Conjugated elimination versus [1,2]-Wittig rearrangement of unsaturated diox(ol)anes

Article abstract of DOI:10.1016/S0040-4039(02)02568-6

A set of unsaturated dioxanes and dioxolanes has been prepared in three steps from acetoxy-1-isoprene. When reacted with two equivalents of t-BuLi in THF, these compounds provided, under various conditions, a mixture of two types of 1,3-dienes. The first one, derived from a conjugated elimination reaction resulting in the heterocycles opening, is relatively unstable but could be trapped as an acrylate. The second one resulted probably from a [1,2]-Wittig rearrangement. The competition between these two reactions has been observed for both dioxanes and dioxolanes (but not for acyclic ketals) and is the object of a strong temperature effect, the Wittig rearrangement being favored over the elimination at room temperature. A difference between kinetic and thermodynamic deprotonation sites is proposed to be at the origin of this competition on the base of both experimental and theoretical results.

Full text of DOI:10.1016/S0040-4039(02)02568-6

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 373–377
Conjugated elimination versus [1,2]-Wittig rearrangement of
unsaturated diox(ol)anes
a
a
b
a,
Lo ¨ı c Lemi e` gre, Thomas Regnier, Jean-Claude Combret and Jacques Maddaluno *
a
Laboratoire des Fonctions Azot e´ es & Oxyg e´ n e´ es Complexes, IRCOF, Universit e´ de Rouen, 76821 Mont St Aignan Cedex,
France
b
Laboratoire des Sciences et M e´ thodes S e´ paratives, IRCOF, Universit e´ de Rouen, 76821 Mont St Aignan Cedex, France
Received 4 October 2002; accepted 2 November 2002
Abstract—A set of unsaturated dioxanes and dioxolanes has been prepared in three steps from acetoxy-1-isoprene. When reacted
with two equivalents of t-BuLi in THF, these compounds provided, under various conditions, a mixture of two types of
1
,3-dienes. The first one, derived from a conjugated elimination reaction resulting in the heterocycles opening, is relatively unstable
but could be trapped as an acrylate. The second one resulted probably from a [1,2]-Wittig rearrangement. The competition
between these two reactions has been observed for both dioxanes and dioxolanes (but not for acyclic acetals) and is the object of
a strong temperature effect, the Wittig rearrangement being favored over the elimination at room temperature. A difference
between kinetic and thermodynamic deprotonation sites is proposed to be at the origin of this competition on the base of both
experimental and theoretical results. © 2002 Elsevier Science Ltd. All rights reserved.
Albeit the object of less attention than its [2,3] counter-
part, the [1,2]-Wittig rearrangement is a useful synthetic
been obtained lately thanks to enantioselective chiral-
base induced processes. Good advantage has also been
3
1
process that remains a reaction of experimental and
taken of acetals in this reaction, in particular for the
2
1a,4
theoretical interests. Because it is a convenient source
synthesis of C-glycosides,
and the influence of sub-
of carbonꢀcarbon bonds, its asymmetric version has
also been extensively investigated, and high e.e. have
strate structure and conditions on this particular case
has been subjected to in-depth studies.
5
Figure 1.
*
Corresponding author.
0
040-4039/03/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02568-6

