to find that no reaction had taken place. However, adding
the base 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) to the
reaction mixture triggered the accumulation of oxazolidinone
3 as the only observed product. We assumed that the source
of carbon in this reaction was CO2, likely present in the
nondegassed commercial anhydrous THF used as solvent.
The fixation of CO2 in a zwitterionic DBU-CO2 complex
is a known process, as is the subsequent transcarboxylation
of this complex and amines.6 Thus, in the errant reaction it
appeared that DBU promoted the carboxylation of amine 1
to produce a carbamic acid intermediate that underwent
Mitsunobu cyclization to provide 3.
Following from this and Kodaka’s earlier observations,3
we became interested in further expanding the scope of a
tandem amino alcohol carboxylation and Mitsunobu reaction
strategy for preparing carbamates. Given the mildness with
which an amino alcohol could be carboxylated to give a
carbamic acid intermediate (e.g., 4 f 5, Scheme 1), the
DBU, tributylphosphine, and di-tert-butyl azodicarboxylate
under a balloon of CO2 gas resulted in 55% conversion to
an equal mixture of the desired oxazolidinone 6a and the
hydrazine adduct 7 (entry 1). Using substoichiometric DBU
gave the same result (entry 2). These results implied the
incomplete formation of carbamic acid intermediate (e.g., 4
f 5), allowing Mitsunobu alkylation of di-tert-butyl hydra-
zine-1,2-dicarboxylate to compete (4a f 7). Gratifyingly,
pretreating 4a with DBU-CO2 for 45 min prior to adding
the n-Bu3P and DBAD completely suppressed the formation
of byproduct 7 (entry 3).8 Finally, increasing the number of
equivalents of Mitsunobu reagents (entries 4 and 5) led to
complete and clean conversion of 4a to 6a, isolated in 92%
yield after purification.
These reaction conditions were applied to a variety of
substrates (Table 2). After the successful cyclization of the
simple N-substituted ethanolamine 4a (entry 1), we were
gratified to find that a variety of substitution patterns are
accommodated in high-yielding reactions. The aniline 4b
gave N-phenyloxazolidinone 6b in almost quantitative yield
(entry 2). Primary amines gave N-unsubstituted oxazolidi-
nones with substituents at either the 4-position (6c, 6d), the
5-position (6e), or both (6e, 6f) in good yields. Consistent
with earlier work,3 primary amine substrates 4e-g produced
6e-g with retention of configuration at the oxygen-bearing
center (entries 5-7). However, this was not the case for
analogous substrates containing secondary amines, which
unexpectedly gave oxazolidinones with inVersion of config-
uration at the oxygen-bearing center.9 This stereochemical
divergence was most evident in comparisons of 4g with 4h
(entry 7 vs 8) and 4i with 4j (entry 9 vs 10). Further exploring
the scope of the reaction, we found that 4k produced the
bicyclic oxazolidinone 6k efficiently and that the 1,3-amino
alcohol 8 cyclized readily to provide the six-membered
oxazinone 9 in high yield.
Scheme 1
stereoselectivity of the Mitsunobu transformation (e.g., 5 f
6), and the potential for adapting the reaction to other
substrate types, this method was anticipated to compliment
existing carbamate synthesis strategies.
The conversion of aminoethanol 4a to oxazolidinone 6a
was chosen to optimize reaction conditions (Table 1).
Intrigued by the N-substitution-dependent stereochemical
divergence observed in these reactions, we undertook isotope
labeling experiments designed to differentiate pathways
involving either retention or inversion of configuration in
the Mitsunobu cyclization.10 In the reaction of a carbamic
acid 5 to produce 6 (Scheme 1), the use of 18O-labeled carbon
Table 1. Optimization of Reaction Conditions
(6) (a) Mizuno, T.; Okamoto, N.; Ito, T.; Miyata, T. Tetrahedron Lett.
2000, 41, 1051. (b) Perez, E. R.; da Silva, M. O.; Costa, V. C.; Rodrigues-
Filho, U. P.; Franco, D. W. Tetrahedron Lett. 2002, 43, 4091.
(7) Alternatively, the use of Cs2CO3, either in the presence or absence
of added CO2, met with limited success, affording only partial conversion
of 4a to 6a. For reports of amine carboxylation using carbonate salts, see:
(a) Butcher, J. K. Synlett 1994, 825. (b) Inesi, A.; Mucciante, V.; Rossi, L.
J. Org. Chem. 1998, 63, 1337. (c) Salvatore, R. N.; Chu, F.; Nagle, A. S.;
Kapxhiu, E. A.; Cross, R. M.; Jung, K. W. Tetrahedron 2002, 58, 3329.
(8) In this and related reactions a precipitate often appears during the
CO2 pretreatment period, implying that an alkylammonium N-alkylcarbamate
salt (RNH3+‚-O2CNHR) is formed. The extent to which complete conver-
sion to carbamic acid 5 occurs before n-Bu3P and DBAD are added is
unknown. See: Hampe, E. M.; Rudkevich, D. M. Tetrahedron 2003, 59,
9619.
load timea
CO2
equiv of
conversionb ratiob
entry
DBU/n-Bu3P/DBAD
(%)
6a :7
1
2
3
4
5
0 min
0 min
45 min
45 min
45 min
1.1/1.1/1.1
0.1/1.1/1.1
0.1/1.1/1.1
0.1/1.5/1.5
0.1/2.1/2.1
55
57
74
95
100
52:48
50:50
100:0
100:0
100:0
a Time for which solution of 4a and DBU are pretreated with CO2 via
balloon, before n-Bu3P/DBAD were added. b Determined by LC-MS (UV,
λ ) 215 nm).
(9) Kodaka et al.3a had found that although 4c and 4i were converted to
6c and 6i with retention of configuration, using Ph3P instead of n-Bu3P
promoted inversion in the case of 4c f 6c. No secondary amine substrates
were evaluated for stereochemical outcome.
(10) Retention of configuration in the Mitsunobu reaction: (a) Ahn, C.;
DeShong, P. J. Org. Chem. 2002, 67, 1754. (b) Smith, A. B.; Safonov, I.
G.; Corbett, R. M. J. Am. Chem. Soc. 2002, 124, 11102. (c) Liao, X.; Wu,
Y.; De Brabander, J. K. Angew. Chem., Int. Ed. 2003, 42, 1648.
Consistent with literature examples of amine carboxylation,
we selected acetonitrile as solvent and DBU-CO2 as the
carboxylation promoter.6,7 Subjecting 4a to 1 equiv each of
2886
Org. Lett., Vol. 6, No. 17, 2004