Light-mediated multi-target protein degradation using arylazopyrazole photoswitchable PROTACs (AP-PROTACs)

Light-activable spatiotemporal control of PROTAC-induced protein degradation was achieved with novel arylazopyrazole photoswitchable PROTACs (AP-PROTACs). The use of a promiscuous kinase inhibitor in the design enables this unique photoswitchable PROTAC to selectively degrade four protein kinases together with on/off optical control using different wavelengths of light.


General methods
All reagents were purchased from commercial sources (Sigma-Aldrich, Merck, Fluorochem UK) and were used without further purification. Pan-BET bromodomain inhibitor JQ1-COOH were provided by GlaxoSmithKline Medicines Research Centre, Stevenage. Multikinase inhibitor CTx-0294885 were synthesised with flow chemistry approaches as previously described. 1 Analytical thin-layer chromatography (TLC) on aluminium sheets with silica gel 60 F254 (Merck) were used and visualized with UV light (254 nm) or appropriate TLC stain to monitor reactions. Flash chromatography with silica gel Geduran Si 60 (0.040-0.063 mm, Merck) was performed for general compound purification. Nuclear Magnetic Resonance (NMR) spectra were recorded on a BRUKER AV-400 spectrometer at 400 MHz ( 1 H-NMR) and 101 MHz ( 13 C-NMR) in deuterated solvents at 298 K. Chemical shifts are given in parts per million (ppm), coupling constants J are given in Hertz, and spin multiplicities are given as s (singlet), d (doublet), t (triplet), q (quartet) or m (multiplet). Spectra were analysed with MestReNova.

Cell culture and treatment
HeLa cells and MDA-MB-231 cells were obtained from the Francis Crick Institute cell banking. Cells were cultured in low-glucose Dulbecco's Modified Eagle's Medium (DMEM, Sigma, D6046) supplemented with 10% fetal bovine serum (FBS, Gibco) and grown in a humidified incubator at 37 °C, 5% CO2. Cells were plated at 3 × 10 5 cells per well in a 6-well plate, allowed to adhere for 24 h before being treated with the indicated compounds and time, with 1 min irradiation of the cell flask at either 365 nm or 457 nm using LED light every 2 to 3 hours.

Cell Lysis
After cell treatment, cells were washed twice with PBS, added the respective cold lysis buffer and collected into eppendorf. Cells were lysed on ice for 30 min. The collected suspensions were centrifuged for 10 min at 17,000 g, 4 °C, and then the supernatant was collected. The protein concentration of each sample was determined using the DC TM Protein assay kit (BIO-RAD) with a bovine serum albumin standard curve in lysis buffer and normalized. The lysates were used in the subsequent immunoblotting, and proteomics sample preparation steps, or stored at −80 °C.

Proteomics Sample Preparation
The lysates of treated cells (100 μg, 100 μL) were reduced and alkylated by the addition of TCEP (to 5 mM final concentration) and CAA (to 15 mM final concentration) for 45 min at RT under vigorous shaking. Proteins were precipitated by adding 2 sample volumes of methanol, 0.5 volume of chloroform and 1 volume of purified water. The mixture was vortexed and centrifuged at 6,000 rpm f or 2 min resulting in pellets of precipitated proteins. The pellets were washed with 3 sample volumes of methanol by vortexing and sonicating and then centrifuged at 8,000 rpm for 4 min. The supernatant was discarded, and the washing procedure was repeated an additional four times. The protein precipitate was air-dried for 5 min and redissolved in 100 μL of 50 mM HEPES pH 8.0 with sonication and vortexing. The protein solution was digested with trypsin protease (0.4 μg, 1:250 to protein, Pierce™ 90057) and incubated overnight in a thermoshaker with 800 rpm shaking at 37 °C. Trypsin digestion was quenched with the addition of 0.5% TFA. Peptide quantification assay (Pierce Fluorometric Peptide Assay) was performed. The same volume of peptide digest from each condition was aliquoted and dried with a SpeedVac Vacuum concentrator. The dried peptide samples were labelled with tandem mass tags (TMT10plex ref 90110, Thermo Scientific) for 2 h at RT. The TMT labelling reaction was quenched with the addition of 5% hydroxylamine and the samples were vortexed for 5 min and centrifuged at 8,000 rpm for 4 min at rt. An equal volume of supernatant from each condition was combined into a single Eppendorf. The combined peptides were fractionated (Pierce high pH Reversed-Phase Peptide Fractionation kit) into 8 fractions and dried with a SpeedVac Vacuum concentrator. The dried fractions were resuspended in 2% acetonitrile and 0.5% trifluoroacetic acid in water to around 1 μg/μL.

