Gram-scale production and applications of optically pure 13C-labelled (+)-catechin and (-)-epicatechin

Article abstract of DOI:10.1002/1099-0690(200106)2001:12<2379::AID-EJOC2379>3.0.CO;2-D

In this paper gram-scale asymmetric total syntheses of pure (+)-4-[¹³C]catechin (1) and (-)-4-[¹³C]epicatechin (2) are described. Labelling was introduced through acylation of a phloroglucinol derivative 4 by 1-[¹³C]acetic acid, after activation by TFAA. Condensation of the resulting C₆-C₂ acetophenone building block 6 with a C₁-C6 benzaldehyde unit 9 provided the C₆-C₃-C₆ flavonoid skeleton, with benzyl ethers as phenol protecting groups. Asymmetry was introduced by an inexpensive and very efficient resolution process, performed at the penultimate step towards naturally occurring flavan-3-ols, by using tartaric acid derivatives. Both enantiomers of 4-[¹³C]catechin were obtained with a high level of enantiomeric purity, especially (+)-1 of the natural series. The other enantiomer (-)-1 was a valuable precursor of natural (-)-4-[¹³C]epicatechin by epimerisation at C-2. These two natural flavanols are thus obtained for the first time in large quantities in optically pure form as labelled compounds and will be useful tools for biological studies using NMR or isotopic mass spectrometry in forthcoming experiments.

Full text of DOI:10.1002/1099-0690(200106)2001:12<2379::AID-EJOC2379>3.0.CO;2-D

FULL PAPER
13
Gram-Scale Production and Applications of Optically Pure C-Labelled
(؉)-Catechin and (؊)-Epicatechin[‡]
[a]
[a]
[a]
´
Bastien Nay, Valerie Arnaudinaud, and Joseph Vercauteren*
Keywords: Asymmetric synthesis / Natural products / Isotopic labelling / Total synthesis
In this paper gram-scale asymmetric total syntheses of pure
(+)-4-[13C]catechin (1) and (−)-4-[13C]epicatechin (2) are de-
scribed. Labelling was introduced through acylation of a
curring flavan-3-ols, by using tartaric acid derivatives. Both
enantiomers of 4-[13C]catechin were obtained with a high
level of enantiomeric purity, especially (+)-1 of the natural
phloroglucinol derivative 4 by 1-[13C]acetic acid, after activa- series. The other enantiomer (−)-1 was a valuable precursor
tion by TFAA. Condensation of the resulting C6−C2 aceto- of natural (−)-4-[13C]epicatechin by epimerisation at C-2.
phenone building block 6 with a C1−C6 benzaldehyde unit 9
provided the C6−C3−C6 flavonoid skeleton, with benzyl time in large quantities in optically pure form as labelled
ethers as phenol protecting groups. Asymmetry was intro- compounds and will be useful tools for biological studies
duced by an inexpensive and very efficient resolution pro- using NMR or isotopic mass spectrometry in forthcoming
These two natural flavanols are thus obtained for the first
cess, performed at the penultimate step towards naturally oc-
experiments.
Introduction
at C-2 to (Ϫ)-2 in 50% yield. Therefore, these labelled com-
pounds, obtained in usable amounts and for the first time
in enantiomerically pure native forms (free phenolic
groups), will definitely allow us to assess their fate in
humans.
As we have already presented in previous papers,[1,2] it
would be very useful to dispose of 13C-labelled catechin and
epicatechin (1 and 2, Scheme 1) to assess the question of
their resorption and to increase the knowledge of their
metabolism in humans. Introduction of 13C labelling re-
quires the total synthesis of the skeleton, as described in
earlier syntheses (monomer[1] and dimer B3[3]). In addition,
the problems posed by the stereochemistry at C-2 and C-3
must be overcome,[4] as has been achieved, for example, by
chemical resolution of unlabelled racemic catechin with an
-tartrate derivative.[2]
Results and Discussion
Choice of Synthetic Strategy
Flavonoid synthesis refers to two possible routes. The
first route consists of coupling a C6 phenolic unit to a
C3ϪC6 cinnamic moiety to form the C6ϪC3ϪC6 skeleton.
We recently applied this to the synthesis of racemic 4-
[13C]catechin (40 mg).[1] The more efficient and convenient
second strategy was applicable on a larger scale, consisting
in the crotonization reaction between a C6ϪC2 aceto-
phenone and a C1ϪC6 benzaldehyde to form a chalcone.
This formally allowed us to synthesize 13C-labelled (Ϫ)-pro-
cyanidin B3, a dimer of catechin (30 mg).[3] We applied it
herein to the production of gram amounts of catechins.
Scheme 1
This paper describes the first total synthesis of 13C-la-
belled and optically pure natural (ϩ)-catechin (1) simultan-
eously. The total synthesis of the biologically interesting
(Ϫ)-epicatechin (2) incorporating 13C labelling, could not
be obtained in an optically pure form through ent-catechin,
as expected from catechin -tartrate.[2] We also report
herein, the first total synthesis of 13C-labelled (Ϫ)-epicate-
chin (2) with 99% ee thanks to the -tartrate derivatives.
After saponification, the pure (Ϫ)-catechin was epimerized
Phenol Protections
Benzyl protecting groups are commonly used in phenol
chemistry. A difficulty that often arises during benzylation
of nucleophilic polyphenols such as phloroglucinol is the
undesirable C-alkylation in addition to O-alkylation. As
phloroglucinol was our starting material, this problem was
avoided by using the method of Kawamoto et al.,[5] through
phloroglucinol triacetate (Scheme 2): 40 g of dried phlorog-
lucinol (3) was acetylated before benzylation, giving 106 g
of tri-O-benzylphloroglucinol (4) (87%, 2 steps). Addition-
[‡]
Total Synthesis of Isotopically Labelled Flavonoids, 4. Ϫ
Part 3: Ref.[3]
[a]
Laboratoire de Pharmacognosie (GESNIT, E. A. No. 491),
´
´
Faculte de Pharmacie, Universite Victor Segalen Bordeaux 2
´
146 Rue Leo Saignat, 33076 Bordeaux Cedex, France
Fax: (internat.) ϩ 33-5/56960975
E-mail: Joseph.Vercauteren@gnosie.u-bordeaux2.fr
Eur. J. Org. Chem. 2001, 2379Ϫ2384
WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001
1434Ϫ193X/01/0612Ϫ2379 $ 17.50ϩ.50/0
2379

B. Nay, V. Arnaudinaud, J. Vercauteren
FULL PAPER
ally, 58 g of 3,4-bis(benzyloxy)benzaldehyde (9) was pro-
duced from 30 g of 3,4-dihydroxybenzaldehyde (8) (83%).
