952
(Re) face
A. Krief et al. / Tetrahedron Letters 46 (2005) 951–953
of HMPA prior to the aqueous quench ((i) THF,
À78 °C, 1 h (ii) HMPA, À78 °C, 3 h (iii) aq NH4Cl).
CO2Me
a) R3SnM 4, Solvent, -78°C
O
O
Me
Me
b) aq. NH4Cl, -78°C
(Si) face
It is even more interesting to compare the behaviour of
the trialkylsilyl metals to that of trialkylstannylmetals3
which belong to the same family of compounds.
1
R3Sn
CO2Me
R3Sn
O
CO2Me
O
Me
Me
Me
Both silyl-4a,b and stannyl-5a,b lithiums react in THF
with extremely high stereocontrol on methyl Z-3-(2,2-di-
methyl-[1,3]dioxolan-4-yl)-acrylate 1 but from different
faces (Re and Si, respectively; Scheme 1, entry c and
Scheme 2, entries a,b). The presence of HMPA in the
reaction medium reverses the face of attack in the for-
mer case (Scheme 1, entries b,c) whereas as we have dis-
covered in the present study it has no effect on the
reaction involving tributylstannyllithium (Scheme 1,
compare entries b,c).
+
Me
O
O
5Re
5Si
Yield
5 (%) anti/syn
5Si/5Re
Entry
1
R3SnM
4
a
b
c
d
e
f
Z
Z
Z
Z
E
E
Me3SnLi
74
58
67
84
73
50
100/00
100/00
100/00
0/100
(n-Bu)3SnLi
(n-Bu)3SnLi,HMPA
(n-Bu)3Sn(Et2)ZnLi
(n-Bu)3Sn(Et2)ZnLi
(n-Bu)3SnLi
Silyl-4c and stannyl-3 zincates however both add to the
same Re-face of methyl Z-3-(2,2-dimethyl-[1,3]dioxo-
lan-4-yl)-acrylate 1 (Scheme 1, entry d; Scheme 2, entry
d) and this trend is not reversed with its E-stereoisomer
although the stereofacial differentiation is poorer
(Scheme 1, entry h; Scheme 2, entry e).
10/90
74/26
Scheme 2.
the previous case, it generates the syn-adduct 3bRe in
good yield from the Z-enoate 1, and with extremely high
de (Scheme 1, entry c; compare to entry a).
The stereochemistry of each of the two diastereoisomers
of 3, has been ascertained by transforming them to the
corresponding
4-(dimethylphenylsilanyl)-5-hydroxy-
We performed, in order to confirm the important role of
HMPA in this facial stereocontrol, the reaction of 2b
with the Z-enoate Z-1 in THF-HMPA (7 equiv., THF,
À78 °C, 4 h). We found, as expected that it produces
the anti-adduct 3bSi resulting from the stereoselective at-
tack on its Si-face (Scheme 1, entry b). The stereoselec-
tivity proved to be much better than that observed from
2a under the same experimental conditions (Scheme 1,
compare entries a,b).
methyldihydrofuran-2-ones 6anti and 6syn on tandem
acid catalyzed dioxolane ring opening c-hydroxyester
ring closure (CF3CO2H, THF, 20 °C, 3 h, Scheme 3).6
The 1H NMR spectra of 6anti compares successfully with
that already described in the literature for the same com-
pound prepared from the silylcuprate and 5-hydroxy-
methyl-5H-furan-2-one.6 That of its stereoisomer 6syn
is as expected (Ha: d = 2.22 ppm, JHAÀHB = 8.4 Hz;
Hb d = 4.674 ppm, JHAÀHB = 8.4 Hz).
Diethylzincate4c 2c behaves as its lithium analogue 2b
when the reactions are performed in THF. It reacts by
the Re-face of the Z-enoate Z-1 and offers the advantage
of providing the syn-adduct 3bRe with even higher ste-
reocontrol (de 96 % instead of 90 %) and higher yield
(Scheme 1, entry d, compare to entry c).
Our results compare favourably with those already
described by Smadja and co-workers, involving the
tris(trimethylsilyl)silyl radical,7 since they allow the stereo-
selective synthesis of each of the two stereoisomers
of 3.
We were rather surprised to find that silyl metals react
with very poor stereocontrol with the E-enoate E-1,
whatever are the conditions used (de 8–28%; Scheme
1, entries e–h). This contrasts with the results reported
above for the Z-diastereoisomer Z-1 (de 62–96%;
Scheme 1, compare entries e–h to a–d).
At present, we are unable to rationalize the whole set of
results involving addition and cycloaddition reactions
on c-alkoxy-a,b-unsaturated esters bearing a chiral
c-carbon on the basis of any known model (degenerated
H
a
PhMe2Si
PhMe2Si
O
CO2Me
This whole set of these results is quite exceptional,
especially, the high control of the face of attack of sily-
llithium 2b towards methyl Z-3-(2,2-dimethyl-[1,3]-
dioxolan-4-yl)-acrylate 1 in THF and the remarkably
completely reversed facial stereocontrol found if the
reaction is carried out in the presence of HMPA. We
have ascertained, at least in the latter case that the reac-
tion occurs under kinetic control. This has been
achieved by determining the structures of the mixtures
of products generated as described in Scheme 1, entry
c ((i) THF, À78 °C, 4 h (ii) aq NH4Cl) or after addition
Hb
CF3CO2H, aq. THF,
20°C, 3h
Me
O
O
O
Me
O
OH
96 %
3Si
6anti
H
a
PhMe2Si
CO2Me
PhMe2Si
O
Hb
CF3CO2H, aq. THF,
20°C, 3h
Me
O
Me
O
OH
93 %
3Re
6Syn
Scheme 3.