Selective USP7 inhibition elicits cancer cell killing through a p53-dependent mechanism AbstractUbiquitin specific peptidase 7 (USP7) is a deubiquitinating enzyme (DUB) that removes ubiquitin tags from specific protein substrates in order to alter their degradation rate and sub-cellular localization. USP7 has been proposed as a therapeutic target in several cancers because it has many reported substrates with a role in cancer progression, including FOXO4, MDM2, N-Myc, and PTEN. The multi-substrate nature of USP7, combined with the modest potency and selectivity of early generation USP7 inhibitors, has presented a challenge in defining predictors of response to USP7 and potential patient populations that would benefit most from USP7-targeted drugs. Here, we describe the structure-guided development of XL177A, which irreversibly inhibits USP7 with sub-nM potency and selectivity across the human proteome. Evaluation of the cellular effects of XL177A reveals that selective USP7 inhibition suppresses cancer cell growth predominantly through a p53-dependent mechanism: XL177A specifically upregulates p53 transcriptional targets transcriptome-wide, hotspot mutations in TP53 but not any other genes predict response to XL177A across a panel of ~500 cancer cell lines, and TP53 knockout rescues XL177A-mediated growth suppression of TP53 wild-type (WT) cells. Together, these findings suggest TP53 mutational status as a biomarker for response to USP7 inhibition. We find that Ewing sarcoma and malignant rhabdoid tumor (MRT), two pediatric cancers that are sensitive to other p53-dependent cytotoxic drugs, also display increased sensitivity to XL177A. IntroductionUbiquitination is a post-translational modification that regulates myriad fundamental cellular processes, most notably protein degradation1, and is implicated in numerous disease settings, including cancer2,3, infection4, and neurodegeneration5. In particular, the ubiquitin-proteasome system (UPS) has become a target of interest in oncology, as proteasome inhibitors and ligands targeting the E3 ubiquitin ligases are approved as cancer therapies6,7. As the enzymes responsible for proteolytically cleaving ubiquitin moieties, deubiquitinating enzymes (DUBs) regulate virtually all physiological and pathophysiological processes1,8,9,10. Ubiquitin-specific peptidase 7 (USP7) is one of the most widely studied DUBs, and it has been associated with multiple substrates, cellular pathways, and disease states. USP7 was first discovered as an interacting partner and stabilizer of the Herpesvirus E3 ligase ICP011. Since then, USP7 has also been reported to interact with and regulate numerous mammalian E3 ligases, epigenetic modifiers, and transcription factors, among other targets12.Of these potential substrates, USP7鈥檚 stabilization of MDM2, the E3 ligase for p53, has garnered the most interest from a mechanistic and therapeutic standpoint given the role of p53 as tumor suppressor across many cancer types. Although many cancers are driven by mutant TP53, roughly half of all adult malignancies and 95% of pediatric cancers harbor intact WT TP53 and may benefit from therapeutic approaches that stabilize p5313,14. Indeed, MDM2 inhibitors such as idasanutlin and the dual MDM2/MDM24 inhibitor ATSP-7041 are currently undergoing clinical evaluation15,16, supporting investigation of additional p53 stabilizing strategies such as USP7 inhibition. However, USP7鈥檚 promiscuity has raised questions over the relative importance of p53 in its overall cellular function. USP7 has more than 20 reported substrates (see recent reviews12,17,18), several of which (PTEN19, FOXO420, N-Myc21, PCNA22, Claspin23, and others) play a key role in proliferation and tumorigenesis. Indeed, USP7 has now been proposed as a therapeutic target independent of TP53 mutational status in multiple cancers including bortezomib-resistant multiple myeloma24, neuroblastoma21, T-cell acute lymphoblastic leukemia25, and acute myeloid leukemia26. These findings have spurred interest in the development of specific USP7 inhibitors, as there now appear to be several cancer indications that may benefit from USP7 modulation. There have been a number of small molecule USP7 inhibitors reported to date26,27, and these compounds have consistently demonstrated the ability to stabilize p53 protein levels in cyto, although they do not exhibit p53-dependent growth suppression26,28. These findings have largely been credited to USP7鈥檚 role in other cellular pathways, although they could also be due to unidentified off-targets of these compounds.We recently reported XL18829, a potent and selective USP7 inhibitor that, in contrast to other reported small molecules, induces p53-dependent growth suppression in TP53-WT Ewing Sarcoma cell lines30. This somewhat surprising finding suggested to us that the functional significance of the USP7-p53 axis in Ewing Sarcoma, and perhaps cancer more broadly, is not completely understood and that improved research tools are needed to enable reliable on-target analysis of USP7 biology. To that end, we undertook the development of novel irreversible XL188 analogs that could be comprehensively profiled for USP7 selectivity and then used as chemical tools to elucidate cellular responses to selective USP7 inhibition.Here, we report the discovery of a 340 pM irreversible inhibitor of USP7, XL177A. XL177A labels the catalytic cysteine, C223, of USP7 with exquisite selectivity for USP7 across the DUBome and human proteome. Evaluation of XL177A and its enantiomer, XL177B, as a negative control, across a broad panel of 484 cancer cell lines and in a series of targeted confirmatory studies, found that TP53 mutational status predicted response to selective USP7 inhibition across multiple cancer lineages. We also found that, in TP53-WT cell lines, TP53 knockout (KO) rescues XL177A and USP7-KO-mediated growth suppression. Overall, our findings demonstrate that covalent targeting of the active-site is a viable strategy for development of DUB inhibitors, provide a well-validated chemical tool for investigation of USP7 biology, and demonstrate that p53 is a relevant biomarker for cellular responses to USP7 inhibition.ResultsRational design of irreversible USP7 inhibitor XL177AWe recently reported the structure-guided development of XL188, a noncovalent inhibitor of USP7 that binds in the thumb-palm cleft that guides the ubiquitin C-terminus into the active site. Specifically, a co-crystal structure of XL188 and the USP7 catalytic domain shows the compound bound within the S4-S5 pocket of enzyme about 5鈥壝?removed from the catalytic cysteine (Fig.1a). The proximity of the compound to the catalytic Cys residue suggested that elaboration of XL188 may yield a new series of covalent (irreversible) USP7 inhibitors. Chemical synthesis and biochemical characterization of 50 XL188 analogs (Liu et al., in preparation) culminated in development of XL177A as a highly potent and selective irreversible inhibitor of USP7 (Fig.1b). In an enzymatic assay using full-length USP7 and fluorogenic substrate ubiquitin-7-amino-4-methylcoumarin (Ub-AMC), XL177A inhibited USP7 with an IC50 of 0.34鈥塶M (Fig.1c). An accurate measure of the enzyme inactivation rate (kinact) and inhibition constant (KI) for XL177A was not feasible due to rapid and complete labeling of USP7 by the compound. However, determination of these parameters for XL041, a precursor compound containing the same electrophilic warhead, confirmed an irreversible mode of inhibition for the compound series (Fig.S1). A covalent binding mode was also confirmed for XL177A using mass spectrometry. We incubated purified USP7 catalytic domain with vehicle (DMSO) or XL177A for 15鈥塵inutes and analyzed samples using capillary electrophoresis-mass spectrometry (CE-MS). Quantitative labeling of USP7 by XL177A with a mass shift corresponding to inhibitor mass minus HCl, was observed (Fig.1d). MS/MS analysis confirmed binding to the catalytic residue, C223 (Fig.1e). An analog of XL177A without the chloride leaving group, XL058, was unable to label the USP7 catalytic domain by MS (Fig.S2), and no mass shift was detected when XL177A was incubated with USP7 C223A (Fig.S3). To test USP7 target engagement in a cellular context, we used competitive activity-based protein profiling (ABPP) with MCF7 crude cell extracts and the DUB activity-based probe (ABP) hemagglutinin (HA)-ubiquitin vinylmethylsulfone (HA-Ub-VS): XL177A inhibited HA-Ub-VS labeling with IC50s of 85 and 8鈥塶M following 30鈥塵in and 4鈥塰r compound preincubations, respectively (Fig.1f,g). We also employed live cell treatment and competitive ABPP to demonstrate that XL177A inhibits USP7 in cyto, with an IC50 of 39鈥塶M after 6鈥塰r treatment (Fig.2a,b). The enantiomer of XL177A, XL177B (Fig.1b), also demonstrated time-dependent USP7 inhibition, but with ~500-fold less potency against USP7 in experiments using purified enzyme, crude lysate, and live cell treatment (Figs.1c,e, 2a,b). XL177B is thus a useful matched control for assessing the USP7-specific effects of XL177A.Figure 1Design of a potent irreversible USP7 inhibitor. (a) Co-crystal structure of XL188 bound to the USP7 catalytic domain, highlighting the ligand鈥檚 solvent accessibility and distance to the catalytic cysteine (PDB: 5VS6). (b) Chemical structures of XL188, XL177A, and XL177B. (c) IC50 plots of full-length USP7 cleavage of Ub-AMC following 6-hour pre-treatment with XL177A or XL177B (n鈥?鈥? experimental replicates, error bars鈥?鈥?5% CI). (d) Intact MS spectra of USP7 catalytic domain treated with DMSO (red) or a 2.5-fold molar excess XL177A for 15鈥塵inutes (green), showing single labeling of USP7 with a mass shift corresponding to the inhibitor (n鈥?鈥? experimental replicates). (e) MS-MS spectra of USP7 catalytic domain treated with 2.5-fold molar excess XL177A for 15鈥塵inutes. The XL177A-labeled cysteine (C223) is indicated in the peptide sequence shown. (f) Western blots showing USP7 labeling by the DUB ABPP HA-Ub-VS in whole cell lysate after 30-minute or 4-hour pre-treatment with XL177A or XL177B. Full blots are presented in Fig.S14. (g) Densitometry from F (n鈥?鈥? experimental replicates, error bars鈥?鈥塖D).Full size imageFigure 2XL177A is a potent USP7 inhibitor and p53 stabilizer in cyto. (a) Western blots showing USP7 labeling by HA-Ub-VS after 6-hour cell treatment with XL177A or XL177B. (b) Densitometry from A (n鈥?