Vojtěch
Dočekal‡
a,
Bedřich
Formánek‡
a,
Ivana
Císařová
b and
Jan
Veselý
*a
aDepartment of Organic Chemistry, Faculty of Science, Charles University, Hlavova 2030/8, 128 43 Prague 2, Czech Republic. E-mail: jan.vesely@natur.cuni.cz; Web: http://orgchem.cz/vesely/
bDepartment of Inorganic Chemistry, Faculty of Science, Charles University, Hlavova 2030/8, 128 43 Prague 2, Czech Republic
First published on 2nd August 2019
In this study, we report an enantioselective synthesis of sulfur heterocycles containing the dihydro-2H-pyran moiety. The synthesis strategy is based on quinidine catalyzed formal [4 + 2] cycloaddition between 3-alkylidene benzo[b]thiophenes and allenoates, affording the corresponding cycloadducts in high yields (up to 95%) with high enantioselectivities (up to 99% ee). Moreover, cycloadducts can be isolated by simple filtration from the reaction media.
Over the past few decades, Lewis base promoted cycloaddition reaction has been extensively explored. Among the most studied substrates used for catalytic formal cycloadditions are allenoates. After the pioneering work of Lu,6 reporting a phosphine-catalyzed [3 + 2] cycloaddition of allenoates with electron-deficient alkenes, this methodology was adopted to investigate various substrates with polarized CX double bonds, such as imines, carbonyls, and other activated olefins.7 Allenoates represent C2 or C3 synthons in cycloadditions depending on the Lewis base used. According to the different stabilities of amines or phosphorus zwitterionic adducts formed between Lewis base and allenoate in the initial step, cycloadditions provide different products.8 Generally, the reaction catalyzed by tertiary phosphine prefers “1,3-dipole” formation, leading to formal [3 + 2] cycloadducts. In contrast, tertiary amine catalysis generates “1,2-dipole”, resulting in formal [4 + 2] or [2 + 2] cycloaddition depending on the characteristics of the electrophile. α,β-Unsaturated carbonyl compounds are commonly used heterodienes together with allenoates in amine-catalyzed formal [4 + 2] cycloadditions (Fig. 2).9 In sharp contrast to the well-developed cycloaddition reactions of allenoates with various alkylidene nitrogen- and oxygen-containing heterocycles,10 there is no report on tertiary amine catalyzed formal [4 + 2] cycloaddition between alkylidene sulfur-containing heterocycles and allenoates.
Entrya | Catalyst | Solvent | Time (h) | Convers.b (%) | 3a/4ab | Yieldc (3a,%) | eed (3a, %) |
---|---|---|---|---|---|---|---|
a The reactions were conducted with 0.01 mmol of 1a, 0.12 mmol of 2a, and 20 mol% QD in 0.5 mL of solvent at rt. b Determined by 1H NMR of the crude product. c Isolated yields. d Determined by chiral HPLC analysis. e The reaction was conducted with 10 mol% 2,4-dinitrobenzoic acid as an additive. f The reaction was conducted with 5 mol% QD and 2.5 mol% 2,4-dinitrobenzoic acid as an additive. g Product precipitated from the reaction mixture. | |||||||
1 | QD | CHCl3 | 15 | 100 | 20:1 | 83 | 73 |
2 | CN | CHCl3 | 65 | 75 | 15:1 | 72 | 39 |
3 | C1 | CHCl3 | 15 | 100 | >20:1 | 91 | 66 |
4 | β-lCD | CHCl3 | 15 | 100 | >20:1 | 94 | 48 |
5 | (DHQD)2AQN | CHCl3 | 15 | 100 | 20:1 | 83 | 63 |
6 | (DHQD)2PHAL | CHCl3 | 65 | 70 | 8:1 | 47 | 64 |
7 | (DHQD)2PYR | CHCl3 | 15 | 90 | 11:1 | 43 | 71 |
8 | (DHQ)2PYR | CHCl3 | 15 | 100 | 20:1 | 86 | −68 |
9 | QD | Benzene | 15 | 100 | 20:1 | 93 | 79 |
10 | QD | THF | 15 | 100 | 19:1 | 90 | 78 |
11 | QD | DMF | 15 | 100 | 4:1 | 62 | 86 |
12 | QD | iPrOH | 15 | 100 | >20:1 | 77g | 67 |
13 | QD | MeOH | 15 | 100 | >20:1 | 82g | 89 |
14e | QD | MeOH | 15 | 97 | >20:1 | 92g | 90 |
15f | QD | MeOH | 72 | 90 | >20:1 | 75g | 86 |
Conversely, (DHQD)2AQN was identified to be most selective for the production of exo-olefinic product 3a, but unfortunately, with the lowest enantiocontrol from these types of catalyst. Further catalyst screening conducted with pseudoenantiomers of the abovementioned Cinchona alkaloid derivatives revealed preferential formation of the product with opposite enantiocontrol, ent-3a. For example, the model reaction of 1a with 2a catalyzed with (DHQ)2PYR afforded ent-3a in 86% yield with moderate enantioselectivity (68% ee, entry 8). For more details, see Table S1 in the ESI.†
