Asymmetric synthesis of cyclopropane-1,1-dicarboxylates from a -γ- Alkoxy-alkylidenemalonate

Article abstract of DOI:10.1055/s-1999-3173

2-Metallo-2-sulfonylpropanes and 2-metallo-2-nitropropanes react with the dimethyl alkylidenemalonate derived from the acetonide of D- glyceraldehyde to produce the Michael adduct or the corresponding dimethyl cyclopropane-1,1-dicarboxylate. The nature of

Full text of DOI:10.1055/s-1999-3173

1936
LETTER
Asymmetric Synthesis of Cyclopropane-1,1-Dicarboxylates from a g-Alkoxy-
Alkylidenemalonate
Alain Krief,* Laurent Provins, Alexandre Froidbise
Laboratoire de Chimie Organique de Synthèse, Département de Chimie, Facultés Universitaires Notre-Dame de la Paix, 61 rue de
Bruxelles, B-5000 NAMUR, Belgium
Fax+32 81 724536; E-mail: alain.krief@fundp.ac.be
Received 22 September 1999
Extension of this reaction to 2-metallo-2-nitropropanes4
Abstract: 2-Metallo-2-sulfonylpropanes and 2-metallo-2-nitropro-
5N and 2-metallo-2-sulfonylpropanes 5S should provide
interesting informations on the intimate mechanism of
these specific reactions as well as on that of organometal-
panes react with the dimethyl alkylidenemalonate derived from the
acetonide of D-glyceraldehyde to produce the Michael adduct or the
corresponding dimethyl cyclopropane-1,1-dicarboxylate. The na-
ture of the product formed and the relative configuration of the cy- lics with g-heterosubstituted-a,b-unsaturated esters which
still remains unclear.3,5
clopropane derivative depends upon the nature of the leaving group
and the conditions used (solvent and counter ion).
We were rather surprised to find that 2-lithio-2-sulfonyl-
Key words: asymmetric synthesis, Michael adduct, cyclopropane
propane 5SLi and 2-lithio-2-nitropropane 5NLi behave dif-
derivative
ferently with the dimethyl alkylidenemalonate D-2. Thus,
2-lithio-2-sulfonylpropane 5SLi reacted in THF almost ex-
clusively on the (Si)-face and lead to the g-sulfonyl
malonate6a 7SSi (THF, 20 °C, 0.5 h) whereas 2-lithio-2-ni-
tropropane 5NLi reacted, under similar conditions, exclu-
sively on the (Re)-face, providing after hydrolysis the g-
nitro malonate6a 7NRe (Scheme 2).
2-Metallo-nitroalkanes and 2-metallo-alkylsulfones add
to a,b-unsaturated esters and alkylidenemalonates in a
Michael mode, under a large variety of conditions and
produce g-nitro and g-sulfonyl esters.1,2 The adducts de-
rived from alkylidenemalonates further react under forced
conditions, in polar solvents, and lead to cyclopropane-
1,1-dicarboxylates in reasonably good yields.2 The latter
transformation offers, over the one involving phosphorus
and sulfur ylides and the same electrophiles, the advan-
tage to transfer the same isopropylidene moiety from re-
agents of lower molecular weight.
We recently reported3 that isopropylidene diphenylsul-
furane 3S and isopropylidene triphenylphosphorane 3P re-
act on the (Re)-face of the dimethyl alkylidenemalonate
D-2 derived from D-glyceraldehyde acetonide D-1 and
produce the corresponding dimethyl cyclopropane-1,1-di-
carboxylate 4Re almost as a single diastereoisomer
(Scheme 1).
Scheme 2
The same starting materials afforded, as expected,
dimethyl cyclopropane-1,1-dicarboxylates 4 when the re-
actions were instead carried out in DMSO at 80 °C
(Scheme 3).6b Both reactions provided the same stereoiso-
mer resulting from the cyclopropanation of dimethyl alky-
lidenemalonate D-2 by its (Si)-face but their
diastereoselectivity are very different: 4Si was produced
with very high stereocontrol from 2-lithio-2-sulfonylpro-
pane 5SLi (DMSO, 80 °C, 66h)6b,7 and as a mixture of ste-
reoisomers from 2-lithio-2-nitropropane 5NLi (DMSO,
80 °C, 24h).3,6b,7 Thus whereas the open chain derivative
and the cyclopropane arose from the attack of the same
face of D-2 from 5SLi they resulted from a different face of
attack from 5NLi
.
Scheme 1
In order to understand these conflicting results, we under-
took a more detailed study aimed to determine which of
the reactions takes place under kinetic or thermodynamic
Synlett 1999, No. 12, 1936–1938 ISSN 0936-5214 © Thieme Stuttgart · New York

