Intramolecular Povarov Reactions for the Synthesis of Chromenopyridine Fused 2-Pyridone Polyheterocycles Binding to α-Synuclein and Amyloid-β Fibrils

A BF3·OEt2 catalyzed intramolecular Povarov reaction was used to synthesize 15 chromenopyridine fused thiazolino-2-pyridone peptidomimetics. The reaction works with several O-alkylated salicylaldehydes and amino functionalized thiazolino-2-pyridones, to generate polyheterocycles with diverse substitution. The synthesized compounds were screened for their ability to bind α-synuclein and amyloid β fibrils in vitro. Analogues substituted with a nitro group bind to mature amyloid fibrils, and the activity moreover depends on the positioning of this functional group.

T he thiazolino fused 2-pyridone represents a privileged scaffold which can be modified to display various biological activities. 1 It was initially designed as a peptidomimetic to combat the virulence of uropathogenic E. coli by inhibiting the formation of pili. 1a With alternative substitution patterns, compounds based on this scaffold have also demonstrated activity against Chlamydia trachomatis, 1b Listeria monocytogenes, 1c and Mycobacterium tuberculosis. 1d Rigidifying the scaffold by equipping it with sterically demanding aryl groups provides compounds with the ability to modulate formation of bacterial and human amyloid fibrils. 2 Extension of the bicyclic thiazolino fused 2-pyridone with nitrogen containing aromatic heterocycles offers another way of rigidifying the peptidomimetic scaffold, 3 as exemplified by compounds 1−4 ( Figure 1A). These analogues are able to modulate α-synuclein amyloid fibril formation, which is associated with Parkinson's disease, a human neurodegenerative disorder, 2a,3a,b or bind to mature α-synuclein fibrils. 3c Chromenopyridine is a versatile structural motif present in a variety of polyheterocycles with applications in biology as estrogenic, antibacterial, and anticancer agents and biosensors (compounds 5−8, respectively, Figure 1B). 4 Recently, the merging of two different active fragments to develop scaffolds with improved biological properties has received considerable attention. 5 As mentioned above, annulation of bicyclic thiazolino fused 2-pyridone with different heterocycles has resulted in scaffolds capable of modulating amyloid fibrils. 3 We envisaged that fusing thiazolino 2-pyridone and chromenopyridine, by combining units 9 and 10, could afford scaffolds with the ability to target amyloid structures ( Figure 1B) and, in addition, constitute a new central fragment with great potential for drug discovery in general.
Since its discovery, the Povarov reaction 6 has been used widely for the synthesis of nitrogen containing, six-membered heterocycles with biological relevance. 7 Recently, we utilized the Lewis acid catalyzed Povarov reaction to construct a tricyclic pyridine fused 2-pyridone peptidomimetic scaffold, 3c whose analogues have been shown to bind α-synuclein and Aβ fibrils by ThT displacement 8 in vitro (e.g., compound 4, Figure  1A). We envisioned that performing the Povarov reaction in an intramolecular fashion 9 could result in the desired chromenopyridine annulated 2-pyridones 11 in a single operation ( Figure 1B). The new scaffold 11 being equipped with the tetrahedral carbon, and the oxygen, could decrease planarity of the structures and enable increased hydrogen bonding, respectively. 10 In addition, these features could potentially confer selectivity between different amyloid structures, a very desired property in diagnostic and therapeutic applications of amyloid binding small molecules. 11 We perceived that chromenopyridine ring fused 2-pyridone scaffold 11 would be accessible from amino 2-pyridone 9 and O-alkylated salicylaldehyde 10. Hence, we began our work by investigating the feasibility of the reaction between 9a and Ocinnamyl salicylaldehyde 10a in an intramolecular Povarov setup using BF 3 ·OEt 2 as catalyst, followed by oxidation with DDQ (Scheme 1). 3c As hypothesized, the intramolecular reaction worked smoothly, and the desired product 11a was isolated in excellent yield. To explore the substrate scope of the reaction, substituted O-cinnamyl salicylaldehydes 10b−i (Schemes S1 and S2) were allowed to react with 6-amino-2pyridones 9a,b to construct 11b−k in good to excellent yields (Scheme 1).
