Cryo-EM structures of Gid12-bound GID E3 reveal steric blockade as a mechanism inhibiting substrate ubiquitylation

Abstract

Protein degradation, a major eukaryotic response to cellular signals, is subject to numerous layers of regulation. In yeast, the evolutionarily conserved GID E3 ligase mediates glucose-induced degradation of fructose-1,6-bisphosphatase (Fbp1), malate dehydrogenase (Mdh2), and other gluconeogenic enzymes. "GID" is a collection of E3 ligase complexes; a core scaffold, RING-type catalytic core, and a supramolecular assembly module together with interchangeable substrate receptors select targets for ubiquitylation. However, knowledge of additional cellular factors directly regulating GID-type E3s remains rudimentary. Here, we structurally and biochemically characterize Gid12 as a modulator of the GID E3 ligase complex. Our collection of cryo-EM reconstructions shows that Gid12 forms an extensive interface sealing the substrate receptor Gid4 onto the scaffold, and remodeling the degron binding site. Gid12 also sterically blocks a recruited Fbp1 or Mdh2 from the ubiquitylation active sites. Our analysis of the role of Gid12 establishes principles that may more generally underlie E3 ligase regulation.

Fig. 1. Gid12 binds to the Gid4-containing GID E3 ligase complexes.

a Quantitative MS identification of proteins interacting with Ipf1-3×FLAG (originally YDL176W, now called Gid12) or an untagged control strain under glycolytic conditions (n = 3 biologically independent samples). Data are log-transformed ratios of protein LFQ intensities versus −log10-transformed P values of two-tailed Student’s t tests. The hyperbolic curve separates specifically interacting proteins from background (square; false-discovery-rate-adjusted P = 0.05; minimal fold change S0 = 0.1). The bait protein Gid12 is highlighted in blue. b SDS-PAGE of GST- or Strep-affinity purifications after co-infecting insect cells with all GID E3 ligase subunits except the one indicated above the lane. Lane 1 shows the result from a control co-infection with all subunits (n = 3 biologically independent experiments). c Coordinates of Gid12 bound to the substrate-receptor and scaffolding module (Gid12-SRS) were docked into the cryo-EM density maps of Gid12-SRS, Gid12 bound to GID^SR4 (Gid12-GID^SR4) and Chelator-GID^SR4 (Gid12-Chelator-GID^SR4) at 3.3, 9.8, and 19.4 Å resolution, respectively, from left to right. Gid12 is colored in blue, Gid4 in gold, Gid5 in magenta, Gid8 in salmon, and Gid1 in green.

Fig. 2. Gid12 binds both the substrate receptor Gid4 and the tip of Gid5 in the scaffolding module.

a Overall structure of Gid12-SRS. Gid12 is colored in blue, Gid4 in gold, Gid5 in magenta, Gid8 in salmon, and Gid1 in green. Dotted wedge highlights that Gid12 projects from the scaffolding module such that the substrate receptor, Gid4, is enwrapped around 305° in the complex. Closeup highlights interactions between Gid12, Gid4, and Gid5. b Gid12 forms a 7-bladed β-propeller, with its top face and central pore extensively interacting with Gid4, and edge binding Gid5. The β-propeller blades are labelled 1–7. c Gid12 structural elements, showing blades 1–7 from bottom and top faces of the β-propeller, color-coded in blue, green, hot pink, burnt orange, slate gray, khaki, and red, respectively. A meandering insertion in the 3rd blade (Ins3), indicated by the arrow, plugs the central pore from the bottom face of β-propeller (left); Insertion between the 3rd and 4th β-sheets in blade 4 (Ins4), insertion between blades 4 and 5 (Ins4-5), and insertion between blades 5 and 6 (Ins5-6) project outward from the center of the β-propeller (right). d Gid12 surface representation, colored by electrostatic surface potential. The bottom face is highly charged (red), whereas the top face, the surface of Ins3-Ins4 5-helix bundle, and central pore (see closeup), which interact with Gid4, are substantially hydrophobic (white). e Closeups of the superposition of Gid12-SRS with a prior structure without Gid12 (cyan, PDB ID: 6SWY) show that a ≈30-residue Gid5 disordered region (DR)—invisible in the published cryo-EM maps without Gid12—binds an edge of Gid12 formed by blades 1 and 7. f Closeups showing extensive interactions between Gid12, shown as blue cryo-EM density, and Gid4 shown as gold ribbon. The Gid4 β-barrel interacts with the top face of Gid12. The Gid4 3-helix bundle (3HB) interacts with Ins4-5 and Ins5-6 from Gid12.

