A ubiquitin-specific, proximity-based labeling approach for the identification of ubiquitin ligase substrates

Abstract

Over 600 E3 ligases in humans execute ubiquitination of specific target proteins in a spatiotemporal manner to elicit desired signaling effects. Here, we developed a ubiquitin-specific proximity-based labeling method to selectively biotinylate substrates of a given ubiquitin ligase. By fusing the biotin ligase BirA and an Avi-tag variant to the candidate E3 ligase and ubiquitin, respectively, we were able to specifically enrich bona fide substrates of a ligase using a one-step streptavidin pulldown under denaturing conditions. We applied our method, which we named Ub-POD, to the really interesting new gene (RING) E3 ligase RAD18 and identified proliferating cell nuclear antigen and several other critical players in the DNA damage repair pathway. Furthermore, we successfully applied Ub-POD to the RING ubiquitin ligase tumor necrosis factor receptor-associated factor 6 and a U-box-type E3 ubiquitin ligase carboxyl terminus of Hsc70-interacting protein. We anticipate that our method could be widely adapted to all classes of ubiquitin ligases to identify substrates.

Fig. 1.. Using Ub-POD to identify substrates of RING E3 ligase RAD18.

(A) Structural model of Ub (salmon) in complex with a Ub-conjugating enzyme E2 (yellow) and a Ub ligase E3 (RING) (green). The distance between the N termini of E3 and Ub is mentioned. Placing AP-tag at the N terminus of Ub and BirA at the catalytic end of RING-E3 ligase brings AP and BirA in proximity to each other. (B) Schematic representation of the Ub-POD method. (C) RAD18 Ub-POD workflow (left). HEK-293 cells transfected for 16 to 24 hours with AP-Ub and HA-BirA-RAD18 WT or HA-BirA-RAD18 C64D were exposed to UV (10 mJ/cm²) (right). Biotin was added at the time of transfection, and cells were allowed to recover 6 hours after UV irradiation. Untreated cells served as control. Lysates were subjected to SDS-PAGE and immunoblotting. Results are representative of two independent biological replicates. (D) Volcano plot of streptavidin pulldown enriched proteins isolated from cells transfected with HA-BirA-RAD18 WT and AP-Ub in the absence or presence of UV exposure (n = 3 biological replicates). Significantly altered proteins are shown in dark red or blue [false discovery rate (FDR) <0.05, log₂FC >I0.6I] and light red or blue (FDR <0.05, 0 > log₂FC < I0.6I) (moderated t test).

Fig. 2.. A variant of AP-tag increases the efficiency of Ub-POD.

(A) Left: Sequences of various AP-tag variants. Right: AP-tag variants were expressed in HEK-293 cells along with HA-BirA or HA-BirA-RAD18. Biotin was added at the time of transfection, and cells were allowed to recover 6 hours after UV irradiation. Lysates were subjected to SDS-PAGE and immunoblotting. Results are representative of three independent biological replicates. (B) HEK-293 cells were transfected with (−2)AP-Ub and HA-BirA or HA-BirA-RAD18 followed by UV exposure (10 mJ/cm²). Cells were treated the same as in (A). Volcano plot of streptavidin enriched proteins in the two abovementioned conditions (n = 3 biological replicates). Significantly altered proteins are shown in dark red or blue (FDR <0.05, log₂FC >I0.6I) and light red or blue (FDR <0.05, 0 > log₂FC < I0.6I) (moderated t test). (C) GO term analysis of potential RAD18 ubiquitination substrates. Bar graph shows significantly enriched pathways. (D) HEK-293 are transfected with indicated plasmids. Cells were treated, and lysates were immunoblotted the same as in (A). Results are representative of two independent biological replicates. (E) HEK-293 cells were transfected with indicated small interfering RNAs (siRNAs). After 48 hours, cells were exposed to UV (10 mJ/cm²). The UV-treated and untreated cells were allowed to recover for 6 hours. Lysates were subjected to SDS-PAGE followed by immunoblotting with the indicated antibodies. Results are representative of three independent biological replicates. (F) Confocal microscopy of HEK-293 cells expressing either HA-BirA or HA-BirA-RAD18, along with (−2)AP-Ub, immunostained with anti–phospho histone H2AX (Ser¹³⁹) (green), antistreptavidin (red). Cells with notable Streptavidin signal are circled using white dotted lines. Scale bar, 12 μm. DAPI, 4′,6-diamidino-2-phenylindole. (G) Percentage of transfected cells showing colocalization (Pearson’s coefficient > 0.5) of streptavidin and phospho histone H2AX (Ser¹³⁰) signals (indicated by yellow arrows) is represented (****P < 0.0001) (n, biological replicates = 2, 10 different fields).

Fig. 3.. Using Ub-POD on the RING E3 ligase TRAF6.

