RING domains act as both substrate and enzyme in a catalytic arrangement to drive self-anchored ubiquitination

Abstract

Attachment of ubiquitin (Ub) to proteins is one of the most abundant and versatile of all posttranslational modifications and affects outcomes in essentially all physiological processes. RING E3 ligases target E2 Ub-conjugating enzymes to the substrate, resulting in its ubiquitination. However, the mechanism by which a ubiquitin chain is formed on the substrate remains elusive. Here we demonstrate how substrate binding can induce a specific RING topology that enables self-ubiquitination. By analyzing a catalytically trapped structure showing the initiation of TRIM21 RING-anchored ubiquitin chain elongation, and in combination with a kinetic study, we illuminate the chemical mechanism of ubiquitin conjugation. Moreover, biochemical and cellular experiments show that the topology found in the structure can be induced by substrate binding. Our results provide insights into ubiquitin chain formation on a structural, biochemical and cellular level with broad implications for targeted protein degradation.

Fig. 1. Structure of initiation of RING-anchored ubiquitin chain elongation.

a Side and top view of the Ub-R:Ube2N~Ub:Ube2V2 structure (Ub-R, Ub in red, R in blue, Ube2N~Ub, Ube2N in green, Ub in orange, and Ube2V2 in teal). Chains drawn as cartoon represent the asymmetric unit. b The canonical model of initiation of RING-anchored ubiquitin chain elongation. c Schematic cartoon, representing the canonical model of RING-anchored ubiquitin chain elongation shown in b. Symmetry mates are denoted by ′ next to the label. Ub ubiquitin, Ub-R ubiquitin-RING.

Fig. 2. Chemical mechanism of ubiquitination.

a Magnified regions of the active site of Ube2N~Ub/Ube2V2 (Ube2N in green, donor Ub in orange, acceptor Ub in red, and Ube2V2 in teal). Stereo-images are shown in Supplementary Fig. 1b. b Chemical scheme for the activation of the acceptor lysine. c Acid coefficients (pK_a), d K_M, and e k_cat of di-ubiquitin formation by Ube2N/V2 are presented as best fit + standard error. Fits were performed using data of n = 3 technical replicates, and are shown in Supplementary Fig. 2 and Source data. Ub ubiquitin.

Fig. 3. The mechanism of RING-anchored ubiquitination in trans.

a Surface representation of the canonical model of the Ub-R:Ube2N~Ub:Ube2V2 structure (Ub-R, Ub in red, R in blue, Ube2N~Ub, Ube2N in green, Ub in orange, and Ube2V2 in teal). b Domain architecture of TRIM21 constructs used in biochemical assays. c Cartoon models of substrate (Fc, gray) engagement by TRIM21 constructs (blue). The structural basis for these models is shown in Supplementary Fig. 8. d Substrate (Fc)-induced self-ubiquitination assay of 100 nM Ub-TRIM21 constructs. Reactions were incubated for 5 min at 37 °C. Further data can be found in Supplementary Fig. 9. * (asterisk) indicates a TRIM21 degradation product that could not be removed during purification. Western blots are representative of n = 3 independently performed experiments. Uncropped blots are provided in Source data. Ub ubiquitin, R RING, B Box, CC coiled-coil, PS PRYSPRY, kDa kilo Dalton.

Fig. 4. The mechanism of RING-anchored ubiquitination in cis.

a For ubiquitination in cis, the RING-anchored (blue) ubiquitin chain (red) must be sufficiently long to reach the active site on Ube2N~Ub/Ube2V2 (Ube2N in green, Ub in orange, and Ube2V2 in teal). The chain can go around two different routes, one shown here and the other in Supplementary Fig. 10. The ubiquitin chain was modeled using the Ub-R:Ube2N~Ub:Ube2V2 structure and a K63-linked Ub₂ structure (2JF5 (ref. )) using PyMol. b Substrate (Fc)-induced self-ubiquitination assay of 100 nM Ub_n-TRIM21 constructs. Reactions were incubated for 5 min at 37 °C. Western blots are representative of n = 3 independently performed experiments. Uncropped blots are provided in Source data. Ub ubiquitin, R RING, PS PRYSPRY, kDa kilo Dalton.

Fig. 5. Catalytic RING topology drives targeted protein degradation.

a Schematic cartoons showing the topology of TRIM21 (blue) on GFP-Fc (green and gray, respectively). b, c GFP-Fc degradation assay. b Western blot of RPE-1 TRIM21-knockout cells transiently expression GFP-Fc and a series of TRIM21 constructs. Western blots are representative of n = 2 independently performed experiments. c Shown is the flow cytometry analysis of green fluorescence of RPE-1 TRIM21-knockout cells transiently expressing GFP-Fc and a series of TRIM21 constructs. After electroporation, each population of cells was split in two and either treated with MG132 or DMSO. Data are presented as mean ± standard error of the mean. Each data point in the graph represents one biologically independently performed experiment (n = 3 (for mCh-CC-PS, R-R-B-CC-PS, and R-PS) or 4 (R-B-C-C-PS and R-R-PS)). A two-tailed unpaired Student’s T test was performed to assess the significance of fluorescence reduction relative to mCh-CC-PRYSPRY (P values: R-B-CC-PRYSPRY, 0.0797 (ns); R-R-B-CC-PRYSPRY, 0.02366 (ns); R-PRYSPRY, 0.4964 (ns); and R-R-PRYSPRY, 0.0035 (**)). d, e Trim-Away of Caveolin-1-mEGFP (Cav1-GFP) in NIH 3T3 GFP-Cav-1-knock in cells. Shown in d is the normalized GFP fluorescence (error bars represent ± SEM of four images) and in e the western blot after the experiment. Data in d, e are representative of n = 2 independent experiments. Uncropped blots and raw data are provided in Source data. R RING, B Box, CC coiled-coil, PS PRYSPRY, mCh mCherry, kDa kilo Dalton, ns not significant.

Fig. 6. TRIM protein assembly on viruses.

Cartoon models of the assembly of TRIM5 (a, b) and TRIM21 (c, d) on viral capsids. a Shown is the hexagonal assembly of TRIM5 on HIV-1 capsid as imaged by cryo-electron tomography. c Assembly of TRIM21:antibody complexes on adenovirus capsid (adenoviral measurements are based on 6B1T). b, d Cartoons visualizing how the TRIM protein assembly on the viral capsid enables the formation of the catalytic RING topology. R RING, B Box, CC coiled-coil, PS PRYSPRY, Ub ubiquitin.