The aza-Cope/Mannich reaction is the transformation of an appropriately substituted unsaturated iminium ion to an acyl-substituted pyrrolidine via [3,3]-sigmatropic rearrangement followed by Mannich cyclization. The reaction represents a synthetically useful method for synthesizing pyrrolidines with high diastereoselectivity.
The [3,3]-sigmatropic or Cope rearrangement involves the reorganization of σ and π bonds among six contiguous atoms. It has been observed that placing positive charge on one of these atoms decreases the activation barrier of the rearrangement signficantly. Thus, unsaturated iminium ions rearrange significantly more rapidly than their corresponding neutral all-carbon analogues. However, in the absence of a thermodynamic bias for the rearrangement equilibrium, useful amounts of products cannot be isolated. One strategy for introducing such a bias involves designing a substrate such that Cope rearrangement leads to proximal enol and iminium groups. Upon rearrangement, an irreversible Mannich cyclization may then occur to afford acyl-substituted pyrrolidines (Eq. 1).(1)
This sequence, known as the aza-Cope/Mannich reaction, has become a synthetically useful method for pyrrolidine synthesis. Because the aza-Cope rearrangement proceeds through a chair-like transition state, high levels of diastereoselectivity are common when chiral substrates are used.
Mechanism and Stereochemistry
Two plausible mechanisms for the overall transformation are outlined in Eq. 2. One involves an aza-Cope rearrangement followed by a Mannich cyclization, as described above. Alternatively, an aza-Prins process involving electrophilic attack of the iminium ion on the carbon-carbon double bond may occur first. Subsequent pinacol rearrangement leads to the same product as the aza-Cope/Mannich sequence.(2)
Two observations support the aza-Cope/Mannich mechanism for a wide variety of substrates. The first is that as the reaction proceeds, epimerization occurs at the carbon bearing R1. Only the aza-Cope/Mannich mechanism involves planarization of the carbon bearing R1, as the aza-Prins/pinacol mechanism involves retentive migration of the stereocenter bearing R1. Thus, the aza-Prins/pinacol mechanism is unlikely.
The second is that the rate and yield of the reaction are independent of the nature of the R2 substituent. The reaction proceeds well even when R2 is a strongly electron-withdrawing sulfone substituent. This observation is inconsistent with the formation of a cation at the carbon bearing R2, which is required for the aza-Prins/pinacol mechanism.
Electronically biased substrates, such as N-sulfonyliminium ions and enol ethers, may react via the aza-Prins/pinacol mechanism.
The aza-Cope rearrangement proceeds through a highly ordered, chair-like transition state. Because stereomutation of the intermediate enol iminium ion is slow relative to Mannich cyclization, high diastereoselectivity is often observed when chiral substrates are employed. This stereoselectivity is a direct consequence of the chair-like transition state of rearrangement. Large groups prefer to occupy pseudoequatorial positions in the transition state (Eq. 3). Consistent with the idea that the reaction occurs through a chair-like transition structure, substrates containing (Z) double bonds react more slowly and with lower stereoselectivity than (E) substrates.(3)
Scope and Limitations
The aza-Cope/Mannich reaction generally proceeds well, provided the unsaturated iminium ion substrate can be prepared. Thus, the scope of the reaction is limited primarily by methods for preparing 1-aminobut-3-en-2-ol precursors and generating iminium ions.
Preparing 1-Aminobut-3-en-2-ol Precursors
1-Aminobut-3-en-2-ols required for the aza-Cope/Mannich reaction can be prepared through the reaction of amines with unsaturated epoxides, which in turn can be generated in situ from 1-bromo-2-hydroxybut-3-enes. Subsequent methylation with potassium hydride and iodomethane leads to enol ether precursors (Eq. 4).(4)
Direct reactions of amines with alkyl bromides that cannot form epoxides (such as ethers) often result in over-alkylation. Amide bond formation followed by reduction avoids these problems (Eq. 5).(5)
Whereas the previous two methods have involved carbon-nitrogen bond formation, carbon-carbon bond formation from vinylation of ketones or aldehydes also represents a viable route to aza-Cope/Mannich precursors. The diastereoselectivity is high in additions to cyclic ketones (Eq. 6).(6)
Iminium Ion Generation
The most common method for generating iminium ions for the aza-Cope/Mannich reaction involves acid-promoted condensation of a sterically unhindered aldehyde with a 1-aminobut-3-en-2-ol precursor (Eq. 7). Typically, only sub-stoichiometric quantities of acid are needed. Camphorsulfonic acid (CSA) is commonly employed in this role.(7)
Ketones and sterically hindered aldehydes are not suitable for the method in Eq. 7. Dehydrative oxazoline formation followed by heating in the presence of a full equivalent of acid is an alternative method that works for these classes of substrates (Eq. 8).(8)
Alkene and Amine Scope
Both 1,1- and 1,2-disubstituted alkenes undergo the reaction efficiently. A single account of successful reaction of a trisubstituted alkene has been reported. Chiral (E)-alkenes typically react with high diastereoselectivity (Eq. 9); stereoselectivity in reactions of (Z)-alkenes is typically lower (see Stereochemistry above).(9)
Amine substituents have a profound influence on the stereoselectivity of the reaction, because they influence the preferred configuration of the iminium ion prior to sigmatropic rearrangement. Large substituents favor the (Z)-iminium ion, which reacts to give the syn diastereomer preferentially. Small substituents favor the (E)-iminium ion, which reacts to give the anti diastereomer preferentially (Eq. 10).(10)
Because the aza-Cope/Mannich reaction is intramolecular and requires only heat, it may be used to prepare sterically crowded, pyrrolidine-containing cyclic frameworks. Although the starting material may require several steps to prepare, the high stereoselectivity of the aza-Cope/Mannich reation often justifies this extra effort. A prominent example of the use of this reaction is the first asymmetric synthesis of (–)-strychnine (Eq. 11). Heating the precursor in paraformaldehyde and sodium sulfate afforded the aza-Cope/Mannich product in 98% yield. Under these conditions, iminium ion formation is catalyzed by trace formic acid present in the paraformaldehyde.(11)
Comparison to Other Methods
A variety of alternative methods exist to synthesize the pyrrolidine skeleton, but few match the mildness, stereoselectivity, and substrate scope of the aza-Cope/Mannich method. Perhaps the most closely related method is the aza-Cope/aza-Prins reaction, which involves [3,3]-sigmatropic rearrangement followed by aza-Prins cyclization (Eq. 12). The formation of a carbocation after aza-Prins cyclization limits the scope of this reaction.(12)
1,3-Dipolar cycloaddition of azomethine ylides and alkenes is a second alternative for pyrrolidine synthesis (Eq. 13). Electron-deficient alkenes are necessary, but catalytic, enantioselective variants of this reaction exist.(13)
Palladium-catalyzed cycloaddition of trimethylenemethane and imines also produces the pyrrolidine skeleton via 1,3-dipolar cycloaddition. Because of the nature of the dipole, only pyrrolidines containing an alkylidene group in the C4 position are typically accessible.
