Multi-dimensional high-throughput molecular glue screening via gas phase affinity selection native mass spectrometry and cryo-EM analysis

Significance of the topic

Targeted protein degradation using molecular glues (MGs) is a paradigm-shifting approach in drug discovery because it can induce degradation of traditionally undruggable proteins by stabilizing weak interactions between a target protein and an E3 ubiquitin ligase. Rapid, reliable identification and validation of MGs is therefore critical for both academic research and pharmaceutical development. Native mass spectrometry (MS) combined with structural methods such as cryo-EM provides a direct route to detect intact ternary complexes (E3–MG–target) and to validate binding modes, but widespread adoption has been hampered by low throughput and sensitivity to weak binders. This work demonstrates an integrated, high-throughput workflow that combines online buffer exchange native MS, gas-phase affinity selection, multiplexed compound handling and cryo-EM to accelerate MG discovery and characterization.

Objectives and overview of the study

The primary goals were to (1) increase the throughput of native-MS-based molecular glue screening, (2) enable identification of bound small molecules (including weak binders) by combining gas-phase ligand release and tandem MS, and (3) provide orthogonal structural validation using cryo-EM. The study focused on screening for MGs that stabilize interaction between the CRBN–DDB1 E3 ligase complex and the kinase target WEE1. The workflow was evaluated in both single-compound and multiplexed screening formats and benchmarked for speed, sensitivity to weak binders, and the ability to assign bound ligands unambiguously.

Materials and methods

Key experimental elements and workflow components described in the study:

• Proteins and compounds: CRBN–DDB1 complex and WEE1 were provided by Dana-Farber; compound libraries were prepared as 10 mM stocks and diluted into 200 mM ammonium acetate (AmAc) for native MS.
• Mixtures and concentrations: Samples were typically prepared with CRBN–DDB1 and WEE1 mixed at a 1:1 volume ratio to achieve ~1:2 molar ratio in the final mix. Reported working concentrations were on the order of low micromolar for proteins (CRBN–DDB1 ≈ 3.3 µM; WEE1 ≈ 6.6 µM) and low-to-mid micromolar for compounds (typical per-compound ≈ 5 µM; some conditions used higher compound excess).
• Incubation conditions: Two binding regimes were compared—long incubation (25 °C, 1 h) and rapid incubation (25 °C, 5 min) to assess kinetics and practical throughput.
• Online buffer exchange (OBE) and LC-native MS: Rapid OBE (~0.7 min per run) was used to remove non-volatile buffers prior to native MS, followed by a short MS acquisition (~1.5 min per run).
• Screening formats: Single-compound-per-well (96 compounds = 96 injections) and multiplexed (up to 4 compounds/well; 96 compounds = 24 injections) formats were tested to optimize throughput while minimizing inter-compound competition.
• Gas-phase affinity selection MS (workflow): (1) full Orbitrap native MS injection; (2) isolation of the intact target complex in Q1; (3) controlled dissociation of the complex in the higher-energy collision cell (FHCD) to liberate bound ligands; (4) detection of dissociated complexes and released ligands in the Orbitrap; (5) MS/MS fragmentation of released ligands in the ion trap for structural information.

Used instrumentation

The principal mass spectrometer platform used was a Thermo Scientific Orbitrap Ascend Structural Biology Tribrid mass spectrometer configured for native MS and top-down experiments. Relevant capabilities and settings referenced include:

• m/z range detection up to m/z 16,000 with quadrupole isolation up to m/z 8,000
• Native MS-compatible ion optics and online buffer exchange (OBE) front end
• Fragmentation modes available: PTCR, HCD/FHCD, CID, ETD, and UVPD
• Integration with autosampler, fraction collector, UV detector and an Easy-spray interface for LC-native-MS workflows
• Data handling using ProSight Native for native/top-down analysis and deconvolution

Results and discussion

Throughput and formats:

• The integrated LC-native MS approach with rapid OBE and short MS acquisitions enabled very high throughput screening; the study reports the capacity to analyze more than 2,500 compounds per day under optimized conditions.
• Multiplexing (4 compounds/well) reduced injections for a 96-compound set to 24 injections, while single-compound-per-well preserved maximal sensitivity for weak binders but required more injections.

