Characterisation of class VI TRIM RING domains: linking RING activity to C-terminal domain identity

Abstract

TRIM E3 ubiquitin ligases regulate multiple cellular processes, and their dysfunction is linked to disease. They are characterised by a conserved N-terminal tripartite motif comprising a RING, B-box domains, and a coiled-coil region, with C-terminal domains often mediating substrate recruitment. TRIM proteins are grouped into 11 classes based on C-terminal domain identity. Class VI TRIMs, TRIM24, TRIM33, and TRIM28, have been described as transcriptional regulators, a function linked to their C-terminal plant homeodomain and bromodomain, and independent of their ubiquitination activity. It is unclear whether E3 ligase activity is regulated in family members where the C-terminal domains function independently. Here, we provide a detailed biochemical characterisation of the RING domains of class VI TRIMs and describe the solution structure of the TRIM28 RING. Our study reveals a lack of activity of the isolated RING domains, which may be linked to the absence of self-association. We propose that class VI TRIMs exist in an inactive state and require additional regulatory events to stimulate E3 ligase activity, ensuring that associated chromatin-remodelling factors are not injudiciously degraded.

Figure 1.. Multiple sequence alignment of TRIM RING domains.

Multiple sequence alignment of the class VI RING domains with those TRIM RING domains for which there are structures in the PDB. PDB codes are indicated in brackets. The secondary structure of the TRIM25 RING is shown above the alignment. The red asterisk denotes the conserved hydrophobic residue, which is important for E2 binding in various RINGs. The blue asterisk denotes the conserved “linchpin” residue, which forms hydrogen bonds with both E2 and ubiquitin.

Figure 2.. Isolated class VI RING domains lack catalytic activity.

(A, C, E) E2–Ub^Atto discharge assays with class VI RING domains and different E2 enzymes as indicated. Assays were carried out with TRIMs as indicated, and reactions were monitored over 60 min. (B, D, F) Quantification of the discharge assays: the loss of E2–Ub^Atto is plotted as the average of experimental triplicates or duplicates (±SD). The discharge assays were run in parallel to those shown in Figs 3D and S2. Some of the panels that act as negative and positive controls are shown in two figures. Source data are available for this figure.

Figure S1.. Concentration dependence of E2–Ub discharge assays.

(A, C, E) E2–Ub^Atto discharge assays with increasing concentrations of class VI RING domains. Assays were carried out with 4 μM of either TRIM25 or TRIM32 as a positive control, and up to 100 μM of either TRIM28 RING, TRIM24 RING, or TRIM33 RING as indicated. Reactions were monitored over 15 min. (B, D, F) Quantification of discharge assays: the loss of E2–Ub^Atto is plotted as the average of experimental duplicates. Source data are available for this figure.

Figure 3.. Oligomeric state of class VI RINGs and link to catalytic activity.

(A–C) SEC‐MALLS traces of TRIM28, TRIM24, and TRIM33. The constructs were analysed over different concentration ranges, as indicated. (D, F) UBE2D–Ub^Atto discharge assays with various TRIM28 constructs ± MAGE as indicated. Reactions were monitored over 60 min. (E, G) Quantification of the discharge assays: the loss of UBE2D–Ub^Atto is plotted as the average of experimental triplicates (±SD). The discharge assays were run in parallel to those shown in Figs 2A, C, E, and S2. Some of the panels that act as negative and positive controls are shown in two figures. Source data are available for this figure.

Figure S2.. Catalytic activity of TRIM28 RING mutants.

(A, C) E2–Ub^Atto discharge assays with various TRIM28 RING constructs and different E2 enzymes as indicated. Reactions were monitored over 60 min. (B, D) Quantification of the discharge assays: the loss of E2–Ub^Atto is plotted as the average of experimental triplicates or duplicates (±SD). The discharge assays were run in parallel to those shown in Figs 2A, C, E, and 3D. Some of the panels that act as negative and positive controls are shown in two figures. Source data are available for this figure.

Figure S3.. Amide regions of the 1D ¹H-NMR spectra of different RING constructs.

Comparison of the amide regions of the 1D ¹H-NMR spectra of TRIM28 RING, RING–RING fusion, and R69L mutant (A), TRIM25 RING and its L17R mutant (B), TRIM32 RING and its M24R mutant (C).

Figure S4.. Auto-ubiquitination activity of full-length TRIM28 with a range of E2 conjugating enzymes.

Upper panel: full-length TRIM28 was incubated with E1, ubiquitin, ubiquitin^Atto, and a range of E2 enzymes as indicated. The reactions were monitored for auto-ubiquitination over 60 min. The TRIM25 RBCC was included in conjunction with UBE2D as a positive control. Lower panel: the results of the same assay carried out in the absence of any E3 ligase. The band labelled by an asterisk is likely an auto-ubiquitinated form of UBE2K.

Figure S5.. NMR interaction analysis of TRIM28 RING domain with UBE2D, UBE2D-Ub, and Ubiquitin.

(A) ¹H-¹⁵N HSQC spectrum of ¹⁵N-TRIM28 RING. (B–D) Magnification of the area of the HSQC spectrum highlighted in (A) overlaid with the spectra resulting from the end-point titration with either unlabelled UBE2D (B), UBE2D–Ub (C), or ubiquitin (D).

Figure 4.. Structural characterisation of the TRIM28 RING domain.

(A) Residue-specific ¹⁵N NMR relaxation parameters obtained for TRIM28 RING. (B) Left panel shows the superimposition of the NMR-derived 20 lowest energy conformers of TRIM28 RING. Right panel shows the lowest energy conformer with structural features and zinc ions indicated. (C) The solvent-subtracted SAXS profile, Kratky plot and P(r) distribution for TRIM28 RING (left panel), and the SAXS-derived envelope superimposed to the lowest energy conformer calculated from NMR data (right panel). (D) Superimposition of the TRIM28 RING NMR structure with the crystal structure of the TRIM25 RING (5FER) (left panel) and the TRIM32 RING (5FEY) (right panel).

Figure 5.. Interaction of TRIM28 RING with E2–ubiquitin.

(A, C) ¹H-¹⁵N chemical shift perturbations (δΔ) of ¹⁵N-TRIM28 RING residues upon titration with unlabelled UBE2D–Ub (A) or ubiquitin (C) plotted against the residue number. (B, D) Residues that experience chemical shift changes upon titration of either UBE2D (B) or ubiquitin (D) are mapped onto the surface of TRIM28 RING. Residues are coloured to represent the relative change in chemical shift and denoted by the coloured horizontal lines on the plots (A, C).

Figure 6.. Mutational analysis of RING domain catalytic activity.

(A, C, E) UBE2D–Ub^Atto discharge assays with TRIM28, TRIM25, and TRIM32 RING mutants. Reactions were monitored over 60 min. (B, D, F) Quantification of the discharge assays: the loss of UBE2D–Ub^Atto is plotted as the average of experimental triplicates (±SD). Source data are available for this figure.