New phenylglycine-derived primary amine organocatalysts for the preparation of optically active warfarin

Article abstract of DOI:10.1002/ejoc.200900664

In this work we present new, fully synthetic phenylglycinederived primary amine organocatalysts useful, for the onestep preparation, of optically active warfarin, an. important anticoagulant. Both enantiomeric forms of the catalysts are equally available and can be prepared by robust procedures without recourse to chromatographic purification. Together with a co-catalyst, particularly acetic acid or2,4-dinitrophen- ol, they can furnish warfarin in approximately 80 % ee and represent inexpensive alternatives to other primary amine organocatalysts such as the chiral diamines and. Cinchona-derived primary amines.

Full text of DOI:10.1002/ejoc.200900664

FULL PAPER
DOI: 10.1002/ejoc.200900664
New Phenylglycine-Derived Primary Amine Organocatalysts for the
Preparation of Optically Active Warfarin
[a]
[a]
[a]
[a]
Tor E. Kristensen, Kristian Vestli, Finn K. Hansen, and Tore Hansen*
Keywords: Amines / Amino alcohols / Asymmetric synthesis / Michael addition / Organocatalysis
In this work we present new, fully synthetic phenylglycine-
derived primary amine organocatalysts useful for the one-
step preparation of optically active warfarin, an important
anticoagulant. Both enantiomeric forms of the catalysts are
equally available and can be prepared by robust procedures
without recourse to chromatographic purification. Together
with a co-catalyst, particularly acetic acid or 2,4-dinitrophen-
ol, they can furnish warfarin in approximately 80% ee and
represent inexpensive alternatives to other primary amine or-
ganocatalysts such as the chiral diamines and Cinchona-de-
rived primary amines.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
alone or in conjunction with other functionalities,[4f,4g] is a
rapidly expanding field within organocatalysis with (among
others) primary amino acids, Cinchona-derived amines and
Introduction
Warfarin is a Vitamin K antagonist, inhibiting Vita-
min K epoxide reductase and thereby decreasing blood co-
agulation by preventing the Vitamin K-dependent synthesis
of blood-clotting proteins. In the form of its sodium salt
[4]
chiral diamines being the preferred catalysts. Unlike sec-
ondary amine organocatalysts, which can react with alde-
hydes, ketones and α,β-unsaturated aldehydes, the primary
amine catalysts can also easily convert the less active α,β-
unsaturated ketones.
Herein we wish to report our studies on the asymmetric
organocatalytic preparation of warfarin, its preference for
primary amine organocatalysts over the standard pyrrol-
idine-derived secondary amines and the development of
new phenylglycine-derived organocatalysts that are cheaper
and more large-scale-adaptable than previous primary
amine organocatalysts, as well as nearly equally affordable
in both enantiomeric forms.
®
®
(
Coumadin , Marevan ) it is one of the most widely pre-
scribed anticoagulants for prevention of thrombosis and
embolism. Although currently prescribed as the racemate,
activity and metabolism are markedly dissimilar for the two
enantiomers, the (S) enantiomer being more active than the
(
R) enantiomer by a factor of 2–5.[1] Efficient asymmetric
syntheses of warfarin are therefore of great interest, and
organocatalysis in particular has proven itself a valuable in-
strument in the preparation of enantiomerically enriched
[
2]
warfarin.
The two key building blocks for virtually all preparations
of warfarin are benzylideneacetone and 4-hydroxycouma-
rin. They are inexpensive industrial products, benzyl-
ideneacetone being simply the aldol condensation product
of acetone and benzaldehyde, whereas 4-hydroxycoumarin
Results and Discussion
The research group of Jørgensen in Denmark reported
in 2003 that the imidazolidine catalyst 4 was an efficient
asymmetric catalyst for the formation of warfarin (3) from
is readily available from, among others, acetylsalicylic acid
or o-hydroxyacetophenone.[
3]
Through organocatalysis,
these two simple building blocks can be forged directly into
optically active warfarin.[ The asymmetric Michael ad-
dition of 4-hydroxycoumarin to benzylideneacetone is of
special interest within organocatalytic reactions since it not
only provides a product of crucial medical importance, but
also because it mechanistically favours primary amine or-
4
-hydroxycoumarin (1) and benzylideneacetone (2), as de-
2]
[
2a]
picted in Scheme 1.
In 2006, a report argued convincingly that the efficiency
of imidazolidine catalyst 4 in the preparation of warfarin
originated in its decomposition into chiral diamine 5 (Fig-
ure 1) and that this chiral diamine was the genuine catalyst
[
2]
[4]
ganocatalysts. Primary amine organocatalysis, either
[
2b]
for the transformation.
diamine directly, using acetic acid as co-catalyst, and ob-
The authors then employed this
[
a] Department of Chemistry, University of Oslo,
P. O. Box 1033 Blindern, 0315 Oslo, Norway
Fax: +47-2285-5441
[
2b]
tained similar or better results.
Importantly, the authors
postulated a diimine intermediate as essential for the high
E-mail: tore.hansen@kjemi.uio.no
[
2b]
degree of asymmetric induction.