3
74
L. Lemi e` gre et al. / Tetrahedron Letters 44 (2003) 373–377
The propensity of a,b-unsaturated acetals to undergo
Reacting acetals 8 in THF with 2 equiv. of t-BuLi led
to two types of 1,3-dienes. The first one (10) was
obviously derived from the conjugated elimination/ring
opening sequence that could be anticipated on the basis
of previous results, while the second one (11) was
bearing a p-methoxyphenyl substituent. This product
results more than likely from the dehydration of an
intermediate alcohol 9, itself derived from a [1,2]-Wittig
conjugated eliminations upon deprotonation is well
known and offers a handy access to functionalized
6
1
,3-dienes. However, these carbanions can also be
candidates to the [1,2]- or [2,3]-Wittig rearrangements
(
Fig. 1). A substrate such as 1 carries indeed two
4 6
possible deprotonation sites (C and C ). While the
carbanion at C is likely to undergo either a [1,2]- or
6
10,11
rearrangement
(Scheme 2). This unexpected reac-
[
2,3]-Wittig rearrangement, yielding 2a and 2b, respec-
4
tion could not be dodged even under basic work-up
conditions, rendering perfectly vain any attempt of
diastereocontrol, but leaving all its interest to the con-
trol of the competitive pathways. The alcohols 10 are
relatively sensitive and could not be efficiently isolated
by chromatography over silica gel. We found the direct
trapping of the intermediate alcoholates by trifl-
uoroethyl acrylate furnishes the corresponding esters 12
tively, that at C can provides the [1,2]-Wittig or the
d-elimination products 3 and 4, respectively. Defining
conditions favoring, at will, one of these pathways over
the others could be of obvious synthetic interest. We
also contemplated that introducing a remote C -sym-
2
metrical group could influence the diastereoselectivity
of the rearrangement through an inductor not directly
7
involved in the reaction. This would increase the syn-
12
in good yields (up to 70%). On a stereochemical point
thetic potential of these substrates since: (i) the pro-
tected aldehyde function is preserved after the
transposition; (ii) a broad and rapid tuning of the
selectivity can be easily undertaken thanks to the rela-
tively large set of chiral diols commercially available.
We present in this communication our preliminary
findings about the influence of a few parameters on the
chemioselectivity.
6
c,d
of view, the dienic part of 12 adopts the expected
1E,3Z) configuration.
(
The sensitivity of these competitive reactions to experi-
mental conditions has been evaluated and the results
are gathered in Table 1. The three first entries concern-
The unsaturated C -symmetrical dioxanes and diox-
2
olanes necessary to this study have been prepared from
isoprenylacetate 5 and NBS in THF/MeOH (Scheme
8,9
1), following the efficient protocol of Venturello et al.
The bromoacetal 6, obtained as a :75:25 E/Z mixture,
was then substituted by potassium p-methoxybenzyl
alcoholate and the methyl acetal 7 underwent transac-
etalization reacting it with the diols retained in methyl-
ene chloride in the presence of catalytic amounts of
6
a
camphorsulfonic acid (CSA). Six cyclic acetals 8a–f
were thus prepared in medium to reasonable overall
yields (49–87%), the configuration of their double bond
remaining unchanged during these steps. Substrates
8
a–d were designed to probe the sensitivity of the
reaction to dioxolane structural features, while achiral
8
e,f have been considered as easy-to-prepare bench-
marks for their dioxanic counterparts.
Scheme 2.
Table 1. Elimination versus Wittig in THF (Scheme 2)
Entry
SM
R
n
Conf.
T (°C)
9/10
1
2
3
4
5
8a
8a
8a
8b
8b
8b
8c
8d
8e
8e
8f
Me
Me
Me
Me
Me
Me
Ph
Ph
H
0
0
0
0
0
0
0
0
1
1
1
−
anti
anti
anti
syn
syn
syn
anti
syn
–
−78
−40
0
B5:95
33:67
67:33
13:87
67:33
0:100
0:100
0:100
0:100
49:51
58:42
B5:95
0
20
20
20
0
a
6
7
8
9
1
1
1
0
0
1
2
H
Me
–
–
syn
–
20
20
20
7
a
Scheme 1.
2 equiv. of TMEDA were added to the medium.

