhibitors, emphasizing controlled exploitation of subpocket flexibility and stabilization of key gatekeeper residues as critical determinants of potency and selectivity.