LETTER
Asymmetric Synthesis of Cyclopropane-1,1-Dicarboxylates from a g-Alkoxy-Alkylidenemalonate
1937
Scheme 3
control. We found that addition of 2-lithio-2-sulfonylpro-
pane 5SLi to D-2 is slightly dependent upon the conditions at room temperature. Under these conditions the dia-
used. Complete diastereoselection was observed when the stereoisomeric ratio changed with the reaction time to
reaction was performed in THF-HMPA (D-2/HMPA:1/4; give 7NRe slowly from 2-lithio-2-nitropropane (Table 2,
-78 °C, 0.5 h, yield in 7S:76%, 7SSi/7SRe:100/0)8 instead of entry 1) and more rapidly when the reaction was instead
THF alone and slightly lower selectivity was found when performed with its potassium analogue (Table 2, entry 3).
the reaction was performed in ether (-78 °C, 0.5 h, 72%, This suggests that this reaction is under thermodynamic
7
SSi/7SRe :92/8).
control.
We have proven that these reactions occur under kinetic
control since the same 7SSi/7SRe ratio of stereoisomers is
obtained whether the reaction is quenched at -78 °C or at
room temperature, after short or long reaction time. We
have also secured that no epimerization is taking place
when each stereoisomer 7SRe or 7SSi is treated with a base
at room temperature (K2CO3 in DMSO, 20 °C, 10 h) or
even at 80 °C. Under the latter conditions however, ring
closure was readily achieved and lead with complete ste-
reocontrol (Scheme 4) to the dimethyl cyclopropane-1,1-
dicarboxylates 4Si from 7SSi (K2CO3, DMSO, 80 °C, 3 h,
84%, 100% de) and to its stereoisomer 4Re from 7SRe
(K2CO3, DMSO, 80 °C, 0.7 h, 77%, 100% de)
We have independently proven the presence of an equilib-
rium by reacting a 45/55 mixture of 7NRe/7NSi artificially
enriched in 7NSi, with K2CO3 in DMSO 20 °C and found
(Scheme 5) that it provides, after 18 h, almost exclusively
the 7NRe adduct (98% de).
Scheme 4
We have also found that the rate of this reaction depends
not only upon the nature of the counter-ion (it is much
faster with cesium than with lithium, Scheme 4, Table 1)
but also upon the nature of the stereoisomer used (7SRe re-
acts faster than 7SSi, Scheme 4, Table 1).9
The reaction of 2-metallo-2-nitropropane 5N with alky-
lidenemalonate D-2 is more complex. The 7NSi/7NRe ratio
of stereoisomers is highly dependent upon the solvent and
the conditions used (temperature and time). Although a
single stereoisomer (7NRe) was formed in THF10 (Scheme
Scheme 5
The cyclization of 7NRe to 4, carried out in DMSO under
different conditions implying alkali metal carbonates at
80 °C and potassium carbonate at 40 °C, lead every time
2), substantial amounts of the other stereoisomer (7NSi
)
was produced when the reaction was performed in DMSO
Synlett 1999, No. 12, 1936–1938 ISSN 0936-5214 © Thieme Stuttgart · New York

