Nitrone-olefin 3+2 cycloaddition

The correct title of this article is Nitrone-olefin [3+2] cycloaddition. The substitution or omission of any < > [ ] { } is due to technical restrictions.

The nitrone-olefin [3+2] cycloaddition reaction is the combination of a nitrone with an alkene or alkyne to generate an isoxazoline or isoxazolidine via a [3+2] cycloaddition process.^[1]

Introduction

When nitrones are combined with either alkenes or alkynes, [3+2] cycloaddition leads to the formation of a new C-C bond and a new C-O bond. The cycloadditions is stereospecific with respect to the configuration of the alkene; however, diastereoselectivity in reactions of C-substituted nitrones is often low.^[2] Regioselectivity is controlled by the dominant frontier orbitals interacting during the reaction, and substrates with electronically distinct substituents tend to react with high regioselectivity. Intramolecular versions of the reaction have been used to synthesize complex polyclic carbon frameworks. Reduction of the N-O linkage leads to 1,3-aminoalcohols.

(1)

Mechanism and Stereochemistry

Synthesis of Nitrones

Nitrones are generated most often either by the oxidation of hydroxylamines or condensation of monosubstituted hydroxylamines with carbonyl compounds (ketones or aldehydes). The most general reagent used for the oxidation of hydroxylamines is mercury(II) oxide.^[3]

(2)

Carbonyl condensation methods avoid issues of site selectivity associated with the oxidation of hydroxylamines with two sets of (alpha) hydrogens.^[4]

(3)

A significant problem associated with many reactive nitrones is dimerization.^[5] This issue is alleviated experimentally by employing an excess of the nitrone or increasing the reaction temperature to exaggerate entropic factors.

(4)

[3+2] Cycloaddition

The [3+2] cycloaddition itself is a concerted, pericyclic process whose regiochemistry is controlled by the frontier molecular orbitals on the nitrone (the dipole) and the dipolarophile.^[6] When R' is an electron-donating group, alkyl, or aryl, the dominant FMOs are the HOMO of the dipolarophile and the LUMO of the nitrone. Thus, connecting the atoms whose coefficients in these orbitals are largest, the 5-substituted isoxazolidine is predicted to predominate. On the other hand, when the dipolarophile is electron poor, the HOMO_nitrone-LUMO_{dipolarophile} interaction is most important, and the 4-substituted product is favored.

(5)

Scope and Limitations

Alkyl and aryl terminal alkenes react with high regioselectivity to give 5-substituted isoxazolidines. This outcome is consistent with frontier molecular orbital (kinetic) control of the distribution of isomers: the nitrone oxygen, which possesses the largest orbital coefficient in the HOMO of the nitrone, forms a bond with the inner carbon of the alkene, which possesses the largest orbital coefficient in the LUMO of the alkene.^[7]

(6)

The configuration of 1,2-disubstituted alkenes is maintained in the products of cycloaddition. Consistent with FMO control of the reaction, the more electron-withdrawing substituent on these substrates ends up in the 4 position of the product. Put another way, the carbon with the largest LUMO coefficient in the dipolarophile (distant from the electron-withdrawing group) forms a bond with the nitrone oxygen, which possesses the largest HOMO coefficient in the nitrone.^[8]

(7)

Alkynes can also serve as dipolarophiles in this reaction. The rules for predicting alkene cycloaddition products based on the relevant FMOs apply to substituted alkynes as well—electron-poor alkynes tend to give 4-substituted products, while electron-rich, alkyl, and aryl alkynes give 5-substituted products.^[9]

(8)

Intramolecular variants of the reaction are very useful for the synthesis of complex polycyclic frameworks. These reactions generally take place at much lower temperatures than intermolecular cycloadditions. Regiochemistry is more difficult to predict for intramolecular reactions: either bridged or fused products can result, and both cis- and trans-fused rings are possible.^[10]

(9)

An existing stereocenter in the tether between the alkene and nitrone often leads to the generation of a single diastereomer of product. In this example, the bulkier phenyl substituent ends up on the exo face of the bicyclic ring system.^[11]

(10)

Synthetic Applications

The use of 2,3,4,5-tetrahydropyridine-1-oxide (THPO) is common for the construction of fused heterocycles in alkaloids and other natural products. A synthesis of (+)-lupinine uses a ring-expanding rearrangement of a mesylate, providing rapid access to the target.^[12]

(11)

The structure of hydroxycotinine, a human metabolite of nicotine, was confirmed via an independent synthesis employing nitrone-olefin cycloaddition.^[13]

(12)

Rearrangement of a [3+2] cycloadduct provides (+)-porantheridine.^[14] The cycloadduct is subjected to hydrogenation, acid hydrolysis, oxidation, basic hydrolysis, and cyclization to give the target.

