The interface between the E ligase
The interface between the E3 ligase and the E2 enzyme can vary, and ZNF451 and SP-RING ligases stabilize this interaction via noncovalent binding to a scaffold SUMO (SUMOB) on the backside of the E2 (Cappadocia et al., 2015; Eisenhardt et al., 2015; Streich & Lima, 2016). By contrast, RanBP2 does not involve such a scaffold SUMO and interacts directly with the backside of the E2 (Pichler, Knipscheer, Saitoh, Sixma, & Melchior, 2004; Reverter & Lima, 2005). Hence, E2 backside interaction is also important for efficient substrate modification of most but not all E3 ligases and thus can also be addressed by using the in vitro sumoylation assays (see below).
Substrate sumoylation can be visualized by SDS polyacrylamide gel electrophoresis (PAGE) due to the stability of the isopeptide bond in SDS and DTT. SUMO attachment results in a ~15–20kDa size shift in substrate mobility as measured by SDS-PAGE (the predicted ~10kDa SUMO migrates more slowly in gels because of its flexible N-terminus). Thus, SUMO–protein conjugates can be distinguished from noncovalent SUMO interactions such as those that occur through a SUMO interaction motif (SIM) or with the backside of the E2 (Pichler et al., 2017). Also thioester bonds, such as the linkage in SUMOD~E2, can be visualized by SDS-PAGE, but in Pranlukast to isopeptide bonds, this linkage is sensitive to reducing agents and should not exceed a final concentration of 0.1mM DTT.
Substoichiometric substrate modification E3 ligases catalyze the rapid transfer of the SUMOD from the E2 enzyme to the substrate. By definition, enzymes are recycled in the reaction, allowing multiple rounds of substrate modification by a single enzyme. Thus, enzymes function at substoichiometric amounts relative to the substrate (low enzyme to substrate ratio) in a concentration- and time-dependent manner. To monitor sumoylation in vitro we use purified recombinant enzymes following the protocols described in detail by the Melchior lab (Werner, Moutty, Moller, & Melchior, 2009). Ideally, all assay components are purified from Escherichia coli as bacteria lack a SUMO system, thus avoiding copurification of SUMO E3 ligases or proteases which at undetectable concentrations could affect the outcome of the in vitro reaction. Any known and putative SUMO substrate can be tested in this system if can be purified. All known SUMO E3 ligases show some substrate specificity in vitro, but they usually have broad substrate spectra; this is especially when working with a truncated E3 ligase (Cox et al., 2015; Koidl et al., 2016; Pichler et al., 2004). Here, we use GST-Sp100 as model substrate as it is a promiscuous substrate that can be sumoylated by all known classes of SUMO E3 ligases as well as by the sumoylated E2 (S*E2) without an E3 (Knipscheer et al., 2008; Koidl et al., 2016; Pichler et al., 2002, Pichler et al., 2004; Sternsdorf, Jensen, & Will, 1997). For in vitro sumoylation reactions, all components can simply be added together in a test tube with ATP and Mg and incubated for up to 30min in a multiturnover reaction. Detection of sumoylation is performed by immunoblotting using substrate-specific antibodies. We use anti-GST antibodies that detect the N-terminally fused GST-tag of Sp100, and thus we can rule out that sumoylation interferes with detection. Because the E2 enzyme directly recognizes a SCM, it can modify substrates in vitro at high enzyme concentrations as we illustrate in Fig. 1A. Hence, it is important to use a low E2 concentration in E3-dependent reactions that either show no or only marginal substrate modification by the E2 alone. At high enzyme concentrations, the mammalian E2 itself gets sumoylated at Lys 14 (S*E2). We have shown that this particular modification stabilizes the E2 interaction with GST-Sp100 due to a SIM in the substrate that is in close proximity to the SCM. The E2*S adduct therefore also enhances GST-Sp100 modification to a degree comparable to some E3 ligases (Knipscheer et al., 2008) (Fig. 1A). However, E3 ligases are usually more potent and promote higher sumoylation rates when at low (substoichiometric) levels; this can be seen both by varying E3 concentrations (Fig. 1B) and by following sumoylation over time (Fig. 1C).