Fig. 3. Gid12 remodels the N-degron binding pocket in Gid4.

a Superposition of Gid12-bound Gid4 (gold) with prior structures of Gid4 from GID^SR4 complexes without substrate (“Apo”, PDB ID: 6SWY, purple) and bound to the Pro/N-degron of Fbp1 (PDB ID: 7NS3, cyan) reveals conformational heterogeneity of Gid4 L2-loop. b Closeup showing structural differences in Gid4 L2-loop when bound to Gid12 (Gid4 in gold) or to Pro/N-degron of Fbp1 substrate (Gid4 in cyan, PDB ID: 7NS3). c L2-loop structural similarity amongst human Gid4 structures with various N-degron peptides (PDB ID: 6CCT, 6CCU, 6CD8, 6CD9, 6CDC, 6CDG, 6WZX, 6WZZ; colored in cyan, pink, lawn green, salmon, gray, yellow, deep pink and blue respectively). Yeast Gid4 from apo GID^SR4 (6SWY) is shown in purple with its disordered L2-loop indicated with arrow and dotted line, and human Gid4 Leu164, which corresponds to yeast Gid4 Leu172, are shown for reference. d Closeups showing Gid4 L2-loop (gold) bound to Gid12 (blue), with Gid12 depicted as ribbon (left) or surface (right). The Gid4 L2-loop conformation complements the Gid12 pocket at the junction between blades 7 and 1 in the central pore.

Fig. 4. Gid12 fully and partially obstructs substrate placement in Chelator-GID^SR4 and GID^SR4 assemblies, respectively.

a Superposition of 3D reconstructions of Gid12-Chelator-GID^SR4 (transparent gray, with Gid12 in blue) and Chelator-GID^SR4-Fbp1 (solid gray, with substrate Fbp1 in red) shows that Gid12 fills much of the central region of Chelator-GID^SR4 that encapsulates the globular domain of the recruited Fbp1 substrate. b Superposition of 3D reconstruction of Gid12-GID^SR4 with the corresponding region of the map of Chelator-GID^SR4-Fbp1. The relatively open structure of GID^SR4 does not encapsulate Fbp1, whose globular domain may thus occupy various relative positions between the Pro/N-degron recruited to Gid4 and RING-based catalytic module. c Cartoon representation showing how GID^SR4 could dynamically accommodate various orientations of Fbp1 between Gid4 and RING-activated Ubc8~ubiquitin intermediate, some of which could be compatible with Gid12 bound to Gid4.

Fig. 5. Cryo-EM maps of a ≈ 5 MDa 60-subunit Cage-GID^SR4 assembly.

a Overall architecture of Gid12-Cage-GID^SR4 is seen from its transparent cryo-EM density fit with three copies of models of Gid12-Chelator-GID^SR4 in solid surfaces in different shades of gray, with models for Gid12 (this study) in blue. b 19.8 Å-resolution cryo-EM density for Cage-GID^SR4 is shown as a transparent surface fit with three copies of models of Chelator-GID^SR4 (EMD-12541) in solid surfaces in different shades of gray. c Models of Gid1 and Gid7 domains fitted at junctions of the three individual Chelator-GIDs. These interaction at the junction constitute the cage architecture. The resolution of the density, together with prior maps and models, allows attributing domains but not their specific orientations or interactions. d 12 Å-resolution cryo-EM density map of Cage-GID^SR4-Fbp1 (here, the N-terminal degron of Fbp1 - PTLVNG was exchanged with Mdh2 degron - PHSVTP to increase stability). Fbp1 tetramer encapsulated in the oval center of each Chelator-GID^SR4 is color-coded in red.

