Cullin-RING ligase BioE3 reveals molecular-glue-induced neosubstrates and rewiring of the endogenous Cereblon ubiquitome

Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 https://doi.org/10.1186/s12964‑025‑02091‑5 RESEARCH Open Access © The Author(s) 2025. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Cell Communication and Signaling Cullin‑RING ligase BioE3 reveals molecular‑glue‑induced neosubstrates andrewiring oftheendogenous Cereblon ubiquitome Laura Merino‑Cacho1, Orhi Barroso‑Gomila1,8, Mónica Pozo‑Rodríguez1, Veronica Muratore1,9,10, Claudia Guinea‑Pérez1, Álvaro Serrano2,3, Coralia Pérez1, Sandra Cano‑López1, Ainhoa Urcullu1, Mikel Azkargorta1,4, Ibon Iloro1,4, Carles Galdeano3,5, Jordi Juárez‑Jiménez2,3, Ugo Mayor6,7, Felix Elortza1,4, Rosa Barrio1* and James D. Sutherland1* Abstract Background The specificity of the ubiquitination process is mediated by the E3 ligases. Discriminating genuine substrates of E3s from mere interacting proteins is one of the major challenges in the field. We previously devel‑ oped BioE3, a biotin‑based approach that uses BirA‑E3 fusions together with ubiquitin fused to a low‑affinity AviTag to obtain a site‑specific and proximity‑dependent biotinylation of the substrates. We proved the suitability of BioE3 to identify targets of RING and HECT‑type E3 ligases. Methods BioE3 experiments were performed in HEK293FT and U2OS stable cell lines expressing TRIPZ‑bioGEFUb transiently transfected with BirA‑cereblon (CRBN). Cells were seeded using biotin‑free media, followed later by a short‑biotin pulse. We evaluated the applicability of the BioE3 system to CRBN and molecular glues by Western blot and confocal microscopy, blocking the proteasome with bortezomib, inhibiting NEDDylation with MLN4924 and treating the cells with pomalidomide. For the identification of endogenous substrates and neosubstrates we ana‑ lyzed the eluates of streptavidin pull‑downs of BioE3 experiments by LC–MS/MS. Analysis of targets for which ubiquit‑ ination changes significantly upon treatment was done using two‑sided Student’s t‑test. Orthogonal validations were performed by histidine pull‑down, GFP‑trap and computational modelling. Results Here we demonstrate that BioE3 is suitable for the multi‑protein complex Cullin‑RING E3s ligases (CRLs), the most utilized E3‑type for targeted protein degradation (TPD) strategies. Using CRBN as proof of concept, one of the substrate receptors of CRL4 E3 ligase, we identified both endogenous substrates and novel neosubstrates upon pomalidomide treatment, including CSDE1 which contains a G‑loop motif potentially involved in the binding to CRBN in presence of pomalidomide. Importantly, we observed a major rearrangement of the endogenous ubiquit‑ ination landscape upon treatment with this molecular glue. *Correspondence: Rosa Barrio [email protected] James D. Sutherland [email protected] Full list of author information is available at the end of the article Page 2 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 Conclusions The ability of BioE3 to detect and compare both substrates and neosubstrates, as well as how sub‑ strates change in response to treatments, will facilitate both on‑target and off‑target identifications and offer a broader characterization and validation of TPD compounds, like molecular glues and PROTACs. Keywords Ubiquitin, Targeted Protein Degradation, Molecular Glue, Immunomodulatory drugs, E3 ligases Background Protein ubiquitination is a post-translational modification involved in almost all cellular processes and plays a crucial role in the regulation of protein homeostasis [1]. Ubiquitin (Ub) is covalently attached to the target protein in an highly regulated enzymatic cascade that involves two activating E1s, around 40 conjugating E2s and more than 600 E3 ligases [2]. Furthermore, ubiquitination can be reversed by deubiquitinating enzymes (DUBs) [3]. Depending on their mechanism of Ub transfer to the substrate protein, E3s are classified in three main families: RING (Really Interesting New Gene; around 600 members), HECT (Homology to E6AP C Terminus, around 30 members) and RBR (RING-Between-RING, around 15 members) [4, 5]. Cullin-RING E3 ligases (CRLs) are the most abundant class of RING E3s and are composed by multiple subunits [6]. These complexes contain a cullin scaffold, a RING finger protein (RBX1 or RBX2) that binds the E2, and distinct sets of adaptors and substrate receptors that specifically recruit target proteins. CRL activity requires cullin NEDDylation and is downregulated by deNEDDylation mediated by the COP9 signalosome. CRLs are presently the most used subfamily for targeted protein degradation (TPD), a strategy that uses chemicals to recruit undesired proteins-of-interest to an E3 ligase for ubiquitination and degradation. CRLs based on cereblon (CRBN) and Von Hippel-Lindau tumor suppressor (VHL) are currently the most commonly used in TPD. Specifically, CRBN is the substrate receptor of the CUL4–RBX1–DDB1–CRBN (CRL4CRBN) E3 complex. CRBN was identified as the target of immunomodulatory imide drugs (IMiDs) [7], which include thalidomide, pomalidomide and lenalidomide. These compounds and their derivatives have been the basis for many described protein degradation drugs, like monovalent molecular glues (MGs) and bivalent proteolysis targeting chimeras (PROTACs) [8, 9]. IMiDs bind to CRBN and alter its substrate specificity, acting as MGs and leading to ubiquitination of non-native substrates (neosubstrates). Ubiquitination by a particular E3 or on particular substrates can occur in diverse ways (on a single site, multiple sites, and with extended ubiquitin-chains of different linkages and topologies). This influences substrate fate and often, but not always, leads to protein degradation. Therefore, surveying endogenous ubiquitinated substrates and of neosubstrates upon cellular treatment with molecular glues and PROTACs should be a crucial step in validation of this promising drug class. Understanding substrate recognition by particular E3 ligases is a relevant area of research in the Ub field, especially in the light of new developments in TPD [10]. Assays that aid in the discovery and/or characterization of substrate specificity of E3 ligases will be an important addition to the chemical biology and drug discovery toolbox. Different strategies to identify targets of E3 ligases have been developed. Proximity proteomics, which has been applied to members of the multi-subunit RING SCF (Skp, Cullin, F-box) complex and others, may identify interactors, some of which may be targets [11–13]. Other strategies involve the direct fusion of E3 ligases to Ub-like proteins (UbLs) (UBAIT, TULIP, and SATT) or Ub-binding domains [14–17]. Overexpression of an E3 in combination with epitope-tagged Ub has also been used to identify candidate E3 substrates [18, 19], but some of these substrates might not