The DUSP-Ubl domain of USP4 enhances its catalytic efficiency by promoting ubiquitin exchange

Abstract

Ubiquitin-specific protease USP4 is emerging as an important regulator of cellular pathways, including the TGF-β response, NF-κB signalling and splicing, with possible roles in cancer. Here we show that USP4 has its catalytic triad arranged in a productive conformation. Nevertheless, it requires its N-terminal DUSP-Ubl domain to achieve full catalytic turnover. Pre-steady-state kinetics measurements reveal that USP4 catalytic domain activity is strongly inhibited by slow dissociation of ubiquitin after substrate hydrolysis. The DUSP-Ubl domain is able to enhance ubiquitin dissociation, hence promoting efficient turnover. In a mechanism that requires all USP4 domains, binding of the DUSP-Ubl domain promotes a change of a switching loop near the active site. This 'allosteric regulation of product discharge' provides a novel way of regulating deubiquitinating enzymes that may have relevance for other enzyme classes.

Figure 1. The USP domain of USP4 is formed by the D1 and D2 domains.

(a) Domain representation of human USP4 and constructs used in this study. The position of the catalytic residues (C, H, D) is indicated in red. SL=switching loop (residues 385–392); BL1 and BL2=blocking loops 1 and 2 (residues 831–834 and 874–880, respectively) (b) Cartoon representation of USP4 D1D2 domain crystal structure coloured as in a. The catalytic triad and zinc-coordinating residues are represented as sticks. Blocking loops 1 and 2 and switching loop are indicated as BL1, BL2 and SL, respectively. Dotted lines indicate the position of the USP4 insert. (c,d) Zoom on the superposition of the six non-crystallographic symmetry-related USP4 copies in the asymmetric unit showing the flexibility in the zinc-finger ribbon (c) and in the BL1, BL2 and SL loops (d).

Figure 2. The activity of USP4 is significantly reduced in the absence of the DUSP–Ubl domain.

(a) Product accumulation in time for USP4 FL and CD reactions on ubiquitin–rhodamine (Ub–Rhod). The panel on the right zooms on USP4 CD curves. (b) Michaelis–Menten plots showing the catalytic rate of USP4 FL and CD as a function of substrate concentration (error bars represent s.e.m. for at least three independent replicates). The panel on the right zooms on the USP4 CD curve. (c) Product accumulation for USP4 CD reaction at three different enzyme concentrations and 150 nM Ub–Rhod showing that the amplitude of the ‘burst’ phase is equal to the enzyme concentration. (d) Michaelis–Menten plot for USP4 FL and CD reaction on the model substrate-ubiquitinated PCNA (error bars represent s.e.m. for at least two independent replicates).

Figure 3. The DUSP–Ubl domain promotes ubiquitin dissociation and association.

(a,b) Pre-steady-state kinetics of ubiquitin dissociation (a) and association (b) for USP4 CD (upper panel) and FL (lower panel) measured by fluorescence polarization in a stopped-flow device. (c) Equilibrium binding of ubiquitin to USP4 CD (upper panel) and FL (lower panel) measured by fluorescence polarization (error bars represent s.e.m. for at least three independent replicates).

Figure 4. The DUSP domain, the linker and the insert are required for USP4 activity.

(a) Michaelis–Menten plots showing the catalytic rate of USP4 FL, CD, ΔDUSP, Δlinker and Δinsert as a function of substrate concentration (error bars represent s.e.m. for at least three independent replicates). The domain representation of USP4 indicating the DUSP domain, the insert and the linker is given as reference. (b) Surface plasmon resonance (SPR) traces of DUSP–Ubl at different concentrations binding to USP4 CD, insert and D1D2. The overlay of the normalized responses is given in the lower right panel.

Figure 5. Single-residue mutations in the DUSP domain and in the switching loop hamper USP4 activation.

(a) Mutations mapped onto structure of the murine DUSP–Ubl domain (PDB 3JYU). Mutated residues that cause a reduction in USP4 FL activity are depicted as yellow sticks and labelled in red. Hydrogen bonds are indicated as black dotted lines. (b) Normalized activities of USP4 FL DUSP mutants measured at 5 μM Ub–Rhod concentration (error bars represent s.e.m. for at least three independent replicates). (c) Michaelis–Menten plots showing the catalytic rate of F386G and PQ385AV mutants in comparison with USP4 CD (left panel) and USP4 FL (right panel) wild type (error bars represent s.e.m. for at least three independent replicates). (d) Comparison of the normalized pre-steady-state kinetics of USP4 CD and USP4 CD F386G ubiquitin dissociation. (e) Structure of the USP4 switching loop (SL) represented as yellow sticks (the rest of USP4 is coloured as in Fig. 1b). Residues mutated in c are labelled in red. Hydrogen bonds are indicated as black dotted lines. Blocking loop 2 (BL2) is shown with stick representation. Superposed ubiquitin is represented as a grey cartoon. (f) Schematic model representing the USP4 enzymatic turnover. The ubiquitinated target (T) is recognized by USP4 (1); ubiquitin hydrolysis is catalysed by USP4 (2); the deubiquitinated target dissociates from USP4 (3); ubiquitin dissociates from USP4 allowing the access of a new substrate (4). The dissociation of the DUSP–Ubl domain from the catalytic domain allows the switching loop (SL) to retain ubiquitin in the USP4 active site, preventing the access of more substrates (3a).