鈥? experimental replicates, error bars鈥?鈥塖D). (c) Whole cell lysate Western blots of MCF7 cells after 2-hr treatment with XL177A or XL177B (n鈥?鈥? experimental replicates). Full blots are presented in Fig.S14. (d) Whole cell lysate Western blots of MCF7 cells after 16-hr treatment with XL177A or XL177B (n鈥?鈥?). (d) Quantitative real-time PCR of MCF7 cells treated for 24鈥塰ours with 1 渭M XL177A or XL177B (n鈥?鈥? experimental replicates with 3 technical replicates each, error bars鈥?鈥塖EM). E) Cell cycle analysis based on propidium iodide staining of MCF7 cells after 24-hr treatment with 1 渭M XL177A or XL177B (n鈥?鈥? experimental replicates, error bars鈥?鈥塖D).Full size imageTo further confirm inhibition of cellular USP7, we determined the impact of XL177A/B on the HDM2-p53 signaling axis, which is a well validated target of USP7 DUB activity. As expected, treatment of MCF7 cells, which express WT TP53, with XL177A induced rapid degradation of HDM2 within 2鈥塰ours, followed by increases in p53 and downstream p21 protein levels (Fig.2c,d) and stimulating transcription of p53 target genes related to both cell cycle arrest (CDKN1A and GADD45A) and apoptosis (BAX and DDB2) (Fig.2e). Indeed, 1 渭M XL177A induced complete G1 arrest in MCF7 cells after 24鈥塰ours (Figs.2f and S4). In contrast, negative control XL177B did not exert any of the same effects except at 10 渭M, where it demonstrated USP7 inhibition and on-target cell signaling effects (Figs.2c鈥揺, and S4). Due to negative feedback signaling, whereby p53 transcriptionally upregulates MDM2, after 24鈥塰ours of treatment with XL177A, p53 and p21 protein levels remained high, but MDM2 protein levels matched DMSO control (Fig.2d).XL177A binds USP7 in the XL188 binding pocketXL188 binds the S4-S5 pocket of USP7 (Fig.1a), and we hypothesized that XL177A was still binding this pocket. Unfortunately, extensive efforts to crystallize the USP7-XL177A complex for structure determination by X-ray were unsuccessful. We thus investigated binding mode using structure-activity-relationship (SAR) studies, USP7 mutant enzyme studies, hydrogen-deuterium exchange mass spectrometry (HDX), and molecular dynamics (MD) simulations. The 4-hydroxy-piperidine group of XL188 forms hydrogen-bonding interactions with the sidechain carboxylic group of USP7 Q297 and the peptide backbone of V296 and is required for USP7 inhibition: XL024 (IC50鈥?鈥? 渭M), which has a hydrogen atom instead of this hydroxyl group, is ~1,000-fold less potent than XL112 (a racemic mixture of XL177A and XL177B, IC50鈥?鈥?.0059鈥壩糓) and is ~10-fold less potent than XL058 (IC50鈥?鈥?.904 渭M), which lacks the Cys223-reactive chloro atom (Fig.3a and TableS1). In addition, two XL188-resistant USP7 mutants, F291N and Q351S27,29, are inhibited by XL177A with 100-fold loss in potency compared to wild-type enzyme (Figs.3b and S5). To gain more direct structural information about the interaction of XL177A with USP7, we performed HDX to monitor changes in protein dynamics31. Both XL188 and XL177A strongly protected the BL1 and 伪-4/5 loops surrounding the S4-S5 pocket from exchange (Fig.3c,d and S6). Consistent with the benzyl moiety of XL177A being buried in the S4 pocket, an XL177A analog lacking the benzyl group (XL041) does not protect 尾-sheet residues 410鈥?23 that engage the benzyl group on XL188, which may explain the 100-fold loss of potency for XL041 compared to XL177A (Fig.3a,e, and S6). While the regions of protection for XL188 and XL177A are similar we observed enhanced exchange in the region from 伪2 to 伪4 of USP7, putatively due to changes in dynamics resulting from covalent bond formation. These helices likely became destabilized upon XL177A binding, leading to exposure of the backbone hydrogens. We note that this XL177A-induced change in USP7 conformation may help explain why crystallization efforts with this compound were unsuccessful, with the caveat that differences between solution and solid-phase dynamics may complicate this picture. MD simulations of XL177A binding with covalent labeling of C223 and H-bonding to Q297 taken as priors showed a decrease in water contacts of the 伪4/5 and S4-S5 residues as well as an increase in water contacts for the region from 伪2 to 伪4 when compared to the unbound USP7 construct, consistent with HDX experimental data. Additionally, final snapshots after 150鈥塶s production stages show a binding mode similar to that of the XL188-USP7 construct (Fig.S7). Taken together, these data suggest that XL177A possesses similar binding sites and a similar binding mode to that of XL188 but induces additional conformation changes in protein dynamics compared to XL188. Similar effects were observed when full length USP7 was bound to XL177A (data not shown).Figure 3XL177A has a similar USP7 binding mode to XL188. (a) Structure-activity-relationship of XL177A showing the potency of the indicated compounds against full-length USP7 in Ub-AMC assays (n鈥?鈥? or 3 experimental replicates). (b) Comparative Ub-AMC IC50 curves of USP7-WT, USP7-Q351S, and USP7-F291N after 6-hr pre-incubation with XL177A (n鈥?鈥? experimental replicates, error bars鈥?鈥?5% CI). (c鈥揺) Structure of USP7 CD highlighting regions with increased (green) or decreased (purple) hydrogen exchange after treatment with XL188 (c), XL041 (d), and XL177A (e) (PDB: 5VS6; n鈥?鈥? experimental replicates).Full size imageXL177A is a highly selective USP7 chemical probeThe selectivity of XL177A was first assessed by determining its inhibitory activity across a panel of 41 recombinant DUBs using in vitro activity assays. At a concentration of 1鈥壜礛 ( 1000-fold higher than its IC50 for USP7), XL177A completely inhibited USP7 enzymatic activity but did not exhibit significant activity against any other DUBs (Fig.4a). The DUB enzymes in this panel primarily consist of only domains or binding partners that are sufficient for in vitro activity, and many DUBs are large multi-domain proteins and/or exist in macromolecular complexes. Furthermore, the standard conditions for this panel include compound pre-incubations of 15鈥塵inutes, limiting our ability to assess off-targets that are inhibited with time-dependent kinetics. We thus utilized competitive ABPP with quantitative MS to explore the selectivity of XL177A in a more native context. Briefly, either DMSO or XL177A was pre-incubated with HEK293 crude cell extract for 5鈥塰ours. The lysate was then incubated with a 1:1 mixture of biotin-ubiquitin-propargylic acid (Bio-Ub-PA) and biotin-ubiquitin-vinyl methyl ester (Bio-Ub-VME), an ABP combination that maximized DUB biotin labeling in our hands (Fig.S8). The labeled lysates were enriched by streptavidin resin, tandem mass tag (TMT)-labeled, combined and analyzed by LC/MS. We found that XL177A significantly blocked USP7 labeling by DUB ABPs in a dose-dependent manner while remaining selective against 59 other DUBs (Fig.4b). Collectively, using state-of-the-art DUB in vitro and activity-based proteomic profiling, we demonstrated that XL177A binds to and inhibits USP7 with 10-fold selectivity over closely related DUBs.Figure 4XL177A is selective for USP7. (a) Remaining activity of 41 purified recombinant DUBs against Ub-Rho after 15-minute pre-treatment with XL177A (n鈥?鈥?). (b) Ratio of Bio-Ub-PA/VME labeling for 60 DUBs in HEK293AD lysate between samples pre-treated for 5鈥塰ours with DMSO v. 1 渭M XL177A (n鈥?鈥? experimental replicates with two technical replicates, reported values are medians of all replicates). (c) Ratio of XL177A-DTB labeling for 566 proteins in HEK293AD lysate between samples pre-treated for 5鈥塰ours with DMSO v. 1 渭M XL177A (n鈥?鈥? experimental replicates with two technical replicates, reported values are medians of all replicates).Full size imageTo define XL177A specificity proteome-wide, we assessed binding partners using an unbiased chemical proteomics screen. First, we synthesized XL177A-DTB, an XL177A analog with a desthiobiotin (DTB) affinity tag, and demonstrated that it retained USP7 inhibitory activity (TableS1). Streptavidin affinity pulldowns of XL177A-DTB followed by on-bead digest and proteomic analysis identified 566 proteins including USP7. Since this approach not only identifies targets of XL177A but also endogenously biotinylated or highly abundant native proteins, we performed a competition experiment with XL177A to identify proteins that were bona fide targets of XL177A. Specifically, we treated HEK293 cell lysates with XL177A for 5鈥塰ours at 1 or 10 渭M, incubated with XL177A-DTB, and quantified concentration-dependent blocking of XL177A-DTB binding throughout the proteome. Of the 566 proteins covalently modified by XL177A-DTB, only USP7 exhibited 3-fold inhibition of labeling when treated with XL177A at 1 or 10 渭M (Fig.4c). These findings corroborate that XL177A is highly specific for USP7 relative to other DUBs and further, the rest of the proteome. The tetrahydroacridine warhead of XL177A is closely related to the biologically quiescent cysteine electrophile 4-chloro-pyridine32, which may help to explain XL177A鈥檚 selectivity.Taken together, these data demonstrate that XL177A possesses cellular permeability, highly potent target engagement with native USP7, and exquisite proteome-wide selectivity for USP7. This compound, in combination with XL177B, meets the criteria of potency, selectivity, and cellular activity set out by multiple organizations to describe well-characterized chemical probes33,34. We thus sought to use XL177A/B to gain a deeper understanding of the role of USP7 in cancer using unbiased profiling experiments.Transcriptional profiling of XL177A reveals strong enrichment of p53 target genesUSP7 directly and indirectly regulates protein levels and subcellular localization of multiple transcription factors and epigenetic modifiers including p5335, N-Myc21, FOXO420, DNMT136, EPOP37, and PRC138. In order to systematically evaluate the cellular effects of USP7 inhibition in an unbiased manner, we first evaluated the transcriptome-wide effects of XL177A and XL177B in MCF7 cells. We tested the MDM2/p53 interaction inhibitor Nutlin-3A in parallel to benchmark USP7 inhibition against a validated MDM2 inhibitor. To this end, we treated MCF7 cells for 24鈥塰ours with low or high doses of XL177A, XL177B, or Nutlin-3A, then used high-throughput 3鈥?Digital Gene Expression (DGE) RNA-seq to analyze the transcriptome-wide