Charlie
Verrier
b and
Paolo
Melchiorre
*ab
aICREA – Catalan Institution for Research and Advanced Studies, Pg. Lluís Companys 23, 08010 Barcelona, Spain. E-mail: pmelchiorre@iciq.es; Tel: +34 977920208
bICIQ – Institute of Chemical Research of Catalonia, Av. Països Catalans 16, 43007 Tarragona, Spain
First published on 27th April 2015
Disclosed herein is a stereoselective method for the synthesis of 2,3-furan fused carbocycles bearing adjacent quaternary and tertiary carbon stereocenters. The chemistry is based on an asymmetric addition of β-ketoesters to 2-(1-alkynyl)-2-alkene-1-ones catalysed by natural cinchona alkaloids followed by a silver-catalysed intramolecular cycloisomerisation. By exploiting the distinct catalysis modes of quinine, which can act either as a general base or, upon opportune modifications, as a phase transfer catalyst, a complete switch of the enforced sense of diastereoinduction is achieved. The stereodivergent systems enable access to the full matrix of all possible stereoisomeric products.
Herein, we describe a straightforward synthetic strategy for accessing six- and seven-membered-ring furan derivatives bearing adjacent quaternary and tertiary carbon stereocentres in very high yields and stereoselectivities. The chemistry, which uses readily available substrates and catalysts, is based on a two-step sequential process whereby an organocatalytic asymmetric addition of β-ketoesters 2 to cyclic 2-(1-alkynyl)-2-alkene-1-ones 1 (ref. 5 and 6) is followed by a silver-catalysed intramolecular cycloisomerisation of the transient allenyl ketone7 intermediate 3 (Scheme 1). Significantly, we have identified two distinct catalytic systems which infer complementary diastereoselectivities, thereby enabling access to the full complement of stereoisomers of the annulated products 4 and 5 at will.8
We focused on identifying chiral organocatalysts that could infer high stereocontrol in the initial 4,5′-addition reaction. Representative results of our extensive studies are listed in Table 1, with more details reported in the ESI.† Intriguingly, natural cinchona alkaloid derivatives,12 acting as general base catalysts,13 afforded impressive levels of stereoselectivity (4a formed with a dr of up to 19:1, and an ee of up to 99%, entries 1–4), largely outperforming any other synthetic catalyst tested (see ESI†).
Entry | Catalyst | Yieldb (%) | 4a:5a ratioc | ee (%) major isomerd |
---|---|---|---|---|
a Reactions were performed in DCM at −10 °C on a 0.2 mmol scale using 1.2 equiv. of 2a, with [1a]0 = 0.1 M. After 48 hours, the 4,5′-addition was quenched by filtration through a pad of silica. Upon evaporation of the solvent, the cycloisomerisation of the intermediate 3a was conducted by dissolving the crude residue in 2 mL of AcOEt and adding 10 mol% of AgNO3. b Yield of the isolated products 4a and 5a (diastereomeric mixture). c Diastereomeric ratio determined by 1H NMR analysis of the crude mixture upon cycloisomerisation. d Enantiomeric excess, as determined by HPLC analysis on chiral stationary phases, refers to the major diastereoisomer; the absolute configuration is specified between brackets. e [1a]0 = 0.5 M. f Performed at −20 °C in a 5:1 mixture of DCM/33% K2CO3 aq., with [1a]0 = 0.2 M. g Performed using 10 mol% of catalyst. | ||||
1 | QN | 66 | 19:1 | 99 (1R,2R) |
2 | CD | 62 | 12:1 | 99 (1R,2R) |
3 | QD | 82 | 18:1 | 98 (1S,2S) |
4 | CN | 84 | 17:1 | 98 (1S,2S) |
5e | QN | 83 | 19:1 | 99 (1R,2R) |
6 | OMe-QN | 15 | 1:1.5 | <5 |
7 | QN-OH | <5 | — | — |
8f,g | PTC-QN | 73 | 1:13 | 86 (1R,2S) |
9f | PTC-QD | 89 | 1:19 | 87 (1S,2R) |