Scheme 3. Resolution process of compound 13
Resolution of 13 and Subsequent Synthesis of Labelled (؉)-
and (؊)-Catechin
Recently, we described that coupling 13 (rac) to a derivat-
ive of -tartaric acid provided an efficient method of syn-
thesizing optically pure native catechins.[2] Esterification
(Scheme 3) of the free 3-hydroxy group of 13 with di-
benzoyltartaric acid monomethyl ester (L-14) in the pres-
Scheme 2. Total synthesis of labelled racemic compound 13
ence of DCC and DMAP gave a mixture of diastereomers
15 and 16 (35 g, 92%) which were inseparable by preparat-
ive HPLC. One, (؊)-15 from the (ϩ)-catechin series, crystal-
lized slowly but efficiently in hexane/dichloromethane (3:1)
in a diastereomerically pure form (14.6 g after recrystalliza-
tion in the same solvent, de Ͼ 99%). The other, (؊)-16 from
the nonnatural (Ϫ)-catechin series (20 g, de ϭ 88%), re-
mained in solution, making the separation successful. A di-
astereomerically pure product from this last series was ob-
tained by hydrolysis of (؊)-16 to (؊)-13 (de ϭ 88%) fol-
Synthesis of Labelled Phloroacetophenone Derivatives
Tri-O-benzylphloroacetophenone (5), labelled at the car-
bonyl group, was synthesized by acylating 4 (117 g) with 1-
[13C]acetic acid (99% 13C enrichment) in the presence of
trifluoroacetic anhydride (Scheme 2),[3,6,7] giving 75 g of
pure 5 after silica gel column chromatography (58% from
4). Selective deprotection of 5 was achieved by the action
of TiCl4, yielding 47 g of pure 6 (80%) after purification
from the C-benzylated by-product 7 (7%).
lowed by esterification by D-14, yielding 11.5 g of pure (؉)-
15 after column chromatography and crystallization (de Ͼ
99%).
Employing this method, we could prepare (؊)-15 and
(؉)-15 separately in optically pure forms. Both enantiomers
were subjected to hydrolysis and phenol deprotection
(Scheme 4 and Scheme 5) to give the two optically pure (ee
Ͼ 99%, checked by HPLC analysis on β-cyclodextrin bon-
ded column) 13C-labelled catechins 1 [3.60 g of (؉)-1 from
Synthesis of Labelled Intermediates Chalcone 10 and
Racemic Catechin 13
Crotonization of 6 (47 g) and 9 (45 g) led to chalcone 10
(75 g) in 86% yield after crystallization (Scheme 2). Reduc-
tion of 10 with NaBH4 and cyclization in the presence of
BF3·OEt2 (ClarkϪLewis method)[8,9] led to the labile 2H-
chromene 11. These steps were particularly delicate, prob-
ably due to the ability of 11 to form an allylic cation[10] as
well as a stabilized methylenequinone by Claisen rearrange-
ment.[11] No purification was carried out on 11 which was
immediately oxidized by an osmium-catalyzed diastereose-
lective dihydroxylation, providing 31 g of racemic 4-
[13C]flavan-3,4-diol 12 from 74 g of 10 (41% yield after crys-
tallization). Deoxygenation at C-4 was carried out with
NaBH3CN,[1,12] giving 24.3 g of racemic 4-[13C]catechin
tetrabenzyl ether 13 (81% yield).
Scheme 4. Synthesis of optically pure labelled (ϩ)-catechin (1)
Numerous attempts were made to make the synthesis of
13 asymmetric. However, stereoselective reduction of 10
was not possible, only a racemic mixture of 11 was observed
in all cases {NMR analysis in the presence of Pr(hfc)3 in
[D6]benzene}. Moreover, neither epoxidation nor enantiose-
lective dihydroxylation of 10 and 11 were successful. There-
fore, we finally turned to chemical resolution of 13.
Scheme 5. Synthesis of labelled (Ϫ)-epicatechin (2)
Eur. J. Org. Chem. 2001, 2379Ϫ2384
2380

Optically Pure 13C-Labelled (ϩ)-Catechin and (Ϫ)-Epicatechin
(؊)-15 and 2.91 g of (؊)-1 from (؉)-15]. Figure 1a shows
FULL PAPER
Conclusion
1
the 4-H section of the H NMR spectrum of (؉)-1. It dis-
The total synthesis of gram amounts of 13C-labelled nat-
ural (ϩ)-catechin [(؉)-1] and (Ϫ)-epicatechin [(؊)-2] has
been completed in 6.2% overall yield (as asymmetric flavan-
3-ols) from phloroglucinol (3). Therefore, we significantly
improved the yields in comparison with the other strategy
(C3 ϩ C3ϪC6)[1] and furthermore, this route is asymmetric.
Indeed, an efficient resolution of racemic benzylated cate-
chin was used as the key step, allowing us to obtain the two
antipodal series in high optical purity. In particular, the
(Ϫ)-epicatechin series was synthesized for the first time in
its native labelled form. Nonracemic compounds are abso-
lutely necessary to carry out future biological studies, in-
cluding pharmacokinetic studies in humans, 13C being very
useful in this work as an isotopic tracer.
plays the large coupling constants of 130 Hz within the ddd
observed at δ ϭ 2.50 and 2.85 for each proton directly
bound to labelled C-4.