be direct. To complement these approaches, we recently developed BioE3 [20], a biotin-based strategy based on two elements: (1) the fusion of the BirA enzyme, a biotin ligase that labels specifically a biotin acceptor peptide (AviTag), to the E3 ligase of interest; and (2) a UbL fused to an AviTag with lower affinity for BirA (bioGEF) [21]. The use of bioGEFUbLs allows a site-specific and proximity-dependent biotinylation that leads to the specific labeling of the ubiquitinated substrates. Those can be captured by streptavidin pull-down and identified using liquid chromatography-mass spectrometry (LC–MS/MS) proteomics. BioE3 was applied to the RING non-associated to cullins (RNF4, MIB1, MARCH5 and RNF214) and HECT (NEDD4) E3 ligases. Similar methods (E-STUB and UbPOD) have been described that support this bioUb-based approach to identify targets [22, 23]. Here we demonstrate how BioE3 can be used for the identification of both CRBN endogenous substrates and neosubstrates in presence of iMiDs. By fusing BirA to the N-terminus of CRBN, we show specific biotinylation in HEK293FTand U2OS-TRIPZ-bioGEFUb cells. We validated Spalt-like 4 (SALL4) as a neosubstrate upon pomalidomide treatment, thus confirming the capacity of CRBN BioE3 to identify neosubstrates. Our proteomic study identified known and novel endogenous substrates of CRBN, and potential pomalidomide-induced Page 3 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 neosubstrates, including CSDE1 (Cold Shock Domain Containing E1), with orthogonal validation and computational modelling to explore binding sites. Importantly, we discovered global differences in the ubiquitination of endogenous substrates upon pomalidomide treatment. By revealing changes in both endogenous substrates and neosubstrates of particular E3-drug combinations, we anticipate that BioE3 will be a very useful tool in the future development of TPD. Methods Cell culture U2OS (ATCC HTB-96) and HEK293FT (Invitrogen) were cultured at 37ºC and 5% CO2 in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS, Biowest) and 1% penicillin/streptomycin (Gibco). HEK293FT cells were used for Western blot and mass spectrometry experiments, whereas U2OS cells were used for confocal microscopy. For all BioE3 experiments, cells were pre-cultured for 24h in biotinfree media supplemented with 10% dialyzed FBS (3.5kDa MWCO; 150mM NaCl; filter-sterilized) prior to transfections, to allow robust labelling during biotin pulses. Cultured cells were maintained for a maximum of 20 passages and tested negative for mycoplasma. Cloning Plasmids were generated by standard cloning or Gibson Assembly (NEBuilder HiFi Assembly, NEB). XL10Gold bacteria (Agilent) were used. Depending on the construction, we used plasmid backbones derived from TRIPZ (Open Biosystems/Horizon) or Lenti-Cas9blast (Addgene #52,962, kindly provided by F. Zhang). TRIPZ-bioGEFUb and TRIPZ-bioGEFUbnc were previously described (Addgene #208,045, 208,044) [20]. CRBN ORF was amplified from hTERT-RPE1 cell cDNA by high-fidelity PCR (Platinum SuperFi DNA Polymerase; Invitrogen #12,351,010) and was inserted into the EcoR1Not1 sites of Lenti-EFS-BirAopt-GSQ-RBXN-P2A-blast (Addgene #208,048) [20]. CRBN mutation described in the text was introduced by 2-fragment overlap PCR and Gibson assembly or using primers: CRBN.W386A. qc.for (agctggtttcctgggtatgccGCTactgttgcccagtgtaagatc) and CRBN.W386A.qc.rev (gatcttacactgggcaacagtAGCggcatacccaggaaaccagct). Constructions were validated by Sanger sequencing. Further construct details are available upon request. Lentiviral transduction Packaging of lentiviral expression constructs was done in HEK293FT cells by transfecting psPAX2 and pMD2.G (kindly provided by D. Trono; Addgene #12,260, #12,259) and pTAT (kindly provided by P. Fortes; for TRIPZ-based vectors) using calcium phosphate. After 12–18h transfection media were removed and replaced with fresh media. Lentiviral supernatants were collected twice (24h each), pooled, filtered (0.45 µm), supplemented with sterile 8.5% PEG6000, 0.3 M NaCl, and incubated for 12–18h at 4°C. Lentiviral particles were concentrated by centrifugation (1, 500 × g, 45min, 4°C). HEK293FT and U2OS cells were transduced with non-concentrated or 5 × concentrated virus, respectively. Drug selection was performed with 1µg/ml puromycin (ChemCruz). Transfections anddrug treatments HEK293FT and U2OS cells were transfected using calcium phosphate or Lipofectamine 3000 (Thermo Fisher), respectively. TRIPZ cell lines stably transduced were induced with DOX (doxycycline hyclate 1µg/ml; 24h; Sigma-Aldrich) prior to biotin treatment (50µM; 2h; Sigma-Aldrich). BTZ (200 nM; MedChemExpress), MLN4924 (1µM, MedChemExpress) and pomalidomide (10µM, MedChemExpress) treatments were performed without biotin prior to cell lysis or immunostaining at the indicated time-points. Western blot analysis To remove excess biotin, we washed cells with 1 × PBS and then we lysed them in highly stringent washing buffer (WB) 5 (WB5: 8M urea, 1% SDS in 1 × PBS) containing 1 × protease inhibitor cocktail (Roche) and 50µM N-Ethylmaleimide (NEM, Alfa Aesar). Samples were sonicated and centrifuged (16,000 × g, 30min at room temperature, RT). Protein concentration was determined by the BCA Protein Assay (Pierce) following manufacturer’s instructions. For SDS-PAGE, 20µg of protein was loaded and transferred to nitrocellulose membranes. PBT (1 × PBS, 0.1% Tween-20) was used for blocking, except for antibiotin blots, where casein-based blocking solution (Sigma) was used. Primary antibodies were incubated for 2h at RT or overnight at 4ºC and secondary antibodies for 45min at RT. Antibodies were used as follows: anti-biotin-HRP (1/1,000, Cell Signaling Technology Cat#7075S); anti-BirA (1/1,000, SinoBiological Cat#11,582-T16); antiAviTag (1/1,000, GenScript Cat#A00674); anti-NEDD8 (1/1,000, Abcam Cat# ab81264); anti-GAPDH (1/5,000, Proteintech Cat# 60,004–1-Ig); anti-CSDE1 (1/1,000, Proteintech Cat#13,319–1-AP); anti-GFP (1/1,000, Roche Cat#11,814,460,001); anti-GFP (1/2,000, rabbit polyclonal serum; generated in-house against recombinant GFP protein); anti-HA-tag (1/2,000, Cell Signaling Technology Cat#3724); anti-Mouse-HRP (1/5,000, Jackson ImmunoResearch Cat#115–035–062); anti-Rabbit-HRP (1/5,000, Jackson ImmunoResearch Cat#111–035–045). We used Super Signal West Femto (ThermoFisher) or Clarity ECL (BioRad) to detect the proteins using an Page 4 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 iBright CL1500 (Thermo Fisher). Uncropped blots are provided as Supplementary Material 12. Immunostaining andconfocal microscopy U2OS cells were seeded on 11 mm coverslips (20,000 cells per well; 24 well plate). After washing the cells with 1 × PBS they were fixed with 4% PFA supplemented with 0.1% Triton X-100 in 1 × PBS for 20min at RT. Then, coverslips were washed 3 times with 1 × PBS and incubated in blocking buffer (2% fetal calf serum, 1% BSA in 1 × PBS) for 30min at RT. Primary antibodies were incubated for 1h at 37ºC and cells were washed 3 times with 1 × PBS. Primary antibodies were used as follow: anti-BirA (1/200, Novus Biologicals Cat#NBP2-59,939); anti-NEDD8 (1/100, Abcam Cat# ab81264). Secondary antibodies and fluorescent streptavidin were incubated for 30min at 37ºC. After that, nuclei were stained with DAPI (300ng/ml in 1 × PBS, Sigma Aldrich) for 10min. Secondary antibodies (ThermoFisher) were all used at 1/200: anti-Mouse Alexa Fluor 488 (Cat#A-11029), antiRabbit Alexa Fluor 647 (Cat#A-21244), anti-Mouse Alexa Fluor 647 (Cat#A-31571). Streptavidin Alexa Fluor 594 (1/200, Cat#016–290-084, Jackson ImmunoResearch) was also used. Images were taken with a confocal microscope (Leica SP8 Lightning) using 63 × Plan ApoChromat NA1.4 objective. Pull‑down ofbiotinylated proteins The lysates cleared in WB5 were normalized to the same protein concentration and incubated overnight at RT with equilibrated NeutrAvidin-agarose beads (ThermoFisher) at a ratio of 1/50 (Vbeads/ Vlysate). The high affinity between biotin and streptavidin allows stringent series of washes, as follows (VWB/2Vlysate): 2 × WB1 (8M urea, 0.25% SDS); 3 × WB2 (6 M Guanidine-HCl); 1 × WB3 (6.4M urea, 1M NaCl, 0.2% SDS); 3 × WB4 (4M urea, 1M NaCl, 10% isopropanol, 10% ethanol and 0.2% SDS); 1 × WB1; 1 × WB5; and 3 × WB6 (2% SDS; WB1-6 prepared in 1 × PBS). Biotinylated proteins were eluted from the beads using 1 volume of Elution Buffer (4 × Laemmli buffer, 100mM DTT; 80µl for LC–MS/MS experiments) by heating at 99°C for 5min twice, followed by vortexing. Beads were separated using 0.8µm clarifying filters (2,000 × g, 2min; Vivaclear Mini, Sartorius). Liquid Chromatography Mass Spectrometry (LC–MS/MS) Pull-down experiments for mass-spectrometry were performed independently in triplicates. For each replicate, four confluent 15cm dishes (8 × 107 cells, 2ml of lysis per plate; 8ml total) were analyzed by LC–MS/ MS. Samples eluted from the NeutrAvidin beads were separated in SDS-PAGE and stained with Sypro Ruby (Invitrogen) following manufacturer’s instructions. Gel lanes were carefully cut to ensure consistency and reproducibility. Slices were subsequently washed in milli-Q water. Reduction and alkylation were performed (10mM DTT in 50mM ammonium bicarbonate, 56°C, 20min, followed by 50mM chloroacetamide in 50mM ammonium bicarbonate, 20 min, protected from light). Gel pieces were dried and incubated with trypsin (12.5µg/ ml in 50mM ammonium bicarbonate, 20min, ice-cold). After rehydration, the trypsin supernatant was discarded. After hydration with 50 mM ammonium bicarbonate, gel pieces were incubated at 37°C overnight. Following digestion, 0.1% TFA was used to clean acidic peptides, which were dried in a RVC2 25 SpeedVac concentrator (Christ). Peptides were resuspended in 10µl 0.1% formic acid (FA) and sonicated for 5min prior to analysis. Samples were analyzed using a timsTOF Pro mass spectrometer (trapped ion mobility spectrometry/quadrupole time of flight hybrid, Bruker Daltonics) coupled online to a EVOSEP ONE (Evosep), which uses parallel accumulation–serial fragmentation (PASEF), at the proteomics platform of CIC bioGUNE. Sample (200ng) was directly loaded in a 15cm performance column (Evosep) applying a 30 samples per day method and data dependent acquisition mode. Mass spectrometry data analysis DIA data was processed with DIA-NN software for protein identification and quantification using default parameters. Searches were carried out against a database consisting of Homo sapiens protein entries from Uniprot in library-free mode. We considered carbamidomethylation of cysteines and oxidation of methionines as fixed and variable modifications, respectively. Match between runs was applied and precursor FDR was set at 1%. Data was processed and analyzed by Perseus (version 1.6.15) [24]. Proteins identified by at least 2 peptides and present in at least 2 out of 3 replicates in at least one group were included in the analysis. Statistical significance was assessed using a two-sided Student’s t-test. Data were loaded into GraphPad Prism 10 version 10.0.2 to build the corresponding volcano-plots. Network analysis was conducted using STRING version 1.4.2 in Cytoscape version 3.9.1, applying a high-confidence interaction score of 0.7 [25, 26]. Size, transparency and width of the edges were continuously mapped to the Log2 fold change. The Molecular COmplex DEtection (MCODE) plug-in version 1.5.1 was used to identify highly connected subclusters of proteins (degree cutoff of 2; Cluster finding: Haircut; Node score cutoff of 0.2; K-Core of 2; Max. Depth of 100) [27]. Gene ontology analysis was performed using g:Profiler web server version e108_eg55_p17_0254fbf and REVIGO [28, Page 5 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 29]. Venn diagrams were drawn using InteractiVenn web tool [30]. Immobilized metal affinity chromatography forvalidations For orthogonal validations, HEK293FT cells were co-transfected with the indicated constructs and pcDNA3-6xHIS-Ub (gift from M. Rodriguez, CRNSLCC, Toulouse), lysed in lysis buffer (8M urea, 0.1M Na2HPO4/NaH2PO4 pH 8.0, 0.01M Tris–HCl pH 8.0, 20 mM imidazole pH 8.0, 5 mM β-mercaptoethanol, and 0.1% Triton X-100), supplemented with 1 × protease inhibitor cocktail (Roche) and 50µM PR619 DUB inhibitor (Merck). Samples were then sonicated and cleared by centrifugation (25, 000 × g, 30 min at RT). Cleared lysates were adjusted to the same protein concentration before incubating them with 1/50 (volbeads/vollysate) equilibrated Ni–NTA agarose beads (Invitrogen) overnight at RT. Beads were then washed three times using WBB (8M urea, 0.1M Na2HPO4/NaH2PO4 pH 8.0, 0.01M Tris–HCl pH 8.0, 20 mM imidazole pH 8.0, 2.5 mM β-mercaptoethanol, and 0.1% Triton X-100), and two times using WBC (8M urea, 0.1M Na2HPO4/NaH2PO4 pH 6.3, 0.01M Tris–HCl pH 6.3, 10mM imidazole pH 7.0, 2.5mM β-mercaptoethanol, and 0.1% Triton X-100). Proteins were eluted with 1 volbeads of Elution Buffer (4M urea, 50mM NaH2PO4/Na2HPO4, 5mM Tris/HCl pH 8, 500mM, imidazole pH 7.0, 1.25mM β-mercaptoethanol, and 0.05% Triton X-100). Eluates were analyzed by western blotting with appropriate antibodies. GFP‑trap pull‑down All steps were performed at 4ºC. HEK293FT cells were collected 48h after transfection, washed 3 times with 1 × PBS and lysed in RIPA lysis buffer (50mM Tris–HCl pH 8, 150mM NaCl, 1% IGEPAL CA-630, 0.5% sodium deoxycholate, and 0.1% SDS) supplemented with 1 × protease inhibitor cocktail (Roche), BTZ (MedChemExpress) and 50µM PR619 (Merck). Lysates were kept on ice for 30min and cleared by centrifugation (25,000 × g, 30 min at 4ºC). Cleared lysates were incubated with 15µl of equilibrated GFP-Selector beads (Nanotag Biotechnologies) overnight at 4ºC in a rotating wheel. Beads were washed 4 times with RIPA lysis buffer, twice