Tandem Ring Opening/Intramolecular [2 + 2] Cycloaddition Reaction for the Synthesis of Cyclobutane Fused Thiazolino-2-Pyridones

Reaction of thiazoline fused 2-pyridones with alkyl halides in the presence of cesium carbonate opens the thiazoline ring via S-alkylation and generates N-alkenyl functionalized 2-pyridones. In the reaction with propargyl bromide, the thiazoline ring opens and subsequently closes via a [2 + 2] cycloaddition between an in situ generated allene and the α,β-unsaturated methyl ester. This method enabled the synthesis of a variety of cyclobutane fused thiazolino-2-pyridones, of which a few analogues inhibit amyloid β1–40 fibril formation. Furthermore, other analogues were able to bind mature α-synuclein and amyloid β1−40 fibrils. Several thiazoline fused 2-pyridones with biological activity tolerate this transformation, which in addition provides an exocyclic alkene as a potential handle for tuning bioactivity.


References S65
Optimization of ring opening reaction: We commenced our investigation with treatment of thiazolino fused 2-pyridone 1a with methyl iodide and K2CO3. The reaction proceeded slowly and cleanly at rt. and led to isolation of 2a in 26% yield (along with 69% of unreacted 1a) in 10 days. Elevation of reaction temp. was considered first for speeding up the transformation, therefore a solution of 1a in DMF was allowed to react with methyl iodide in presence of various bases at 60 °C in a sealed tube. Increasing the temperature from room temperature to 60 °C did speed up the reaction, but it was still incomplete after 5 days (entry 1). The organic bases DMAP and DBU did not afford more than trace amounts of 2a. Replacing K2CO3 with Cs2CO3 however, provided full, clean conversion in just 24 hours. We also observed that we could replace DMF with THF without penalties, allowing for a simpler workup procedure, and the amount of MeI could be reduced to 3.0 equivalents for 1a.  [3,2-a]pyridine-3-carboxylate (1n): C-10 (200 mg, 0.529 mmol), DMAP (6.47 mg, 0.053 mmol) and DCC (163 mg, 0.794 mmol) were dissolved in DCM (5 mL) at 25 °C and L-menthol (124 mg, 0.794 mmol) was added to the mixture. The reaction mixture was then left on stirring at 40 °C overnight. After 24 h, reaction mixture was diluted with DCM (100 mL) washed with aqueous NH4Cl (saturated) followed by washing with brine (150 mL), and dried over anhydrous Na2SO4, filtered, and concentrated. The crude product was purified by automated flash column chromatography (25 g SNAP Cartridge) eluting with 0-40% ethyl acetate in heptane, to provide 170 mg (67%) of 1n, as a white powder. The product was purified by automated flash column chromatography (25 g SNAP Cartridge) eluting with 0-40% ethyl acetate in heptane, and 200 mg of C-10 was converted to 160 mg (59%) of 1n, isolated as a

Separation of enantiomers of 10b
Scheme S7: Separation of enantiomers of 10b using chiral HPLC After establishing the method, semipreparative chiral HPLC was used to separate the enantiomers of 10b. 40 mg of compound 10b was dissolved in 2.0 ml of DMSO and injected in four iterations. MeCN/H2O with 0.15% TFA was used as mobile phase. An isocratic gradient of 55% MeCN (0.15% TFA) and 45% H2O (0.15% TFA with was run for 65 min. at a constant flow rate of 18 ml/min. The first peak eluted after 20 min. and followed by the second peak ( Figure S5). Fractions were collected manually, concentrated under vacuo and freeze dried. Four separate injections of racemic 10b were performed in order to get enough material for biological testing. NMR confirmed that the first peak was a pure enantiomer (also confirmed by analytical HPLC, (Figure S7). and the second peak was 95% pure enantiomer ( Figure S8).     Figure S8. The high-performance liquid chromatography (HPLC) of injected racemic 10b (+) (10 μL, 1 mg/mL in DMSO), with Lux 5 μm i-Amylose-1 (250 x 4.6 mm) chiral column, isocratic 55% MeCN+ 45% H2O (with 0.15% TFA) eluting-solvent system, 0.8 mL/min flow rate and 254 nm detection wavelength at ambient temperature.

