Asymmetric organocatalytic Michael addition of cyclopentane-1,2-dione to alkylidene oxindole

An asymmetric Michael reaction between cyclopentane-1,2-dione and alkylidene oxindole was studied in the presence of a multifunctional squaramide catalyst. Michael adducts were obtained in high enantioselectivities and in moderate diastereoselectivities.

Herein, we report the results of an asymmetric organocatalytic Michael addition of CPD to alkylidene oxindoles.

Results and Discussion
Chiral multifunctional thioureas [26,27] and squaramides [28] are extensively used as catalysts in asymmetric Michael additions. We believed that a bifunctional hydrogen-bonding catalyst would activate both CPD via a tertiary amino group of a quinuclidine moiety acting as a base via anion-binding, and an oxindole through the squaramide or thiourea moieties of the catalyst as hydrogen bond donors ( Figure 1) [29][30][31][32]. Therefore, squaramide and thiourea catalysts were screened in a model reaction between CPD 1 and Boc-protected benzylidene oxindole 2a at room temperature in the presence of 10 mol % of catalyst ( Figure 2).
First, the quinidine-derived squaramide A was used and the desired product was obtained as a mixture of chromatographically inseparable diastereoisomers in 53% yield but in low enantiomeric excess for both diastereomers ( Table 1, entry 1). With the quinine-derived thiourea B, the reaction was slow and the yield was very low, 12% (Table 1, entry 2). For that reason, we focused on the screening of squaramides. Squaramides were found to be more selective catalysts than thioureas. When squaramide C was used as a catalyst, the product was isolated in 80%/87% ee (major/minor diastereoisomer) ( Table 1, entry 3). The enantioselectivity was even higher with the cinchoninederived squaramide D, 85%/92% (major/minor) ( Table 1, entry 4). To further optimise the reaction, we screened different solvents (apolar, polar aprotic, and chlorinated solvents) ( Table 1, entries 5-7). According to the obtained results chloroform was clearly superior to other solvents. Previously the isolated yield of the product had been moderate and to increase the yield the substrate concentration was varied. A substantial excess of CPD (five equivalents) led to a very slow reaction and a decrease in enantioselectivity ( Table 1, entry 8). It was assumed that the binding between CPD and the catalyst was stronger than the binding between the substituted oxindole and the squaramide decreasing the effective concentration of the catalyst. Taking this into consideration, 2 equiv of substituted oxindole was used and the reaction proceeded smoothly in 2 h in high enantioselectivity (90%/94% ee), in high yield (74%) but in moderate diastereoselectivity (Table 1, entry 9). Next, we looked onto the effect of lower temperature on the reaction. At 0 °C the reaction was approximately 10 times slower and only the ee of the minor diastereoisomer increased by 3% (Table 1, entry 10), so there was no justification for carrying out the reaction at a lower temperature because of the longer time needed.
Next, we screened different protecting groups for the oxindole. Previously, Boc-protected oxindole 2a gave us the product in 75% yield, in dr 2.6:1 and in ee 90%/94% (Scheme 1, 3a). With a Cbz-protecting group the enantioselectivity decreased to 82%/ 88% (Scheme 1, 3b). The use of a sterically more demanding Fmoc-protecting group decreased the ee values even more for the minor diastereoisomer (Scheme 1, 3c). Surprisingly, with benzyl-protected oxindole, the reaction did not proceed (Scheme 1, 3d), which implies that the carbonyl group of the carbamate moiety in the N-protecting group and electron-withdrawing properties of the protection groups are essential for coordination with the catalyst and for the reactivity of the Michael acceptor. Using a tosyl-protected oxindole the reaction was sluggish, the yield was low and the enantioselectivity could not be determined (Scheme 1, 3e). These experiments revealed that the best results were achieved in chloroform at room temperature with catalyst D, using 1 equiv of diketone and 2 equiv of N-Boc-substituted oxindole 2a.