Interestingly, replacing
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.200900664.
diaryldiamine 5 with alkyldiamine 6 gave reduced stereose-
Eur. J. Org. Chem. 2009, 5185–5191
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5185

T. E. Kristensen, K. Vestli, F. K. Hansen, T. Hansen
FULL PAPER
Scheme 1. First organocatalytic preparation of warfarine.[2a]
[
11c]
lectivity and a different reaction pattern, an effect that was HN /CHCl ,
used in their preparation through the Mit-
3
3
attributed to the greater basicity of 6 vs. 5.[ More recently, sunobu inversion/Staudinger reduction, as well as the need
the combination of chiral diamine 5 (or its enantiomer) and for chromatographic purification.
acid co-catalyst has also proven successful in the archetypi- We reasoned that the most convenient starting material
2b]
[
11]
cal aldol reactions of substituted benzaldehydes with for primary amine organocatalysts would be a natural or
ketones, a manifestation of the growing utility of primary readily available synthetic amino acid, just as is the case for
amine organocatalysts.[
5]
the more common secondary amine organocatalysts 9–12
Figure 2).
(
Figure 1. Organocatalysts for the asymmetric preparation of warfa-
rin.
Figure 2. Traditional secondary amine organocatalysts.
[
2]
As for secondary amines, proline is reported to give ex-
In 2007, a third report on the organocatalytic prepara- cellent conversion in the preparation of warfarin, but race-
tion of warfarin was published,[2c] this time by using the mic product.
[2a]
We tested all the four most common sec-
Cinchona-derived amines 9-amino-9-deoxyepiquinine (7) ondary amine organocatalysts in the preparation of warfa-
and 9-amino-9-deoxyepicinchonine (8), iminium catalysts rin from 1 and 2, namely -proline (9), the typical MacMil-
previously developed by the same research group.[ More lan-type imidazolidinone 10 and the Jørgensen/Hayashi di-
6]
than 90% ee was obtained by reaction in CH Cl2 with arylprolinols 11 and 12. As can been clearly seen from
2
CF CO H as co-catalyst, although still at 20 mol-% of cata- Table 1, these standard organocatalysts show negligible
3
2
[
2c]
lyst loading. Interestingly, in the original report for these conversion (even after more than 9 d of reaction) and give
Cinchona-derived amines, the authors also described a di- racemic or near racemic products. As mentioned, conver-
arylamino alcohol derived from -valine as a useful, but less sion is good for proline, but enantioselectivity is non-exist-
selective catalyst for Michael addition to α,β-unsaturated ent. Given the extremely poor results of the secondary
[
6]
ketones. Such Cinchona-derived primary amines have re- amines in terms of stereoselectivity, it is tempting to specu-
cently also proven useful for the Michael addition of 1,3- late whether an organocatalytic asymmetric enamine/imin-
diaryl-1,3-propanedione to nitro olefins.[7]
ium ion pathway does in fact operate at all (or only to a
Although chiral diamines such as 5 and Cinchona-de- very small extent), or whether these transformations may
rived amines such as 7 and 8 are effective iminium catalysts, simply be general acid/base catalysed (probably depending
their preparations are unfortunately adversely ineffective. A on co-catalysts), as is the case in the classical preparation
racemic mixture of (1R,2R)-1,2-diamino-1,2-diphenyle- of racemic warfarin with bases such as Et N, Na PO or
3
3
4
[
12]
thane (5) and (1S,2S)-1,2-diamino-1,2-diphenylethane (ent- alkali fluorides. This has very important implications for
) can be resolved by diastereomeric salt formation with the more active primary amine organocatalysts as well, be-
tartaric or mandelic acids. However, the synthesis of the cause alongside the asymmetric organocatalytic pathway a
racemate starting material is surprisingly cumbersome, a racemic pathway may operate, albeit slowly. With very long
5
[
8]
[9]
[10]
fact only rarely mentioned in the literature.
An efficient reaction times, the racemic mechanism may start to domi-
synthesis was a long-standing problem,[ and better pro- nate, lowering the stereoselectivity, even though overall
10]
cedures are still needed. On the other hand, the use of Cin- yields are improving.
chona-derived amines on a preparative scale, although orig-
By taking readily available primary amino acids as the
inating from relatively cheap Cinchona alkaloids, is also se- natural starting point for the preparation of new primary
verely hampered by the costly and/or hazardous azodicar- amine organocatalysts, we prepared a series of bulky amino
[
11a]
boxylates and azide reagents, such as HN /benzene,
or alcohols (13–17) (Figure 3). They are easily prepared on a
3
5186
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 5185–5191

New Organocatalysts for the Preparation of Optically Active Warfarin
Table 1. Traditional secondary amine organocatalysts in the prepa- and approximately 80% ee in only 24 h, almost identical
ration of optically active warfarin from 1 and 2.
[2b]
to the previous reports on diamine 5.
It was originally
Catalyst[a] Reaction time [h]
Yield [%][b]
ee [%][c]
anticipated by us that introducing sterically more de-
manding side chains through cyclohexylglycine or tert-leu-
cine (amino alcohols 16–17) would enhance the enantio-
selectivity even further, a hypothesis obviously invalidated
by experiment. Clearly, the presence of an aromatic moiety
9[d]
24
88
5
19
34
racemic
racemic
14
10
11
12
235
235
235
racemic
[
a] All reactions, except with 9, were undertaken with 20 mol-% is crucial. Among the amino acids, phenylglycine is in an
catalyst at room temperature in THF with 1 equiv. of CH
as co-catalyst. [b] Isolated yield. [c] Determined by chiral HPLC.
d] In DMSO with 50 mol-% catalyst.
