tertiary alcohol products proceed with modest (ca. 3:1)
diastereoselectivities.9,10 In this communication, we describe
our preliminary studies on the scope and mechanism of new
tandem Wittig rearrangement/aldol reactions of O-alkyl
methyl glycolate derivatives.
[1,2]-Wittig rearrangement of boron ester enolate 4 to
generate 5 (Scheme 2). This hypothesis is supported by the
Scheme 2. Proposed Mechanism
Scheme 1
fact that treatment of methyl O-benzyl glycolate (1) with
Bu
2 3
BOTf/Et N for 15 min at room temperature followed by
12
aqueous workup affords rearranged product 7 in 81% yield.
Following the initial Wittig rearrangement, conversion of
to boron enolate 6 presumably occurs with high selectivity
In our preliminary experiments, we sought to determine
which of the two sequential reactions that lead to the
conversion of 1 to 3 occurs first, as this knowledge could
aid in the development of optimal conditions for this
transformation. The possibility that the sequence is initiated
by an initial aldol reaction of 1 was rapidly discounted. The
expected product of the boron-aldol reaction, â-hydroxy
5
for E(O)-enolate generation due to chelation between the
ester carbonyl and the adjacent boron alkoxide. Enolate 6
then undergoes aldol reaction to provide the observed syn-
diol product 3 with excellent stereoselectivity. Evidence for
the intermediacy of doubly borylated ester enolate 6 was
obtained through HRMS analysis of a reaction mixture
1
1
ester 2, was prepared as a mixture of diastereomers and
resulting from treatment of 1 with Bu
at room temperature. A signal was observed for m/z 428.3650
calculated mass ) 428.3633) with an isotopic distribution
2 3
BOTf/Et N for 15 min
treated with a mixture of Bu BOTf and Et N in CH Cl at
2
3
2
2
room temperature for 15 min. As shown in eq 1, these
conditions resulted predominantly in the decomposition of
the starting material and provided only trace amounts (<5%)
of 3.
(
in accord with the calculated pattern for 6.
With knowledge of the reaction mechanism in hand,
improved conditions were developed in which 1 was treated
with excess Bu
rangement before introduction of the aldehyde. In a repre-
sentative experiment, 1 was added to a solution of Et N (4
equiv) and Bu BOTf (3.2 equiv) in CH Cl at 0 °C and then
2 3
BOTf/Et N and allowed to undergo rear-
3
2
2
2
warmed to room temperature for 15 min. The mixture was
then cooled to 0 °C. Acrolein (1.5 equiv) was added, and
the reaction mixture was allowed to warm to room temper-
ature and stir for 1 h. Upon workup, the diol product 3 was
obtained in 67% yield with >20:1 diastereoselectivity (Table
1, entry 1).
The failure of 2 to undergo clean conversion to 3 suggests
that the transformation of 1 to 3 likely proceeds via an initial
(7) For examples of [1,2]-Wittig rearrangements of enolates, see: (a)
Curtin, D. Y.; Proops, W. R. J. Am. Chem. Soc. 1954, 76, 494. (b) Paquette,
L. A.; Zeng, Q. Tetrahedron Lett. 1999, 40, 3823. (c) Vilotijevic, I.; Yang,
J.; Hilmey, D.; Paquette, L. A. Synthesis 2003, 1872. (d) Garbi, A.; Allain,
L.; Chorki, F.; Ourevitch, M.; Crousse, B.; Bonnet-Delpon, D.; Nakai, T.;
Begue, J.-P. Org. Lett. 2001, 3, 2529.
To probe the scope of the tandem Wittig rearrangement/
aldol reaction, 1 was treated with a variety of different
aldehydes using the optimized conditions described above.
As shown in Table 1, the diol products were all obtained in
good yields with excellent diastereoselectivities.13 The
transformation is effective with aromatic aldehydes
(8) For examples of [2,3]-Wittig rearrangements of boron ester enolates,
see: (a) Oh, T.; Wrobel, Z.; Rubenstein, S. M. Tetrahedron Lett. 1991, 32,
4
4
647. (b) Fujimoto, K.; Matsuhashi, C.; Nakai, T. Heterocycles 1996, 42,
23.
(
9) Kumagai, N.; Matsunaga, S.; Kinoshita, T.; Harada, S.; Okada, S.;
Sakamoto, S.; Yamaguchi, K.; Shibasaki, M. J. Am. Chem. Soc. 2003, 125,
169.
10) Aldol reactions of O-protected 2-hydroxycarbonyl compounds that
2
(
afford tertiary alcohol products also generally proceed with modest
diastereoselectivity unless BHT esters or chiral auxiliaries are employed.
See: (a) Murata, Y.; Kamino, T.; Hosokawa, S.; Kobayashi, S. Tetrahedron
Lett. 2002, 43, 8121. (b) Kamino, T.; Murata, Y.; Kawai, N.; Hosokawa,
S.; Kobayashi, S. Tetrahedron Lett. 2001, 42, 5249. (c) Montgomery, S.
H.; Pirrung, M. C.; Heathcock, C. H. Carbohydr. Res. 1990, 202, 13. (d)
Heathcock, C. H.; Pirrung, M. C.; Young, S. D.; Hagen, J. P.; Jarvi, E. T.;
Badertscher, U.; M a¨ rki, H.-P.; Montgomery, S. H. J. Am. Chem. Soc. 1984,
(12) Efforts to directly effect the boron-mediated aldol reaction of
R-hydroxy ester 7 via treatment with excess Et3N/Bu2BOTf or 1 equiv of
KH followed by excess Et3N/Bu2BOTf were unsuccessful. Boron-mediated
aldol reactions of unprotected glycolate esters have also not been described
in the literature. This suggests that 7 is not an intermediate along the reaction
pathway and that formation of enolate 6 from borylated ester 5 occurs more
rapidly than protonolysis of 5 to afford 7.
1
(13) In most cases, only one stereoisomer was observed by H NMR
1
06, 8161.
11) Ester 2 was prepared through an aldol reaction between acrolein
and the lithium enolate of 1 at -78 °C.
analysis of crude reaction mixtures. Product stereochemistry was assigned
through 1H NMR nOe analysis of an acetonide derivative of 3. See the
Supporting Information for complete details.
(
4662
Org. Lett., Vol. 8, No. 20, 2006
Bertrand, Myra Beaudoin