According to our observations (Table 1), the model reaction tolerates a wide range of polar and non-polar solvents, such as chlorinated, aromatic, ethereal solvent as well as alcohols and DMF. Performing the model reaction in most of these solvents led to the formation of product 3a in high yields with moderate to good enantioselectivities (entries 1 and 9–13). The highest yields and enantioselectivities of 3a were achieved in alcohols (entries 12 and 13), where the highly enantiomerically enriched product 3a precipitated directly from the reaction mixtures. For example, the reaction in methanol afforded 3a in 82% yield with high enantioselectivity (89% ee, entry 13). For more details, see Table S2 in the ESI.† Subsequently, the presence of an additional Brønsted acid had a substantial effect on the reaction yield. When 2,4-dinitrobenzoic acid was used, product 3a was obtained in higher yield (92%, entry 14, for other additives see Table S3 in the ESI†). On the other hand, the reaction temperature and concentration did not have a marked effect on reaction efficiency and enantioselectivity (for details see Tables S4 and S5 in the ESI†). A slight decrease of reaction efficiency was observed upon reduction of catalyst loading and additives (entry 15, for details see Table S6 in the ESI†). The abovementioned results indicated that the most suitable conditions for formal [4 + 2] cycloaddition between 1a and 2a were achieved with quinidine and 2,4-dinitrobenzoic acid in methanol at room temperature (entry 13, Table 1).
After optimizing the reaction conditions, we began to explore the scope of organocatalytic cycloaddition reactions by varying sulfur-containing alkylidene 1 (Scheme 1). Cycloadducts 3b, c, and e derived from alkylidenes 1b, c, and e containing electron-withdrawing groups (EWGs) at position 5 of the benzo[b]thiophene ring were produced in high yields (72–87%) with a high degree of enantioselectivity (80–89% ee). Similarly, alkylidenes 1f–g substituted at position 6 or 7 of the benzo[b]thiophene ring showed comparable reactivity and similar enantiocontrol under the reported reaction conditions. Significantly reduced reactivity was observed, when the electron-donating group (EDG) was present at the benzo[b]thiophene ring. The observed low conversion of 1d may be explained by the electron-rich behavior of the starting material.
Next, we surveyed the scope of the reaction with various 3-subsituted vinyl benzothiopen-2-ones. As shown in Scheme 1, a variety of aryl vinyl derivatives bearing either EWG or EDG at the para-position are tolerated giving the desired products 3h–n in high yields (72–95%). 3-Benzylidene benzothiopen-2-ones with EWG substituents produced the corresponding products (3h–j) with a good level of enantiomeric purity (83–85% ee). When the strongly electron-withdrawing nitro group was present, enantioselectivity significantly dropped to 75% ee. Conversely, substrates bearing EDG (3l–m) were obtained in excellent yields with high enantiocontrol. For example, the reaction between p-methylbenzylidene benzothiopen-2-one 1m and 2a gave 3m in 95% yield with the enantioselectivity of 89% ee. Whereas the meta-substituted aryl derivative (1o) did not show any change in reactivity and selectivity from para-substituted analogues, ortho-substituted aryl vinyl benzothiophen-3-one (1p) produced cyclic products with diminished exo/endo selectivity (3p/4p = 6:1). Nevertheless, product 3p was isolated in acceptable yield (58%) with high enantiomeric purity (89% ee). Furthermore, the developed protocol was also applied to benzothiopen-2-ones bearing the heterocyclic moiety (1q) and alkyl groups (1r–s) affording cycloadducts 3q–s in good to acceptable yields. To our disappointment, the reaction with n-propyl derivative 1r did not reach full conversion even with a prolonged reaction time, and the corresponding product 3r was obtained in low yield.
