Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 62 of the 62 references it cites were found in a registry or an index; 10 of the 19 sources its claims cite were checked for support.

Cereblon induces G3BP2 neosubstrate degradation using molecular surface mimicry

Open access

Abstract

Molecular glue degraders (MGDs) are small-molecule compounds that divert E3 ligases to degrade nonnatural substrates called neosubstrates. Clinically effective MGDs bind cereblon (CRBN), a substrate receptor of the Cullin 4–RING E3 ubiquitin ligase (CRL4CRBN), and recruit neosubstrates to an MGD-induced neosurface on the CRBN CULT domain through molecular mimicry of a natural CRBN degron. Here, we identify G3BP2 (Ras–GAP SH3 domain-binding protein 2), a neosubstrate that bypasses canonical interactions with CRBN by engaging an unconventional binding site on the CRBN LON domain. The ternary complex interface does not resemble known interactions with CRBN. Instead, CRBN leverages a preexisting protein–protein interaction (PPI) hotspot on the target protein by mimicking an endogenous binding partner of G3BP2. Our findings suggest that composite neosurfaces that mimic and stabilize the footprint of natural PPIs (in short, ‘glueprints’) could become a viable strategy for the rational expansion of the MGD target repertoire. Annunziato, Quan and Donckele et al. identify G3BP2 (Ras–GAP SH3 domain-binding protein 2) as a molecular glue-induced neosubstrate of the CRL4CRBN E3 ubiquitin ligase. The CRBN–glue neosurface uses a molecular surface mimicry mechanism to recruit and degrade G3BP2 in a compound-dependent manner.

Overview

CoChat Summary

## TL;DR

The paper identifies G3BP2 as a bona fide CRBN neosubstrate degraded by the molecular glue degrader MRT-5702, and shows that recruitment occurs through an unconventional CRBN LON-domain interaction rather than the canonical G-loop/CULT-domain mechanism. Structural, biochemical, computational, and cellular data support a “surface mimicry” model in which CRBN mimics a natural G3BP2 partner to hijack a preexisting protein–protein interaction hotspot.

## Key findings

  • In CAL51 cells, MS-based global proteomics after 6 h of treatment with 10 µM MRT-5702 identified G3BP2 as significantly reduced, with USP10 also reduced to a lesser extent; both effects were dependent on cullin neddylation, proteasome activity, and CRBN.
  • MRT-5702 produced a dose-dependent NanoBRET signal with NanoLuc–G3BP2 in HEK293T cells, but not with NanoLuc–USP10 or NanoLuc–G3BP1, supporting G3BP2 as a specific CRBN neosubstrate.
  • Domain swapping showed that the NTF2L domain of G3BP2 is necessary and sufficient for MRT-5702-induced recruitment and degradation; recombinant G3BP2NTF2L (residues 1–139) formed a ternary complex with CRBN and MRT-5702 in TR-FRET.
  • Cryo-EM of the DDB1ΔBPB–CRBNΔ1–40–MRT-5702–G3BP2NTF2L complex reached 2.5 Å overall resolution, with a locally refined ternary interface of 2.8 Å, and showed that MRT-5702D is the only active configuration for G3BP2; ligand density favored the (S,S)-configuration.

## Methods

  • The study combined global proteomics, cellular NanoBRET, HiBiT assays, mutagenesis/domain swaps, TR-FRET, cryo-EM, and computational modeling to map the CRBN–G3BP2 interaction and validate the ternary complex.
  • Cellular experiments used CAL51 and HEK293T cells; proteomics profiling treated CAL51 cells with 10 µM compound for 6 h, while NanoBRET and degradation assays tested MRT-5702-dependent target engagement and protein loss.
  • Structural work used purified CRBN–DDB1 and G3BP2NTF2L proteins, with cryo-EM grid preparation and single-particle analysis in cryoSPARC, followed by atomic model building and refinement in PHENIX, Coot, ChimeraX, and PyMOL.