Experimental Conditions and Procedure
Experimentally, the aza-Cope/Mannich reaction is straightforward. The aldehyde, 1-aminobut-3-en-2-ol precursor, and a Brønsted acid may be added in any order. No reaction ordinarily takes place until the reaction mixture is heated to 50-85 °C. Anhydrous solvents are typically used, and sodium sulfate may be added to promote dehydrative condensation; however, the strict exclusion of water and air does not appear to be essential. A variety of solvents have been employed, although benzene, toluene, and acetonitrile are most common.
A mixture of 1-benzylamino-2-methylbut-3-en-2-ol (450 mg, 2.35 mmol), paraformaldehyde (78 mg, 2.6 mmol), camphorsulfonic acid (495 mg, 2.13 mmol), and anhydrous benzene (5 mL) was heated at reflux for 10 hours under Ar. After cooling to room temperature, 1 N NaOH (3 mL) was added, the organic layer was separated, and the aqueous layer was extracted with ether (3 X 5 mL) . The combined organic extracts were dried (Na2SO4) and concentrated, and the residue was purified by bulb-to-bulb distillation (oven temperature 95 °C, 0.01 mm) to give 425 mg (89%) of the product as a colorless oil, which was >95% pure by GC analysis. A specimen was purified by preparative GC to yield an analytical sample: IR (film) 1712 cm-1; 1H NMR (250 MHz, CDCl3) 7.2-7.4 (m, Ph), 3.62 (br s, 2H) , 3.05-3.2 (m , 1H) , 2.75-2.9 (m, 1 H) , 2.45-2.75 (m , 3H) , 2.15 (s, 3H) , 1.95-2.2 (m, 2H) ; 13C NMR (63 MHz, CDCl3) 208.8, 139.0, 128.7, 128.3, 127.1, 60.2, 55.7, 53.9, 50.5, 28.5, 26.6; CI-HRMS (m/z) [M + H<sup>+</sup>] calcd for C13H18NO, 204.1388; found, 204.1380.
- ↑ Horowitz, R. M.; Geissman, T. A. J. Am. Chem. Soc. 1950, 72, 1518.
- ↑ a b Jacobsen, E. J.; Levin, J.; Overman, L. E. J. Org. Chem. 1988, 110, 4329.
- ↑ Armstrong, A.; Shanahan, S. E. Org. Lett. 2005, 7, 1335.
- ↑ a b Overman, L. E.; Mendelson, L. T; Jacobsen, E. J. J. Am. Chem. Soc. 1983, 105, 6629.
- ↑ Overman, L. E.; Kakimoto, M.-A. J. Am. Chem. Soc. 1979, 101, 1310.
- ↑ Doedens, R. J.; Meier, G. P.; Overman, L. E. J. Org. Chem. 1988, 53, 685.
- ↑ Overman, L. E.; Mendelson, L. T.; Flippin, L. A. Tetrahedron Lett. 1982, 23, 2733.
- ↑ Overman, L. E.; Kakimoto, M.; Okawara, M. Tetrahedron Lett. 1979, 20, 4041.
- ↑ Overman, L. E.; Wild, H. Tetrahedron Lett. 1989, 30, 647
- ↑ Overman, L. E.; Trenkle, W. C. Isr. J. Chem. 1997, 37, 23.
- ↑ Knight, S. D.; Overman, L. E.; Pairaudeau, G. J. Am. Chem. Soc. 1993, 115, 9293.
- ↑ Overman, L. E.; Mendelson, L. T. J. Am. Chem. Soc. 1981, 103, 5579.
- ↑ Agami, C.; Couty, F.; Persoulis, M. Synlett 1992, 847.
- ↑ Padwa, A.; Chen, Y. Y.; Dent, W.; Nimmesgern, H. J. Org. Chem. 1985, 50, 4006.
- ↑ Yamago, S.; Nakamura, E. Org. React. 2002, 61, 1.
- ↑ Overman, L. E.; Kakimoto, M.; Okazaki, M. E.; Meier, G. P. J. Am. Chem. Soc. 1983, 105, 6622.