Moreover, the gram-scale synthesis of 13C-labelled procy-
anidin dimer B3 is underway. Therefore, after a preliminary
study showing that racemic flavan-3,4-diol 12 cannot be
used in such a large-scale synthesis,[3] it is now conceivable
that we could oxidize labelled (؉)-13 at C-4, as a source of
optically pure diol 12, the precursor of procyanidolic oligo-
mers.[10,16]
Experimental Section
General Remarks: CH313COOH was purchased from Euriso-top
(Gif-sur-Yvette, France), with a 99% 13C enrichment. All other
commercial materials were used without further purification. All
reactions were performed under nitrogen and monitored by TLC.
Compounds 4,[5] 9 (reaction was performed at room temp., 83%
yield),[17]
L
-14 and
D
-14[2] were prepared according to the literature.
1
Crystallizations were performed as often as possible. Ϫ All H and
13C NMR spectra were recorded with a Bruker AMX-500 spectro-
meter at 500.13 and 125.73 MHz, respectively (proton decoupling
mode for carbon). Ϫ 1H NMR spectra were referenced to the signal
at δ ϭ 7.27 of residual CHCl3 or to CHD2OD at δ ϭ 3.31. Ϫ 13C
NMR spectra were referenced to signals of CDCl3 (δ ϭ 77.0) or of
CD3OD (δ ϭ 49.1) according to the solvent used. Ϫ UV/Vis spec-
tra were recorded using a Hitachi U2000 spectrometer. Ϫ FT-IR
spectra were recorded with a Bomem MB100 spectrometer. Ϫ Low-
and high-resolution mass spectra were obtained with Finnigan
MAT TSQ 700 and Micromass (UK) ZAB2-SEQ spectrometers,
respectively. Ϫ Elemental analyses were measured by the Service
Central d’Analyses du CNRS (69390 Vernaison, France). Ϫ Op-
tical rotations were measured with a PerkinϪElmer 141 polari-
meter equipped with a sodium lamp (589 nm). Ϫ Analytical TLC
was performed on Merck silica gel 60 F254 plates, and column
1
Figure 1. Expansion of H NMR spectra (500 MHz) showing sig-
nals of protons 4-Hα and 4-Hβ of (ϩ)-[4-13C]catechin (a) and (؊)-
[4-13C]epicatechin (b) with their broad characteristic coupling con-
1
stant J(4-H,13C-4)
Epimerisation of (؊)-4-[13C]Catechin into Natural (؊)-4-
[13C]Epicatechin
It is well known that catechin (1) can undergo epimeris-
ation at C-2 to form ent-epicatechin (2), through reversible
opening of ring C in basic medium.[13Ϫ15] We found more
straightforward practical conditions than those of Foo and
Porter[15] to synthesize (؊)-2 from (؊)-1, using 1% (w/v)
aq. Na3PO4 (pH ϭ 11.4) under nitrogen. This reaction led
to an equilibrium mixture of (؊)-1 and (Ϫ)-4-[13C]epicate-
chin [(؊)-2] in an approximate 3:1 ratio after 20 h at room
temperature (Scheme 5); (؊)-2 was purified by chromato-
graphy on Sephadex LH-20 and (؊)-1 could be recycled
sequentially four times to obtain 1.23 g of enantiomerically
pure (؊)-2 (ee Ͼ 99%, checked by HPLC on β-cyclodex-
trin), as well as 434 mg of recovered (؊)-1. Figure 1b shows
˚
chromatography using the indicated solvents on silica gel 60 A
70Ϫ200 µm (SDS, 13124 PEYPIN, France). Ϫ Melting points were
obtained with a Tottoli apparatus and are uncorrected.
[13CO]-1-[2,4,6-Tris(benzyloxy)phenyl]ethanone (5): CH313COOH
(21.9 mL, 379 mmol) and TFAA (64 mL, 454 mmol) were stirred
for 5 min at room temp. Phloroglucinol tribenzyl ether (4) (50 g,
126 mmol) was dissolved in CH2Cl2 (200 mL) and then added at 0
°C. After 1.5 h, the deep purple solution was diluted with AcOEt
and poured into saturated aq. NaHCO3 solution. The organic layer
was washed with brine, dried with Na2SO4 and concentrated. The
crude extract was purified by silica gel column chromatography,
1
a part of the H NMR spectrum of (؊)-2 with the large
coupling constants (126 and 132 Hz) of the two 4-H bound
to the labelled carbon atom.
Eur. J. Org. Chem. 2001, 2379Ϫ2384
2381

B. Nay, V. Arnaudinaud, J. Vercauteren
FULL PAPER
eluent CH2Cl2/cyclohexane (6:4, v/v), to give 30 g of 5 (55% yield
(CDCl3): δ ϭ 7.68 (dd, J ϭ 15.5, 6.2 Hz, β-H), 7.78 (dd, J ϭ 15.5,
from 4). 75 g of 5 was produced from 106 g of 4 (64%). Pale yellow 5.1 Hz, α-H). Ϫ 13C NMR (CDCl3): δ ϭ 106.6 (d, J ϭ 58 Hz, C-
˜
oil. Ϫ UV/Vis (MeOH): λmax ϭ 267 nm. Ϫ IR (thin film): ν ϭ
1Ј), 125.9 (d, J ϭ 55 Hz, C-α), 129.0 (d, J ϭ 6.1 Hz, C-1), 142.7
1
1694 cmϪ1 (CϭO). Ϫ H NMR (CDCl3): δ ϭ 2.50 (d, J ϭ 6.3 Hz, (C-β), 192.6 (labelled CO).
