Generation of Vps23 and Vps36 mutants incapable of binding Ub. (A) Model of the Vps23 UEV domain in complex with Ub (Protein Data Bank accession no. 1UZX). The residues mutated to create the vps23^ΔUb alleles are shown in red. (bottom) A schematic of the position of the UEV domain and the core region of Vps23 involved in complex formation with other ESCRT-I subunits is shown. (B) Recombinant V5 epitope–tagged wild-type and mutant Vps23 UEV domains were used for binding experiments with GST, Ub-GST, or GST-Vps27 C terminus. Bound proteins were immunoblotted with 1 or 10% of the input lysates. (C) Cell lysates from strains (PLY335, PLY3529, and PLY3530) bearing the wild-type (WT) or mutant alleles of VPS23 were immunoblotted with α-Vps23 and α-PGK (3-phosphoglycerate kinase) antibodies. (D) Schematic of Vps36 showing the GLUE domain with an insertion of two NZF domains, the second of which binds Ub. The core region of Vps36 interacts with the rest of ESCRT-I. (E) Summary of HSQC NMR experiments with ¹⁵N-labeled Ub (25 µM) with the GLUE domains from wild-type Vps36 GLUE domain (40 µM) or Vps36^ΔUb (240 µM). Chemical shift differences were quantified and plotted on the Ub sequence (right) or mapped on the Ub surface (left). Red indicates (0.2δN² + δH²)^1/2 ≥ 0.03. Black residues were not observed. The models below show the NMR structure of Ub with the Npl4 NZF domain and a predicted structure of the Vps36 NZF domain. Yellow residues are the position of T₁₈₇F₁₈₈. (F) Portion of the HSQC spectra of ¹⁵N-labeled Ub in the presence (red) and absence (green) of GST fusions of Vps36 and Vp36^ΔUb. ppm, parts per million. (G) V5 epitope–tagged Vps36 GLUE domains of wild-type or ΔUb were assayed for Ub binding with Ub-GST. Bound proteins were immunoblotted with 10% of the input lysates. (H) Cell lysates from wild-type strains expressing HA epitope–tagged alleles of wild-type VPS36 and vps36^ΔUb from low copy plasmids were immunoblotted with anti-HA and α-PGK.