Visible Light Control over the Cytolytic Activity of a Toxic Pore-Forming Protein

Enabling control over the bioactivity of proteins with light, along with the principles of photopharmacology, has the potential to generate safe and targeted medical treatments. Installing light sensitivity in a protein can be achieved through its covalent modification with a molecular photoswitch. The general challenge in this approach is the need for the use of low energy visible light for the regulation of bioactivity. In this study, we report visible light control over the cytolytic activity of a protein. A water-soluble visible-light-operated tetra-ortho-fluoro-azobenzene photoswitch was synthesized by utilizing the nucleophilic aromatic substitution reaction for installing a solubilizing sulfonate group onto the electron-poor photoswitch structure. The azobenzene was attached to two cysteine mutants of the pore-forming protein fragaceatoxin C (FraC), and their respective activities were evaluated on red blood cells. For both mutants, the green-light-irradiated sample, containing predominantly the cis-azobenzene isomer, was more active compared to the blue-light-irradiated sample. Ultimately, the same modulation of the cytolytic activity pattern was observed toward a hypopharyngeal squamous cell carcinoma. These results constitute the first case of using low energy visible light to control the biological activity of a toxic protein.


Experimental Section 1.General information
All chemicals for synthesis were obtained from commercial sources and used as received unless stated otherwise.Technical grade solvents were used for extraction and chromatography.Thin Layer Chromatography (TLC) was performed using commercial Kiesegel 60 F254 silica gel plates with fluorescence-indicator UV254 (Merck, TLC silica gel 60 F254).For detection of components, UV light at λ=254 nm or λ=365 nm was used.Alternatively, oxidative staining was performed using a basic solution of potassium permanganate in water or aqueous cerium phosphomolybdic acid solution (Seebach's stain).Merck silica gel 60 (230-400 mesh ASTM) was used in normal phase flash chromatography.Büchi Reveleris® X2 automatic column was used with Büchi EcoFlex silica columns (4 -40 g, 40−63 μM, 60 Å).Spectroscopic measurements were made in Uvasol® grade solvents using a quartz cuvette (path length 10.0 mm).UV-Vis measurements were performed on an Agilent 8453 UV-Visible absorption Spectrophotometer.UV-Vis irradiation experiments were carried out using a custom-built (Prizmatix/Mountain Photonics) multi-wavelength fiber coupled LED-system (FC6-LED-WL) with LED lights (425A and 530B), 530 nm and 430 nm LED light source (3x 530 nm, 3x 430 nm, LED Nichia NCSB219B-V1, Sahlmann Photochemical Solutions).The temperature was controlled with a Quantum Northwest TC1 temperature controller.The data was processed using Agilent UV-Vis ChemStation B.02.01 SP1, Spectragryph 1.2, OriginPro 2016 and all images were assembled in Adobe Illustrator.NMR spectra were obtained using Agilent Technologies 400-MR (400/54 Premium Shielded) (400 MHz) and Bruker Innova (1H: 600 MHz, 13C: 151 MHz) spectrometers at room temperature (22-24 °C).Chemical shift values (δ) are reported in parts per million (ppm) with the solvent resonance as the internal standard (CDCl3: δ 7.26 for 1 H, δ 77.16 for 13 C; DMSO: δ 2.05 for 1 H, δ 39.52 for 13 C, CD3OD: δ 3.31 for 1 H, δ 49.0 for 13 C).The following abbreviations are used to indicate signal multiplicity: s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), brs (broad signal) or dd (doublet of doublets).Exact mass spectra were recorded on an LTQ Orbitrap XL (ESI+).All reactions requiring an inert atmosphere were carried out under a nitrogen atmosphere using oven dried glassware and standard Schlenk techniques.Dichloromethane and toluene were used from solvent purification system using an MBraun SPS-800 column.Melting points were determined using Stuart analogue capillary melting point SMP11 apparatus.All errors are given as standard deviations.NB.For all sulfonated compounds, purification proved to be challenging and tedious, requiring reversed phase chromatography using mildly basic buffer/ACN mixtures as eluents.The buffer used, ammonium hydrogencarbonate, was chosen for easy removal of the salt via freeze-drying.Furthermore, decomposition of the photoswitches was observed upon removal of water on the rotary evaporator at higher temperatures, thus requiring complete removal of the solvent exclusively by a freeze-drying procedure.