Aware of the importance of the 4-nitrophenyl substituent for amyloid binding activity of the tricyclic scaffold 4, 3c we prepared 11b−f and 11k with nitro groups in various positions. Notably, salicylaldehydes with electron withdrawing R 3 substituents underwent faster Povarov reaction due to lowering of LUMO in the electrophilic imine intermediate, providing 11b −d, 11g, and 11k. 12 Electron donating R 4 substituents rendered the alkene more nucleophilic and likewise decreased the reaction times, while electron with-drawing R 4 substituents increased them. Heating was required to achieve synthetically useful reaction times for synthesis of 11e and 11h, where, in the former case, the moderate yield reflects a less clean conversion. In the preparation of 11f, the effect of the electron withdrawing R 3 substituent compensated for poor nucleophilicity of the alkene moiety, and heating was not required to complete the reaction in 1 day. To test the scalability, 11b was also prepared from 1.6 mmol of 9a. The yield (88%) is comparable to that at 0.4 mmol scale (86%).
Next, we turned our attention toward the synthesis of C-13 unsubstituted target molecules 13 (Scheme 2). SAR from our previous study suggested that the best amyloid binding properties are achieved when the corresponding position is unsubstituted. 3c It was approached by Povarov reaction The Journal of Organic Chemistry pubs.acs.org/joc Note between 9a and O-allyl salicylaldehyde 12. To our dismay, we were only able to isolate small amounts (7%) of the desired product 13a, from the complex reaction mixture after 4 days at 70°C. Microwave irradiation, 120°C for 3 h, shared the same lack of success, and 13a was isolated in only 11% yield. Though fruitful results are reported, 9e attempts with terminal allyl moieties often suffer from low yields. 7a,9a,12a,b Familiar with the mechanistic features of the Lewis acid catalyzed Povarov reaction, 7a,d,e,12c,13 we realized that use of the terminal allyl group as alkene components would require the reaction to go via high energy carbocations or operate via an alternative mechanism. 9a,b,d,14 With our previous strives in mind where we had made use of ethyl vinyl ether as alkene component, for synthesis of unsubstituted tricyclic analogues 4, 3c we naturally thought of employing a vinyl ester moiety as electron donating auxiliary. With 3-bromopropenyl benzoate, 15 we were able to alkylate salicylic aldehydes to synthesize the required intermediates 14a−d (Schemes 3 and S3). With the alkene component now armed with an ester group, capable of mesomeric contributions, we attempted Povarov reactions between 14a−d and 9a,b. Still, heating the reaction mixtures to 70°C was required to achieve reasonable reaction times.
The desired compounds 13a−e were isolated in low to moderate yields after 24 h of heating, followed by oxidation with DDQ at room temperature. The unusually low yield of 13e results from the formation of side products, which complicated the purification. The low to moderate yields motivated us to try a slightly different approach, using CuBr 2 catalysis and O-propargyl salicylaldehyde (Scheme S4). 7b,c,f,9a Unfortunately, the method suffered from undesired side reactions and 13a was only isolated in 31% yield.
Having both sets of compounds in hand, we hydrolyzed the methyl ester to deprotect the carboxylic acid and reveal the peptidomimetic (Scheme 4). The limited solubility of 16a−e in organic solvents complicated handling, and these compounds were thus obtained in lower yields.
The carboxylic acids 15a−k and 16a−e were initially screened in an α-synuclein fibrilization assay with Thioflavin T (ThT) to probe for fibril binding and modulation of amyloid fibril formation (Figure 2A and Figure S1). 8,16 Compounds 15b,c,15e,f,15k,16a,b,and 16e bind to the α-synuclein fibrils and had a significant effect on the fluorescence intensity, by competing with ThT for binding. In addition, compound 15e has a mild inhibitory effect upon the fibril formation, as the lag phase was slightly extended. In order to verify that amyloid fibrils were indeed present, samples were taken upon the end point of the assay and visualized with transmission electron microscopy (TEM) ( Figure S2). No visible difference was observed between the control experiment (black trace) and the mixture with 16a (green trace) (Figures 2B and S2).
In a similar assay, compounds (15a−c, 15e,f, 15k, 16a−c, 16e) were added after 70 h, when the fluorescence traces had reached the plateau phase ( Figures 3A and S3). The compounds were observed to bind to the mature α-synuclein fibrils and displace bound ThT, indicated by reduction of the ThT fluorescence, to varying degrees ( Figures 3B and S4).
