Abstract
The Aster proteins (encoded by the Gramd1a-c genes) contain a ligand-binding fold structurally similar to a START domain and mediate nonvesicular plasma membrane (PM) to endoplasmic reticulum (ER) cholesterol transport. In an effort to develop small molecule modulators of Asters, we identified 20α-hydroxycholesterol (HC) and U18666A as lead compounds. Unfortunately, both 20α-HC and U18666A target other sterol homeostatic proteins, limiting their utility. 20α-HC inhibits sterol regulatory element-binding protein 2 (SREBP2) processing, and U18666A is an inhibitor of the vesicular trafficking protein Niemann-Pick C1 (NPC1). To develop potent and selective Aster inhibitors, we synthesized a series of compounds by modifying 20α-HC and U18666A. Among these, AI (Aster inhibitor)-1l, which has a longer side chain than 20α-HC, selectively bound to Aster-C. The crystal structure of Aster-C in complex with AI-1l suggests that sequence and flexibility differences in the loop that gates the binding cavity may account for the ligand specificity for Aster C. We further identified the U18666A analog AI-3d as a potent inhibitor of all three Aster proteins. AI-3d blocks the ability of Asters to bind and transfer cholesterol in vitro and in cells. Importantly, AI-3d also inhibits the movement of low-density lipoprotein (LDL) cholesterol to the ER, although AI-3d does not block NPC1. This finding positions the nonvesicular Aster pathway downstream of NPC1-dependent vesicular transport in the movement of LDL cholesterol to the ER. Selective Aster inhibitors represent useful chemical tools to distinguish vesicular and nonvesicular sterol transport mechanisms in mammalian cells.