Sulfonation reaction conditions
We conducted several qualitative test reactions where we looked for formation of water-soluble products during the SNAr reaction of non-water soluble azobenzene 4 with sodium sulfite at different reaction conditions.If the sulfonation occurs, sulfonated azobenzene eluted on reversed phase silica.However, since over-sulfonation can take place and we have previously observed formation of complex mixtures of water-soluble products, in these experiments we specifically aimed to optimize the reaction conditions to yield the least possible number of water-soluble side products.Therefore, here we qualitatively investigated the number of water-soluble products formed by reversed phase TLC and in some examples determined the conversion.To summarize the results, the use of solvent mixtures containing 40-60% of water and a polar solvent, such as ethanol, ACN or DMF, resulted in less complex water-soluble product mixtures with the highest content of the mono sulfonated product.This is likely due to the optimal solvent mixture required to dissolve both the non-polar, aromatic azobenzene and the highly polar water-soluble sodium sulfite salt.While higher temperatures increased the conversion of the sulfonation reaction, unfortunately they resulted in complex mixtures of multi-sulfonated products and by-products which were extremely difficult to separate.Therefore we opted for conditions yielding the highest amount of the mono-sulfonated compound 5, namely 1/1 = water/ACN at 50 °C (overnight), resulting in minimal formation of multi-sulfonated species, especially since the starting material could easily be recovered by extraction of the crude mixture with DCM.

Haemolytic activity assay
Defibrinated sheep blood (Fisher Scientific) was washed several times with wash buffer (15 mM Tris-HCl pH 7.5, 150 mM NaCl) by centrifugation (8000 rpm for 10 s) until the supernatant was clear.The washed blood was resuspended and diluted in wash buffer to an OD650 of 0.8-0.9.The activity assays were performed in a dark room under red light.The diluted blood (100 μL) was added to a 96-well plate containing several solutions with different concentrations of toxin (labelled FraC monomers) in either cis or trans-state.Immediately after adding the red blood cells suspension, the haemolytic activity was measured by monitoring the decrease in OD650 using the Synergy H1 Hybrid-Multimode reader (BioTek).Percentage of haemolysis was calculated as followed for a specific time point: % hemolysis = 100*(AbsC-FraC -Absbuffer)/(Abstriton -Absbuffer).
where AbsC-FraC is the absorbance measured for the blood sample were the labelled FraC monomers were added, Absbuffer is the absorbance measured for the blood sample were only buffer was added (negative control) and Abstriton is the absorbance measured for the blood sample were Triton, a strong detergent which causes complete lysis of blood cells, was added (positive control).Photoisomerization of the visible light switch to the cis-state was achieved by irradiation for 30 min with the 530 nm wavelength LED and to the trans-state by irradiation for 30 min with the 430 nm wavelength LED.
For the reversibility test, the same sample of labelled FraC monomers was consecutively irradiated with 430 nm and 530 nm lamps for 30 min, while removing three aliquots after each irradiation step.The aliquoted samples were measured in the haemolytic activity assay.

Figure S1 .
Figure S1.Scheme representing the attachment reaction of visible light-responsive switch A to the thiol group of cysteine in the water-soluble conformation of the FraC monomer.The alpha helix which gets extended into the lipid bilayer is highlighted in orange (PDB 4TSP 2 ).

Figure S15 .
Figure S15. 1 H-NMR of compound 5 (with impurities present) in CD3CN with a drop DCl.

Figure S20 .
Figure S20.LCMS traces of isolated compound 5 measured in negative mode (Conditions: 0.1% NH3 in water and 0.1% NH3 in ACN as eluents, HSST3 C18 column).A) Measured absorbance at 365 nm and the total ion current (TIC) measured in negative mode.B) Measured m/z of observed peaks at noted retention times and the zoom in of the observed m/z (C) to visualize the isotope pattern due to the presence of bromine.

Figure S32 .
Figure S32.LCMS traces of isolated azobenzene A measured in negative mode (Conditions: 0.1% NH3 in water and 0.1% NH3 in ACN as eluents, HSST3 C18 column).A) Measured absorbance at 365 nm and the total ion current (TIC) measured in negative mode.B) Measured m/z of observed peaks at noted retention time.