Bioorthogonal Photocatalytic Protein Labeling and Cross-Linking Enabled by Stabilized Ketyl Radicals

Abstract

Radical reactions offer transformative potential in biological contexts but remain constrained by poor selectivity and off-target reactivity. We address these limitations through visible-light photocatalytic generation of diaryl ketyl radicals from benzophenones. This strategy circumvents traditional UV excitation pathways by suppressing triplet diradical formation─which drives nonspecific [2 + 2] cycloadditions and H atom abstraction─in favor of bioorthogonal radical-radical coupling. Our platform enables precise live-cell protein labeling with minimal cytotoxicity, including in sensitive primary neuronal cultures, and achieves site-specific modification via genetically incorporated benzophenone-based unnatural amino acids Bpa. The spatial selectivity of this approach exceeds conventional UV-based cross-linking methods, facilitating site-to-site analysis of tertiary protein interactions in structurally defined complexes. We demonstrate these capabilities by (1) quantifying dimerization interfaces of the Diels-Alderase PyrI4 and (2) resolving Bcl-X_L/Bid interactions critical for apoptotic regulation. This photocatalysis-driven methodology establishes a robust alternative to cycloaddition-based bioorthogonal chemistry for spatiotemporally controlled interrogation of dynamic biomolecular processes.

1

Bioorthogonal protein labeling via photocatalytic ketyl radicals. (A) Conventional methods use highly reactive radicals causing nonspecific modifications (top) versus our stabilized ketyl radicals enabling selective bioorthogonal coupling (bottom). (B) UV photoexcitation generates reactive triplet diradicals that undergo uncontrolled H-abstraction/[2 + 2] cycloadditions (red). Visible-light photocatalysis produces stabilized diaryl ketyl radicals via steric/resonance effects, directing selective radical–radical coupling (blue). (C) Benzophenone-functionalized proteins undergo site-specific labeling and cross-linking through photocatalytic ketyl radical homocoupling.

2

Bioorthogonal ketyl radical coupling in aqueous media. (A) Screening of photocatalyst (10 μM) and reductant (10 mM). Yields were determined by HPLC quantification. Standard (std) conditions: 100 μM benzophenone 1, 10 μM Ru­(bpy)₃Cl₂, 10 mM VcH, 468 nm light irradiation (4.5 mW/cm²) for 10 min in PBS pH 7.4/MeCN (1:1). (B) Radical trapping experiments. Benzophenone 1 (0.1 mmol), VcH (0.15 mmol), Ru­(bpy)₃Cl₂ (0.002 mmol), and 4-cyanopyridine (0.3 mmol) in pH 7.4 buffered PBS/MeCN (1:1) under 468 nm light irradiation (20 mW/cm²) for 12 h under a nitrogen atmosphere at room temperature. (C) Reactivity comparison: photocatalytic ketyl radicals versus UV-induced triplet diradicals. Left: HRMS detection confirms the formation of signature [2 + 2] cycloaddition products from reactions between benzophenone (1 mM) and dThd or vinyl ether (1 mM) under UV light (365 nm). Right: Quantitative comparison of alkene consumption. HPLC analysis shows >95% of alkene substrates remain unreacted under photocatalytic conditions (468 nm) versus significant consumption under UV photoexcitation (365 nm), demonstrating high selectivity. Reaction conditions: Photocatalysis: Benzophenone (1 mM), VcH (10 mM), Ru­(bpy)₃Cl₂ (10 μM) in MeCN/PBS (pH 7.4, 1:1 v/v), 468 nm light (4.5 mW/cm², 10 min). Photoexcitation: Benzophenone (1 mM) with alkene substrates in MeCN/PBS, 365 nm UV light (0.2 mW/cm², 10 min).

