Most enzymes involved in the addition
Most Vanoxerine mg involved in the addition and removal of ubiquitin bind weakly to an overlapping surface on ubiquitin. Previously, Sidhu and colleagues diversified this surface and then used phage display to screen these libraries of ubiquitin variants (UbVs) for binding to DUBs and E3 ligases (Ernst et al., 2013). Using this approach, selective and tight-binding UbVs that modulate the activity of DUBs and E3 ligases have been identified (Gorelik and Sidhu, 2017). However, UbVs that bind tightly and specifically to targets are not always forthcoming. In this issue, Teyra et al., (2019) take two different approaches to improve UbV development. First, they use the process of affinity maturation, where prior rounds of selection are used to guide further library development, to develop tight binders. In a second approach aimed at improving specificity, they link two UbVs that bind at distinct sites, with a surprisingly desirable outcome. As a first step to the development of USP15 inhibitors, UbVs that were specific for each domain, including the DUSP and catalytic domains, were identified (Figure 1). By focusing on the UbVs that bound the catalytic domain, they then set out to develop more effective USP15 inhibitors. To do this, two rounds of affinity maturation and selection of the library were completed. This resulted in UbVs that bound to USP15 with an EC50 of ∼1 nM. The UbVs were then screened in biochemical assays, which showed that after each step of library maturation the UbVs became more potent USP15 inhibitors. However, the UbVs concurrently lost specificity and inhibited other members of the USP family (Figure 1). This promiscuity was especially true for the closest homologs of USP15, USP4 and USP11, and underscores the difficulties in developing specific inhibitors for DUBs. In an alternative approach, the initial UbVs that bound each domain of USP15 were revisited with the goal of improving affinity by linking two UbVs together. Fusion of the UbV that was specific to the N-terminal DUSP domain to a UbV that targeted the catalytic domain resulted in a potent USP15 inhibitor (Figure 1). Notably, this dimeric UbV displayed inhibitory activity comparable to the UbVs from their third library, but it had minimal cross-reactivity with other members of the USP family. As a result, the dimeric UbV was the most effective USP15 inhibitor generated in the study. To understand more about how the UbVs interact with USP15, structures of some of the UbV-USP15 complexes were solved. In one crystal structure, the DUSP-binding UbV forms a strand-swapped dimer whereby the first β strand of each monomer is flipped 180° to interact with its partner. The exchange creates a pocket, which embraces the DUSP domains of two USP15 molecules. This is not the first time that swapping of the first β strand in UbVs has been observed, suggesting that this kind of rearrangement is more commonly selected for than expected (Gabrielsen et al., 2017). Not surprisingly, the UbV-DUSP interface relies on the diversified residues of ubiquitin. On USP15 the interaction includes residues that are not conserved in USP4 and USP11, which probably accounts for the preference of DUSP-binding UbV for USP15. Teyra et al., (2019) also obtained structures of the UbVs from each successive library that bound to the catalytic domain of USP15. Surprisingly, none of the UbVs bind in the expected ubiquitin-binding pocket. Instead, all UbVs bind in a manner that is likely to occlude entry of ubiquitin into the active site of USP15 (Ward et al., 2018). While the UbV-USP15 interface is similar in all three of the structures, the interaction of the UbVs from libraries two and three is enhanced by an extended loop included as part of these advanced libraries. This loop makes extra contacts with USP15, but the contact site includes residues that are conserved with USP4, explaining why these UbVs increased non-specific binding.