Importantly, the use of the “pseudoenantiomeric” catalysts quinine (QN) and quinidine (QD) secured access to both antipodes of the adduct 4 with excellent selectivity (entries 5 & 3, respectively). Protection of the quinine hydroxy moiety resulted in a greatly reduced reactivity along with a complete loss of stereocontrol (entry 6). Mechanistically, this suggests that the cinchona catalysts might operate through a bifunctional activation mode, simultaneously binding and activating the two reacting partners.13a,14 Interestingly, the cupreidine derivative QN-OH, a catalyst with a proven ability to promote the highly stereoselective addition of β-ketoesters to cyclic enones, remained completely inactive in our system (entry 7).15
We then modified the cinchona alkaloid scaffold by alkylating the basic bridgehead nitrogen of the quinuclidine core, the classical approach for achieving catalysts suitable for use as phase transfer catalysts (PTCs).16 Among the many PTCs tested in the model reaction (see ESI† for details), the cinchona-derived trimeric species PTC-QN and PTC-QD,17 easily obtained by the poly-alkylation of quinine and quinidine with 1,3,5-tris(bromomethyl)benzene, provided the most interesting results. When performing the reaction in DCM and in the presence of 33% K2CO3 aq., a complete switch of the enforced sense of diastereoinduction was achieved, so that the adduct 5 was almost exclusively formed in high optical purity (entries 8 & 9).
These findings allowed us to fully control the stereochemical outcome of the process, enabling the generation of any stereoisomer of the annulated furans 4 and 5 at will. This is considered a challenging goal because when asymmetric catalysis is applied to processes that generate two stereogenic centres in one product, there is generally no obvious means of modifying a catalyst to modulate the relative sense of those two centres.8 In this case, the two diastereodivergent systems are based on different organocatalysts, but are derived from the common chiral core of natural cinchona alkaloids. The divergent stereocontrol arises from the ability of the cinchona catalysts to execute distinct modes of catalysis for activating the reagents (base catalysis vs. phase transfer catalysis, Scheme 2).
Having identified two distinct catalytic systems that can selectively channel the reaction manifolds toward complementary diastereochemical outcomes, we examined the scope of the two-step process using QN (5 mol%) as a general base catalyst. As revealed in Table 2, the method shows a wide substrate generality and an excellent level of stereoselectivity, providing access to a variety of complex annulated furans 4 adorned with two vicinal quaternary and tertiary stereocentres. We first tested the possibility of modifying the cyclic scaffold of the 2-(1-alkynyl)-2-alken-1-one component 1. The cycloheptenone derivative reacted smoothly to provide the seven-membered-ring furan 4b in high optical purity (entry 2). In contrast, the cyclopentenone derivative was not a suitable substrate, since a complete lack of reactivity in the silver-catalysed intramolecular cycloisomerisation step was observed. Different substitution patterns on the aromatic moiety of 1 were well-tolerated, regardless of their electronic properties (entries 3–5). In addition, an alkyne bearing a vinylic substituent (entry 6) provided the corresponding furan 4f with high stereocontrol, albeit with a moderate chemical yield. A limitation of the system is that we have thus far failed to react alkynes bearing alkyl or TMS groups, and linear substrates.