(13)

Experimental Conditions and Procedure

Example Procedure^[15]

(14)

C-Phenyl-N-methylnitrone (1.00 g, 7.4 mmol) was dissolved in an excess of freshly distilled ethyl vinyl ether (25.0 mL, 161 mmol), and the solution was sealed in a thick-walled glass reaction tube. The tube was heated at 80° for 72 hours and then cooled, and the excess ethyl vinyl ether was removed by evaporation under high vacuum. The residue was chromatographed (CH₂Cl₂) to remove unreacted nitrone followed by bulb-to-bulb distillation at 145° (1 torr) to give a yellow oil (1.19 g, 78%). As evidenced by GLC (SE-30 column/140°) and NMR analysis, the reaction mixture contained a 50:50 mixture of cis and trans isomers (retention times 4.0 and 5.4 minutes, respectively). The isomers were separated by column chromatography (120 g; 10:1 hexane:ethyl acetate). The less polar isomer was the cis isomer, while the more polar compound was the trans isomer.

cis Isomer (oil): ¹H NMR (CDCl₃, 200 MHz) δ:1.25 (t, J = 7 Hz, 3H), 2.32 (ddd, J = 3, 10, and 13 Hz, 1H), 2.55 (s, 3H), 2.86 (ddd, J = 6, 10, and 13 Hz, 1H), 3.34 (t, J = 10 Hz, 1H), 3.45 (dq, J = 7 and 14 Hz, 1H), 3.93 (dq, J = 7 and 14 Hz, 1H), 5.15 (dd, J = 3 and 6 Hz, 1H), 7.33 (m, 5H); IR (CCl₄) 3100–3000 (m), 3000–2800 (vs), 1455 (s), 1370 (s), and 1100 (vs) cm⁻¹; mass spectrum (EI), m/z (relative intensity) 207 (M+, 4), 161 (36), 118 (100), 77 (75). Anal. Calcd. for C₁₂H₁₇NO₂: C, 69.54; H, 8.27; N, 6.76. Found: C, 69.44; H, 8.10; N, 6.69.

trans Isomer (oil): ¹H NMR (CDCl₃, 200 MHz) δ: 1.23 (t, J = 7 Hz, 3H), 2.41 (ddd, J = 5, 9, and 13 Hz, 1H), 2.57 (dd, J = 6 and 13 Hz, 1H), 2.78 (s, 3H), 3.48 (dq, J = 7 and 10 Hz, 1H), 3.86 (dq, J = 7 and 10 Hz, 1H), 4.03 (dd, J = 6 and 9 Hz, 1H), 5.16 (d, J = 5 Hz, 1H), 7.30 (m, 5H); IR (CCl₄) 3100–3000 (m), 3000–2800 (vs), 1455 (s), 1370 (s), and 1100 (vs) cm⁻¹; mass spectrum (EI), m/z (relative intensity): 207 (M+, 9), 161 (100), 134 (63), 118 (53), 77 (48). Anal. Calcd. for C₁₂H₁₇NO₂: C, 69.54; H, 8.27; N, 6.76. Found: C, 69.37; H, 8.10; N, 6.77.

References

1. ^ Confalone, P. N.; Huie, E. M. Org. React. 1988, 36, 1. doi: (10.1002/0471264180.or036.01)
2. ^ Tufariello, J. J.; Ali, S. A.; Klingele. H. O. J. Org. Chem. 1979, 44, 4213.
3. ^ Thesing, J.; Sirrenberg, W. Justus Liebigs Ann. Chem. 1957, 609, 46.
4. ^ Exner, O. Collect. Czech. Chem. Commun. 1951, 16, 258.
5. ^ Thesing, J.; Maver, H. Chem. Ber. 1956, 89, 2159.
6. ^ Houk, K. N. J. Am. Chem. Soc. 1972, 94, 8953.
7. ^ Huisgen, R.; Hauck, H.; Grashey, R.; Seidl, H. Chem. Ber. 1968, 101, 2568.
8. ^ Joucla, M. Tetrahedron 1973, 29, 2315.
9. ^ Winterfeldt, E.; Krohn, W.; Stracke, H. Chem. Ber. 1969, 102, 2346.
10. ^ LeBel, N. A.; Post, M. E.; Whang, J. J. J. Am. Chem. Soc. 1964, 86, 3759.
11. ^ Vinick, F. J.; Fengler, I. E.; Gschwend, H. W. J. Org. Chem. 1977, 42, 2936.
12. ^ Tufariello, J. J.; Tegeler, J. J. Tetrahedron Lett., 1976, 4037.
13. ^ Dagne, E.; Castagnoli, Jr., N. J. Med. Chem. 1972, 15, 356.
14. ^ Gossinger, E. Tetrahedron Lett. 1980, 21, 2229.
15. ^ DeShong, P.; Dicken, C. M.; Staib, R. R.; Freyer, A. J.; Weinreb, S. M. J. Org. Chem. 1982, 47, 4397.