Organocatalytic Approach to Benzofused Nitrogen‐Containing Heterocycles: Enantioselective Total Synthesis of (+)‐Angustureine

Abstract

Simply selective: The organocatalytic intramolecular aza-Michael reaction (IMAMR) was employed as an enantioselective tool to generate enantiomerically enriched isoindolines, indolines, tetrahydroquinolines, and tetrahydroisoquinolines from o-substituted alkylidenes, anilines and benzylamines (see scheme). The methodology was applied to the synthesis of the natural product (+)-Angustureine. Pyrrolidine and piperidine benzofused heterocyclic ring systems have become attractive targets in organic and medicinal chemistry. These families of compounds are extremely valuable scaffolds due to their widespread occurrence in Nature, diverse biological activity, and interesting chemical properties.1 Therefore, the development of new strategies for the generation of these heterocyclic frameworks, especially in a chiral nonracemic form, is of great interest in organic synthesis. Regarding the piperidine family, strategies for the catalytic enantioselective preparation of tetrahydroquinolines are mainly based either on the asymmetric hydrogenation of the corresponding aromatic quinoline derivatives by using chiral metal catalysts2 or, more recently, chiral Brønsted acids;3 or alternatively an enantioselective aza-Diels–Alder reaction.4 In the same way, one of the common strategies used to access enantiomerically enriched tetrahydroisoquinolines is based on the asymmetric transfer hydrogenation of isoquinoline and dihydroisoquinoline skeletons.5 However, probably the most popular method used to access these derivatives relies on the addition of nucleophilic carbon species to the CN bond of dihydroisoquinolines.6 In the pyrrolidine benzofused family, the indoline core is the most important substructure, and their synthesis is nowadays the subject of intensive research. In recent years, several reports for the enantioselective formation of indolines have appeared in the literature.7 However, the construction of the isoindoline core is clearly underdeveloped and only one method for its catalytic enantioselective preparation has been reported very recently.8 The addition of nitrogen-centered nucleophiles to α,β-unsaturated systems, that is, the so-called aza-Michael reaction, is one of the simplest and most direct ways to create CN bonds. The intramolecular version is particularly relevant because it allows the direct generation of nitrogen-containing heterocycles. Despite its synthetic potential, examples of the enantioselective version of this intramolecular reaction9 are very scarce and all rely on the use of organocatalysts. One of these examples described the organocatalytic synthesis of tetrahydroisoquinolines by using amides as nitrogen nucleophiles, although with poor enantiocontrol.10 Very recently, the synthesis of pyrazolo–indole compounds was performed by using the indole nitrogen as a nucleophile in the presence of a cinchonidine-derived catalyst.11 The third example of an organocatalytic intramolecular aza-Michael reaction (IMAMR) reported to date was developed by our research group,12 and allowed for the generation of several five- and six-membered heterocycles with excellent enantioselectivities by using carbamates as nitrogen nucleophiles.13 We envisioned the possibility of performing this organocatalytic IMAMR on ortho-substituted anilines and benzylamines that have a pendant α,β-unsaturated moiety. Herein we report the easy and enantioselective preparation of tetrahydroquinolines, tetrahydroisoquinolines, indolines, and isoindolines by following the aforementioned transformation. The common synthetic strategy for the synthesis of these four valuable heterocyclic derivatives is depicted in Scheme 1. Our approach started with a cross metathesis (CM) reaction of the terminal alkenylic chain of ortho-substituted N-protected anilines and benzylamines with acrolein to afford the corresponding α,β-unsaturated derivatives that, in turn, were subjected to an organocatalytic IMAMR and subsequent aldehyde reduction. The starting N-protected amines 1–4 (Table 1) were prepared according to procedures previously described in the literature (see the Supporting Information). The synthetic strategy. PG=Protecting group. Entry Substrate n m PG[a] Product Yield[b] [%] 1 1 a 0 1 Cbz 5 a 60 2 1 b 0 1 Boc 5 b 70 3 1 c 0 1 Ts 5 c 65 4 1 d 0 1 Ac 5 d 35 5 2 1 0 Cbz 6 50 6 3 1 1 Cbz 7 60 7 4 2 0 Cbz 8 73 Michael acceptors (5–8) for the organocatalytic conjugate addition were assembled by a CM reaction between compounds 1–4 and acrolein, catalyzed by second-generation Hoveyda–Grubbs catalyst 9 Cl2(IMes)RuCH(o-iPrOC6H4). The reaction proceeded at room temperature in dichloromethane, giving