Stereoselective alkylation of tartrate derivatives. A concise route to (+)-O-methylpiscidic acid and natural analogues

Article abstract of DOI:10.1039/b606362b

The process of stereoselective alkylation of tartrate derivatives, was investigated. Tartaric acid is an abundant versatile chiral small building block for the asymmetric synthesis of natural products. The problem associated with the reduction of menopaus

Full text of DOI:10.1039/b606362b

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
Stereoselective alkylation of tartrate derivatives. A concise route to
(+)-O-methylpiscidic acid and natural analogues†
Anthony J. Burke,a Christopher D. Maycock*a,b and M. Rita Venturaa
Received 4th May 2006, Accepted 5th May 2006
First published as an Advance Article on the web 24th May 2006
DOI: 10.1039/b606362b
The lithium enolate of (2S,3S,5S,6S)-dimethoxy-2,3-di-
methyl-1,4-dioxane-5,6-dithiocarboxylate 8 undergoes stere-
oselective mono- and/or dialkylations to afford two new stere-
ogenic centers. The alkylation products obtained possessed a
cis stereochemistry, which was confirmed by the synthesis of
natural 4ꢀ-O-methylpiscidic acid dimethyl ester 2.
Most of these structures can be seen basically as alkylated
tartaric acid derivatives and although several syntheses of these
compounds have been reported,1,2,4 the work of Seebach5 is notable
for its concise approach. They reported the alkylation of isopropy-
lidene protected tartaric acid esters 6 and isolated a mixture of the
monoalkylated products. The major diastereoisomer, 7, obtained
by this approach, did not have the stereochemistry of natural
piscidic acid and was therefore not a precursor for its synthesis
or that of stereochemically related natural analogues.
(+)-Piscidic acid 1 is one of the constituents of the hypnotic and
narcotic extracts of Piscidia erythrina L. (Jamaica dogwood) and
is equally a component of the antitussic extracts isolated from
Dioscorea nipponica, a medicinal plant used for treating chronic
bronchitis.1,2 Piscidic acid 1 and O-methylpiscidic acid have been
shown to be linked to phosphorous uptake in pigeon peas, Cajanus
cajan (L.) Millsp., an important crop in India.1,2 Furthermore, a
commercial remedy for the reduction of menopausal symptoms
is based upon an extract of dried Cimicifuga racemosa rhizome,
which, among other compounds, contains the closely related
piscidic and fukiic acid 3 esters.3 4ꢀ-O-Methylpiscidic acid dimethyl
ester 2 has been isolated from Narcissus poeticus L.1,2 and the
glucoside loroglossin 4 is a characteristic constituent of orchids.4
Structurally related eucomic acid 5 has been isolated from the
bulbs of Eucomis punctata L’He´rit.2
Tartaric acid is an abundant versatile chiral small building block
for the asymmetric synthesis of natural products.6 The dioxane
8, readily available from tartaric acid,7 has a rigid structure and
undergoes stereoselective aldol reactions possibly via a dienolate.8
We predicted that the lithium enolate of (2R,3R,5R,6R)-
dimethoxy-2,3-dimethyl-1,4-dioxane-5,6-dithiocarboxylate
8
would undergo stereoselective mono- or dialkylations to afford
new stereogenic centers. If the dienolate was involved then this
acetal would provide a chiral memory. It was also expected that
this would provide us with the correct stereochemistry for the
synthesis of the natural product 4ꢀ-O-methylpiscidic acid dimethyl
ester 2,1,2 intermediate for the synthesis of (+)-piscidic acid 1.1,2
Other natural products of the same family, such as fukiic acid 3
and its esters,1–3 loroglossin 44 and eucomic acid 52 could also
be available using the same methodology. The synthesis of the
well characterised ester 2 would provide a way of confirming the
stereochemical outcome of the alkylation reactions.
Alkylation of the lithium dienolate of 8 afforded mono and
dialkylated products (Table 1). The degree of alkylation was
dependent upon the nature and reactivity of the halide used
and thus the proportion of monoalkylated and/or dialkylated
compounds obtained. The effect of the size of the halide, the
amount of reagent, the presence of HMPA, and the reaction
temperature and time were also studied and Table 1 represents
a selection of alkylation reactions which gave significant results.
HMPA was essential for the successful outcome of the reaction
and although DMPU could also be used, the yield of alkylated
product was lower and the product was more difficult to purify.
Alkylation with MeI afforded the dialkylated product 9 ex-
clusively (Table 1). Our assignment of the cis stereochemistry
aInstituto de Tecnologia Qu´ımica e Biolo´gica, Universidade Nova de Lisboa,
Apartado 127, 2780-156, Oeiras, Portugal
bDepartamento de Qu´ımica e Bioqu´ımica, Faculdade de Cieˆncias, Univer-
sidade de Lisboa, Rua Ernesto de Vasconcelos, Lisboa, Portugal E-mail:
maycock@itqb.unl.pt.
† Electronic supplementary information (ESI) available: experimental
procedures and spectroscopic data for all new compounds. See DOI:
10.1039/b606362b
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2361–2363 | 2361
©

