Bouchaib
Mouhsine
ab,
Anthony
Saint Pol
a,
Abdallah
Karim
b,
Maël
Penhoat
*d,
Clément
Dumont
ac,
Isabelle
Suisse
a and
Mathieu
Sauthier
*a
aUniv. Lille, CNRS, Centrale Lille, ENSCL, Univ. Artois, UMR 8181 – UCCS Unité de Catalyse et Chimie du Solide, F-59000 Lille, France. E-mail: mathieu.sauthier@univ-lille.fr
bÉquipe de Chimie de Coordination et de Catalyse, Département de Chimie, Faculté des Sciences Semlalia, Université Cadi Ayyad, BP 2390, Marrakech, Morocco
cICAM, Site de Lille, 6 rue Auber, 59016 Lille Cedex, France
dUSR 3290, MSAP (Miniaturisation pour la Synthèse, l'Analyse et la Protéomique), Univ. de Lille, F-59000 Lille, France
First published on 14th June 2023
2-Oxindoles and 2-coumaranones could be selectively allylated in the presence of low cost catalysts based on nickel, leading selectively to different polysubstituted derivatives. Allyl alcohol was used as an allylating reagent and allowed the synthesis of compounds under neutral conditions with water produced as the sole by-product. Experiments have been performed in batch and in flow chemistry. The latter protocol in flow led to a very efficient and straightforward allylic alkylation of 2-oxindoles in only a few minutes with high chemoselectivities towards either the C,C-bisallylated product or the C,C,N-trisallylated one.
2-Oxindoles and 2-coumaranones belong to an important class of molecules with a privileged aromatic heterocyclic scaffold. 2-Oxindoles are for example found in a wide range of bioactive natural products and pharmaceuticals.9–15 The introduction of substituents on the nucleophilic positions (N1 or C3) of 2-oxindoles allows the synthesis of a large panel of valuable molecules with unique biological activities such as anti-cancer, anti-microbial or anti-bacterial properties.16–19 In this context, the introduction of an allyl functionality on oxindole or coumaranone frameworks is interesting because of further possible reactions on the terminal double bond such as oxidation, hydroformylation or metathesis and the possibility of producing new valuable building blocks. The allylic alkylation reaction is one of the most relevant and efficient reactions for the construction of C–C and C–heteroatom (N/O/S) bonds.20,21 Despite the abundant studies in this field, allylation of 2-oxindoles has been poorly explored. This reaction has been performed in the presence of transition metals as palladium,22–26 molybdenum27 and iridium/copper28 as catalysts. 2-Coumaranone allylation has only been performed by stoichiometric means, typically with allyl bromide in the presence of K2CO3/Bu4NHSO4 or NaH, thus generating large amounts of salts.29,30 To the best of our knowledge, the use of nickel in the catalytic allylation of 2-oxindoles and 2-coumaranones has never been reported.31–34
Herein, we report the allylic alkylation of 2-oxindoles and 2-coumaranones with a particular focus on the sustainability of the process. The N-unprotected 2-oxindole skeleton bears two nucleophilic sites (N1 and C3) which are in competition for the allylation reaction (Scheme 1). For the first time, catalytic oxindole allylation is performed in the presence of a nickel based catalyst as a non-critical material and low cost metal. In order to prevent waste, allyl alcohol is used as an allylation reagent leading to water as the sole reaction by-product.35–39 It is also noteworthy that allyl alcohol is industrially produced40 and is also accessible from vegetal feedstock as glycerol.41–44
Scheme 1 The nickel-catalyzed allylation of 2-oxindole and 2-coumaranone with allyl alcohol: sustainability of the reaction/selectivity issue. |
Depending on the sites that are allylated and the substitution degree, several products are accessible. Diallylated products are particularly interesting as precursors of pharmaceutically active compounds.16 Herein, we wish to report that nickel-based catalytic systems are able to promote the C,C-bisallylation and C,C,N-trisallylation of 2-oxindoles with allyl alcohol under additive-free conditions. This protocol also proved to be suitable for the C,C-bisallylation of 2-coumaranones. As part of our interest in developing efficient continuous flow protocols,45,46 this reaction was implemented under flow conditions, thus allowing better activities and a highly improved selectivity control toward the products of C,C-bisallylation or C,C,N-trisallylation of 2-oxindoles.
