obtained at room temperature could only be moderately
increased to 4.1:1 at low temperature, with no further
improvement on adding chelating InCl3 (Table 1, entries
1-3). While we were concerned with the stability of both
the starting hydrazone and the final adduct under the
H-abstracting photochemical conditions, we observed neither
dimerization nor change of the diastereomeric ratio on
prolonged irradiation under the reaction conditions. It thus
appears that the solvent was adequately protecting these
systems from hydrogen abstraction and subsequent side
reactions, making meaningful the use of the same photo-
chemical radical generation procedure for further studies on
diastereoselectivity. For these, we selected hydrazone 4a,
which incorporates not an electron-withdrawing activating
substituent but rather an ethyl group. Although, as expected,
its reactivity against the 1,3-dioxolanyl radical diminished
as compared to 1, the addition still proceeded quickly at room
temperature (0.5 h) and at a convenient rate even at low
temperatures (4.5 h at -78 °C) (Table 1, entries 4 and 5).
Regarding stereoselectivity, it was only slightly higher than
that obtained for 1, both at room temperature and at -78
°C, but now, for 4a, the addition of InCl3 (2 equiv) had a
remarkable effect and the diastereomeric ratio reached a value
of 10.1:1 in a considerably reduced reaction time (Table 1,
compare entries 5 and 6).
derivative 8 allowed for additional stereomeric enrichment
by crystallization and the subsequent preparation of 10 with
properties identical to those of an authentic sample prepared
from pure (R)-2-aminobutyric acid.10,16 It thus follows that
the stereochemical outcome of the reaction results from the
overwhelming preferential addition of the dioxolanyl radical
on the In-chelated N-acyl hydrazone from the opposite face
to the benzyl group.8b,17 Noteworthy, the same face is the
most exposed one for the nonchelated hydrazone, a factor
that undoubtedly contributes to the high net stereochemical
induction observed.
We finally made a preliminary survey on the scope and
limitations of the method applying the one-pot protocol to
relevant aldehydes 6b-e. Aldehydes 6b and 6c were selected
to study the influence of R-branching, either with alkyl or
with oxygenated substituents, and also because of the
significance of the corresponding amino acids derived from
them. The process worked well for 6b, which afforded adduct
5b with good diastereoselection (dr ) 97:3), thus opening a
new pathway for the asymmetric synthesis of cyclopropyl-
substituted amino acids.18 For the special case of 6c, carrying
a dimethyl acetal function, the overall yield and the stereo-
selectivity were lower than those for 6b but still sufficient
(75%, 95:5) to be of practical synthetic value; thus, good
prospects exist for the application of this methodology to
the preparation of R-acetal R-amino acids.19,20
The process could be done in one pot starting from
propionaldehyde (6a).12 Its treatment with (S)-3-amino-4-
benzyl-1,3-oxazolan-2-one (7S)13 in 1,3-dioxolane in the
presence of catalytic amounts of p-TsOH for 1 h, followed
by UV irradiation (1-1.5 h) in the presence of Ph2CO and
InCl3 at -78 °C, led to 5a (241-681 mg scale) with
improved yields (93-99% overall) and selectivities (dr )
99:1-98:2, Scheme 1). To demonstrate the usefulness of this
As for the aromatic aldehydes 6d and 6e, the reaction was
quick and efficient to render the corresponding adducts in
good yields.21 Chiral induction at the R center was, however,
either nonexistent for 4-methoxy-benzaldehyde or of the
(12) For the nonasymmetric one-pot intermolecular radical addition to
aldimines derived from aromatic aldehydes, see: Yamada, K.; Yamamoto,
Y.; Tomioka, K. Org. Lett. 2003, 5, 1797-1799.
(13) S)-3-Amino-4-benzyl-1,3-oxazolan-2-one (7S) was prepared from
commercial (S)-4-benzyl-1,3-oxazolan-2-one (Supporting Information). For
a recent work on the comparison of electrophilic amination reagents for
N-amination of 2-oxazolidinones, see: Shen, Y.; Friestad, G. K. J. Org.
Chem. 2002, 67, 6236-6239.
Scheme 1
(14) Numerous methods have been reported for the reductive cleavage
of N-N bonds; see: Enders, D.; Lochtman, R.; Meiers, M.; Mu¨ller, S.;
Lazny, R. Synlett 1998, 1182-1184 and references therein.
(15) Deslongchamps and Moreau first reported that aldehyde acetals
reacted with ozone to give esters in high yield: Deslongchamps, P.; Moreau,
C. Can. J. Chem. 1971, 49, 2465-2467. For additional methods, see: Curini,
M.; Epifano, F.; Marcotullio, M. C.; Rosati, O. Synlett 1999, 777-779 and
references therein.
(16) Commercially available (R)-2-aminobutanoic acid was converted
to 10 according to: Nosho, Y.; Seki, T.; Kondo, M.; Ohfuji, T.; Tamura,
M.; Okai, H. J. Agric. Food Chem. 1990, 38, 1368-1373.
(17) Access to the enantiomer of 10 would be possible starting with 7R,
accessible in turn from (R)-4-benzyl-1,3-oxazolan-2-one (see ref 13), which
is also commercially available at approximately the same price as its
enantiomer.
(18) First stereocontrolled synthesis of (S)-cleonin and related cyclopro-
pyl-substituted amino acids has been recently published: Esposito, A.; Piras,
P. P.; Ramazzotti, D.; Taddei, M. Org. Lett. 2001, 3, 3273-3275.
(19) For the first asymmetric synthesis of (R)- and (S)-2-amino-3,3-
dimethoxypropanoic acid (R-formylgycine dimethylacetal), including refer-
ences for its previous use in synthesis, see: DeMong, D. E.; Williams, R.
M. Tetrahedron Lett. 2002, 43, 2355-2357.
(20) Although still not tested, controlled oxidation of 5c should only
affect the dioxolane ring, as acyclic dialkoxy acetals react much more slowly
than cyclic ones due to stereoelectronic control: (a) Deslongchamps, P.;
Atlani, P.; Frehel, D.; Malaval, A.; Moreau, C. Can. J. Chem. 1974, 52,
3651-3664. See also: (b) Li, S.; Deslongchamps, P. Tetrahedron Lett. 1993,
34, 7759-7762, (c) Sueda, T.; Fukuda, S.; Ochiai, M. Org. Lett. 2001, 3,
2387-2390. (d) Plesnicˇar, B.; Cerkovnik, J.; Tuttle, T.; Kraka, E.; Cremer,
D. J. Am. Chem. Soc. 2002, 124, 11260-11261 and references therein.
See also ref 15.
addition as the key step for the asymmetric synthesis of
R-amino acids, we transformed 5a into 10 as indicated in
Scheme 1.14,15 Initial conversion of 5a into its benzamide
Org. Lett., Vol. 5, No. 14, 2003
2463