effects of these compounds39,40. Overall, we detected transcripts for 7,000鈥?0,000 genes (7276 genes detected across all conditions). Interestingly, hierarchical clustering revealed that XL177A and Nutlin-3A transcriptional profiles were correlated at both low (0.1 渭M XL177A and 1 渭M Nutlin-3A) and high (1 渭M XL177A and 10 渭M Nutlin-3A) doses, while XL177B clustered with the vehicle (DMSO)-treated control cells (Fig.5a). While XL177A did upregulate a set of genes not affected by Nutlin-3A, the overall transcriptional profile of these compounds led us to systematically evaluate whether they were enriched for similar gene sets.Figure 5XL177A activates p53 signaling transcriptome-wide. (a) Heat map showing hierarchical clustering of transcriptional profiles from MCF7 cells treated for 24鈥塰ours with 0.1 or 1 渭M XL177A, 0.1 or 1 渭M XL177B, or 1 or 10 渭M Nutlin-3A. (b) Volcano plots of genes enriched or depleted after 24-hr treatment with 1 渭M XL177A. (c) Expression levels of all genes (clear) or direct p53 target genes (filled) from MCF7 cells treated for 24鈥塰ours with 1 渭M XL177A (adjusted p鈥?lt;鈥?.0001), Nutlin-3A (adjusted p鈥?lt;鈥?.0001), or XL177B (ns). Box: inter-quartile range; Whiskers: gens within 3鈥塖D of mean. (d) Top 5 datasets enriched in the set of genes significantly up- or down-regulated by 1 渭M XL177A treatment and their ranking in the set of genes up- or down-regulated by Nutlin-3A. (e,f) Waterfall plot of all detected transcripts from 1 渭M XL177A ranked by normalized expression score and compared to DREAM complex targets (e) or direct p53 targets (f).Full size imageTo identify specific gene sets that were enriched under different treatment conditions, we computed the overlap between significantly up- or down-regulated genes and the gene sets in the Molecular Signatures Database41. A recent meta-analysis of p53 target genes had been generated in part by RNA-seq data from Nutlin-3A-treated MCF7 cells16, and we found that the upregulated genes from both low- and high-dose Nutlin-3A were enriched for the direct p53 target genes from this meta-analysis. The same report had also identified a set of dimerization partner, Retinoblastoma-like, E2F and multi-vulval class B (DREAM) complex target genes by combining p53 repression data (generated in part with Nutlin-3A-treated MCF7 cells) with cell cycle expression and DREAM component binding. Again, we found that this DREAM target geneset was most strongly enriched in the downregulated genes from both Nutlin-3A treatments. Interestingly, we also found that both low- and high-dose XL177A strongly upregulated the p53 targets (most significant for both doses) and strongly downregulated the DREAM complex targets (most significant for high dose and second most significant for low dose) (Fig.5b鈥揻).USP7 inhibition by XL177A thus elicits a transcriptional signature that strongly phenocopies that of direct MDM2 inhibition by Nutlin-3A, suggesting that p53 upregulation may be the most relevant consequence of USP7 inhibition in TP53-WT MCF7 cells. These results were somewhat surprising, as we expected to identify multiple gene signatures relevant to other USP7 targets. We considered that XL177A may produce different effects in TP53-WT versus TP53-mutant cells and thus sought to profile XL177A in a panel of cancer cell lines with diverse genomic backgrounds.Selective USP7 inhibition is correlated to TP53 mutational status in a diverse panel of cell linesWe screened XL177A, XL177B, and Nutlin-3A in 8-point dose response against a panel of 484 cancer cell lines using barcoded cancer cell lines/PRISM technology42. Briefly, cell lines in average pool size of 25 were treated with inhibitor for 5 days, then the relative abundance of each cell line was determined by high-throughput sequencing. Calculated area under curve (AUC) for each cell line was determined and utilized to determine sensitivity. Correlation of viability profiles between XL177A and XL177B was only observed at the two highest doses of XL177B, which is consistent with an on-target effect overall (Fig.S9). The viability profile observed for XL177A was compared to the viability profiles in the genetic knockout profiles from the DepMap cancer dependency map (Fig.6a)43. Sensitivity to XL177A correlated most highly with USP7 KO, supporting the conclusion that XL177A鈥檚 effects are on-target. Interestingly, XL177A sensitivity was highly correlated with MDM2 and UBE2D3 (an E2 conjugating enzyme for p53) KO and highly anti-correlated with TP53 KO, implying that p53 is a key predictor of response to USP7 inhibition across the entire panel of cells. In addition to genetic modulation, features that correlate with viability response to XL177A were interrogated using publicly available gene expression, copy number, and mutation profiles from the DepMap portal. Notably, a mutation in TP53 predicted decreased sensitivity to XL177A (鈭嗏€?鈥?.0512, p鈥?鈥?.0014) (Fig.6b). The strength of association was similar to that of Nutlin-3A (鈭嗏€?鈥?.0919, p鈥?lt;鈥?.0001) (Fig.S10). No other mutations were associated with the sensitivity profile of XL177A, and TP53 mutations were not significantly associated with a differential response to XL177B. Taken together, these data suggest that the mechanism of action of XL177A is USP7 inhibition and that sensitivity to USP7 inhibition is highly associated with TP53 mutational status.Figure 6XL177A acts through a p53-dependent mechanism. (a) Correlation between relative proliferation profiles of 484 cancer cell lines after XL177A treatment for 5 days or CRISPR KO with the indicated sgRNA. (b) Depmap was queried to determine whether mutations in any of 18,000 genes correlated with XL177A AUC. Only TP53 mutant cell lines displayed significantly altered response relative to WT. (c) Lineages enriched as sensitive (left of 0) or resistant (right of 0) to XL177A in the PRISM data set. (d) Relative Cell Titer Glo luminescence of a panel of TP53-WT (red) or TP53-mutant (black) Ewing Sarcoma cell lines after treatment with XL177A for 3 days. (e) Relative Cell Titer Glo luminescence of TC32 cells expressing the indicated sgRNA after treatment with XL177A for 3 days. (f) Correlation between USP7 DEMETER and AVANA dependency scores in the Broad Depmap portal for TP53-WT (red) and TP53-mutant (black) cell lines. (g) Log10 ratio of A549-FF to A549-sgTP53-Renilla cells after treatment of an initial 1:1 mixture of the two cell lines with the indicated sgRNA for the indicated number of days.Full size imageSelective USP7 inhibition exerts p53-dependent growth suppression in individual cell linesTo validate our results from the unbiased screen, we tested effects of XL177A, XL177B and Nutlin-3A using individual cell lines. We chose to prioritize studies in Ewing Sarcoma30, acute myeloid leukemia (AML)26, and multiple myeloma (MM)24 because these lineages have been evaluated with previous generations of USP7 inhibitors. As rhabdoid tumors were identified as the lineage most sensitive to both XL177A and Nutlin-3A in the cell profiling experiment we performed (Fig.6c), we also included rhabdoid cell lines in our testing. We assessed the effect of XL177A in panels of these four cell lineages and found that TP53 functional status predicted response to USP7 inhibition: 9 out of 10 TP53-WT cell lines were XL177A-sensitive, compared to only 1 out of 17 TP53-mutant cell lines (Figs.6d and S11). XL177B was ~100-fold less potent in the cell lines tested, consistent with the effects being USP7-dependent (Figs.S11 and S12).We had previously reported a series of stable TP53-KO TC32 cell lines30, and given the strong growth suppression of TC32 by XL177A, we assessed its effect in the matched TP53-KO cell lines. Interestingly, we found that TP53 KO completely rescued TC32 growth except at high doses, indicating that this effect is p53-dependent (Fig.6e). Meanwhile, GNE-6640, a USP7 inhibitor that does not exhibit global p53-dependent effects26, showed little selectivity for TP53-WT Ewing Sarcoma lines or TP53-WT TC32 cells versus TP53-mutant or TP53-KO cells (Fig.S12), a result that parallels our previous findings with the USP7 inhibitor P509130. These results suggest that, while the globally selective USP7 inhibitor XL177A acts through a p53-dependent mechanism, P5091 and GNE-6640 may have as yet unknown targets that lead to p53-independent effects.Genetic studies validate TP53 status as a predictor of response to USP7Given the highly p53-dependent nature of response to XL177A, we sought to identify whether genetic modulation of USP7 also produces p53-dependent effects in cancer cell lines. We first queried the DepMap portal43, which houses genome-scale loss-of-function datasets interrogating gene essentiality across hundreds of genomically characterized cancer cell lines. Both RNAi (DEMETER2, n鈥?鈥?11) and CRISPR-Cas9 (CERES, n鈥?鈥?36) targeting USP7 produced datasets that were strongly correlated with a number of p53-network genes, including MDM2 (RNAi R2鈥?鈥?.598; CRISPR R2鈥?鈥?.575), MDM4 (DEMETER2 R2鈥?鈥?.522; CERES R2鈥?鈥?.356) and PPM1D (DEMETER2 R2鈥?鈥?.346; CERES R2鈥?鈥?.515), TP53 (DEMETER2 R2鈥?鈥夆垝0.534; CERES R2鈥?鈥夆垝0.496), and CHEK2 (DEMETER2 R2鈥?鈥夆垝0.372; CERES R2鈥?鈥夆垝0.406) (TableS3). In order to specifically assess the role of p53 in response to genetic modulation of USP7, we annotated individual cell lines as TP53-functional-WT or functional-mutant using the method in Giacomelli et al44. and performed two-way T-tests comparing the WT and mutant populations. In both the DEMETER2 (mean diff.=0.344鈥壜扁€?.027, p鈥?lt;鈥?.0001, n鈥?鈥?84) and CERES (mean diff.=鈭?.273鈥壜扁€?.025, p鈥?lt;鈥?.0001, n鈥?