1938
A. Krief et al.
LETTER
to a mixture of the two diastereoisomers 4Re/4Si (45/55 to
35/65, Scheme 6, Table 3). Careful monitoring by H
NMR of the reaction described in Table 3, entry 5 shows
that D-2 is formed as an intermediate.
Acknowledgement
1
The authors thank the FRIA (Fonds pour la Formation à la Recher-
che dans l’Industrie et dans l’Agriculture, Bruxelles, Belgium) for
supporting this work (fellowships to A.F. and L.P.)
References and Notes
(1) (a) Yechezkel, T.; Ghera, E.; Ramesh, N. G.; Hassner, A.
Tetrahedron: Asymmetry 1996, 7, 2423. (b) Ferreira, A. R. G.;
Dias, A. G.; Pinto, A. C.; Costa, P. R. R.; Miguez, E.; da Silva,
A. J. R. Tetrahedron Lett. 1998, 39, 5305. (c) Costa, J. S.;
Dias, A. G.; Anholeto, A. L.; Monteiro, M. D.; Patrocinio, V.
L.; Costa, P. R. R. J. Org. Chem. 1997, 62, 4002.
(2) Krief, A.; Hevesi, L.; Chaboteaux, G.; Mathy, P.; Sevrin, M.;
DeVos, M.-J. J. Chem. Soc., Chem. Commun. 1985, 1693.
(b) Chaboteaux, G.; Krief, A. Bull. Soc. Chim. Belg. 1985, 94,
495. (c) Krief, A.; DeVos, M.-J. Tetrahedron Lett. 1985, 26,
6115. (d) Ono, N.; Kamimura, A.; Kaji, A. Synthesis 1984,
226.
Scheme 6
(3) Krief, A.; Provins, L.; Froidbise, A. Tetrahedron Lett. 1998,
39, 1437 and references cited.
(4) 2-metallo-2-nitropropanes are better described as nitronates
anions. Description as 2-metallo-2-nitropropanes in the text is
only used for consistency and does not reflect the reality.
(5) (a) Krief, A. in "Organic Synthesis. A Chemistry for the 21st
Century Monograph", International Union of Pure and
Applied Chemistry, Trost, B. M., Ed; Blackwell Scientific
Publications 1994, 337. (b) Krief, A.; Dumont, W.; Pasau, P.;
Lecomte, P. Tetrahedron 1989, 45, 3039. (c) Mulzer, J.;
Kappert, M. Angew. Chem. Int. Ed. Engl. 1983, 22, 63.
(6) [5] (a) The two stereoisomers of 7S have been separated (Thick
layer preparative liquid chromatography on SiO2, pentane/
ether 1/1, 7SSi Rf:0.4, 7SRe Rf :0.55) and the stereochemistry of
the major stereoisomer (7NRe and 7SSi) of each of the above
mentioned reactions has been unambiguously assessed by X-
ray crystallography.7 (b) The stereochemistry of the dimethyl
cyclopropane-1,1-dicarboxylates 4Re and 4Si have been
assessed by comparison with an authentic sample3. (c) The
stereochemistry of the dimethyl cyclopropane-1,1-
In conclusion, we have shown that 2-lithio-2-sulfonylpro-
pane reacts under kinetic control with alkylidenemalonate
D-2 in DMSO to provide dimethyl cyclopropane-1,1-di-
carboxylate 4Si almost as a single diastereoisomer. This is
diastereoisomeric to the one (4Re) formed when isopropy-
lidene diphenylsulfurane-LiBF4 3S or isopropylidene
triphenylphosphorane-LiI 3P are used instead.3
dicarboxylates 4Si has been also assessed by X-ray
crystallography.7
(7) (a) Baudoux, G.; Norberg, B.; Provins, L.; Froidbise,
A.; Krief, A.; Evrard G. Acta Cryst. 1998, C54, 834.
(8) 4Si is produced (65%) as a single stereoisomer by performing
the reaction in one pot by (i) 1,4-addition in THF-HMPA at
-78 °C for 2 h, (ii) heating the mixture to 20 °C, (iii) adding
DMSO and heating the mixture at 80 °C for 60 h. Notice that
7SSi does not cyclize when heated at 80 °C in THF-HMPA.
(9) These reactions have been followed by GC2 using 1,3,5-
trimethoxybenzene as an internal standard (Hewlett-Packard
5890A, capillary column, SE 30 HL; 30m x 0.2 mm, film
thickness 0.33 mm, injector and detector :250 °C, oven from
100 °C to 220 °C at 10 °C/min., Rt = 3.5 min.). Under similar
conditions the cyclopropane derivatives have the following
retention times: 4Re Rt = 5.1 min. 4Si Rt = 5.3 min.
Dimethyl cyclopropane-1,1-dicarboxylate 4Si was also
formed from 2-lithio-2-nitropropane but with much lower
diastereoselection. This result is even more puzzling since
the open chain 1,4-adduct 7NRe which is isolated if the re-
action is performed under milder conditions, has the op-
posite stereochemistry (it is derived from the (Re)-face
attack) and is formed under thermodynamic control.
Apparently, the formation of the 4Si adduct from 7NRe re-
quires the formation of 7NSi as an intermediate which is
the least stable of both diastereoisomers and which must
cyclize to 4Si faster than 7NRe do to produce4Re. We would
have expected that 7N behaves as 7S : this is apparently not
the case. Results are opposite to the one observed with the
sulfone derivatives! What are the reasons for such dis-
crepancies? We are working towards this end in trying to
understand the intimate mechanism of each of these reac-
tions.
(10) A 97/3 mixture of 7NRe/7NSi is formed in 70% yield when the
reaction is performed in THF-HMPA.
Article Identifier:
1437-2096,E;1999,0,12,1936,1938,ftx,en;G18299ST.pdf
Synlett 1999, No. 12, 1936–1938 ISSN 0936-5214 © Thieme Stuttgart · New York