Strategies to Target ISG15 and USP18 Toward Therapeutic Applications
ISGylation
ISG15 is one of the genes most strongly induced by type I interferon and was the first Ubiquitin-like modifier (Ubl) identified (Blomstrom et al., 1986; Haas et al., 1987). Analogous to ubiquitin, Ubls like ISG15, small ubiquitin-related modifier (SUMO), human leukocyte antigen (HLA)-F adjacent transcript 10 (FAT10) or neural precursor cell expressed, developmentally down-regulated 8 (NEDD8) can be covalently linked to target proteins to alter a variety of biological processes.
ISG15 is composed of two Ubl domains connected by a flexible polypeptide hinge region. Each domain is formed by four β-sheets and a single α-helix (Narasimhan et al., 2005) reminiscent of the ubiquitin structure. The C-terminal tail of ISG15 contains the LRLRGG motif which is essential for the conjugation to target proteins. Like ubiquitin, ISG15 can be covalently attached to lysine residues of target proteins (through the ε-amino group) via the LRLRGG motif (Loeb and Haas, 1992).
Analogous to the ubiquitin conjugation system, ISGylation is mediated by the consecutive action of a three-step catalytic cascade, where all the enzymes are induced by type I IFNs (Figure 1). E1-activating enzymes bind to Ub (or ISG15) and, mediated by ATP-Mg, form a complex that catalyzes Ub (or ISG15) C-terminal acyl adenylation (Tokgoz et al., 2006). Subsequently, a catalytic cysteine on the E1 enzyme interacts with the ubiquitin-AMP or ISG15-AMP complex undergoing acyl substitution that leads to thioester bond formation and the release of an AMP group. After that, through a transthiolation reaction, an E2 cysteine residue replaces the E1 enzyme. E2-conjugating enzymes catalyze the isopeptide bond formation but also contribute to substrate specificity. E3-ligase enzymes bind the E2-ubiquitin thioester, recognize the protein substrate and catalyze the transfer of ubiquitin or ISG15 from the E2 enzyme to the target protein (Zhang and Zhang, 2011).
Type-I interferon signaling and ISG15. Type I interferon (IFN) binds its receptor causing the dimerization of the two subunits IFNAR1 and IFNAR2 and thus the activation of the JAK-STAT pathway. The receptor associated kinases TYK2 and JAK1 induce recruitment and phosphorylation of STAT1 and STAT2. The phosphorylated proteins translocate to the nucleus and together with IRF9 form a trimer called ISGF3. This trimer acts as a transcriptional activator and is capable of binding to the ISRE of IFN response genes activating their expression. ISG15 and its three conjugating enzymes E1-activating enzyme (UBE1L), E2-conjugating enzyme (hUBCH8, mUBCM8) and E3 ligases (hHERC5/mHERC6, EFP, HHARI, TRIM25), as well as the ISG15 protease USP18 are all IFN-response genes. ISG15 is linked to target proteins via its conjugation system, which is counteracted by USP18 protease activity. Moreover, free ISG15 can act as a cytokine binding to LFA-1, subsequently inducing IFN-γ secretion by natural killer cells and T lymphocytes. Furthermore, USP18 also plays an important role as a negative regulator of IFN type I signaling. USP18 can interact with IFNAR2 and STAT2, competing with JAK1 for receptor binding and thus inhibiting signal transduction. The SCF complex binds USP18 by mediating its poly-ubiquitination and proteasomal degradation, which is inhibited by ISG15 in human cells only. IFNAR, Interferon alpha/beta receptor; Tyk2, Tyrosine kinase 2; JAK1, Janus Kinase 1; STAT1/2, Signal transducer and activator of transcription 1/2; IRF9, Interferon regulatory factor 9; ISGF3, Interferon-stimulated gene factor 3; ISRE, Interferon-sensitive response element; ISG15, IFN-response gene 15; UBE1L, Ubiquitin-activating enzyme E1 homolog; h, human; hUBCH8, Ubiquitin/ISG15-conjugating enzyme E2 L6 in human; m, mouse; mUBCM8, Ubiquitin/ISG15-conjugating enzyme E2 L6 in mouse; mHERC6, E3 ISG15-protein ligase HERC6 in mouse; hHERC5, E3 ISG15-protein ligase HERC5 in mouse; EFP, Estrogen-responsive finger protein; HHARI, Human homolog of Drosophila ariadne-1; TRIM25, Tripartite motif-containing protein 25; USP18, Ubiquitin-specific protease 18; LFA-1, Lymphocyte function-associated antigen 1; ub, ubiquitin; SCF, Skp, cullin, F-box protein; Skp2, S-phase kinase-associated protein 2.