rise to the desired aminoaldehydes 5–8 as stable products in all cases and in moderate to good yields (Table 1). With α,β-unsaturated aldehydes 5–8 in hand, the first step of our study was the optimization of the cyclization process in terms of the reaction conditions and catalyst employed. We decided to examine the process on N-Cbz benzylamine 5 a as a model substrate. Diarylprolinols I and II and imidazolidinone III were chosen as catalysts for the organocatalytic process.14 The screening process to identify the optimum conditions is summarized in Table 2.0 Entry Catalyst Solvent T [°C] t [h] 10 a[a] [%] ee[b] [%] 1 I CHCl3 0 1 78 82 2 I CHCl3 −20 12 70 92 3 I CHCl3 −30 12 72 98 4 I CHCl3 −40 12 56 >99 5 I toluene 0 48 45 78 6 I THF −40 to RT 24 trace – 7 II CHCl3 −30 16 68 >99 8 III CHCl3 −30 12 72 54 The first attempt was performed at 0 °C with catalyst I (20 mol %) in CHCl3, in the presence of benzoic acid (0.2 equiv) as the cocatalyst. The reaction took 1 h to run to completion, affording the desired isoindoline 10 a in 78 % yield and 82 % enantiomeric excess (ee) after aldehyde reduction (Table 2, entry 1). With this encouraging result, the next reaction was carried out at −20 °C to give 10 a in 70 % yield and 92 % ee after 12 h (Table 2, entry 2). To our delight, when the temperature was lowered to −30 °C, the product was obtained in 72 % yield with an excellent 98 % ee (Table 2, entry 3). Lower temperatures (−40 °C) led to complete selectivity (>99 % ee) although the yield was slightly inferior. Other solvents, such as toluene, made the process less efficient because 10 a was obtained after 48 h in 45 % yield and the ee dropped to 78 %. In THF the process did not take place even after the reaction mixture was warmed up to room temperature (Table 2, entries 5 and 6). It is worth noting that catalyst II also afforded the final product with an excellent 99 % ee at −30 °C (Table 2, entry 7). However, under the same reaction conditions catalyst III afforded isoindoline 10 a in a modest 54 % ee (Table 2, entry 8). Therefore, optimized conditions were found to involve exposure of substrate 5 a to either catalyst I or II in the presence of benzoic acid in CHCl3 at −30 °C, which made the preparation of isoindoline (+)-10 a in good yield and excellent ee possible (Table 2, entries 3 and 7). To explore the effects of nitrogen substitution on enantioselectivity, substrates 5 a–d, which have carbamates, sulfonamides, or acetamides as nitrogen-protecting groups, were synthesized. When the Boc protecting group was used (i.e., 5 b), the process took place with a similar yield but with a slight decrease in the ee compared to 5 a (Table 3, entries 1,2 vs. 3,4). The use of tosylamide 5 c as the starting material followed the same trend, giving rise to isoindoline 10 c in good yield and ee with catalysts I and II (Table 3, entries 5 and 6). Finally, the use of acetamide 5 d as the nitrogen nucleophile in the IMAMR produced a dramatic decrease of the yield and rendered 10 d with modest ee (Table 3, entries 7 and 8). Entry 5 PG Catalyst 10 Yield [%] ee[a] [%] 1 5 a Cbz I 10 a 72 98 2 5 a Cbz II 10 a 68 99 3 5 b Boc I 10 b 72 94 4 5 b Boc II 10 b 68 91 5 5 c Ts I 10 c 61 94 6 5 c Ts II 10 c 63 91 7 5 d Ac I 10 d 36 64 8 5 d Ac II 10 d 28 58 With these results in hand, it became apparent that the decrease of the nucleophilicity of the nitrogen source in the IMAMR resulted in lower yields and ee values of the final products. Thus, conjugate addition with acetyl amides is much slower (if compared with carbamates and tosyl amides), and alternative reaction pathways promoted by protons liberated during the process (Brønsted acid catalysis) apparently became more important. This is a nonselective process that competes with the iminium activation by the organocatalyst, which translates into the decrease of the selectivity in the overall process. The extension of this protocol to the synthesis of several enantiomerically enriched benzofused heterocycles was examined next. Thus, when aniline derivative 6 was subjected to the optimized conditions, indoline (+)-11 was obtained in good yield and excellent ee (93 % and 92 % with catalysts I and II, respectively; Table 4, entries 1 and 2). When benzylamine derivative 7 was employed, the process gave tetrahydroisoquinoline (+)-12 in a complete enantioselective fashion (99 % ee; Table 4, entries 3 and 4). Finally, when the organocatalytic protocol was performed on aniline derivative 8, tetrahydroquinoline (+)-13 was formed in up to 92 % ee (Table 4, entries 5 and 6). Entry n m Substrate Catalyst Alcohol Yield [%] ee[a] [%] 1 1 0 6 I 70 93 2 1 0 6 II 67 92 3 1 1 7 I 67 99 4 1 1 7 II 63 99 5 2 0 8 I 70[b] 92 6 2 0 8 II 60[b] 91 It is worth noting that the IMAMR