3 H, α-H), 5.01 (s, 2 H, 4-OCH2C6H5), 5.06 (s, 4 H, 2- and 6-
[4؊13C]-5,7-Dibenzyloxy-2-[3,4-bis(benzyloxy)phenyl]-2H-chromene
OCH2C6H5), 6.27 (d, J ϭ 1.0 Hz, 2 H, 3-H, 5-H), 7.36Ϫ7.41 (m,
15 H, 3 OCH2C6H5). Ϫ 13C NMR (CDCl3): δ ϭ 32.3 (d, J ϭ
47.0 Hz, C-α), 70.3 (4-OCH2C6H5), 70.7 (2- and 6-OCH2C6H5),
93.7 (C-3, C-5), 115.4 (d, J ϭ 55.0 Hz, C-1), 127.1, 127.4, 127.9,
128.1, 128.5, 128.6 (OCH2C6H5 ortho/meta/para), 136.4 (4-
OCH2C6H5 ipso), 136.5 (2- and 6-OCH2C6H5 ipso), 157.2 (C-2, C-
6), 161.1 (C-4), 201.2 (labelled CO). Ϫ HRMS (EI): m/z calcd. for
13C12C28H26O4 439.1859; found 439.1864.
(11): According to Kawamoto et al.,[9] compound 10 (20 g,
30.7 mmol) was dissolved in 1,2-dimethoxyethane (300 mL) at 85
°C and then NaBH4 (1.2 g, 31.5 mmol) was added. After 5 min at
85 °C, the mixture was cooled, diluted with ethyl acetate and
washed 3 times with brine. The organic layer was dried with sodium
sulfate before adding a solution of BF3·OEt2 (380 µL) in CH2Cl2
(5 mL). After 20 min of stirring at room temp., the solution was
washed 3 times with brine, dried with sodium sulfate and concen-
[13CO]-1-[2,4-Bis(benzyloxy)-6-hydroxyphenyl]ethanone (6): TiCl4 trated to give an orange resin (19.8 g) that was not purified due
(1  in CH2Cl2, 102 mL, 102 mmol) was added dropwise at 0 °C
to the instability of product 11. Ϫ Data as described previously.[1]
1
within 1 h to a solution of 5 (74.7 g, 170.1 mmol) in CH2Cl2 (1.1 ). Alterations due to labelling: Ϫ H NMR (CDCl3): δ ϭ 6.88 (ddd,
After 160 min of stirring, the mixture was slowly poured into aque- J ϭ 166, 9.9, 1.7 Hz, 4-H). Ϫ 13C NMR (CDCl3): δ ϭ 105.0 (d,
ous saturated NaHCO3, washed with water and brine, dried with
Na2SO4 and concentrated. The residue was crystallized from ether,
and the mother liquors were purified by column chromatography
on silica gel (eluent CH2Cl2/cyclohexane, 6:4), providing 47.1 g of
6 (80%) after elution of 5.1 g of by-product 7 (7%).
J ϭ 53 Hz, C-4a), 118.9 (labelled C-4), 119.7 (broad s, C-3). Ϫ MS
(EI, 70 eV): m/z (%) ϭ 633 [Mϩ] (40), 542 (30), 91 (100).
2,3-trans-3,4-cis-[4؊13C]-5,7-Dibenzyloxy-2-[3,4-bis(benzyloxy)-
phenyl]chroman-3,4-diol (12): According to Kawamoto et al.,[9]
a
solution of 11 (19.8 g) in THF (240 mL) was added to a mixture
6: White crystals. Ϫ M.p. 104 °C (ref.[9] 103 °C). Ϫ UV/Vis of N-methylmorpholine N-oxide (4.6 g, 34 mmol), water (12 mL),
Ϫ1
˜
(MeOH): λmax ϭ 287 nm. Ϫ IR (KBr): ν ϭ 1604 cm (CϭO). Ϫ THF (160 mL) and OsO4 (0.5 mmol, 6.2 mL of a 2.5 wt-% solution
1H NMR (CDCl3): δ ϭ 2.57 (d, J ϭ 6.1 Hz, 3 H, H-α), 5.06, 5.07 in 2-methyl-2-propanol, purchased from Aldrich (ref. 20,886-8).
(4 H, 4- and 6-OCH2C6H5), 6.12 (d, J ϭ 2.2 Hz, 1 H, 5-H), 6.19 After 6 h at room temp., the pale yellow precipitating mixture was
(d, J ϭ 2.2 Hz, 1 H, 3-H), 7.38Ϫ7.43 (m, 10 H, 2 OCH2C6H5), dissolved in CH2Cl2 (500 mL) and washed with an aqueous solu-
14.02 (s, 1 H, 2-OH). Ϫ 13C NMR (CDCl3): δ ϭ 33.2 (d, J ϭ tion of Na2S2O3 (10 g/100 mL), water and brine. Racemic com-
43.0 Hz, C-α), 70.2, 71.1 (4- and 6-OCH2C6H5), 92.3 (C-5), 94.8
pound 12 was crystallized from ether after dissolution in a min-
(C-3), 106.3 (d, J ϭ 57.0 Hz, C-1), 127.6, 127.9, 128.3, 128.4, 128.7, imum CH2Cl2. Yield: 10.7 g (52% from 10). We obtained 31 g of
128.7 (OCH2C6H5 ortho/meta/para), 135.6, 135.9 (OCH2C6H5 12 from 74 g of 10 (41% yield) in several runs via 11. Ϫ Data as
ipso), 162.0 (C-6), 165.1 (C-4), 167.5 (C-2), 203.1 (labelled CO). Ϫ described previously.[1,9] Ϫ M.p. 175 °C. Ϫ Alterations due to label-
1
MS (EI, 70 eV): m/z (%)
ϭ
349 (30) [M]ϩ, 91 (100).
Ϫ
ling: Ϫ H NMR (CDCl3): δ ϭ 5.09 (dd, J ϭ 152.0, 3.9 Hz, 4-H).
13C1C21H20O4 (349.4): calcd. C 75.91, H 5.77; found C 75.65, H
5.76.
Ϫ
13C NMR (CDCl3): δ ϭ 61.5 (labelled C-4), 70.2 (d, J ϭ 34 Hz,
C-3), 105.2 (d, J ϭ 48 Hz, C-4a). Ϫ HRMS (FABϩ, nitrobenzyl
alcohol): m/z calcd. for 13CC42H38O7Li [M ϩ Li] 674.2811; found
674.2868.