In sharp contrast to ubiquitin, which is highly conserved among different species, the amino acid (AA) composition of ISG15 and its effector functions can differ substantially among species. ISG15 has only been identified in vertebrates, and murine ISG15 and its human counterpart share only 64% homology and 76% similarity on the AA level. This is most likely caused by high evolutionary pressure on anti-pathogenic immune effector functions which need to adapt to immune evasion mechanisms from rapidly mutating pathogens.
One of the features that substantially differ in the ISGylation mechanisms between murine and human ISG15 is the use of certain enzymes. The E1 ubiquitin-activating enzyme E1 homolog (UBE1L/UBA7) is a common enzyme for human and mouse in the ISG15 system (Kim et al., 2006), whereas the E1 counterparts for ubiquitin are ubiquitin-like modifier activating enzyme 1 (UBA1) and ubiquitin-like modifier activating enzyme 6 (UBA6) (Pelzer et al., 2007). Ubiquitin/ISG15-conjugating enzyme E2 L6 (UBCH8) and UBCM8 represent the human and murine E2 conjugating enzymes in ISGylation, respectively. Both share only 76% AA identity, whereas E2 conjugating enzymes for other ubl systems show 95–100% identity (Kim et al., 2004). UBCH8 also interacts with the E1-activating enzyme from the Ub conjugation system which indicates an overlap of both conjugation systems at the level of the E2 enzyme (Zhao et al., 2004). However, the enzyme ubiquitin-conjugating enzyme E2 L3 (UBE2L3/UBCH7) represents the dominant conjugating enzyme in ubiquitination as the KM values uncover a 36-fold higher affinity of UBE1L to UBCH7 as compared to UBCH8 (Durfee et al., 2008). Four cellular ISG15 E3 ligases have been identified so far. Human E3 ISG15–protein ligase HERC5 (HERC5) and the murine counterpart E3 ISG15–protein ligase HERC6 (HERC6) are the dominant E3 ligases in ISGylation that coordinate the conjugation of ISG15 to substrates. Interestingly, both mISG15 and hISG15 can be conjugated either by hHERC5 or mHERC6 (Wong et al., 2006; Ketscher et al., 2012). Furthermore, the E3 ubiquitin-protein ligases estrogen-responsive finger protein (EFP) (Zou and Zhang, 2006), human homolog of Drosophila ariadne-1 (HHARI) (Okumura et al., 2007), and tripartite motif-containing protein 25 (TRIM25) (Park et al., 2016) were also reported to mediate ISGylation.
It was shown that ISGylation can occur in a cotranslational process favoring modification of newly synthesized proteins. As in infected cells mainly viral proteins are translated, ISGylation can interfere with pathogen protein function as shown for capsid assembly of the papilloma virus (Durfee et al., 2010). Furthermore, cellular proteins involved in antiviral defense or export of viral particles were shown to be ISGylated (Perng and Lenschow, 2018).
USP18 Functions: DeISGylation and Negative Regulation of the IFN Response
Ubiquitination and Ubl-conjugation pathways can be reversed by the action of deubiquitinating enzymes (DUBs). These proteases remove or trim Ub/Ubl residues from target proteins. Most of the endogenous proteases from the USP family recognize and deconjugate ubiquitin. However, a small group of proteins from the USP family have been reported to show cross-reactivity and deconjugate ISG15 and ubiquitin, as is the case for USP2, USP5, USP13, USP14, and USP21 (Catic et al., 2007; Ye et al., 2011; Basters et al., 2017).
In addition, many viruses and bacteria have evolved ways to revoke ISGylation as an immune evasion mechanism. Examples of these viral ISG15 proteases were found in the Middle East respiratory syndrome coronavirus (MERS-CoV) (Mielech et al., 2014); Crimean-Congo hemorrhagic fever virus (CCHFV) (Frias-Staheli et al., 2007) or severe acute respiratory syndrome coronavirus (SARS-CoV) (Bekes et al., 2016). They all encode papain-like proteases (PLPs) that impair the host innate immune response.