Entry | [Ni] (mol%) | L | T (°C) | Solvent (0.5 mL) | Equiv. 2a | Conv. 1a (%) | Yieldb (%) | ||
---|---|---|---|---|---|---|---|---|---|
3a | 4a | 5a | |||||||
a Conditions: 1a (1.8 mmol), Ni(cod)2/ligand (1:2), 17 h in a sealed Schlenk tube. b Determined by GC of the crude product using anisole as an internal standard. | |||||||||
1 | 1.5 | L1 | 80 | MeOH | 2 | 83 | — | 50 | 33 |
2 | 1.5 | L2 | 80 | MeOH | 2 | 85 | — | 60 | 25 |
3 | 1.5 | L3 | 80 | MeOH | 2 | 77 | — | 27 | 50 |
4 | 3 | L1 | 80 | MeOH | 3 | 100 | — | 70 | 30 |
5 | 3 | L1 | 100 | EtOH | 3 | 100 | — | — | >99 |
6 | 3 | L1 | 100 | EtOH | 2 | 81 | — | 56 | 25 |
7 | 3 | L1 | 100 | EtOH | 1 | 61 | 18 | 43 | — |
8 | 3 | L1 | 100 | Toluene | 3 | 100 | — | — | 97 |
9 | 3 | L1 | 100 | Dioxane | 3 | 100 | — | — | >99 |
10 | 3 | L1 | 100 | THF | 3 | 100 | — | — | >99 |
11 | 3 | L1 | 100 | Neat | 3 | 100 | — | — | 98 |
12 | 3 | L1 | 100 | DMSO | 3 | 100 | — | 75 | 25 |
The 1a conversions were not complete (77 to 85%) because of the total consumption of allyl alcohol to produce the polysubstituted products 4a and 5a. The ratio of allyl alcohol 2a/2-oxindole 1a and the catalyst amount were then increased (3 equiv. of 2a and 3 mol% Ni) resulting in a total 2-oxindole conversion (entry 4). This also led to the formation of 4a as the major product of reaction (yield = 70%) in the mixture with the trisallylated derivative 5a (yield = 30%). An increase of reaction temperature to 100 °C allowed the selective formation of the trisallylated product 5a with 100% 2-oxindole conversion (entry 5). In this case, EtOH was chosen instead of MeOH due to its higher boiling point. In order to evaluate the possibility of synthesizing the monosubstituted product 3a, the quantity of allyl alcohol was then decreased (entries 6 and 7). The bisallylated derivative 4a was always the major product of reaction even with only 1 equiv. of allyl alcohol. In the latter case, 3a was obtained with a low yield of 18% and the conversion of the starting material was not complete, thus showing rapid bisallylation shortly after the first allylation of 2-oxindole (entry 7). For further investigation, various solvents were also evaluated. The reaction yielded more than 97% of the trisallylated product in most solvents such as toluene, dioxane and THF as well as under solvent-less conditions (entries 8–11). However, DMSO as a polar aprotic solvent proved to be the convenient choice for the allylation of indoles38 and allowed a more selective synthesis of the C,C-bisallylated 2-oxindole 4a which was obtained with 75% yield (entry 12).
In summary, the C,C,N-trisallylated 2-oxindole could be easily produced in different solvents with 3 equiv. of allyl alcohol and 3 mol% Ni catalyst at 100 °C. The use of DMSO clearly improves the selectivity of the reaction toward the C,C-bisallylated derivative. At this point, we decided to study the scope of the reaction for the synthesis of more challenging compounds. Various 2-oxindoles have then been scrutinized as substrates under the optimized conditions for the synthesis of C,C-bisallylated derivatives (Ni(cod)2/dppf/DMSO/100 °C) (Table 2). The C,C-bisallylated 4a–g and C,C,N-trisallylated 5a–g derivatives were separated and purified by silica gel column chromatography, thus allowing access to isolated yields for both families of compounds. 2-Oxindole 1a was transformed into the 4a and 5a derivatives and the products were isolated with yields of 60% and 20% respectively (entry 1). No clear tendency corresponding to electronic effects of the oxindole 5-substitution could be highlighted. 5-Fluorooxindole 1b led to the formation of both bis and trisallylated derivatives with respectively 67% and 29% yields (entry 2). The C,C-bisallylated product 4c was isolated as the sole product with a yield of 70% from the reaction with 5-bromooxindole 1c (entry 3) and the C,C-bisallylated product 4d was also selectively obtained from 5-methoxyoxindole 1d with a yield of 95% (entry 4). The bis and trisallylated products were respectively obtained with 63% and 32% yields from the reaction with 5-aminooxindole 1e. It is very noteworthy that no allylation occurred on the amino group of the aniline moiety, very likely because of the lower acidity of this group in comparison to the amide group of the 2-oxindole (entry 5). The introduction of substituents at positions 4 and 6 of 2-oxindole led to very similar results. For example, only 36% of the C,C-bisallylated product 4f was isolated from 4-bromooxindole derivative 1f along with 30% 5f (entry 6). 4g and 5g were obtained with respectively 73% and 17% yields from 6-acyloxindole 1g (entry 7).
