Structural basis of the specificity of USP18 toward ISG15

Abstract

Protein modification by ubiquitin and ubiquitin-like modifiers (Ubls) is counteracted by ubiquitin proteases and Ubl proteases, collectively termed DUBs. In contrast to other proteases of the ubiquitin-specific protease (USP) family, USP18 shows no reactivity toward ubiquitin but specifically deconjugates the interferon-induced Ubl ISG15. To identify the molecular determinants of this specificity, we solved the crystal structures of mouse USP18 alone and in complex with mouse ISG15. USP18 was crystallized in an open and a closed conformation, thus revealing high flexibility of the enzyme. Structural data, biochemical and mutational analysis showed that only the C-terminal ubiquitin-like domain of ISG15 is recognized and essential for USP18 activity. A critical hydrophobic patch in USP18 interacts with a hydrophobic region unique to ISG15, thus providing evidence that USP18's ISG15 specificity is mediated by a small interaction interface. Our results may provide a structural basis for the development of new drugs modulating ISG15 linkage.

Figure 1. Structure of USP18 in the unbound and ISG15-bound state.

(a) USP18 crystallized with two molecules in the asymmetric unit. Superposition of the two molecules in the asymmetric unit (dark blue, chain A; green, chain B). For comparison, ISG15-bound USP18 (light blue) is structurally aligned. The Zn²⁺ ions in the finger domain are shown as spheres and illustrate the extent of the movement. Residues 255–267 corresponding to blocking loop 1 are disordered in unbound USP18 and are not resolved in the electron density, whereas they form a short antiparallel β-sheet in ISG15-bound USP18. (b–d) Close-up views of the catalytic triad in ISG15-bound (b) and unbound (c,d) USP18. In all structures of USP18, the residues Cys61, His314 and Asn331, which form the catalytic triad, are in proximity. (b) Only in ISG15-bound USP18 do the side chains of the three residues exhibit the correct orientation and distances to stabilize a thiolate state of the Sγ of Cys61. (c,d) In both molecules of unbound USP18, the imidazole of His314 is slightly shifted away from Cys61, thereby weakening the interaction of the two side chains.

Figure 2. The C-terminal Ubl domain of ISG15 is sufficient for USP18 binding and activity.

(a) Structural alignment of both USP18–ISG15 complexes present in the asymmetric unit. For clarity, only one USP18 molecule (blue) is shown. Left, surface representation of USP18 (blue) with ISG15 aligned from both complexes (red and orange). Right, the molecules are rotated by 180°. USP18 is shown as a cartoon with an outline of the surface. The C-terminal Ubl domain of ISG15 is embedded between the finger and the thumb domains. In contrast, the N-terminal Ubl domain of ISG15 exhibits no contacts to USP18. In both complexes, the N-terminal domain adopts a different position with regard to the C-terminal domain and is located 8 Å or 18 Å from the USP18 thumb domain (distance between the Cα atoms of ISG15 Ser22 and USP18 Glu91). (b) Coomassie-stained SDS–PAGE showing that USP18 reacts with full-length ISG15 and the C-terminal domain of ISG15 (ct-ISG15) with the same efficiency. The reaction products of stoichiometric amounts of USP18 with full-length ISG15-PA, ct-ISG15-PA or ubiquitin-PA (Ub-PA) were visualized as a shift in molecular mass (MM). (c) Immunoblot showing cleavage of ct-ISG15 from cellular substrates by recombinant USP18. Cell lysates with cellular substrates modified by hemagglutinin (HA)-tagged ct-ISG15 or HA-tagged-full-length ISG15 were incubated with recombinant USP18. Top, deconjugation of ISG15 from cellular proteins, as monitored with an anti-HA antibody. Middle, β-actin was used as a loading control. Bottom, Ponceau S staining shows the presence of USP18.

Figure 3. Conformational changes in USP18 after ISG15 binding.

