Photochemistry synthesis. Part 2: Enantiomerically pure polyhydroxy-1,1,3-triarylpropan-2-ols

Article abstract of DOI:10.1016/j.jphotochem.2011.10.019

A new method to open the heterocyclic ring of flavan-3-ols via photolytic cleavage of the ether bond, with stereoselective trapping of the intermediates with phloroglucinol to obtain phloroglucinol grafted derivatives of flavan-3-ols, was developed. Photolysis of catechin and epicatechin, respectively, in the presence of phloroglucinol yielded the enantiomeric (1S,2S)- and (1R,2R)-1,3-di(2,4,6-trihydroxyphenyl)-1-(3,4-dihydroxyphenyl) propan-2-ols, respectively. The absolute configuration at C-1 and C-2 was determined by electronic circular dichroism (experimental and calculated) and these results confirmed that the trapping mechanism is controlled by the C-3 configuration of the flavan-3-ol.

Full text of DOI:10.1016/j.jphotochem.2011.10.019

Journal of Photochemistry and Photobiology A: Chemistry 227 (2012) 18–24
Contents lists available at SciVerse ScienceDirect
Journal of Photochemistry and Photobiology A:
Chemistry
journal homepage: www.elsevier.com/locate/jphotochem
Photochemistry synthesis. Part 2: Enantiomerically pure
polyhydroxy-1,1,3-triarylpropan-2-olsଝ
Anke Wilhelm-Moutona, Susan L. Bonneta, Yuanqing Dingb, Xing-Cong Lic, Daneel Ferreirac,d
,
Jan H. van der Westhuizena,∗
a Department of Chemistry, University of the Free State, Nelson Mandela Avenue, Bloemfontein 9301, South Africa
b National Center for Natural Products Research, The University of Mississippi, MS 38677, USA
c Research Institute of Pharmaceutical Sciences, The University of Mississippi, MS 38677, USA
d Department of Pharmacognosy, The University of Mississippi, MS 38677, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 20 July 2011
Received in revised form 20 October 2011
Accepted 22 October 2011
Available online 29 October 2011
A new method to open the heterocyclic ring of flavan-3-ols via photolytic cleavage of the ether bond, with
stereoselective trapping of the intermediates with phloroglucinol to obtain phloroglucinol grafted deriva-
tives of flavan-3-ols, was developed. Photolysis of catechin and epicatechin, respectively, in the presence
of phloroglucinol yielded the enantiomeric (1S,2S)- and (1R,2R)-1,3-di(2,4,6-trihydroxyphenyl)-1-(3,4-
dihydroxyphenyl)propan-2-ols, respectively. The absolute configuration at C-1 and C-2 was determined
by electronic circular dichroism (experimental and calculated) and these results confirmed that the
trapping mechanism is controlled by the C-3 configuration of the flavan-3-ol.
Keywords:
Photochemistry
Flavan-3-ol
© 2011 Elsevier B.V. All rights reserved.
Polyhydroxy-1,1,3-triarylpropan-2-ols
Electronic circular dichroism
1. Introduction
7
and trace amounts of the 4␤-phloroglucinol-fisetinidol
9
[Scheme 2].
Flavan-3-ols and their oligomers are ubiquitous in higher plants
[1,2] and are commercially important constituents of red wine [3],
leather [4], adhesives [5–7] and black tea [2]. Due to their antioxi-
dant properties, free phenolic flavan-3-ols exhibit protection from
diseases attributed to oxidative stress such as cancers [8,9], cardio-
vascular [10] and neurogenerative diseases [11]. Other biological
effects include antitumor activity [12], improved blood flow [13],
the inhibition of cholesterol absorption [14] and protection from
damage by ultraviolet B radiation [15].
Much of the chemistry and photochemistry of catechin 1a are
explained by the formation of reactive B-ring quinone methide
intermediates, e.g. 2a, or A-ring o- or p-quinone methide interme-
diates, e.g. 3 and 4, respectively [Scheme 1] [16].
The formation of compound 7 requires cleavage of the C-2 ether
bond to form the B-ring quinone methide 6 followed by stereose-
lective intramolecular recyclization via the stronger nucleophylic
hydroxy group of the phloroglucinol D-ring. Inversion of configu-
ration at C-4 from 4␣ to 4␤ is due to rotation about the C-2-C-3
bond in 6 (interconversion of the A and D-rings). Inversion of con-
figuration at C-2 from 2R to 2S is attributed to the C-3 absolute
configuration of 5 (C-2 of 6) that requires anti-recyclization of 6.
Intramolecular recyclization is slow enough to permit bond
rotation to yield the observed isomerization at C-2, but too fast
to allow intermolecular trapping of intermediate 6 by methanol or
2-propanol. The formation of 9 requires the formation of the A-
ring quinone methide 8. An ionic or radical mechanism could not
be ruled out for the formation of 7 but could not explain the for-
mation of 9. This reaction represented the first and hitherto only
synthetic access to the thermodynamically less stable 2,3-cis-3,4-
cis-4-arylflavan-3-ols.
Forest and coworkers [15] investigated the photochemistry of
catechin 1a and epicatechin 10 in water to understand the photo-
chemistry of humic substances in aquatic systems. They observed
ultraviolet light induced epimerization of epicatechin 10 and ent-
catechin 11 in a 1:1 mixture of water and acetonitrile [Scheme 3]
to a photostationary ratio of 1:19. This ratio correlates with the fact
that 2,3-trans-flavan-3-ols are thermodynamically more stable
Van der Westhuizen and coworkers [17] investigated the photo-
chemistry and electronic circular dichroism (ECD) of enantiomer-
ically pure free phenolic 4-arylflavan-3-ols. Sensitized photolysis
(0.05 M benzophenone) of the 4␣-phloroglucinol-fisetinidol 5 in
methanol gave the first access to the 4␤-resorcinol-ent-epicatechin
ଝ
Part 1 is: Z. Han, S.L. Bonnet, J.H. Van der Westhuizen, Tetrahedron 64 (2008)
2619–2625.
