br E E backside interaction The
E3–E2 backside interaction The E2 possesses an important regulatory interface which is termed its backside as it is opposite to the catalytic cleft that bears the active-site cysteine forming the thioester with SUMOD. This backside site interacts noncovalently with a scaffold SUMOB and was originally shown to be important for E2-mediated SUMO chain formation in vitro (Capili & Lima, 2007; Knipscheer, van Dijk, Olsen, Mann, & Sixma, 2007). Moreover, it partially overlaps with the E2–E1 interface (Duda et al., 2007) and is required for direct or indirect E2–E3 interactions. SUMOB is essential for the E3 activity of ZNF451 family members (Cappadocia et al., 2015; Eisenhardt et al., 2015) and strongly enhances the activity of Siz/Pias family members (Mascle et al., 2013; Streich & Lima, 2016), whereas RanBP2 directly interacts with the backside of the E2 independent of a SUMO (Pichler et al., 2004; Reverter & Lima, 2005). The requirement of the E2-SUMOB interaction for E3 activity can be tested in vitro by employing a SUMO or E2 mutant which are specifically impaired in this interaction: SUMO2 D62R abolishes, and E2 F22A weakens this particular interaction (Capili & Lima, 2007; Knipscheer et al., 2007). Here we show RanBP2 as an example of an E3 that displays no major defects with either mutant (Fig. 3A, upper panel) as SUMOB is dispensable for its catalytic activity. However, E2 F22A has mild effects on RanBP2’s activity as E2-Phe22 contributes to the larger RanBP2–E2 interface (Pichler et al., 2004; Reverter & Lima, 2005). By contrast, PIAS1 clearly requires the SUMOB-E2 interaction for efficient Sp100 sumoylation (Fig. 3A, lower panel) as does ZNF451 (Eisenhardt et al., 2015; Koidl et al., 2016). At low PIAS1 concentrations, SUMO2 D62R results in minimal GST-Sp100 modification. The E2 F22A BG45 structure only partially interrupts the E2-SUMOB interaction and hence causes a more mild reduction in PIAS1-dependent substrate sumoylation. Of note, Sp100 modification in the absence of an E3 (E2 or S*E2 dependent) remains unaffected (Fig. 3B). In vitro sumoylation assays used here are similar to substrate modification described in Section 2 and also set up in a multiturnover reaction.
Donor SUMO positioning Initially, E3 ligases were thought to interact simultaneously with the charged E2 enzyme and the substrate to bring them in close distance for an efficient SUMOD transfer. As this definition may better describe a cofactor than an enzymatic activity, the question was raised as to whether E3 ligases also involve a catalytic component. The first evidence in this direction came from the crystallographic analysis of RanBP2 interacting with a donor SUMOD-charged E2 mimic; this revealed that the E3 ligase also binds the modifier SUMOD (Reverter & Lima, 2005). Subsequent biochemical analysis demonstrated that this SIM-dependent interaction is indeed essential for RanBP2’s catalytic activity (Reverter & Lima, 2005). Meanwhile, all three classes of SUMO E3 ligases were shown to depend on this feature that is also called the “closed conformation” as it positions the SUMOD modifier optimally for the nucleophilic attack of the incoming substrate lysine (Cappadocia et al., 2015; Eisenhardt et al., 2015; Reverter & Lima, 2005; Streich & Lima, 2016; Yunus & Lima, 2009). Based on the SUMO–SIM interaction involved in SUMOD positioning, a SUMO2ΔSBD (SIM-binding domain; Q30A, F31A, I33A) mutant can be investigated that disrupts this important binding interface (Eisenhardt et al., 2015; Meulmeester, Kunze, Hsiao, Urlaub, & Melchior, 2008). In Fig. 4A, multiturnover assays are presented for RanBP2 and PIAS1 that both show clearly that substrate sumoylation depends on interactions of SUMO with the SIM of the E3. However, such assays are only conclusive for E3 ligases that do not require the SUMOB-E2 backside interaction, which is true for RanBP2 (Fig. 3). E3 ligases such as ZNF451 or Siz/PIAS family members have additional SBD(SUMO)–SIM(E3) interactions involving the scaffold SUMOB (see Fig. 3, Fig. 4); analysis of these E3s demands single-turnover assays to clearly distinguish between SIM-dependent SUMOD positioning and scaffold SUMOB binding. For such assays, the E2 is charged with SUMO2 wt or the SUMO2ΔSBD mutant, and discharge reactions are performed in the presence of SUMO2 wt added along with the substrate and the E3 (Fig. 4B) to allow E3 scaffold SUMOB binding. E3-independent sumoylation of Sp100 by the E2 or S*E2 is independent of both SUMOD positioning (Fig. 4C) and scaffold SUMOB binding (Fig. 3B).