An asymmetric approach to coumarin anticoagulants via hetero-Diels-Alder cycloaddition

Article abstract of DOI:10.1016/S0957-4166(01)00124-0

We have developed a general, two-step protocol for the synthesis of chiral non-racemic coumarin anticoagulants (e.g. warfarin, coumachlor and acenocoumarol). This approach features a one-pot three-component tandem Knoevenagel-hetero-Diels-Alder reaction between in situ generated 3-arylidene-2,4-chromanediones and iso-propenyl ether derived from (-)-menthol.

Full text of DOI:10.1016/S0957-4166(01)00124-0

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 12 (2001) 707–709
An asymmetric approach to coumarin anticoagulants via
hetero-Diels–Alder cycloaddition
Giancarlo Cravotto,a,* Gian Mario Nano,a Giovanni Palmisanob,* and Silvia Tagliapietraa
aDipartimento di Scienza e Tecnologia del Farmaco, Via Giuria 9, 10125 Turin, Italy
bDipartimento di Scienze Chimiche, Fisiche e Matematiche, Via Valleggio 11, 22100 Como, Italy
Received 26 February 2001; accepted 9 March 2001
Abstract—We have developed a general, two-step protocol for the synthesis of chiral non-racemic coumarin anticoagulants (e.g.
warfarin, coumachlor and acenocoumarol). This approach features a one-pot three-component tandem Knoevenagel–hetero-
Diels–Alder reaction between in situ generated 3-arylidene-2,4-chromanediones and iso-propenyl ether derived from (−)-menthol.
© 2001 Elsevier Science Ltd. All rights reserved.
A number of coumarin compounds possessing antico-
agulant activity (like their prototype, dicoumarol) have
been synthesised as potential drugs for the management
of myocardial infarction. Starting from dicoumarol, a
number of modifications (“1a–1c) of the 4-hydroxy-
coumarin moiety have been performed,1 including the
introduction of a (het)arylmethyl group at the 3-posi-
tion and the stereochemical control of the newly cre-
ated stereocentre.
both enantiopure warfarin enantiomers3 is feasible
through ‘classical’ racemate resolution, by crystallising
the quinidine/quinine salts, as well as by direct chro-
matographic separation on a chiral stationary phase.
Our interest in this area was stimulated by the prospect
of designing an enantioselective entry to this class of
oral anticoagulants, which includes compounds such as
1a–1c, based on chiral auxiliary-directed p-facial dis-
crimination in an intermolecular hetero-Diels–Alder
(HDA) reaction (Scheme 1). At the outset of our inves-
tigations there were no published reports of any asym-
metric synthesis of 1a. However, since 1996 Li et al.
reported enantioselective hydrogenation in the presence
of Duphos–Rh(I) catalyst as the key step in the synthe-
sis of (−)-1a.4
Warfarin 1a is today the dominant coumarin anticoag-
ulant owing to its excellent potency and good pharma-
cokinetic profile. While its marketed form is the
racemic sodium salt (Coumadin®), the anticoagulant
activity of the (S)-(−) enantiomer is known to be six
times higher than that of the (+)-enantiomer.2 Access to
X
X
X
O
O
O
O
4
2
3
O
O
O
OR
Me
+
Me
OR
O
OH
O
Chiral auxiliary
removal
[4+2] / [endo]
1
4
2
3
a X=H; b X=Cl; c X=NO2
A R=Me; B R=(-)-menthyl
Scheme 1. Asymmetric synthesis of chiral non-racemic coumarin anticoagulants 1a–1c.