with benzylamines 5 and 7 led to slightly better ee values than aniline derivatives 6 and 8. This difference between both types of substrates would be rationalized again in terms of nitrogen nucleophilicity because benzylamines possess a more nucleophilic carbamate and afford the corresponding bicyclic derivatives with nearly complete selectivity. The application of the present method to the enantioselective synthesis of biologically active tetrahydroquinoline alkaloid Angustureine15 proved its utility.16 Therefore, after the organocatalytic aza-Michael reaction on compound 8, under the aforementioned conditions, aldehyde 14 was obtained in 68 % yield. Wittig reaction with propyltriphenylphosphonium bromide, followed by carbamate reduction with LiAlH4 and palladium-catalyzed hydrogenation, afforded the desired natural product (S)-(+)-15 in a very simple manner (Scheme 2). Additionally, oxidation of aldehyde 14 and further esterification with trimethylsilyldiazomethane gave rise to tetrahydroquinoline β-amino acid derivative (+)-16 (Scheme 2). Enantioselective synthesis of (+)-Angustureine 15. The synthesis of (S)-(+)-Angustureine allowed us to indirectly determine the absolute configuration of tetrahydroquinoline 13. The newly created stereocenter was thus determined to be R, which is in agreement with the commonly accepted mechanism to rationalize organocatalytic processes through iminium activation. Once the iminium ion is formed in its preferred E conformation (Scheme 3, intermediate A), the attack on the nitrogen source that gives rise to enamine intermediate B (Scheme 3) takes place from the bottom side (Re face) of the diene moiety because the upper side is shielded by the pyrrolidine substituent.17 The stereochemical outcome of the IMAMR on substrates 5–8. In conclusion, we have developed an efficient and simple method for the enantioselective synthesis of indolines, isoindolines, tetrahydroquinolines, and tetrahydroisoquinolines by means of an organocatalytic IMAMR of the corresponding aniline and benzylamine derivatives. The process took place with good yields and excellent ee values if diarylprolinols I and II were used as catalysts. The application of this method to the synthesis of the alkaloid (+)-Angustureine has also been presented. An investigation into further applications of this methodology is currently underway. Synthesis of adducts 10–13: In a flame-dried, 10 mL, round-bottomed flask, unsaturated aldehydes 5–8 (1 equiv) were dissolved in dry chloroform (0.1 M) and the solution was cooled to −30 °C. A mixture of catalyst I or II (20 mol %) and benzoic acid (0.2 equiv) in chloroform was added to this solution, and the resulting solution was stirred at this temperature for 12 h (except for substrate 8, which was maintained for 24 h). The mixture was then diluted with methanol and NaBH4 (3 equiv) was added portionwise. The mixture was allowed to reach 0 °C and, after 30 min at this temperature, the reaction was quenched with saturated NH4Cl and extracted with dichloromethane (3×10 mL). The organic extracts were washed with brine, dried over anhydrous Na2SO4, and concentrated to dryness under vacuum. After flash chromatography over silica gel with mixtures of hexane/ethyl acetate as the eluent, the corresponding alcohols 10–13 were obtained as colorless oils. The enantiomeric ratios were determined by using HPLC analysis with a Chiracel IC column (25×0.46 cm). Representative example: (1S)-N-Benzyloxycarbonyl-1-(2-hydroxyethyl)isoindoline (+)-10 a was isolated by following the aforementioned procedure and using flash chromatography with hexanes/ethyl acetate 3:1 as the eluent to give a colorless oil (72 % yield, 98 % ee with catalyst I; 68 % yield, 99 % ee with catalyst II). The ee was determined by using HPLC analysis with a Chiralpack IC column (eluent: hexane/isopropanol 87:13); flow rate=1.1 mL min−1, tmajor=36.9 min, tminor=39.2 min. [α]=+21.0 (c=1.0 in CHCl3); 1H NMR (300 MHz): δ=1.60–1.69 (m, 1 H), 2.16–2.27 (m, 1 H), 3.70 (br s, 2 H), 3.96 (br s, 1 H), 4.63 (d, J=14.9 Hz, 1 H), 4.91 (d, J=14.9 Hz, 1 H), 5.22 (s, 2 H), 5.38 (d, J=10.0 Hz, 1 H), 7.21–7.40 ppm (m, 9 H); 13C NMR (75 MHz): δ=40.4 (CH2), 51.8 (CH2), 59.0 (CH2), 60.6 (CH), 67.5 (CH2), 122.4 (CH), 127.6 (2 CH), 128.0 (CH), 128.2 (CH), 128.6 (CH), 135.8 (C), 136.3 (C), 141.3 (C), 156.6 ppm (C); HMRS (EI+): m/z calcd for C18H19NO3 [M+]: 297.1365; found: 297.1360. We thank the Ministerio de Educación y Ciencia (CTQ2007-61462) and the Generalitat Valenciana (GV/2007/013), Spain, for their financial support. J.M., M.D.P.-C., and D.J. express their thanks for predoctoral fellowships, M.S.-R. for a postdoctoral contract, and C.P. for a Ramón y Cajal contract. Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.