## Limitations

  • The paper states that the local resolution at the CRBN–G3BP2 interface was only 3.5–4.5 Å in the initial cryo-EM reconstruction because of conformational flexibility; this limits direct atomic interpretation of the interface in that region.
  • The Methods section refers to a “Nature Portfolio Reporting Summary” and supplementary materials for further research design details, but those details are not included in the provided text; this is evident from the text.
  • The provided text does not state in vivo validation, disease-model testing, or clinical relevance experiments for G3BP2 degradation; this absence is evident from the text.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41594-025-01738-8

Cites

62 of 62 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 G3BP2 is a G-loop-independent CRBN neosubstrate. a, Chemical structure of MRT-5702. Asterisks indicate chiral centers in the compound tail. b, Global TMT proteomics in CAL51 cells using MRT-5702 (10 μM, 6 h) performed in duplicates. Proteins with predicted β-hairpin G-loops in C2H2 and non-C2H2 domains are labeled. Data were analyzed using a two-sided moderated t-test (linear models for microarray). c, Western blot in CAL51 cells using MRT-5702 (1 μM, 24 h) and rescue of degradation by cotreatment with pathway inhibitors. Representative result of three independent experiments. d, Ternary complex formation between CRBN and G3BP2, G3BP1 and USP10. Technical replicates were measured by NanoBRET and are presented as the mean ± s.d. (n = 3). EC50, half-maximal effective concentration. e, Domain architecture of G3BP proteins. f. Levels of overexpressed HiBiT–G3BP2 variants in CAL51 cells measured by luminescence after 24 h of compound treatment. Technical replicates are presented as the means ± s.d. (n = 3). Half-maximal degradation concentration (DC50) and maximum degradation (Dmax) values for wild-type G3BP2 are indicated. g, Ternary complex formation between recombinantly purified DDB1–CRBN and the NTF2L domain of G3BP2, measured by in vitro TR-FRET and presented as mean of two representative measurements. h. Structure of the G3BP1 NTF2L homodimer bound to the FGDF motif of USP10 (PDB 7XHF). Predicted PPI propensity plotted onto the surface of the NTF2L domain shown at the bottom. i, Surface similarity between the USP10 interface with G3BP proteins, a surface patch in the CRBN LON domain and the CRBN F150A mutant. The USP10 self-comparison served as a positive control reference. Fingerprint score is a combined score of shape similarity (shown in surface plots) and electrostatics. Details about surface alignments can be found in the Methods. j, Ternary complex formation between wild-type and interface mutant forms of full-length CRBN and G3BP2. Technical replicates were measured by NanoBRET and are presented as the mean ± s.d. (n = 3). Source data
Fig. 1 G3BP2 is a G-loop-independent CRBN neosubstrate. a, Chemical structure of MRT-5702. Asterisks indicate chiral centers in the compound tail. b, Global TMT proteomics in CAL51 cells using MRT-5702 (10 μM, 6 h) performed in duplicates. Proteins with predicted β-hairpin G-loops in C2H2 and non-C2H2 domains are labeled. Data were analyzed using a two-sided moderated t-test (linear models for microarray). c, Western blot in CAL51 cells using MRT-5702 (1 μM, 24 h) and rescue of degradation by cotreatment with pathway inhibitors. Representative result of three independent experiments. d, Ternary complex formation between CRBN and G3BP2, G3BP1 and USP10. Technical replicates were measured by NanoBRET and are presented as the mean ± s.d. (n = 3). EC50, half-maximal effective concentration. e, Domain architecture of G3BP proteins. f. Levels of overexpressed HiBiT–G3BP2 variants in CAL51 cells measured by luminescence after 24 h of compound treatment. Technical replicates are presented as the means ± s.d. (n = 3). Half-maximal degradation concentration (DC50) and maximum degradation (Dmax) values for wild-type G3BP2 are indicated. g, Ternary complex formation between recombinantly purified DDB1–CRBN and the NTF2L domain of G3BP2, measured by in vitro TR-FRET and presented as mean of two representative measurements. h. Structure of the G3BP1 NTF2L homodimer bound to the FGDF motif of USP10 (PDB 7XHF). Predicted PPI