Entry | 2-Oxindole | Conv.b (%) | Isolated yield (%) | |
---|---|---|---|---|
1a–g | 4a–g | 5a–g | ||
a Reaction conditions: 1a–g (1.8 mmol), 2a (5.4 mmol), Ni(cod)2 (0.054 mmol), dppf (0.108 mmol), DMSO (0.5 mL), 17 h, T = 100 °C in a sealed Schlenk tube. b Conversions determined by GC using anisole as an internal standard. | ||||
1 | 1a | 100 | 60 | 20 |
2 | 1b | 100 | 67 | 29 |
3 | 1c | 82 | 70 | — |
4 | 1d | 100 | 95 | — |
5 | 1e | 100 | 63 | 32 |
6 | 1f | 95 | 36 | 30 |
7 | 1g | 100 | 73 | 17 |
C- and/or N-substituted 2-oxindoles 1h–k were also reacted (Table 3). N-Phenyl-2-oxindole 1h was efficiently converted into the C,C-bisallylated derivative 4h with 82% isolated yield (entry 1). However, with an electron donating group on the nitrogen, the N-methyl-2-oxindole 1i was bisallylated with a moderate yield of 51% (entry 2). The 3-methyl-N-phenyl-2-oxindole 1j was readily allylated and the corresponding product 4j was isolated with a high yield of 98% (entry 3). Finally, a very good yield of 85% for derivative 4k was obtained from 3-methyl-2-oxindole 1k and only 10% of the C,C,N-trisallylated derivative 4k′ was observed (entry 4).
Entry | 2-Oxindole | Conv.b (%) | Isolated yield (%) |
---|---|---|---|
1h–k | 4h–k | ||
a Reaction conditions: 1h–k (1.8 mmol), 2a (5.4 mmol), Ni(cod)2 (0.054 mmol), dppf (0.108 mmol), DMSO (0.5 mL), 17 h, T = 100 °C in a sealed Schlenk tube. b Conversions determined by GC using anisole as an internal standard. | |||
1 | 85 | 82 | |
2 | 1i | 56 | 51 |
3 | 100 | 98 | |
4 | 1k | 100 | 4k: 85% |
4k′: 10% |
Fig. 1 Evolution curves of allylation reaction of 2-oxindole 1a with allyl alcohol 2a in batch mode. |
We first evaluated the 2-oxindole allylation under flow. After preparative experiments, we set up a reactor with FEP tubing (length = 1 meter; tubing i.d. = 800 μm; heated volume = 0.5 mL). The tube was immersed in a water bath at 80 °C. Substrates (2-oxindole and allyl alcohol) were introduced thanks to a unique syringe. A quenching module at the tube exit was not necessary since the reaction is immediately stopped in the presence of oxygen. A back pressure regulator (20 psi) should be installed because the reaction temperature was higher than the boiling point of the solvent (Scheme 3).
The experimental catalytic conditions were identical to those of the batch procedures (Ni(cod)2/dppf/THF/80 °C). Thanks to the faster heat transfer, the catalytic transformation was greatly accelerated in the flow reactor with a very high chemoselectivity. Indeed, within 3 minutes, the bisallylated 2-oxindole product 4a was obtained with 100% conversion (Fig. 3). Increasing the residence time to 25 minutes in the reactor tube by using a lower flow rate allowed converting 4a to trisallylated product 5a (Fig. 3). As in the batch (Fig. 1), the monoallylated product was not observed, which shows that the second allylation step is considerably much faster than the first one. This can be explained by an increased inductive effect between the two steps. In that way, by tuning the residence time of the substrates in the reactor, either C,C-bisallylated product 4a or C,C,N-trisallylated 2-oxindole 5a could be obtained selectively in a very short time.
Using the same conditions, a similar experiment was carried out with 2-coumaranone. The solution needed to be first passed into an ultrasound bath to dissolve reagents before introduction into the flow reactor and to avoid any clogging of the system. Surprisingly, at a flow rate of 100 μL min−1, we observed that more than 80% of the C,C-bisallylated derivative 4a′ was produced after 1 minute and 95% of 2-coumaranone had been transformed to the C,C-bisallylated 2-coumaranone 4a′ after 5 minutes. In view of this noticeable result, we thought to perform the allylation reactions in batch at ambient temperature to compare. Indeed, only 2-coumaranone could be completely transformed into the diallylated product in 17 h while the other 2-oxindole substrates are unreactive under these conditions. This shows that these catalytic transformations are very sensitive to heat transfer conditions and the high level of control offered by flow chemistry enables greater selectivity to a chosen synthetic product just by varying the flow rate (Fig. 4).
All experimental details and product characterization data can be found in the ESI.†
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3re00192j |
This journal is © The Royal Society of Chemistry 2023 |