Interestingly, all compounds equipped with a nitro group except 15d showed binding activity. The strongest binding was observed for compounds with the nitro group situated in position C-9 (15b, 15c, 16a, and 16b). Moving the nitro group to position C-8 (16e) or to the 4′ position of the C-13 aryl group (15e,f) is accompanied by a decrease in fibril binding ability. Notably, the presence of the C-13 phenyl group abolishes all binding activity rendered by the nitro group in position C-8 (15d vs 16e) but not C-9 (15b vs 16a, Figure  S3). The cyclopropyl group appears to be the favored C-14 substituent, over methoxy and proton (15b vs 15c and 15k, and 16a vs 16b). This observation is in agreement with the SAR on scaffold 4, where cyclopropyl as R 1 substituent was found somewhat superior to phenyl, hydrogen, and methoxy substituents. 3c Taken together, these observations fit with the hypothesis that binding sites on amyloid fibrils are made up of shallow hydrophobic clefts flanked by polar groups, situated between the rows of side chains and running parallel to the fiber axis. 11a,c,e Compounds were also evaluated against Aβ40 in a similar manner (Figure 4 and Figures S5−S7). Compounds 15b,c, 15f, and 16a, which bind strongly to α-synuclein fibrils, were found to bind mature Aβ fibrils as well, indicated by reduced ThT fluorescence, compared to the control experiments. In summary, we have developed methods to fuse two privileged scaffolds, namely, chromenopyridines and thiazolino 2-pyridones. These new peptidomimetic polyheterocycles constitute a new scaffold with potential for diverse substitution patterns. The intramolecular Povarov reaction between Oalkylated salicylaldehydes and amino functionalized thiazolino-2-pyridones afforded chromenopyridine fused 2-pyridone polyheterocycles in moderate to excellent yields. Rewardingly, biological evaluation of chromenopyridine fused thiazolino 2pyridones revealed compounds capable of binding to αsynuclein and Aβ40 amyloid fibrils in vitro. An interesting SAR was observed with respect to groups at position C-9. The polyheterocycles equipped with a nitro group at position C-9 showed the strongest binding to α-synuclein and Aβ40 amyloid fibrils. However, changing the position of the nitro group resulted in decreased binding ability versus both amyloid fibrils. As binding to mature amyloid fibrils is a property of pharmacological relevance, 11 we intend to investigate these promising compounds further in future biological studies.

■ EXPERIMENTAL SECTION
General. Unless otherwise stated, purchased reactants and reagents were used as received from commercial suppliers. Molecular sieves and LiCl were dried at 300°C under high vacuum for 4 h prior to use. Acetonitrile was dried over activated 3 Å molecular sieves (5% w/v) for 48 h, then transferred via syringe to new 3 Å molecular sieves (5% w/v) for storage until use. DMF, THF, and diethyl ether were dried using an SG Water solvent drying tower according to the manufacturer's instructions and stored over activated 3 Å (DMF) or 4 Å (THF and diethyl ether) MS for 48 h or more before use. Amberlyst was rinsed prior to use with THF/MeOH 1:1 in a cylindrical sintered funnel until the filtrate was transparent, then dried briefly by passing air through. Microwave reactions were performed in     17 Under an atmosphere of nitrogen, cinnamyl bromide II (2.34 mmol, 1.30 equiv) was dissolved in DMF (1.0 mL) and transferred with a syringe to a mixture of salicylaldehyde I (1.80 mmol, 1.00 equiv) and K 2 CO 3 (496 mg, 2.00 equiv) under nitrogen. The reaction mixture was stirred at r.t. until completion was indicated by TLC analysis. The mixture was then transferred to a beaker of ice-cold water (60 mL) while stirring; the resulting precipitate was filtered off and washed with water (5 mL). The precipitate was then dissolved in EtOAc (50 mL) and washed with brine (5 × 25 mL), dried with sodium sulfate, filtered, and evaporated. Unless otherwise stated, the compounds were used in the following synthetic step without further purification.
General Procedure for Synthesis of O-Alkylated Salicylaldehydes 14a−d. Under an atmosphere of nitrogen, 3-bromopropenyl benzoate (868 mg, 3.60 mmol, 1.80 equiv) was dissolved in DMF (2.0 mL) and transferred with a syringe to a mixture of salicylaldehyde I (2.00 mmol, 1.00 equiv) and K 2 CO 3 (553 mg, 2.00 equiv) under nitrogen. The reaction mixture was stirred at r.t. until completion was indicated by TLC analysis. The mixture was diluted with CH 2 Cl 2 (50 mL) and washed with brine (5 × 50 mL), filtered through sodium sulfate, and evaporated. The crude product was purified with automated flash column chromatography.