propensity plotted onto the surface of the NTF2L domain shown at the bottom. i, Surface similarity between the USP10 interface with G3BP proteins, a surface patch in the CRBN LON domain and the CRBN F150A mutant. The USP10 self-comparison served as a positive control reference. Fingerprint score is a combined score of shape similarity (shown in surface plots) and electrostatics. Details about surface alignments can be found in the Methods. j, Ternary complex formation between wild-type and interface mutant forms of full-length CRBN and G3BP2. Technical replicates were measured by NanoBRET and are presented as the mean ± s.d. (n = 3). Source dataCC BY-NC-ND · PubMed Central
Fig. 2 Cryo-EM structure of the DDB1–CRBN–MRT-5702–G3BP2 ternary complex. a, Global cryo-EM envelope of the DDB1ΔBPB–CRBN–MRT-5702–G3BP2NTF2L ternary complex. Cryo-EM data collection was carried out once. b, In vitro TR-FRET between DDB1–CRBN and G3BP2NTF2L in the presence of racemic MRT-5702 or its separated enantiomers. Data are presented as the mean of two representative measurements. c, Cellular NanoBRET between CRBN and full-length G3BP2 in the presence of racemic MRT-5702 or its separated stereoisomers. Technical replicates are presented as the means ± s.d. (n = 3). d, Levels of overexpressed HiBiT–G3BP2 in CAL51 cells measured by luminescence after 24 h of compound treatment. DC50 and Dmax values for MRT-5702D are indicated. Technical replicates are presented as the means ± s.d. (n = 3). e, Volume overlap of calculated stereoisomer conformations with the ligand cryo-EM density (SMOC: segment-based Manders overlap coefficient). Each point represents an Omega-generated low-energy conformer. Box-and-whisker plots show the median (horizontal line), interquartile range (IQR; box size) and 1.5× the IQR (whisker length). Upper outliers indicate best fits to the experimental map. f, Close-up view of the PPI interface between G3BP2NTF2L and the CRBN LON-loop in the derived cryo-EM model. Source data
Fig. 2 Cryo-EM structure of the DDB1–CRBN–MRT-5702–G3BP2 ternary complex. a, Global cryo-EM envelope of the DDB1ΔBPB–CRBN–MRT-5702–G3BP2NTF2L ternary complex. Cryo-EM data collection was carried out once. b, In vitro TR-FRET between DDB1–CRBN and G3BP2NTF2L in the presence of racemic MRT-5702 or its separated enantiomers. Data are presented as the mean of two representative measurements. c, Cellular NanoBRET between CRBN and full-length G3BP2 in the presence of racemic MRT-5702 or its separated stereoisomers. Technical replicates are presented as the means ± s.d. (n = 3). d, Levels of overexpressed HiBiT–G3BP2 in CAL51 cells measured by luminescence after 24 h of compound treatment. DC50 and Dmax values for MRT-5702D are indicated. Technical replicates are presented as the means ± s.d. (n = 3). e, Volume overlap of calculated stereoisomer conformations with the ligand cryo-EM density (SMOC: segment-based Manders overlap coefficient). Each point represents an Omega-generated low-energy conformer. Box-and-whisker plots show the median (horizontal line), interquartile range (IQR; box size) and 1.5× the IQR (whisker length). Upper outliers indicate best fits to the experimental map. f, Close-up view of the PPI interface between G3BP2NTF2L and the CRBN LON-loop in the derived cryo-EM model. Source dataCC BY-NC-ND · PubMed Central
Fig. 3 Conformational flexibility of the CRBN LON-loop enhances complementarity with G3BP2. a, Side view of the ternary complex interface involving G3BP2NTF2L, MRT-5702 and the CRBN LON domain. The CRBN LON domain is shown in light gray and the CRBN CULT is shown in dark gray. b, Superposition of G3BP2-engaged CRBN (light gray) with CRBN models from PDB 4TZ4, 6H0G, 5V3O and 6H0F highlights conformational flexibility of the CRBN LON-loop. c, Superposition of the G3BP2–CRBN docking model (generated with CRBN chain B of PDB 5FQD) with the cryo-EM-derived structural model. CRBN residue F150 is shown as spheres. Differences in LON-loop conformation and overall pose of the interacting G3BP2NTF2L protomer are highlighted.
