triisobutylamine, 1:14, all under the standard conditions.3,4
We believe that the mechanistic pathway for pyridine as base
that favors â-lactone formation is most likely to be the
following:
120 °C, [R]23 -48.9 (c ) 0.65, CHCl3)) that exhibits a
D
carbonyl stretching band at 1826 cm-1 but no amide carbonyl
band (COOMe at 1722 cm-1). Physical characterization data
and procedures for the synthesis of 53 and acrylamide 2a4
appear below. The optimum base for formation of â-lactone
was found to be pyridine which resulted in a yield of 5 of
80% and a ratio of â-lactone to acrylamide of 8:1. An
ORTEP diagram for the bicyclic â-lactone 5 is displayed in
Figure 1.
If that is the case, it follows that the likely pathway from 4a
to the acrylamide 2a involves N-acylation of 4a by acrylyl
chloride itself, the function of the amine being simply to
serve as a stoichiometric proton acceptor. This interpretation
provides a logical explanation for the fact that the most
hindered bases, 2,6-di-tert-butylpyridine and triisobutylamine,
do not lead to â-lactone, but instead to acrylamide 2a. These
highly hindered bases would not be expected to provide
activated N-acrylylammonium complexes analogous to the
pyridinium complex 7. Nucleophilic attack by the secondary
amine on the activated electrophile 7, which could in
principle occur either at the CO group or at the carbon â to
CO, probably occurs preferentially at the latter for steric
reasons (the amino group in 4a is encumbered by the attached
substituents). Sterically preferred Michael addition of the
amine 4a to 7 leads to a keto ketene (8) which then cyclizes
by internal [2+2]-cycloaddition to give the â-lactones 5 and
6. This internal cycloaddition process, which could be very
useful in synthesis, is precedented by a small number of
examples in the literature.5
Figure 1. ORTEP diagram for 5.
The observation that the course of the reaction of acrylyl
chloride with the amino ketone 4a (or 4b) depends critically
on the nature of the amine coreactant has important
mechanistic implications. The ratio of acrylamide 2a to
â-lactones (5 + 6) has been found to vary with the catalytic
amine as follows: for 2,6-lutidine, 2.2:1; for pyridine, 8:1;
for Et3N, 2.5:1; for 2,6-di-tert-butylpyridine, >1:40; for
(4) (R)-Methyl-2-(benzyloxymethyl)-2-[N-(4-methoxybenzyl)acryl-
amido)]-3-oxobutanoate, 2a: To a solution of the amino ketone 4a (1.85
g, 5.0 mmol) in CH2Cl2 (10 mL) was added 2,6-di-tert-butylpyridine (2.81
g, 15.0 mmol) at 0 °C and then acrylyl chloride (0.80 mL, 10 mmol)
(dropwise with stirring at 0 °C). After 12 h at 23 °C (reaction complete by
TLC analysis), the reaction mixture was diluted with ethyl acetate (25 mL)
and washed with 2 N HCl and brine, dried over Na2SO4, and concentrated
under reduced pressure to afford the crude product. This mixture was
purified by flash chromatography (silica gel, ethyl acetate/hexane, 1:10 then
1:4) to afford pure amide 2a as a colorless solid (1.74 g, 82%). Found: Rf
) 0.80 (50% ethyl acetate in hexane); mp ) 85-86 °C; [R]23D -12.75 (c
1.45, CHCl3); FTIR (film) νmax 3030, 2995, 1733, 1717, 1510, 1256, 1178,
1088, 1027, 733, 697 cm-1; 1H NMR (CDCl3, 500 MHz) δ 7.30 (2H, d, J
) 8.0 Hz), 7.25 (3H, m), 7.11 (2H, m), 6.88 (2H, d, J ) 9.0 Hz), 6.38 (2H,
m), 5.63 (1H, dd, J ) 8.5, 3.5 Hz), 4.93 (1H, d, J ) 18.5 Hz), 4.78 (1H,
(3) (4R,5R,6R)-4-Benzyloxymethyl-3-(4-methoxybenzyl)-5-methyl-7-
oxo-6-oxa-3-aza-bicyclo[3.2.0]heptane-4-carboxylic Acid Methyl Ester,
5: To a solution of amino ketone 4a (1.85 g, 5.0 mmol) in CH2Cl2 (10 mL)
was added pyridine (1.22 mL, 15.0 mmol) at 0 °C and then acrylyl chloride
(0.80 mL, 10 mmol) (added dropwise with stirring at 0 °C). The reaction
mixture was stirred for 12 h at 23 °C, and the solvent was removed in
vacuo to afford the crude product. This mixture was purified by column
chromatography (silica gel, ethyl acetate-hexane, 1:10 then 1:4) to give
bicyclic â-lactone 5, the diastereomer 6 (1.70 g, 80%), and the acrylamide
2a (0.20 g, 10%). Crystallization of the mixture of lactones from ether-
hexane provided colorless crystals of 5. Found for 5: Rf ) 0.81 (50% ethyl
acetate in hexanes); mp ) 118-120 °C; [R]23D ) -48.9 (c 0.65, CHCl3);
FTIR (film) νmax 3050, 2990, 1826, 1722, 1511, 1241, 1102, 1032, 825,
1
699 cm-1; H NMR (CDCl3, 400 MHz) δ 7.30 (5H, m), 7.12 (2H, d, J )
8.8 Hz), 6.81 (2H, d, J ) 8.8 Hz), 4.55 (2H, dd, J ) 21.6, 12.0 Hz), 4.02
(1H, d, J ) 9.2 Hz), 3.92 (1H, d, J ) 13.6 Hz), 3.87 (1H, d, J ) 8.8 Hz),
3.81 (3H, s), 3.78 (3H, s), 3.52 (1H, d, J ) 6.4 Hz), 3.19 (1H, d, J ) 10.0
Hz), 3.16 (1H, d, J ) 13.6 Hz), 2.82 (1H, dd, J ) 10.0, 6.4 Hz), 1.83 (3H,
s); 13C NMR (CDCl3, 100 MHz) δ 170.09, 169.62, 159.07, 138.25, 130.31,
129.43, 129.17, 128.82, 128.52, 128.00, 127.77, 127.69, 114.09, 88.15,
74.13, 73.40, 70.82, 58.47, 55.50, 53.54, 51.70, 50.67, 19.80; HRMS (ESI)
calcd for C24H28NO6 (M + H)+ 426.1916, found 426.1918.
d, J ) 18.5 Hz), 4.27 (2H, m), 3.78 (3H, s), 3.76 (3H, s), 2.42 (3H, s); 13
C
NMR (CDCl3, 125 MHz) δ 198.12, 169.23, 168.62, 158.01, 136.95, 130.64,
130.38, 128.63, 128.13, 127.77, 127.32, 114.33, 77.49, 73.97, 70.66, 55.49,
53.09, 49.03, 28.24; LRMS (m/z, %) 426 (M+ + 1) (100); HRMS (ESI-
MS) calcd for C24H28NO6 426.1916, found 426.1909 (M + H)+. Because
triisobutylamine is much less expensive than 2,6-di-tert-butylpyridine and
the yield of 2a is only slightly less (78%), it is the preferred reagent for
larger-scale preparations.
1718
Org. Lett., Vol. 8, No. 8, 2006