with RIPA/4M urea, and once more with RIPA. The samples were eluted in 2 × Laemmli buffer by boiling for 5min at 95ºC. Eluates were analyzed by western blotting with appropriate antibodies. Molecular modelling The G-loop regions of CSDE1 were identified by comparing the position of α-carbon atoms between the CSDE1 AlphaFold model AF-O75534-F1 and the crystallographic structure of SALL4 with PDB id: 7BQU. The structure for the first CSDE1 domain was obtained from the AlphaFold model AF-O75534-F1 and the second domain was extracted from the NMR model with PDB id: 2YTV. The ternary complexes were modeled by superimposing each of the G-loop sequences of CSDE1 to the structure of CK1α in complex with CRBN (PDB id 5FQD) as reference and the pomalidomide-binding mode in the thalidomide-binding domain was taken from the crystallographic structure with PDB id: 6H0F. The AMBERff14SB and GAFF2 force fields were used to assign atom types for the proteins and pomalidomide, respectively. The partial charges for pomalidomide were derived using the RESP protocol at the HF/6-31G(d) level of theory, calculated using Gaussian16 [31]. The Zn2 + cation coordinated with CRBN was modeled with the bound model, using the ZAFF parameters [32]. The systems were solvated on a truncated octahedral box of TIP3P water molecules, and neutralized with salt counterions. Following the protocol we have recently reported [33], each system was minimized, heated to 298K, and equilibrated to 1bar. Each H-bond in the ternary complex interface was evaluated using 100 independent steered molecular dynamics trajectories. Starting positions and velocities for each system were sampled from independent classical MD 10ns trajectories, using a flatbottom restraint to keep the H-bonds between 2.5 and 3.5 Å. Then the H-bonds were brought to a 2.5 Å distance to start a constant speed steering at a speed of 0.5 Å/ ns, using the stiff spring approximation. All simulations were performed with the CUDA accelerated version of PMEMD from the Amber23 package [34]. The potential mean force (PMF) of each H-bond was computed using the Jarzynski equality on the resulting work profiles [35]. The error estimations for the PMF profiles were obtained by bootstrapping 25 times with 25 replica subsamples. Statistical analysis Statistical analysis was performed with GraphPad Prism 10 using two-tailed unpaired Student’s t-test. Differences were considered significant when p-value < 0.05 (*). Quantification of protein levels in Western blots and statistical analysis are shown in Supplementary Material 13. Results Applying BioE3 tostudy Cullin‑RING E3 ligases (CRL) We previously used BioE3 for detecting specific targets of RING and HECT type E3 ligases [20], but the potential of BioE3 to detect targets of E3s multi-protein complexes like the CRLs was not tested. First, we considered the orientation of the BirA enzyme within the fusion protein, which could produce steric problems. We performed BioE3 experiments with BirA fused to the Nor C-terminus of CRBN and we observed a similar pattern Page 6 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 of biotinylated proteins (Fig. S1a). Based on this and previous reports [36, 37], we continued using a fusion of BirA to the N-terminus of CRBN (Fig.1a, b) and transiently transfected it into stable cell lines expressing the bioGEFUb in a doxycycline (DOX)-dependent manner (TRIPZ-bioGEFUb). The low-affinity bioGEF AviTag enables site-specific and proximity-dependent biotinylation [21]. In this way, after DOX induction for 24h and controlled biotin pulses, substrates modified by the CRL4BirA−CRBN and labelled with biotinylated bioGEFUb can be purified using streptavidin pull-down for identification by LC–MS/MS, or imaged by immunofluorescence and confocal microscopy (Fig.1a, b). To test the specificity of the system, we performed BioE3 using BirA-CRBN and HEK293FT-TRIPZ-bioGEFUb cells, inducing the expression of bioGEFUb with DOX and performing a 2-h biotin pulse. We observed biotinylated proteins, indicating the activity of BirA-CRBN fusion (Fig.1c), similarly to what was previously reported by Huang and collaborators for CRBN-BirA [22]. Additionally, when proteasome activity was blocked with bortezomib (BTZ) we observed a further accumulation of biotinylated proteins that was reversed upon inhibition of NEDDylation with MLN4924 (Fig.1c). We then confirmed co-localization of the BirA-CRBN fusion protein with NEDD8 in the cytoplasm by confocal microscopy in U2OS-TRIPZ-bioGEFUb cells (Fig.1d), which suggests that the fusion was correctly incorporated into the CRL complex. Upon proteasomal inhibition we observed an increase in biotinylated proteins (Fig. 1d, Strep panel) that co-localized with the BirA enzyme, indicating specific biotinylation. MLN4924 treatment reduced the biotin labeling and dispersed the NEDD8 signal. Taken together, these data suggest that the BioE3 system is biotinylating CRBN substrates in a CRLdependent manner. Next, we tested wild type (WT) bioGEFUb for its use in CRBN-BioE3. In previous experiments [20], we used a non-cleavable version of Ub bearing the L73P mutation (Ubnc) to prevent the recycling of biotinylated bioGEFUb by suppressing access of DUBs. However, to analyze HECT-type E3s (e.g. NEDD4), we found that it is necessary to use the bioGEFUb, since bioGEFUbnc was not efficiently passed from E2 to E3 [20]. We performed BioE3 experiments both in HEK293FT-TRIPZ-bioGEFUb and HEK293FT-TRIPZ-bioGEFUbnc stable cell lines (Fig. S1b) and observed higher abundance of biotinylated proteins when using bioGEFUb. Importantly, as shown in Fig.1d, biotinylated material was restricted to BirA-CRBN localization, indicating the specificity of labelling in our conditions, so bioGEFUb was used for the rest of experiments shown here. BioE3 identifies SALL4 asaneosubstrate ofCRBN The transcription factor SALL4 is one of the best-known neosubstrates of CRBN induced by pomalidomide [38– 40], so we aimed to identify SALL4 by BioE3 upon treatment with pomalidomide as a proof of concept. We used a CRBN mutant deficient in IMiD-binding (BirA-CRBNW386A) [41] that should act as a negative control for the identification of neosubstrates upon pomalidomide treatment. We performed BioE3 experiments in HEK293FTand U2OS-TRIPZ-bioGEFUb cell lines transiently transfected with the WT or mutant versions of BirACRBN, blocking the proteasome with BTZ and treating the cells with the IMiD drug pomalidomide (Fig. S1c). By Western blot we noticed a change in the biotinylation pattern when treating the samples with pomalidomide, suggesting that BioE3 is sensitive to molecular glues and can differentiate the endogenous targets from the neosubstrates. BirA-CRBNW386A also produced biotinylation of proteins, but without changing the pattern observed after adding the IMiD (Fig. S1c). By confocal microscopy, we could validate the suitability