鈥?13), TP53-functional status was significantly correlated to response to USP7 genetic modulation (Fig.6f). Overall, available genetic data confirm that response to USP7 modulation is correlated with TP53 status across multiple cancer cell lines.To further assess the functional role of p53 on USP7 modulation in TP53-WT cell lines, we employed a dual luciferase reporter assay to assess the comparative fitness of TP53-WT and TP53-KO cells in response to USP7 KO44. Briefly, we expressed Firefly (FF) luciferase in parental A549 and RKO cells and Renilla luciferase in stable TP53-KO A549 and RKO cells. The FF- and Renilla-expressing cells were mixed in a 1:1 ratio and then exposed to Cas9 and sgRNAs targeting MDM2, USP7, TP53, CDKN1A, LacZ, or FF luciferase for 17 days. Both sgMDM2 and sgUSP7 led to sustained reductions in the parental cells of both A549 and RKO (Figs.6G and S13), indicating that TP53-KO improves the fitness of these cells in response to USP7 or MDM2 modulation. Thus, the cell killing effect of USP7 KO is, as with MDM2 KO and XL177A treatment, at least partially mediated by p53 in TP53-WT cells.DiscussionDUBs are an emerging class of potential drug targets. Multiple lines of evidence have led to the classification of numerous DUBs as drivers of disease across diverse therapeutic areas. In oncology, many DUBs have been classified as oncogenic based on their regulatory role on substrate proteins involved in tumorigenesis45. Until recently, the excitement originating from these studies has been counterbalanced by uncertainty regarding the targetability of DUBs. However, recent studies describing development of highly potent and selective inhibitors of DUBs, including USP7 and CSN546, have led to a turning point in the field, re-energizing interest in DUBs as tractable targets for drug discovery. The prevailing current approach to DUB inhibitor development is based on designing non-covalent, allosteric inhibitors. Here, we approach DUB inhibitor development from a different angle and report the first highly potent (sub-nanomolar) and selective (proteome-wide) irreversible DUB inhibitor. The compound, XL177A, inhibits USP7 with an IC50 of 0.34鈥塶M and binds and inhibits native USP7 with potencies in the low nanomolar range. Notably, proteome-wide chemoproteomic experiments revealed exquisite selectivity for USP7, with no proteins other than USP7 being substantially labeled by XL177A. Consistent with this high degree of selectivity, several lines of evidence strongly support the notion that phenotypes observed with XL177A treatment are on-target. First, we paired XL177A with its inactive enantiomer (XL177B), in a wide range of signaling and proliferation experiments and observe loss of phenotype with XL177B. Moreover, comparison of the sensitivity profile of XL177A across over 400 cancer cell lines with sensitivity profiles generated in those same cell lines with KD / KO of over 18,000 genes identified USP7 KO as exhibiting the most similar profile. The combination of potency, proteome-wide selectivity, and evidence of on-target mechanism across large panels of cancer cell lines nominate XL177A as a best-in-class cellular probe of USP7. We note that XL177A is not an in vivo probe, and we thus have a limited understanding of its selectivity, potency, and potential toxicity in more complex animal models. Nonetheless, the cellular properties of XL177A make it a highly valuable tool for the research community in unraveling USP7 biology and addressing fundamental questions surrounding DUB function more broadly.Using XL177A, paired with XL177B, we focused our investigations on USP7 substrates or cellular features that predict growth suppressive effects in cancer in response to pharmacological inhibition of USP7. Overwhelmingly, our studies revealed mutational status of the tumor suppressor p53 to be the key cellular feature predicting sensitivity to USP7 by XL177A. Out of more than 18,000 genes analyzed, only mutations in TP53 significantly correlated with response to XL177A across a panel of more than 400 cell lines. Tested individually, 9 out of 10 TP53-WT cell lines were XL177A-sensitive, compared to only 1 out of 17 TP53-mutant cell lines. In addition, TP53 KO rendered TP53-WT cells resistant to XL177A, and transcriptome-wide response to XL177A phenocopied effects of Nutlin-3A, a validated inhibitor of MDM2-mediated p53 ubiquitination and therefore a compound expected to stabilize p53 and inhibit its proteasomal degradation.TP53 is one of the most commonly mutated genes in cancer. Due to its role in regulating cell cycle arrest and apoptosis, its suppression is a critical node in cancer progression, and p53 re-activation is a key therapeutic strategy across all cancer types. Despite the high TP53 mutation rate in cancer, several cancer types have a low rate of TP53 mutation, especially hematopoietic and pediatric malignancies. For these TP53-WT cancers, p53 activation can be achieved using direct p53 stabilizers such as MDM2 inhibitors, MDMX inhibitors, and Wip1 inhibitors. These p53 activating agents act synergistically with each other and with DNA-damaging radiation and chemotherapy in TP53-WT tumors, leading to clinical trials that combine p53 activators and cytotoxic drugs in multiple myeloma, acute myeloid leukemia, and B cell lymphoma. However, clinical trials with MDM2 inhibitors have already revealed acquired resistance, and additional pathways that can be modulated to activate p53 are of great interest for the treatment of TP53-WT tumors. We previously reported that the USP7 inhibitor XL188 acted as a p53 activator in TP53-WT Ewing sarcoma, but studies of USP7 inhibition in other contexts have typically focused on p53-independent mechanisms. Here, we establish that the p53-dependent activity of USP7 inhibition is of great importance for growth suppression of cancer cells across all lineages. We propose that USP7 inhibitors are p53 activators that should be explored clinically as single agents or in combination studies in hematopoietic and pediatric malignancies with a low rate of TP53 mutation. While USP7 inhibition may stabilize p53 by inducing MDM2 degradation, we note that the p53-dependent effects of USP7 inhibition may not be solely driven by MDM2. Previous studies have shown that USP7 inhibition or knockdown also leads to decreased DNA damage tolerance via degradation of RAD18 (replication-associated repair)47, CSB (nucleotide excision repair)48, and ALKBH2/3 (alkylation repair)49, and USP7 has been shown to regulate both p14/Arf (which mediates p53 stability) and Tip60 (which mediates p53 transcriptional activity)50,51. Further insights into these USP7 substrates may furnish a better understanding of the molecular mechanism of the p53 dependence we observe for our inhibitors.In conclusion, here we report the first highly potent (sub-nanomolar) and selective (proteome-wide) irreversible DUB inhibitor. Our results demonstrate that covalent targeting of the active site of DUBs is a viable strategy for probe and drug discovery and further, support exploration of a target-class approach to rapidly discover covalent hits, leads and probes for multiple DUBs. Moreover, using the newly developed selective USP7 inhibitor we now provide an answer to a long-standing question surrounding the USP7/p53 regulator axis and demonstrate that TP53 status determines the response to selective USP7 inhibition, with TP53-WT cell lines found to be sensitive. Together, the probe we developed and the insights into the p53 dependency of its activity provide valuable information for further advancement of both basic biology and clinical translation of USP7.MethodsAntibodies, cell lines, and reagentsHDM2 (sc-965) antibody was obtained from Santa Cruz. p53 (9282s), p21 (2947s), GAPDH (2118s), and USP7 (4833s) antibodies were obtained from Cell Signaling Technology. Ub-AMC (U-550) and HA-Ub-VS (U-212) were obtained from Boston Biochem. Bio-Ub-PA (UbiQ-076) and Bio-Ub-VME (UbiQ-054) were obtained from UbiQ Bio. BAX (Hs00180269_m1), CDKN1A (Hs00355782_m1), DDB2 (Hs03044953_m1), GADD45A (Hs00169255_m1), GAPDH (402869), MDM2 (Hs00540450_s1), and TP53 (Hs01034249_m1) Taqman probes were obtained from Thermo-Fisher. MCF7 cells were a generous gift from Jean Zhao鈥檚 laboratory, and NCI-H1975 cells were a generous gift from Pasi J盲nne鈥檚 laboratory. HEK 293AD, G401, G402, and MESSA were purchased from ATCC. Stable TP53 knockout TC32 and A549 cells were previously described in Stolte et al. and Giacomelli et al., respectively30,44.USP7 Cloning, expression, and purificationThe constructs encoding USP7 full length (amino acids 1鈥?102) and catalytic domain (208鈥?60) used were cloned as described29. Both constructs were overexpressed in E. coli BL21 (DE3). Cells were grown at 37鈥壜癈 to an OD of 0.9, cooled to 16鈥壜癈, induced with 500鈥壩糓 isopropyl-1-thio-D-galactopyranoside (IPTG), incubated overnight at 16鈥壜癈, collected by centrifugation, and stored at 鈭?0鈥壜癈. Cell pellets were sonicated in lysis buffer (25鈥塵M Tris pH 8, 1鈥塎 NaCl, and 10鈥塵M BME) supplemented with 10鈥壩糶/ml phenylmethanesulfonylfluoride (PMSF) and the resulting lysate was centrifuged at 30,000 xg for 40鈥塵in. Ni-NTA beads (Qiagen) were mixed with lysate supernatant for 2鈥塰ours, and washed with lysis buffer supplemented with 25鈥塵M imidazole. The bound protein was eluted with lysis buffer supplemented with 300鈥塵M imidazole. The sample was then concentrated to 1鈥塵l (30鈥塳Da concentrator; Amicon Ultra, Millipore), and run on a Superdex 200 (GE healthcare) in buffer containing 25鈥塵M HEPES pH 7.5, 200鈥塵M NaCl, and 1鈥塵M DTT. Fractions were pooled, concentrated and frozen at 鈭?0鈥壜癈. The construct encoding the catalytic domain of USP7 C223A was generated by site-directed mutagenesis, and the protein was purified as described for the WT enzyme.Ub-AMC Enzymatic and kinetic assaysFull length USP7 was tested for its activity in Ubiquitin-AMC assay in presence or absence of inhibitors. USP7 (5鈥塶M) was pre-incubated for 6鈥塰ours at room temperature with different concentrations of inhibitors or DMSO as a control in 50鈥塵M HEPES pH 7.5, 0.5鈥塵M EDTA, 11 uM ovalbumin, and 5鈥塵M DTT. Ubiquitin-AMC (Boston Biochem) was then added to a final concentration of 500鈥塶M. The initial rate of the reaction was measured by collecting fluorescence data at one minute intervals over 30-minute period using a Clariostar fluorescence plate reader at excitation and emission wavelength of 345 and 445鈥塶m respectively. The calculated initial rate values were plotted against inhibitor concentrations to determine IC50s. All the experimental data were plotted using Prism GraphPad. All assays for each compound were performed at least twice for each compound.To calculate the ki and kianct values for XL-9678-041, the procedure above was used, but different concentrations of inhibitor were incubated with USP7 for different time points (5鈥塵in - 3鈥塰ours) before adding Ubiquitin-AMC. To determine kobs, the time course curves were fit to the equation y鈥夆€?鈥夆€墆max(1鈥夆垝鈥塭xp(鈭?i>kobs鈰?/span>x)). The kobs values were then plotted against the inhibitor concentrations and fit to the equation y鈥?