L. Lemi e` gre et al. / Tetrahedron Letters 44 (2003) 373–377
375
ing the anti dioxolane 8a indicate that the [1,2]-Wittig
rearrangement, leading to 9 then to diene 11, is dramat-
ically favored when increasing the temperature, even on
a relatively small 20°C range. However, the conditions
of entry 3 are associated with a chemical yield in 11
that remains, after flash chromatography, as low as
37%. Comparing entries 3 and 4 show that the syn
dioxolane 8b is more prone to an elimination than its
anti counterpart 8a. This can be related to the strain
undergone by the five-membered heterocycle bearing
two syn 4,5-substituents. Interestingly, an opposite
effect is described in literature for 4,6-disubstituted
dioxanes: anti isomers open more easily than their
Figure 2. Steric strains in syn and anti dioxolanes and diox-
anes.
1
3
triequatorial syn counterparts (Fig. 2).
The dioxolane derived from the meso butanediol isomer
has been employed in entries 4–6, in an attempt to
probe the sensitivity of these reactions to the exact
structure of the acetal. The same alteration of the
chimioselectivity in favor of the Wittig rearrangement
was observed with temperature, a 67:33 ratio being
obtained at room temperature, in an identically poor
yield (38%). Entry 6 shows that the addition of 2 equiv.
(
with respect to the substrate) of tetramethyl-
ethylenediamine (TMEDA) to the medium renders
the reaction elimination-selective. The yield in 12 then
reaches values in the 70% range. This result can be
related to a significant alteration of the solvation and/
1
4
or aggregation of the original allyllithium intermedi-
ate by the diamine, that would suppress an eventual
reversibility of this elimination by preventing the back-
addition of the alcoholate on the dienic system. Entries
7
and 8 further indicate the sensitivity of this system to
structural parameters since the acetals 8c,d, derived
from hydrobenzoin, yields selectively, between 0 and
Scheme 3.
2
0°C, the ring-opening products 10c,d. The case of
dioxanes has been evaluated in entries 9–11. With the
,3-propanediol derived dioxane 8e, the same tempera-
1
recovered quantitatively. The steric hindrance due to
the gem-dimethyl group is possibly responsible for the
inaccessibility of the allylic protons to the base.
ture effect is noted. It is worthy of note that at 20°C,
the 2,4-pentanediol derivative 8f provides more or less
the same ratio between the competitive reactions (com-
pare entries 5 and 11). However, the dioxanes seem
more fragile under these conditions, since the isolated
yields after chromatography on silica gel are extremely
low (:10%). Finally, the acyclic ketal 7 turned out to
be elimination-selective, giving an efficient access to the
corresponding dienol ether (entry 12), in line with previ-
We next tried to evaluate the relative stabilities of the
4
6
two C /C delocalized carbanions from a semiempirical
theoretical point of view. We are well aware that the
figures we can get from such calculations are quantita-
tively meaningless since they do not take into account
+
the cation (Li ) nor the solvent (THF), and are limited
6
,15
ous results.
to the AM1 level. In addition, these computations can
obviously provide only with thermodynamic data.
However, even qualitative indications on the general
behavior of these systems could help to improve the
discussion. The resulting optimized anions are dis-
played on Figure 3(A,B), respectively. The energy val-
Let us now consider the mechanistic implications of
these observations. Since we were not able to isolate
neither the original carbanion (quenching the reaction
medium with Bu SnCl) nor the alcoholate(s) 2b and/or
3
6
3
deriving from the two possible [1,2]-Wittig rearrange-
ues show that the C anion is more stable by :2.5
ments (with TMSCl), it was not easy to determine the
pathways followed by the system. Preventing the ben-
zylic deprotonation was thus attempted resorting to
cumbersome acetal 16 (Scheme 3). This latter was pre-
pared in 51% yield condensing the potassium alcoholate
of 14 (obtained from p-methoxybenzophenone 13 and
an excess methyllithium in 35% yield) on allylic bro-
mide 15 (itself obtained in 91% from 6). However, 16 is
unreactive toward t-BuLi and the starting material was
kcal/mol, suggesting a higher thermodynamical acidity
of this position.
This energy difference, together with the data in Table
1, suggest that the competition between d-elimination
and [1,2]-Wittig rearrangement could stem from the
4
6
difference between C /C deprotonation sites. The low
temperature conditions (−78°C, entry 1) would favor
4
the kinetic C carbanion that would mainly evolve

