Activation of Sirtuin 2 Inhibitors Employing Photoswitchable Geometry and Aqueous Solubility

Abstract Because isoenzymes of the experimentally and therapeutically extremely relevant sirtuin family show high similarity, addressing the unique selectivity pocket of sirtuin 2 is a promising strategy towards selective inhibitors. An unrelated approach towards selective inhibition of isoenzymes with varied tissue distribution is targeted drug delivery or spatiotemporal activation by photochemical activation. Azologization of two nicotinamide‐mimicking lead structures was undertaken to combine both approaches and yielded a set of 33 azobenzenes and azopyridines that have been evaluated for their photochemical behaviour and bioactivity. For some compounds, inhibitory activity reached the sub‐micromolar range in their thermodynamically favoured E form and could be decreased by photoisomerization to the metastable Z form. Besides, derivatization with long‐chain fatty acids yielded potent sirtuin 2 inhibitors, featuring another intriguing aspect of azo‐based photoswitches. In these compounds, switching to the Z isomer increased aqueous solubility and thereby enhanced biological activity by up to a factor of 21. The biological activity of two compounds was confirmed by hyperacetylation of sirtuin specific histone proteins in a cell‐based activity assay.


Figure S1
: Setup for UV radiation.
As visible light source 1.5 m of a 11 W•m -1 RGB-LED strip (Paulmann FlexLED 3D Set, 63 LEDs) wound in a hollow glass cylinder (h = 12 cm, ø 10 cm) was applied. The quartz glass cuvette was put on a slice of aluminium foil and the cylinder placed over it for irradiation ( Figure S2).

UV/Vis Spectra and Half-Life of Thermal Isomerization
The half-life of the thermal Z→E isomerization after 5 minutes of 365 nm irradiation was determined by UV/Vis-spectroscopy. Therefore, solutions of the respective compounds (50 µM) in 5%, 50% or 90% DMSO in enzyme assay buffer (v/v) were prepared. An initial spectrum of the thermal equilibrium was recorded. After 5 minutes of irradiation with 365 nm the increase of absorbance at the thermal equilibrium maximum (λ ∆ max) was recorded as exemplified by compound 15c (Figure S3). Absorbance at λ = λ ∆ max was plotted as a function of time. Non-linear regression yielded the half-life of the thermal Z→E isomerization (Table   S1). For extrapolation of the reaction kinetics, the plateau value was set to the initial absorbance at λ=λ ∆ max. The following equation was used for non-linear regression:  before and after 5 minutes of 365 nm and 452 nm irradiation, respectively. Isocratic elution using water/methanol mixtures gave complete separation of the isomers ( Figure S4).
Integration of the peak areas yielded the relative percentages of (E)-and (Z)-photoisomers.  Fluorescence-Based Activity Assay (ZMAL-Assay) The influence of the synthesized compounds on the deacetylase activity of the class I sirtuins Sirt1-3 was determined in a homogenous fluorescence-based enzyme assay, using the artificial substrate Z-(Ac)Lys-AMC (ZMAL). ZMAL is processed by sirtuins yielding Z-Lys-AMC.
The latter is subsequently cleaved by tryptic digestion, releasing 7-amino-4-methylcoumarin (AMC), excitation of which leads to a fluorescent read out (λex=390 nm; λem=460 nm). Instead, non-processed ZMAL, does not serve as a substrate for the tryptic digestion and does not contribute to the fluorescence signal. Therefore, the fluorescence intensity is proportional to the activity of the respective sirtuin.
After subtraction of the blank fluorescence signal (without enzyme) from the signal of the respective probe, the residual signal intensity is related to a 100 % conversion control (AMC), yielding the percentage substrate conversion. Comparison of the determined substrate conversion with a no-inhibition DMSO control yields the percentage inhibition at the respective inhibitor concentration.
For determination of IC50, inhibitory activity of the respective compound was determined at 8 to 10 distinct concentrations. The measured values were analysed by a non-linear fit to the dose-response curve using a four parameter logistic model ( Figure

FP-Based Binding Assay
To get more insight on the binding mode of the compounds 7c, 7f and 9a a competition assay with a SirReal based fluorescent probe that binds to the selectivity pocket of Sirt2 was where Is/IP are the fluorescence intensities measured in the s-and the p-plane respectively, both blank corrected. G (grating factor) is an instrument specific constant. Inhibition was calculated from fluorescence polarization in relation to the maximal polarization and minimal polarization controls and are displayed in % inhibition of binding of the fluorescent probe 21 (Table S1).