2-Hydroxy-4-nitrobenzaldehyde (Ib). 2-(Hydroxymethyl)-5-nitrophenol (8.20 g, 1.00 equiv) was dissolved in THF (100 mL). (Diacetoxyiodo)benzene (17.27 g, 1.11 equiv) and TEMPO (158 mg, 0.02 equiv) were weighed up and added. The resulting suspension was stirred at r.t. After 30 min, a clear solution remained, but TLC indicated some remaining acid. No further progress was indicated after 65 min, whereupon (diacetoxyiodo)benzene (3.14 g, 0.20 equiv) was added. After 50 min of stirring, the reaction was found complete by TLC. Saturated aqueous sodium thiosulfate solution (100 mL) was added, and the mixture was stirred for 5 min, then transferred to a separation funnel and extracted with EtOAc (150 + 50 + 50 mL). The organic phases were combined, washed with brine (200 mL), and evaporated. The residue was redissolved in chloroform (300 mL) and extracted with aqueous potassium hydroxide solution (5% w/v, 3 × 200 mL). The combined extracts were cooled in an ice-water bath and neutralized with HCl (12 M, 44 mL). The precipitate was filtered off and dissolved in EtOAc, dried with sodium sulfate, filtered, and evaporated. The residue was redissolved in DCM/MeOH (99:1) and purified with flash column chromatography (70 × 120 mm silica gel, DCM/MeOH 99:1). 5.23 g (64.5%) isolated as a light-yellow powder. NMR data recorded in (CD 3 ) 2 CO are in close agreement with published data, 18 and the phenolic proton is visible as well. 1 3224,3110,3046,2885,1676,1537,1476,1353,1299,1219,1172,1116,958,882,827,793,740 1, 162.0, 152.6, 134.9, 123.8, 114.5, 113. General Procedure for Synthesis of Cinnamyl Esters IVa−c. The cinnamyl esters were synthesized according to a literature procedure with slight modifications. 19 LiCl (509 mg; 12 mmol, 1.2 equiv) was weighed up in an oven-dried 100 mL round-bottom flask and put under vacuum. The flask was heated with a hot air gun for several minutes and allowed to cool down to room temperature, then back-filled with nitrogen. The flask was put under vacuum and backfilled with nitrogen twice more. Dried MeCN (30 mL) was added with a syringe, and the resulting suspension was cooled to 0°C in an ice-water bath. Triethylphosphonoacetate (2.4 mL, 12 mmol, 1.2 equiv) was added with a syringe while stirring. After 20 min was added benzaldehyde III (10 mmol, 1.0 equiv), followed by DBU (1.8 mL, 12 mmol, 1.2 equiv) dropwise. The resulting reaction mixture was allowed to warm up to room temperature and stirred until completion was indicated by TLC analysis. Upon complete consumption of the aldehyde, the mixture was concentrated under reduced pressure and partitioned between diethyl ether (30 mL) and NH 4 Cl solution (sat. aq.) (40 mL). The aq. phase was extracted with diethyl ether (2 × 30), and the combined organic phase was dried over sodium sulfate, filtered, and concentrated by rotary evaporation. The residue was redissolved in DCM and purified with automated flash column chromatography.
General Procedure for Synthesis of Cinnamyl Alcohols Va− d. The cinnamyl alcohols were synthesized according to a literature procedure. 19c The cinnamyl ester IV (10 mmol, 1.0 equiv) was weighed up in an oven-dried 100 mL round-bottom flask, put under nitrogen, and dissolved in dried THF (30 mL). The reaction mixture was cooled to −78°C in a dry ice/acetone bath, and DiBAl-H (1 M in toluene, 24 mL, 24 mmol. 2.4 equiv) was added slowly with a syringe. The mixture was stirred at −78°C while the reaction was monitored with TLC. Upon complete consumption of the cinnamyl ester, saturated Rochelle salt solution (20 mL) was added to the mixture at −78°C. The mixture was allowed to warm up to room temperature, transferred to a separation funnel, and extracted with EtOAc (5 × 40 mL). The organic phases were combined, dried over sodium sulfate, filtered through a pad of EtOAc-wet Celite, and concentrated by rotary evaporation.