3
76
L. Lemi e` gre et al. / Tetrahedron Letters 44 (2003) 373–377
References
1. Recent examples: (a) Tomooka, K.; Yamamoto, H.;
Nakai, T. Angew. Chem., Int. Ed. Engl. 2000, 39, 4500;
(
b) Tomooka, K.; Kikuchi, M.; Igawa, K.; Suzuki, M.;
Keong, P.-H.; Nakai, T. Angew. Chem., Int. Ed. Engl.
000, 39, 4502; (c) Katritzky, A. R.; Fang, Y. Heterocy-
2
cles 2000, 53, 1783; (d) Kitagawa, O.; Momose, S.-I.;
Yamada, Y.; Shiro, M.; Taguchi, T. Tetrahedron Lett.
2001, 42, 4865; (e) Barluenga, J.; Fananas, F. J.; Sanz,
R.; Marcos, C.; Trabada, M. Org. Lett. 2002, 4, 1587.
. Inter alia: (a) Antoniotti, P.; Tonacchini, G. J. Org.
Chem. 1998, 63, 9756; (b) Jursic, B. S. Int. J. Quantum
Chem. 1999, 73, 299; (c) Sheldon, J. C.; Taylor, M. S.;
Bowie, J. H.; Dua, S.; Chia, C. S. B.; Eichinger, P. C. H.
J. Chem. Soc., Perkin Trans. 2 1999, 333.
2
3
4
5
6
. Recent contributions in the field: (a) Maleczka, R. E.;
Geng, F. J. Am. Chem. Soc. 1998, 120, 8551; (b)
Tomooka, K.; Yamamoto, K.; Nakai, T. Angew. Chem.,
Int. Ed. Engl. 1999, 38, 3741.
. (a) Tomooka, K.; Yamamoto, H.; Nakai, T. J. Am.
Chem. Soc. 1996, 118, 3317; (b) Tomooka, K.; Kikuchi,
M.; Igawa, K.; Keong, P.-H.; Nakai, T. Tetrahedron Lett.
4
Figure 3. (A) AM1 optimized C anion (E=2.533 kcal/mol);
(
6
B) AM1 optimized C anion (E=0.000 kcal/mol).
1999, 40, 1917.
. (a) G a¨ rtner, P.; Letschnig, M. F.; Knollm u¨ ller, M.; V o¨ l-
lenkle, H. Tetrahedron: Asymmetry 1999, 10, 4811; (b)
G a¨ rtner, P.; Letschnig, M. F.; Knollm u¨ ller, M.; Mereiter,
K. Tetrahedron: Asymmetry 2000, 11, 1003.
. (a) Gaonac’h, O.; Maddaluno, J.; Chauvin, J.; Duhamel,
L. J. Org. Chem. 1991, 56, 4045; (b) Maddaluno, J.;
Gaonac’h, O.; Le Gallic, Y.; Duhamel, L. Tetrahedron
Lett. 1995, 36, 8591; (c) Martin, C.; Maddaluno, J.;
Duhamel, L. Tetrahedron Lett. 1996, 37, 8169; (d)
Deagostino, A. M.; Maddaluno, J.; Mella, M.; Prandi,
C.; Venturello, P. J. Chem. Soc., Perkin Trans. 1 1998,
along the eliminative pathway, while higher tempera-
tures would lead preferentially to a C carbanion that
could only undergo the Wittig steps. In such a case an
intramolecular?) C /C carbanion interchange would
6
4
6
(
be required that remains to be demonstrated.
In conclusion, the above results indicate that the reac-
tion of a,b-unsaturated diox(ol)anes with t-BuLi trig-
gers a competition between the conjugated elimination
reaction and the [1,2]-Wittig rearrangement. Depending
mainly on the temperature, one or the other of these
two pathways becomes exclusive or preponderant. We
have been unable yet to establish, on experimental
grounds, if these two reactions stem from a same
lithiated intermediate or not. Conclusions are particu-
larly difficult to draw from the experiments in Table 1
since both the regioselectivity of the deprotonation and
the relative stabilities of the various diox(ol)anes con-
sidered certainly influence the kinetic of the two reac-
tions. However, the temperature effect and the results
of semiempirical calculations suggest that the allylic
deprotonation could take place under kinetic control
while the benzylic proton abstraction would rather
happen under thermodynamic conditions. Experiments
to confirm or invalidate these hypotheses are currently
underway.
881; (e) Guillam, A.; Toupet, L.; Maddaluno, J. J. Org.
Chem. 1998, 63, 5110; (f) Guillam, A.; Toupet, L.; Mad-
daluno, J. J. Org. Chem. 1999, 64, 9348.
7. A comparable concept relying on protected diols has
been put forward recently by Maleczka and Geng (Ref.
3a).
8. (a) Deagostino, A.; Balma Tivola, P.; Prandi, C.; Ven-
turello, P. Synthesis 1999, 1841. Experimental details are
given in: (b) Bataille, C.; B e´ gin, G.; Guillam, A.;
Lemi e` gre, L.; Lys, C.; Maddaluno, J.; Toupet, L. J. Org.
Chem. 2002, 67, 8054.
9. For another access to these compounds see Ref 6a.
1
0. A [1,2]-Wittig/b-elimination competition, as well as a
temperature effect is reported by G a¨ rtner et al. (Ref. 5b).
1. No [2,3]-Wittig rearrangement was observed on acetals 8
by contrast with similar unsaturated aminals that yield
significant amounts of the corresponding products: Blan-
chard, N.; Mangeney, P., unpublished results.
1
1
2. Details on the conditions providing ester 11, a useful
triene for intramolecular Diels–Alder reactions, are
described in Lemi e` gre, L. Ph.D. Thesis, Universit e´ de
Rouen, 2002 and will be published later on.
Acknowledgements
1
3. (a) Caballero, M.; Garcia-Valverde, M.; Pedrosa, R.;
Vicente, M. Tetrahedron: Asymmetry 1996, 7, 219; (b)
Ballini, R.; Bosica, G.; Cossu, S.; De Lucchi, O.; Peluso,
L.L. is grateful to the MRT for a Ph.D. fellowship. We
thank Dr. Jean-Erik Ancel (Rh oˆ ne-Poulenc Industriali-
sation) for generous samples of acetoxyisoprene.

L. Lemi e` gre et al. / Tetrahedron Letters 44 (2003) 373–377
377
P. Tetrahedron 2001, 57, 4461.
4. Collum, D. B. Acc. Chem. Res. 1992, 25, 448.
5. (a) Naito, K.; Rickborn, B. J. Org. Chem. 1980, 45, 4061;
hedron Lett. 1984, 25, 519; (d) Fujiwara, S.; Katsumura,
S.; Isoe, S. Tetrahedron Lett. 1990, 31, 691; (e) Ven-
turello, P. J. Chem. Soc., Chem. Commun. 1992, 1032; (f)
Deagostino, A.; Prandi, C.; Venturello, P. Tetrahedron
1996, 52, 1433.
1
1
(b) Moss, R. J.; Rickborn, B. J. Org. Chem. 1984, 49,
3
694; (c) Mioskowski, C.; Manna, S.; Falck, J. R. Tetra-