Yanji
Song
,
Yuqiao
Zhou
,
Zheng
Tan
,
Ziliang
Jiao
,
Xiaoming
Feng
and
Xiaohua
Liu
*
Key Laboratory of Green Chemistry & Technology, Ministry of Education, College of Chemistry, Sichuan University, Chengdu 610064, China. E-mail: liuxh@scu.edu.cn
First published on 30th September 2024
An interesting asymmetric organocatalytic cascade reaction of 3-hydroxyoxindoles with coumarins was accomplished using a chiral bisguanidinium hemisalt. A series of enantioenriched spirobenzofuran indolinones were afforded in high yields with high diastereo- and enantioselectivities. The reaction was found to involve several steps: an intermolecular Michael reaction, transesterification, a retro-Michael reaction, substitution and decarboxylation, and then an intramolecular Michael reaction to yield the product. The final step of this cascade is found to be the stereo-determining step. In addition, this reaction provides a facile route for the late-stage modification of several drug molecules via the installation of the spirobenzofuran indolinone scaffold.
3-Hydroxyoxindole, which bears two nucleophilic sites, is a useful synthon for the construction of this spiro-heterocyclic skeleton,6 and has been reported to participate in asymmetric Michael additions following a lactamization cascade reaction with methyleneindolinone to yield multicyclic derivatives (Scheme 1b).7 An interesting multi-step reaction with coumarin-3-carboxylates in the presence of organic bases to create racemic benzofuran spiro-indolinones diastereomers had been reported recently by Zhang, Bu and coworkers (Scheme 1c).8 To realize an enantioselective version of this reaction is challenging because the formation of the products proceeds over several steps, seemingly involving a Michael addition, transesterification, substitution and the emission of CO2. Each step generates an intermediate bearing two or three stereogenic centers, whose mutual interference and complicated interactions with the chiral catalyst leads to an unpredictable stereo-arrangement of the products.
In view of the bifunctional catalytic capability of guanidine-amide catalysts in asymmetric cascade reactions and others,9 we investigated their performance in the enantioselective synthesis of benzofuran spiro-indolinones from 3-acyl coumarins and 3-hydroxyoxindoles (Scheme 1c). Herein, we report the identification of a chiral bisguanidinium hemisalt for this purpose, which affords the desired products in excellent yields with high enantio- and diastereoselectivities (up to 99% yield, 96:4 er and >19:1 dr). Tracing the enantioselectivity of the intermediates as well as the products, in connection with control experiments, it was revealed that the stereo-determining step is the final step, regardless of the stereogenic centers generated in the preceding steps, because of the occurrence of a retro-Michael reaction of the spiro-lactone intermediates. The reaction has wide substrate generality and also enables the late-stage modification of drug molecules under mild conditions.
Entry | G, solvent | Yield (%) | dr | er |
---|---|---|---|---|
a Unless otherwise specified, all reactions were carried out with A1 (1.0 equiv.), B1 (0.10 mmol) and chiral guanidine (10 mol%) under an N2 atmosphere at 30 °C for 12–48 hours. Isolated yields are shown. er and dr values were determined using UPCC analysis on a chiral stationary phase. b 5 Å MS (20 mg) were used. c A1 (1.5 equiv.). HBArF = HB[3,5-(CF3)2C6H3]4. | ||||
1 | G1, EtOAc | 68 | 84:16 | 42:58 |
2 | G2, EtOAc | 61 | 80:20 | 38:62 |
3 | G6, EtOAc | 99 | 76:24 | 34:66 |
4 | GS1, EtOAc | 81 | 83:17 | 48:52 |
5 | BG1, EtOAc | 99 | 60:40 | 54:46 |
6 | G6, toluene | 99 | 76:24 | 31:69 |
7 | BG1·HBArF, toluene | 99 | >19:1 | 85:15 |
8 | BG3·HBArF, toluene | 66 | 76:24 | 42:58 |
9 | BG4·HBArF, toluene | 28 | >19:1 | 58:42 |
10 | BG5·HBArF, toluene | 99 | 47:53 | 47:53 |
11 | BG6·HBArF, toluene | 99 | 19:1 | 86:14 |
12b | BG6·HBArF, toluene | 90 | 19:1 | 88:12 |
13b,c | BG6·HBArF, toluene | 99 | >19:1 | 87:13 |
14b,c | BG6·HBArF, iPrOAc | 94 | >19:1 | 95:5 |
With the optimized reaction conditions in hand (Table 1, entry 14), the scope of the 3-hydroxyoxindoles and coumarins in the asymmetric cascade reaction was investigated (Scheme 2). A series of N-ethyl-3-hydroxyindolinones, regardless of the electronic and steric properties of the substituents located at the C5–C7 positions on the phenyl group mostly proceeded well, thus the desired products C2–C14 were isolated in moderate to good yields (69–99% yield) with high enantioselectivity (91:9–96:4 er) and excellent diastereoselectivity (>19:1). The exception to these results was the substrate bearing a 6-methoxyl substituent at the C6-position (C9, 78:22 dr). However, possibly due to steric hindrance, the substituent located at the C4-position was less reactive and almost no product was generated. The absolute configuration of C13 was determined to be (2R,3S) by X-ray diffraction analysis.10 In addition, the N-substitution of the 3-hydroxyoxindoles obviously affected the outcome (C14–C22). It was shown that sterically hindered N-substitution was disadvantageous to the reactivity (C17, C18 and C21), whilst less hinderance at the N-position was not good for the enantioselectivity (C15 and C16).
