Hirotsugu
Suzuki
,
Yuya
Kawai
,
Yosuke
Takemura
and
Takanori
Matsuda
*
Department of Applied Chemistry, Tokyo University of Science, 1-3 Kagurazaka, Shinjuku-ku, Tokyo 162-8601, Japan. E-mail: mtd@rs.tus.ac.jp
First published on 16th March 2022
We developed a rhodium-catalysed decarbonylative C(sp2)–H alkylation method for indolines. This reaction facilitates the use of alkyl carboxylic acids and their anhydrides as a cheap, abundant and non-toxic alkyl source under redox-neutral conditions, featuring the introduction of a primary alkyl chain, which cannot be addressed by previous radical-mediated decarboxylative reaction. Through a mechanistic investigation, we revealed that an initially formed C-7 acylated indoline was transformed into the corresponding alkylated indoline via a decarbonylation process.
The C-7 alkylation has also attracted considerable attention within the field of C–H functionalisation of indolines, utilising a series of alkylating reagents: activated alkenes,7 diazo compounds bearing an electron-withdrawing group,8 cyclopropanols9 and aziridines.10,11 However, these alkylating reagents cannot yield indolines bearing a simple alkyl chain without using any specific functional groups. Recently, Punji et al. addressed this problem by using simple primary and secondary alkyl chlorides in their iron-catalysed alkylation, but its highly basic conditions might narrow the substrate scope (Scheme 1a).12
Alkyl carboxylic acids have recently garnered significant attention as an alkylating reagent owing to their advantages, such as being inexpensive, abundant in nature, stable under benchtop conditions and easy to handle.13,14 Despite these advantages, only one precedent of C-7 alkylation of indolines has been reported to date: the palladium-catalysed oxidative decarboxylative alkylation of indolines with alkyl carboxylic acids (Scheme 1b).15 In this method, primary alkyl carboxylic acids could not be coupled with indolines because of the low stability of the primary alkyl radical species. Moreover, the oxidative conditions sometimes reduce the generality of the reaction and complicate the isolation procedure. Considering these limitations, we developed the redox-neutral decarbonylative alkylation of indolines by employing (in situ-formed) carboxylic anhydrides (Scheme 1c).16
Inspired by previous work on decarbonylative alkylation,13 we began our investigation by employing 1-(pyrimidin-2-yl)indoline (1a), propionic acid (2a) and pivalic anhydride with various rhodium salts as a catalyst (Table 1). An initial experiment was performed in 1,4-dioxane at 130 °C for 18 h in the presence of [RhCl(CO)2]2 (2.5 mol%), producing the desired 7-ethyl-1-(pyrimidin-2-yl)indoline (3a) in good yield (entry 1). Other rhodium catalysts, namely [RhCl(cod)]2, [Rh(OAc)(cod)]2, [Rh(cod)2]BF4, [Rh(cod)2]OTf and RhCl(PPh3)3, yielded little of the desired C–7 alkylated product (entries 2–6). Among the solvents tested, 1,2-dichloroethane (DCE) was proved to be the best choice (entries 7–10). Other anhydrides, acetic anhydride, di-tert-butyl dicarbonate (Boc2O) and the combination of Boc2O and pivalic acid, decreased the yield (entries 11–13). Lower reaction temperature (110 °C) was inadequate for the reaction to promote (entry 14). Increasing the amount of 2a and pivalic anhydride resulted in an improved yield of the coupling product (entry 15). A control experiment showed that both the rhodium catalyst and pivalic anhydride are essential for the reaction to proceed (entries 16 and 17).
Entry | Rh catalyst | Solvent | Anhydride | Yield (%)b |
---|---|---|---|---|
a Reaction conditions: 1a (0.2 mmol), 2a (0.24 mmol), Rh catalyst (5 mol% of [Rh]) and Piv2O (0.3 mmol) were reacted in solvent (0.5 mL) at 130 °C for 18 h, unless otherwise noted. b Yield was determined by 1H NMR analysis using 1,2,4,5-tetramethylbenzene as an internal standard. The value in parentheses indicates isolated yield. c At 110 °C. d 2a (0.4 mmol) and Piv2O (0.5 mmol) were used. | ||||
1 | [RhCl(CO)2]2 | 1,4-Dioxane | Piv2O | 69 |
2 | [RhCl(cod)]2 | 1,4-Dioxane | Piv2O | 0 |
3 | [Rh(OAc)(cod)]2 | 1,4-Dioxane | Piv2O | 8 |
4 | [Rh(cod)2]BF4 | 1,4-Dioxane | Piv2O | Trace |
5 | [Rh(cod)2]OTf | 1,4-Dioxane | Piv2O | Trace |
6 | RhCl(PPh3)3 | 1,4-Dioxane | Piv2O | Trace |
7 | [RhCl(CO)2]2 | Toluene | Piv2O | 59 |
8 | [RhCl(CO)2]2 | DCE | Piv2O | 83 |
9 | [RhCl(CO)2]2 | MeCN | Piv2O | 24 |
10 | [RhCl(CO)2]2 | DMF | Piv2O | Trace |
11 | [RhCl(CO)2]2 | 1,4-Dioxane | Ac2O | 50 |
12 | [RhCl(CO)2]2 | 1,4-Dioxane | Boc2O | 31 |
13 | [RhCl(CO)2]2 | 1,4-Dioxane | Boc2O + PivOH | 16 |
14c | [RhCl(CO)2]2 | 1,4-Dioxane | Piv2O | 41 |
15d | [RhCl(CO)2]2 | DCE | Piv2O | 91(92) |
16 | – | DCE | Piv2O | 0 |
