led us to investigate the reactivity of the structurally related
N-(ethoxycarboxymethyl)oxazolidines under similar reaction
conditions.
Isolation of (5R)-4-N-methyl-5-phenylmorpholin-2-one 3a
suggested that the ring expansion did take place. However,
spectroscopic evidence indicated the absence of the ethoxy-
carboxymethyl fragment and the presence of the methyl
group attached to the nitrogen atom, thus leading us to
conclude that the ring expansion did not happen through the
ring expansion-carbonylation sequence. Rather, it may have
proceeded through an intramolecular nucleophilic attack of
the oxygen atom of the oxazolidine at the carbon of the ester
functional group, resulting in the formation of iminium
intermediate 2 as illustrated in Scheme 2.
We now describe the synthesis of new N-(ethoxycar-
boxymethyl)oxazolidines and their reaction with the [Rh-
(COD)Cl]2/KI catalytic system under carbonylative condi-
tions. We further show that the reactivity of the N-(ethoxycar-
boxymethyl)oxazolidines differs significantly from their thio-
analog in that instead of ring expansion-contraction, it
undergoes intramolecular reductive ring expansion with the
formation of morpholin-2-ones. This approach constitutes a
new route to these particular types of heterocycles.5
Substituted N-(ethoxycarboxymethyl)oxazolidines were
synthesized starting from the corresponding enantiomerically
pure amino alcohols6 by alkylation with ethyl bromoacetate
and subsequent cyclization with paraformaldehyde,7 attaining
the desired oxazolidines as depicted in Scheme 1.
Scheme 2. Proposed Mechanism for the Intramolecular Ring
Expansion of 1a via the Intermediate Iminium Ion 2
Scheme 1. Synthesis of N-(Ethoxycarboxymethyl)oxazolidines
The next step in the reaction is most likely a two-electron
one-proton reduction (formal hydride addition) of 2. Al-
though it is not clear how this step occurred since there was
no source of hydride in the reaction media, we rationalized
that the conversion could be improved by the use of a mixture
of carbon monoxide and hydrogen instead of pure carbon
monoxide to satisfy the need for the rhodium hydride species
in the reaction. As illustrated in Table 1 (entry 2), the use of
CO/H2 mixture dramatically improves the conversion of 1a
into (5R)-4-N-methyl-5-phenylmorpholin-2-one 3a affording
the product in nearly quantitative yield. Additional evidence
for our mechanistic hypothesis was obtained by carrying the
reaction out under an atmosphere of deuterium gas and
carbon monoxide. The 13C NMR spectrum of the isolated
compound shows a signal for the carbon nucleus of the
N-methyl functional group as a triplet due to carbon-
deuterium coupling.8 The use of pure hydrogen gas (no CO),
however, results in a reduction of the rhodium catalyst and
deposition of rhodium metal in the form of “rhodium mirror”
(Table 1, entry 5). It is evident, therefore, that although
carbon monoxide does not participate in the reaction directly,
its presence is necessary to stabilize the catalyst.
An initial attempt employing 1a as a model substrate and
5 mol % of [Rh(COD)Cl]2 and 10 mol % of KI as the
catalyst, in toluene at 180 °C under 65 atm of carbon
monoxide, gave only 11% of the reaction product which
appeared to be (5R)-4-N-methyl-5-phenylmorpholin-2-one 3a
(Table 1, entry 1). In the absence of the catalyst, or using
Table 1. Catalytic Reductive Ring Expansion of 1aa
entry
catalystb
CO, atm H2, atm t, °C yield,c %
1
2
3
4
5
6
7
[Rh(COD)Cl]2/KI
[Rh(COD)Cl]2/KI
65
30
65
50
180
180
180
180
100
100
100
11
95
d
d
e
30
Further optimization showed that the reaction temperature
could be decreased from 180 to 100 °C (Table 1, entry 6),
Co2(CO)8/Ru3(CO)12
[Rh(COD)Cl]2/KI
[Rh(COD)Cl]2/KI
[Rh(COD)Cl]2/KI
30
30
5
(4) Khumtaveeporn, K.; Alper, H. J. Am. Chem. Soc. 1994, 116, 5662-
6.
30
5
97
97
(5) Alternative routes to (5R)-4-N-methyl-5-phenylmorpholin-2-one 3a
were described in following articles: (a) Alker, D.; Harwood, L. M.;
Williams, C. E. Tetrahedron 1997, 53, 12671-12678. (b) Agami, C.; Couty,
F.; Hamon, L.; Prince, B.; Puchot, C. Tetrahedron 1990, 46, 7003-7010.
(c) Agami, C.; Couty, F.; Daran, J. C.; Prince, B.; Puchot, C. Tetrahedron
Lett. 1990, 31, 2889-2892.
a All experiments were carried out in anhydrous toluene for 15 h. b Entries
1, 2, 5-7: [Rh(COD)Cl]2, 5 mol %, KI, 10 mol %; entry 4: Co2(CO)8, 30
mol %, Ru3(CO)12, 14 mol %. c Isolated yield after purification by column
chromatography. d Starting material was decomposed. e Rhodium metal
deposition was observed.
(6) McKennon, J. M.; Meyers A. I. J. Org. Chem. 1993, 58, 3568-
3571.
(7) Compounds 1a and N-(cyanomethyl)-4-phenyloxazolidine: Deprez,
P.; Royer, J.; Husson, H.-P. Tetrahedron 1993, 49, 3781-3792.
(8) The fact that only 63% of the compound has N-CH2D structural
fragment (calculated according to the integral values of the NMR spectrum,
(S11 (inset), Supporting Information) may be due to hydrogen-deuterium
exchange which occurs in rhodium hydride bonded to 1,5-cyclooctadiene.
Co2(CO)8/Ru3(CO)12 as the catalytic system, reaction resulted
in the formation of decomposition products (Table 1, entries
3 and 4).
1358
Org. Lett., Vol. 10, No. 7, 2008