Structural alignment of USP18 (light blue) bound to ISG15 (orange) and unbound USP18 (dark blue), revealing a conformational change in the finger domain and in the blocking and switching loops. (a) Residues 255–267 (blocking loop 1), which are unordered in unbound USP18, fold into a short β-sheet that interacts with ISG15. (b) The switching loop connects helix 4 and helix 5 of the thumb domain and comprises residues 129–135. In unbound USP18, residues 131–135 fold into a short α-helix that blocks the access of the C-terminal LRLRGG motif of ISG15 to the catalytic site of USP18. In ISG15-bound USP18, the α-helix unwinds, and the entire loop comprising these residues is displaced.

Figure 4. ISG15-binding boxes 1 and 2 in USP18.

(a) Structure of the USP18–ISG15 complex. Blue, USP18; orange, ISG15; green, IBB-1 of USP18; purple, IBB-2 of USP18. (b) Structural alignment of the USP18–ISG15 complex with the USP7–Ub complex (PDB 1NBF). Blue, USP18; orange, ISG15; gray, USP7; yellow, Ub. The orientation of Ub bound to USP7 closely resembles the orientation of the ISG15 C-terminal domain in the USP18–ISG15 complex. (c) Close-up view of IBB-1. The residues of USP18 forming IBB-1 are shown in green and form a hydrophobic pocket that accommodates the bulky aromatic side chain of Trp121 from ISG15 (orange). (d) Superposition of IBB-1 with the respective region of the USP7–Ub complex. The labeling refers to residues in USP18 and ISG15. The interaction between USP18 and ISG15 is mediated by hydrophobic residues, whereas USP7 and ubiquitin display polar residues in this region. (e) Close-up view of IBB-2 (purple), showing several residues of a short antiparallel β-sheet of USP18 forming hydrogen bonds (dotted lines) with ISG15. For clarity, the side chains of the interacting residues are omitted. (f) Superposition of IBB-2 with the respective region of the USP7–Ub complex. The distances between ubiquitin and USP7 are larger than those in the USP18–ISG15 complex.

Figure 5. IBB-1 but not IBB-2 is critical for USP18 activity.

(a) Coomassie blue–stained SDS–PAGE showing the reactivity of USP18 WT and variants toward ISG15-PA and ct-ISG15-PA probes after incubation for the indicated times. The figure shown is representative of three independent experiments. (b) Catalytic activity of USP18 WT and variants toward the ISG15-FP substrate, as shown by substrate cleavage, monitored on the basis of the change in fluorescence polarization (in millipolarization units (mP)). Different amounts of USP18 proteins were incubated with ISG15-FP. (c) Immunoblot analyses showing deISGylation of endogenous substrates by USP18 WT or variants. ISG15 cleavage by USP18 was monitored with an antibody directed against ISG15.

Figure 6. Zebrafish USP18 recognizes ISG15 and ubiquitin.

(a) Immunoblot of lysates from HEK 293T cells transfected with S-tagged versions of mouse USP18 (mUSP18), mouse USP18 with the active site cysteine replaced by alanine (mUSP18 C61A), human USP21 (hUSP21), zebrafish USP18 (drUSP18) and zebrafish USP18 with the active site cysteine replaced by an alanine (drUSP18 C38A) or from untransfected cells (control). Protein expression was visualized with an antibody directed against the S tag. (b) Immunoblot of lysates from cells transfected with the indicated expression constructs, incubated with the active site–directed probe ISG15-PA. Complex formation was monitored on the basis of a size shift detected with an anti-S-tag antibody. (c) Immunoblot analyses of protein lysates from cells transfected with the indicated expression constructs, incubated with the active site–directed probe Ub-PA. Complex formation was monitored on the basis of a size shift detected with an anti-S-tag antibody. USP21 is cross-reactive for ISG15, and ubiquitin served as a positive control for Ub binding. Results shown in a–c are representative of three independent experiments.