鈥?i>kinact/(1鈥?鈥?ki/x)) to obtain the values for ki and kianct.MS LabelingPurified USP7 catalytic domain was diluted to 20 渭M in 10 渭L labeling buffer (20鈥塵M HEPES pH 7.5, 150鈥塵M NaCl, 1鈥塵M TCEP) and incubated for the indicated times with 50 渭M (2.5脳) compound. After incubation, samples were flash frozen in liquid nitrogen and stored at 鈭?0鈥壜癈 until analysis.Intact MS AnalysisIntact mass analysis was performed by injecting 5鈥壜礸 USP7 catalytic domain onto a self-packed reversed phase column (1/32鈥?O.D.鈥壝椻€?00 um I.D., 5鈥塩m of POROS 10R2 resin). After desalting, protein was eluted with an HPLC gradient (0鈥?00% B in 4鈥塵inutes, A鈥?鈥?.2鈥塎 acetic acid in water, B鈥?鈥?.2鈥塎 acetic acid in acetonitrile, flow rate ~30鈥壜礚/min) into an LTQ ion trap mass spectrometer (ThermoFisher Scientific, San Jose, CA). Mass spectra were processed using MagTran1.03b252.CE-MS AnalysisTo identify sites of covalent modification, treated protein was reduced (10鈥塵M dithiothreitol), alkylated (22.5鈥塵M iodoacetamide), and digested with trypsin overnight at 37鈥壜癈. Peptides were desalted using SP353, dried by vacuum centrifugation, and reconstituted in 1% formic acid/50% acetonitrile with 100鈥塵M ammonium acetate. Peptides were then analyzed by CE-MS using a ZipChip CE system and autosampler (908 Devices, Boston, MA) interfaced to a QExactive HF mass spectrometer (ThermoFisher Scientific, San Jose, CA). Peptide solution was loaded for 30鈥塻econds, and the mass spectrometer was operated in data dependent mode and subjected the 5 most abundant ions in each MS scan (60k resolution, 3E6 target, lock mass enabled) to MS/MS (15k resolution, 1E5 target, 100鈥塵s max inject time). Dynamic exclusion was enabled with a repeat count of 1 and an exclusion time of 6鈥塻econds. MS/MS data was extracted to.mgf using mulitplierz scripts54,55 and searched against a forward-reverse human NCBI refseq database using Mascot version 2.6. Search parameters specified fixed carbamidomethylation of cysteine, and variable oxidation (methionine) and XL177A modification (cysteine). Precursor mass tolerance was set to 10 ppm and product ion tolerance was 25 mmu. Spectral validation was performed using mzStudio56.Cell cultureMCF7 cells were cultured in RPMI-1640 growth medium supplemented with 10% FBS. HEK293AD cells were cultured in DMEM鈥?鈥?0% FBS鈥?鈥?% penicillin-streptomycin. A673 cells were cultured in DMEM鈥?鈥?0% FBS鈥?鈥?鈥塵M sodium pyruvate鈥?鈥?%PSQ. TC32 and TC71 cells were cultured in RPMI鈥?鈥?0% FBS鈥?鈥?% PSQ. TC138 and CHLA258 cells were cultured in IMDM鈥?鈥?0% Fetal Bovine Serum鈥?鈥?mM L-Glutamine鈥?鈥?X ITS (5鈥壜礸/mL insulin, 5鈥壜礸/mL transferrin, 5鈥塶g/mL selenous acid). G401, G402, and MES-SA cells were cultured in McCoy鈥檚 Modified Media (ATCC)鈥?鈥?0% FBS. SKNO-1-luc鈥?鈥? SKM-1, NB4-luc鈥?鈥? HL60, HEL, MV4,11, OCI-AML3, and MOLM14 cells were cultured in RPMI鈥?鈥?0% FBS鈥?鈥?% L-glutamine鈥?鈥?% penicillin-streptomycin. All cell lines were maintained in 10 cm2 tissue-culture treated dishes 37鈥壜癈 in a 5% CO2 incubator. All cell lines were verified Mycoplasma-free by the MycoAlert test kit.HA-Ub-VS LabelingHA-Ub-VS experiments were performed as previously described in Lamberto et al29. Briefly, target engagement lysis buffer (50鈥塵M Tris pH 8.0, 150鈥塵M NaCl, 5鈥塵M MgCl2, 0.5鈥塵M EDTA, 0.5% NP-40, 10% glycerol, 1鈥塵M TCEP, protease and phosphatase inhibitors) was added to cell pellets on ice. Lysate was cleared by centrifugation and diluted to 1.67鈥塵g/mL. Where indicated, 30 渭L lysate was then incubated with inhibitors or DMSO for the indicated timepoints. 1 渭M HA-Ub-VS was then added to the lysate and incubated at RT for the indicated time points. Labeling reactions were quenched with 4x LDS sample buffer (Thermo Fisher B0007) supplemented with 10% BME, vortexed vigorously, and heated to 95鈥塁 for 5鈥塵inutes. Samples were resolved by SDS-PAGE and analyzed by Western blot with the indicated antibodies.Real-Time PCRAfter cell treatment under the indicated conditions, total cellular RNA was purified using a Qiagen RNEasy kit. 1 渭g of RNA was then converted to cDNA using SuperScript III First-Strand Synthesis (Invitrogen). cDNA from each sample was then combined with the indicated TaqMan probe and 2x MasterMix in a 96-well Fast RT-PCR plate (Invitrogen). qPCR was performed on an Invitrogen 7500 Fast qPCR instrument and gene expression was calculated using the 2-鈭嗏垎Ct method on Graphpad Prism.Cell cycle analysisFor propidium iodide (PI) staining, treated cells (~1 million per condition) were washed with cold PBS, then fixed in 80% ethanol overnight at 鈭?0鈥壜癈. After fixing, cells were pelleted, washed with PBS, and reconstituted in 500 渭L FxCycle PI / RNAse A staining solution (Thermo Fisher). Cells were stored overnight at 4鈥壜癈 and analyzed using a BD Fortessa flow cytometer.Hydrogen deuterium exchangeExperiments were performed as previously described29. Briefly, USP7 (50鈥壜礛 in 20鈥塵M Hepes (pH 7.5), 200鈥塵M NaCl, 1鈥塵M TCEP, 5% glycerol H2O) was pre-incubated at room temperature (21鈥壜癈) with the indicated compound in a 10:1 compound:USP7 ratio (XL041 for 60鈥塵in, XL177A for 30鈥塵in), then diluted with 15-fold D2O buffer (pD 7.5) at room temperature. At each deuterium exchange time point (from 10鈥塻 to 4鈥塰ours) an aliquot from the exchange reaction was removed and acidified (pH to 2.5) with an equal volume of quench buffer (0.8% formic acid and 0.8鈥塎 guanidine hydrochloride, H2O) to quench the reaction. Quenched samples were immediately injected into the LC/MS system. Samples were digested online using a Poroszyme immobilized pepsin cartridge (2.1鈥塵m鈥壝椻€?0鈥塵m, Applied Biosystems) at 15鈥壜癈 for 30鈥塻, then injected into a custom Waters nanoACQUITY UPLC HDX Manager and analyzed on a XEVO G2 mass spectrometer (Waters Corp., USA). Peptides were compared to highly deuterated peptide standards to determine the average amount of back-exchange (~20鈥?0%). Deuterium levels were not corrected for back-exchange and are therefore reported as relative31. All experiments were performed in duplicate, and the error of measuring the mass of each peptide averaged鈥壜扁€?.15鈥塂a. The HDX MS data were processed using PLGS 3.0 and DynamX 3.0 (Waters Corp., USA). Common peptides that were compared between the USP7 catalytic domain alone and bound to the compounds led to a sequence coverage of 85.4% with 81 peptic peptides for XL-041 and 94% with 105 peptic peptides for XL-177A (Supplemental Fig.S7A,B).Molecular dynamics simulationsThe models of the XL188-bound USP7 complex were constructed from the co-crystalized PDB crystal structure (PDBID: 5VS6). The model of the XL177A- and XL177B-bound USP7 complexes were constructed from the apo and Ub-Aldehyde bound USP7 crystal structures (PDBIDs: 1NB8 and 1NBF, respectively). With the preparation and simulation protocols tested in our prior studies57,58,59, all protein structures were prepared using Protein Preparation Wizard, solvated in SPC water with sodium counter ions by the System Builder, and simulated with the NPT ensemble (300鈥塊, 1鈥塧tm, Martyna-Tuckerman-Klein coupling scheme) using the OPLS3e force field60 in Desmond v5.461. Each construct underwent minimization, equilibration, and 150-ns production stages with a time step of 2鈥塮s. The Ewald technique with a 9鈥壝?cutoff was used for the Van der Waals and other electrostatic calculations. Hydrogen atoms were constrained using the SHAKE algorithm. At least two replicas were collected for each construct. Trajectory visualization and data analysis was done with the VMD program62 and in-house TCL and Python scripts.In Vitro DUB ProfilingCompounds were screened using the Ubiquigent Drug Profiler SPT system (www.ubiquigent.com/drug-discovery-services/dubprofiler/). Each of 41 purified DUBs was incubated with compound for 15鈥塵inutes, then ubiquitin rhodamine 110 (Ub-Rho) was added and percent inhibition determined based on fluorescence relative to a DMSO control.In Situ DUB ProfilingDUB profiling was performed using conditions similar to those in Lawson et al63. HEK 293AD cells were lysed using target engagement lysis buffer (50鈥塵M Tris pH 8.0, 150鈥塵M NaCl, 5鈥塵M MgCl2, 0.5鈥塵M EDTA, 0.5% NP-40, 10% glycerol, 1鈥塵M TCEP, protease and phosphatase inhibitors), and the lysate was cleared by centrifugation. Samples were diluted to 2鈥塵g/mL, and 1鈥塵L lysate was incubated with the indicated concentration of XL177A for 5鈥塰ours at RT. Excess inhibitor was removed using a 30鈥塊 Amicon spin filter, then the resulting lysate was incubated with 1 渭M each of Biotin-Ub-PA and Biotin-Ub-VME for 90鈥塵inutes at RT. SDS was added to a final concentration of 1.2%, and samples were heated to 80鈥壜癈 for 5鈥塵inutes. After cooling to RT, 1X PBS was added to dilute the final SDS concentration to 0.2%. 100 渭L streptavidin agarose slurry was added to each sample, followed by incubation at RT for 3鈥塰ours. After streptavidin enrichment, samples were washed vigorously (2脳0.2% SDS, 3x PBS, 3x ddH2O). After the final wash, all supernatant was removed using a flat-bottom tip, and the resin was flash frozen and stored at 鈭?0鈥壜癈 until workup for TMT labeling.XL177A-DTB ProfilingHEK 293AD cells were lysed as described above, and the lysate was cleared by centrifugation. Samples were diluted to 10鈥塵g/mL, and 200 渭L lysate (2鈥塵g protein total) was incubated with the indicated concentrations of XL177A for 4鈥塰ours at RT, then 2 渭M of WH114A for 4 additional hours. SDS was added to a final concentration of 1.2% (27.2 渭L of a 10% stock), and denatured by heating to 80鈥壜癈 for 5鈥塵inutes. After cooling to RT, 1125 渭L 1X PBS was added to dilute the final SDS concentration to 0.2%. 50 渭L streptavidin agarose slurry was added to each sample, followed by incubation at RT for 3鈥塰ours. After streptavidin enrichment, samples were washed vigorously (2脳0.2% SDS, 3x PBS, 3x ddH2O). After the final wash, all supernatant was removed using a flat-bottom tip, and the resin was flash frozen and stored at 鈭?0鈥壜癈 until workup for TMT labeling.Sample prep for mass spectrometry analysisStreptavidin beads were resuspended in 100鈥塵M Tris pH 8.0 and bound proteins were denatured with 0.1% rapigest, reduced (10鈥塵M dithiothreitol), alkylated (22.5鈥塵M iodoacetamide), and digested with trypsin overnight at 37鈥壜癈. To remove rapigest, recovered supernatants were acidified with 10% TFA, incubated at 37o C for 45鈥塵inutes, and centrifuged at 14,000鈥塺pm for 15鈥塵inutes at 4o C. Peptides were then desalted by C18, and dried by vacuum centrifugation. Dried peptides were reconstituted in 50鈥塵M pH8.0 TEAB and labeled with TMT reagent at RT for 1鈥塰our. TMT reactions were pooled and treated with hydroxylamine according to the manufacturer鈥檚 instructions. Peptide mixtures were then dried, reconstituted in 50鈥塵M ammonium bicarbonate and desalted by SP353. Eluted peptides were then analyzed by nanoflow LC-MS/MS as described64 with a NanoAcquity UPLC system (Waters, Milford, MA) interfaced to a QExactive HF mass spectrometer (Thermofisher Scientific, San Jose, CA). TMT labeled peptides were injected onto a precolumn (4鈥塩m POROS 10R2, Applied Biosystems, Framingham, MA), resolved on an analytical column (30鈥壜祄 I.D.