7: White crystals. Ϫ M.p. 117 °C. Ϫ UV/Vis (MeOH): λmax
ϭ
287 nm. Ϫ IR (thin film): ν˜ ϭ 1621 cmϪ1 (CϭO). Ϫ 1H NMR
(CDCl3): δ ϭ 2.60 (d, J ϭ 6.1 Hz, 3 H, α-H), 4.05 (s, 2 H, 3-
(؎)-[4؊13C]-5,7-Dibenzyloxy-2-[3,4-bis(benzyloxy)phenyl]chroman-
3-ol (13): According to Brown and Fuller,[12] NaBH3CN (44 g,
698 mmol) was carefully added to a suspension of 12 (31 g) in
CH2C6H5), 5.08 (s,
2
H, 6-OCH2C6H5), 5.12 (s,
2 H, 4-
OCH2C6H5), 6.11 (s,
1
H, 5-H), 7.25Ϫ7.43 (m, 15 H,
3
OCH2C6H5), 14.11 (s, 1 H, 2-OH). Ϫ 13C NMR (CDCl3): δ ϭ 28.1 acetic acid (1.5 ). After 2 d, the acetic acid was evaporated and
(3-CH2C6H5), 33.3 (d, J ϭ 43.0 Hz, C-α), 70.1 (4-OCH2C6H5), 71.0 the residue was dissolved in CH2Cl2 before washing with saturated
(6-OCH2C6H5), 88.7 (C-5), 106.4 (d, J ϭ 56.7 Hz, C-1), 110.2 (C- aqueous NaHCO3, water, and brine. The crude extract containing
3), 125.4, 127.2, 127.8, 128.0, 128.1, 128.4, 128.6, 128.7, 128.8 13 and residual 12 was purified by centrifugal chromatography on
(CH2C6H5 ortho/meta/para), 135.8 (6-OCH2C6H5 ipso), 136.2 (4-
OCH2C6H5 ipso), 141.6 (3-CH2C6H5 ipso), 161.0 (C-6), 162.4 (C- 19.55 g of racemic 13. The remaining 12 (7.93 g) was recycled and
4), 164.0 (C-2), 203.3 (labelled CO). Ϫ MS (EI, 70 eV): m/z (%) ϭ resubmitted to reduction, giving 6.52 g of 13 after purification. As
439 (10) [M]ϩ, 91 (100). Ϫ 13C1C28H26O4 (439.5): calcd. C 79.48, some impurities remained in 13 (26.1 g in total), it was purified
silica gel (Merck, ref. 1.07749), eluent CH2Cl2/AcOEt (97:3), giving
H 5.96; found C 79.11, H 5.87.
once more by silica gel column chromatography (eluent CH2Cl2)
to give 24.4 g of 13 (80% yield). Data as described previously.[1,10,18]
[
13CO]-1-[2,4-Bis(benzyloxy)-6-hydroxyphenyl]-3-[3,4-bis(benzyl-
1
Alterations due to labelling: Ϫ H NMR (CDCl3): δ ϭ 2.68 (ddd,
oxy)phenyl]propenone (10): To a stirred solution of 6 (47 g,
135 mmol) in DMF (800 mL), NaH (60%), dispersed in mineral oil
(22 g, 540 mmol) was added and then 9 (45 g, 142 mmol) in DMF
(200 mL) was added dropwise over a period of 15 min at 0 °C.
After stirring for 2 h at room temp., the reaction was carefully
quenched with water, DMF was evaporated under vacuum, and the
J ϭ 130.5, 16.5, 8.8 Hz, axial 4β-H), 3.13 (ddd, J ϭ 133.2, 16.5,
5.6 Hz, equatorial 4α-H). Ϫ 13C NMR (CDCl3): δ ϭ 27.6 (labelled
C-4), 68.1 (d, J ϭ 36 Hz, C-3), 102.3 (d, J ϭ 44 Hz, C-4a).
Resolution of (؎)-13: A solution of (؎)-13 (24.3 g, 37.4 mmol) in
CH2Cl2 (400 mL) was refluxed for 2.5 h in the presence of the -
residue was dissolved in dichloromethane (800 mL) before washing tartaric acid derivative
L-14 (27.8 g, 74.8 mmol), DCC (15.4 g,
with water and brine. The extract was concentrated and the residue
74.8 mmol), and 4-DMAP (228 mg, 1.87 mmol). After cooling and
filtering the mixture, the CH2Cl2 layer was washed with water,
was crystallized from ether to give 10 as yellow crystals (75 g, 86%
yield). Ϫ M.p. 137Ϫ138 °C (ref.[9] 140 °C). Ϫ Spectroscopic data brine, and dried with Na2SO4. The crude extract was purified by
1
as described previously.[1] Alterations due to labelling: Ϫ H NMR
silica gel column chromatography (eluent CH2Cl2/hexane, 9:1), fur-
2382
Eur. J. Org. Chem. 2001, 2379Ϫ2384

Optically Pure 13C-Labelled (ϩ)-Catechin and (Ϫ)-Epicatechin
FULL PAPER
nishing 35 g of a mixture of inseparable (؊)-15 and (؊)-16. (؊)-15 para and OCOC6H5 ortho/meta/para), 128.6 and 128.7 (2 OC-
was crystallized from hexane/CH2Cl2 (3:1), giving 14.6 g of white OC6H5 ipso), 130.5 (C-1Ј), 136.8 (5-OCH2C6H5 ipso), 136.9 (7-
crystals after one recrystallization from the same solvent (83% of
expected, de Ͼ 99% based on silica gel HPLC, eluent hexane/
CH2Cl2, 85:15, 1 mL/min), and 20 g of diastereomerically enriched
OCH2C6H5 ipso), 137.1 and 137.3 (3Ј- and 4Ј-OCH2C6H5 ipso),
148.9 and 149.0 (C-3Ј and C-4Ј), 154.4 (C-8a), 157.5 (C-5), 159.0
(C-7), 164.8 (3ЈЈ-OCOC6H5), 165.1 (2ЈЈ-OCOC6H5), 165.3 (C-1ЈЈ),
(؊)-16 (de ϭ 88%). Hydrolysis of (؊)-15 in MeOH/H2O/KOH 166.3 (C-4ЈЈ). Ϫ MS (FABϩ, nitrobenzyl alcohol): m/z (%) ϭ 1006
(270 mL/30 mL/3 g) led to a precipitate of enantiomerically pure [MHϩ] (75), 634 (100), 542 (70). Ϫ 13C1C61H52O13 (1006.1): calcd.
