Background

The small ubiquitin-related modifier (SUMO) can be covalently transferred onto proteins, altering their subcellular localization, protein–protein interactions, and/or activity. The conjugation of SUMO moieties onto substrates occurs via the processive actions of a unique E1 activation enzyme (composed of a heterodimer containing SAE1/AOS1 and SAE2/UBA2) and an E2 conjugating enzyme (UBC9), with many E3 ligases assisting [2]. Within higher eukaryotes, there are three variants of conjugatable SUMO: SUMO1, SUMO2, and SUMO3. SUMO2 and SUMO3 share a 97% sequence identity and are virtually indistinguishable [3]; as such, they are often designated SUMO2/3. SUMO1 and SUMO2/3 share many substrates, but proteomic studies have revealed that SUMO2/3-modified substrates do not exclusively overlap with those modified by SUMO1 [1,4]. The divergence between SUMO1 and SUMO2/3 is particularly evident in mitosis, where SUMO1 localizes to the mitotic spindle, whereas SUMO2/3 co-occupies areas associated with centromeres and condensed chromosomes [5].

SUMOylation has emerged as a vital regulator of mitosis, with defects in mitotic progression and chromosome segregation observed upon deficiency, depletion, or chemical inhibition of core components within the SUMO conjugation machinery [6–10]. While SUMO conjugation is required for efficient and accurate mitotic progression, it is notable that global SUMO conjugates are lower during this phase of the cell cycle [5], which can make identifying alterations in SUMOylation during mitosis challenging. Therefore, obtaining a large and relatively “pure” mitotic population is essential to accurately pinpoint mitotic SUMOylation events that are not skewed by asynchronous contaminants. Due to the short duration of this phase of the cell cycle (approximately 1 h in RPE, HeLa and MCF10A) [11], cell synchronization is commonly used to obtain sufficient cell numbers to study. This has classically been achieved using microtubule poisons such as nocodazole [12,13] and mitotic shake-off [14].

The identification of SUMO conjugates requires cell lysates to be prepared under denaturing conditions. This encourages the stability of the SUMO modification by inhibiting the actions of SUMO-specific proteases/isopeptidases, as well as reducing the co-precipitation of non-covalent SUMO interactors [15]. The simplest method of identifying SUMOylated proteins is through the detection of a mobility shift on denaturing gels that is sensitive to SUMO-specific proteases and/or disruption to SUMO conjugation machinery. Whilst this may be a valid starting point for identifying a SUMO substrate, it is indirect, biased toward abundant or highly modified targets, and only speaks to the modification of one protein. To increase the abundance of SUMO-modified substrates and to enable the identification of SUMOylated sites on target proteins, affinity purification with overexpressed SUMO constructs has frequently been employed. The use of 6xHis-tagged exogenous proteins and Ni-NTA resin is particularly used due to its compatibility with denaturing conditions [4,15–20]. However, the use of overexpression is an inherent issue, as it intrinsically alters the balance of SUMO variants (i.e., SUMO1 and SUMO2/3), which has the potential to alter their usage [7,21]. A further caveat to using overexpression is the prerequisite for cells to have an acceptable degree of transfectability. This can limit the cell lines and types of samples that can be used to identify SUMOylated substrates.

A number of methods have been developed to enrich SUMO conjugates under endogenous conditions [1,22–28]. Many of these techniques rely on mass spectrometry–based proteomics to identify SUMO moieties or remnants on peptides [22–26]. These methods are integral to our understanding of SUMO biology, enabling the site-specific identification of SUMOylation on substrates. However, the main limitation in adopting these approaches for more frequent use, among the wider molecular biology community, is the high degree of specialism they require. The level of information that these techniques are able to acquire may also lie beyond the scope of the specific research question under investigation.

The enrichment of endogenous SUMOylated proteins (as opposed to peptides) has been described using SUMO-traps [alternatively named SUMO binding entities (SUBEs)] [28] and antibodies [1]. Both approaches require less specialism to perform, and protein identification can occur through targeted western blotting or mass spectrometry. SUMO-traps can capture endogenous poly-SUMOylated proteins using a GST-fusion protein bearing tandem SUMO interacting motif repeats from the SUMO E3 RNF4. However, these are not suited to the identification of mono-SUMOylated proteins and consequently bias results toward SUMO2/3-containing substrates. Additionally, as it is currently presented in the literature, this assay is performed under non-denaturing conditions, and therefore, non-covalent SUMO interactors and interacting proteins can also be co-precipitated with SUMOylated substrates [28]. Becker et al. [1] described a procedure that used commercially available antibodies and denaturing conditions to enrich SUMOylated substrates. The main benefits of this technique derive from its ease of use, accessibility of reagents, and relative affordability. Due to the initial denaturation of lysates, misidentification of non-covalent interactors should be avoided. However, the reliance of this assay on antibodies is a major limitation. Antibodies are inherently limited by the specificity of their antigen and may not capture all SUMO substrates or bind other SUMO paralogs to those intended. Batch-to-batch variation within polyclonal antibodies can also contribute to variation in results. Lastly, if a significant enough enrichment is not achieved through immunoprecipitation, this assay may struggle to detect modifications on lowly expressed or modified proteins.

Here, we describe an optimized procedure to assess SUMO conjugation in mitosis. This assay was successfully employed to characterize an alteration in SUMO variant usage during mitosis when an acetyl mimic of the SUMO E1 component SAE2 (SAE2-K164Q) was expressed in cells [7]. Building on the assay described by Becker et al. [1], cells are enriched for mitosis using nocodazole and further isolated via a mitotic shake-off. Lysis and denaturation then proceed, using a high-percentage SDS buffer and treatment with DTT. The denatured lysate is diluted tenfold with a 0% SDS lysis buffer before a standard immunoprecipitation with commercial antibodies. This method bears the same underlying disadvantages as the Becker et al. technique, described above, with some additional caveats: the requirement for a large number of cells and the use of agents to synchronize cells before harvesting. The benefits of using this assay over other techniques include its ease of use, high adaptability to investigate specific research questions, lack of requirement for specialist equipment, and the absence of SUMO variant imbalance generated by overexpression.