of the mutant for BioE3 experiments: interestingly, BirA-CRBNWT displayed a nuclear localization after pomalidomide treatment that was not observed with BirA-CRBNW386A, together with an increase in the nuclear biotinylated material (Fig.2a, Strep panel). (See figure on next page.) Fig. 1 BioE3 labels CRL‑dependent ubiquitinated substrates of CRBN. a, b Schematic representation of the BioE3 strategy adapted to the substrate receptor CRBN (a) and the constructs used in this work (b). TRIPZ, all‑in‑one inducible lentiviral vector; bioGEF, low affinity AviTag (see text); DOX, doxycycline; TetON, tetracycline inducible promoter; PUROR, puromycin resistant cassette; EFS, elongation factor 1α short promoter. c Western blot of BioE3 experiment performed on HEK293FT stable cell line expressing TRIPZ‑bioGEFUb and transfected with EFS‑BirA‑CRBN. Indicated samples were treated with 100 nM bortezomib (BTZ) for 24 h, 200 nM BTZ for 6 h and/or 1 µM MLN4924 for 6 or 24 h. Molecular weight markers are shown to the left of the blots in kDa, antibodies used are indicated to the right. AviTag antibodies highlight all the ubiquitinated proteins, while biotin shows those ubiquitinated by CRBN. d Confocal microscopy images of BioE3 experiment performed on U2OS stable cell line expressing TRIPZ‑bioGEFUb transfected with BirA‑CRBN. Indicated samples were treated with 200 nM BTZ for 6 h and/or 1 µM MLN4924 for 24 h. Biotinylated material is stained with fluorescent streptavidin (Strep, magenta) and BirA (green) and NEDD8 (blue) with specific antibodies. Scale bar: 8 µm. All BioE3 experiments were performed by pre‑incubating the cells in dialyzed FBS‑containing media prior to transfections, doxycycline (DOX) induction at 1 µg/ml for 24 h and biotin supplementation at 50 µM for 2 h Page 7 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 Fig. 1 (See legend on previous page.) Page 8 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 Considering that the endogenous expression level of SALL4 in HEK293FT cells is low [42], we decided to perform the BioE3 experiment with exogenously expressed SALL4-YFP and treating the cells with BTZ and/or pomalidomide. After isolating the biotinylated proteins by streptavidin pull-down, we observed an enrichment of polyubiquitinated SALL4-YFP upon pomalidomide treatment with the CRBNWT version but not with the IMiDbinding mutant CRBNW386A (Fig.2b). We confirmed this result using confocal microscopy, observing co-localization of BirA-CRBNWT and biotin signal at nuclear bodies formed by SALL4-YFP only in those cells treated with pomalidomide (Fig.2c). Furthermore, BirA-CRBNW386A did not co-localize to the SALL4 nuclear bodies, neither biotinylated SALL4-YFP, even after treatment with the IMiD. Altogether, we concluded that BioE3 is sensitive to the responses induced by molecular glues and can be used to discriminate between endogenous substrates and neosubstrates. BioE3 identifies endogenous targets ofCRBN Once we confirmed the suitability of BioE3 to biotinylate both putative substrates and neosubstrates after molecular glue treatment, we performed a large-scale experiment in triplicates, treating the cells with BTZ only, BTZ and MLN4924 or BTZ and pomalidomide (Figs.3and4; Supplementary Tables1–3). We isolated the biotinylated proteins by streptavidin pull-downs and confirmed by Western blot the enrichment of the eluates in biotinylated proteins when the proteasome was blocked with BTZ (Fig.3a). By inhibiting NEDDylation the amount of biotinylated proteins was reduced, supporting MLN4924 as a useful negative control. We also observed a reduction in biotinylated proteins after pomalidomide treatment. The eluates were then analyzed by LC–MS/MS to identify specific ubiquitinated targets of CRBN. First, by comparing the samples treated or not with BTZ (BTZ versus DMSO), we identified 376 putative targets of CRBN that are proteasome-targeted (Fig. 3b; Supplementary Table1). Among them, we found glutamine synthetase (GLUL), an endogenous target of the substrate receptor previously described [43]. We also interrogated the effect of inhibiting NEDDylation in the identification of endogenous substrates of CRBN, by comparing BTZ versus BTZ and MLN4924 treated cells, and identified 267 putative targets of CRBN that require NEDD8 activation of the CRL complex (Fig.3c; Supplementary Table2). Among them, 177 (66%) were also present in the CRBN BioE3 BTZ versus DMSO comparison, including GLUL (Fig. 3d). Interestingly, NEDD8 was significantly enriched when inhibiting NAE1 (left quadrant of Fig.3c), indicating that mixed NEDD8Ub conjugates were accumulating in that sample. This was also confirmed by Western blot, showing an enrichment in high molecular weight NEDDylated proteins (Fig.3e). These results are in line with previous reports showing the use of NEDD8 by the Ub machinery when the NEDDylation machinery is blocked [44, 45]. In addition to the aforementioned hits, we identified proteins related to the UbL machinery, including ubiquitination (E3s and DUBs), SUMOylation (SUMO1, SUMO2, PIAS1, PIAS4), NEDDylation (NEDD8, COPS6) and proteasome components. We also found components of the cAMP signaling pathway like MTOR, PRKAR1A and CREB3 [46], components of the Wnt signaling pathway such as CSNK1E and CTNNB1 [47], chloride channels as CLCC1 and CLCN3 [48], BSG for which a ubiquitination-independent, chaperone-like function of CRBN was described [49], and components of CRLs (DDB1) and the COP9 signalosome (COPS6, COPS7A, COPS3) [50]. To determine the functional role of CRBN we performed a STRING network analysis of the potential substrates (Figs. S2 and S4). The network showed a major interconnected core-cluster composed of 71% and 64% of the identified substrates respectively in BTZ versus Fig. 2 Identification of SALL4 as a neosubstrate of CRBN upon pomalidomide treatment. a Confocal microscopy of BioE3 experiment performed in U2OS stable cell line stably expressing TRIPZ‑bioGEFUb. Cells were transfected with EFS‑BirA‑CRBNWT or EFS‑BirA‑CRBNW386A (IMiD‑binding deficient mutant). Indicated samples were treated with 10 µM pomalidomide (POM) and/or with 200 nM bortezomib (BTZ) for 6 h. Biotinylated material is stained with fluorescent streptavidin (Strep, magenta) and BirA (green) with specific antibodies. b, c BioE3 experiment performed in HEK293FT‑TRIPZ‑bioGEFUb (b) or U2OS‑TRIPZ‑bioGEFUb (c) stable cell lines transiently transfected with CMV‑SALL4‑YFP and EFS‑BirA‑CRBNWT or EFS‑BirA‑CRBNW386A and treated with 10 µM POM and/or 200 nM BTZ for 6 h. b Western blot validation of SALL4 as a neosubstrate upon pomalidomide treatment. The dot represents the possible monoubiquitinated protein, whereas the bar represents the polyUb‑modified SALL4‑YFP. Strep PD: streptavidin pull‑down. Molecular weight markers are shown to the left of the blots in kDa, antibodies used are indicated to the right. Quantification