A Chemical and Enzymatic Approach to Study Site-Specific Sumoylation

Abstract

A variety of cellular pathways are regulated by protein modifications with ubiquitin-family proteins. SUMO, the Small Ubiquitin-like MOdifier, is covalently attached to lysine on target proteins via a cascade reaction catalyzed by E1, E2, and E3 enzymes. A major barrier to understanding the diverse regulatory roles of SUMO has been a lack of suitable methods to identify protein sumoylation sites. Here we developed a mass-spectrometry (MS) based approach combining chemical and enzymatic modifications to identify sumoylation sites. We applied this method to analyze the auto-sumoylation of the E1 enzyme in vitro and compared it to the GG-remnant method using Smt3-I96R as a substrate. We further examined the effect of smt3-I96R mutation in vivo and performed a proteome-wide analysis of protein sumoylation sites in Saccharomyces cerevisiae. To validate these findings, we confirmed several sumoylation sites of Aos1 and Uba2 in vivo. Together, these results demonstrate that our chemical and enzymatic method for identifying protein sumoylation sites provides a useful tool and that a combination of methods allows a detailed analysis of protein sumoylation sites.

Fig 1. A new chemical and enzymatic approach to analyze protein sumoylation site.

(A) Schematic of the conventional method using intact SUMO remnant as a signature for MS analysis. (B) Schematic of our new method (see text for details). (C) In vitro sumoylation reaction was performed using purified E1, E2 and Mms21 enzymes to generate poly-SUMO chains. After acetylation, treatment by Ulp1 disassembled poly-SUMO chains. (D) Method to evaluate acetylation efficiency of known SUMO targets purified from yeast cells. Ulp1 was used prior to acetylation reaction. (E) Comparison of acetylated versus unmodified lysine identified for known SUMO targets after treated by increasing concentration of acetic anhydride. The numbers of peptides containing unmodified lysine and acetylated lysine (in parenthesis) are indicated.

Fig 2. Multiple methods are used to identify sumoylation of E1 and E2 enzymes.

(A) In vitro sumoylation using E1, E2 and acetylated SUMO, which cannot form poly-SUMO chains, reveals higher molecular weight sumoylated species of Uba2, Aos1 and Ubc9. Ulp1 removal of SUMO allowed the identification of unmodified lysine as sumoylation sites. (B) Peptides that led to the identification of sumoylation sites of E1 and E2 enzymes are listed. Positions of the unmodified lysine (sumoylation site) are shown in parenthesis while all labeled lysine are acetylated. The periods flanking each peptide indicate the site of trypsin cleavage. (C) Comparison between WT Smt3 and Smt3-I96R as substrate for in vitro sumoylation by Aos1-Uba2 and Ubc9-K153R. Both WT Smt3 and Smt3-I96R are acetylated to prevent poly-SUMO formation. (D) Sumoylated peptides identified for Aos1 and Uba2 using EQIGG and GG remnants with the position of lysine indicated in parenthesis. * The first methionine was removed from Aos1, which is fused with N-terminal 6xHIS tag for bacterial expression.

Fig 3. Characterization of the effect of smt3-I96R mutation on cell growth, accumulation of GCRs and intracellular sumoylation.

(A) Plate assay showed that smt3-I96R mutation does not cause any appreciable growth defect in the presence of genotoxic agents. The mec1Δ mutant is used as a positive control for its hypersensitivity to genotoxic agents including hydroxyurea (HU), camptothecin (CPT) and UV radiation. (B) Rates of accumulating GCRs in smt3-I96R mutant (strain HZY4247) and smt3GG mutant (strain HZY3392), both containing a C-terminal stop codon to expose the Gly-Gly repeat and thus mature SUMO in cells. (C) Abundance ratios of the purified sumoylated proteins in WT (HZY2101) and smt3-I96R (HZY4248) strains, measured by quantitative MS using the method described previously [16].

Fig 4. Proteome-wide analysis of protein sumoylation sites using GG-remnant method and the chemical labeling and Ulp1-cleavage method.

(A) Sumoylated proteins were purified using Ni-NTA and anti-FLAG affinity resins, eluted by denaturing method and then divided for various treatments as indicated. (B) Venn diagram shows the number of sumoylation sites identified by the GG-remnant method, acetylation without Ulp1 cleavage, and acetylation with Ulp1-cleavage method. (C) Sumoylation sites identified with the position of lysine for each protein shown in parenthesis. Previously known sumoylation sites are underlined.

Fig 5. Effect of K4R and K7R mutation on Aos1 sumoylation and function.

(A) Effect of K4R and K7R mutations on sumoylation of Aos1 in vitro. (B) Effect of Effect of K4R and K7R mutations on Aos1 sumoylation in vivo. Epitope-tagged Aos1-HA in the HF-SUMO strain was purified using anti-HA affinity resins and its sumoylated forms was detected using anti-FLAG Western blot. (C) Growth and sensitivity of various sumoylation-defective aos1 mutants to genotoxic agents. 10-fold dilution of the same amounts of cells were plated and incubated at 30°C for 3 days. The mec1Δ mutant is used as a positive control for its hypersensitivity to genotoxic agents including hydroxyurea (HU), camptothecin (CPT) and UV.

Fig 6. Characterization of Uba2 sumoylation.

(A) Effect of uba2-K229R mutations on the ability of Aos1-Uba2 in catalyzing sumoylation reaction. (B) Effect of uba2-K229R and uba2-3KR on Uba2 sumoylation in vivo. Whole cell lysates from the indicated yeast strains were subjected to Western blot analysis. (C) Effect of uba2-K229R and uba2-3KR mutations on cell growth. The mec1Δ mutant is used as a positive control for sensitivity to hydroxyurea (HU), camptothecin (CPT) and UV.