* Corresponding authors. Fax: +39(11)6707687; e-mail: cravotto@pharm.unito.it
0957-4166/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(01)00124-0

708
G. Cra6otto et al. / Tetrahedron: Asymmetry 12 (2001) 707–709
3-Arylidene-2,4-chromanediones 3, deriving from the
Knoevenagel condensation of 4-hydroxycoumarin with
aldehydes, undergo an easy HDA cycloaddition with
suitable iso-propenyl ethers 4. Heterodienes 3 are elec-
tron-deficient 4p-systems that may undergo inverse-
electron-demand Diels–Alder reactions (which are
favoured by a dominant interaction between the
LUMO of 1-oxo-1,3-butadiene and the HOMO of elec-
tron-rich 2p-systems such as enol ethers, strained
unsubstituted olefinic and acetylenic dienophiles). These
are usually concerted non-synchronous transformations
that preserve the configuration of the dienophile and
usually exhibit high regioselectivity.5
reaction have been described, with the chiral auxiliary
attached to the heterodiene10,11 or to the dienophile.12,13
A survey of chiral auxiliaries was undertaken in order
to optimise product diastereoselectivities and to
increase yields such that this process might become
synthetically useful. Specifically, iso-propenyl ether 4B,
derived from (−)-(1R,2S,5R)-menthol, has thus far
given the best results among all of the chiral non-
racemic vinyl ethers tested. A key advantage of using
the menthyl group as a chiral auxiliary is that both
enantiomers are available commercially, the auxiliary is
cheap and is recyclable if required.
In a preliminary attempt to synthesise ( )-1a–1c, the
isolation of the requisite 3-arylidene-2,4-chromane-
diones 3a–3c under classical Knoevenagel conditions,
even at low temperature, proved unsuccessful, being
uniformly frustrated by the formation of 2:1 adducts as
the sole products.6
Since the methods previously reported for the prepara-
tion of these compounds [i.e. Hg(II)-promoted trans-
etherification reaction with n-butyl vinyl ether]14 were
not applicable, the required dienophile 4B was prepared
in an unoptimised yield of 51% with e.e. of 96% by
modification of a literature method [(i) iso-propenyl
methyl ether, dichloroacetic acid, rt; (ii) TMSOTf,
TEA, DCM, 0°C“rt].15
Therefore, we needed to develop different conditions so
that Knoevenagel adducts 3 would not react further.
Pilot studies disclosed that the one-pot three-compo-
nent reaction of 4-hydroxycoumarin, benzaldehyde (1.1
equiv.) and 2-methoxypropene (2.2 equiv.) in dry diox-
ane (90°C, 4 hours, screw cap pressure tube) in the
presence of catalytic ethylenediammonium diacetate
The reaction of 4B (1.1 equiv.) with benzaldehyde (1.2
equiv.) and 4-hydroxycoumarin in dry dioxane in the
,
presence of Tietze base and 5 A molecular sieves at
80°C was complete in 10 hours (TLC). Prior to chro-
matographic separation, analysis of the crude reaction
mixture by 1H NMR and HPLC analysis (Waters,
m-Porasil 7.8×300 mm, column) revealed that the reac-
tion proceeded with endo-selectivity (with an endo/exo
ratio of 4.1:1).
,
(Tietze base) and powdered oven-dried 5 A molecular
sieves proved to be the most convenient and satisfac-
tory, giving a mixture of endo- and exo-adducts 2Aa in
a 3:1 ratio (83% yield). These isomers correspond to the
cis- and trans-products, respectively, regarding the 2-
OMe and 4-Ph groups.
After release of the chiral auxiliary (19:1 TFA/H2O),
the enantioselectivity (76% e.e.) of the HDA reaction
was determined by chiral HPLC of the crude product
(on an R,R Whelk-O type column). In chromatographic
separation of the crude product on a neutral alumina
4-(4-Chlorophenyl)- 2Ab and 4-(4-nitrophenyl)-pyra-
nocoumarin 2Ac were similarly obtained from the cor-
responding aldehydes in 78–84% range yield, with good
diastereoselectivity favouring the cis(endo)-isomer (cis/
trans ratio from 1.3:1 to 2.1:1). For all cycloadducts
chemical shifts of H-(4) for the major isomers were
consistently 0.1–0.3 ppm higher than those for minor
isomers. The OMe group is c-equatorial in a half-chair
conformation (3H2) for cis adducts and c-axial for trans
adducts, considering the 4-aryl substituent to be c-
equatorial in all cases.7
column
(hexane/EtOAc
19:1)
cis(endo)-adducts
(2R*,4S*)-2Ba eluted first, followed by trans(exo)-
adducts (2R*,4R*)-2Ba. It was found that the early
eluted cycloadducts as
a rule can be efficiently
upgraded to diastereomeric purity by recrystallisation.