graph of the ubiquitinated GFP‑SALL4 level in the GFP panel is shown under the Western blots; *: p‑value < 0.05. c) Biotinylated material is stained with fluorescent streptavidin (Strep, magenta) and BirA (blue) with a specific antibody. SALL4‑YFP can be found in green. Scale bar: 8 µm. Yellow dotted lines indicate the nuclei. Insets show the amplification of the area indicated by a white dotted square in each panel. All BioE3 experiments were performed by pre‑incubating the cells in dialyzed FBS‑containing media prior to transfections, doxycycline (DOX) induction at 1 µg/ml for 24 h and biotin supplementation at 50 µM for 2 h (See figure on next page.) Page 9 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 Fig. 2 (See legend on previous page.) Page 16 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 Abbreviations bioGEF Low affinity AviTag BTZ Bortezomib CRBN Cereblon CRL Cullin‑RING E3 ligase CRL4CRBN CUL4–RBX1–DDB1–CRBN E3 complex CSDE1 Cold Shock Domain Containing E1 DOX Doxycycline DUB Deubiquitinating enzyme FBS Fetal Bovine Serum GLUL Glutamine Synthetase HECT Homology to E6AP C Terminus IMiD Immunomodulatory Imide Drug LC–MS/MS Liquid chromatography‑mass spectrometry MG Molecular Glue PMF Potential Mean Force POM Pomalidomide PROTAC Proteolysis Targeting Chimeras RBR RING‑Between‑RING RING Really Interesting New Gene SALL4 Spalt‑like 4 TPD Targeted Protein Degradation Ub Ubiquitin UbL Ubiquitin‑like UPS Ubiquitin Proteasome System Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s12964‑ 025‑ 02091‑5. Supplementary Material 1. Fig. S1. Optimization of the experimental conditions for CRBN BioE3. a, b, c) Western blot of BioE3 experiments performed on HEK293FT stable cell lines expressing TRIPZ‑bioGEFUb or TRIPZ‑bioGEFUbnc and transfected with EFS‑BirA‑CRBN, EFS‑CRBN‑BirA or EFS‑BirA‑CRBNW386A. Indicated samples were induced with doxycycline at 1 µg/ml for 24 hours, treated with 200 nM bortezomib for 6 hours, 1 µM MLN4924 for 24 hours or 10 µM pomalidomide for 6 hours and sup‑ plemented with 50 µM biotin for 2 hours. Molecular weight markers are shown to the left of the blots in kDa, antibodies used are indicated to the right. All BioE3 experiments were performed by pre‑incubating the cells in dialyzed FBS‑containing media prior to transfections. Supplementary Material 2. Fig. S2. STRING network analysis of CRBN ubiquitinated substrates. Substrates defined in Figure 3b show a highly interconnected network composed of 71% of the proteins. Highly inter‑ connected sub‑clusters were derived and characterized using MCODE. Color, transparency and size of the nodes were discretely mapped to the Log2 enrichment value as indicated. Supplementary Material 3. Fig. S3. Gene ontology scatterplot of the CRBN ubiquitinated substrates. REVIGO plots were generated for targets defined in Figure 3b. Colors indicate the ‑Log10 p‑value as shown in the Figure, and size of the bubble indicates the size of each term. Only terms with a p‑value < 0.05 are represented. Supplementary Material 4. Fig. S4. STRING network analysis of CRBN NEDDylation‑dependent substrates. Substrates defined in Figure 3c show a highly interconnected network composed of 64% of the proteins. Highly interconnected sub‑clusters were derived and characterized using MCODE. Color, transparency and size of the nodes were discretely mapped to the Log2 enrichment value as indicated. Supplementary Material 5. Fig. S5. Gene ontology scatterplot of CRBN NEDDylation‑dependent substrates. REVIGO plots were generated for targets defined in Figure 3c. Colors indicate the ‑Log10 p‑value as shown in the Figure, and size of the bubble indicates the size of each term. Only terms with a p‑value < 0.05 are represented. Supplementary Material 6. Fig. S6. STRING network analysis of CRBN neosubstrates. The neosubstrates upon pomalidomide treatment defined in Figure 4a show a highly interconnected network composed of 71% of the proteins. Highly interconnected sub‑clusters were derived and characterized using MCODE. Color, transparency and size of the nodes were discretely mapped to the Log2 enrichment value as indicated. Supplementary Material 7. Fig. S7. Gene ontology scatterplot of the neosubstrates of CRBN. REVIGO plots were generated for targets defined in Figure 4a. Colors indicate the ‑Log10 p‑value as shown in the Figure, and size of the bubble indicates the size of each term. Only terms with a p‑value < 0.05 are represented. Supplementary Material 8. Fig. S8. Effect of pomalidomide on endogenous protein levels of CSDE1. a) Endogenous CSDE1 levels in HEK293FT cells upon bortezomib, MLN4924 and pomalidomide treatments. b) Quantita‑ tive data and statistical analysis of three independent experiments as in Figure S8a. CSDE1 levels in cells treated with DMSO or pomalidomide 10 µM were taken in consideration for the analysis. c) Endogenous CSDE1 levels in HEK293FT cells transiently transfected with EFS ‑BirA or EFS‑BirA‑ CRBN. d) Quantitative data and statistical analysis of four independent experiments overexpressing BirA‑CRBN as in Figure S8c. CSDE1 levels were normalized to BirA‑CRBN expression. e) BioE3 experiment performed using HEK293FT stable cells that express TRIPZ‑bioGEFUb transiently transfected with BirA‑CRBN. BTZ and POM treatments were performed as indicated. Biotinylated proteins were purified using streptavidin beads. Molecular weight markers are shown to the left of the blots in kDa, anti‑ bodies used are indicated to the right Supplementary Material 9. Supplementary Table 1. LC‑MS/MS processed data of CRBN BioE3 treated or not with proteasomal inhibitor bortezomib. Gene Ontology analysis of CRBN targets and selected GO terms by Revigo Supplementary Material 10. Supplementary Table 2. LC‑MS/MS processed data of CRBN BioE3 with proteasomal inhibitor bortezomib, with or without NEDDylation inhibitor MLN4924. Gene Ontology analysis of CRBN targets and selected GO terms by Revigo Supplementary Material 11. Supplementary Table 3. LC‑MS/MS processed data of CRBN BioE3 with proteasomal inhibitor bortezomib, with or without pomalidomide treatment. Gene Ontology analysis of CRBN targets and selected GO terms by Revigo Supplementary Material 12. Uncropped Western blots. Images with the same exposures as used in the main Figures are shown to the left. Merged images show the positioning of the molecular weight markers and might show different exposure times. Supplementary Material 13. Quantification of protein levels in Western blots and statistical analysis Acknowledgements R.B., J.D.S., U.M. and C.G. acknowledge networking support from the European Cooperation for Science & Technology ProteoCure COST Action (CA20113). C.G. and J.J.‑J. are members of the Computational Biology Drug Design Consolidated Research Group supported by the Generalitat de Catalunya (2021SGR00671). Authors’ contributions L.M.‑C., O.B.