∗
Corresponding author. Tel.: +27 51 401 2782; fax: +27 51 444 6384.
E-mail address: vdwestjh@ufs.ac.za (J.H. van der Westhuizen).
1010-6030/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.jphotochem.2011.10.019

A. Wilhelm-Mouton et al. / Journal of Photochemistry and Photobiology A: Chemistry 227 (2012) 18–24
19
Scheme 1. Formation of o- and p-quinone methides.
Scheme 4. Acid catalyzed condensation of catechin 1a with phloroglucinol 12a and
resorcinol 12c, respectively.
3-phenyl-2,3-dihydrobenzofuran 14 [Scheme 4]. The postulated
1,1,3-triphenylpropanol intermediate 13a was proven during acid
catalyzed condensation of catechin 1a and diethylphloroglucinol
12b to yield 13b. The stereochemistry of the product was not deter-
mined.
Weinges and coworkers [19] investigated the acid-catalyzed
condensation of tetra-O-methylcatechin 1c and resorci-
nol 12c (dioxane/HCl(c); 2:3, v/v). The reaction yielded the
1,1,3-triphenylpropanol 13c and
a
2-diphenylmethyl-2,3-
dihydrobenzofuran 15 after methylation [Scheme 4].
Laks et al. [20] found that when catechin 1a was dissolved in
an alkaline solution under mild conditions (ambient temperature,
pH 12) in the presence of phloroglucinol 12a, a mixture of cat-
echinic acid 16 and the phloroglucinol adduct 13a was formed
[Scheme 5]. Catechinic acid 16 is formed via nucleophilic attack of
the phloroglucinol A-ring on the B-ring quinone methide via C-8.
Mitsunaga and co-workers [21–23] reported a BF3-catalyzed
condensation between phenol 17 and catechin 1a that yielded an
acyclic phenolated product 18 and the corresponding dehydration
product 19 [Scheme 6].
Herein, we report the photolytic cleavage of the pyran
heterocycle of catechin 1a and epicatechin 10, and the stereos-
elective trapping of the intermediates with phloroglucinol 12a
to obtain synthetic access to enantiopure polyhydroxy-1,1,3-
triarylpropan-2-ols. We determined the absolute configuration of
the 1,1,3-triarylpropan-2-ols via the comparison of calculated and
experimental ECD spectra.
Scheme 2. Photolysis of 4␣-phloroglucinol-fisetinidol.
than the 2,3-cis diastereomers. Since the absorption coefficients
of 10 and 11 are identical, the correlation suggests that photoiso-
merisation of 10 and 11 takes place via the same quinone methide,
zwitterionic or biradical intermediate. The sunscreen properties of
catechin and epicatechin were attributed to this epimerization.
Mayer and coworkers [18] reported acid-catalyzed [HCl (aq)]
fission of the heterocyclic ring of catechin 1a with simul-
taneous grafting of
a nucleophilic phenolic species such as
phloroglucinol 12a followed by dehydration to yield a 2-benzyl-
Scheme 3. Photochemical isomerization of epicatechin 10 and ent-catechin 11.
Scheme 5. Base-catalyzed condensation of catechin 1a with phloroglucinol 12a.

20
A. Wilhelm-Mouton et al. / Journal of Photochemistry and Photobiology A: Chemistry 227 (2012) 18–24
20b: ıH (600 MHz, CDCl3, Me4Si) 7.19 (1H, d, J = 2.0 Hz, H-2ꢀꢀꢀ),
7.17 (1H, dd, J = 2.0, 8.5 Hz, H-6ꢀꢀꢀ), 7.10 (1H, d, J = 8.5 Hz, H-5ꢀꢀꢀ), 6.85
(2H, s, H-3ꢀꢀ, H-5ꢀꢀ), 6.82 (2H, broad s, H-3ꢀ, H-5ꢀ), 5.84 (1H, m, H-2),
4.75 (1H, d, J = 9.0 Hz, H-1), 3.19 (1H, dd, J = 5.7, 14.4 Hz, H-3␣), 3.00
(1H, dd, J = 4.6, 14.4 Hz, H-3␤), 2.27 (3H, s, 1× OAc), 2.26 (3H, s, 1×
OAc), 2.25 (3H, s, 1× OAc), 2.23 (3H, s, 1× OAc), 2.04 (6H, broad s,
2× OAc), 2.00 (6H, s, 2× OAc), 1.72 (3H, s, 1× OAc); ıC (150 MHz,
CDCl3, Me4Si) 170.6–168.0 (9× O-COCH3), 150.5 (C-3ꢀꢀ/C-5ꢀꢀ), 149.6
(C-3ꢀ/C-5ꢀ), 149.3 (C-3ꢀꢀꢀ), 149.2 (C-4ꢀ), 142.0 (C-4ꢀꢀ), 140.6 (C-4ꢀꢀꢀ),
138.2 (C-1ꢀ), 126.2 (C-2ꢀꢀꢀ), 123.5 (C-5ꢀꢀꢀ), 123.0 (C-1ꢀꢀ), 122.8 (C-6ꢀꢀꢀ),
118.8 (C-1ꢀꢀꢀ), 114.3 (C-2ꢀ/C-6ꢀ), 113.7 (C-2ꢀꢀ/C-6ꢀꢀ), 71.2 (C-2), 42.1
(C-3␣/␤), 26.1 (C-1), 21.1–20.3 (9× O-COCH3).; HR-MS: [M+H]+,
found m/z 795.2139. C39H38O18 requires m/z 795.2135.; IR (KBr): ꢀ
max = 1774, 1371, 1193, 1025 cm−1; ECD: [ꢁ]278.4 2.4 × 103, [ꢁ]260.0
1.6 × 103, [ꢁ]230.0 4 × 103, [ꢁ]209.2 8.7 × 103, [ꢁ]193.0 −9.4 × 101,
[ꢁ]187.0 1.5 × 103.