3 2
CO H
especially fortunate position. -Phenylglycine is a side-
chain moiety in several important penicillins (e.g. ampicil-
lin) and cephalosporins (e.g. cephalexin), and several thou-
sand tons of this compound are prepared every year by
[
[
18]
large scale, with no chromatographic purification involved, DSM and others, primarily for this very purpose. In the
by treating the appropriate methyl or ethyl ester hydrochlo- traditional industrial synthesis, racemic phenylglycine is
ride salt of the amino acid with a five- to tenfold excess of prepared by a Strecker reaction of benzaldehyde, followed
PhMgBr in Et O or THF. In this way, naturally occurring by hydrolysis of the nitrile. This racemate is then resolved
2
-leucine and -phenylalanine gave 13 and 14,[ whereas by diastereomeric salt formation,
13]
[18a]

with (+)-(1S)-cam-
synthetic or semisynthetic -phenylglycine, -cyclohexylgly- phor-10-sulfonic acid being a very effective resolving agent
[
19]
cine and -tert-leucine gave increasingly bulky amino for this compound.
Therefore, phenylglycine is particu-
-Cyclohexylglycine is larly useful as a starting material for primary amine organo-
alcohols 15–17, respectively.[
14–16]
[
15]
available by simple hydrogenation of -phenylglycine,
catalysts since both enantiomers are nearly equally inexpen-
and -tert-leucine is produced on a ton scale by Evonik De- sive.
gussa GmbH through an enzymatic process.[ We tested
17]
All efficient catalysts for the preparation of warfarin [di-
these five bulky amino alcohols in the preparation of warfa- amine 5, cinchona amines 7/8 and diphenylphenylglycinol
rin using acetic acid as co-catalyst, and the results are pre- (15)] then seem to have a common structural motif where
sented in Table 2.
the primary amine has an aromatic moiety in α-position,
preferably together with a somewhat bulky group. Diimine
formation does not seem to be a prerequisite for good
enantioselectivity. To probe the importance of the bulky
side group, we tried amine 18, amino alcohol 19, and silyl
ether 20, in the preparation warfarin (Table 3). α-Methyl-
benzylamine (18) is commercially available in both enantio-
meric forms at very low cost, the racemate being efficiently
[
20]
resolved with malic or tartaric acid.
is conveniently prepared by direct reduction of phenylgly-
Phenylglycinol (19)
[
21–23]
cine with NaBH /I or NaBH /H SO .
The silyl ether
4
2
4
2
4
20 has to the best of our knowledge not been previously
reported, but is prepared from amino alcohol 15 in one
Figure 3. Bulky amino alcohols 13–17.
non-chromatographic step by using the iodine-catalysed
[
24]
silylation with hexamethyldisilazane.
This excellent
method eliminated any need for TMS triflate. The complete
synthesis of the new primary amine organocatalyst 20, with
amino alcohol 15 as an intermediate, is shown in Scheme 2.
Because of the synthetic nature of phenylglycine, the enan-
tiomeric form of catalyst 20 is equally available by starting
with -phenylglycine (syntheses of both amino alcohols are
provided in the Experimental Section) (Figure 4).
Table 2. Bulky amino alcohols 13–17 in the preparation of optically
active warfarin from 1 and 2.
Catalyst[a]
Yield [%][b]
ee [%][c]
Reaction time [h]
1
1
1
1
1
1
1
3
4
5
6
6
7
7
24
24
24
24
192
24
11
18
76
12
65
2
4
11
81
–8
–6
racemic
racemic
The results in Table 3 provide interesting insights. First,
by comparing catalysts 18, 19, 15 and 20 (in the standard
system of THF with 1 equiv. CH CO H), we found the
192
10
3
2
[
a] All reactions were undertaken with 20 mol-% catalyst at room
temperature in THF with 1 equiv. of CH CO H as co-catalyst. [b]
Isolated yield. [c] Determined by chiral HPLC.
enantioselectivity to increase from 33% ee for the simple
α-methylbenzylamine (18), through 51% for phenylglycinol
3
2
(19) to 81% for diphenylphenylglycinol (15) (after 24 h of
As can be clearly seen from Table 2, most of these amino reaction time), all of them with fairly high yields (Entries 1–
alcohols, like the traditional secondary amine organocata- 3). Then, for silyl ether 20, the enantioselectivity reaches
lysts, show no promise for the preparation of warfarine. 86% ee under the same conditions, but only with a modest
However, the phenylglycine-derived amino alcohol 15 is re- 23% isolated yield (Entry 5). This seems to fit nicely in a
markably effective in this transformation, giving good yield pattern where introduction of increasingly bulky groups in-
Eur. J. Org. Chem. 2009, 5185–5191
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5187

T. E. Kristensen, K. Vestli, F. K. Hansen, T. Hansen
Table 3. Primary amine organocatalysts 15 and 18–20 in the preparation of optically active warfarin from 1 and 2.
FULL PAPER
Entry
Catalyst[a]
Co-catalyst (1 equiv.)
