SUMO-activated target traps (SATTs) enable the identification of a comprehensive E3-specific SUMO proteome

Abstract

Ubiquitin and ubiquitin-like conjugation cascades consist of dedicated E1, E2, and E3 enzymes with E3s providing substrate specificity. Mass spectrometry-based approaches have enabled the identification of more than 6500 SUMO2/3 target proteins. The limited number of SUMO E3s provides the unique opportunity to systematically study E3 substrate wiring. We developed SUMO-activated target traps (SATTs) and systematically identified substrates for eight different SUMO E3s, PIAS1, PIAS2, PIAS3, PIAS4, NSMCE2, ZNF451, LAZSUL (ZNF451-3), and ZMIZ2. SATTs enabled us to identify 427 SUMO1 and 961 SUMO2/3 targets in an E3-specific manner. We found pronounced E3 substrate preference. Quantitative proteomics enabled us to measure substrate specificity of E3s, quantified using the SATT index. Furthermore, we developed the Polar SATTs web-based tool to browse the dataset in an interactive manner. Overall, we uncover E3-to-target wiring of 1388 SUMO substrates, highlighting unique and overlapping sets of substrates for eight different SUMO E3 ligases.

Fig. 1.. SUMO E3 overexpression affect endogenous SUMO2/3 levels.

(A) E3s studied in this article. The mutations performed on each E3 to construct the catalytic dead mutant controls are indicated. (B) In vitro SUMOylation assays including ZMIZ2 SUMO E3 enzyme and different concentrations of the SUMO E2. Assays were carried out using either SUMO1 or SUMO2. (C) Representative immunofluorescence image of U2OS cells transiently transfected with green fluorescent protein (GFP)–LAZSUL immunostained for SUMO2/3. (D) Superplot depicting relative SUMO2/3 nuclear intensities after immunostaining of individual U2OS cells transiently transfected with GFP-tagged constructs of different E3s. Values were normalized to the average SUMO2/3 nuclear intensity of GFP negatives from each individual experiment. Values from three independent experiments are depicted. (E) Stable-inducible GFP-LAZSUL–expressing U2OS cells were treated with control or RNF4-targeting small interfering RNAs (siRNAs). Thirty-six hours after siRNA transfection, GFP-LAZSUL expression was induced with doxycycline (20 μg/ml). Cells were fixed 48 hours after siRNA transfection and analyzed by immunostaining. (F) Quantification of the normalized nuclear SUMO2/3 intensities from the cells in (E). Independent values from two independent experiments are depicted. (G) Analysis by immunoblotting of the cells in (C). Size bars in fluorescence microscopy images represent 10 μm. DAPI, 4′,6-diamidino-2-phenylindole; IB, immunoblot; SIM, SUMO interaction motifs; SAP, SAF-A/B, Acinus and PIAS domain; S-P RING, Siz/Pias Really Interesting New Gene; TAD, Trans Activator Domain; UIM, Ubiquitin Interacting Motif.

Fig. 2.. SUMO-activated target traps (SATTs) enable the identification of E3-specific SUMOylatio substrates.

(A) SATTs screen rationale. SUMO moieties covalently attach to the C-term of an E3 of interest, which will be attached to E3 substrates, enabling the copurification of the E3 together with the SUMOylation target, which will be later identified by mass spectrometry (MS). (B) SATT negative controls rationale. While ∆GG SATTs lack the C-terminal SUMO diGly motif, unable to conjugate to the substrate, catalytic dead mutants prevent interaction with the SUMO E2. (C) Heatmap of z scores for different SATT targets. Only HIS-SUMO1 and HIS-SUMO2_Q87R targets are included. LC-MS/MS, liquid chromatography tandem MS.

Fig. 3.. SUMO1-SUMO2/3 overlap and Gene Ontology.

(A) Overlap between HIS-SUMO1 and HIS-SUMO2_Q87R targets. (B) Overlap between SUMO1–SUMO-activated target trap (SATT) and SUMO2-SATT substrates for the indicated E3s. (C) Gene Ontology (GO) analysis for the SUMOylation substrates of the different E3 SATTs analyzed in this study. ER, endoplasmic reticulum; ADP, adenosine 5′-diphosphate; ATP, adenosine 5′-triphosphate; LSU-rRNA, large subunit ribosomal–ribonucleic acid; dsDNA, double-stranded DNA.

Fig. 4.. SUMO-activated target trap (SATT)–specific SUMO chains.

(A) Total peptide intensities of the different peptides corresponding to different acceptor lysines in SUMO1, SUMO2, SUMO3, and SUMO4 depicted in fig. S4 in the different SUMO2_Q87R SATT samples. * indicates that this site corresponds to a peptide present in the three indicated SUMO types, thus not enabling distinction. WT, wild type.

Fig. 5.. Complement analysis.

Multiple volcano plots (Hawaii plot) depicting statistical differences of different unbiased SUMO1–SUMO-activated target trap (SATT) (A) or SUMO2-SATT (B) SATT targets using the values from the rest of wild-type SATTs as complement negative control. Cutoffs represent a Pearson of 100 and a false discovery rate = 0.05 (*) or 0.01 (**) and an S0 = 0.1.

Fig. 6.. Polar SUMO-activated target traps (SATTs) application.

SATT Polar plots extracted from the polarVolcaNoseR web application. Some of the most prominent or specific substrates from different E3s are indicated. Depicted plots use ∆GG as control, which include the SATT index, and only consider HIS-SUMO targets.

Fig. 7.. Orthogonal validation.

(A) Experimental setup. Parental and NSMCE2–knockout (KO) U2OS cells rescued or not with either wild-type (WT) NSMCE2 or a catalytic dead mutant and expressing or not HIS-SUMO2_Q87R were cultured, incubated for 5 hours with MG132, lysed, and analyzed by liquid chromatography tandem mass spectrometry (LC-MS/MS). Next, data were processed together with the NSMCE2-SUMO2–SUMO-activated target trap (SATT)data by MaxQuant and Perseus. (B) Venn diagram depicting overlap among NSMCE2-SUMO2-SATT substrates, HIS-SUMO2_Q87R substrates affected by lack of NSMCE2 catalytic activity, and NSMCE2-SUMO2-SATT substrates shared with other E3-SATTs. (C to E) Immunoblot analysis of HIS-SUMO2 substrates decreasing upon NSMCE2 catalytic activity (C), LAZSUL-KO (D), or PIAS4 knockdown (E).