(؉)-13 that was extracted with CH2Cl2 and washed with water and
brine and then concentrated. (؉)-13 was used without further puri-
fication (97% yield for hydrolysis, ee Ͼ 99%). The same procedure
for (؊)-16 led to enantiomerically enriched (؊)-13 (12.1 g, 93%
yield, ee ϭ 88%) that was esterified with -tartaric acid derivative
C 74.12, H 5.21; found C 73.78, H 5.29.
(؉)-15: [α]2D0 ϭ ϩ35 (c ϭ 1, CH2Cl2). Ϫ M.p. 150 °C. Ϫ Same
spectroscopic data as (؊)-15. Ϫ 13C1C61H52O13 (1006.1): calcd. C
74.12, H 5.21; found C 74.22, H 5.20.
D
-14 (13.9 g, 37.3 mmol) by the above-mentioned method to give
(؉)-13: [α]2D0 ϭ ϩ1.5 (c ϭ 1, CH2Cl2; ref.[18] ca. 0, c ϭ 3.3, CHCl3).
Data as previously described [cf. compound (؎)-13].
(؉)-15 after crystallization in an optically pure form (11.5 g, 70%
of expected, de Ͼ 99%) as for (؊)-15. Hydrolysis of (؉)-15 gave
optically pure (؊)-13 (7.1 g, 96% yield, ee Ͼ 99%).
(؊)-13: [α]2D0 ϭ Ϫ1.5 (c ϭ 1, CH2Cl2). Data as previously described
[cf. compound (؎)-13].
(؊)-15: [α]2D0 ϭ Ϫ35 (c ϭ 1, CH2Cl2). Ϫ M.p. 150 °C. Ϫ UV/Vis
(؉)- and (؊)-[4؊13C]Catechin (1): A suspension of (؉)-13 (7.9 g)
in MeOH (120 mL) was submitted to hydrogenolysis at room temp.
in the presence of palladium (790 mg) on activated carbon (Ald-
rich, ref. 20,569-9), under hydrogen. After 6 h, the methanol was
evaporated and the residue was dissolved in ethyl acetate before
filtration through Celite and concentration. The crude extract
(4.15 g) was purified by chromatography on Sephadex LH-20
(water/ethanol, 85:15, 5 mL/min), yielding 3.60 g of pure (؉)-1
(90% yield). The same procedure was applied to produce 2.91 g of
(؊)-1 from 7.12 g of (؊)-13 (92% yield).
˜
(CH3OH): λmax ϭ 274 nm. Ϫ IR (KBr): ν ϭ 3063, 3032, 2939,
2870, 1746 (s), 1619, 1594, 1513, 1498, 1452, 1439, 1378, 1259 (s),
1180, 1124 (s), 1019, 809, 742, 705 cmϪ1. Ϫ 1H NMR (CDCl3):
δ ϭ 2.62 (ddd, J ϭ 4.8, 17.0, 132.0 Hz, 4-Ha), 2.69 (ddd, J ϭ 4.9,
17.0, 134.0 Hz, 4-Hb), 3.73 (s, CH3 ester), 4.73 and 4.83 (2 d, J ϭ
11.9 Hz, 5-OCHaHbC6H5), 5.00 (s, 7-OCH2C6H5), 5.05 and 5.12 (2
s, 3Ј- and 4Ј-OCH2C6H5), 5.13 (d, J ϭ 3.5 Hz, 2-H), 5.38 (large dd,
J ϭ 4.8, 11.0 Hz, 3-H), 5.97 (m, 2ЈЈ- and 3ЈЈ-H), 6.09 (d, J ϭ
2.2 Hz, 6-H), 6.21 (d, J ϭ 2.2 Hz, 8-H), 6.79 (dd, J ϭ 1.9, 8.4 Hz,
6Ј-H), 6.86 (d, J ϭ 8.4 Hz, 5Ј-H), 6.91 (d, J ϭ 1.9 Hz, 2Ј-H),
7.22Ϫ7.48 (m, 24 H, 4 OCH2C6H5 and 2 OCOC5H6 meta), 7.53
and 7.58 (2 m, 2 H, 2 OCOC5H6 para), 8.02 and 8.06 (2 m, 4 H, 2
OCOC5H6 ortho). Ϫ 13C NMR (CDCl3): δ ϭ 22.3 (labelled C-4),
52.9 (CH3 ester), 69.6 (5-OCH2C6H5), 70.1 (7-OCH2C6H5), 71.1
(d, J ϭ 46 Hz, C-3), 71.3 (C-2ЈЈ and C-3ЈЈ), 71.4 (3Ј- and 4Ј-
OCH2C6H5), 77.2 (C-2), 93.8 (C-6), 94.3 (C-8), 100.3 (C-4a), 112.9
(C-2Ј), 115.1 (C-5Ј), 119.4 (C-6Ј), 126.9, 127.2, 127.4, 127.5, 127.7,
127.9, 128.3, 128.4, 128.5, 128.6 (OCH2C6H5 ortho/meta/para, and
OCOC6H5 meta), 128.5 and 128.6 (OCOC6H5 ipso), 130.0 and
130.1 (OCOC6H5 ortho), 130.7 (C-1Ј), 133.5 and 133.6 (OCOC6H5
para), 136.8 (5-OCH2C6H5 ipso), 136.9 (7-OCH2C6H5 ipso), 137.0
and 137.2 (3Ј- and 4Ј-OCH2C6H5 ipso), 149.1 (C-3Ј, C-4Ј), 154.3
(C-8a), 157.4 (C-5), 158.9 (C-7), 164.9 (3ЈЈ-OCOC6H5), 165.0 (2ЈЈ-
OCOC6H5), 165.2 (C-1ЈЈ), 166.3 (C-4ЈЈ). Ϫ MS (FABϩ, nitrobenzyl
alcohol): m/z (%) ϭ 1006 [MHϩ] (100), 634 (50), 542 (20). Ϫ
13C1C61H52O13 (1006.1): calcd. C 74.12, H 5.21; found C 74.14,
H 5.09.