Reaction time [h]
Yield [%][b]
ee [%][c]
1
2
3
4
5
6
7
8
9
18
19
15
15
20
20
20
20
15
15
15
15
15
20
20
20
20
20
20
20
20
20
20
20
20
20
20
CH
CH
CH
CH
CH
CH
CH
3
3
3
3
3
3
3
CO
CO
CO
CO
CO
CO
CO
2
2
2
2
2
2
H
H
H
H
H
H
H
24
24
24
48
24
48
72
96
24
24
24
24
24
24
24
48
72
96
96
336
24
96
24
24
24
24
24
90
73
76
90
23
68
80
93
67
62
80
37
64
54
40
46
72
70
54
76
11
18
–
–33
–51
81
65
86
70
68
70
66
79
79
50
69
80
80
83
67
63
79
40
84
81
–
2
[
d]
CH
3
CO
none
CO
CO[
2
H
H[
2
e]
f]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
CH
CH
3
H[
2
3
g]
HFP
2,4-DNP[
none
h]
2,4-DNP
2,4-DNP[
h]
[h]
2,4-DNP
2,4-DNP[
2,4-DNP
2,4-DNP[
h]
[i]
j]
CF
CF CO
CCl
3
CO
2
H
[
k]
3
2
H
3
CO
2
H
[
l]
CSA
HCO
PhCO
–
–
2
H
38
47
44
68
70
73
2
H
PCP[
m]
[
a] All reactions were undertaken with 20 mol-% catalyst at room temperature in THF (unless other solvents are noted) together with
equiv. of co-catalyst (unless otherwise noted). [b] Isolated yield. [c] Determined by chiral HPLC. [d] 10 equiv. of CH CO H. [e] Reaction
in THF/CH Cl (1:1). [f] Reaction in THF/PhF (7:5). [g] HFP = Hexafluoroisopropanol. [h] 2,4-DNP = 2,4-dinitrophenol (20 mol-%).
i] 4 equiv. of DNP. [j] 20 mol-% DNP, reaction in CH Cl . [k] 2 equiv. of CF CO H. [l] CSA = (+)-(1S)-camphor-10-sulfonic acid. [m]
PCP = 2,3,4,5,6-pentachlorophenol.
1
3
2
2
2
[
2
2
3
2
reaction time. It is important to note that 4-hydroxycouma-
rin has an unusually selective solubility, making THF one
of extremely few possible solvents for this reaction. Modi-
fied solvent systems, prepared by adding CH Cl or PhF
2
2
(Entries 10–11), do not seem to offer any added benefits.
The use of CH Cl as the reaction medium is interesting
2
2
because 4-hydroxycoumarin has a poor solubility in this
solvent. Nevertheless, previous disclosures have reported
[
2a,2c]
the successful preparation of warfarin in CH Cl .
Es-
2
2
pecially the Cinchona derivatives 7 and 8 work remarkably
[
2c]
well in this solvent, whereas the silyl ether 20 shows only
very modest results in CH Cl (Entry 20). We hypothesized
2
2
that phase-transfer effects could be at play, but addition of
the phase-transfer catalyst benzyltriethylammonium chlo-
ride to our reaction mixtures in CH Cl was of no use.
Scheme 2. Preparation of primary amine organocatalyst 20.
2
2
As for additives, strongly acidic ones, such as CCl CO H,
3
2
CF CO H and camphorsulfonic acid (Entries 21–24) dras-
3
2
tically reduced the conversion for silyl ether 20, but can give
good selectivity (84% ee for CF CO H, the only one of
3
2
them to give an isolable yield). Additives of medium acidity,
such as HCO H, PhCO H and pentachlorophenol (En-
Figure 4. Structurally similar primary amine organocatalysts.
2
2
tries 25–27) gave higher conversions, but showed no added
creases the selectivity, but at the cost of catalytic efficiency, benefits over standard CH CO H, and hexafluoro-2-propa-
3
2
and then especially for silyl ether 20. However, if the reac- nol (Entry 12) seemed entirely uninteresting. On the other
tion time is lengthened to improve overall conversion, the hand, 2,4-dinitrophenol is an additive that has been recently
enantioselectivity drops from 81% ee to 65% for diphenyl- reported to be unusually effective for primary amine organo-
phenylglycinol (15) after 48 h of reaction time (Entries 3–4) catalysts in aldol reactions,[25] with interesting results also
and from 86% ee to 68% for silyl ether 20 after 72 h of being obtained in this work, although more so for the silyl
5188
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 5185–5191

New Organocatalysts for the Preparation of Optically Active Warfarin
ether 20 (Entries 15–20) than for the more reactive di- 4H
2
O in aqueous H
SO and glacial CH
KMnO , K CO and NaOH in water, all followed by heating.
2
SO
4
, a solution of p-anisaldehyde, concd.
H
2
4
3
CO
2
H in 96% EtOH or a solution of
phenylphenylglycinol 15 (Entry 13). However, also in this
case, reductions in enantioselectivity occurred at prolonged
reaction times (Entries 16–18), and large excesses of the ad-
ditive lowered the yields (Entries 15 and 19) but seemed to
preserve the enantioselectivity somewhat better at long re-
action times. Nevertheless, 2,4-dinitrophenol cannot be
considered a practical additive. It has explosive properties
4
2
3
Merck silica gel (60, 40–63 µm) was used for flash chromatography,
either manually or with a Teledyne Isco CombiFlash Compan-
ion with PeakTrak software (v. 1.4.10), by using EtOAc/hexanes
of technical quality. H and C NMR spectra were recorded with
®
®
1
13
a Bruker Avance DPX-300 or -200 spectrometer operating at 300/
1
13
200 MHz ( H) or 75/50 MHz ( C). Chemical shifts are reported in
(
particularly in the form of its metal salts), as well as a parts per million (δ) and are, unless otherwise noted, reported rela-
considerable toxicity associated with it, and a strong yellow tive to the internal references of the solvent: δ = 2.49/39.7 ppm for
discolouration that is easily carried over to the reaction [D ]DMSO and δ = 7.26/77.0 for CDCl . Electrospray ionisation
6
3
[
26]
mass spectra were recorded with a Micromass Q-Tof-2 mass spec-
trometer. Infrared spectra were recorded with a Perkin–Elmer Spec-
trum One FTIR spectrometer. Melting points were determined
with a Büchi Melting Point B-545 apparatus. Optical rotation was
recorded by using a Perkin–Elmer Instruments 341 Polarimeter at
products.