Scheme 2 Scope of the 3-hydroxyoxindoles and coumarins. Using the same conditions as Table 1, entry 14. dr values were determined by 1H NMR. a5 Å MS (30 mg). bBG6·HBArF (20 mol%). |
The electronic properties and positions of the substituents on the phenol ring of the coumarins appeared to affect the reaction to some extent (C22–C38). Coumarins bearing electron-withdrawing groups partly had an impact on the stereo-outcomes of the cascade reaction. When strong electron-withdrawing substituents were located on the aromatic ring (C28, C31), both the diastereo- and enantioselectivities decreased dramatically (73:27–75:25 er). The coumarins derived from sesamol or 1-hydroxy-2-naphthaldehyde both proceeded smoothly under the optimized reaction conditions (C39, C40), but the er and dr values dropped slightly for C40 (86:14 er, 90:10 dr).
In addition, the reaction was compatible with different acyl groups. Coumarins containing aryl or alkyl acyl groups underwent the cascade reaction well, delivering the corresponding products (C41–C45) smoothly in good yields with high enantioselectivities (88:12–96:4 er, >19:1 dr). Additionally, a series of drug molecules based on 3-benzoyl-6-hydroxy-2H-chromen-2-ones, including indomethacin, fenofibric acid, mefenamic acid, and (S)-naproxen, were employed and all the reactions proceeded in good to excellent yields. This demonstrated the potential of the methodology for late-stage modification.
To prove the synthetic value of this protocol, a gram-scale synthesis of C1 and its further transformation were carried out. As shown in Scheme 3a, the reaction worked well with 4.0 mmol of 3-benzoyl-2H-chromenone (B1) at 30 °C, and the desired product C1 could be isolated in 99% yield (1.52 g) with a nearly maintained er value (94:6 er) and a >19:1 dr. Moreover, the enantiopure product (99:1 er) was readily generated upon recrystallization. The product C1 could be reduced smoothly by NaBH4 to generate the alcohol product D1. In the presence of hydroxylammonium chloride and pyridine, the oxime derivative D2 was obtained in 88% yield and 99:1 er. In addition, treating C1 with phenyl magnesium bromide afforded the diphenyl alcohol derivative D3 in 71% yield and 99:1 er.
The mechanism for the generation of the spirobenzofuran indolinone product was proposed to proceed via multiple steps (Scheme 3b), including a Michael addition to generate Int1, then transesterification to give spirooxindolelactone Int2. Following emission of CO2 and substitution, a new spiro-backbone was formed. Given the fact that all intermediates bear stereogenic centers, it raises questions as to how the chirality is generated and delivered. Tracing the process via controlling the reaction time revealed that the yield of Int2 increased to a maximum in the first six hours and then dissipated, meanwhile, the yield of product C1 gradually increased as the reaction time was prolonged (see ESI† for details). We isolated Int2 from the catalytic system in the middle of the reaction using racemic or chiral guanidine catalysts. Only one diastereomer of Int2 was observed whose relative configuration was identified via X-ray crystal analysis.10 This is consistent with the DABCO promoted racemic reaction of a coumarin-3-carboxylate ester.11 The enantioselectivity of Int2 was not higher than 50% ee in the asymmetric catalysis, which decreased as the highly enantiomerically enriched product C1 was generated.
The direct transformation from Int2 (>19:1 dr) into the final product C1 was investigated (Scheme 3b). Interestingly, when putting the racemic intermediate Int2 into the standard reaction, the desired product C1 could be isolated with high enantio- and diastereocontrol regardless of the reaction time, and nearly quantitative conversion was achieved after longer reaction times. When enantiomerically enriched Int2 was used, there was nearly no difference in terms of reactivity and stereoselectivity in comparison with the racemic reactant, manifesting that the racemization process is fast, and that a kinetic resolution process was not important. No reaction occurred without a catalyst, implying the reconstruction of intermediate Int2 must be accelerated by the catalyst. These results suggested that the stereochemistry of the product was not determined by the stereo-arrangement of Int2 which should undergo a retro-Michael reaction to lose the stereogenic centers generated in the former steps.
Given the observed diastereoselectivity in the racemate synthesis and optimization process (Table 1), direct intramolecular SN2 substitution following decarboxylation could be rule out.8 As depicted in Scheme 4, we propose that Int2 undergoes a retro-Michael reaction and enolization to give achiral Int3 due to the rearomatization of indolinone. With the assistance of the base, the deprotonated phenoxy anion attacks the ester substitution to release CO2via transition state TS1, following by protonation to yield achiral enone intermediate Int4. Finally, in the presence of the chiral guanidinium catalyst, a diastereo- and enantioselective intramolecular Michael addition occurs preferably via a Si/Si-facial approach, leading to the formation of the (2R,3S)-product as the major product. In comparison, the racemic reaction of the 3-ester-substituted coumarin occurred at higher reaction temperatures.8 In our study, the acyl substituted coumarin is used instead of the ester-substituted compound, which enhances the reactivity due to the higher reactivity of enone Int4 in comparison with the related α,β-unsaturated ester intermediate (see ESI for details).
Footnote |
† Electronic supplementary information (ESI) available: Experimental details, characterization data and crystallographic data in CIF. CCDC 2302490 and 2321099. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4qo01471e |
This journal is © the Partner Organisations 2024 |