(؉)-1: ee Ͼ 99% based on HPLC analysis on a chiral Cyclobond
I (β-cyclodextrin bonded) column (1 mL/min, eluent H2O/MeOH,
85:15 from 0 to 10 min, linear gradient from 85:15 at 10 min to
50:50 at 60 min). Ϫ [α]2D0 ϭ ϩ16 (c ϭ 1, water/acetone, 1:1; ref.[13]
ϩ15.4). Ϫ M.p. 205Ϫ210 °C (dec.). Ϫ UV/Vis (MeOH): λmax
ϭ
279 nm. Ϫ 1H NMR (CD3OD): δ ϭ 2.51 (ddd, J ϭ 130.0, 16.2,
8.0 Hz, 4β-H), 2.84 (ddd, J ϭ 130.0, 16.2, 5.4 Hz, 4α-H), 3.98 (m,
3-H), 4.57 (dd, J ϭ 7.4, 3.3 Hz, 2-H), 5.87 (d, J ϭ 2.2 Hz, 8-H),
5.94 (d, J ϭ 2.2 Hz, 6-H), 6.72 (dd, J ϭ 8.1, 2.0 Hz, 6Ј-H), 6.76 (d,
J ϭ 8.1 Hz, 5Ј-H), 6.84 (d, J ϭ 2.0 Hz, 2Ј-H). Ϫ 13C NMR
(CD3OD): δ ϭ 28.4 (labelled C-4), 68.8 (d, J ϭ 37 Hz, C-3), 82.7
(C-2), 95.6 (C-6), 96.4 (C-8), 100.9 (d, J ϭ 43 Hz, C-4a), 115.3 (C-
2Ј), 116.2 (C-5Ј), 120.1 (C-6Ј), 132.2 (C-1Ј), 146.2 (C-3Ј, C-4Ј),
156.8 (d, J ϭ 5 Hz, C-8a), 157.5 (C-7), 157.7 (d, J ϭ 6 Hz, C-5).
Ϫ HRMS (FABϩ): m/z calcd. for [13C12C14H14O6ϩH] 292.0902;
found 292.0901. Ϫ 13C1C14H16O7,H2O (309.3): calcd. C 58.57, H
5.21; found C 58.65, H 5.26.
(؊)-16: [α]2D0 ϭ Ϫ29 (c ϭ 1, CH2Cl2). Ϫ UV/Vis (CH3OH): λmax ϭ
275 nm. Ϫ IR (film): ν˜ ϭ 3063, 3032, 2932, 2862, 1767, 1734 (s),
1618, 1594, 1510, 1498, 1452, 1438, 1376, 1249 (s), 1122 (s), 1025,
734, 698 cmϪ1. Ϫ 1H NMR (CDCl3): δ ϭ 2.85 (dd, J ϭ 5.0,
(؊)-1: ee Ͼ 99%. Ϫ [α]2D0 ϭ Ϫ16 (c ϭ 1, water/acetone, 1:1). Ϫ
Same spectroscopic data as (؉)-1. Ϫ HRMS (FABϩ): m/z calcd.
for [13C12C14H14O6 ϩ H] 292.0902; found 292.0909.
132 Hz, two 4-H), 3.75 (s, CH3 ester), 4.96, 5.01 and 5.08 (2 d, J ϭ (؊)-[4؊13C]Epicatechin (2): Compound (؊)-1 (2.1 g) was dissolved
12 Hz, and s, respectively, 3Ј- and 4Ј-OCH2C6H5), 5.02 and 5.03 in a solution of Na3PO4 (2 g) in distilled water at pH ϭ 11.4, in a
(5- and 7-OCH2C6H5), 5.08 (d, J ϭ 11.4 Hz, 2-H), 5.42 (dd, J ϭ flask purged with nitrogen. After 20 h at room temp., HPLC ana-
5.0, 11.4 Hz, 3-H), 5.79 (d, J ϭ 2.9 Hz, 2ЈЈ-H), 6.99 (d, J ϭ 2.9 Hz, lysis showed a mixture of 24% epicatechin and 76% catechin (RP18,
3ЈЈ-H), 6.16 (d, J ϭ 2.2 Hz, 8-H), 6.27 (d, J ϭ 2.2 Hz, 6-H), 6.72 gradient from H2O/MeOH/TFA, 85:15:0.0025 to 20:80:0.0025
(dd, J ϭ 1.9, 8.7 Hz, 6Ј-H), 6.76 (d, J ϭ 8.7 Hz, 5Ј-H), 6.84 (d, J ϭ within 1 h, detection at 280 nm). The solution was then acidified
1.9 Hz, 2Ј-H), 7.26Ϫ7.50 (m, 24 H, 4 OCH2C6H5 and 2 OCOC6H5 to pH ϭ 6.5 with aq. 1 HCl, and extracted 5 times with ethyl
meta), 7.53 and 7.62 (2 m, 2 OCOC6H5 para), 8.10 (m, 2 OCOC6H5 acetate. The combined organic extracts were dried with Na2SO4
ortho). Ϫ 13C NMR (CDCl3): δ ϭ 22.8 (labelled C-4), 52.9 (CH3
and concentrated. The crude extract was purified by chromato-
ester), 70.0 and 70.1 (5- and 7-OCH2C6H5), 71.0, 71.1, 71.4 and graphy on Sephadex LH-20 (water/ethanol, 85:15, 5 mL/min),
71.5 (C-3, C-2ЈЈ, C-3ЈЈ, 3Ј- and 4Ј-OCH2C6H5), 76.9 (C-2), 93.9 (C- yielding pure (Ϫ)-2 in the first fractions [386 mg, 18% from starting
8), 94.5 (C-6), 100.5 (d, J ϭ 43 Hz, C-4a), 112.8 (C-2Ј), 114.7 (C-
(؊)-1] and recovered (؊)-1 (1.45 g, 69%) which was epimerized
5Ј), 119.2 (C-6Ј), 127.1, 127.2, 127.3, 127.6, 127.7, 127.9, 128.0, once more. In this way, we obtained 1.23 g of (؊)-2 from 2.9 g of
128.3, 128.4, 128.5, 128.6, 133.5, 133.6 (OCH2C6H5 ortho/meta/
Eur. J. Org. Chem. 2001, 2379Ϫ2384
(؊)-1 [50% yield based on recovered (؊)-1] while 434 mg of (؊)-1
2383

B. Nay, V. Arnaudinaud, J. Vercauteren
FULL PAPER
lon, J. P. Monti, J. Vercauteren, Tetrahedron Lett. 2001, 42,
1279Ϫ1281.
