henhao
Sign in
BiochemistryNature Communications
Peer-reviewed journal article

Unbiased mapping of cereblon neosubstrate landscape by high-throughput proteomics

Martin StegeriD, Gisele NishiguchiiD, Qiong WuiD, Bjoern Schwalb, Bachuki ShashikadzeiD, Kevin M. McGowaniD, Marisa Actis, Anup AggarwaliD, Zhe ShiiD, Jeanine E. Price, Anand Mayasundari, Lei YangiD, Anastasia Bednarz, Sophie Machata, Tobias Graef, Denis Bartoschek, Vadim DemicheviD, Uli OhmayeriD, Jun J.J. YangiD, Henrik DaubiD, Zoran RankovićiD

Abstract

Molecular glue degraders (MGDs) are small molecules that harness the ubiquitin-proteasome system to induce degradation of target proteins, including those lacking conventional druggable pockets. Given the challenges in their rational design, MGD discovery predominantly relies on screening-based approaches, such as cell viability assays. However, one potential limitation of such screening methods is the risk of overlooking non-essential neosubstrates of potential therapeutic value. To address this concern, we present a high-throughput proteome-wide MGD screening platform utilizing label-free, data-independent acquisition mass spectrometry (DIA-MS) for integrated proteomics and ubiquitinomics analysis. Processing a diverse set of 100 CRBN-ligands across two cancer cell lines reveals a broad array of neosubstrates, including 50 novel candidates validated by MS-based ubiquitinomics. These findings considerably expand the current landscape of CRBN-mediated neosubstrates. Comprehensive hit validation and structure-degradation relationship analyses guided by global proteomics, identifies highly selective and potent phenyl glutarimide-based degraders of novel neosubstrates, including KDM4B, G3BP2 and VCL, none of which contain the classical CRBN degron motif. This study demonstrates that comprehensive, high-throughput proteomic screening offers new opportunities in MGD drug discovery.

PDFDOI
Full-text digest ↓

02 /Specification & Classification

Click any tag to browse matching papers
Citations
40
FWCI
10.49
Field percentile
top 1%
OpenAlex Topics

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (19 Oct 2024) · CC BY 4.0

Proteomics and ubiquitinomics screening of 100 CRBN ligands in two cell lines validated 50 novel neosubstrates and yielded selective phenyl glutarimide degraders of KDM4B, G3BP2 and VCL.

Viability-based screens of molecular glue degraders can miss neosubstrates that cancer cells do not need for survival. This work adds an unbiased proteome-wide screen with ubiquitinomics validation, which widens the known CRBN neosubstrate landscape.

Key results

  1. ShownCombined proteomics, neddylation-inhibitor rescue and ubiquitinomics validated 50 previously unreported CRBN neosubstrates from a selection of 100 compounds; the novel KDM4B, G3BP2 and VCL targets lack the classical degron motif.

    Huh-7: 32 of 100 compounds lowered at least one protein by over 25%, giving 84 candidates; 16 compounds at 6 h downregulated 26; MLN4924 blocked 23 (over 70% reduction); 16 validated by ubiquitinomics. 50 novel in total across Huh-7, NB-4 and HEK293-CRBNoe. · Systematic discovery of novel CRBN neosubstrates; Overexpression of CRBN enhances neosubstrate…; Conclusions

  2. ShownThe label-free diaPASEF screen gave deep, precise proteome coverage and re-detected many previously reported neosubstrates, supporting platform sensitivity.

    720 LC-MS runs; on average 10,200 protein groups per compound in Huh-7 and 9,500 in NB-4; over 98% retained for testing; median CV about 6%. Of 35 known neosubstrates, 33 were quantified and 25 were significantly downregulated by at least one compound. · Deep proteomic screening of a molecular glue…

  3. ShownStable CRBN overexpression in HEK293 cells increased degradation and ubiquitination, confirming weakly degraded candidates missed by ubiquitinomics in the primary lines and revealing extra targets.

    HEK293-CRBNoe had about eight-fold higher CRBN than parental HEK293. SJ10040 ubiquitinomics found 13 neosubstrates (incl. ZBTB21, CSNK1E, ZNF324, ZBTB41, STK35, CHD7); the promiscuous SJ42101 gave about 30. Of 11 candidates undetected earlier, several were confirmed. · Overexpression of CRBN enhances neosubstrate…

  4. ShownChemical core determined neosubstrate profile: classical IMiD-like cores hit most known targets, while phenyl glutarimide (PG) compounds targeted distinct, more selective neosubstrate sets.