Scheme 6. BF3 catalyzed condensation between phenol 17 and catechin 1a.
2-[(2S,3R)-2-Acetoxy-3-(3,4-diacetoxyphenyl)-3-methoxypropyl]
benzene-1,3,5-triyl triacetate 21b, and 2-[(2S,3S)-2-acetoxy-3-(3,
4-diacetoxyphenyl)-3-methoxypropyl]benzene-1,3,5-triyl triacetate
21d.
2. Experimental
2.1. Analytical
Acetylation of the reaction mixture yielded 21b and 21d as a
mixture which could not be separated. However, since there is
almost no overlap of resonances in the 1H and 13C NMR spectra,
respectively, both acetate diastereoisomers could be fully charac-
terized with NMR.
21b (in mixture): ıH (600 MHz, CDCl3, Me4Si) 7.10 (1H, d,
J = 8.2 Hz, H-5ꢀ), 7.06 (1H, d, J = 2.0 Hz, H-2ꢀ), 7.03 (1H, dd, J = 2.0,
8.2 Hz, H-6ꢀ), 6.87 (2H, s, H-3ꢀꢀ/H-5ꢀꢀ), 5.06 (1H, m, H-2), 4.12 (1H, d,
J = 3.7 Hz, H-3), 3.23 (3H, s, 1× OMe), 2.85 (1H, dd, J = 6.7, 13.8 Hz,
H-1␣), 2.61 (1H, dd, J = 8.2, 13.8 Hz, H-1␤), 2.20 (3H, s, 1× OAc),
2.21 (3H, s, 1× OAc), 2.19 (6H, s, 2× OAc), 2.18 (3H, s, 1× OAc), 1.85
(3H, s, 1× OAc); ␦C (100.6 MHz, CDCl3) 170.2–167.9 (6× O-COCH3),
150.2–141.7 (C-2ꢀꢀ/C-6ꢀꢀ, C-4ꢀꢀ, C-3ꢀ, C-4ꢀ), 136.5 (C-1ꢀꢀ), 125.3–122.5
(C-2ꢀ, C-5ꢀ, C-6ꢀ), 199.7 (C-1ꢀ), 113.9 (C-3ꢀꢀ/C-5ꢀꢀ), 81.1 (C-3), 74.1
(C-2), 57.6 (OCH3), 24.7 (C-1), 21.2–20.5 (6× O-COCH3).
All NMR experiments were recorded on a Bruker 600 MHz spec-
trometer. ECD spectra were recorded on a Jasco J-710 spectropo-
larimeter in spectrophotometric grade methanol (∼1 mg/10 mL
MeOH). Mass spectra were recorded on a Waters Quattro Ultima
triple quadrupole mass spectrometer, operated in the positive
mode. Solid state FR-IR spectra were recorded as KBr pellets on
a Bruker Tensor 27 spectrometer in the range of 3000–600 cm−1
.
All photochemical reactions were carried out inside a RAYON pho-
tochemical reactor manufactured by Southern N.E. Ultraviolet Co.,
Middletown, CT, USA, equipped with Rayonet Photochemical Reac-
tor lamps Cat. No. RPR-250, 300, and 350 nm, respectively.
2.2. Synthesis
2.2.1. Photolysis of catechin
21d (in mixture): ıH (600 MHz, CDCl3, Me4Si) 7.21 (1H, dd,
J = 2.0, 8.3 Hz, H-6ꢀ), 7.15 (1H, d, J = 2.0 Hz, H-2ꢀ), 7.13 (1H, d,
J = 8.3 Hz, H-5ꢀ), 6.75 (2H, s, H-3ꢀꢀ/H-5ꢀꢀ), 5.00 (1H, m, H-2), 4.33 (1H,
d, J = 3.5, H-3), 3.28 (3H, s, 1× OMe), 3.00 (1H, dd, J = 10.9, 14.3 Hz,
H-1␣), 2.53 (1H, dd, J = 2.90, 14.3 Hz, H-1␤), 2.16 (3H, s, 1× OAc),
2.15 (3H, s, 1× OAc), 2.07 (3H, s, 1× OAc), 2.05 (6H, s, 2× OAc), 2.03
(3H, s, 1× OAc); ıC (100.6 MHz, CDCl3) 170.2–167.9 (6× O-COCH3),
150.2–141.7 (C-2ꢀꢀ/C-6ꢀꢀ, C-4ꢀꢀ, C-3ꢀ, C-4ꢀ), 137.0 (C-1ꢀꢀ), 125.0–122.3
(C-2ꢀ, C-5ꢀ, C-6ꢀ), 120.0 (C-1ꢀ), 113.6 (C-3ꢀꢀ/C-5ꢀꢀ), 83.6 (C-3), 76.1
(C-2), 58.0 (OCH3), 22.9 (C-1), 21.2–20.5 (6× O-COCH3).
(1S,2S)-1,3-Di(2,4,6-trihydroxyphenyl)-1-(3,4-
dihydroxyphenyl)propan-2-ol
dihydroxyphenyl)-2-hydroxy-3-methoxypropyl]benzene-1,3,5-triol
21a, and 2-[(2S,3S)-3-(3,4-dihydroxyphenyl)-2-hydroxy-3-
20a,
2-[(2S,3R)-3-(3,4-
methoxypropyl]benzene-1,3,5-triol 21c.
Catechin 1a (450 mg, 1.55 mmol) and phloroglucinol 12a
(600 mg, 4.74 mmol) were irradiated in MeOH (40 mL) at 250 nm for
20 h. The solvent was evaporated in vacuo to give the crude product,
which was purified by column chromatography (Sephadex LH-20,
EtOH as eluant) to give 20a (95 mg, 15%) as an amorphous solid, a
mixture of 21a and 21c (75 mg, 15% combined yield), and unreacted
catechin 1a (155 mg, 34%).