Experimental procedure for evaluation of fibrillation modulating properties.
Human wild-type α-synuclein was expressed and purified as described previously, 2 and denatured and refolded according to the published procedure, 1a in order to enhance the reproducibility of the experiments. A 96-well plate (Corning 3650, Kennebunk, ME, USA) was loaded with samples containing wild-type a-synuclein (70 μM) and compound (100 μM) solubilized in PBS (10 mM) and DMSO (100 μM), followed by addition of 20 μM ThT (Sigma-Aldrich, Saint Louis, MO, USA) and a 2 mm glass bead. The plate was incubated at 37 °C using a FLUOstar Omega microplate reader (BMG Labtech GmbH, Ortenberg, Germany) set to orbital averaging, 500 cycles, and a cycle time of 600 seconds, during 70 hours. Each experiment was performed in triplicate. The formation of amyloid fibers was followed by ThT fluorescence as a function of time (lex = 440 nm, lem = 480 nm) 3 .
Recombinant Aβ1-40 was obtained from AlexoTech AB (Umeå, Sweden) and fibril formation was monitored according to a previously published procedure.   Note: To verify that the low fluorescence intensity did indeed correspond to a competitive binding with ThT, mature fibrils were allowed to form in the absence of compound for 70 h, whereupon the supposed fibril binder was added. The remaining ThT fluorescence was then measured (Figure S11). Figure S11: Compounds 7, 10a and 10e showing low ThT amplitudes in the α-synuclein experiments displace bound ThT. Displacement of ThT was monitored after 70 hours when fibers were fully formed by addition of 100 µM of compound and further incubation for 20 hours. The decrease in signal intensity can be compared before (70 hours) and after addition (75 hours) of compound, when the signals had reached a steady plateau.
Note: In order to distinguish between inhibitors and compounds binding to mature fibrils, compounds 8b, 10a and 10b were added to mature Amyloid b1-40 fibrils ( Figure S12). Figure S12: Compounds 8b, 10a and 10b showing an inhibitory effect of Aβ1-40 fiber formation were investigated for displacement of bound ThT. The ThT amplitude was monitored after 47 hours when fibers were fully formed by addition of 20 µM of compound and further incubation for 13 hours. The signal intensity can be compared before (47 hours) and after addition (50 hours) of compound, when the signals had reached a steady plateau. Only compound 10a show minor displacement of ThT indicated by a reduction of amplitude.  Figure S15: ThT plots for tested inhibitory compounds acquired for Amyloidb1-40. Plots of the previously reported inhibitor TTR (Transthyretin) 5 and curcumin 6 were added as controls for Amyloidb1-40 inhibition. For the controls, 200 nM of transthyretin was used, whereas for curcumin, 20 uM was used. Due to a higher sensitivity setting for these two controls, the signal amplitude has been adjusted for clarity.