鈥壝椻€?0鈥塩m packed with 5鈥壜祄 Monitor C18) and introduced to the mass spectrometer by ESI (spray voltage = 3.5鈥塳V, flow rate ~30 nL/min). The mass spectrometer was operated in data dependent mode such that the 15 most abundant ions in each MS scan (m/z 300鈥?000, 120鈥塊 resolution, target鈥?鈥?E6, lock mass for 445.120025 enabled) were subjected to MS/MS (m/z 100鈥?000, 30鈥塊 resolution, target鈥?鈥?E5, max fill time鈥?鈥?00鈥塵s). Dynamic exclusion was selected with a repeat count of 1 and an exclusion time of 30鈥塻econds. MS/MS data was extracted to.mgf using mulitplierz scripts54,55 and searched against a forward-reverse human NCBI refseq database using Mascot version 2.6. Search parameters specified fixed carbamidomethylation of cysteine, fixed N-terminal and lysine TMT labelling, and variable oxidation (methionine). Additional multiplierz scripts were used to filter results to 1% FDR and derive protein-level aggregate reporter ion intensities using peptides mapping uniquely into the genome65.DGE-RNAseqThe DGE RNA sequencing was performed following the published method39,40 with modifications as described below. MCF7 cells were seeded in a 384 well plate at 2500 cells per well, allowed to adhere for 24鈥塰ours, then treated with two doses in triplicate of XL177A, XL177B, Nutlin-3, or DMSO. After 24鈥塰ours of treatment, an EL405x plate washer (BioTek) was used to aspirate the media and wash once with PBS. Cells were lysed in the plate with 10鈥壜祃 of lysis buffer (1x Qiagen TCL, 1% BME) for 5鈥塵inutes at room temperature, then stored at 鈭?0鈥壜癈. Automated liquid handling was performed with assistance from ICCB-Longwood Screening Facility. A BRAVO Automated Liquid Handling Platform (Agilent) was used for RNA extraction as follows. The lysate was mixed and 10鈥壜祃 was transferred to a 384 well PCR plate. 28鈥壜祃 of homemade SPRI66 (solid-phase reversible immobilization) beads were added to the lysate and mixed. After 5鈥塵inutes, the beads were pulled down by placing the plate on a magnet, then the beads were washed twice with 80% ethanol. Beads were air dried for 1鈥塵inute, then 20鈥壜祃 of nuclease free water was added, the plate was removed from the magnet, and the beads were resuspended. After 5鈥塵inutes, the plate was placed back on the magnet to pull down the beads. The supernatant was transferred to a fresh 384 well plate. RNA quantity was checked with the Qubit Fluorometer and RNA quality was assessed using the Agilent BioAnalyzer RNA 6000 Pico Kit. The BRAVO platform was again used to transfer 5鈥壜祃 of supernatant was transferred to a fresh 384 well plate with RT master mix and 1鈥壜祃 of barcoded E3V6NEXT adapters for reverse transcription and template switching. The plate was incubated for 90鈥塵inutes at 42鈥壜癈, then cDNA was pooled, purified with the QIAquick PCR purification kit. The full-length cDNA was treated with Exonuclease I to remove excess primers for 30鈥塵inutes at 37鈥壜癈 then amplified x5 cycles with Advantage 2 PCR Enzyme System using the SINGV6 primer. The amplified full-length cDNA was purified with Agencourt AMPure XP magnetic beads and quantitated by Qubit Fluorometer. The sequencing library was prepared using the Nextera DNA kit following the manufacturer鈥檚 instructions. Briefly, 55鈥塶g of cDNA was tagmented for 5鈥塵inutes at 55鈥壜癈 then purified using Zymo DNA Clean Concetrator-5 column, then amplified x7 cycles and purified using a 0.9x ratio of AMPure XP magnetic beads. The library size distribution was assessed by Agilent BioAnalyzer HS DNA Kit before it was quantified by qPCR and sequenced on an Illumina NextSeq at the Biopolymers Facility at Harvard Medical School.AnalysisThe bcbio-nextgen single cell RNA-seq analysis pipeline (https://bcbio-nextgen.readthedocs.io/en/latest/) was used to deconvolve the well barcodes and convert reads to counts. Any detected barcode that differed by more than one base from an expected barcode was removed. Unique reads were identified using unique molecular identifiers (UMIs) in order to remove PCR duplicates. The reads remaining after these filters were aligned to the transcriptome (GRCh38) using RapMap67. The R package DESeq. 2 version 1.20.0 (R version 3.5.1) was used for differential expression analysis68. The Molecular Signatures Database (v6.2)41,69,70 was used to compute overlap for the significantly upregulated and downregulated genes (adjusted p value 0.05 and fold change 1.5) for XL177A and Nutlin-3 at both doses.Cancer cell line profilingCancer cell line profiling was performed using the Broad鈥檚 PRISM platform. Cell treatment and data analysis were performed as described in Yu et al42. Raw Luminex signal was converted to AUC values for each cell line as described. The resulting AUC profile for each compound was modeled against DepMap using the limma package in R71. Because hotspot mutation analysis was performed against DepMap, TP53 annotations in Fig.6c,d are directly from DepMap. TP53 annotation using the functional scores in Giacomelli et al44. also produced significant differences between the functional WT and functional mutant populations (XL177A: mean diff.鈥?鈥?.072鈥壜扁€?.016, p鈥?lt;鈥?.0001; Nutlin-3: mean diff.鈥?鈥?.165鈥壜扁€?.011, p鈥?lt;鈥?.0001).Cell viabilityCells were plated in 384-well culture-treated plates and allowed to settle overnight. After drug treament and appropriate incubation time, cell viability was assessed using the CellTiter-Glo Luminescent Cell Viability Assay (Promega). Luminescence was read on a Fluostar Omega Reader (BMG Labtech).Computational cancer profilingThe Broad Institute鈥檚 public DepMap portal was accessed via www.depmap.org/portal/interactive. The top 100 gene profile correlations for USP7 CRISPR (Avana) and Combined RNAi (Broad, Novartis, Marcotte) were downloaded directly from the DepMap portal. TP53 profiling was performed offline after downloading the CRISPR (Avana) and Combined RNAi (Broad, Novartis, Marcotte) gene dependency sets. TP53 annotation was performed manually using the methods in Giacomelli et al.44, with cell lines with functional scores 0 annotated as TP53 functional WT and cell lines with functional scores 0 annotated as TP53 functional mutant. Parametric, unpaired, two-tailed t-tests were performed using GraphPad PRISM.Dual reporter luciferase competition assayp53WT and p53NULL A549 cells constitutively expressing firefly luciferase or Renilla luciferase have been described44. Each cell line was infected with lentivirus encoding S. pyrogenes Cas9 under control of the human EF1alpha promoter (pLX311) and selected in blasticidin (InvivoGen) (1鈥塵g/mL) (10 渭g/mL). To perform the competition assay, Cas9-expressing p53WT cells were mixed at a 1:1 ratio with complementarily-labeled Cas9-expressing p53NULL cells and seeded at 2,500 cells/well in 96-well dishes in 200鈥壜礚 of normal culture media. The following day, cells were infected with an array of sgRNA-expressing lentiviruses (pXPR003). Twenty-four hours thereafter, the supernatant was removed and fresh media containing puromycin (InvivoGen) (1 渭g/mL) was added to select for infected cells. Two days later, cells were split into two new replica plates, and incubated for four more days. One replica plate was subjected to a dual luciferase assay72 and luminescence readings were obtained using a Wallac EnVision (Perkin-Elmer). Readings from wells infected with experimental sgRNAs were normalized to wells infected with control sgRNAs, and firefly:Renilla luminescence ratios were calculated to estimate the relative effects of sgRNAs on p53WT versus p53NULL cells within a well. To continue the assay, the second replica plate was passaged at a 1/16 dilution44. The process of reading and re-plating the cells was repeated every 4 days.ChemistryAll commercially available starting materials were purchased from Sigma Aldrich, Fisher Scientific, Oakwood Chemical and Combi Blocks, and anhydrous solvents were purchased from Fisher Scientific. All reagents and solvents were used as received without further purification. If necessary, air or moisture sensitive reactions were carried out under an inert nitrogen atomsphere.Removal of solvents was accomplished on a B眉chi R-300 rotary evaporator connected to a Welch 1400B-01 vacuum line and Labconco FreeZone 6 plus system. Compound purification was achived using normal phase column chromatography (Teledyne CombiFlash chromatography system) and/or reversed phase chromatography (Shimadzu system with SunFire庐 Prep C18 OBDTM 5鈥壩糓 column). Purity was assessed by UPLC (Waters Acquity system) and analytical thin layer chromatography (TLC, EMD Millipore TLC Silica Gel60 F254). TLC visualization was accomplished by irradiation under UV light (254鈥塶m).All 1H-NMR spectra were recorded at 298鈥塊 on a Bruker ARX 500 (500鈥塎Hz) spectrometer, and all 13C-NMR spectra were recorded on a Bruker ARX 500 (126鈥塎Hz) spectrometer. Samples were dissolved in CDCl3, DMSO-d6, or CD3OD, and spectra were referenced to the residual solvent peak (chlorofrom-d: 7.26 ppm for 1H-NMR and 77.16 ppm for 13C-NMR; DMSO-d6: 2.50 ppm for 1H-NMR and 39.25 ppm for 13C-NMR, CD3OD: 3.31 ppm (-Me) for 1H NMR and 49.00 ppm for 13C NMR or tetramethylsilane (TMS) as the internal standard. NMR peaks are reported below with chemical shift, multiplicity (s鈥?鈥塻inglet, d鈥?鈥塪oublet, t鈥?鈥塼riplet, q鈥?鈥塹uartet, m鈥?鈥塵ultiplet, br鈥?