17 | [RhCl(CO)2]2 | DCE | Trace |
Having determined the optimised reaction conditions, we explored the decarbonylative alkylation of various alkyl carboxylic acids 2 with 1a (Table 2).17 Acetic acid, n-octanoic acid, isovaleric acid and 3-phenylpropionic acid delivered the corresponding C-7 alkylated indolines 3b–e in 73–85% yields. Moreover, the reaction with 3-phenylpropionic acid proceeded smoothly on 1 mmol scale, showing the feasibility for a large scale synthesis (3e). Alkyl substituents bearing phenoxide, phthalimide and methyl ester moieties were tolerated, providing the corresponding alkylated indolines in moderate to good yields (3f–h). 2-(4-Methoxyphenyl)acetic acid afforded an acceptable yield of 3i. Subsequently, the substrate scope of the indoline derivatives was examined. The reaction of 2- and 3-substituted indolines produced the desired alkylated indolines 3j–l in 81–96% yields. 4-Methylindoline rendered the desired product 3m in high yield. Indolines bearing an electron-donating and -withdrawing group at the 5-position, including methoxy, methyl, fluoro and chloro groups, underwent the alkylation with high yields (3n–q). 6-Fluoroindoline was compatible with the reaction conditions (3r), while 6-methylindoline failed to react with 3-phenylpropionic acid. Notably, the reaction was not restricted to indolines; a 2-substituted indole and a carbazole reacted with an alkyl carboxylic acid in a similar fashion to yield 3t and 3u in 73% and 52%, respectively. Further, C-2 alkylation proceeded with 1-(pyrimidin-2-yl)indole, furnishing 73% of 3v. The decarbonylative arylation of 1a proceeded smoothly by subjecting benzoic acids to the optimised reaction conditions (3w and 3x).
Alkyl carboxylic anhydrides underwent this decarbonylative alkylation reaction in a similar manner (Table 3). The alkylation of propionic and acetic anhydrides with 1a proceeded smoothly to provide the corresponding products in moderate to high yields (3a and 3b, respectively). To our delight, branched carboxylic anhydrides participated in the reaction to furnish the products 3y–A in 71–85% yields. Application of the reaction conditions to benzoic anhydride derivative successfully led to the formation of the C-7 arylated indoline 3w.
a Reaction conditions: 1a (0.3 mmol), 4 (0.6 mmol) and [RhCl(CO)2]2 (2.5 mol%) were reacted in DCE (0.75 mL) at 150 °C for 18 h. |
---|
A series of control experiments were performed to gain insight into the reaction mechanism (Scheme 2). Initially, a set of H/D exchange experiments was conducted in the presence/absence of 2a and pivalic anhydride (Scheme 2a). The addition of D2O to the standard reaction conditions resulted in the incorporation of deuterium into the recovered starting material. Moreover, H/D exchange of 1a was observed by heating 1a, [RhCl(CO)2]2 and D2O (5.0 equiv.) in DCE at 130 °C. These results indicate that the C(sp2)–H bond activation step is reversible. When the alkylation reaction was halted after 1 h, a significant amount of 7-acylated indoline 5 was detected (Scheme 2b). As ketone 5 appears to be a putative intermediate in the formation of 3a, 5 was subjected to the standard reaction conditions without an alkyl carboxylic acid and pivalic anhydride (Scheme 2c). The clean conversion of 5 into 3a was confirmed, indicating that 3a arose from the catalytic decarbonylation of 5.18 In the radical trapping experiment with BHT, no significant decrease of the yield was observed in both alkyl carboxylic acid and anhydride cases,19 suggesting the involvement of an alkyl radical species is unlikely (Scheme 2d). Moreover, the kinetic isotope effect experiment revealed that the C–H bond cleavage step is likely to be the rate-determining step (Scheme 2e).
Based on the aforementioned experimental results and previous reports,13,14,16 we propose the reaction mechanism depicted in Fig. 1. The reaction starts with the formation of mixed anhydride 2′ followed by oxidative addition of the C–O bond to rhodium(I) A to furnish acylrhodium(III) carboxylate B. Coordination of the indoline 1 to B promotes a C-7 selective C–H activation, which might proceed via an electrophilic mechanism,13c yielding six-membered rhodacycle C. Reductive elimination from C affords the acylated product 5. Besides, acylrhodium(III) C undergoes deinsertion of CO and subsequent reductive elimination, leading to the formation of the C-7 alkylated indole 3 along with the regeneration of active Rh(I) catalyst A.
Footnote |
† Electronic supplementary information (ESI) available: Experimental procedures and characterisation data for new compounds. See https://doi.org/10.1039/d2ob00249c |
This journal is © The Royal Society of Chemistry 2022 |