2.2.2. Photolysis of epicatechin
(1R,2R)-1,3-Di(2,4,6-trihydroxyphenyl)-1-(3,4-
dihydroxyphenyl)propan-2-ol 22a.
20a: ıH (600 MHz, DMSO, Me4Si, 120 ◦C) 8.93 (1H, s, 1× OH),
8.87 (1H, s, 1× OH), 8.80 (2H, s, 2× OH), 8.74 (1H, s, 1× OH),
8.47 (1H, s, 1× OH), 8.38 (1H, s, 1× OH), 6.61 (1H, d, J = 2.0 Hz,
H-2ꢀꢀꢀ), 6.53 (1H, d, J = 8.3 Hz, H-5ꢀꢀꢀ), 6.50 (1H, dd, J = 2.0, 8.3 Hz, H-
Epicatechin 10 (580 mg, 2.0 mmol) and phloroglucinol 12a
(750 mg, 5.95 mmol) were irradiated at 250 nm in MeOH (40 mL)
under nitrogen for 20 h. The solvent was evaporated in vacuo to give
the crude product, which was purified by column chromatography
(Sephadex LH-20, ethanol as eluant) to give the title compound 22a
(137 mg, 16%) as an amorphous solid, ent-catechin (58 mg, 10%)
[ECD spectra of acetylated catechin and ent-catechin Figs. S2 and S3,
respectively], and unreacted epicatechin 10 (181 mg, 31%).
22a: ıH (600 MHz, DMSO, Me4Si) identical to 20a. ıC (100.6 MHz,
CDCl3) identical to 20a. HR-MS: [M+H]+, found m/z 417.1194.
6
ꢀꢀꢀ), 5.85 (2H, s, H-3ꢀ, H-5ꢀ), 5.84 (2H, s, H-3ꢀꢀ, H-5ꢀꢀ), 4.58 (1H, d,
J = 4.3 Hz, H-1), 4.54 (1H, m, H-2), 2.70 (1H, dd, J = 2.5 Hz, 14.3 Hz,
H-3␣), 2.43 (1H, dd, J = 9.6 Hz, 14.3 Hz, H-3␤); ıC (100.6 MHz, Act-
D6) 157.5, 157.2, 157.0, 156.9 (6× C-OH), 144.4, 142.9 (2× C-OH),
134.3 (C-1ꢀꢀꢀ), 119.5 (C-2ꢀꢀꢀ), 115.6 (C-5ꢀꢀꢀ), 114.6 (C-6ꢀꢀꢀ), 105.7 (C-
1ꢀꢀ), 104.8 (C-1ꢀ), 95.2 (C-3ꢀ/5ꢀ and C-3ꢀꢀ/5ꢀꢀ), 76.6 (C-2), 45.4 (C-1),
29.3 (C-3); m/z (ESI) 417 (30, MH+), 257 (25), 159 (100), 141 (50),
109 (37); HR-MS: [M+H]+, found m/z 417.1189. C21H20O9 requires
m/z 417.1185; IR (KBr): ꢀ max = 3419, 1614, 1151, 1023, 989 cm−1
;
C21H20O9 requires m/z 417.1185. IR (KBr): ꢀ max = 3419, 1614,
ECD: [ꢁ]286.6 8.3 × 102, [ꢁ]262.6 6.3 × 102, [ꢁ]248.0 3.0 × 102, [ꢁ]239.0
−2.2 × 102, [ꢁ]220.4 3.9 × 103, [ꢁ]207.8 −3.4 × 102, [ꢁ]201.6 3.3 × 102,
[ꢁ]190.4 1.7 × 102.
1151, 1023, 989 cm−1; ECD: [ꢁ]293.8 6.8 × 102, [ꢁ]276.6 8.6 × 102,
[ꢁ]260.0 −2.6 × 102, [ꢁ]238.2 6.9 × 102, [ꢁ]222.2 −3.8 × 103, [ꢁ]206.8
1.2 × 103, [ꢁ]199.8 −4.1 × 102, [ꢁ]188.0 5.1 × 102.
Acetylation of 20a yielded (1S,2S)-1,3-di(2,4,6-triacetoxy-
phenyl)-1-(3,4-diacetoxyphenyl)propan-2-ol 20b.
Acetylation of 22a yielded (1R,2R)-1,3-di(2,4,6-triacetoxyphenyl)-
1-(3,4-diacetoxyphenyl)propan-2-ol 22b.

A. Wilhelm-Mouton et al. / Journal of Photochemistry and Photobiology A: Chemistry 227 (2012) 18–24
21
Scheme 8. Photolysis of epicatechin 10 and phloroglucinol 12a.
Scheme 7. Photolysis of catechin 1a and phloroglucinol 12a.
22b: ıH (600 MHz, CDCl3, Me4Si) identical to 20b. ıC (150 MHz,
CDCl3, Me4Si) identical to 20b. HR-MS: [M+H]+, found m/z
795.2145. C39H38O18 requires m/z 795.2135. IR (KBr): ꢀ max
identical to 20b; ECD: [ꢁ]292.4 2.3 × 103, [ꢁ]251.4 8.7 × 102, [ꢁ]230.4
−2.7 × 103, [ꢁ]209.6 −9.0 × 103, [ꢁ]195.8 2.3 × 103, [ꢁ]185.2 6 × 102.
Scheme 9. Photolysis of tetra-O-methylcatechin 1c in MeOH/Acetone.
2.2.3. Photolysis of 3ꢀ,4ꢀ,5,7-tetra-O-methylcatechin
2-(2-Hydroxy-2-methylpropyl)-3,5-dimethoxyphenol 25.
Photolysis of catechin 1a and epicatechin 10 in the presence of
resorcinol and phenol, respectively, at 250 nm showed no coupling
products. Grafting of a phenolic moiety at C-2 of the presumed B-
ring quinone-methide intermediate (e.g. 2a) was only observed in
the case of phloroglucinol.