Conclusions
In essence, we have screened a number of amino acid room temperature. The enantiomeric excess was determined by
®
using a Thermo Scientific SpectraSYSTEM P2000 pump with a
derived primary amines for the preparation of warfarin by
asymmetric Michael addition of 4-hydroxycoumarin to
benzylideneacetone and identified potentially highly useful
SpectraSYSTEM® UV3000 UV/Vis detector and a Chiralcel AD-
®
H column from Daicel Chemical Industries.
phenylglycine-derived structural motifs that offer a useful (R)-Diphenylphenylglycinol (Enantiomer of 15): (–)--α-Phenylgly-
cine (36.26 g, 240 mmol) was suspended in MeOH (350 mL), and
starting point for the preparation of new primary amine
the reaction flask was cooled in an ice/water bath. Dropwise, SOCl
2
organocatalysts. We found diphenylphenylglycinol, espe-
cially in combination with co-catalysts such as acetic acid
or 2,4-dinitrophenol, to be a surprisingly effective organ-
ocatalyst for this transformation, far better than other
amino alcohols. This may be an indication that not only
steric, but also more subtle electronic perturbations most
probably influence the performances of primary amine or-
ganocatalysts substantially. From this, we prepared a new
primary amine organocatalyst from diphenylphenylglycinol
by straightforward silylation. This catalyst may be perceived
as representing a type of primary amine analogue of the
(
26.0 mL, 358 mmol) was added over a period of 20 min. The reac-
tion flask was removed from the ice/water bath, and the solution
was stirred at room temperature for 46 h. The volatiles were evapo-
rated in vacuo, and the residual crystals were slurried in Et
2
O
(
200 mL), vacuum-filtered and dried at room temperature for 28 h
to give -phenylglycine methyl ester hydrochloride (48.12 g, 99%)
as colourless, fibrous crystals.
A three-necked 1000 mL round-bottomed flask, equipped with an
addition funnel, reflux condenser and glass stopper, was charged
with magnesium turnings (34.37 g, 1414 mmol) and an egg-shaped
2
stirring bar (1.5ϫ0.625 in). Dry Et O (80 mL) was added, followed
Jørgensen/Hayashi diarylprolinols 11/12. This catalyst is by a small portion of a solution of bromobenzene (222.51 g,
conveniently available in both enantiomeric forms by a non- 1417 mmol) in dry Et O (430 mL). When the reaction had initiated,
2
chromatographic procedure and together with its parent the rest of the PhBr solution was added over a period of 1 h 20 min,
keeping the mixture at reflux. The mixture was stirred an additional
alcohol represents affordable alternatives to chiral diamines
and Cinchona-derived primary amines for the preparation
of optically active warfarin. This is potentially very useful,
1
h at reflux to dissolve most of the Mg and then diluted with dry
2
Et O (300 mL). The reaction flask was cooled in an ice/water bath,
while -phenylglycine methyl ester hydrochloride (47.98 g,
as the optical purity of warfarin can be further enhanced
2
1
38 mmol) was added carefully in small portions over a period of
0 min. The mixture was refluxed in a heating mantle for 3 h, the
by simple recrystallization[
2a]
and as preparation of op-
tically active warfarin by classical resolution is very cum-
reaction flask cooled in an ice/water bath, and its contents were
carefully poured onto crushed ice (1250 mL) in a 3000 mL beaker.
[
27]
bersome.
Optically active warfarin is therefore by now
one of the relatively few medicinal compounds for whose Saturated aqueous NH Cl (300 mL), EtOAc (300 mL) and water
4
preparation a catalytic asymmetric synthesis may actually (200 mL) were added. The mixture was stirred for 1 h and sepa-
be the preferred route, even on a large scale.
rated in a large separation funnel. The aqueous phase was extracted
with EtOAc (2ϫ300 mL), and the combined organic extracts were
4
washed with brine (200 mL), dried with anhydrous MgSO , filtered,
and the solvents were evaporated in vacuo to give a yellow solid.
Experimental Section
The solid was dissolved in boiling MeOH (500 mL). Crystallisation
at room temperature for 2 h gave a first crop of crystals, while sub-
sequent crystallisation in an ice/water bath for 2 h gave a second
crop. The two isolated crops were combined and dried at room
temperature to give (R)-diphenylphenylglycinol (45.55 g, 66%).