‑G., R.B. and J.D.S. conceived and designed the work; L.M.‑C., O.B.‑ G., M.P.‑R., V.M., C.G.‑P., A.S., C.P., S.C.‑L., A.U., M.A., I.I., F.E. and J.D.S. acquired and analyzed the data; L.M.‑C., O.B.‑G., C.G., J.J.‑J., U.M., R.B. and J.D.S. interpreted the data; L.M.‑C., M.P.‑R. and R.B. drafted the manuscript; and J.D.S., O.B.‑G. and C.G., J.J.‑J. substantively revised it. All authors read and approved the final manuscript. Funding R.B. and J.D.S. were funded by MCIN/AEI/https://doi. org/10.13039/501100011033, projects PID2023‑147399NB‑I00 and PID2020‑ 114178 GB‑I00 and CEX2021‑001136‑S and CEX2021‑001202‑M Severo Ochoa Excellence Program and additional support provided by the Diputación Foral de Bizkaia, Programa Transferencia Tecnológica 2023, project 6/12/ TT/2023 /00001, and the Department of Industry, Tourism, and Trade of the Basque Country Government (Elkartek Research Programs); J.J.‑J. was funded by grant CNS2022‑135307 funded by the MCIN/AEI/https://doi. org/10.13039/501100011033and the European Union NextGeneration EU/ PRTR; V.M., C.G.‑P., M.P.‑R. and S.C.‑L. were funded by FPI grants funded by Page 17 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 MCIN/AEI /https://doi.org/10.13039/501100011033and FSE, PRE2018‑086230, PRE2021‑099359, PRE2022‑104553 and PREP2023‑000294, respectively; L.M.‑C. was funded by FPU grant FPU20/05282 (funded by Ministerio de Universi‑ dades para la Formación de Profesorado Universitario). O.B.‑G., J.D.S. and R.B. acknowledge funding by the grant 765445‑EU (UbiCODE Program). U.M. was funded by the Spanish MCIU (PID2020‑117333 GB‑I00 (FEDER/EU)). Access to HPC facilities were granted through the Red Española de Supercomputación (BCV‑2021–3‑0006 and BCV‑2021–2‑0005). Data availability The datasets generated and/or analyzed during the current study are avail‑ able in the ProteomeXchange Consortium via the PRIDE partner repository [57] with the dataset identifier PXD055877. All the other data generated or analyzed during this study are included in this published article [and its sup‑ plementary information files]. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests C.G. is co‑founder and Drug Discovery Scientific Advisor of Oniria Therapeu‑ tics. The other authors declare that they have no competing interests. Author details 1 Center for Cooperative Research in Biosciences (CIC Biogune), Basque Research and Technology Alliance (BRTA), 48160 Derio, Spain. 2 Depar t‑ ment of Pharmacy and Pharmaceutical Technology and Physical‑Chemistry, School of Pharmacy, University of Barcelona, 08028 Barcelona, Spain. 3 Institute of Theoretical and Computational Chemistry (IQTC), School of Chemistry and Physics, University of Barcelona, 08028 Barcelona, Spain. 4 Instituto de Salud Carlos III, C/ Monforte de Lemos 3‑5, Pabellón 11, Planta 0, 28029 Madrid, Spain. 5 Institute of Biomedicine (IBUB), University of Barcelona, 08028 Barcelona, Spain. 6 Biochemistry and Molecular Biology Department, University of the Basque Country, Leioa, Spain. 7 Ikerbasque ‑Basque Founda‑ tion for Science, Bilbao, Spain. 8 Present address: Biobizkaia Health Research Institute, Barakaldo 48903, Spain. 9 Present address: Telethon Institute of Genet‑ ics and Medicine (TIGEM), Pozzuoli, Italy. 10 Department of Clinical Medicine and Surgery, Federico II University, Naples, Italy. Received: 25 October 2024 Accepted: 8 February 2025 References 1. Dikic I, Schulman BA. An expanded lexicon for the ubiquitin code. Nat Rev Mol Cell Biol. 2023;24(4):273–87. 2. Cappadocia L, Lima CD. Ubiquitin‑like protein conjugation: structures, chemistry, and mechanism. Chem Rev. 2018;118(3):889–918. 3. Lange SM, Armstrong LA, Kulathu Y. Deubiquitinases: From mechanisms to their inhibition by small molecules. Mol Cell. 2022;82(1):15–29. 4. Zheng N, Shabek N. Ubiquitin ligases: structure, function, and regulation. Annu Rev Biochem. 2017;86:129–57. 5. Morreale FE, Walden H. Types of ubiquitin ligases. Cell. 2016;165(1):248–e1. 6. Baek K, Scott DC, Schulman BA. NEDD8 and ubiquitin ligation by cullin‑ RING E3 ligases. Curr Opin Struct Biol. 2021;67:101–9. 7. Ito T, Ando H, Suzuki T, Ogura T, Hotta K, Imamura Y, et al. Identifi‑ cation of a primary target of thalidomide teratogenicity. Science. 2010;327(5971):1345–50. 8. Oleinikovas V, Gainza P, Ryckmans T, Fasching B, Thoma NH. From thalido‑ mide to rational molecular glue design for targeted protein degradation. Annu Rev Pharmacol Toxicol. 2024;64:291–312. 9. Tsai JM, Nowak RP, Ebert BL, Fischer ES. Targeted protein degradation: from mechanisms to clinic. Nat Rev Mol Cell Biol. 2024;25(9):740–57. 10. Belcher BP, Ward CC, Nomura DK. Ligandability of E3 ligases for targeted protein degradation applications. Biochemistry. 2023;62(3):588–600. 11. Roux KJ, Kim DI, Raida M, Burke B. A promiscuous biotin ligase fusion protein identifies proximal and interacting proteins in mammalian cells. J Cell Biol. 2012;196(6):801–10. 12. Coyaud E, Mis M, Laurent EM, Dunham WH, Couzens AL, Robitaille M, et al. BioID‑based identification of Skp Cullin F‑box (SCF)β‑TrCP1/2 E3 ligase substrates. Mol Cell Proteomics. 2015;14(7):1781–95. 13. Zhang H, Li S, Liu P, Lee FHF, Wong AHC, Liu F. Proteomic analysis of the cullin 4B interactome using proximity‑dependent biotinylation in living cells. Proteomics. 2017;17(8). https:// doi. org/ 10. 1002/ pmic. 20160 0163. 14. O’Connor HF, Lyon N, Leung JW, Agarwal P, Swaim CD, Miller KM, et al. Ubiquitin‑Activated Interaction Traps (UBAITs) identify E3 ligase binding partners. EMBO Rep. 2015;16(12):1699–712. 15. Kumar R, Gonzalez‑Prieto R, Xiao Z, Verlaan‑de Vries M, Vertegaal ACO. The STUbL RNF4 regulates protein group SUMOylation by targeting the SUMO conjugation machinery. Nat Commun. 2017;8(1):1809. 16. Salas‑Lloret D, Jansen NS, Nagamalleswari E, van der Meulen C, Gracheva E, de Ru AH, et al. SUMO‑activated target traps (SATTs) enable the identification of a comprehensive E3‑specific SUMO proteome. Sci Adv. 2023;9(31):eadh2073. https:// doi. org/ 10. 1126/ sciadv. adh20 73. 17. Watanabe M, Saeki Y, Takahashi H, Ohtake F, Yoshida Y, Kasuga Y, et al. A substrate‑trapping strategy to find E3 ubiquitin ligase substrates identi‑ fies Parkin and TRIM28 targets. Commun Biol. 2020;3(1):592. 18. Ramirez J, Lectez B, Osinalde N, Sivá M, Elu N, Aloria K, et al. Quantitative proteomics reveals neuronal ubiquitination of Rngo/Ddi1 and several proteasomal subunits by Ube3a, accounting for the complexity of Angel‑ man syndrome. Hum Mol Genet. 2018;27(11):1955–71. 19. Alduntzin U, Lectez B, Presa N, Osinalde N, Fernandez M, Elu N, et al. Angelman Syndrome causing UBE3A ligase displays predominantly synaptic ubiquitination activity in the mouse brain. Research Square. 2023. https:// doi. org/ 10. 21203/ rs.3. rs‑ 26658 03/ v1. 20. Barroso‑Gomila O, Merino‑Cacho L, Muratore V, Perez C, Taibi V, Maspero E, et al. BioE3 identifies specific substrates of ubiquitin E3 ligases. Nat Commun. 