The dimeric deubiquitinase USP28 integrates 53BP1 and MYC functions to limit DNA damage

Abstract DNA replication is a major source of endogenous DNA damage in tumor cells and a key target of cellular response to genotoxic stress. DNA replication can be deregulated by oncoproteins, such as transcription factor MYC, aberrantly activated in many human cancers. MYC is stringently regulated by the ubiquitin system - for example, ubiquitination controls recruitment of the elongation factor PAF1c, instrumental in MYC activity. Curiously, a key MYC-targeting deubiquitinase USP28 also controls cellular response to DNA damage via the mediator protein 53BP1. USP28 forms stable dimers, but the biological role of USP28 dimerization is unknown. We show here that dimerization limits USP28 activity and restricts recruitment of PAF1c by MYC. Expression of monomeric USP28 stabilizes MYC and promotes PAF1c recruitment, leading to ectopic DNA synthesis and replication-associated DNA damage. USP28 dimerization is stimulated by 53BP1, which selectively binds USP28 dimers. Genotoxic stress diminishes 53BP1–USP28 interaction, promotes disassembly of USP28 dimers and stimulates PAF1c recruitment by MYC. This triggers firing of DNA replication origins during early response to genotoxins and exacerbates DNA damage. We propose that dimerization of USP28 prevents ectopic DNA replication at transcriptionally active chromatin to maintain genome stability.

(B) Immunoblot analysis of protein levels of USP25, USP28 and MYC in different clones of HLF cells with sgRNA against USP25 and/or USP28.Note that a faint band is detected by the USP28 antibody in USP28-KO clones, as was also the case for the Cre-mediated knockout of USP28 in murine cells (45).This is most likely due to cross reactivity of the antibody with the highly homologous USP25 protein, as the band is absent in the USP28/USP25 double knockout cells.
(C) Immunoblotting analysis of MYC protein level in HLF USP28-Ctrl/KO cells, treated with cycloheximide (100 μg/ml) for the indicated time points.Image shows one representative biological experiment (n=3).
(D) Top five gene sets for the indicated databases, enriched within the USP28-deregulated genes based on the analysis of RNA-seq data in HLF USP28-KO and USP28-Ctrl cells.Analysis by the Enrichr portal (98).
(F) PLA assays with antibodies against MYC and USP28 in HLF USP28-KO cells expressing USP28-WT/M or a control vector.Quantification shows data points for one representative experiment (n=2).At least 30 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, *P < 0.05, ***P < 0.001.Scale bar = 10 μm.
(H) Immunoblotting analysis of USP28-KO MEFs and HeLa cells, expressing USP28-WT or USP28-M, treated with cycloheximide (100 μg/ml) for the indicated time points.Image shows one of two representative experiments (n=2).
(I) Ubiquitin pulldown assays with HeLa USP28-KO cells expressing MYC, WT His-Ub and USP28-WT/M, immunoblotting with antibodies against total ubiquitin.Image shows one representative experiment (n=2).(A) PLA assays with antibodies against MYC and CTR9 or PAF1 in HLF USP28-KO cells expressing USP28-WT/M or a control vector.Quantification shows data points for one representative experiment (n=2).At least 65 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, *P < 0.05, ****P < 0.0001.
(D) Colony formation assay in soft agar for USP28-KO MEFs, expressing an shRNA against p19Arf and vectors encoding USP28-WT, USP28-M or a control vector.Quantification shows data points for one representative experiment (n=4).At least 6 view fields were quantified.The data were analyzed with ordinary one-way ANOVA followed by Tukey's multiple comparison test of selected pairs, *P < 0.05, ns P > 0.05.Error bars denote S.D. Scale bar = 10 μm.
(C) Immunofluorescence analysis documenting γH2AX and Cyclin A intensity in HLF USP28-WT or USP28-M cells with or without serum deprivation.Quantification shows data points for one representative experiment (n=2).At least 93 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, ****P < 0.0001, ns P > 0.05.Scale bar = 10 μm.