In contrast to cross-reactive isopeptidases, USP18 is an endogenous ISG15-specific protease that shows no reactivity toward ubiquitin (Malakhov et al., 2002; Basters et al., 2014; Ronau et al., 2016) and it represents the major ISG15 isopeptidase in vivo (Ketscher et al., 2012). In order to define the structural function relationship for this specificity, Basters et al., identified the molecular determinants by solving the crystal structures of mouse USP18 alone and in complex with mouse ISG15. USP18 specificity toward ISG15 is mediated by a small interaction interface of two defined areas within the USP18 sequence, termed ISG15-binding box1 and box2 (IBB-1 and IBB-2, respectively). IBB-1 interacts through hydrophobic contact with ISG15. In ISG15, the side chain of His149 stablizes π-π stacking contact to the aromatic AA Trp121. The IBB1 region, which comprises the USP18 residues Ala138, Leu142, and His251, forms a hydrophobic pocket that specifically accommodates the bulky aromatic side chains of ISG15. Furthermore, the side chains of Pro128 (ISG15) and Leu142 (USP18) contribute to further stability. Of note, replacement of the USP18 residues corresponding to the IBB-1 region, by the homologous residues of the ubiquitin specific protease USP7, resulted in lower affinity toward ISG15. Within the IBB-2 region, the USP18 residues Thr262 and Gln259 interact with the ISG15 residues Gln114, His116, and Gln119 through hydrogen bonds. Likewise, replacement of the USP18 residues corresponding to the IBB-2 region, by the homologous residues of the ubiquitin specific protease USP7, resulted in lower affinity toward ISG15. Moreover, only the ISG15 C-terminal domain (AA residues 77-155) is necessary and sufficient for USP18 binding and activation. Structural data demonstrated that only the ISG15 C-terminal but not the N-terminal UBL domain binds USP18. In vitro assays revealed that USP18 cleaved the ISG15 C-terminal domain as effectively as it cleaved full-length ISG15 (Basters et al., 2017).
Independent of its deconjugating activity, USP18 binds to the IFN-α/β receptor 2 (IFNAR2) complex, where it competes with Janus kinase 1 (JAK1), and thereby negatively regulates type I IFN signaling (Malakhova et al., 2006). Remarkably, USP18 requires Signal transducer and activator of transcription 2 (STAT2) for exerting its inhibitory effect on IFN signaling and IFN-stimulated gene expression (Arimoto et al., 2017) (Figure 1). In humans, binding of free ISG15 prevents proteasomal degradation of USP18 by the S-phase kinase-associated protein 2 (SKP2) and thus is critical to ensure negative regulation of IFN-α/β immunity by stabilizing USP18 (Tokarz et al., 2004; Zhang et al., 2015). However, murine ISG15 appears not to influence the stability of mouse USP18 or IFNAR signaling underlining species specific peculiarities (Knobeloch et al., 2005; Osiak et al., 2005; Zhang et al., 2015).
ISG15 as a Secreted Protein
ISG15 in its unconjugated form has been reported to be released from cells exerting cytokine like activity. Although ISG15 does not have a leader signal sequence to direct its secretion, it has been shown that certain cell types are capable of releasing ISG15 to the extracellular space. Such cell types are epithelial-derived cell lines, fibroblasts, monocytes, neutrophils and lymphocytes (Knight and Cordova, 1991; Bogunovic et al., 2012; Sun et al., 2016). Extracellular ISG15 has been detected in the media of cells as well as in the serum of patients treated with IFN-α/β (D'Cunha et al., 1996). Early work suggested that secreted ISG15 elicits IFN-γ secretion from lymphocytes (Recht et al., 1991). Bacillus Calmette–Guérin (BCG) can also induce IFN-γ secretion from control peripheral blood mononuclear cells (PBMCs) when stimulated with recombinant human ISG15 (Bogunovic et al., 2012). In normal control patients, extracellular interleukin (IL)-12 played a synergistic role with ISG15 stimulating the release of IFN-γ and IL-10. Both, natural killer (NK) cells and T lymphocytes secreted IFN-γ in response to IL-12 and ISG15 (Bogunovic et al., 2012). However, IFN-γ secretion was not detected in PBMCs from ISG15-deficient patients and that defect leads to susceptibility to mycobacterial disease and autoinflammation (Bogunovic et al., 2012).
Recently, the adhesion molecule leukocyte function associated antigen-1 (LFA-1) has been identified as the receptor for extracellular ISG15 (Swaim et al., 2017) (Figure 1). To identify this receptor, ISG15 ubiquitin-activated interaction trap (UBAIT) was employed (O'Connor et al., 2015).
Conflict of Interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Abstract
The interferon (IFN)-stimulated gene product 15 (ISG15) represents an ubiquitin-like protein (Ubl), which in a process termed ISGylation can be covalently linked to target substrates via a cascade of E1, E2, and E3 enzymes. Furthermore, ISG15 exerts functions in its free form both, as an intracellular and as a secreted protein. In agreement with its role as a type I IFN effector, most functions of ISG15 and ISGylation are linked to the anti-pathogenic response. However, also key roles in other cellular processes such as protein translation, cytoskeleton dynamics, exosome secretion, autophagy or genome stability and cancer were described. Ubiquitin-specific protease 18 (USP18) constitutes the major ISG15 specific protease which counteracts ISG15 conjugation. Remarkably, USP18 also functions as a critical negative regulator of the IFN response irrespective of its enzymatic activity. Concordantly, lack of USP18 function causes fatal interferonopathies in humans and mice. The negative regulatory function of USP18 in IFN signaling is regulated by various protein–protein interactions and its stability is controlled via proteasomal degradation. The broad repertoire of physiological functions and regulation of ISG15 and USP18 offers a variety of potential intervention strategies which might be of therapeutic use. Due to the high mutation rates of pathogens which are often species specific and constantly give rise to a variety of immune evasion mechanisms, immune effector systems are under constant evolutionarily pressure. Therefore, it is not surprising that considerable differences in ISG15 with respect to function and sequence exist even among closely related species. Hence, it is essential to thoroughly evaluate the translational potential of results obtained in model organisms especially for therapeutic strategies. This review covers existing and conceptual assay systems to target and identify modulators of ISG15, ISGylation, USP18 function, and protein–protein interactions within this context. Strategies comprise mouse models for translational perspectives, cell-based and biochemical assays as well as chemical probes.