3ꢀ,4ꢀ,5,7-Tetra-O-methylcatechin 1c (50 mg, 0.14 mmol) and
phloroglucinol 12a (82 mg; 0.65 mmol) in MeOH/acetone (25 mL)
(90/10, v/v) were irradiated at 300 nm for 2.5 h. The solvent was
evaporated in vacuo to give the crude product, which was purified
by silica column chromatography (T/A 7:3) to give the title com-
pound 25 (7.0 mg, 14%) as an amorphous solid; Rf (T/A 7:3) 0.60,
and 3,4-dimethoxybenzaldehyde 24 (4.0 mg, 8%); Rf (T/A 7:3) 0.46.
25: ıH (600 MHz, CDCl3, Me4Si) 8.48 (1H, s, OH-1), 6.18 (1H,
d, J = 2.4 Hz, H-3), 6.08 (1H, d, J = 2.4 Hz, H-5), 3.77 (3H, s, OMe),
3.75 (3H, s, OMe), 2.82 (2H, s, CH2), 1.94 (1H, s, OH-2ꢀ), 1.28 (6H,
s, 2×-CH3). ıC (150 MHz, CDCl3, Me4Si) 159.84 (C-2), 159.29 (C-
4), 157.54 (C-6), 106.05 (C-1), 94.56 (C-3), 91.17 (C-5), 75.25 (C-2ꢀ),
55.53, 55.25 (2×-OCH3), 35.74 (C-1ꢀ), 29.26 (C-3ꢀ); HR-MS: [M+H]+,
found m/z 227.1287. C12H18O4 requires m/z 227.1283.
4. Discussion
The respective 1,1,3-triarylpropan-2-ols 20a and 22a, obtained
from photolysis of catechin 1a and epicatechin 10 in the pres-
ence of phloroglucinol, have identical 1H NMR spectra, but mirror
image ECD spectra. The pronounced mirror image Cotton effects
observed for 20a and 22a in the 215–280 nm region, indicate their
enantiomeric relationship [Fig. 1].
The aromatic region of the 1H NMR spectra of the enan-
tiomers 20a and 22a show one two-proton singlet at ı 5.84 for the
free-rotating phloroglucinol moiety (A-ring), one ABX system cor-
responding to the B-ring (ı 6.48–6.64), and two broadened singlets
at ı 5.82 and 5.90 indicating the two protons of the rotationally
restricted [20] D-ring phloroglucinol moiety. At 120 ◦C the H-3(D)
3. Results
Photolysis of catechin 1a in MeOH at 250 nm in the presence of
phloroglucinol 12a resulted in the isolation of the optically active
product 20a (15% yield) with a (1S,2S) configuration, and unre-
acted optically active starting material 1a (34%) [Scheme 7]. Two
side products, diastereoisomers 21b and 21d, were detected upon
acetylation of the reaction mixture, but could not be separated.
Since there is no overlap of resonances, the 1H NMR spectrum of the
mixture could be interpreted. The ratio of 21b:21d was 4:1 (yields
of 12% and 3%, respectively).
Photolysis of epicatechin 10 under the same conditions resulted
in the isolation of the optically active products 22a (15% yield)
and ent-catechin 11, in a 2:1 ratio, and unreacted optically active
starting material 10 (25%) [Scheme 8].
Photolysis of 3ꢀ,4ꢀ,5,7-tetra-O-methylcatechin 1c in MeOH and
in the presence of 5 eq. of phloroglucinol gave no products. Addition
of acetone (10%), a triplet sensitizer, to the reaction mixture leads
to the formation of fragmentation products 24 (8%) and 25 (15%)
[Scheme 9].
Fig. 1. ECD spectra of 20a and 22a.

22
A. Wilhelm-Mouton et al. / Journal of Photochemistry and Photobiology A: Chemistry 227 (2012) 18–24
Fig. 3. ␣-Hydrogen atom occupies the 1,3-allylic-strain position.
Fig. 2. Quinone methide 2a and b and ionic intermediate 27a and b tautomers.
associated with ionic reaction products and ␲ → ␲*-triplet states
with radical reaction products [28], we postulate a singlet mecha-
nism. This is supported by the observation that efforts to sensitize
the reaction with acetophenone or acetone to improve yields failed
and that the use of 5 equivalents of phloroglucinol, a known singlet
inhibitor, leads to lower yields.
and H-5(D) resonances coalesce into a sharp two-proton singlet at
ı 5.85. In the aliphatic region H-1 and H-2 overlap at ı 4.46 at room
temperature. At 120 ◦C H-1 appears as a doublet at ı 4.58 (J = 4.3 Hz)
and H-2 as a multiplet at ı 4.50–4.46. One of the diastereotopic C-3
protons resonates as a doublet of doublets at ı 2.43 (J = 9.6, 14.3 Hz),
while the resonance of the other proton overlaps with the DMSO
resonance at ı 5.84.
Acetylation of 20a and 22a yielded the enantiomers 20b and
22b, respectively. The presence of nine acetoxy resonances in their
1H NMR spectra, confirms the fission of the heterocyclic C-rings and
the addition of a phloroglucinol moiety at C-1. The shielded acetoxy
resonances (ı 1.72) for 20b and 22b correspond to the C-2 acetoxy
group. The HMBC spectra show a correlation between H-1 (ı 4.46)
and C-1(D) (ı 26.2) indicating that the additional phloroglucinol
rings are indeed attached to C-1.
Photolysis of epicatechin 10 also resulted in epimerization at
C-2 to additionally yield ent-catechin 11. The 1H NMR spectrum of
11 is identical to that of catechin 1a, but mirror image ECD spectra
shows their enantiomeric relationship [24]. The photolytic epimer-
ization at C-2 of catechin 1a to the thermodynamically less stable
ent-epicatechin is not observed upon monitoring the reaction via
TLC. This correlates with Forest’s conclusion that the 2,3-trans iso-
mer (e.g. ent-catechin) is thermodynamically more stable than the
2,3-cis isomer, epicatechin [15].