Cell-Based Activity Assay
To qualify inhibitory activity of compounds on sirtuins we used the western blot technique.
Therefore, we used precast "Mini-Protean TGX Stain-Free Gels" and corresponding "Trans-Blot Turbo Transfer Pack Mini" PVDF membranes (Biorad, Germany). For sample preparation, we used the urinary bladder cancer cell line RT-4, which showed the highest expression of Sirt2 in our cell line stock (Data not shown), obtained from "Deutsche Sammlung von Mikroorganismen und Zellkulturen" (DSMZ), Germany. Cells were grown in RPMI 1640 cell medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin purchased from PAN Biotec, Germany and stored in a humidified incubator at 37 °C with 5% CO2 atmosphere. For sample preparation, cells were harvested with Trypsin/EDTA and counted in a Beckman's Coulter Counter. In six well plates, 5·10 5 cells were seeded out in 3 mL medium and let allow to attach for 24 h. Stock solutions of compounds (1000-fold) were prepared with DMSO and illuminated with UV light (365 nm) for 5 min. Then, cells were exposed to compounds at a final concentration of 50 µM and incubated for 24 h. Cells were detached with Trypsin/EDTA and washed with phosphate buffered saline (PBS) by centrifugation at 500 g for 5 min and resuspended in PBS. Cell pellets were lysed with a lysis buffer containing Tris 50 mM (pH 7.4), 100 mM NaCl, 100 mM NaF, 5 mM EDTA, 0.2 mM Na3VO4, 0.1% Triton X and freshly added 1% protease inhibitor cocktail (Sigma Aldrich, Germany) on ice, followed by sonification for 30 minutes. The protein concentration was determined by the Bradford method against bovine serum albumin (BSA) as standard. Each slot of the precast gel was loaded with Blots were blocked with 10% non-fat milk powder in Tris-buffered saline containing 20 mM Tris, 145 mM NaCl and Tween 0.5% (TBST) for 2 h and incubated with primary antibody in 1:1000 dilutions (in TBST and 1% BSA) over night at 4 °C. After washing procedure, blots were incubated with secondary antibody conjugated with horse radish peroxidase (1:5000 in TBST and 1% BSA) for 2 h. All used antibodies were obtained from cell signalling (UK): anti ac.H3K9 (#9649), anti ac. H3K56 (#4243), anti H3 (#4499), anti ac. H4K8 (#2594), anti H4 (#13919), anti ac.α-tubulin (#5335), anti α-tubulin (#2125) and anti rabbit-HRP (#7074). Target protein bands were detected with Clarity Western ECL Substrate (Biorad, Germany) and recorded at an Advanced Fluorescence Imager (INTAS, Germany). Band intensity was related to internal standard of the TGX Stain-Free gels as loading control.

Dihedral Analysis
In order to verify the conformations of Z-15c and Z-15f obtained from our molecular docking experiments, a two-dimensional dihedral analysis was performed using MOE with 3-(2phenyldiazenyl)-pyridine as reference compound. Both dihedrals were rotated by 360° in 5° steps. At each step, the potential energy was calculated with AMBER force field parameters and AM1-BCC partial charges applied. The diazenyl double bond of both compounds is nearly perfectly planar with values of 2.5° ((Z)-15c) and 2.1° ((Z)-15f), respectively. All values were truncated to 10 kcal/mol after subtracting the value of the lowest energy conformation.