Experimental procedure for investigation of elevated ThT amplitudes.
In contrast to outcompeting ThT, causing the fluorescence signal amplitude to decrease, some traces show a higher fluorescence signal compared to the control experiment. A test was therefore performed to distinguish whether these fibrils are modulated by the compound, or if the compound rather alter the binding mode of ThT to the fibril, causing the fluorescnence to increase. Fibril formation of α-synuclein propagates via a primary-nucleation, fibrilbreakage mechanism at neutral pH. 2 The process of polymerization occurs according to a template-dependent effect where free monomers accommodate into the end of the parental fibril. The fibrillar architecture is hence preserved. 3 To selectively probe the template-dependent polymerization of α-synuclein and hence investigate if an alternative fibrillar architecture has been formed as a result of the 2-pyridone compound present, fibril-propagation during high seed conditions was performed. Using this approach, the rate of nucleation can be neglected since fibril-elongation by far dominates the process during the logarithmic-phase. 7 To asses this, 400 μl of mature fibrils were taken after 70 h, centrifuged at 17000 x g for 20 min. using a Micro Star 17 centrifuge (VWR, USA). The supernatant was then discarded. To remove any traces of 2-pyridone compound, 1 ml of Milli-Q water was added and the centrifugation was repeated as described above. This washing procedure was repeated in twice more. To maximize the number of free fibrillar ends, the fibrils were then sonicated using a Qsonica Q500 Sonicator (Qsonica, USA) (20% amplitude, 8 cycles, and 3 sec on, 3 sec off). 5% (v/v) of the washed and sonicated fibrils were then added to a new ThT experiment together with monomeric α-synuclein in a new fibrilisation experiment, without 2-pyridone compound. The maximum fluorescence intensities of the seeded experiments vs. control experiment were then compared to investigate whether the compound modulate the fibrils formed or not ( Figure  S16). No significant differences in ThT fluorescence between the seeded experiments and the control was observed. Thus, the greatly increased fluorescence measured at the plateau phase of the fibrilisation experiments with 8a, 10c or 10e present are not caused by any modulation of the fibril form, exerted by the 2-pyridone compound. Figure S16: Experiments with α-synuclein in the presence of some 2-pyridone compounds showed greatly increased ThT amplitudes at the plateau phase (after about 30 h incubation). These observations are not due to modulation of the fibrils, exerted by the compounds. Monomeric α-synuclein was incubated together with 5% (v/v) washed fibrils taken from a previous experiment where the 2-pyridone-compounds were present. No significant difference in the ThT signal at the end of the incubation (70 h) was seen. The dotted line represents the difference in ThT signal between unseeded and seeded control experiment. WT = control experiment (α-synuclein only), WT* = seeded control experiment (seeded with fibrils formed in the absence of 2-pyridonecompound), 8a = a-synuclein incubated in the presence of 8a, 8a* = α-synuclein monomers seeded with fibrils generated together with compound 8a, 8b = a-synuclein incubated in the presence of 8b, 8b* = α-synuclein monomers seeded with fibrils generated together with compound 8b.

Preparation of samples for TEM visualization of amyloid fibrils.
At the end point of the fibrilization experiments, samples (3.5μL) were applied to glow discharged formvar and carbon coated Cu-grids. The grids were washed and then negatively stained in 1.5% uranyl acetate for 2 x 15 s. A Talos 120C microscope (FEI, Eindhoven, The Netherlands) was used for sample examination, operating at 120kV. Micrographs were acquired with a Ceta 16M CCD camera (FEI, Eindhoven, The Netherlands) using TEM Image & Analysis software ver. 4.17 (FEI, Eindhoven, The Netherlands). Pictures were taken at 12 000 X and 40 000 X magnification, and are shown below. Fibers formed in the presence of compound 10a, 10d and in the absence of compound showed no visual differences (control).
Figure S17: Transmission electron microscope images of fibers formed after the addition of 10a to human wildtype α-synuclein Figure S18: Transmission electron microscope images of fibers formed after the addition of 10d to human wildtype α-synuclein Figure S19: Transmission electron microscope images of human wild-type α-synuclein fibers S15 Crystallography. X-ray quality crystals of 5e were obtained as a racemic mixture through recrystallization from absolute ethanol. Intensity data was collected with an Oxford Diffraction Excalibur 3 system, using ω-scans and Mo Kα (λ = 0.71073 Å) radiation. 8 The data was extracted, integrated and empirically absorption corrected using Crysalis RED. 9 The structure was solved by direct methods and refined by full-matrix least-squares calculations on F 2 using SHELXL and WinGX. 10,11 Molecular graphics were generated using Crystal Maker 9.2. 12 CCDC deposition number 2087355 Figure S20. Crystal Structure of both enantiomers of 5e. 30% ellipsoid contour probability.