Fig. 3 Conformational flexibility of the CRBN LON-loop enhances complementarity with G3BP2. a, Side view of the ternary complex interface involving G3BP2NTF2L, MRT-5702 and the CRBN LON domain. The CRBN LON domain is shown in light gray and the CRBN CULT is shown in dark gray. b, Superposition of G3BP2-engaged CRBN (light gray) with CRBN models from PDB 4TZ4, 6H0G, 5V3O and 6H0F highlights conformational flexibility of the CRBN LON-loop. c, Superposition of the G3BP2–CRBN docking model (generated with CRBN chain B of PDB 5FQD) with the cryo-EM-derived structural model. CRBN residue F150 is shown as spheres. Differences in LON-loop conformation and overall pose of the interacting G3BP2NTF2L protomer are highlighted.CC BY-NC-ND · PubMed Central
Fig. 4 Detailed interactions at the ternary complex interface. a, Interactions between G3BP2NTF2L helix α1 and α2 residues and the CRBN LON domain. b, Interactions between G3BP2NTF2L and MRT-5702D. c, Interactions among G3BP2NTF2L helix α1, MRT-5702D and the CRBN sensor loop of the CULT domain.
Fig. 4 Detailed interactions at the ternary complex interface. a, Interactions between G3BP2NTF2L helix α1 and α2 residues and the CRBN LON domain. b, Interactions between G3BP2NTF2L and MRT-5702D. c, Interactions among G3BP2NTF2L helix α1, MRT-5702D and the CRBN sensor loop of the CULT domain.CC BY-NC-ND · PubMed Central
Fig. 5 G3BP2 recruitment bypasses canonical PPIs between CRBN and G-loop-like motifs. a, MGD-induced ternary complex interface involving CRBN, pomalidomide and the β-hairpin G-loop target ZNF692 (top). The primary protein–protein interface is formed between the G-loop residues (shown in yellow) of ZNF692 zinc finger 4 (ZF4) and CRBN residues N351, H357 and W400 (bottom; PDB 6H0G). b, Top, MGD-induced ternary complex interface between CRBN and the G-loop surface mimic VAV1. Bottom, the primary protein–protein interface is formed between RT-loop residues (shown in cyan) of the VAV1 SH3c domain and CRBN residues N351, H357 and W400. The VAV1 nSrc-loop residue K815 forms a secondary contact with the CRBN LON-loop (PDB 9NFR). c, Top, G3BP2 ternary complex interface highlighting key contacts with MRT-5702D and the CRBN LON-loop. Bottom, the CRBN LON domain is shown in light gray, the CRBN CULT domain is shown in dark gray, the sensor loop is shown in green and the MGD is shown in pink. Bottom, CRBN residues N351, H357 and W400 are not participating in PPIs. d, Total buried molecular surface area within ternary complexes represented by the footprints of ZNF692 (PDB 6H0G), VAV1 SH3c (PDB 9NFR) and G3BP2 NTF2L (PDB 9OS2) mapped onto the respective CRBN or MGD neosurfaces. Footprints are colored according to electrostatics of the CRBN region involved in the ternary complex interface. The pink contour indicates the surface area contributed by the MGD. Quantification of the total footprint area and its relative distribution over the LON and CULT domains and the MGD is shown in the table below.
Fig. 5 G3BP2 recruitment bypasses canonical PPIs between CRBN and G-loop-like motifs. a, MGD-induced ternary complex interface involving CRBN, pomalidomide and the β-hairpin G-loop target ZNF692 (top). The primary protein–protein interface is formed between the G-loop residues (shown in yellow) of ZNF692 zinc finger 4 (ZF4) and CRBN residues N351, H357 and W400 (bottom; PDB 6H0G). b, Top, MGD-induced ternary complex interface between CRBN and the G-loop surface mimic VAV1. Bottom, the primary protein–protein interface is formed between RT-loop residues (shown in cyan) of the VAV1 SH3c domain and CRBN residues N351, H357 and W400. The VAV1 nSrc-loop residue K815 forms a secondary contact with the CRBN LON-loop (PDB 9NFR). c, Top, G3BP2 ternary complex interface highlighting key contacts with MRT-5702D and the CRBN LON-loop. Bottom, the CRBN LON domain is shown in light gray, the CRBN CULT domain is shown in dark gray, the sensor loop is shown in green and the MGD is shown in pink. Bottom, CRBN residues N351, H357 and W400 are not participating in PPIs. d, Total buried molecular surface area within ternary complexes represented by the footprints of ZNF692 (PDB 6H0G), VAV1 SH3c (PDB 9NFR) and G3BP2 NTF2L (PDB 9OS2) mapped onto the respective CRBN or MGD neosurfaces. Footprints are colored according to electrostatics of the CRBN region involved in the ternary complex interface. The pink contour indicates the surface area contributed by the MGD. Quantification of the total footprint area and its relative distribution over the LON and CULT domains and the MGD is shown in the table below.CC BY-NC-ND · PubMed Central