To extend the substrate scope, other allenic substrates 2b–c were also examined (Scheme 1). As expected, methyl ester derivate 3t was obtained in high yield and with high enantioselectivity. On the other hand, the reaction with allenic ketones failed. No conversion to cycloadduct 3u was observed, probably due to increased tendency to be protonated prior to the formation of zwitterionic intermediates.
The developed organocatalytic procedure was also applied to alkylidenes derived from other sulfur-containing heterocycles 5a–c. Regioisomeric 2-benzylidenebenzo[b]thiophen-3(2H)-one (5a) exhibited lower reactivity, but the desired cycloadduct 6a was obtained after the prolonged reaction time in good yield with excellent selectivity (98% ee, Scheme 2).
Scheme 2 Substrate scope of the formal cycloaddition reaction with diverse sulfur containing alkylidenes 5. |
5-Membered heteroaromatic derivatives containing sulfur and nitrogen atoms can also be accommodated in the cycloaddition process. 5-Benzylidene-2-phenylthiazol-4(5H)-one 5b, as well as its isomeric analogue 5c, provided the corresponding products 6b and 6c, respectively, in good yields with high to excellent enantioselectivity. Noteworthily, cycloadducts 3 and 6 can also be isolated by simple filtration from the reaction performed in methanol, which supports the practical use of the developed process (Schemes 1 and 2). Using this separation technique, the corresponding products were obtained in high yields (up to 88%) with high enantioselectivities (up to 98% ee).
The applicability of the prepared cycloadducts was demonstrated on the following set of transformations. Cycloadducts 3 can be easily converted into the corresponding 4H-pyran derivatives 4. For example, compound 3a was converted by the treatment of DBU with pyran 4a in good yield (77%) without losing the enantiomeric purity (99% ee, Scheme 3). Similar results were obtained, when 3a was oxidized to sulfone 8 using MCPBA or selectively reduced to 9 under catalytic hydrogenation conditions. Furthermore, bromoderivative 3e was used in Suzuki coupling with aryl boronic acid, affording biphenyl compound 10 without losing the enantiomeric purity (Scheme 3).
The absolute configuration of cycloadducts 3 and 6 was assigned (3R) on the dihydropyran ring using X-ray diffraction analysis of compounds 3a and 6b (Fig. 3, for details see the ESI†).11
Fig. 3 X-ray structure of cycloadducts 3a and 6b.11 |
Based on the determined absolute configuration of 3 and the previous mechanistic studies,8,12 a reaction mechanism for an asymmetric formal [4 + 2] cycloaddition can be proposed (Scheme 4). Initially, zwitterionic intermediate II is formed via an addition of a tertiary amine catalyst to allenoate 2. Subsequently, an intermolecular attack of II to electron-deficient β carbon of 1 results in the formation of enolic intermediate III. The origin of a high stereocontrol of the attack in the case of arylidene derivatives can be explained by possible aromatic stacking stabilization between the aryl moiety of 1 and the quinoline ring of II. In the final step, enolic intermediate III undergoes 6-endo-trig ring-closure and the elimination of the catalyst affording the final product 3.
Footnotes |
† Electronic supplementary information (ESI) available. CCDC 1916993 and 1916994. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c9qo00886a |
‡ These authors contributed equally to this work. |
This journal is © the Partner Organisations 2019 |