ISG15 is a 17 kDa ubiquitin-like protein that functions in the innate immune response to pathogens. Structurally, ISG15 consists of two ubiquitin-like domains connected by a short linker. The canonical function of ISG15 is that of an intracellular ubiquitin-like protein modifier. Like ISG15, the human E1 (Ube1L), E2 (Ube2L6) and E3 (Herc5) enzymes for ISG15 conjugation are all induced at the transcriptional level by type I IFN.

An interesting feature of the human ISG15 conjugating system is that a single E3 enzyme (human Herc5) is responsible for the ISGylation of thousands of substrate proteins. We have shown that hHerc5 comigrates with polyribosomes by sucrose gradient sedimentation and that it binds to to a so far unidentified determinant (protein and/or RNA) on the 60S subunit. Nascent polypeptides are ISGyated while they are still being translated, indicating that ISGylation is a cotranslational process. Our working model is that Herc5 modifies a large fraction of the proteins that a cell is translating, accounting for how a single ligase can target thousands of proteins. We further hypothesize that, in the context of a Type I Interferon response, newly translated viral proteins may be the biologically relevant targets of ISGylation. ISGylation does not target proteins for proteasomal degradation; we hypothesize that it sterically interferes with the function of proteins to which it is conjugated. Unlike ubiquitin, ISG15 does not form poly-ISG15 chains.

Like ubiquitination, ISGylation occurs via formation of an isopeptide bond between an epsilon-amino group of lysine side chains of substrate proteins and carboxyl group of the terminal amino acid (glycine) of ISG15.

DiGly proteomics has been used to study the sites of ubiquitination. Tryptic digestion of ISG15-modified protein produces peptides containing a di-glycine (DiGly)-modified lysine as shown in the left figure. A general motif antibody that recognizes the K-ε-GG motif has been developed to enrich the site-containing peptides which are analyzed by mass spectrometry.

To exclude the DiGly peptides coming from ubiquitination, a de-ubiquitiunase (DUB) Usp2 is used prior to trypsin digestion to remove ubiquitin.

In our recent proteomics, HEK293T cells were transfected with the four components for ISG15 conjugation including Ube1L, UbcH8, hHerc5 and ISG15. Herc5-knock out cells were also transfected as the controls. Following the Usp2 treatment and the DiGly proteomics wokrflow, we have identified 3733 ISGylation sites corresponding to 1524 proteins.

Several features of the ISGylation sites have been studied. Significant differences between the ISG15-modified lysine and those of the proteome background have been found. For example, specific amino acids flanking the Herc5 modification sites are either enriched or depleted relative to the amino acid frequency of the proteome. This finding makes it possible to distinguish potential targets and has implication for the evolution of viral protein sequences.

While hHerc5 is the major ISG15 E3 ligase in human, mice do not have a Herc5 gene. Instead, a closely related gene, mouse Herc6, encodes the major ISG15 ligase. A recent study showed that the Herc5-related ISG15 ligases are under strong positive evolutionary selection. However, little is know about if these ligases function via similar mechanisms or if they share similar substrate specificities.

In this project, ISGylated and un-ISGylated sites together with seven upstream (N-terminal side) amino acids and seven downstream (C-terminal side) amino acids (7by7 seq) were extracted. The AAIndex1 database was utilized to find different physicochemical and biochemical properties of each amino acid at each position. A deep-learning based prediction tool (SPIDER3) was also used for four backbone torsion angles calculation and secondary structure prediction of each residue. All these properties and torsion angles of each amino acid residue were used as features to train the models.