3

Site-specific labeling of proteins via photocatalytic ketyl radical coupling. (A) Schematic illustration and chemical structures of the benzophenone-based probes (e.g., Biotin-BP 6) used for photocatalytic radical–radical coupling. (B) Analysis of protein biotinylation efficiency. Western blot analysis confirms successful photocatalytic biotinylation of Bcl-X_L protein. Comparison of reactions with and without the benzophenone precursor (SulfoNHS-BP 5) demonstrates target-specific labeling. Reaction conditions: Bcl-X_L (10 μM), SulfoNHS-BP 5 (100 μM), Ru­(bpy)₃Cl₂ (10 μM), VcH (20 mM), Biotin-BP 6 (100 μM) in PBS (pH 7.4), irradiated with 468 nm light (8.2 mW/cm², 10 min). (C) Live-cell membrane labeling. Confocal microscopy and flow cytometry analysis of HeLa cells show efficient cell-surface biotinylation dependent on photocatalytic labeling. Procedure: Cells were incubated with SulfoNHS-BP 5 (200 μM, 37 °C, 30 min), followed by treatment with Ru­(bpy)₃Cl₂ (10 μM), VcH (20 mM), and Biotin-BP 6 (200 μM) in PBS and irradiation with 468 nm light (14.4 mW/cm², 15 min, RT). Cells were then fixed and stained with streptavidin-TAMRA (imaging) or streptavidin-Cy5 (flow cytometry). Scale bar: 50 μm. (D) Orthogonality and biocompatibility of photocatalytic labeling in primary neurons. Left: Schematic of the labeling strategy. Middle: Confocal microscopy shows specific biotinylation (red) colocalized with neuronal markers (green), demonstrating precise targeting. Right: Quantitative MTT assay reveals high cell viability (>84%) under standard conditions.

4

Photocatalytic benzophenone coupling for site-specific protein modification by genetic code engineering with unnatural amino acid Bpa. (A) Comparison of photocatalytic site-specific protein labeling and UV-induced nonspecific labeling. Proteins (Bcl-X_L-136Bpa/WT Bcl-X_L) were incubated and irradiated under two conditions: Photocatalytic group: 468 nm light (4.5 mW/cm², 10 min) with the addition of 100 μM Biotin-BP VcH (20 mM) and Ru­(bpy)₃Cl₂ (10 μM); UV photoexcitation group: 365 nm UV light (0.2 mW/cm², 10 min) directly with the addition of 100 μM Biotin-BP. (B) Benzophenone modified with various imaging and enrichment moieties using in-gel fluorescence and Western blot separately. Reaction conditions: 10 μM Bcl-X_L-136Bpa, 10 μM Ru­(bpy)₃Cl₂, 20 mM VcH, and 100 μM moieties 7–12 in PBS (pH 7.4) irradiated with 468 nm blue light (8.2 mW/cm²) at room temperature for 10 min. (C) Photocatalytic labeling demonstrates specific membrane targeting versus UV-induced nonspecific background. Top: Schematic of the labeling strategy: EGFR-EGFP-expressing 293T cells were treated with the nanobody 7D12–76Bpa and subsequently labeled with Biotin-BP 6 under uniform illumination. Middle: Photocatalytic labeling (468 nm) shows specific colocalization of EGFP (green, target) and SA-TAMRA (red, label) signals after fixation and staining, demonstrating precise membrane targeting. Bottom: Photolytic labeling (365 nm) results in widespread nonspecific background labeling without target colocalization, highlighting the superior specificity of the photocatalytic approach.

5

Photocatalytic benzophenone coupling to explore protein–protein interactions at atomic-level precision. (A) Schematic illustration of protein–protein cross-linking. Reaction conditions: 10 μM genetically encoded protein with Bpa, 10 μM Ru­(bpy)₃Cl₂, 20 mM VcH, 468 nm light irradiation in PBS buffer (pH 7.4) for 10 min. (B) Homodimeric cross-linking of PyrI4–138Bpa. (C) Heterodimeric cross-linking of Trx-62Bpa and PAPS-191Bpa. (D) Heterodimeric cross-linking of Bcl-X_L-126Bpa and Bid-83Bpa. (E) Proximity cross-linking of Bid-83Bpa Bcl-X_L-126Bpa/171Bpa. (F) Binding free energy of Bcl-X_L and Bid by molecular dynamics simulation. (G) K _D of Bcl-X_L-126Bpa and Bid-83Bpa by bio-layer interferometry (BLl).

6

Orthogonal analysis of the ketyl radical coupling with cycloaddition reactions. Bcl-X_L-136Bpa (10 μM), GST-TCO (10 μM), and Oval-DBCO (10 μM) reacted with Biotin-BP 6 (100 μM), Biotin-tetrazine 13 (100 μM), and Biotin-azide 14 (100 μM) under standard photocatalytic conditions (10 μM Ru­(bpy)₃Cl₂, 20 mM VcH, 468 nm light irradiation for 10 min). The results were analyzed by Western blots.