Extended Data Fig. 1 CRBN architecture, MGD-induced allostery and neosubstrate engagement. a, Domain architecture of human Cereblon (CRBN). b, Compound binding to the CRBN Trp-cage causes major conformational changes that create a neosurface on closed CRBN primed for neosubstrate recruitment. c, View onto the MGD-induced neosurface of closed CRBN (PDB: 5fqd) and areas of high predicted PPI propensity are indicated and separated by domain on the right hand side. d, Ternary complex interface with the CK1α β-hairpin G-loop engaging the two CRBN sensor loop residues N351 and H357, and the Trp-cage residue W400 (PDB: 5fqd). e, Ternary complex interface with the helical G-loop of mTOR-FRB (PDB: 9ngt). f, Ternary complex interface with a G-loop surface mimic, the RT-loop of the VAV1 SH3c domain (PDB: 9nfr).
Extended Data Fig. 1 CRBN architecture, MGD-induced allostery and neosubstrate engagement. a, Domain architecture of human Cereblon (CRBN). b, Compound binding to the CRBN Trp-cage causes major conformational changes that create a neosurface on closed CRBN primed for neosubstrate recruitment. c, View onto the MGD-induced neosurface of closed CRBN (PDB: 5fqd) and areas of high predicted PPI propensity are indicated and separated by domain on the right hand side. d, Ternary complex interface with the CK1α β-hairpin G-loop engaging the two CRBN sensor loop residues N351 and H357, and the Trp-cage residue W400 (PDB: 5fqd). e, Ternary complex interface with the helical G-loop of mTOR-FRB (PDB: 9ngt). f, Ternary complex interface with a G-loop surface mimic, the RT-loop of the VAV1 SH3c domain (PDB: 9nfr).CC BY-NC-ND · PubMed Central
Extended Data Fig. 2 G3BP2 is a G-loop-independent CRBN target. a, G3BP2 Western blot in CAL51 cells using MRT-5702 (24 h treatment) and rescue of degradation by genetic CRBN knock-out. Experiment carried out once. b, Sequence alignment of human G3BP1 and G3BP2 generated with the ENDscript server37. Protein domains are annotated. c, Luminescent signal derived from overexpressed wild-type HiBiT-G3BP1 and a domain-swapped variant in CAL51 cells measured after 24 h compound treatment. Technical replicates presented as means ± s.d. (n = 3). Source data
Extended Data Fig. 2 G3BP2 is a G-loop-independent CRBN target. a, G3BP2 Western blot in CAL51 cells using MRT-5702 (24 h treatment) and rescue of degradation by genetic CRBN knock-out. Experiment carried out once. b, Sequence alignment of human G3BP1 and G3BP2 generated with the ENDscript server37. Protein domains are annotated. c, Luminescent signal derived from overexpressed wild-type HiBiT-G3BP1 and a domain-swapped variant in CAL51 cells measured after 24 h compound treatment. Technical replicates presented as means ± s.d. (n = 3). Source dataCC BY-NC-ND · PubMed Central
Extended Data Fig. 3 Recognition of FGDF and FxFG motifs by the NTF2L domains of G3BP proteins and ternary complex docking model with CRBN. a, Structures of G3BP1 and G3BP2 NTF2L domains bound to FGDF motifs from various endogenous interaction partners. b, FxFG motif recognition by the G3BP1 NTF2L domain. c, NTF2L homodimer of G3BP1 bound to two individual USP10 peptides (PDB: 7xhf). d, Model of a G3BP2 NTF2L domain (PDB: 5drv) bound to the FGDF motif of USP10. e, Mimicry-guided docking model of CRBN engaging the FGDF binding site on the G3BP2 NTF2L domain through the LON domain residues F150 and I152. f, Mimicry-guided docking model of the G3BP2 NTF2L homodimer with closed CRBN (PDB: 5fqd). The three tryptophan residues of the MGD-binding pocket (Trp-cage) in the CRBN CULT domain are shown as spheres. g, Binding of fluorescent CRBN Tracer to WT, F150A or Trp-cage mutant (W386A) form of CRBNΔ191-248 by NanoBRET TE. h-i, Percentage signal of Tracer NanoBRET upon treatment by h, Lenalidomide or i, MRT-5702 in CRBNΔ191-248 WT or the F150A mutant. g-i, Technical replicates presented as mean ± s.d. (n = 3). Source data