General: All commercially available reagents were used as received,
and all solvents were used without further purification. -tert-Leu-
cine and -cyclohexylglycine were generously donated by Evonik
Degussa GmbH. All other reagents were obtained from Sigma–
Aldrich or VWR. The heating mantles used in this work were of
the type Heat-On® from Radleys, either fluoropolymer-coated or
with anodized finish. Thin layer chromatography (TLC) was per-
formed on Merck silica gel 60 F254 TLC plates, either on alumin-
ium sheets or glass. They were visualized by UV light, or by immer-
High-purity product was obtained by recrystallisation of this prod-
uct from MeOH (500 mL) with crystallisation at room temperature
for 2 h and in an ice/water bath for 2 h, followed by isolation by
filtration and drying at room temperature to give (R)-diphenyl-
phenylglycinol as beautiful needle-shaped crystals. M.p. 131–
4 6 7 2 4 2
sion in a solution of either (NH ) Mo O24·4H O and Ce(SO ) ·
Eur. J. Org. Chem. 2009, 5185–5191
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5189

T. E. Kristensen, K. Vestli, F. K. Hansen, T. Hansen
FULL PAPER
1
(
5
32 °C (dec.). [α]20
300 MHz, [D ]DMSO): δ = 1.81 (s, 2 H, NH
.60 (s, 1 H, OH), 6.93–7.34 (m, 13 H, Ph-H), 7.71–7.75 (m, 2 H, concentration-dependant NMR spectra. The integrals in the NMR
D
= +240 (c = 1.030, CHCl
3
). 1H NMR
Warfarin exists in a rapid equilibrium with a pair of pseudo-dia-
6
2
), 4.99 (s, 1 H, 2-H), stereomeric hemiketals and therefore gives highly complicated and
13
Ph-H) ppm. C NMR (75 MHz, [D
1
1
6
]DMSO): δ = 61.1, 80.0,
25.5, 125.8, 126.1, 126.3, 126.6, 126.7, 127.1, 127.9, 129.0, 142.2,
45.7, 147.4 ppm. This is a known compound with spectroscopic
spectroscopic data below are therefore given relative to the species,
either the keto form or one of the two pseudo-diastereomeric hemi-
ketal forms, to which it belongs and not absolute relative to the
other species.
[14]
properties in accordance with those reported.
(
S)-Diphenylphenylglycinol (15): In a completely analogous pro-
cedure to the one above, (+)--α-phenylglycine (37.24 g, 246 mmol)
and SOCl (27.0 mL, 372 mmol) in MeOH (350 mL) gave -phenyl-
2
glycine methyl ester hydrochloride (48.35 g, 97%) as colourless, fi-
brous crystals. A portion of this ester hydrochloride (48.17 g,
2
39 mmol) was treated with the Grignard reagent prepared from
Mg (35.53 g, 1462 mmol) and bromobenzene (229.94 g,
464 mmol), by using the same amounts of Et O as above, in the
1H NMR (300 MHz, CDCl
H, CH ketal), 1.92–2.03 (m, 1 H, CH
keto), 2.33–2.55 (m, 3 H, CH
.6 Hz, 1 H, CH
and 10.0 Hz, 1 H, CH
3
3
): δ = 1.66 (s, 3 H, CH ketal), 1.67 (s,
1
2
3
3
2 3
ketal), 2.27 (s, 3 H, CH
same manner to give (S)-diphenylphenylglycinol (15; 43.18 g, 62%)
after recrystallisation from MeOH (500 mL). Very pure material
of needle-shaped crystals was likewise obtained by an additional
recrystallisation from MeOH (500 mL), as was done for the other
ketal), 3.31 (dd, 3J = 19.3 and
2
3
2
2
keto), 3.43 (s, 1 H, OH ketal), 3.83 (dd, J = 19.3
3
2
keto), 3.89 (s, 1 H, OH ketal), 4.16 (dd, J
3
20
= 11.3 and 6.9 Hz, 1 H, CH ketal), 4.25 (dd, J = 6.9 and 3.7 Hz,
enantiomer. M.p. 132–133 °C (dec.), [α]
D
= –241 (c = 1.036,
]DMSO): δ = 1.81 (s, 2 H, NH ),
.99 (s, 1 H, 2-H), 5.60 (s, 1 H, OH), 6.93–7.34 (m, 13 H, Ph-H),
3
1
1 H, CH ketal), 4.71 (dd, J = 10.0 and 2.6 Hz, 1 H, CH keto),
CHCl
3
). H NMR (300 MHz, [D
6
2
7.15–7.35 (m, 21 H, 7 H Ar-H ketal, 7 H Ar-H ketal, 7 H Ar-H
4
7
.71–7.75 (m, 2 H, Ph-H) ppm. 1 C NMR (75 MHz, [D
3
keto), 7.42–7.60 (m, 3 H, 1 H Ar-H ketal, 1 H Ar-H ketal, 1 H Ar-
6
]DMSO):
3
H keto), 7.79 (dd, J = 8.0 and 1.4 Hz, 1 H, Ar-H ketal), 7.89 (dd,
δ = 61.1, 80.0, 125.5, 125.8, 126.1, 126.3, 126.6, 126.7, 127.1, 127.9,
3
3
J = 8.0 and 1.4 Hz, 1 H, Ar-H ketal), 7.93 (dd, J = 8.0 and
1
29.0, 142.2, 145.7, 147.4 ppm. This is a known compound with
1
3
[14]
1.4 Hz, 1 H, Ar-H keto), 9.52 (s, 1 H, OH keto) ppm. C NMR
75 MHz, CDCl ): δ = 27.5, 28.0, 30.0, 34.3, 34.8, 35.3, 40.0, 42.6,
5.1, 99.0, 100.5, 101.2, 104.1, 107.8, 115.5, 115.9, 116.1, 116.4,
spectroscopic properties in accordance with those reported.
