Cells are able to respond to external stimuli and sense the surrounding environment by varying the lipid and protein composition of the plasma membrane. One way this is achieved is through receptor downregulation via the endocytic pathway. Once receptors are at the early endosome, the endosomal sorting complexes required for transport (ESCRT) machinery identifies and clusters ubiquitinylated receptors, induces membrane deformation, and drives vesicle scission. This intricate process generates intralumenal vesicles (ILVs) within an endosome, known as a multivesicular endosome (MVE). Upon ILV formation, receptors are sequestered from cytoplasmic effectors. A MVE fuses with the lysosome where the contents are degraded. A great deal of emphasis has been placed on understanding how the ESCRT machinery carries out this membrane deformation and scission process, but a cohesive mechanism remains unknown. In this thesis, I demonstrate that ESCRT-II and Vps20 generate a curvature sensing supercomplex that restricts Vps32 filament formation to highly curved membranes. Membrane remodeling is observed through this filament formation. I also show that ESCRT-II and Vps20 interact in solution, which raises the possibility that a curvature sensing supercomplex may exist within the cytoplasm. Furthermore, I demonstrate that Vps20 is in an open conformation in solution unlike the downstream ESCRT-III subunit, Vps24. This crucial finding suggests that all ESCRT-III subunits are not in the same conformation, which could help explain the distinct function of each ESCRT-III protein. Finally, I show that unmodified Vps32 can self-assemble independent of upstream ESCRT components and characterize a mechanism of Vps32 filament assembly, in vitro.