Scheme 3
.
Boc-Activation of a Lactam and Reduction-
Scheme 4. Alternative Retro-Mannich Ring Opening
Solvolysis Retro-Mannich
retro-Mannich reaction under a variety of conditions, includ-
ing silica gel chromatography. Stirring 14 with silica gel
suspended in a methanol-methylene chloride mixture gave
the acyl-imine-enamide 15. The enolizable nature of all three
stereogenic centers in 15 made their preservation uncertain,
and the flexibility of this structure made NMR determination
of its structure problematic. A crystal structure of 15, grown
from methanol, however, confirmed that the stereochemical
relationship of these stereocenters was intact, as well as
formation of a new stereocenter resulting from addition of
methanol to the acylimine (Figure 2, 16).12
An alternative method for cleaving the cyclobutane also
took advantage of the divergent amide reactivities of 12
(Scheme 4). Treatment of 12 with sodium hydride and
iodomethane gave N-methylation of the less hindered amide.
The hindered nature of the remaining amide is evidenced
by the O-benzylation product 17 (2:1 O-benzyl:N-benzyl),
generated with sodium hydride and benzyl bromide. Treat-
ment of 17 with sodium cyanoborohydride reduced the
imidate to secondary amine 18. This amine was stable, but
when treated with sodium methoxide, a single product 20
was formed. This product had a new stereogenic center,
presumably set by kinetic protonation of 19. Reduction of
the imine 20 with sodium borohydride gave a 5:8 mixture
of two amines 21 and 22. The oxalate salt of 22 provided
an X-ray structure that defined the five stereogenic centers
resulting from this reaction sequence (Figure 2).13 Notably,
the three stereocenters originating in 12 remained in the same
relative configurations in product 22.
In these reaction paths, a [4 + 4]-photocycloaddition sets
four stereogenic centers in 10, fully controlled by the single
stereogenic center of the tether isobutyl group. None of these
four new stereocenters remain in the ultimate products 15,
16, 20, 21, or 22. Cope rearrangement of the [4 + 4] adduct
10 transfers two of those stereogenic centers to the cyclobu-
tane 11. The other two stereogenic centers disappear during
the retro-Mannich reactions. Nevertheless, in each case one
of the stereogenic centers that transiently disappear during
the retro-Mannich reaction is reformed with complete ste-
reocontrol (structures 16, 20). This control may be a function
of the seven-membered ether ring biasing the conformational
options during protonation of 19 and the addition of methanol
to imine 15. In the latter case, one can assume a reversible
addition to the imine, driven by either thermodynamics or
the crystallization, leads to formation of product 16, whereas
in the protonation of the amide enol(ate) 19, this stereo-
chemical outcome is most likely kinetic in origin. On the
other hand, the reduction of imine 20 to give a 5:8 mixture
of diastereomers 21 and 22 is more consistent with the
absence of stereochemical bias engendered by the seven-
membered ring. The crystal structures of 16 and 22 reveal
the rather open oxepin conformation (Figure 2).
(10) Compound 12, crystallized from methanol as the hydrate,
j
C16H24N2O3·H2O, in the rhombohedral space group R3 with a ) 28.990(2)
Å, c ) 10.1877(7) Å, V ) 7415.0(9) Å3, Z ) 18, and dcalc ) 1.251 g/cm3,
determined from a 0.30 mm × 0.28 mm × 0.18 mm crystal. X-ray intensity
data were collected on a Rigaku Mercury CCD area detector employing
graphite-monochromated Mo Ka radiation (λ ) 0.71069 Å) at a temperature
of 143 K. Refinement converged to R1 ) 0.0372 and wR2 ) 0.1008 for
2455 reflections for which F > 4σ(F) and R1 ) 0.0432, wR2 ) 0.1053 and
GOF ) 1.075 for all 2920 unique, nonzero reflections, and 202 variables.
CCDC 714967 contains the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cambridge
(11) Flynn, D. L.; Zelle, R. E.; Grieco, P. A. J. Org. Chem. 1983, 48,
2424–2426.
(12) Compound 16, C21H31N2O4, crystallized from methanol in the
(13) The oxalic acid salt of compound 22, C21H38N2O7, crystallized from
j
j
triclinic space group P1 with a ) 8.8340(10) Å, b ) 10.278(2) Å, c )
ethanol/ethyl acetate in the triclinic space group P1 with a ) 10.3811(13)
13.298(2) Å, R ) 90.009(11)°, ꢀ ) 106.126(7)°, γ ) 98.122(9)°, V )
1147.3(3) Å3, Z ) 2, and dcalc ) 1.087 g/cm3, determined from a 0.38 ×
0.18 × 0.005 mm crystal. X-ray intensity data were collected on a Rigaku
Mercury CCD area detector employing graphite-monochromated Mo Ka
radiation (λ ) 0.71069 Å) at a temperature of 143 K. Refinement converged
to R1 ) 0.0459 and wR2 ) 0.0999 for 1925 reflections for which F > 4σ(F)
and R1 ) 0.0886, wR2 ) 0.1240 and GOF ) 0.984 for all 3553 unique,
nonzero reflections, and 269 variables. CCDC 714968 contains the
supplementary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic Data Centre
Å, b ) 11.062(2) Å, c ) 11.7553(14) Å, R ) 109.649(2)°, ꢀ )
93.3740(10)°, γ ) 115.456(2)°, V ) 1114.6(2) Å3, Z ) 2, and dcalc ) 1.283
g/cm3, determined from a 0.40 × 0.23 × 0.06 mm crystal. X-ray intensity
data were collected on a Rigaku Mercury CCD area detector employing
graphite-monochromated Mo Ka radiation (λ ) 0.71069 Å) at a temperature
of 143 K. Refinement converged to R1 ) 0.0384 and wR2 ) 0.0983 for
2745 reflections for which F > 4σ(F) and R1 ) 0.0541, wR2 ) 0.1049 and
GOF ) 1.019 for all 3910 unique, nonzero reflections, and 278 variables.
CCDC 714969 contains the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cambridge
4542
Org. Lett., Vol. 11, No. 20, 2009