(
3
(
(
S)-Diphenylphenylglycinol Trimethylsilyl Ether (20): A portion of
S)-diphenylphenylglycinol (10.096 g, 34.9 mmol) was dissolved in
Cl (120 mL), and iodine (123 mg, 0.48 mmol) and hexa-
4
1
1
1
2
16.6, 122.7, 123.0, 123.5, 123.8, 123.9, 126.4, 126.6, 126.9, 127.0,
27.1, 127.9, 128.1, 128.5, 129.1, 131.4, 131.7, 131.9, 139.6, 141.5,
43.2, 152.7, 152.8, 152.9, 158.9, 159.7, 161.0, 161.4, 162.1, 162.2,
CH
2
2
methyldisilazane (5.639 g, 34.9 mmol) were added. The reaction
mixture was stirred at room temperature for 4 h, and MeOH
12.5 ppm. This is a known compound with spectroscopic proper-
(
3.0 mL) was added to quench residual HMDS. The mixture was
stirred for 5 min, and the volatiles were removed by evaporation in
vacuo. The residual oil was redissolved in CH Cl (120 mL), a solu-
tion of Na ·5H O (10.405 g) in H O (120 mL) was added, and
the mixture was stirred for 5 min and separated. The organic phase
was dried with anhydrous MgSO , filtered through a short pad of
Celite® 535 by using a Büchner funnel, and the filter cake was
washed with CH Cl (50 mL). The filtrate was concentrated in
vacuo to give (S)-diphenylphenylglycinol trimethylsilyl ether (20) in
[2]
ties in accordance with those reported.
Supporting Information (see footnote on the first page of this arti-
2
2
1
13
cle): H and C NMR spectra for compound 20.
2
S
2
O
3
2
2
4
Acknowledgments
2
2
We would like to acknowledge Evonik Degussa GmbH, who gener-
ously donated samples of -tert-leucine and -cyclohexylglycine for
this work.
2
0
quantitative yield. Light yellow, honey-like oil. [α]
D
= –11.1 (c =
): δ = –0.15 (s, 9 H,
), 4.95 (s, 1 H, 2-H), 6.79–6.85 (m, 2 H,
Ph-H), 7.02–7.14 (m, 3 H, Ph-H), 7.23–7.33 (m, 8 H, Ph-H), 7.35–
1
3 3
1.198, CHCl ). H NMR (300 MHz, CDCl
SiMe ), 1.69 (s, 2 H, NH
3
2
[
1] For example, see: T. Meinertz, W. Kasper, C. Kahl, E.
Jähnchen, Br. J. Clin. Pharmacol. 1978, 5, 187–188 and refer-
ences cited therein.
13
7
6
1
2
3
.43 (m, 2 H, Ph-H) ppm. C NMR (75 MHz, CDCl ): δ = 1.8,
2.6, 84.5, 126.5, 126.6, 126.6, 127.4, 127.7, 128.9, 129.1, 129.2,
41.5, 142.3, 143.7 ppm. IR (film): ν˜ = 3390, 3087, 3060, 3032,
955, 2897, 1602, 1494, 1447, 1251 cm . HRMS (ESI): calcd. for
27NOSi + H] 362.1940; found 362.1949.
[
2] a) N. Halland, T. Hansen, K. A. Jørgensen, Angew. Chem. Int.
Ed. 2003, 42, 4955–4957; b) H. Kim, C. Yen, P. Preston, J.
Chin, Org. Lett. 2006, 8, 5239–5242; c) J.-W. Xie, L. Yue, W.
Chen, W. Du, J. Zhu, J.-G. Deng, Y.-C. Chen, Org. Lett. 2007,
–1
[C
23
H
9, 413–415; for a review of organocatalytic synthesis of drugs
General Procedure for the Preparation of Warfarin (3): A small vial
was charged with 4-hydroxycoumarin (1, 81.1 mg, 0.50 mmol),
benzylideneacetone (2, 87.7 mg, 0.60 mmol), additive (if appropri-
ate) and the appropriate catalyst (0.10 mmol, 20 mol-%). A small
magnetic stirring bar and THF (1 mL) were added, and the reac-
tion mixture was stirred at room temperature for 24 h. The volatiles
were removed by flushing with pressurized air. The residue was
purified by column chromatography on silica gel with 15% EtOAc
in hexanes to give pure warfarin (3) as a white solid. The enantio-
meric excess was determined by HPLC analysis of the purified
product with an AD-H chiral column (20% iPrOH in isohexane,
and bioactive natural products, see: d) R. M. de Figueiredo, M.
Christmann, Eur. J. Org. Chem. 2007, 2575–2600.
[3] a) V. Shah, J. Bose, R. Shah, J. Org. Chem. 1960, 25, 677–678;
b) J.-C. Jung, J.-C. Kim, O.-S. Park, Synth. Commun. 1999, 29,
3587–3595.
[
4] For reviews, see: a) F. Peng, Z. Shao, J. Mol. Catal. A 2008,
2
85, 1–13; b) G. Bartoli, P. Melchiorre, Synlett 2008, 1759–
1772; c) Y.-C. Chen, Synlett 2008, 1919–1930; d) L.-W. Xu, Y.
Lu, Org. Biomol. Chem. 2008, 6, 2047–2053; e) L.-W. Xu, J.
Luo, Y. Lu, Chem. Commun. 2009, 1807–1821; for combina-
tions of primary amines with additional functionalities, see: f)
F. Xue, S. Zhang, W. Duan, W. Wang, Adv. Synth. Catal. 2008,
350, 2194–2198; g) C. G. Kokotos, G. Kokotos, Adv. Synth. Ca-
tal. 2009, 351, 1355–1362 and references cited therein.