Accordingly, the endo-mixture was dissolved in the
minimum volume of EtOAc and diluted with twice the
volume of di-iso-propyl ether. The resulting suspension
was warmed to 40°C until a clear solution was obtained
and allowed to stand at rt until crystals were formed.
Two crystallisations and acid hydrolysis of enriched
(2R,4S)-2Ba provided (S)-warfarin 1a with e.e. of 95%
by chiral HPLC in 61% overall yield referred to 4-
hydroxycoumarin. The cleaved chiral auxiliary can eas-
ily be recovered at this stage if required.
With the cycloaddition step under control, unmasking
of the carbonyl function in 2Aa–2Ac was next
addressed. As expected, exposure of ketals to 3N HCl
in the presence of SiO2 as a promoter8 or 19:1 TFA/
H2O (overnight, rt) furnished the targeted compounds
1a (warfarin), 1b (coumachlor) and 1c (acenocoumarol)
in near to quantitative yields. It should be noted that
compounds 1a–1c exist in solution as equilibrium mix-
tures of open-chain keto and cyclic hemiketal forms (as
a diastereomeric mixture) as observed previously.9
Although the C-(2) stereochemistry is irrelevant to the
present work, the (2R,4S)-absolute configuration in
upgraded 2Ba was evident from: (i) comparison of CD
spectra (which showed the presence of an inherently
dissymmetric chromophore with [q]:1.2×105 observed
at ca. 220 nm) with those of the (2R,4S)-cyclocoumarol
2Aa and (S)-warfarin;16 (ii) the elution order in chiral
HPLC of the hydrolysis products; (iii) NMR studies
(1H NMR, COSY and NOESY) and our previous
From this finding we recognised the potential of the
two-step cycloaddition/hydrolysis protocol for the gen-
eral and enantioselective synthesis of coumarin antico-
agulants.
A few examples of chiral non-racemic
2-alkoxy-3,4-dihydro-2H-pyran synthesis via HDA

G. Cra6otto et al. / Tetrahedron: Asymmetry 12 (2001) 707–709
709
results on other related cycloadducts.7b Under our opti-
mised conditions, 4-chloro- and 4-nitrobenzaldehyde
reacted with 4B to give (S)-coumachlor 1b in 56%
overall yield with 93% e.e. and (S)-acenocoumarol 1c
with 95% e.e. in 59% overall yield.
3. (a) Bargmann-Leyder, N.; Tambute, A.; Begos, A.;
Caude, M. Chromatographia 1993, 37, 433; (b) Soini, H.;
Stefansson, M.; Riekkola, M.-L.; Novotny, M. V. Anal.
Chem. 1994, 66, 3477; (c) Armstrong, D. W.; Zhou, E.
Y.; Chen, S.; Le, K.; Tang, Y. Anal. Chem. 1994, 66,
4278; (d) See also: Jeyaraj, G. L.; Porter, W. R. J.
Chromatogr. 1984, 315, 378 and references cited therein.
4. (a) Robinson, A.; Li, H.-Y.; Feaster, J. Tetrahedron Lett.
1996, 46, 8321; (b) US Pat. 5686631, 1996 (DuPont);
Chem. Abstr. 1997, 126, 199455a;; (c) Demir, A. S.;
Tanyeli, C.; Gulbeyaz, V.; Akgun, H. Turk. J. Chem.
1996, 20, 139.
5. (a) Boger, D. L.; Weinreb, S. M. In Hetero Diels–Alder
Methodology in Organic Synthesis; Wasserman, H. H.,
Ed. Oxobutadienes. Academic Press: San Diego, 1987;
Vol. 47, pp. 167–204; (b) Jørgensen, K. A. Angew. Chem.,
Int. Ed. 2000, 39, 3558.
6. (a) Appendino, G.; Cravotto, G.; Tagliapietra, S.; Nano,
G. M.; Palmisano, G. Helv. Chim. Acta 1990, 73, 1865;
(b) Appendino, G.; Cravotto, G.; Palmisano, G.; Annun-
ziata, G.; Toma, L. J. Org. Chem. 1994, 59, 5556.