2023;14(1):7656. 21. Fernandez‑Suarez M, Chen TS, Ting AY. Protein‑protein interaction detection in vitro and in cells by proximity biotinylation. J Am Chem Soc. 2008;130(29):9251–3. 22. Huang HT, Lumpkin RJ, Tsai RW, Su S, Zhao X, Xiong Y, et al. Ubiquitin‑ specific proximity labeling for the identification of E3 ligase substrates. Nat Chem Biol. 2024;20(9):1227–36. 23. Mukhopadhyay U, Levantovsky S, Carusone TM, Gharbi S, Stein F, Behrends C, et al. A ubiquitin‑specific, proximity‑based labeling approach for the identification of ubiquitin ligase substrates. Sci Adv. 2024;10(32):eadp3000. 24. Tyanova S, Temu T, Sinitcyn P, Carlson A, Hein MY, Geiger T, et al. The Per‑ seus computational platform for comprehensive analysis of (prote)omics data. Nat Methods. 2016;13(9):731–40. 25. Snel B, Lehmann G, Bork P, Huynen MA. STRING: a web‑server to retrieve and display the repeatedly occurring neighbourhood of a gene. Nucleic Acids Res. 2000;28(18):3442–4. 26. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecu‑ lar interaction networks. Genome Res. 2003;13(11):2498–504. 27. Bader GD, Hogue CW. An automated method for finding molecular complexes in large protein interaction networks. BMC Bioinformatics. 2003;4:2. 28. Reimand J, Arak T, Adler P, Kolberg L, Reisberg S, Peterson H, et al. g:Profiler‑a web server for functional interpretation of gene lists (2016 update). Nucleic Acids Res. 2016;44(W1):W83–9. 29. Supek F, Bosnjak M, Skunca N, Smuc T. REVIGO summarizes and visualizes long lists of gene ontology terms. PLoS ONE. 2011;6(7): e21800. 30. Heberle H, Meirelles GV, da Silva FR, Telles GP, Minghim R. InteractiVenn: a web‑based tool for the analysis of sets through Venn diagrams. BMC Bioinformatics. 2015;16(1):169. 31. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, et al. Gaussian 16 Rev. C.01. Wallingford, CT2016. https:// gauss ian. com/ citat ion/. 32. Peters MB, Yang Y, Wang B, Fusti‑Molnar L, Weaver MN, Merz KM Jr. Struc‑ tural survey of zinc containing proteins and the development of the zinc AMBER force field (ZAFF). J Chem Theory Comput. 2010;6(9):2935–47. Page 18 of 18 Merino‑Cachoetal. Cell Communication and Signaling (2025) 23:101 33. Miñarro‑Lleonar M, Bertran‑Mostazo A, Duro J, Barril X, Juarez‑Jimenez J. Lenalidomide stabilizes protein‑protein complexes by turning labile intermolecular H‑bonds into robust interactions. J Med Chem. 2023;66(9):6037–46. 34. Case DA, Aktulga HM, Belfon K, Cerutti DS, Cisneros GA, Cruzeiro VWD, et al. AmberTools. J Chem Inf Model. 2023;63(20):6183–91. 35. Jarzynski C. Nonequilibrium equality for free energy differences. Phys Rev Lett. 1997;78(14):2690–3. 36. Yamanaka S, Horiuchi Y, Matsuoka S, Kido K, Nishino K, Maeno M, et al. A proximity biotinylation‑based approach to identify protein‑E3 ligase interactions induced by PROTACs and molecular glues. Nat Commun. 2022;13(1):183. 37. Costacurta M, Sandow JJ, Maher B, Susanto O, Vervoort SJ, Devlin JR, et al. Mapping the IMiD‑dependent cereblon interactome using BioID‑proxim‑ ity labelling. FEBS J. 2024;291(22):4892–912. 38. Matyskiela ME, Couto S, Zheng X, Lu G, Hui J, Stamp K, et al. SALL4 medi‑ ates teratogenicity as a thalidomide‑dependent cereblon substrate. Nat Chem Biol. 2018;14(10):981–7. 39. Sievers QL, Petzold G, Bunker RD, Renneville A, Slabicki M, Liddicoat BJ, et al. Defining the human C2H2 zinc finger degrome targeted by thalido‑ mide analogs through CRBN. Science. 2018;362(6414):0572. 40. Donovan KA, An J, Nowak RP, Yuan JC, Fink EC, Berry BC, et al. Thalidomide promotes degradation of SALL4, a transcription factor implicated in Duane Radial Ray syndrome. Elife. 2018;7:e38430. 41. Matyskiela ME, Lu G, Ito T, Pagarigan B, Lu CC, Miller K, et al. A novel cereblon modulator recruits GSPT1 to the CRL4(CRBN) ubiquitin ligase. Nature. 2016;535(7611):252–7. 42. Thul PJ, Akesson L, Wiking M, Mahdessian D, Geladaki A, Ait Blal H, et al. A subcellular map of the human proteome. Science. 2017;356(6340):eaal3321. 43. Nguyen TV, Lee JE, Sweredoski MJ, Yang SJ, Jeon SJ, Harrison JS, et al. Glutamine triggers acetylation‑dependent degradation of glu‑ tamine synthetase via the thalidomide receptor cereblon. Mol Cell. 2016;61(6):809–20. 44. Hjerpe R, Thomas Y, Chen J, Zemla A, Curran S, Shpiro N, et al. Changes in the ratio of free NEDD8 to ubiquitin triggers NEDDylation by ubiquitin enzymes. Biochem J. 2012;441(3):927–36. 45. Hjerpe R, Thomas Y, Kurz T. NEDD8 overexpression results in ned‑ dylation of ubiquitin substrates by the ubiquitin pathway. J Mol Biol. 2012;421(1):27–9. 46. Lee KM, Yang SJ, Choi JH, Park CS. Functional effects of a pathogenic mutation in Cereblon (CRBN) on the regulation of protein synthesis via the AMPK‑mTOR cascade. J Biol Chem. 2014;289(34):23343–52. 47. Shen C, Nayak A, Neitzel LR, Adams AA, Silver‑Isenstadt M, Sawyer LM, et al. The E3 ubiquitin ligase component, Cereblon, is an evolutionarily conserved regulator of Wnt signaling. Nat Commun. 2021;12(1):5263. 48. Chen YA, Peng YJ, Hu MC, Huang JJ, Chien YC, Wu JT, et al. The Cullin 4A/B‑DDB1‑Cereblon E3 ubiquitin ligase complex mediates the degrada‑ tion of CLC‑1 chloride channels. Sci Rep. 2015;5:10667. 49. Eichner R, Heider M, Fernandez‑Saiz V, van Bebber F, Garz AK, Lemeer S, et al. Immunomodulatory drugs disrupt the cereblon‑CD147‑ MCT1 axis to exert antitumor activity and teratogenicity. Nat Med. 2016;22(7):735–43. 50. Schulze‑Niemand E, Naumann M. The COP9 signalosome: A versatile reg‑ ulatory hub of Cullin‑RING ligases. Trends Biochem Sci. 2023;48(1):82–95. 51. Baek K, Metivier RJ, Roy Burman SS, Bushman JW, Lumpkin RJ, Abeja DM, et al. Unveiling the hidden interactome of CRBN molecular glues with chemoproteomics. bioRxiv. 2024. https:// doi. org/ 10. 1101/ 2024. 09. 11. 612438. 52. Ciocia A, Mestre‑Farras N, Vicent‑Nacht I, Guitart T, Gebauer F. CSDE1: a versatile regulator of gene expression in cancer. NAR Cancer. 2024;6(2):zcae014. 53. Petzold G, Fischer ES, Thoma NH. Structural basis of lenalidomide‑ induced CK1alpha degradation by the CRL4(CRBN) ubiquitin ligase. Nature. 2016;532(7597):127–30. 54. Kronke J, Udeshi ND, Narla A, Grauman P, Hurst SN, McConkey M, et al. Lenalidomide causes selective degradation of IKZF1 and IKZF3 in multi‑ ple myeloma cells. Science. 2014;343(6168):301–5. 55. Ichikawa S, Flaxman HA, Xu W, Vallavoju N, Lloyd HC, Wang B, et al. The E3 ligase adapter cereblon targets the C‑terminal cyclic imide degron. Nature. 2022;610(7933):775–82. 56. Fischer ES, Bohm K, Lydeard JR, Yang H, Stadler MB, Cavadini S, et al. Struc‑ ture of the DDB1‑CRBN E3 ubiquitin ligase in complex with thalidomide. Nature. 2014;512(7512):49–53. 57. Vizcaino JA, Csordas A, del‑Toro N, Dianes JA, Griss J, Lavidas I, et al. 2016 update of the PRIDE database and its related tools. Nucleic Acids Res. 2016;44(1):447–56. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ lished maps and institutional affiliations.