USP16 is an ISG15 cross-reactive deubiquitinase that targets pro-ISG15 and ISGylated proteins involved in metabolism

Abstract

Interferon-induced ubiquitin (Ub)-like modifier ISG15 covalently modifies host and viral proteins to restrict viral infections. Its function is counteracted by the canonical deISGylase USP18 or Ub-specific protease 18. Notwithstanding indications for the existence of other ISG15 cross-reactive proteases, these remain to be identified. Here, we identify deubiquitinase USP16 as an ISG15 cross-reactive protease by means of ISG15 activity-based profiling. Recombinant USP16 cleaved pro-ISG15 and ISG15 isopeptide-linked model substrates in vitro, as well as ISGylated substrates from cell lysates. Moreover, interferon-induced stimulation of ISGylation was increased by depletion of USP16. The USP16-dependent ISG15 interactome indicated that the deISGylating function of USP16 may regulate metabolic pathways. Targeted enzymes include malate dehydrogenase, cytoplasmic superoxide dismutase 1, fructose-bisphosphate aldolase A, and cytoplasmic glutamic-oxaloacetic transaminase 1. USP16 may thus contribute to the regulation of a subset of metabolism-related proteins during type-I interferon responses.

Keywords: ISG15; ISGylation; USP16; activity-based probe; metabolism.

Fig. 1.

Activity-based pull-down assay identifies USP16 as an ISG15 cross-reactive DUB (A) Streamlined workflow for identification of ISG15-reactive proteases in lysates of HAP1 cells. (B and C) Volcano plots of a comparative proteomic analysis of the trypsin digests of the materials retrieved by Streptavidin beads. Comparison of biotin-based immunoprecipitation of Biotin-hISG15_CTD-PA probe labeled versus unlabeled IFN-α2 stimulated HAP1 WT cell lysate (B) and comparison of biotin-based immunoprecipitation of IFN-α2 stimulated versus unstimulated cell lysates labeled with Biotin-hISG15_CTD-PA probe (C). The identified Ub/ISG15 proteases are shown in red. The statistical cutoff values used for the proteomic analyses are FDR: 0.01 and s0: 0.1. (D) 5 μM of recombinant human USP16 catalytic domain (CD, aa 196-823) reacts with 5 μM of Rhodamine-tagged Ub-PA, as well as human ISG15_CTD-PA, mouse ISG15_CTD-PA, and mouse full-length ISG15-PA probes in 10 μL volume at 37 °C for 1 h. After reaction with the probe, samples were denatured, resolved by SDS-PAGE, scanned for fluorescence, and stained with InstantBlue Coomassie dye. The probe-labeled USP16 CD is indicated by red asterisks. Representative data of three (n = 3) independent experiments. See also SI Appendix, Fig. S1.

Fig. 2.

USP16 cleaves ISG15-related substrates in vitro. (A) Human proISG15 is cleaved by recombinant human USP18, USP7, USP5, USP16FL, and USP16CD as analyzed by SDS-PAGE and InstantBlue Coomassie staining. The position of marker proteins is indicated. Representative data of two (n = 2) independent experiments. (B) Catalytic activity of recombinant human USP18, USP16FL, and USP16CD toward isopeptide-linked Ub-FP and ISG15-FP substrates. The indicated amounts of USP16 FL/CD were incubated with 200 nM Ub-FP or ISG15-FP. Substrate cleavage was monitored by a change in FP [in millipolarization units (mP)]. Representative data of two (n = 2) independent experiments. (C) ISG15 deconjugation in lysates of HAP1 USP18KO cells. HAP1 USP18KO cells were stimulated with IFN-β to induce ISGylation. 40 μg of cell lysate in 10 μL was incubated with recombinant USP16 CD^WT or USP16CD^C205S at final concentrations of 5 μM at 37 °C for 2 h. Proteins were separated by SDS-PAGE and immunoblotted with anti-human ISG15 antibody. Probing for β-actin served as a loading control. The position of marker proteins is indicated. Representative data of two (n = 2) independent experiments. See also SI Appendix, Fig. S2.

Fig. 3.

Loss of USP16 increases cellular ISGylation. (A) Knockdown of USP16. HAP1 WT cells were transfected with either control siRNA (siCTR) or three different siRNAs (siUSP16#1, siUSP16#2, and siUSP16#3) against USP16 for 72 h. IFN-β (1,000 U/mL) was added for 24 h prior to harvesting. Lysates were resolved by SDS-PAGE. Immunoblot analysis was performed using the indicated antibodies. (B) ISG15 deconjugation in lysates of HAP1 WT and USP16KO cells. Both HAP1 WT and USP16KO cells were stimulated with 1,000 U/mL of IFN-β for 24 h. Cell lysates were resolved by SDS-PAGE and analysed by immunoblotting using the indicated antibodies. For panels A and B, β-actin served as a loading control. Representative data of three (n = 3) independent experiments. See also SI Appendix, Figs. S3 and S4.