LC-MS/MS and Data Analysis
Prepared peptide sample fractions were run using a Thermofisher Q-Exactive LC-MS/MS equipped with a Thermo EASY-SPRAY column as described previously. [2][3][4] The data obtained were processed with MaxQuant version 1.6.17.0, where peptides from the MS/MS spectra were searched in the human proteome database (UniProt accessed January 2021) and identified. The data were further analysed with Perseus 1.6.15.0, Microsoft Office Excel 365, and GraphPad Prism 9. A minimum of 2 unique peptides were selected for protein identification. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE 5 partner repository with the dataset identifier PXD036224.

Cell toxicity assay
MDA-MD-231 cells were seeded in a 96-well plate (7000 cells/well) and grown in low glucose DMEM with 10% FBS overnight. The next day, the media was replaced with media containing compound and SYTOX™ Green Nucleic Acid Stain (dye, 250 nM final concentration). For control, cells were treated with media containing 0.1% DMSO without dye, 0.1% DMSO with dye, 2 µg/mL puromycin and dye, and 2 µg/mL puromycin in 0.1% DMSO with dye. The cell plate was incubated at 37 °C with 5% CO2 in the incubator. Phase and green fluorescence were imaged every 2 h for 3 days with the Incucyte Live-Cell Analysis Systems.

Supplementary Information: Photoswitching properties characterisation General Methods
UV-Vis spectra were recorded on an Agilent Cary 60 UV-Vis spectrometer (wavelength range: 190 -1100 nm, resolution: 1.5 nm, Light source: Xenon Flash Lamp 80 Hz) with a temperature controller. Polystyrene semi-micro cuvettes (1.5 mL, 1 cm) sealed with parafilm were used in the measurement. The data was analysed using WinUV software. LC-MS spectra were recorded on a Waters high-performance liquid chromatography (HPLC) system, including a 2767 autosampler, 515 pump, 2998 photodiode array (PDA) detector and a 3100-electrospray ionization (ESI) mass spectrometer, using MassLynx 4.1 software. Compounds were separated on a 4.6 mm × 100 mm analytical Waters XBridge C18 column using the following gradient: 20% to 98% acetonitrile in water with 0.1% formic acid over 12 min then 98% acetonitrile for 3 min. Two custom-made black boxes (WaveyTech Ltd) installed with 25 mW LED light bulbs of 340 nm, 365 nm, 405 nm, 450 nm and 457 nm wavelength were used to irradiate compounds.   90% Z 10% E

UV-Visible spectroscopy determination of thermal half-life
A solution of 12 μM AP-PROTAC-2 in 0.1% DMSO in water was irradiated with 365 nm LED for 3 min. Following irradiation, the solution was kept in the UV-Vis spectrophotometer in a 37°C chamber for 24 h. A UV-Vis spectrum was recorded every 30 min during the incubation. It was assumed that the compound achieved the same conversion ratio with irradiation as observed in the LC-MS experiment (80% E with 457 nm and 90% Z with 365 nm irradiation). The percentage of Z isomer was plotted against time ( Figure S2), where an exponential trendline was fitted to the plot: [ ] = [ ] 0 − , y = 0.847 −6.08×10 −6 (R 2 = 0.9707) From the rate constant k, the thermal half-life of the Z isomer could be calculated: t 1/2 = 2 = 31.7 ℎ

Supplementary Information: Computational modelling of PROTAC ternary complex
Our structural modelling approach partitions the PROTAC-complex sampling problem into several steps and starts with a templatebased protein-small-molecule docking protocol ClusPro LigTBM, which orients PROTAC end-ligands with corresponding proteins. 7,8 Each PROTAC is bisected in the middle atom of the linker and conformers of the PROTAC halves have been generated by the ETKDG method and relaxed with MMFF ( Figure S9). 9,10 The resulting half-PROTACs are aligned to the respective protein-ligand complex models and filtered for clashes. The middle atom positions of the remaining conformations have been projected to grids. Next, to find energetically favourable protein-protein complex poses, we use the FFT-based docking program PIPER with an additional "silent" term. [11][12][13] This term convolves the aforementioned grids and ensures that the middle atom on both sides of docking overlaps, thus it partially accounts for PROTAC accessible conformations and helps efficiently filter out unfeasible protein-protein complex poses. The structures where the half-PROTACs can be successfully connected into a complete PROTAC are relaxed by Amber energy minimization and clustered to produce complete ternary PROTAC complex models. 14 A pdb file of the representative ternary complex pose shown in the main text was provided as a separate file.

Figure 2D
Rep