Under optimised conditions, the substrate scope of the reaction was examined by using various substituted oxindoles with an E-configuration of the double bond. The results are presented in Scheme 2. Both electron-withdrawing (Scheme 2, 3f-h) and electron-donating groups (Scheme 2, 3m,n) at the phenyl ring of the benzylidene moiety were tolerated. The position of the halide at the aromatic ring did not have a major effect on the yield or the enantioselectivity. Ortho-, meta-and para-chlorophenyl-substituted starting materials afforded products in similar enantioselectivities (Scheme 2, 3f-h). However, the reaction was slower with the sterically more hindered ortho-chloro substrate (Scheme 2, 3f). When instead of a benzylidene-containing substrate an alkylidene with an extra ester moiety was used, the enantioselectivity was lost and the product 3i was obtained as a racemic mixture. Either additional coordination with the catalyst or a lack of π-π-interaction may have been responsible for that. Also, a higher C-H acidity of the proton at the stereogenic centre and possible racemisation can't be excluded. A heteroaromatic oxindole derivative afforded the product 3j in lower yield and high ee values. 4-and 5-bromo oxindole derivatives (2l and 2k, respectively) were also used as starting compounds. If the substituent in the oxindole ring was further from the reaction centre, the outcome was not affected (Scheme 2, 3k). However, when using a 4-bromo-substituted oxindole, the reaction was slower, the yield drastically decreased and the en- antioselectivity was moderate (Scheme 2, 3l). The reaction with an electron-donating p-MeO-substituted benzylidene oxindole was very sluggish and did not reach full conversion (Scheme 2, 3m). The product 3m was obtained with only 36% yield and with undetermined enantiomeric purity, since the peaks were not separable in various HPLC methods. Similarly, the p-Mesubstituted oxindole was also slow in reacting and the yield was moderate, but the enantioselectivity remained high (Scheme 2, 3n). The reaction tolerated alkylidene oxindoles, although the product was obtained in a slightly lower yield and enantioselectivity (Scheme 2, 3o). The reaction did not occur when starting compound 2p was tried. This was probably because of the very poor solubility of the starting material. Generally, the diastereoselectivities of the reactions were moderate (dr 2.1:1-3.6:1) throughout the scope. The diastereoselectivity was missing or was very low for the compounds with non-aromatic substituents at the double bond (3i and 3o).
The relative anti-configuration of the vicinal diastereotopic hydrogens was determined by comparing the 3 J HH coupling constants of the major diastereomer with those of the minor diastereomer. The constants were larger for the major diastereomer, meaning vicinal hydrogens were in anti-configuration.
In all previous experiments only E-isomers were used. In the case of the 3-nitro-substituted starting material 2q we managed to separate isomers and carried out the reaction with both the Eand Z-isomer. In these experiments, both isomers afforded the same major diastereoisomer but opposite enantiomers (Scheme 3, 3q). The diastereoselectivities were similar for the isomers.
Since the diastereoselectivity of the reaction was low, we attempted to increase the ratio of diastereoisomers via enolisation followed by diastereoselective protonation ( Table 2). As the racemate of 3a was obtained in a higher diastereomeric ratio (6.3:1) we applied kinetic and thermodynamic conditions for the epimerisation of it ( Table 2, entries 1 and 2, respectively). Unfortunately, in both cases the diastereomeric ratio decreased and the amount of more stable syn diastereoisomer increased. A similar trend was observed when starting from the enantiomerically enriched 3a (Table 2, entry 3).
It has been shown that substituted oxindoles can be converted to indolopyrans via intramolecular cyclisation [33]. We also tried synthesizing 4H-pyrans in acidic conditions but no cyclised product was detected.
Scheme 2: Scope of the reaction (the relative configuration of the major diastereoisomer is depicted). Reaction conditions: 0.2 M solution of 1 equiv of 1, 2 equiv of 2, 0.1 equiv of catalyst D, chloroform, at room temperature; isolated yields after column chromatography; ee determined by chiral HPLC.

Conclusion
In summary, we have developed a new asymmetric organocatalytic Michael addition of cyclopentane-1,2-dione to alkylidene oxindoles catalysed by bifunctional squaramide which leads to products in high enantioselectivities and moderate diastereose-lectivities. The scope of alkylidene oxindoles is reasonably wide including aromatic and aliphatic substituents at the double bond and also substituents in the oxindole core. The work widens the synthetic utility of cyclopentane-1,2diones.

Scheme 3:
Comparison reactions of E-and Z-isomers (the relative configurations of the major diastereoisomers are depicted). Reaction conditions: 0.2 M solution of 1 equiv of 1, 2 equiv of 2, 0.1 equiv of catalyst D, chloroform, at room temperature; isolated yields after column chromatography; ee determined by chiral HPLC.