View Article Online
Table 1
the alkyl iodide (Fig. 1). The steric hindrance exerted by an alkyl
group in the second alkylation of the monoenolate is much larger
for the allyl group than the methyl group, as seen in the structures
depicted in Fig. 1. The first allyl group can block the attack from
the same side and thus the second allyl group enters trans to the
first. When the installed group is a methyl, the steric hindrance of
the thioester group prevails, the second methyl group entering cis
to the first one.
Fig. 1
Eq.
halide
Steric hindrance also played an important role in the second
alkylation of monobenzylated derivative 16 (Scheme 1). It was
possible to introduce a methyl group in good yield (82%) to furnish
a non-symmetrically dialkylated product 17. We only can assume
that it is the trans product since the installed benzyl group is bulky.
However, it was not possible to introduce a second benzyl group
to the monobenzylated compound indicating that the size of the
electrophilic species is also a limiting factor.
Eq. LDA Halide
Product/yield (%)
2.2
2.2
MeI
5
5
9 (99)
10 (62)
1.5
1.5
11 (82)
2.2
2.2
1.5
12 (65)
5
13 (7), 14 (50)
2.2
2.2
3
3
15 (62)
16 (60)
Scheme 1 a) LDA, HMPA, Mel, THF, −78 ◦C/0 ◦C, 82%.
BnBr
The absolute stereochemistry of the monoalkylated products
needed confirmation and to resolve this problem we attempted a
synthesis of the natural compound 2.
a) HMPA, THF, −78 ◦C to 0 ◦C, 2 h.
20
Monoalkylated product 11, [a]D −14.3 (0.44, CHCl3), (Table 1)
was based upon the non-equivalence of the methyl groups in
the NMR spectrum, indicating that the molecule no longer had
C2-symmetry. To further prove that we had the cis dimethylated
product, cleavage of the acetal afforded a meso diol with a rotation
of zero. This dialkylation is further proof of the formation of a
dienolate and of its higher reactivity since Seebach could only
alkylate with more reactive alkyl halides.5
Interestingly, no traces of monomethylated products were
found, even when the reaction was not complete. The same
reaction conditions were reproducible on a larger scale.
Alkylation using 1-bromo-2-butyne afforded a 62% yield of the
monoalkylated compound 10 (Table 1). No dialkylated product
was detected even when 5 eq. of the halide were used. With the
more bulky halides, such as benzyl bromide and 4-methoxybenzyl
bromide, no dialkylated products were ever formed. With propar-
gyl bromide 7% of the cis dialkylated product was obtained,
however when trying this reaction with a small excess of the
halide (1.5 eq.), a considerable amount of starting material
was recovered. Interestingly, alkylation with allyl iodide afforded
the C2 symmetric trans diallylated product 15 as seen by the
equivalence of protons in the NMR spectrum.
was further transformed, to the enantiomer of natural product
20
24
ent-2, [a]D −43.9 (0.86, CHCl3) (isolated natural product, [a]D
+44 2, c = 0.60, CHCl3;2 synthesised product, [a]D +46 2, c =
1.01, EtOH2), as indicated by the opposite optical rotation for the
natural compound. Using ent-8 from D-tartaric acid, natural 1 and
2 were obtained (Scheme 2).
25
20
Thus, alkylation of dithioester ent-8 afforded ent-11, [a]D
+15.0 (1.2, CHCl3), as expected. Hydrolysis of the bis-acetal
with TFA and water afforded diol 18 in excellent yield (98%).
Transesterification with MeONa in MeOH was also accomplished
in high yield (95%) and the compound obtained was indeed the
20