R
1.0 mL/min, λ = 254 nm, minor enantiomer t = 6.5 min and major
enantiomer t = 11.6 min).
R
5190
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 5185–5191

New Organocatalysts for the Preparation of Optically Active Warfarin
[
[
5] L. Li, L.-W. Xu, Y.-D. Ju, G.-Q. Lai, Synth. Commun. 2009, [16] From tert-leucine, see: W. Behnen, C. Dauelsberg, S.
39, 764–774.
Wallbaum, J. Martens, Synth. Commun. 1992, 22, 2143–2153.
17] A. S. Bommarius, M. Schwarm, K. Stingl, M. Kottenhahn, K.
Huthmacher, K. Drauz, Tetrahedron: Asymmetry 1995, 6,
6] J.-W. Xie, W. Chen, R. Lui, M. Zeng, W. Du, L. Yue, Y.-C.
[
Chen, Y. Wu, J. Zhu, J.-G. Deng, Angew. Chem. Int. Ed. 2007,
46, 389–392.
2851–2888.
[
[
[
[
7] B. Tan, X. Zhang, P. J. Chua, G. Zhong, Chem. Commun. 2009,
[
18] a) M. A. Wegman, M. H. A. Janssen, F. van Rantwijk, R. A.
Sheldon, Adv. Synth. Catal. 2001, 343, 559–576; b) A. Brug-
gink, E. C. Roos, E. de Vroom, Org. Process Res. Dev. 1998, 2,
7
79–781.
8] For example, see: R. R. Fenton, R. S. Vagg, P. A. Williams, In-
org. Chim. Acta 1988, 148, 37–44 and references cited therein.
9] K. Saigo, N. Kubota, S. Takebayashi, M. Hasegawa, Bull.
Chem. Soc. Jpn. 1986, 59, 931–932.
128–133.
[
19] R. Yoshioka, H. Hiramatsu, K. Okamura, I. Tsujioka, S.-i. Ya-
mada, J. Chem. Soc. Perkin Trans. 2 2000, 2121–2128.
10] a) D. C. Braddock, J. M. Redmond, S. A. Hermitage, A. J. P.
White, Adv. Synth. Catal. 2006, 348, 911–916 and references
cited therein; b) D. C. Braddock, S. A. Hermitage, J. M. Red-
mond, A. J. P. White, Tetrahedron: Asymmetry 2006, 17, 2935–
[20] A. W. Ingersoll, Org. Synth. 1937, 17, 80–83.
[21]
/I , see: M. J. McKennon, A. I. Meyers, K. Drauz,
4 2
For NaBH
M. Schwarm, J. Org. Chem. 1993, 58, 3568–3571.
22] For NaBH /H SO , see: A. Abiko, S. Masamune, Tetrahedron
Lett. 1992, 33, 5517–5518.
23] For other NaBH systems, see: a) J. Zhang, X. Gao, C. Zhang,
2
937.
11] a) H. Brunner, J. Bügler, B. Nuber, Tetrahedron: Asymmetry
995, 6, 1699–1702; b) B. Vakulya, S. Varga, A. Csámpai, T.
[
[
4
2
4
[
1
4
Soós, Org. Lett. 2005, 7, 1967–1969; c) W. He, P. Liu, B. L.
Zhang, X. L. Sun, S. Y. Zhang, Appl. Organomet. Chem. 2006,
J. Ma, D. Zhao, Synth. Commun. 2009, 39, 1640–1654; b) A. S.
Demir, I. M. Akhmedov, O. Sesenoglu, Turk. J. Chem. 1999,
2
0, 328–334.
12] I. C. Ivanov, I. Manolov, L. A. Alexandrova, Arch. Pharm.
990, 323, 521–522.
23, 123–126 and references cited in these references.
[
[
[
[24] B. Karimi, B. Golshani, J. Org. Chem. 2000, 65, 7228–7230.
Our workup is a modification of the one disclosed in this paper.
[25] C.-S. Da, L.-P. Che, Q.-P. Guo, F.-C. Wu, X. Ma, Y.-N. Jia, J.
Org. Chem. 2009, 74, 2541–2546.
[
[
1
13] From phenylalanine and leucine, see for example: Z. Zhou, Y.
Guo, Synth. Commun. 2008, 38, 684–696.
14] From phenylglycine, see for example: a) J. Bach, R. Berenguer,
J. Garcia, T. Loscertales, J. Vilarrasa, J. Org. Chem. 1996, 61,
26] S. M. Kaye in Encyclopedia of Explosives and Related Items,
Picatinny Arsenal, New Jersey, 1978, vol. 8, pp.230–232.
27] B. D. West, S. Preis, C. H. Schroeder, K. P. Link, J. Am. Chem.
Soc. 1961, 83, 2676–2679.
9
021–9025; b) R. Fleischer, H. Wunderlich, M. Braun, Eur. J.
Org. Chem. 1998, 1063–1070; c) D. R. Dragoli, M. T. Burdett,
J. A. Ellman, J. Am. Chem. Soc. 2001, 123, 10127–10128.
15] From cyclohexylglycine, see: R. W. Draper (Inventor), Scher-
ing-Plough Corporation, WO 02/22623 A1, 2002.
Received: June 17, 2009
Published Online: August 19, 2009
[
Eur. J. Org. Chem. 2009, 5185–5191
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5191