(A) Biotin time course to estimate appropriate biotin treatment duration. HEK-293 cells were transfected with (−2)AP-Ub together with HA-BirA or HA-BirA-TRAF6 for 24 hours. Biotin (100 μM) was either added together with transfection mixture for 24 hours or added 6 hours, 3 hours, 1 hour, or 15 min before harvest, respectively. Notably, 15-min treatment was sufficient to induce strong biotinylation in HA-BirA-TRAF6 overexpressing cells and was therefore kept as treatment condition throughout all experiments. (B) Experimental setup for Ub-POD-TRAF6-MS. HEK-293 cells transiently overexpressing (−2)AP-Ub together with HA-BirA or HA-BirA-TRAF6 for 24 hours were subjected to biotin (100 μM) and PR-619 (10 μM) for 15 min and harvested. An aliquot was reserved for analysis via immunoblotting, the residual lysate was prepared for streptavidin pulldown and subjected to MS (n = 4 biological replicates). (C) Volcano plot depicting altered biotinylated proteins after enrichment via streptavidin pulldown from HA-BirA-TRAF6 or HA-BirA expressing HEK-293 cells. Known and selected potential TRAF6 substrates are labeled. Validated hits depicted in 3E are highlighted in red. Significantly altered proteins are shown in dark red or blue (FDR < 0.05, log₂FC > I1I) and light red or blue (FDR < 0.05, 0 > Ilog₂FCI < I1I) (moderated t test, n = 4 independent experiments); (D) GO terms of proteins found to be significantly enriched in HA-BirA-TRAF6 versus HA-BirA with FC ≥1. Dot size correlates to number of proteins, dot color to term enrichment (FDR value). (E) Validation of TRAF6 substrates identified by Ub-POD-TRAF6-MS. HEK-293 cells transiently overexpressing (−2)AP-Ub together with HA-BirA, HA-BirA-TRAF6, or dimerization mutant HA-BirA-TRAF6F118A for 24 hours were subjected to biotin (100 μM) and PR-619 (10 μM) for 15 min. Lysates were subjected to streptavidin pulldown followed by SDS-PAGE and immunoblotting.

Fig. 4.. Using Ub-POD to identify substrates of U-box ligase CHIP.

(A) HEK-293 cells transfected for 16 to 24 hours with (−2)AP-Ub and indicated BirA-tagged constructs. Cells were treated with MG132 (10 μM) for 6 hours. Cells were kept in biotin (100 μM) the whole time. Lysates were separated by SDS-PAGE and analyzed by Western blotting. Results are representative of two independent biological replicates. (B) HEK-293 cells transfected for 24 hours with (−2)AP-Ub and BirA-HA or CHIP-GSGS-BirA-HA were incubated with biotin (100 μM) and MG132 (10 μM) for 6 hours. Lysates were subjected to streptavidin pulldown and MS analysis. Volcano plot of proteins labeled by CHIP-GSGS-BirA-HA and HA-BirA in 6 hours (n = 3 biological replicates). Significantly altered proteins are shown in dark red or blue (FDR <0.05, log₂FC > I0.6I) and light red or blue (FDR <0.05, 0 > log₂FC < I0.6I) (moderated t test). (C) Bar graph representation of enriched GO terms for candidate CHIP substrates. (D) HA-BirA or CHIP-GSGS-BirA-HA transfected HEK-293 cells were cotransfected with either (−2)APUb or (−2)AP-UbΔGG. After 6 hours of MG132 treatment, whole-cell lysates were prepared followed by streptavidin pulldown. Pulldown and inputs were run on SDS-PAGE followed by immunoblotting with indicated antibodies. Results are representative of two independent biological replicates. (E) Knockdown of CHIP and immunoblotting for CHIP and ANXA5: HEK-293 cells were reverse transfected with either control siRNA (50 pmol) or different concentrations of CHIP siRNA (25 or 50 pmol). After 48 hours, cells were treated with either vehicle (dimethyl sulfoxide) or MG132 (10 μM). Whole-cell lysates, prepared after 6 hours of treatment, were run on SDS-PAGE followed by immunoblotting with the indicated antibodies. Results are representative of three independent biological replicates.

Fig. 5.. BioID of RAD18 and CHIP.

(A) BioID analysis of RAD18. HEK-293 cells transfected for 24 hours with BirA* or BirA*-RAD18 were exposed to UV (10 mJ/cm²) and allowed to recover for 6 hours. Cells were kept in 100 μM biotin the whole time. Lysates were subjected to streptavidin pulldown followed by MS (n = 3 biological replicates). Volcano plot of proteins labeled by BirA*-RAD18 and BirA*. Significantly altered proteins are shown in dark red or blue (FDR <0.05, log₂FC > I1I) and light red or blue (FDR <0.05, 0 > log₂FC < I1I) (moderated t test). (B) Bar graph depicting significantly enriched GO terms of hits from (A). (C) Overlap between hits identified in RAD18 BioID and Ub-POD experiments. (D) BioID analysis of CHIP. HEK-293 cells transfected for 24 hours with BirA* or CHIP-GSGS-BirA* were incubated with MG132 (10 μM) for 6 hours. Cells were kept in 100 μM biotin the whole time. Streptavidin pulldowns were performed with lysates followed by MS analysis (n = 3 biological replicates). Volcano plot of proteins labeled by CHIP-GSGS-BirA* and BirA*. Significantly altered proteins are shown in dark red or blue (FDR <0.05, log₂FC >I1I) and light red or blue (FDR <0.05, 0 > log₂FC < I1I) (moderated t test). (E) Bar graph representation of significantly enriched GO terms with hits identified in (D). (F) Overlap between hits identified in CHIP BioID and Ub-POD experiments.