Small molecule photocatalysis enables drug target identification via energy transfer

Abstract

Over half of new therapeutic approaches fail in clinical trials due to a lack of target validation. As such, the development of new methods to improve and accelerate the identification of cellular targets, broadly known as target ID, remains a fundamental goal in drug discovery. While advances in sequencing and mass spectrometry technologies have revolutionized drug target ID in recent decades, the corresponding chemical-based approaches have not changed in over 50 y. Consigned to outdated stoichiometric activation modes, modern target ID campaigns are regularly confounded by poor signal-to-noise resulting from limited receptor occupancy and low crosslinking yields, especially when targeting low abundance membrane proteins or multiple protein target engagement. Here, we describe a broadly general platform for photocatalytic small molecule target ID, which is founded upon the catalytic amplification of target-tag crosslinking through the continuous generation of high-energy carbene intermediates via visible light-mediated Dexter energy transfer. By decoupling the reactive warhead tag from the small molecule ligand, catalytic signal amplification results in unprecedented levels of target enrichment, enabling the quantitative target and off target ID of several drugs including (+)-JQ1, paclitaxel (Taxol), dasatinib (Sprycel), as well as two G-protein-coupled receptors-ADORA2A and GPR40.

Keywords: photocatalysis; proteomics; target identification.

Fig. 1.

Photoaffinity labeling comprises a critical component of small molecule target ID. (A) Target ID campaigns are critical for the development of successful drugs, although they often rely on challenging photoaffinity labeling campaigns that employ the stoichiometric activation of diazirine small-molecule conjugates with UV light. (B) Our approach separates the warhead from the small molecule probe, instead employing the photocatalytic activation of diazirines using visible light, giving rise to significant signal enhancement. (C) Development of cell-penetrating, generation 2 photocatalyst suitable for small-molecule conjugation and target ID. (D) Cell permeability of Ir-photocatalysts determined by HaloTag chaser assay; photocatalyst PEG-hexyl chloride conjugate and TAMRA hexyl chloride incubated with HEK293T cells expressing TOM20-HaloTag. Western blot analysis and immunostaining TOM20 with anti-TAMRA reveals off-compete only in the presence of Ir-G2 catalyst.

Fig. 2.

Development of photocatalytic target ID platform for interactome mapping of (+)-JQ1 in HeLa cells. Structure of JQ1-based photocatalyst conjugate and state-of-the-art PAL probe. (A) Labeling of recombinant BRD4 protein vs. spectator protein carbonic anhydrase using free iridium-, (+)-JQ1-, and (–)-JQ1-probe by immunostaining with streptavidin. (B) Comparing permeability of G1- and G2-based (+)-JQ1 probes following irradiation in HeLa cells, streptavidin bead enrichment, and immunostaining with anti-BRD4. (C) BRD4 labeling increases over time (2-min, 5-min, and 15-min irradiation) through photocatalytic signal amplification using (+)-JQ1-G2 probe in HeLa cells following streptavidin bead enrichment and immunostaining with anti-BRD4. (D) TMT-based quantitative chemoproteomic analysis of JQ1-labeling in HeLa cells comparing intracellular labeling by (+)-JQ1-G2 catalysts and unconjugated iridium catalyst (control) reveals BRD proteins as highly enriched in addition to known JQ1 off-targets. (E) TMT-based quantitative chemoproteomic analysis of JQ1-labeling in HeLa cells comparing intracellular labeling by (+)-JQ1-G2 catalysts and unconjugated iridium catalyst + (+)-JQ1 (control). (F) TMT-based quantitative chemoproteomic analysis of JQ1-labeling in HeLa cells comparing intracellular labeling between active (+)-JQ1-G2 and inactive (–)-JQ1-G2 catalysts reveals only (+)-isomer labels BRD proteins. (G) State-of-the-art PAL employing active (+)-JQ1- and inactive (–)-JQ1-Dz-alkyne probes reveals no selective enrichment of BRD4 by Western blotting despite broad biotinylation visible by immunostaining with streptavidin. (H) TMT-based quantitative chemoproteomic analysis in HeLa cells comparing state-of-the-art PAL employing active (+)-JQ1- and inactive (–)-JQ1-Dz-alkyne probes reveals no enrichment of BRD proteins.

Fig. 3.

Intracellular photocatalytic target ID and interactome mapping of dasatinib and paclitaxel. (A) Enrichment of p38 by Western blot for labeling using desHEP-dasatinib-PEG5-G2 labeling in THP1 cells. (B) Label-free proteomic analysis in THP1 cells comparing intracellular labeling by desHEP-dasatinib-PEG5-G2 catalyst vs. Ir-G2-NHEt reveals enrichment of several kinases (red), as well as lysosomal proteins (green) and off-targets (blue). (C) Kinase activity assays reveals dasatinib-G2 retains inhibition activity against Abl and p38, as well as general tyrosine phosphorylation, in K562 cells. (D) TMT-based quantitative chemoproteomic analysis in K562 cells comparing intracellular labeling by dasatinib-G2 catalyst vs. dasatinib-G2 + dasatinib (off-compete control) reveals enrichment of several kinases (red), as well as lysosomal proteins (green) and established off-targets (blue). (E) TMT-based quantitative chemoproteomic analysis in K562 cells comparing intracellular labeling by dasatinib-Dz-alkyne (PAL probe) vs. off-compete control does not reveal enrichment of kinases suitable for conclusive target ID. (F) Initial Western blot studies for paclitaxel-G2 labeling in MCF7 cells following irradiation and streptavidin bead enrichment reveals significant enrichment of a-tubulin by immunostaining compared to unconjugated iridium and DMSO controls. (G) TMT-based quantitative chemoproteomic analysis in MCF7 cells comparing intracellular labeling by paclitaxel–G2 catalyst and unconjugated iridium catalyst (control) reveals enrichment of several tubulin isoforms.

Fig. 4.

Extracellular photocatalytic target ID of GPCRs GPR40 and ADORA2A. (A) Small molecule iridium conjugate of MK-8666-Ir is prepared by Cu-click reaction and localized to the target protein GPR40 (FFAR1) via a ligand binding event on the cell surface of GPR40-expressing HEK293T cells. Irradiation and energy transfer to the diazirine leads to selective labeling of the target protein (PDB: 5TZR). (B) Comparative analysis of labeling of GPR40 through classical UV-based PAL using MK-8666-diazirne and MK-8666-diazirine-biotin probes vs. MK-8666-diazirine-G1 Iridium conjugate. Analysis by Western blot visualization and staining with streptavidin shows no labeling of the target protein following streptavidin-bead based enrichment. (C) Label-free proteomic analysis of MK-8666-Ir photocatalytic µMap reveals enrichment of the target protein compared to off-compete with the unmodified ligand. (D) TMT-based quantitative chemoproteomic analysis in A2a-expressing HEK293T cells comparing extracellular labeling by SCH58261-Dz-alkyne vs. SCH58261-Dz-alkyne + SCH58261 (off-compete control) reveals inconclusive target ID of ADORA2A. (E) TMT-based quantitative chemoproteomic analysis in PC-12 cells comparing extracellular labeling by SCH58261-G1 catalyst vs. SCH58261-G1 + SCH58261 (off-compete control) reveals significant enrichment of ADORA2A.