Entry | R 1, n | R 2 | R 3 | 4 | Yieldb (%) | drc | eed (%) |
---|---|---|---|---|---|---|---|
a Reactions were performed at −10 °C on a 0.2 mmol scale and using 1.2 equiv. of 2. After 48 hours, the organocatalytic 4,5′-addition was quenched by filtration through a pad of silica. Upon evaporation of the solvent, the crude residue was dissolved in 2 mL of AcOEt and 10 mol% of AgNO3 was added. b Yield of the isolated products 4 (diastereomeric mixture). c Diastereomeric ratio determined by 1H NMR analysis of the crude mixture upon cycloisomerization. d Enantiomeric excess determined by HPLC analysis on chiral stationary phases. e The absolute configuration of 4h was unambiguously inferred by X-ray analysis, see ref. 18. | |||||||
1 | Ph, 1 | H | H | a | 83 | 19:1 | 99 |
2 | Ph, 2 | H | H | b | 75 | 12:1 | 97 |
3 | 4-MeOC6H4, 1 | H | H | c | 83 | 19:1 | 99 |
4 | 4-CF3C6H4, 1 | H | H | d | 90 | 19:1 | 99 |
5 | 4-MeC6H4, 1 | H | H | e | 83 | 19:1 | 99 |
6 | H | H | f | 25 | 19:1 | 95 | |
7 | Ph, 1 | MeO | H | g | 62 | 19:1 | 99 |
8 | Ph, 1 | H | Br | h | 82 | 19:1 | 98 |
9 | Ph, 1 | MeO | MeO | i | 67 | 19:1 | 99 |
10 | Ph, 1 | H | Me | j | 87 | 19:1 | 97 |
11 | Ph, 1 | F | H | k | 88 | 19:1 | 98 |
As for the nucleophilic partners 2, electronic variations in the indanone ring were possible, as both electron donating and withdrawing substituents gave the desired products in high yields and excellent diastereo- and enantioselectivities (entries 7–11). Efforts to react six-membered cyclic and linear β-ketoesters need further optimization, since only traces of the corresponding products could be obtained.
We then evaluated the synthetic potential of the PTC-mediated system. As depicted in Table 3, the reactions catalysed by 5 mol% of PTC-QD showed a comparable versatility to the system under general base catalysis, but secured a complementary diastereoselectivity, since the opposite diastereoisomers 5 of the annulated furans were almost exclusively formed with high enantiocontrol.
Entry | R 1, n | R 2 | R 3 | 5 | Yieldb (%) | drc | eed (%) |
---|---|---|---|---|---|---|---|
a Reactions were performed at −20 °C in a 5:1 mixture of DCM/33% K2CO3 aq., on a 0.2 mmol scale and using 1.2 equiv. of 2; with [1a]0 = 0.2 M. After 24 hours, the 4,5′-addition was quenched by filtration through a pad of silica. Upon evaporation of the solvent, the crude residue was dissolved in 2 mL of AcOEt and 10 mol% of AgNO3 was added. b Yield of the isolated products 5 (diastereomeric mixture). c Diastereomeric ratio determined by 1H NMR analysis of the crude mixture upon cycloisomerisation. d Enantiomeric excess determined by HPLC analysis on chiral stationary phases. e The absolute configuration of 5a was unambiguously inferred by X-ray analysis, see ref. 18. f Reactions were performed at −10 °C using 1.5 mol% of catalyst. | |||||||
1e | Ph, 1 | H | H | a | 89 | 19:1 | 87 |
2 | Ph, 2 | H | H | b | 81 | 9.5:1 | 93 |
3 | 4-MeOC6H4, 1 | H | H | c | 87 | 19:1 | 89 |
4 | 4-CF3C6H4, 1 | H | H | d | 84 | 8:1 | 81 |
5 | 4-MeC6H4, 1 | H | H | e | 82 | 19:1 | 91 |
6f | H | H | f | 53 | 19:1 | 92 | |
7f | Ph, 1 | MeO | MeO | i | 62 | 16:1 | 72 |
8f | Ph, 1 | H | Br | h | 78 | 11:1 | 77 |
9 | Ph, 1 | H | Me | j | 87 | 19:1 | 87 |
10 | Ph, 1 | F | H | k | 87 | 15:1 | 85 |
11 | Ph, 1 | Br | H | l | 70 | 19:1 | 83 |
The relative and absolute configuration of products 4h and 5a were unambiguously inferred by anomalous dispersion X-ray crystallographic analysis.18
Footnote |
† Electronic supplementary information (ESI) available: Complete experimental procedures and full compound characterisation, including HPLC traces and NMR spectra (PDF). CCDC 1046068 and 1046069. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c5sc01052g |
This journal is © The Royal Society of Chemistry 2015 |