The OMe-substituted diastereoisomers 21b and 21d appeared
as a single band upon preparative TLC. Efforts to further purify the
mixture failed. The high-resolution 1H NMR resonances of the mix-
ture do not overlap and occur in a consistent 4:1 ratio permitting
the assignment of all the protons of 21b and 21d, respectively, in the
mixture. The characteristic ABX system of the B-ring is observed in
the ı 7.30–7.00 region. Two one-proton doublet of doublets res-
onate at ı 2.85 and 2.61 for the major product (21b), and at ı
3.00 and 2.53 for its isomer (21d), and are assigned to the two
diastereotopic C-3 methylene protons. The C-1 methoxy group that
resonates at ı 3.23 and 3.28, respectively, proves the acyclic nature
of the product. The free phenolic OMe-substituted acetate precur-
sors 21a and 21c were not isolated.
The absolute configuration at C-1 of the phloroglucinol adduct
20a depends on the face of attack of the nucleophile 12a at the pla-
nar C-1 centre of 2a/27a. Attack at the Si-face of 2a/27a will give
rise to 1R absolute configuration, while Re-face attack will yield the
1S diastereomer. Bach and coworkers [29] postulated that ␣-chiral
secondary and tertiary benzylic carbocations will display a pre-
ferred conformation in which the ␣-hydrogen atom occupies the
1,3-allylic strain position. The diastereotopic faces of the carboca-
tions are differentiated by the functional groups at the ␣-carbon. In
the case of intermediate 2a/27a, this would imply that the bulkier
benzyl group at C-2 and not the less bulky hydroxy group direct
attack to the Si-face leading to the 1R,2S diastereomer [Fig. 3].
The C-2 absolute configuration in compound 20a was estab-
lished by theoretical calculation of its electronic circular dichroism
(ECD) spectrum. A systematic conformational search was carried
out with an arbitrarily chosen 1S,2S absolute configuration via
Monte Carlo random search in the SYBYL 8.1 program using an
MMFF94 molecular mechanics force-field calculation. An energy
cutoff of 10 kcal/mol was used to generate a wide window of con-
formers in the Boltzmann population, affording 77 conformers. The
first 20 conformers within an energy cut-off of 4 kcal/mol were
geometrically optimized using density functional theory (DFT) at
the B3LYP/6-31G** level (gas phase) and 19 conformers were relo-
cated. Six predominant conformers, 20c, 20d, 20e, 20f, 20g, and
20h were conformationally analyzed with a Boltzmann distribu-
tion of 57.7, 6.6, 8.4, 4.2, 8.0, and 8.7% by relative energy at the
B3LYP/6-311++G**//B3LYP/6-31G** level [Fig. 4, Table S1].
Their key dihedral reference angles are shown in Table S3.
Time dependent density functional theory (TDDFT) calculations
[30–36] of the six conformers were performed at the B3LYP/6-
311++G**//B3LYP/6-31G** level (gas phase). The overall pattern of
the calculated ECD spectrum is consistent with the experimental
one, although a blue shift [35] of approximately 15 nm is observed
for the three Cotton effects (207, 223, and 236 nm compared to 221,
239, and 250–262 nm) [Fig. 5], confirming the absolute configura-
tion as 1S,2S.
Identification of the methanol-trapped products (21a
and 21c) and not the hydroxymethyl-substituted prod-
uct 26 that would be expected from
a radical mechanism
[25], indicates p-hydroxy stabilized benzylic cation
a
This indicates that the nucleophilic attack at C-1 of 2a/27a takes
place from the Re-face (anti to the C-2 hydroxy group). The C-2
hydroxy group and not the bulkier benzyl group directs nucle-
ophilic attack from the anti-position.
27a or p-quinone methide 2a intermediate mechanism.
It appears from the optimized geometry of 20a, determined to be
20c [Fig. 4], that the high-amplitude Cotton effect at 221 nm in the
experimental ECD spectrum derives from the chiral perturbation of
the closely positioned phloroglucinol and catechol chromophores
at C-1. The spatial orientation of the two chromophores in 20c is
coincidently similar to that of the 4-aryl moiety relative to the A/C-
ring plane of 4␤-arylflavan-3-ols (e.g. 9), which exhibit diagnostic
positive Cotton effects at 220–240 nm [36]. Thus, the C-1 absolute
configuration in 20a was confirmed as 1S.
We could not distinguish unequivocally between the two puta-
tive intermediates 2a and 27a [Fig. 2].
Since both are planar C-1 sp2 hybridized tautomers, we do not
expect much difference in thermodynamic stability and stereose-
lectivity. The pKa values of phenols in the excited state are lower
(more acidic) than in the ground state [26], indicating improved
stabilization of an ionic photolytic intermediate in the excited state
compared to the ground state. The absence of coupling products in
the photolysis of tetra-O-methylcatechin 1c, indicates that a free C-
4ꢀ OH is essential to stabilize the intermediate benzylic carbocation
or quinone methide [16,27]. Since singlet excited ␲ → ␲*-states are
Such exclusive formation of the 1S diastereoisomer indicates
that the directory effect is indeed exerted by the C-2 hydroxy group
and not by the C-2 benzyl group.

A. Wilhelm-Mouton et al. / Journal of Photochemistry and Photobiology A: Chemistry 227 (2012) 18–24
23
Fig. 4. Optimized geometry of 20a.
Scheme 10. Proposed mechanism for the formation of the 1,1,3-triarylpropan-2-ol
20a.
A
feasible mechanism for the formation of the 1,1,3-
triarylpropan-2-ol 20a is summarized in Scheme 10. The same
mechanism accounts for the formation of the epicatechin derived
adduct 22a.