Extended Data Fig. 3 Recognition of FGDF and FxFG motifs by the NTF2L domains of G3BP proteins and ternary complex docking model with CRBN. a, Structures of G3BP1 and G3BP2 NTF2L domains bound to FGDF motifs from various endogenous interaction partners. b, FxFG motif recognition by the G3BP1 NTF2L domain. c, NTF2L homodimer of G3BP1 bound to two individual USP10 peptides (PDB: 7xhf). d, Model of a G3BP2 NTF2L domain (PDB: 5drv) bound to the FGDF motif of USP10. e, Mimicry-guided docking model of CRBN engaging the FGDF binding site on the G3BP2 NTF2L domain through the LON domain residues F150 and I152. f, Mimicry-guided docking model of the G3BP2 NTF2L homodimer with closed CRBN (PDB: 5fqd). The three tryptophan residues of the MGD-binding pocket (Trp-cage) in the CRBN CULT domain are shown as spheres. g, Binding of fluorescent CRBN Tracer to WT, F150A or Trp-cage mutant (W386A) form of CRBNΔ191-248 by NanoBRET TE. h-i, Percentage signal of Tracer NanoBRET upon treatment by h, Lenalidomide or i, MRT-5702 in CRBNΔ191-248 WT or the F150A mutant. g-i, Technical replicates presented as mean ± s.d. (n = 3). Source dataCC BY-NC-ND · PubMed Central
Extended Data Fig. 4 Cryo-EM 3D reconstruction workflow. a, CryoSPARC v4.2.1 was used for image analysis, 2D classification and 3D reconstruction. Multiple rounds of 2D classification were performed to remove particles not representing the ternary complex. b, 3D reconstruction, classification, and global/local refinement workflows. Intermediate local resolution estimations are depicted by a blue-white-red gradient mapped onto the models. Ternary complex models colored according to Fig. 2. c, Gold standard Fourier shell correlation (GSFSC) of the global refinement cryo-EM map. d, Posterior precision directional distribution of the 3D model after global refinement. e-f, GSFSC of the local refinement cryo-EM maps derived from 565k particles (e) and 287k particles (f), respectively.
Extended Data Fig. 4 Cryo-EM 3D reconstruction workflow. a, CryoSPARC v4.2.1 was used for image analysis, 2D classification and 3D reconstruction. Multiple rounds of 2D classification were performed to remove particles not representing the ternary complex. b, 3D reconstruction, classification, and global/local refinement workflows. Intermediate local resolution estimations are depicted by a blue-white-red gradient mapped onto the models. Ternary complex models colored according to Fig. 2. c, Gold standard Fourier shell correlation (GSFSC) of the global refinement cryo-EM map. d, Posterior precision directional distribution of the 3D model after global refinement. e-f, GSFSC of the local refinement cryo-EM maps derived from 565k particles (e) and 287k particles (f), respectively.CC BY-NC-ND · PubMed Central
Extended Data Fig. 5 Representative regions of the cryo-EM envelope with the atomic model. a, Ternary complex structure composed of DDB1ΔBPB/CRBNΔN40, MRT-5702 and G3BP2NTF2L. b, Global cryo-EM map. c, Local refinement cryo-EM map. d, Local refinement cryo-EM map post-processed with DeepEMhancer58. e, Model-density matches between regions of DDB1 and the global cryo-EM map. f, Model-density matches between regions of CRBN and the global cryo-EM map. g, Model-density matches between regions of the ternary complex interface and the three different maps. Arrows indicate areas with improved clarity in the local refinement maps. h, Local refinement map at the interface of the CRBN Trp-cage and helix α1 of the interacting protomer of G3BP2NTF2L. The compound is not modeled into the map volume. i, Map volume attributed to MRT-5702 at the interface of the interacting protomer of G3BP2NTF2L.
Extended Data Fig. 5 Representative regions of the cryo-EM envelope with the atomic model. a, Ternary complex structure composed of DDB1ΔBPB/CRBNΔN40, MRT-5702 and G3BP2NTF2L. b, Global cryo-EM map. c, Local refinement cryo-EM map. d, Local refinement cryo-EM map post-processed with DeepEMhancer58. e, Model-density matches between regions of DDB1 and the global cryo-EM map. f, Model-density matches between regions of CRBN and the global cryo-EM map. g, Model-density matches between regions of the ternary complex interface and the three different maps. Arrows indicate areas with improved clarity in the local refinement maps. h, Local refinement map at the interface of the CRBN Trp-cage and helix α1 of the interacting protomer of G3BP2NTF2L. The compound is not modeled into the map volume. i, Map volume attributed to MRT-5702 at the interface of the interacting protomer of G3BP2NTF2L.CC BY-NC-ND · PubMed Central