Fig. 6. Gid12 biochemically modulates both assembly of Gid4 into the GID E3 ligase complex and its activity toward its substrates.

a Strep-tagged Gid4 in solution can replace untagged Gid4 within GID^SR4 but not in the Gid12-bound complex. A mutant (mut) lacking the C-terminal four residues 359–362 required for incorporation into GID^SR4 is used as a control for specificity of Strep-Gid4 incorporation into the complex (n = 3 biologically independent experiments). b Pulse chase assay examining effects of Gid12 on GID ubiquitin transferase activity in substrate-independent manner. Ubc8 is C-terminally fluorescently labeled with TAMRA for fluorescence detection of the free E2 versus the version thioester-linked to Ub. First, the thioester-linked Ubc8-TAMRA~Ub intermediate was generated in a pulse reaction. Second, the chase reaction monitors formation of the faster-migrating Ubc8-TAMRA upon Ub discharge to free lysine, stimulated by GID^SR4 with or without a bound Gid12 (n = 3 biologically independent experiments). c In vitro ubiquitylation assay of model peptide substrates (with an N-terminal substrate degron, a flexible linker, a single lysine at optimal position to accept ubiquitin, and a C-terminal fluorescein for fluorescent detection) testing effects of adding purified Gid4 alone or Gid4 co-expressed with Gid12 to GID^Ant (n = 3 biologically independent experiments). d 18.9 Å-resolution cryo-EM density map of Chelator-GID^SR4-Mdh2. Mdh2 dimer encapsulated in the oval center of Chelator-GID^SR4 is color-coded in black. e Experiment as in (c), except with the indicated full-length metabolic enzyme substrates, each appended to a C-terminal fluorescein for fluorescent detection (n = 3 biologically independent experiments).

Fig. 7. Overexpression of Gid12 stabilizes Fbp1 in vivo.

a Expression profiles of Gid4 and Gid12 tagged at their endogenous loci, under different growth conditions, monitored by western blotting against hemagglutinin tag (HA). YPD: glycolytic conditions; YPE: overnight non-fermentable carbon source starvation; Carbon Recovery: glucose replenished with YPD. Pgk1 served as loading control (n = 3 biologically independent experiments). b Degradation of 3×HA chromosomally tagged Gid4 in wildtype (WT) and Gid12-overexpressing (GPD promoter) cells were monitored using anti-HA immunoblotting. Gid4 was stable under carbon recovery conditions in a strain continuously overexpressing Gid12 (n = 3 biologically independent experiments). c Quantitative mass spectrometry analysis of the relative levels of Gid2, Gid4 and Gid12 in anti-HA immunoprecipitates of lysates expressing Gid4 with an N-terminal 3×HA tag at the endogenous loci. Yeast cells were grown under glycolytic (YPD), non-fermentable carbon source starvation (YPE), or carbon recovery conditions. Data are mean ± s.d. of n = 3 biologically independent experiments. iBAQ ratios are formed using the respective Gid2 iBAQ value as denominator. d Monitoring degradation of 3×FLAG chromosomally tagged gluconeogenic enzymes in wildtype (WT), Gid12-deficient (gid12∆), and Gid12-overexpressing (ADH1 promoter) cells using anti-FLAG immunoblotting. The effects of GID12 deletion to gluconeogenic enzymes are subtle, however, all four gluconeogenic enzymes are stabilized under carbon recovery conditions in a strain consecutively overexpressing Gid12 (n = 3 biologically independent experiments).