Architecture of the UBR4 complex, a giant E4 ligase central to eukaryotic protein quality control

Abstract

Eukaryotic cells have evolved sophisticated quality control mechanisms to eliminate aggregation-prone proteins that compromise cellular health. Central to this defense is the ubiquitin-proteasome system, where UBR4 acts as an essential E4 ubiquitin ligase, amplifying degradation marks on defective proteins. Cryo-electron microscopy analysis of UBR4 in complex with its cofactors KCMF1 and CALM1 reveals a massive 1.3-megadalton ring structure, featuring a central substrate-binding arena and flexibly attached catalytic units. Our structure shows how UBR4 binds substrate and extends lysine-48-specific ubiquitin chains. Efficient substrate targeting depends on both preubiquitination and specific N-degrons, with KCMF1 acting as a key substrate filter. The architecture of the E4 megacomplex is conserved across eukaryotes, but species-specific adaptations allow UBR4 to perform its precisely tuned quality control function in diverse cellular environments.

Fig. 1. Architecture of the human UBR4 complex.

Composite cryo-EM density map and model colored by protein component and domain as indicated. A schematic domain architecture of UBR4 and KCMF1 is shown below. Under this are detailed views of structural features.

Fig. 2. Mechanism of ubiquitin chain extension by the UBR4 complex

(A) E4 ligase assay showing formation of K0-Ub-Ub* di-ubiquitin from K0-Ub and Ub* catalysed by 200 nM human UBR4 complex when all components of the assay are added for 45 minutes. A schematic explaining the assay and the two ubiquitin components is shown to the left. (B) Cryo-EM density map and model of the structure of human UBR4 complex obtained in the presence of UBE2A. An inset shows a detailed view of the interaction highlighting hydrophobic residues on the E2-binding helix of UBR4 which are mutated in the E2B-5A UBR4 complex variant. (C) Model of the E4 transfer state using our structure overlaid with a crystal structure of the hemiRING-UBE2A complex (PDB pdb_00008BTL) and the cryo-EM structure of the yeast UBR1-UBE2A-Ub_D complex (PDB pdb_00007MEX), both aligned on UBE2A, and an AlphaFold3 model of the UBL-Ub interaction (Fig. S5A) aligned on the UBL domain. The three hydrophobic residues on the UBL domain which interact with ubiquitin and are mutated in the UBL-3A variant are indicated. (D) E4 ligase assay as in panel A with 100 nM of the indicated variant complexes for 45 minutes. (E) Zoomed out image of the E4 transfer state model showing the position of the C-terminal tail of Ub relative to the UBR4 arena.

Fig. 3. KCMF1 mediates selection of UBR4 substrates.

(A) On the left, a cryo-EM map of the human UBR4 complex structure refined with a global mask showing fuzzy central substrate density and on the right, the ΔZZ-DZB UBR4 complex refined with the same approach. (B) MS analysis of insect cell proteins co-purified with the WT and ΔZZ-DZB UBR4 complexes where peak intensity in the WT is plotted against log₂(fold change) between the two datasets to identify proteins which are present in the WT but not the ΔZZ-DZB UBR4 complex. The core complex components are coloured blue and mitochondrial proteins are coloured red. (C) Ubiquitination assay using 200 nM UBR4 complex, UBE2A, UBE2D3 and K0-Ub for 30 minutes in the presence of 2 μM of the indicated recombinantly purified insect cell proteins with or without Ub* fused by a linker to the C-terminus. (D) E4 ligase assay with 200 nM of the human UBR4 complexes for 30 minutes with the indicated MTS sequence, or control with no MTS, fused via a linker to the N-terminus of Ub*. KR indicates a K48R mutation in the fused Ub*. Reactions contained 1:10 DyLight488 K0-Ub which was used to quantify product bands relative to the control substrate and plot these on the right (+/- SD). (E) Two overlaid crystal structures of the KCMF1^{ZZ+DZB(△linker)} domain with either an N-RC(SO₃)K peptide or an N-RTGG peptide. Only the two N-terminal residues are shown for clarity along with important residues on KCMF1 shown in stick representation. (F-G) Quantified E4 ligase assays (+/- SD) as in panel D with the indicated substrates for (F) 200 nM WT UBR4 complex for 30 minutes and (G) 200 nM ΔZZ-DZB UBR4 complex for 15 minutes. Assay gels are shown in Fig. S12A,B.

Fig. 4. Evolutionary conservation of the UBR4 complex

(A) Cryo-EM composite map and model of the C. elegans UBR4 complex coloured by subunit and domain as indicated. The domain architecture is illustrated below. (B) Detailed view of the extended interface of CeKCMF1 with CeUBR4. (C) Low resolution cryo-EM map of the CeUBR4 complex from a global 3D classification (Fig. S14) showing a bound substrate. An AlphaFold3 model of the co-purified protein TnACADM is modelled in the density. (D) Cryo-EM map of the dimer interface core of the AtUBR4 complex coloured by domain and subunit as indicated. The domain architecture derived from bioinformatic comparison with the human UBR4 complex is shown below. (E) Crystal structure of the AtUBR4 ZZ domain in complex with an N-RSS peptide superimposed with the HsKCMF1 ZZ domain crystal structure in complex with an N-RTGG peptide. Sequence differences in the N-degron binding pocket are highlighted.