Sensitivity to weak binders and incubation time:

• Comparisons between 1-hour incubation and 5-minute incubation at 25 °C showed largely comparable hit profiles, indicating rapid formation of many ternary complexes and supporting short incubation times for high-throughput workflows.
• The use of online buffer exchange minimized deleterious matrix effects from non-volatiles, improving detection of intact complexes.

Ligand identification and ambiguity challenges:

• Direct native MS identifies intact E3–target–ligand complexes, but assignment of which compound is bound can be ambiguous in multiplexed wells due to similar compound masses, cofactor presence, and adducts. The study therefore favored one-compound-per-well screening to eliminate competition and simplify assignment when necessary.
• Gas-phase affinity selection MS (isolation of the intact complex followed by dissociation and MS/MS of released ligands) provided a robust route to unambiguously identify bound ligands. Fragmentation spectra collected in the ion trap allowed structural assignment of released small molecules, helping resolve ΔM ambiguities seen in deconvolution of intact masses.

Hit discovery and validation:

• In a demonstration screen of 96 compounds against CRBN–DDB1 and WEE1, 16 candidates were flagged as potential molecular glues under the tested conditions.
• Cryo-EM was applied to characterize ligand-bound complexes and provide orthogonal structural validation of selected hits, confirming binding modes inferred from MS data.

Limitations and practical considerations:

• Non-volatile components in sample buffers can prevent direct native-MS measurements and necessitate rapid and efficient buffer exchange.
• Multiplexing increases throughput but raises the possibility of competitive binding and complicates ligand assignment; gas-phase dissociation and MS/MS mitigate this but add method complexity.
• Deconvolution of native spectra remains sensitive to adducts and cofactors; careful sample preparation and orthogonal validation (MS/MS, cryo-EM) are recommended.

Benefits and practical applications

The integrated approach offers several practical advantages for MG discovery workflows:

• Direct detection of intact ternary complexes reduces false positives that arise in assays that cannot observe the assembled complex.
• Gas-phase affinity selection coupled to MS/MS enables identification and structural interrogation of bound small molecules, even when binding is weak in solution.
• High throughput (enabled by OBE and short acquisition times) allows large libraries to be screened rapidly, supporting early-stage medicinal chemistry and prioritization for structural follow-up.
• Compatibility with cryo-EM provides a pipeline for transitioning from hit identification to mechanistic structural characterization, improving confidence in downstream optimization.

Future trends and potential applications

Expected developments and opportunities building on this workflow include:

• Further automation of sample handling, multiplexing strategies and data analysis pipelines to increase daily compound throughput while maintaining robust ligand assignment.
• Improved informatics for deconvolution and automated MS/MS-based structure annotation of small molecules released from complexes, reducing manual interpretation bottlenecks.
• Adoption of complementary gas-phase fragmentation chemistries (ETD/UVPD/PTCR) tailored to preserve and interrogate labile interactions or to increase structural information on released ligands.
• Broader application to diverse E3 ligases and target classes beyond kinase targets, enabling systematic exploration of MG chemical space across different degradation mechanisms.
• Integration with cell-based degradation assays and proteomics to link in vitro ternary-complex formation to functional degradation outcomes.

Conclusion

This study presents a practical, high-throughput pipeline that combines online buffer exchange native MS, gas-phase affinity selection with MS/MS, multiplexed compound handling, and cryo-EM validation to accelerate molecular glue discovery. The workflow addresses key analytical challenges—detection of weak ternary complexes, unambiguous ligand identification, and throughput—demonstrating the identification of multiple candidate MGs in a 96-compound pilot screen and comparable hit profiles with short incubation times. The approach provides a scalable platform for rapid triage and structural characterization of MG hits, with clear routes for further automation and informatics-driven enhancement.

Reference

1. Liu W., Baek K., Konijnenberg A., Song W., Mullen C., Xiong Y., Durbin K., Bechler S., Fischer E., Viner R., Moehring T. Multi-dimensional high-throughput molecular glue screening via gas phase affinity selection native mass spectrometry and cryo-EM analysis. Thermo Fisher Scientific technical report/poster, 2025.