    98 of 100 compounds were IMiD-like or PG. IKZF1, ZFP91, CSNK1A1, GSPT1, GZF1, RAB28 were hit only by IMiD-like cores; LIMD1, KDM4B, FLCN only by PG compounds. IMiD-like SJ42229 co-depleted known IMiD targets; PG-based SJ41824 did not. · Impact of MGD structure on degradation…

  5. ShownPG degraders SJ41564 (KDM4B) and SJ46479 (VCL) act potently, selectively and CRBN-dependently. SJ41813 degrades G3BP2 directly, with USP10 as a bystander.

    KDM4B DC50 36 nM, Dmax 92% (SK-N-BE(2)-C); no significant change in over 9,000 NB-4 proteins at 10 µM, 6 h. VCL DC50 255 nM, Dmax 96% (Caki-1). G3BP2 DC50 2.8 µM vs USP10 10 µM; TurboID-MS >6-fold G3BP2 enrichment. · Comprehensive characterization and chemical…; Conclusions

Methods, data and reproducibility

100 CRBN ligands were screened at 10 µM for 24 h in triplicate in two cell lines by label-free diaPASEF proteomics. Hits were retested at 6 h with and without MLN4924, then validated by 30-minute K-GG ubiquitinomics; CRBN-overexpressing HEK293 cells added sensitivity.

  • Proteomics screen of 100 CRBN-ligand MGDs in Huh-7 and NB-4 cells, with DMSO controls (selected from a 5,000-compound library) · human (human cell lines; human UniProt database) · n = 100 compounds, triplicates, 2 cell lines, 120 DMSO controls, 720 LC-MS runs
  • HEK293 vs HEK293-CRBNoe proteomics and ubiquitinomics for eight screen-derived MGDs · human · n = eight MGDs
  • Mechanistic validation in SK-N-BE(2)-C (incl. CRBN knockout), Caki-1 and HEK293-TurboID-CRBN cells (Western blot, TurboID-MS) · human · n = blots representative of two or three independent experiments
Methods & tools
DIA-NN, diaPASEF, slicePASEF, LIMMA, MaxLFQ (DIA-NN R package), MLN4924 (neddylation inhibitor), K-GG ubiquitin remnant enrichment, TurboID proximity labeling, UniProt SwissProt human database, t-SNE, R
Reagents & environment
DIA-NN v1.8.2beta9 or 1.9.1 (proteomics) and v1.8.2beta27 (ubiquitinomics); UniProt SwissProt human release 10-2022; downstream statistics in R with LIMMA.
Reproducibility gaps
No repository for analysis code or deposited raw MS data (e.g., PRIDE) is given in the text; compound identities and the 5,000-compound library are described without public access details.
Limitations (5)
  • Eleven neosubstrate candidates showed no compound-induced ubiquitination in Huh-7 or NB-4 cells, attributed to low expression or modest degradation, so sensitivity was limiting even with up to 30,000 ubiquitination sites quantified. · stated by the authors
  • Two strongly promiscuous compounds (SJ6986, SJ41478) were excluded from neosubstrate analysis because they modulated thousands of proteins, and secondary effects of prolonged exposure may confound some regulations. · stated by the authors
  • The mechanisms behind cell line-specific degradation of ubiquitous proteins, and the structural determinants of CRBN recognition of degron-free neosubstrates, remain unresolved. · stated by the authors
  • Only 100 compounds (one representative per scaffold) and a few cell lines were screened; neosubstrates found in CRBN-overexpressing HEK293 cells may not occur at physiological CRBN levels. · noted in this summary
  • Potency estimates (DC50, Dmax) come from Western blot quantification of blots representative of two or three experiments, and no phenotypic or functional consequence of KDM4B, G3BP2 or VCL degradation is tested. · noted in this summary

Who should read:Read this if you work on molecular glue degraders or targeted protein degradation and want a proteomics plus ubiquitinomics screening design for finding CRBN neosubstrates.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISShashikadze '26Zanon '26Annunziato '25Nie '26Petzold '25Galli '26
83% semantic matchBiochemistry
Bachuki Shashikadze, Ines F. Scheller, Denise Winkler+9 more·bioRxiv·Mar 10, 2026· 4 citations