B. C. B. Bezuidenhoudt, D. Ferreira in Plant Polyphenols (Eds.:
R. W. Hemingway, P. E. Lacks), Plenum Press, New York,
1992, p. 143Ϫ165.
H. Kawamoto, F. Nakatsubo, K. Murakami, Synth. Commun.
1996, 26, 531Ϫ534.
E. J. Bourne, M. Stacey, J. C. Tatlow, J. M. Tedder, J. Chem.
Soc. 1951, 718Ϫ720.
J. M. Tedder, Chem. Rev. 1955, 787Ϫ827.
J. W. Clark-Lewis, D. C. Skingle, Aust. J. Chem. 1967, 20,
2169Ϫ2190.
H. Kawamoto, F. Nakatsubo, K. Murakami, J. Wood Chem.
Technol. 1989, 9, 35Ϫ52.
W. Tückmantel, A. P. Kozikowski, L. J. Romanczyk Jr., J. Am.
Chem. Soc. 1999, 121, 12073Ϫ12081.
B. Nay, J. F. Peyrat, J. Vercauteren, Eur. J. Org. Chem. 1999,
2231Ϫ2234.
were recovered. According to the ee of (؊)-1, the ee of (Ϫ)-2 was
Ͼ 99% and was checked by HPLC analysis on a β-cyclodextrin-
[4]
bonded column (1 mL/min, eluent H2O/MeOH, 96:4). Ϫ [α]2D0
ϭ
Ϫ56 (c ϭ 1, acetone/water, 1:1; ref.[14] Ϫ60). Ϫ M.p. 230Ϫ235 °C
(dec.). Ϫ UV/Vis (MeOH): λmax ϭ 279 nm. Ϫ 1H NMR (CD3OD):
δ ϭ 2.72 (ddd, J ϭ 132.1, 16.7, 2.8 Hz, 4α-H), 2.85 (ddd, J ϭ 126.4,
16.7, 4.6 Hz, 4β-H), 4.17 (m, 3-H), 4.81 (large s, 2-H), 5.91 (d, J ϭ
2.2 Hz, 8-H), 5.94 (d, J ϭ 2.2 Hz, 6-H), 6.75 (d, J ϭ 8.1 Hz, 5Ј-H),
6.79 (dd, J ϭ 8.1, 2.0 Hz, 6Ј-H), 6.97 (d, J ϭ 2.0 Hz, 2Ј-H). Ϫ 13C
NMR (CD3OD) δ ϭ 31.7 (labelled C-4), 70.0 (d, J ϭ 36 Hz, C-3),
82.4 (C-2), 98.4, 99.0 (C-8 and C-6), 102.6 (d, J ϭ 44 Hz, C-4a),
117.8 (C-6Ј), 118.4 (C-5Ј), 121.9 (C-2Ј), 134.9 (C-1Ј), 148.2, 148.3
(C-3Ј and C-4Ј), 159.9 (C-8a), 160.2 (C-7), 160.5 (C-5). Ϫ HRMS
(FABϩ): m/z calcd. for [13C12C14H14O6ϩH] 292.0902; found
292.0903.
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
B. R. Brown, M. J. Fuller, J. Chem. Res. 1986, 140Ϫ141.
P. Kiatgrajai, J. D. Wellons, L. Gollob, J. D. White, J. Org.
Chem. 1982, 47, 2910Ϫ2912.
J. A. Kennedy, M. H. G. Munro, H. K. J. Powell, L. J. Porter,
L. Y. Foo, Aust. J. Chem. 1984, 37, 885Ϫ892.
L. Y. Foo, L. J. Porter, J. Chem. Soc., Perkin Trans. 1 1983,
1535Ϫ1543.
Acknowledgments
[14]
[15]
[16]
[17]
´
We thank Region Aquitaine, MENRT and ONIVins for financial
support. We particularly acknowledge Karen Prescott for her help-
ful language corrections. B. N. gratefully acknowledges a scholar-
ship from MENRT.
J. A. Steenkamp, D. Ferreira, D. Roux, Tetrahedron Lett. 1985,
26, 3045Ϫ3048.
Y. Naito, M. Sugiura, Y. Yamaura, C. Fukaya, K. Yokoyama,
Y. Nakagawa, T. Ikeda, M. Senda, T. Jujita, Chem. Pharm.
Bull. 1991, 39, 1736Ϫ1745.
H. Kawamoto, F. Nakatsubo, K. Murakami, Mokuzai Gakkai-
shi 1991, 37, 488Ϫ493; Chem. Abstr. 1991, 115, 279643y.
Received November 16, 2000
[1]
B. Nay, V. Arnaudinaud, J. F. Peyrat, A. Nuhrich, G. Deffieux,
J. M. Merillon, J. Vercauteren, Eur. J. Org. Chem. 2000,
´
[18]
1279Ϫ1283.
[2]
[3]
´
B. Nay, J. P. Monti, A. Nuhrich, G. Deffieux, J. M. Merillon,
J. Vercauteren, Tetrahedron Lett. 2000, 41, 9049Ϫ9051.
´
V. Arnaudinaud, B. Nay, A. Nuhrich, G. Deffieux, J. M. Meril-
[O00582]
2384
Eur. J. Org. Chem. 2001, 2379Ϫ2384