Thus, heterolytic cleavage of the etherocyclic bond in the sin-
glet excited state 28 leads to formation of the quinone methide
2a/benzylic carbocation 27a pair. Stereoselective trapping of these
reactive intermediates with phloroglucinol from the face anti to
the 2-hydroxy group via either the quinone methide 2a or the
benzylic carbocation-derived oxirane 29 yields the enantiopure
1,1,3-triarylpropan-2-ol 20a.
The formation of ent-catechin 11 in the photolysis of epicate-
chin (epimerization at C-2) and the absence of ent-epicatechin in
the photolysis of catechin (no epimerization at C-2) indicates that
recyclization of ring fission intermediates is only feasible from the
face anti to the 2-hydroxy group. This is in agreement with our
conclusion regarding the stereoselectivity of nucleophilic attack by
phloroglucinol on 2a or 27a.
Scheme 11. Photolysis of tetra-O-methylcatechin 1c in pure MeOH.
The formation of both diastereoisomers of the minor MeOH-
trapped products 21a and 21c, may be attributed to the decreased
bulk of MeOH in comparison to that of phloroglucinol. The major
isomer 21a is presumably formed from attack at the Re-face of inter-
mediate 2a/27a, while the less favoured Si-face-attack afforded 21c.
Thus, the stereoselectivity of these photolytic reactions may be
attributed to the C-2 hydroxy group of the intermediate 2a/27a
(for catechin 1a) or 2b/27b (for epicatechin 10), respectively, that
allows the bulky phloroglucinol group to exclusively attack the
quinone methide anti to the C-2 hydroxy group for the catechin
and epicatechin reaction intermediates 2a/27a and 2b/27b, respec-
tively.
Tetra-O-methylcatechin (1c) does not undergo photolytic cou-
pling with phloroglucinol under our conditions. This supports the
presence of a quinone methide intermediate 2a in the transforma-
tion of 1a to 20a (Scheme 10) as 1c cannot form a B-ring quinone
methide. The formation of fragmentation products 24 and 25 in
acetone/MeOH expands the work of Fourie and coworkers [27] who
observed the formation of 30 when 1c was irradiated in pure MeOH
[Scheme 11]. We assume that the methanol adduct 30 is formed via
a singlet and thus an ionic mechanism, and the acetone adduct 25
via a triplet and radical mechanism.
5. Conclusion
We have developed a novel method to cleave the heterocyclic
ring of flavan-3-ols via photolytic fission of the ether bond. The
intermediates are stereoselectively trapped with phloroglucinol to
obtain phloroglucinol-grafted derivatives of flavan-3-ols albeit in
moderate yields. It was demonstrated that the trapping mechanism
Fig. 5. Calculated ECD spectrum of compound 20a (black line) at the B3LYP/6-
311++G**//B3LYP/6-31G**level (gas phase) and its experimental ECD spectrum (blue
line) in MeOH. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of the article.)

24
A. Wilhelm-Mouton et al. / Journal of Photochemistry and Photobiology A: Chemistry 227 (2012) 18–24
is controlled by the C-3 configuration of the flavan-3-ol. By judicial
choice of starting materials both the (R,R) and (S,S) enantiomers of
the grafted phloroglucinol products were accessible.
[18] W. Mayer, F. Merger, G. Frank, Uber Kondensationsprodukte der Catechine,
VI. Zur Konstitution der Reaktionsprodukte des (+)-Catechins mit Phloroglucin
und C-Äthyl-phloroglucin, Liebigs Ann. Chem. 675 (1964) 126–134.
[19] K. Weinges, F. Toribio, Catechin-resorcinol condensation, Liebigs Ann. Chem.
681 (1965) 161–169.
[20] P.E. Laks, R.W. Hemingway, A.H. Conner, Condensed tannins. Base-catalysed
reactions of polymeric procyanidins with phloroglucinol: intramolecular rear-
rangements, J. Chem. Soc. Perkin Trans. I (1987) 1875–1881.
[21] T. Mitsunaga, I. Abe, T. Nogami, The phenolation of monomeric compound
of condensed tannins by Lewis acid catalyst, Mokuzai Gakkaishi 38 (1992)
565–592.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jphotochem.2011.10.019.
[22] T. Mitsunaga, I. Abe, T. Nogami, The formation of a ring-opened compound from
(+)-catechin by BF3 catalyst, Mokuzai Gakkaishi 93 (1993) 328–332.
[23] T. Mitsunaga, I. Abe, T. Nogami, The chemical structure and stereochemistry
of the phenol adduct compound derived from (+)-catechin upon phenolation,
Mokuzai Gakkaishi 38 (1994) 100–106.
[24] O. Korver, C.K. Wilkins, Circular dichroism spectra of flavanols, Tetrahedron 27
(1971) 5459–5465.
[25] Z. Han, S.L. Bonnet, J.H. Van der Westhuizen, Photochemistry synthesis.
Part 1. Synthesis of xanthine derivatives by photolysis of 1-(5ꢀ-oxohexyl)-
3,7-dimethyl-3,7-dihydro-1H-purine-2,6-dione (pentoxifylline): an ambident
chromophore, Tetrahedron 64 (2008) 2619–2625.
[26] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, vol. 1, Springer Sci-
ence + Business Media, New York, 2006, p. 260.
[27] T.G. Fourie, D. Ferreira, D.G. Roux, Flavonoid synthesis based on photolysis of
flavan-3-ols, 3-hydroxyflavanones, and 2-benzylbenzofuranes, J. Chem. Soc.
Perkin Trans. I (1977) 125–133.
References
[1] D. Ferreira, D. Slade, J.P.J. Marais, Flavans and proanthocyanidins, in: Ø.M.
Anderson, K.R. Markham (Eds.), Flavonoids: Chemistry, Biochemistry and
Applications, Taylor & Francis, Boca Raton, 2006, pp. 553–616.
[2] L.J. Porter, J.B. Harborne, The Flavonoids – Advances in Research Since 1986,
Chapman & Hall, London, 1994, pp. 23–56.