Extended Data Fig. 6 Characterization of MRT-5702 through LC-MS, supercritical fluid chromatography (SFC) and NMR. a, LC-MS chromatograms and m/z mass spectrum. b, SFC chromatogram. c, 1H NMR (d6-DMSO) of MRT-5702. Chiral centers of MRT-5702 are indicated with asterisks in the chemical structure.
Extended Data Fig. 6 Characterization of MRT-5702 through LC-MS, supercritical fluid chromatography (SFC) and NMR. a, LC-MS chromatograms and m/z mass spectrum. b, SFC chromatogram. c, 1H NMR (d6-DMSO) of MRT-5702. Chiral centers of MRT-5702 are indicated with asterisks in the chemical structure.CC BY-NC-ND · PubMed Central
Extended Data Fig. 7 SCF-separated stereoisomers of MRT-5702. a-d, SFC chromatogram and 1H NMR (d6-DMSO) of MRT-5702A (a), MRT-5702B (b), MRT-5702C (c), and MRT-5702D (d).
Extended Data Fig. 7 SCF-separated stereoisomers of MRT-5702. a-d, SFC chromatogram and 1H NMR (d6-DMSO) of MRT-5702A (a), MRT-5702B (b), MRT-5702C (c), and MRT-5702D (d).CC BY-NC-ND · PubMed Central
Extended Data Fig. 8 MRT-5702 stereoisomers. a-d, Chemical structures of the different configurations of MRT-5702 (top) and corresponding low energy conformers of each ligand configuration producing best overlaps (SMOC score) with the compound volume in the cryo-EM envelope (bottom). (R,S) configuration in (a), (S,S) configuration in (b), (R,R) configuration in (c) and (S,R) configuration in (d). e, Conformation of MRT-5702D in the final ternary complex model and its match to the cryo-EM envelope of the local refinement map in two different orientations.
Extended Data Fig. 8 MRT-5702 stereoisomers. a-d, Chemical structures of the different configurations of MRT-5702 (top) and corresponding low energy conformers of each ligand configuration producing best overlaps (SMOC score) with the compound volume in the cryo-EM envelope (bottom). (R,S) configuration in (a), (S,S) configuration in (b), (R,R) configuration in (c) and (S,R) configuration in (d). e, Conformation of MRT-5702D in the final ternary complex model and its match to the cryo-EM envelope of the local refinement map in two different orientations.CC BY-NC-ND · PubMed Central
Extended Data Fig. 9 Model and structure of the ternary complex interface. a, Comparison of the G3BP1-USP10 structure (PDB: 7xhf) and the generated G3BP2-USP10 model. b, Comparison of the mimicry-guided docking model and the CRBN-MRT-5702-G3BP2NTF2L cryo-EM structure. Differences in the pose of the interacting G3BP2NTF2L protomer are highlighted. c-f, Model-density matches of different sections of the ternary complex interface. CRBN LON-domain in (c), CRBN LON-loop in (d), MRT-5702D in (e) and CRBN sensor-loop in (f).
Extended Data Fig. 9 Model and structure of the ternary complex interface. a, Comparison of the G3BP1-USP10 structure (PDB: 7xhf) and the generated G3BP2-USP10 model. b, Comparison of the mimicry-guided docking model and the CRBN-MRT-5702-G3BP2NTF2L cryo-EM structure. Differences in the pose of the interacting G3BP2NTF2L protomer are highlighted. c-f, Model-density matches of different sections of the ternary complex interface. CRBN LON-domain in (c), CRBN LON-loop in (d), MRT-5702D in (e) and CRBN sensor-loop in (f).CC BY-NC-ND · PubMed Central
Extended Data Fig. 10 Total buried molecular surface area in MGD-induced ternary complexes with CRBN. a, Total buried molecular surface area (footprint) within ternary complexes represented by the footprints of MGD-dependent neosubstartes mapped onto the respective CRBN/MGD neosurfaces. The footprints are colored according to electrostatics of the CRBN region involved in the ternary complex interface. Pink contour indicates the surface area contributed by the MGD. b, Quantification of the footprints and its relative distribution over the LON and CULT domains, and the MGD.
Extended Data Fig. 10 Total buried molecular surface area in MGD-induced ternary complexes with CRBN. a, Total buried molecular surface area (footprint) within ternary complexes represented by the footprints of MGD-dependent neosubstartes mapped onto the respective CRBN/MGD neosurfaces. The footprints are colored according to electrostatics of the CRBN region involved in the ternary complex interface. Pink contour indicates the surface area contributed by the MGD. b, Quantification of the footprints and its relative distribution over the LON and CULT domains, and the MGD.CC BY-NC-ND · PubMed Central