Fig. 4.

Analysis of the USP16-dependent ISG15 interactome links USP16 to cellular metabolism. (A) Schematic representation of the workflow for ISG15 immunoprecipitation mass spectrometry (IP-MS) used to analyze the ISG15 interactome. HAP1 WT and USP16KO cells were stimulated by 1,000 U/mL of IFN-β for 48 h to produce ISGylated substrates (two biological replicates). (B) Volcano plot showing all proteins identified for the IFN-β stimulated USP16KO samples, compared to the IFN-β stimulated WT HAP1 cells. Dashed lines indicate a cutoff at a difference of the Log2 intensities bigger than 1.5 and a P value of 0.05 (−log10 = 1.3), n = 2 independent experiments. Proteins in red showed increased interaction with ISG15 in the USP16KO cells compared to the WT cells. These are termed the “USP16-dependent ISG15 interactome”. Proteins indicated in blue showed decreased levels in the USP16KO cells compared to WT cells. (C) GO enrichment analysis of the USP16-dependent ISG15 interactome. The bar graph shows the most significantly overrepresented GO terms for CC in light green, MF in purple, and BP in dark blue, compared against the annotated human proteome. The full GO terms for MF are found in SI Appendix, Fig. S7A and BP are found in SI Appendix, Fig. S7B. (D) STRING network analysis of the USP16-dependent ISG15 interactome, with a STRING interaction confidence of 0.4 or higher. Cytoscape software was used to visualize the interaction network. Color and node size indicate the differences (of the Log2-transformed intensities) in abundance for USP16KO HAP1 cells compared with the WT HAP1 control upon IFN-β treatment. (E) Cluster 1 and cluster 2 contain multiple proteins involved in carbon and pyruvate metabolism and hydrogen peroxide metabolic process, respectively. MCODE was used to extract the most highly interconnected clusters (clusters 1 and 2) from the network shown in (D). See also SI Appendix, Figs. S5 and S6.

Fig. 5.

Validation of selected proteins as ISGylated substrates for deISGylation by USP16. (A) DeISGylation assay in HEK293T cells that overexpress USP16 or USP18. Cellular ISGylation was achieved by cotransfecting an ISGylation plasmid mixture including 2 μg HA-UBE1L, 2 μg Flag-UBCH8, 2 μg S-HERC5, and 2 μg Flag ISG15 in HEK293T cells in 6-cm dishes for 24 h. Empty vector (4 μg; lane 3), USP16 WT/C205S, or USP18 WT/C64A were cotransfected with plasmids encoding the ISGylation machinery as indicated. Cell lysates were resolved by SDS-PAGE and analysed by immunoblotting using the indicated antibodies. Representative data of three (n = 3) independent experiments. (B–E) Validation of (Myc)₂-GOT1 (B), (Myc)₂-ALDOA1 (C), (Myc)₂-SOD1 (D), and (Myc)₂-GOT1 (E) as ISGylated substrates for USP16-mediated deISGylation by staining the immunoblot with anti-Flag antibodies to detect Flag-ISG15. Plasmids encoding (Myc)₂-tagged substrates protein were cotransfected with plasmids encoding ISGylation machinery, and GFP empty vector (EV), GFP-USP16WT or GFP-USP16C205S mutant in HEK293T cells in 6 cm dishes for 24 h. (Myc)₂-tagged proteins were immunoprecipitated from the cell lysates by Myc trap beads. The retrieved proteins were interrogated by immunoblot using the indicated antibodies. Quantification of the amounts of ISGylated protein in lanes 2 to 4 by detection with the anti-Myc and anti-Flag antibodies. The ratio of anti-flag vs. anti-Myc protein intensity was normalized to the EV control. Bar graphs report mean; error bars reflect ± SD of three (n = 3) independent experiments. All P values were calculated using Student’s t test: **P < 0.05, ***P < 0.001, and NS = not significant.

Fig. 6.

Overview of the roles of USP16 as a dual deubiquitylating and deISGylating enzyme. Canonical ubiquitylated protein substrates are colored in yellow and novel ISGylated metabolic enzyme substrates in green.