鈥塨road peak), coupling constants (Hz), and number of protons. Mass spectrometry (LCMS) data were obtained on Waters Acquity UPLC system in positive ESI mode.For synthetic routes and analysis see Supplementary Methods and Figs.S15 and S16. All data generated during this study are included in this article (and its Supplementary Files) or are available from the corresponding author upon request. A Life Sciences Reporting Summary for this paper is available. References1.Nandi, D., Tahiliani, P., Kumar, A. Chandu, D. The Ubiquitin-Proteasome System. J. Biosci 31, 137鈥?55 (2016).Article聽Google Scholar聽 2.Senft, D., Qi, J. Ronai, Z. A. Ubiquitin ligases in oncogenic transformation and cancer therapy. Nat. Rev. Cancer 18, 69鈥?8 (2018).Article聽 CAS聽 PubMed聽Google Scholar聽 3.Pinto-Fernandez, A. Kessler, B. M. DUBbing cancer: Deubiquitylating enzymes involved in epigenetics, DNA damage and the cell cycle as therapeutic targets. Front. Genet. 7, 1鈥?3 (2016).Article聽 CAS聽Google Scholar聽 4.Isaacson, M. K. Ploegh, H. L. Ubiquitination, Ubiquitin-like Modifiers, and Deubiquitination in Viral Infection. Cell Host Microbe 5, 559鈥?70 (2009).Article聽 CAS聽 PubMed聽Google Scholar聽 5.Ciechanover, A. Brundin, P. The ubiquitin proteasome system in neurodegenerative diseases: sometimes the chicken, sometimes the egg. Neuron 40, 427鈥?46The ubiquitin proteasome system in neurodeg (2003).6.Manasanch, E. E. Orlowski, R. Z. Proteasome inhibitors in cancer therapy. Nat. Rev. Clin. Oncol. 14, 417鈥?33 (2017).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 7.Bartlett, J. B., Dredge, K. Dalgleish, A. G. The evolution of thalidomide and its IMiD derivatives as anticancer agents. Nat. Rev. Cancer 4, 314鈥?22 (2004).Article聽 CAS聽 PubMed聽Google Scholar聽 8.Hurley, J. H. Stenmark, H. Molecular Mechanisms of Ubiquitin-Dependent Membrane Traffic. Annu. Rev. Biophys. 40, 119鈥?42 (2011).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 9.Shilatifard, A. Chromatin Modifications by Methylation and Ubiquitination: Implications in the Regulation of Gene Expression. Annu. Rev. Biochem. 75, 243鈥?69 (2006).Article聽 CAS聽 PubMed聽Google Scholar聽 10.Jackson, S. P. Durocher, D. Review Regulation of DNA Damage Responses by Ubiquitin and SUMO. Mol. Cell 49, 795鈥?07 (2013).Article聽 CAS聽 PubMed聽Google Scholar聽 11.Everett, R. D. et al. A novel ubiquitin-specific protease is dynamically associated with the PML nuclear domain and binds to a herpesvirus regulatory protein. 16, 1519鈥?530 (1997).12.Kim, R. Q. Sixma, T. K. Regulation of USP7: A high incidence of E3 complexes. J. Mol. Biol., https://doi.org/10.1016/j.jmb.2017.05.028 (2017).13.Kusafuka, T. et al. Mutation analysis of p53 gene in childhood malignant solid tumors. J. Pediatr. Surg. 32, 1175鈥?180 (1997).14.Olivier, M., Hollstein, M. Hainaut, P. TP53 mutations in human cancers: origins, consequences, and clinical use. Cold Spring Harb. Perspect. Biol. 2, 1鈥?7 (2010).Article聽 CAS聽Google Scholar聽 15.Ding, Q. et al. Discovery of RG7388, a potent and selective p53-MDM2 inhibitor in clinical development. J. Med. Chem. 56, 5979鈥?983 (2013).Article聽 CAS聽 PubMed聽Google Scholar聽 16.Chang, Y. S. et al. Stapled 伪鈭抙elical peptide drug development: A potent dual inhibitor of MDM2 and MDMX for p53-dependent cancer therapy. Proc. Natl. Acad. Sci. 110, E3445鈥揈3454 (2013).Article聽 PubMed聽Google Scholar聽 17.Zhou, J. et al. USP7: Target Validation and Drug Discovery for Cancer Therapy. Med. Chem. (Los. Angeles). 14, 3鈥?8 (2018).CAS聽Google Scholar聽 18.Yeasmin Khusbu, F., Chen, F. Z. Chen, H. C. Targeting ubiquitin specific protease 7 in cancer: A deubiquitinase with great prospects. Cell Biochem. Funct. 36, 244鈥?54 (2018).Article聽 CAS聽 PubMed聽Google Scholar聽 19.Song, M. S. et al. The deubiquitinylation and localization of PTEN are regulated by a HAUSP-PML network. Nature 455, 813鈥? (2008).Article聽 ADS聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 20.van der Horst, A. et al. FOXO4 transcriptional activity is regulated by monoubiquitination and USP7/HAUSP. Nat. Cell Biol. 8, 1064鈥?073 (2006).Article聽 CAS聽 PubMed聽Google Scholar聽 21.Tavana, O. et al. HAUSP deubiquitinates and stabilizes N-Myc in neuroblastoma. Nat. Med. 22 (2016).22.Kashiwaba, S. et al. USP7 Is a Suppressor of PCNA Ubiquitination and Oxidative-Stress-Induced Mutagenesis in Human Cells. Cell Rep. 13, 2072鈥?080 (2015).Article聽 CAS聽 PubMed聽Google Scholar聽 23.Faustrup, H., Bekker-Jensen, S., Bartek, J., Lukas, J. Mailand, N. USP7 counteracts SCF尾TrCP- but not APCCdh1-mediated proteolysis of Claspin. J. Cell Biol. 184, 13鈥?9 (2009).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 24.Chauhan, D. et al. A Small Molecule Inhibitor of Ubiquitin-Specific Protease-7 Induces Apoptosis in Multiple Myeloma Cells and Overcomes Bortezomib Resistance. Cancer Cell 22, 345鈥?58 (2013).Article聽 CAS聽Google Scholar聽 25.Arcipowski, K. M. et al. USP7 cooperates with NOTCH1 to drive the oncogenic transcriptional program in T cell leukemia. (2018).26.Kategaya, L. et al. USP7 small-molecule inhibitors interfere with ubiquitin binding. Nature, https://doi.org/10.1038/nature24006 (2017).27.Turnbull, A. P. et al. Molecular basis of USP7 inhibition by selective small-molecule inhibitors. Nature 550, 481鈥?86 (2017).Article聽 ADS聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 28.Lee, G. et al. Small-molecule inhibitors of USP7 induce apoptosis through oxidative and endoplasmic reticulum stress in cancer cells. Biochem. Biophys. Res. Commun. 470, 181鈥?86 (2016).Article聽 CAS聽 PubMed聽Google Scholar聽 29.Lamberto, I. et al. Structure-Guided Development of a Potent and Article Structure-Guided Development of a Potent and Selective Non-covalent Active-Site Inhibitor of USP7. Cell Chem. Biol. 1鈥?1, https://doi.org/10.1016/j.chembiol.2017.09.003 (2017).30.Stolte, B. et al. Genome-scale CRISPR-Cas9 screen identifies druggable dependencies in TP53 wild-type Ewing sarcoma. J. Exp. Med. jem.20171066, https://doi.org/10.1084/jem.20171066 (2018).31.Wales, T. E. Engen, J. R. Hydrogen exchange mass spectrometry for the analysis of protein dynamics. Mass Spectrom. Rev. 25, 158鈥?70 (2006).Article聽 ADS聽 CAS聽 PubMed聽Google Scholar聽 32.Schardon, C. L., Tuley, A., Er, J. A. V., Swartzel, J. C. Fast, W. Selective Covalent Protein Modification by 4-Halopyridines through Catalysis. ChemBioChem 18, 1551鈥?556 (2017).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 33.Frye, S. V. The art of the chemical probe. Nat. Chem. Biol. 6, 159鈥?61 (2010).Article聽 CAS聽 PubMed聽Google Scholar聽 34.Workman, P. Collins, I. Probing the Probes: Fitness Factors For Small Molecule Tools. Chem. Biol. 17, 561鈥?77 (2010).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 35.Li, M., Brooks, C. L., Kon, N. Gu, W. A dynamic role of HAUSP in the p53-Mdm2 pathway. Mol. Cell 13, 879鈥?86 (2004).Article聽 CAS聽 PubMed聽Google Scholar聽 36.Felle, M. et al. The USP7/Dnmt1 complex stimulates the DNA methylation activity of Dnmt1 and regulates the stability of UHRF1. Nucleic Acids Res. 39, 8355鈥?365 (2011).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 37.Liefke, R., Karwacki-Neisius, V. Shi, Y. EPOP Interacts with Elongin BC and USP7 to Modulate the Chromatin Landscape. Mol. Cell 64, 659鈥?72 (2016).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 38.Maertens, G. N., El Messaoudi-Aubert, S., Elderkin, S., Hiom, K. Peters, G. Ubiquitin-specific proteases 7 and 11 modulate Polycomb regulation of the INK4a tumour suppressor. EMBO J. 29, 2553鈥?565 (2010).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 39.Semrau, S. et al. Dynamics of lineage commitment revealed by single-cell transcriptomics of differentiating embryonic stem cells. Nat. Commun. 8 (2017).40.Soumillon, M., Cacchiarelli, D., Semrau, S., van Oudenaarden, A. Mikkelsen, T. S. Characterization of directed differentiation by high-throughput single-cell RNA-Seq. bioRxiv 003236, https://doi.org/10.1101/003236 (2014).41.Subramanian, A., Tamayo, P., Mootha, V. K., Mukherjee, S. Ebert, B. L. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide. Proc Natl Acad Sci USA 102, 15545鈥?5550 (2005).Article聽 ADS聽 CAS聽 PubMed聽Google Scholar聽 42.Yu, C. et al. High-throughput identification of genotype-specific cancer vulnerabilities in mixtures of barcoded tumor cell lines. 34 (2016).43.Tsherniak, A. et al. Resource Defining a Cancer Dependency Map. Cell 170, 564鈥?70.e1644.Giacomelli, A. O. et al. Mutational processes shape the landscape of TP53 mutations in human cancer. Nat. Genet. 1, https://doi.org/10.1038/s41588-018-0204-y (2018).45.Hussain, S., Zhang, Y. Galardy, P. J. DUBs and cancer: The role of deubiquitinating enzymes as oncogenes, non-oncogenes and tumor suppressors. Cell Cycle 8, 1688鈥?697 (2009).Article聽 CAS聽 PubMed聽Google Scholar聽 46.Renatus, M. et al. Targeted inhibition of the COP9 signalosome for treatment of cancer, https://doi.org/10.1038/ncomms13166 (2016).47.Zlatanou, A. et al. USP7 is essential for maintaining Rad18 stability and DNA damage tolerance. 35, 965鈥?76 (2015).48.Schwertman, P. et al. UV-sensitive syndrome protein UVSSA recruits USP7 to regulate transcription-coupled repair. Nat. Genet. 44, 598鈥?02 (2012).Article聽 CAS聽 PubMed聽Google Scholar聽 49.Zhao, Y., Majid, M. C., Soll, J. M., Brickner, J. R. Dango, S. Noncanonical regulation of alkylation damage resistance by the OTUD 4 deubiquitinase. 34, 1687鈥?703 (2015).50.Cai, J.-B. et al. Ubiquitin-specific protease 7 accelerates p14 ARF degradation by deubiquitinating thyroid hormone receptor-interacting protein 12 and promotes hepatocellular carcinoma progression. Hepatology 61, 1603鈥?614 (2015).Article聽 CAS聽 PubMed聽Google Scholar聽 51.Dar, A., Shibata, E. Dutta, A. Deubiquitination of Tip60 by USP7 Determines the Activity of the p53-Dependent Apoptotic Pathway. Mol. Cell. Biol. 33, 3309鈥?320 (2013).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 52.Zhang, Z. Marshall, A. G. A universal algorithm for fast and automated charge state deconvolution of electrospray mass-to-charge ratio spectra. J. Am. Soc. Mass Spectrom. 9, 225鈥?33 (1998).Article聽 CAS聽 PubMed聽Google Scholar聽 53.Hughes, C. S. et al. Ultrasensitive proteome analysis using paramagnetic bead technology. Mol. Syst. Biol. 10, 757鈥?57 (2014).Article聽 CAS聽 PubMed聽Google Scholar聽 54.Alexander, W. M., Ficarro, S. B., Adelmant, G. Marto, J. A. multiplierz v2.0: A Python-based ecosystem for shared access and analysis of native mass spectrometry data. Proteomics 17, 15鈥?6 (2017).Article聽 CAS聽Google Scholar聽 55.Askenazi, M., Parikh, J. R. Marto, J. A. mzAPI: A new strategy for efficiently sharing mass spectrometry data. Nat. Methods 6, 240鈥?41 (2009).