Verification

Claims the paper makes

4 Valid · 1 Not valid · 1 Not checked · 36 Own results and interpretations

14 more claims

Create a free account to see all 42 claims and to verify the 1 reference still open.

See all claims — free account
  • It may state two separate claims as one. 16
  • Probably not a substantial claim. 10

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 59 references

Cited by

See all 9

Metadata

Authors
Annunziato, S., Quan, C., Donckele, E. J., Lamberto, I., Bunker, R. D., Zlotosch, M., Schwander, L., Murthy, A., Wiedmer, L., Staehly, C., Matysik, M., Gilberto, S., Kapsitidou, D., Wible, D., De Donatis, G. M., Trenh, P., SriRamaratnam, R., Strande, V., Lieberherr, R., Lyon, D., Steiner, D., Silva, J., Almeida, R., Dolgikh, E., DeMarco, B., Tsai, J., Sadok, A., Zarayskiy, V., Walter, M., Tiedt, R., Lumb, K. J., Bonenfant, D., Fasching, B., Castle, J. C., Townson, S. A., Gainza, P., & Petzold, G.
Affiliations
Boston Foundation (US)
Corresponding author
Petzold, Georg
Venue
Nature Structural & Molecular Biology
Published
2026-01-20
Volume
33
Issue
3
Pages
479-487
ISSN
  • 1545-9985 (ISSN-L)
  • 1545-9993
OpenAlex
W7124843030
Cited by
9
References
59
Citations by year
2026: 9
FWCI
18.39
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Structural & Molecular Biology (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein Degradation and Inhibitors
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Ubiquitin and proteasome pathways
  • Physical Sciences › Chemistry › Organic Chemistry › Chemical Synthesis and Analysis
Keywords
cereblon, molecular glues, G3BP2, E3 ubiquitin ligase, protein degradation
MeSH terms
Degrons, Binding Sites, Carrier Proteins, Humans, Models, Molecular, Protein Binding, Molecular Mimicry, Ubiquitin-Protein Ligases, Adaptor Proteins, Signal Transducing, Proteolysis

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Cereblon induces G3BP2 neosubstrate degradation using molecular surface mimicry. https://app.cochat.ai/papers/10.1038/s41594-025-01738-8
The paper
Annunziato, S., Quan, C., Donckele, E. J., Lamberto, I., Bunker, R. D., Zlotosch, M., Schwander, L., Murthy, A., Wiedmer, L., Staehly, C., Matysik, M., Gilberto, S., Kapsitidou, D., Wible, D., De Donatis, G. M., Trenh, P., SriRamaratnam, R., Strande, V., Lieberherr, R., . . . Petzold, G. (2026). Cereblon induces G3BP2 neosubstrate degradation using molecular surface mimicry. Nature Structural & Molecular Biology. https://doi.org/10.1038/s41594-025-01738-8

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.