Main Text Post translational modifications by ubiquitin Ub
Main Text Post-translational modifications by ubiquitin (Ub) and ubiquitin-like proteins (Ubl) regulate a diverse array of cellular and physiological processes including protein degradation, localization, and activation. While ubiquitination has primary responsibility for targeting substrates for proteasomal degradation, Ubls like SUMO, Nedd8, and ISG15 often regulate substrate activity by serving as docking sites for protein-protein interactions. Ub/Ubls are covalently linked to their target proteins through a three-enzyme cascade. An E1 activating enzyme forms a thioester conjugate with Ub/Ubl through ATP hydrolysis, which is then transferred to an E2 conjugating enzyme, and finally an E3 ligase transfers Ub/Ubl to its target protein. Different Ub/Ubl proteins have their own cognate E1, E2, and E3 enzyme machinery for protein conjugation (Kerscher et al., 2006). Over the last two decades, Ub- and Ubl-dependent pathways have emerged as attractive targets for inhibitor development. Small molecules targeting E1-activating enzymes and E3 ligases have been developed for specific systems of Ubls. These compounds have shown promise in regulating “undruggable” targets such as transcription factors in therapeutic contexts (Wertz and Wang, 2019). Notably, inhibitors of the E1s for Ub (TAK-243) (Hyer et al., 2018), Nedd8 (MLN4924) (Brownell et al., 2010), and SUMO (ML-792) (He et al., 2017) have been developed and shown promise in the clinic (Shah et al., 2016). These inhibitors are mechanistically similar and form covalent adducts with Ub/Ubl, catalyzed by their respective E1s. This adduct mimics the Ub/Ubl∼adenylate intermediate of the E1 activation cycle, which then binds to E1 and inhibits further Ubl-dependent activation. In this issue of Cell Chemical Biology, Chen and colleagues (Li et al., 2019) expand the scope of E1 inhibition by developing an allosteric inhibitor for small ubiquitin-like modifiers (SUMO) activating enzyme E1. This novel amphetamine sulfate reviews covalently binds in a buried site of the enzyme away from the active site, in contrast to previously characterized E1 inhibitors (Li et al., 2019; Figure 1). The authors identified inhibitors of SUMOylation using two assays to monitor the conjugation of SUMO1 to RanGAP1 by E1 and E2 enzymes: a fluorescence resonance energy transfer (FRET) primary screen that evaluated over 250,000 compounds and a chemiluminescence secondary screen. A counter screen was used to identify hits that selectively inhibited SUMOylation over ubiquitination. The sequence of triaging assays resulted in identification of one compound with high specificity to SUMOylation. A medicinal chemistry optimization campaign yielded an optimized inhibitor of SUMOylation (COH000) with an in vitro IC50 of 0.2 M. The inhibition of SUMO E1 by COH000 was long lasting and resistant to dialysis, and the authors used mass spectrometry and kinetic analysis to confirm that COH000 covalently labeled the enzyme. Unexpectedly, COH000 labeled a buried cysteine (Cys30) in SUMO E1 and not the catalytic cysteine (Cys173). Consistent with a non-competitive allosteric binding modality, the inhibition of SUMO E1 activity was not impacted by the concentration of ATP or SUMO. Cys30 is deeply buried in previously published crystal structures for SUMO E1 (Lois and Lima, 2005, Olsen et al., 2010). However, the SUMO E1 enzyme bound by COH000 has a different thermal denaturation profile and a higher melting temperature compared to the enzyme alone, suggesting the conformation of SUMO E1 is substantially different when bound by the inhibitor (Li et al., 2019). This was confirmed in a parallel publication of the crystal structure for SUMO E1 bound by COH000 (Lv et al., 2018). The structure revealed a dramatic rotation of the “second” catalytic cysteine half (SCCH)-domain, which contains the catalytic cysteine Cys173. This change reveals a cryptic pocket containing Cys30 that formed a covalent bond via a Michael addition reaction with COH000 and locked the enzyme in an inactive conformation (Lv et al., 2018).