(D) Immunoprecipitation analysis with the HA antibodies in HeLa USP28-KO cells, expressing FLAG-tagged USP28-WT and HA-tagged wildtype or mutant USP28.Numbers show the ratio of FLAG grayscale intensity between IP and Lysate normalized to the control group.
(E) PLA assays with antibodies against FLAG and HA or IgG in HeLa USP28-KO cells, expressing FLAG-tagged USP28-WT and HA-tagged USP28-WT/R519W.Quantification shows data points for one representative experiment (n=2).At least 80 cells were quantified.The data were analyzed with two-tailed, Mann-Whitney test, **P < 0.01.
(C) Immunoprecipitation analysis with FLAG antibodies from HeLa cells, transfected with HAand FLAG-tagged USP28 before and after etoposide treatment (5 μM) for indicated time point.Image shows one representative experiment (n=2).
(D) Immunoblotting with Native PAGE documenting expression of USP28-M and the transition between USP28-WT and USP28-M after etoposide (5 μM, 30 min) treatment in HeLa USP28-KO cells transfected with USP28-WT/M.Right panel shows the mean of three independent biological replicates (n=3).The data were analyzed with ordinary one-way ANOVA test followed by Tukey's multiple comparison of selected pairs, *P < 0.05.
(G) qPCR showing the mRNA level of MYC in HLF USP28-Ctrl cells with or without etoposide treatment (5 μM, 30 min).Quantification shows data points for one representative experiment (n=2).The data were analyzed from three technical replicates with two-tailed, unpaired t test, ns P > 0.05.
(C) EdU incorporation assays in HLF USP28-Ctrl/KO cells with shCtrl/sh53BP1.Quantification shows data points for one representative experiment (n=2).At least 204 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, ****P < 0.0001, ns P > 0.05.
(D) EdU incorporation assays in HLF shCtrl/sh53BP1 cells with or without Mirin treatment (25 mM, 2 hr).At least 52 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pair, ns P > 0.05.
(E) Immunofluorescence analysis with γH2AX antibodies in HLF USP28-Ctrl/KO cells with shCtrl/sh53BP1.Quantification shows data points for one representative experiment (n=2).At least 204 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, ****P < 0.0001, ns P > 0.05.
(F) Neutral comet assays showing the DSBs in HLF USP28-Ctrl/KO cells with shCtrl/sh53BP1.Quantification shows data points for one representative experiment (n=2).At least 82 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, ****P < 0.0001, ns P > 0.05.(A) EU incorporation assays in HLF shCtrl/sh53BP1 cells with or without etoposide treatment (5 μM, 30 min).At least 96 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, ****P < 0.0001.
(B) EdU incorporation assays in etoposide-treated (5 μM, 30 min followed by 0 or 16 hr release) HLF cells with or without Mirin treatment (25 mM, 2 hr for unreleased cells and 18 hr for released cells).At least 40 cells were quantified.The data for shCtrl cells were the same as the shCtrl cells in Fig. S6D.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, **P < 0.01, ns P > 0.05.
(C) EdU incorporation assays in HLF USP28-Ctrl/KO cells with or without etoposide treatment (5 μM, 30 min).Quantification shows data points for one representative experiment (n=2).At least 134 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, ****P < 0.0001.
(D) EdU incorporation assays in HLF USP28-WT or USP28-M cells with or without etoposide treatment (5 μM, 30 min).Quantification shows data points for one representative experiment (n=2).At least 43 cells were quantified.The data were analyzed with Kruskal-Wallis test followed by Dunn's multiple comparison of selected pairs, *P < 0.05, **P < 0.01, ns P > 0.05.
(B) Crystal violet staining showing etoposide (5 μM) treated HLF and HeLa cells with indicated time points with or without thymidine (2 mM, 1 hr prior etoposide treatment).Right panels show the mean of three independent biological replicates (n=3).The data were analyzed with ordinary one-way ANOVA followed by Šídák's multiple comparison of selected pairs, *P < 0.05, **P < 0.01, ns P > 0.05.Error bars denote S.D.