[3] R. Corder, W. Mullen, N.Q. Khan, S.C. Marx, E.G. Woods, M.J. Carrier, A. Crozier,
Oenology: redwine procyanidins and vascular health, Nature 444 (2006) 566.
[4] D.G. Roux, D. Ferreira, J.J. Botha, Structural considerations in predicting the
utilization of tannins, J. Agric. Food Chem. 28 (1980) 216–222.
[5] A. Pizzi, Wattle based adhesives for exterior grade particleboard, Forest Prod.
J. 28 (1978) 42–47.
[6] M.A.E. Santana, M.G.D. Baumann, A.H. Conner, Resol resins prepared with tan-
nin liquefied in phenol, Holzforschung 49 (1995) 146–152.
[28] A.K. Zarkadis, V. Georgakilas, G.P. Perdikomatis, A. Trifonov, G.G. Gurzadyan, S.
Skoulika, M.G. Siskos, Triplet- vs. singlet-state imposed photochemistry. The
role of substituent effects on the photo-fries and photodissociation reaction of
triphenylmethyl silanes, Photochem. Photobiol. Sci. 4 (2005) 469–480.
[29] D. Stadler, A. Goeppert, G. Rasul, G.A. Olag, G.K. Surya Prakash, T. Bach, Chiral
benzylic carbocations: low temperature NMR studies and theoretical calcula-
tions, J. Org. Chem. 74 (2009) 312–318.
[30] C. Diedrich, S.J. Grimme, Systematic investigation of modern quantum chemical
methods to predict circular dichroism spectra, J. Phys. Chem. A 107 (2003)
2524–2539.
[31] T.D. Crawford, M.C. Tam, M.L. Abrams, The current state of ab initio calculations
of optical rotation and electronic circular dichroism spectra, J. Phys. Chem. A
111 (2007) 12058–12068.
[32] P.J. Stephens, F.J. Devlin, F. Gasparrini, A. Ciogli, D. Spinelli, B.J. Cosimelli,
Determination of the absolute configuration of a chiral oxadiazol-3-one cal-
cium channel blocker, resolved using chiral chromatography, via concerted
density functional theory calculations of its vibrational circular dichroism,
electronic circular dichroism and optical rotation, J. Org. Chem. 72 (2007)
4707–4715.
[33] Y. Ding, X.-C. Li, D. Ferreira, Theoretical calculation of electronic circular dichro-
ism of the rotationally restricted 3,8-biflavonoid morelloflavone, J. Org. Chem.
72 (2007) 9010–9017.
[34] N. Berova, L.D. Bari, G. Pescitelli, Application of electronic circular dichroism in
configurational and conformational analysis of organic compounds, Chem. Soc.
Rev. 36 (2007) 914–931.
[35] Y. Ding, X.-C. Li, D. Ferreira, 4-Arylflavan3-ols as proanthocyanidin models:
absolute configuration via density functional calculation of electronic circular
dichroism, J. Nat. Prod. 73 (2010) 435–440.
[36] Y. Ding, X.-C. Li, D. Ferreira, Absolute configuration via density functional cal-
culations of electronic circular dichroism, J. Nat. Prod. 72 (2009) 327–335.
[7] N.S. Cetin, N. Ozmen, Use of organosolv lignin in phenol–formaldehyde resins
for particleboard production. I. Organosolv lignin modified resins, Int. J. Adhes.
Adhes. 22 (2002) 477–480.
[8] J. Jankun, S.H. Selman, R. Swiercz, E. Skrypczak-Jankun, Why drinking green tea
could prevent cancer, Nature 387 (1997) 561.
[9] S. Garbisa, S. Biggin, N. Cavallarin, L. Sartor, R. Benelli, A. Albini, Tumor invasion:
molecular shears blunted by green tea, Nat. Med. 5 (1999) 1216.
[10] M. De Lorgeril, P. Salen, Wine ethanol, platelets and Mediterranean diet, Lancet
353 (1999) 1067.
[11] M.A Smith, G. Perry, P.L. Richey, L.M. Sayre, V.E. Anderson, M.F. Beal, N. Kowall,
Oxidative damage in Alzheimer’s, Nature 382 (1996) 120–121.
[12] A.M. Mendoza-Wilson, D. Glossman-Mitnik, Theoretical study of the molecular
properties and chemical reactivity of (+)-catechin and (−)-epicatechin related
to their antioxidant ability, Mol. Struct.: Theochem. 761 (2006) 97–106.
[13] H. Schroeter, C. Heiss, J. Balzer, P. Kleinbongard, C.L. Keen, K. Hollenberg, H.
Sies, C. Kwik-Uribe, H.H. Schmitz, M. Kelm, (−)-Epicatechin mediates beneficial
effects of flavanol-rich cocoa on vascular function in humans, Proc. Natl. Acad.
Sci. U.S.A. 103 (2006) 1024–1029.
[14] R.E. Kreibich, R.W. Hemingway, Condensed tannin–resorcinol adducts in lam-
inating adhesives, Forest Prod. J. 35 (1985) 23–25.
[15] K. Forest, P. Wan, C.M. Preston, Catechin and hydroxybenzhydrols as mod-
els for the environmental photochemistry of tannins and lignins, Photochem.
Photobiol. Sci. 3 (2004) 463–472.
[16] P. Wan, D.W. Brousmiche, C.Z. Chen, J. Cole, M. Lukeman, M. Xu, Quinone
methide intermediates in organic photochemistry, Pure Appl. Chem. 73 (2001)
529–534.
[17] J.H. Van der Westhuizen, D. Ferreira, D.G.J. Roux, Synthesis of condensed
tannins. Part 2. Synthesis of photolytic rearrangement. Stereochemistry, and
circular dichroism of the first 2,3-cis-3,4-cis-4-arylflavan-3-ols, J. Chem. Soc.
Perkin Trans. I (1981) 1220–1226.