natural product 2, [a]D +42.8 (1.02, CHCl3) (isolated natural
product,2 [a]D +44 2, (c = 0.60, CHCl3); synthesised product,
24
25
[a]D +46 2, (c = 1.01, EtOH). Transesterification of the bis-
acetal ent-11 was very slow (>24 h), but the transesterification
of hydroxyester 18 was much faster (15 min). In this way, we
concluded that the stereochemistry of the monoalkylated products
was syn, whereas Seebach5 obtained the anti-product 19.
The obtention of the cis product with methyl iodide and the trans
product with allyl iodide, could be explained by the bulkyness of
2362 | Org. Biomol. Chem., 2006, 4, 2361–2363
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Scheme
2 a) LDA, 4-methoxybenzyl bromide, HMPA, THF,
−78 ◦C/0 ◦C, 82%. b) TFA, CH2Cl2, H2O, D, 98%. c) MeONa, MeOH,
Scheme
3
a) LDA, 3-bromo-2-methylpropene, HMPA, THF,
−78 ◦C/0 ◦C, 65%. b) TFA, CH2Cl2, H2O, D, 99%. c) MeONa,
rt, 95%.
MeOH, rt, 90%. d) Pd/C 10%, AcOEt, 50 atm, 88%.
In order to introduce the isobutyl group present in loroglossin
of 4ꢀ-O-methylpiscidic acid dimethyl ester 2, and in the formal
syntheses of piscidic acid 1, loroglossin 4 and eucomic acid 5.
We thank Fundac¸a˜o para a Cieˆncia e a Tecnologia for a grant
conceded to M. R. V.
4, alkylation of the dioxane ent-8 was carried out using 3-bromo-
20
2-methylpropene (Scheme 3) to afford ent-12, [a]D +41.2 (2.37,
CHCl3), in good yield (65%).
Hydrolysis of the acetal gave diol 20 quantitatively, and
transesterification afforded the dimethyl ester 21. The double
bond was saturated using H2/Pd/C 10%, under 50 atm, and
References
20
the nucleus of loroglossin 22 ([a]D +41.8 (0.29, CHCl3), lit.
1 (a) H. Toshima, M. Saito and T. Yoshihara, Biosci., Biotechnol.,
Biochem., 1999, 63, 964; (b) H. Toshima, M. Saito and T. Yoshihara,
Biosci., Biotechnol., Biochem., 1999, 63, 1934.
2 W. Heller and C. Tamm, Helv. Chim. Acta, 1974, 57, 1766.
3 S. O. Kruse, A. Lohning, G. F. Pauli, H. Winterhoff and A. Nahrstedt,
Planta Med., 1999, 65, 763.
4 R. W. Gray, A. Guggisberg, K. P. Segebarth, M. Hesse and H. Schmid,
Helv. Chim. Acta, 1976, 59, 645.
5 R. Naef and D. Seebach, Angew. Chem., Int. Ed. Engl., 1981, 20, 1030.
6 J. Gawronski and K. Gawronska, in Tartaric and Malic Acids in synthesis,
1999, Wiley-Interscience, New York.
7 (a) M. T. Barros, A. J. Burke and C. D. Maycock, Tetrahedron Lett., 1999,
40, 1538; (b) D. J. Dixon, A. C. Foster, S. V. Ley and D. J. Reynolds,
J. Chem. Soc., Perkin Trans. 1, 1999, 1631.
8 (a) M. T. Barros, C. D. Maycock and M. R. Ventura, Org. Lett., 2003, 5,
4097; (b) M. T. Barros, A. J. Burke, J.-D. Lou, C. D. Maycock and J. R.
Wahnon, J. Org. Chem., 2004, 69, 7847.
[a]D +43 (CHCl3)4) was thus obtained in 88% yield (Scheme 3).
Further transformation to the p-(b-D-glucopyranosyloxy)phenyl
esters would afford loroglossin 4.
All attempts to introduce a protected 3,4-dihydroxybenzyl
group, in order to prepare fukiic acid 3 by a simple alkylation
were surprisingly unsuccessful, due to the instability of 3,4-
dibenzyloxybenzyl bromide and 3,4-dimethoxybenzyl bromide,
under the reaction conditions. The synthesis of eucomic acid 5
from 2 has already been described.2
In summary, we report the stereoselective mono- and dialkyl-
ation of the tartrate derivative 8, the efficiency of which was shown
to be principally dependent upon the nature of the halide em-
ployed. This reaction was used as the key step for a short synthesis
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 2361–2363 | 2363
©