7. (a) Valente, E. J.; Poerter, W. R.; Trager, W. F. J. Med.
Chem. 1978, 21, 231; (b) Annunziata, R.; Raimondi, L.;
Nano, G. M.; Palmisano, G.; Tagliapietra, S. Magn.
Reson. Chem. 1997, 35, 721.
The reaction of 4B with the in situ generated 3a–3c in
the presence of Lewis acids at rt was attempted with
little success: either the products were obtained without
any enantiomeric enrichment, or the Lewis acid was not
an effective promoter. This is presumably due to the
lability of electron-deficient olefin dienophiles, e.g. 3, in
the presence of many Lewis acids.
The major reaction products probably arise from an
intramolecular HDA through an s-trans conformation
of 4B. This reaction exposes the C(a) Re-face of the
double bond to intermolecular attack by (Z)-heterodi-
ene, while positioning the C(a) Si-face so as to be
shielded by the i-Pr group of the chiral auxiliary.
In summary, our strategy based on a thermal inter-
molecular HDA reaction represents a novel method for
an enantiocontrolled installation of the stereogenic cen-
tre in coumarin anticoagulants. Furthermore, the suc-
cessful incorporation of different substituents at C-(4)
(cyclocoumarol numbering) makes it suitable for gener-
ating a library of enantiomerically enriched warfarin-
like analogues.
8. Huet, F.; Lechevallier, A.; Pellet, M.; Conia, J. M. Syn-
thesis 1978, 63.
9. Valente, E. J.; Lingafelter, E. C.; Porter, W. R.; Trager,
W. F. J. Med. Chem. 1977, 20, 1489.
10. Dujardin, G.; Molato, S.; Brown, E. Tetrahedron: Asym-
metry 1993, 4, 193 and references cited therein.
11. For the use of chiral non-racemic enol ethers in HDA
reaction with inverse electron demand, see: (a) Posner, G.
H.; Wettlaufer, D. G. Tetrahedron Lett. 1986, 27, 667; (b)
Choudhury, A.; Franck, R. W.; Gupta, R.-B. Tetra-
hedron Lett. 1989, 30, 4291; (c) Ba¨ckvall, J.-E.; Rise, F.
Tetrahedron Lett. 1989, 30, 5347; (d) Tietze, L. F.; Meier,
H.; Nutt, H. Liebigs Ann. Chem. 1990, 253; (e) Tietze, L.
F.; Schneider, C. J. Org. Chem. 1991, 56, 2476; (f) Tietze,
L. F.; Kettschau, G. Top. Curr. Chem. 1997, 190, 1; (g)
Tietze, L. F. Curr. Org. Chem. 1998, 2, 19.
Acknowledgements
This work was supported by the Ministero dell’Univer-
sita` e della Ricerca Scientifica e Tecnologica (MURST),
project: ‘Chimica dei Composti Organici di Interesse
Biologico’ (Fondi ex 40%).
12. Dujardin, G.; Rossignol, S.; Brown, E. Synthesis 1998,
763.
References
13. Dujardin, G.; Rossignol, S.; Brown, E. Tetrahedron Lett.
1996, 37, 4007.
14. Denmark, S. E.; Schnute, M. E.; Marcin, L. R.; Tho-
1. West, B. D.; Preis, S.; Schroeder, C. H.; Link, K. P. J.
Am. Chem. Soc. 1961, 83, 2676.
2. (a) Hewick, D.; McEwen, J. J. Pharm. Pharmacol. 1973,
25, 458; (b) Baars, L. G. M.; Schepers, M. T.; Hermans,
J. J. R.; Dahlmans, H. J. J.; Thissen, H. H. W. J. Pharm.
Pharmacol. 1990, 42, 861; (c) Manolov, I.; Danchev, N.
D. Eur. J. Med. Chem. 1995, 30, 531.
rarensen, A. J. Org. Chem. 1995, 60, 3208.
15. Dujardin, G.; Rossignol, S.; Brown, E. Tetrahedron Lett.
1995, 36, 1653.
16. Valente, E. J.; Trager, W. F. J. Med. Chem. 1978, 21,
141.
.
.