Glossary
Abbreviations
| IFN | Interferon |
| ISG15 | IFN-stimulated gene product 15 |
| Ubl | Ubiquitin-like protein |
| USP18 | Ubiquitin-specific protease 18 |
| STAT2 | Signal transducer and activator of transcription 2 |
| ABPs | Activity-based probes |
| SUMO | Small ubiquitin-related modifier |
| FAT10 | Human leukocyte antigen (HLA)-F adjacent transcript 10 |
| NEDD8 | Neural precursor cell expressed, developmentally down-regulated 8 |
| AA | Amino acid |
| UBE1L | Ubiquitin-activating enzyme E1 homologue |
| UBA1 | Ubiquitin-like modifier activating enzyme 1 |
| UBA6 | Ubiquitin-like modifier activating enzyme 6 |
| UBCH8 | Ubiquitin/ISG15-conjugating enzyme E2 L6 in human |
| UBCM8 | Ubiquitin/ISG15-conjugating enzyme E2 L6 in mouse |
| UBE2L3/UBCH7 | Ubiquitin-conjugating enzyme E2 L3 |
| HERC5 | E3 ISG15-protein ligase HERC5 |
| HERC6 | E3 ISG15-protein ligase HERC6 |
| h | Human |
| m | Murine |
| EFP | Estrogen-responsive finger protein |
| HHARI | Human homolog of Drosophila ariadne-1 |
| TRIM25 | Tripartite motif-containing protein 25 |
| DUB | Deubiquitinating enzyme |
| MERS-CoV | Middle East respiratory syndrome coronavirus |
| CCHFV | Crimean-Congo hemorrhagic fever virus |
| SARS-CoV | Severe acute respiratory syndrome coronavirus |
| PLP | Papain-like protease |
| IBB1 | ISG15-binding box1 |
| IBB2 | ISG15-binding box2 |
| IFNAR2 | IFN-α/β receptor 2 |
| JAK1 | Janus kinase 1 |
| SKP2 | S-phase kinase-associated protein 2 |
| BCG | Bacillus Calmette–Guérin |
| PBMCs | Peripheral blood mononuclear cells |
| IL | Interleukin |
| NK | Natural killer |
| LFA-1 | leukocyte function associated antigen-1 |
| UBAIT | Ubiquitin-activated interaction trap |
| UPS | Ubiquitin-proteasome system |
| UAE | Ubiquitin activation enzyme (E1) |
| CDC34 | Human ubiquitin-conjugating enzyme E2 R1 |
| UBC13-UEV1A | Ubiquitin-conjugating enzyme variant |
| UBE2N | Ubiquitin-conjugating enzyme E2 N |
| ICAM1 | Intercellular Adhesion Molecule 1 |
| VS | Vinyl methyl sulfone |
| VME | Vinyl methyl ester |
| PRG | Propargylamide |
| ct-ISG15 | C-terminal Ubl domain of ISG15 |
| SPPS | Solid-phase peptide chemistry |
| FMOC | 9-fluorenylmethoxycarbonyl |
| HTS | High-throughput drug screening |
| amc | 7-Amino-4-methylcoumarin |
| FP | Fluorescence polarization |
| TAMRA | 5-carboxytetra-methylrhodamine |
| CHO | Aldehyde |
| AGS | Aicardi-Goutières syndrome |
| PTS | Pseudo-TORCH syndrome |
| MS | Mass spectroscopy |
| BMDMs | Bone marrow derived macrophages |
| VACV | Vaccinia virus |
| OXPHOS | Oxidative phosphorylation |
| FMD | Foot-and-mouth disease |
| IP | Immunoprecipitation |
| LC | Liquid chromatography |
| PROTACs | Proteolysis Targeting Chimeras |
| BRET | Bioluminescence Resonance Energy Transfer |
| FRET | Fluorescence resonance energy transfer. |
Footnotes
Funding. This work was supported by a grant from the Deutsche Forschungsgemeinschaft KN 590/7-1 and BMBF (BacVirISG15).