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 56.Ficarro, S., Alexander, W. Marto, J. mzStudio: A Dynamic Digital Canvas for User-Driven Interrogation of Mass Spectrometry Data. Proteomes 5, 20 (2017).Article聽 CAS聽 PubMed Central聽Google Scholar聽 57.Liao, C., Zhao, X., Brewer, M., May, V. Li, J. Conformational Transitions of the Pituitary Adenylate Cyclase-Activating Polypeptide Receptor, a Human Class B GPCR. Sci. Rep. 7, 1鈥? (2017).Article聽 CAS聽Google Scholar聽 58.Liao, C. et al. Capturing the multiscale dynamics of membrane protein complexes with all-atom, mixed-resolution, and coarse-grained models. Phys. Chem. Chem. Phys. 19, 9181鈥?188 (2017).Article聽 CAS聽 PubMed聽Google Scholar聽 59.Liao, C., May, V. Li, J. PAC1 Receptors: Shapeshifters in Motion. J. Mol. Neurosci., https://doi.org/10.1007/s12031-018-1132-0 (2018).60.Harder, E. et al. OPLS3: A Force Field Providing Broad Coverage of Drug-like Small Molecules and Proteins. J. Chem. Theory Comput. 12, 281鈥?96 (2016).Article聽 CAS聽 PubMed聽Google Scholar聽 61.Bowers, K. J. et al. Scalable Algorithms for Molecular Dynamics Simulations on Commodity Clusters. in Proceedings of the ACM/IEEE Conference on Supercomputing (SC06), Tampa, Florida. 0.1016/0263-7855(96)00018-5 (2006).62.Humphrey, W., Dalke, A. Schulten, K. VMD: visual molecular dynamics. J. Mol. Graph. 14(33鈥?), 27鈥? (1996). Google Scholar聽 63.Lawson, A. P. et al. Identification of deubiquitinase targets of isothiocyanates using SILAC-assisted quantitative mass spectrometry. Oncotarget 5 (2017).64.Ficarro, S. B. et al. Improved electrospray ionization efficiency compensates for diminished chromatographic resolution and enables proteomics analysis of tyrosine signaling in embryonic stem cells. Anal. Chem. 81, 3440鈥?447 (2009).Article聽 CAS聽 PubMed聽Google Scholar聽 65.Askenazi, M., Marto, J. A. Linial, M. The complete peptide dictionary - A meta-proteomics resource. Proteomics 10, 4306鈥?310 (2010).Article聽 CAS聽 PubMed聽Google Scholar聽 66.Rohland, N. Reich, D. Cost-effective, high-throughput DNA sequencing. Genome Res. 939鈥?46, https://doi.org/10.1101/gr.128124.111.22 (2011).67.Srivastava, A., Sarkar, H., Gupta, N. Patro, R. RapMap: A rapid, sensitive and accurate tool for mapping RNA-seq reads to transcriptomes. Bioinformatics 32, i192鈥搃200 (2016).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 68.Love, M. I., Huber, W. Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq. 2. Genome Biol. 15, 1鈥?1 (2014).Article聽 CAS聽Google Scholar聽 69.Liberzon, A. et al. Molecular signatures database (MSigDB) 3.0. Bioinformatics 27, 1739鈥?740 (2011).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 70.Liberzon, A. et al. The Molecular Signatures Database Hallmark Gene Set Collection. Cell Syst. 1, 417鈥?25 (2015).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 71.Ritchie, M. E. et al. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47 (2015).Article聽 CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 72.Dyer, B. W., Ferrer, F. A., Klinedinst, D. K. Rodriguez, R. A noncommercial dual luciferase enzyme assay system for reporter gene analysis. Anal. Biochem. 282, 158鈥?61 (2000).Article聽 CAS聽 PubMed聽Google Scholar聽 Download referencesAcknowledgementsFunding was provided by the Chleck Family Foundation (N.J.S.), the National Science Foundation Graduate Research Fellowship Program (L.D.), the National Institutes of Health R01 CA204915 (K.S.), CA225088 (P.K.S.聽and S.A.B.), CA233800 (J.A.M.),聽CA222218 (J.A.M.),聽and CA211681 (S.J.B.), the St. Baldrick鈥檚 Foundation Robert J. Arceci Innovation Award (K.S.), the DFCI-Northeastern University Joint Program in Cancer Drug Discovery (R.E.I. and S.J.B.), the Mark Foundation (J.A.M. and S.J.B.)聽and a research collaboration with the Waters Corporation (J.R.E.).Author informationAuthor notesThese authors contributed equally: Nathan J. Schauer and Xiaoxi Liu.AffiliationsDepartment of Cancer Biology and the Linde Program in Cancer Chemical Biology, Dana-Farber Cancer Institute, Boston, MA, USANathan J. Schauer,聽Xiaoxi Liu,聽Robert S. Magin,聽Laura M. Doherty,聽Wai Cheung Chan,聽Scott B. Ficarro,聽Wanyi Hu,聽Rebekka M. Roberts,聽Sirano Dhe-Paganon,聽Jarrod A. Marto聽 聽Sara J. BuhrlageDepartment of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA, USANathan J. Schauer,聽Xiaoxi Liu,聽Robert S. Magin,聽Laura M. Doherty,聽Wai Cheung Chan,聽Sirano Dhe-Paganon聽 聽Sara J. BuhrlageDepartment of Systems Biology and Laboratory of Systems Pharmacology, Harvard Medical School, Boston, MA, USALaura M. Doherty,聽Sarah A. Boswell聽 聽Peter K. SorgerDepartment of Chemistry and Chemical Biology, Northeastern University, Boston, MA, USARoxana E. Iacob聽 聽John R. EngenDepartment of Pediatric Oncology, Dana-Farber Cancer Institute and Boston Children鈥檚 Hospital, Boston, MA, USABj枚rn Stolte聽 聽Kimberly StegmaierDr. von Hauner Children鈥檚 Hospital, Department of Pediatrics, University Hospital, LMU Munich, Munich, GermanyBj枚rn StolteThe Broad Institute of MIT and Harvard University, Cambridge, MA, USABj枚rn Stolte,聽Andrew O. Giacomelli,聽William C. Hahn聽 聽Kimberly StegmaierDepartment of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USAAndrew O. Giacomelli,聽Ellen L. Weisberg,聽Arghya Ray,聽Dharminder Chauhan,聽Ken C. Anderson,聽James D. Griffin聽 聽William C. HahnDepartment of Medicine, Harvard Medical School, Boston, MA, USAAndrew O. Giacomelli聽 聽William C. HahnHarvard College, Cambridge, MA, USASumner PereraDepartment of Chemistry, University of Vermont, Burlington, VT, USAKyle McKay聽 聽Jianing LiThe LeBow Institute for Myeloma Therapeutics and Jerome Lipper Myeloma Center, Dana-Farber Cancer Institute, Boston, MA, USAArghya Ray,聽Dharminder Chauhan聽 聽Ken C. AndersonDepartment of Oncologic Pathology and Blais Proteomics Center, Dana-Farber Cancer Institute, Boston, MA, USAScott B. Ficarro聽 聽Jarrod A. MartoDepartment of Pathology, Brigham and Women鈥檚 Hospital and Harvard Medical School, Boston, MA, USAScott B. Ficarro聽 聽Jarrod A. MartoAuthorsNathan J. SchauerView author publicationsYou can also search for this author in PubMed聽Google ScholarXiaoxi LiuView author publicationsYou can also search for this author in PubMed聽Google ScholarRobert S. MaginView author publicationsYou can also search for this author in PubMed聽Google ScholarLaura M. DohertyView author publicationsYou can also search for this author in PubMed聽Google ScholarWai Cheung ChanView author publicationsYou can also search for this author in PubMed聽Google ScholarScott B. FicarroView author publicationsYou can also search for this author in PubMed聽Google ScholarWanyi HuView author publicationsYou can also search for this author in PubMed聽Google ScholarRebekka M. RobertsView author publicationsYou can also search for this author in PubMed聽Google ScholarRoxana E. IacobView author publicationsYou can also search for this author in PubMed聽Google ScholarBj枚rn StolteView author publicationsYou can also search for this author in PubMed聽Google ScholarAndrew O. GiacomelliView author publicationsYou can also search for this author in PubMed聽Google ScholarSumner PereraView author publicationsYou can also search for this author in PubMed聽Google ScholarKyle McKayView author publicationsYou can also search for this author in PubMed聽Google ScholarSarah A. BoswellView author publicationsYou can also search for this author in PubMed聽Google ScholarEllen L. WeisbergView author publicationsYou can also search for this author in PubMed聽Google ScholarArghya RayView author publicationsYou can also search for this author in PubMed聽Google ScholarDharminder ChauhanView author publicationsYou can also search for this author in PubMed聽Google ScholarSirano Dhe-PaganonView author publicationsYou can also search for this author in PubMed聽Google ScholarKen C. AndersonView author publicationsYou can also search for this author in PubMed聽Google ScholarJames D. GriffinView author publicationsYou can also search for this author in PubMed聽Google ScholarJianing LiView author publicationsYou can also search for this author in PubMed聽Google ScholarWilliam C. HahnView author publicationsYou can also search for this author in PubMed聽Google ScholarPeter K. SorgerView author publicationsYou can also search for this author in PubMed聽Google ScholarJohn R. EngenView author publicationsYou can also search for this author in PubMed聽Google ScholarKimberly StegmaierView author publicationsYou can also search for this author in PubMed聽Google ScholarJarrod A. MartoView author publicationsYou can also search for this author in PubMed聽Google ScholarSara J. BuhrlageView author publicationsYou can also search for this author in PubMed聽Google ScholarContributionsN.J.S., X.L., and S.J.B. initiated the project and oversaw all aspects of the project. X.L. and W.H. performed chemical synthesis. R.S.M. and R.R. performed mutagenic and biochemical studies. N.J.S., R.S.M., and W.C.C prepared purified samples for MS, N.J.S. and W.C.C. prepared ABPP and chemoproteomic samples, and W.C.C. and S.B.F. analyzed MS samples. N.J.S. performed cell treatments and analysis of Western blot, qPCR, and cell cycle analysis samples. R.E.I. performed HDX MS experiments. K.M. performed MD simulations. A.O.G. performed luciferase competition assays. N.J.S., B.S., S.P., E.L.W., and A.R. performed cell viability assays. L.D. and S.A.B. performed DGE-RNA-seq. S.d.P. oversaw crystallization efforts and aided in design of USP7 mutants. D.C. and K.C.A. oversaw experiments using multiple myeloma cell lines. J.D.G. oversaw experiments in AML cell lines. K.S. oversaw experiments in Ewing sarcoma cell lines. J.L. oversaw MD simulations. W.C.H. oversaw luciferase competition assays. P.K.S. oversaw RNA-seq. J.R.E. oversaw HDX MS experiments. J.A.M. oversaw all other MS experiments.Corresponding authorCorrespondence to Sara J. Buhrlage.Ethics declarations Competing interests K. Stegmaier participates in the DFCI/Novartis Drug Discovery Program, which included grant support and consulting and has consulted for Rigel Pharmaceuticals on topics unrelated to this manuscript. J.A.M. serves on the SAB of 908 Devices and receives sponsored research support from Vertex and AstraZeneca.聽The authors declare no other potential conflicts of interest. Additional informationPublisher鈥檚 note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary information Arghya Ray, Ting Du, Yan Song, Sara J. Buhrlage, Dharminder Chauhan Kenneth C. Anderson Leukemia (2021) CommentsBy submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate. Sign up for the Nature Briefing newsletter 鈥?what matters in science, free to your inbox daily.