Studies in polyphenol chemistry and bioactivity. 2. Establishment of interflavan linkage regio- and stereochemistry by oxidative degradation of an O-alkylated derivative of procyanidin B2 to (R)-(-)-2,4-diphenylbutyric acid

Article abstract of DOI:10.1021/jo000485+

The assignment of interflavan bond regio- and stereochemistry in oligomeric proanthocyanidins has in the past relied on empirical spectroscopic techniques which are influenced by the conformation of the C rings. Only recently was the 4,8-regiochemistry of procyanidin B₂ (3b) firmly established by 2-dimensional NMR methods. We describe herein the proof of 4β-stereochemistry in 3b by oxidative degradation of the derivative 3d bearing differential (O-benzyl and O-methyl) protecting groups in its 'top' and 'bottom' epicatechin moieties, to (R)-(-)-2,4-diphenylbutyric acid. The key elements of the degradative process are (1) removal of the C-3 alcohol functions through a modified Barton deoxygenation employing hypophosphorous acid as the reducing agent; (2) deprotection of the 'top' unit by hydrogenolysis, followed by exhaustive aryl triflate formation with N,N-bis-(trifluoromethanesulfonyl)aniline and DBU in DMF; (3) hydrogenolytic deoxygenation of the 'top' unit over Pearlman's catalyst with concomitant scission of the O-C2 bond; (4) selective oxidation of the 'bottom' unit with NaIO₄/RuCl₃. The hitherto unreported absolute configuration of (-)-2,4-diphenylbutyric acid was established as R by X-ray crystal structure analysis of the (R)-(+)-α-methylbenzylamine salt. As a corollary, the selectivity of hydrogenolytic and solvolytic reactions of epicatechin-derived tetrasulfonates has been investigated.

Full text of DOI:10.1021/jo000485+

J . Org. Chem. 2000, 65, 5371-5381
5371
Stu d ies in P olyp h en ol Ch em istr y a n d Bioa ctivity. 2.1
Esta blish m en t of In ter fla va n Lin k a ge Regio- a n d Ster eoch em istr y
by Oxid a tive Degr a d a tion of a n O-Alk yla ted Der iva tive of
P r ocya n id in B2 to (R)-(-)-2,4-Dip h en ylbu tyr ic Acid
Alan P. Kozikowski,*,† Werner Tu¨ckmantel,*,† and Clifford George‡
Georgetown University Medical Center, Drug Discovery Program, 3900 Reservoir Road NW,
Washington, D.C. 20007 and Laboratory for the Structure of Matter, Code 6030,
Naval Research Laboratory, 4555 Overlook Avenue SW, Washington, D.C. 20375-5341
tueckmaw@giccs.georgetown.edu
Received March 30, 2000
The assignment of interflavan bond regio- and stereochemistry in oligomeric proanthocyanidins
has in the past relied on empirical spectroscopic techniques which are influenced by the conformation
of the C rings. Only recently was the 4,8-regiochemistry of procyanidin B2 (3b) firmly established
by 2-dimensional NMR methods. We describe herein the proof of 4â-stereochemistry in 3b by
oxidative degradation of the derivative 3d bearing differential (O-benzyl and O-methyl) protecting
groups in its “top” and “bottom” epicatechin moieties, to (R)-(-)-2,4-diphenylbutyric acid. The key
elements of the degradative process are (1) removal of the C-3 alcohol functions through a modified
Barton deoxygenation employing hypophosphorous acid as the reducing agent; (2) deprotection of
the “top” unit by hydrogenolysis, followed by exhaustive aryl triflate formation with N,N-bis-
(trifluoromethanesulfonyl)aniline and DBU in DMF; (3) hydrogenolytic deoxygenation of the “top”
unit over Pearlman’s catalyst with concomitant scission of the O-C2 bond; (4) selective oxidation
of the “bottom” unit with NaIO4/RuCl3. The hitherto unreported absolute configuration of (-)-2,4-
diphenylbutyric acid was established as R by X-ray crystal structure analysis of the (R)-(+)-R-
methylbenzylamine salt. As a corollary, the selectivity of hydrogenolytic and solvolytic reactions of
epicatechin-derived tetrasulfonates has been investigated.
In tr od u ction
was undertaken on the basis of small chemical shift
differences between 6-H of 4,8-linked units and 8-H of
4,6-linked units, as well as other empirical criteria for
which there were neither firmly known structures as
anchoring points nor compelling reasons why the ob-
served parameters had to be influenced by the interflavan
regiochemistry in the way they were postulated to be.6
Only recently have the hypothesized interflavan regio-
chemistries for three more dimeric procyanidins and one
trimer been confirmed by 2D-NMR techniques,7 and the
regiochemistry of procyanidin B3 has been reconfirmed
in the same manner.8 In a different approach, NOE's
were examined between the A2 ring protons and the
adjacent methoxyl protons in the methyl ether acetate
derivatives of fisetinidol-4,6- and -4,8-catechin isomers
(fisetinidol ) 6-deoxycatechin).9 It is our hope that such
rational methods will in future be routinely utilized.
Certainly, the fortuitous confirmation of previously con-
jectured structures should not be construed as a vindica-
tion of questionable empirical rules.
Proanthocyanidins (condensed or nonhydrolyzable tan-
nins) constitute an important group of natural products.1
Although crude preparations containing such compounds
have been known for some time, the isolation and
characterization of individual compounds had to await
the arrival of modern chromatographic and spectroscopic
techniques. As late as 1960, two equally wrong structures
of the simple dimer, epicatechin-4â,8-epicatechin (3b,
procyanidin B2, or then referred to as leucocyanidin 1),
were discussed which contained a supernumerary oxygen
atom and a hemiacetal linkage between the two epicat-
echin units.2 The accepted gross structures of dimeric
proanthocyanidins were subsequently deduced by NMR
spectroscopy3 and by synthesis;4 however, the assignment
of interflavan bond regio- and stereochemistry proved to
be a difficult task. The interflavan bond in procyanidin
B3 (catechin-4R,8-catechin) was located unequivocally by
synthesis of a derivative from 8-bromo-5,7,3′,4′-tetra-O-
methylcatechin,4b in which the position of the Br sub-
stituent was later established by X-ray crystallography.5
In other cases, assignment of interflavan regiochemistry
The stereochemical issue remains a challenge. Exten-
sive use has been made of the coupling constant between
† Georgetown University Medical Center.
(6) The disturbing situation is reviewed in the Introductions of refs
7a and 8.
‡ Naval Research Laboratory.
(1) Part 1: Tu¨ckmantel, W.; Kozikowski, A. P.; Romanczyk, L. L.,
(7) (a) Balas, L.; Vercauteren, J . Magn. Reson. Chem. 1994, 32, 386.
(b) Balas, L.; Vercauteren, J .; Laguerre, M. Magn. Reson. Chem. 1995,
33, 85. (c) Khan, M. L.; Haslam, E.; Williamson, M. P. Magn. Reson.
Chem. 1997, 35, 854.
(8) De Bruyne, T.; Pieters, L. A. C.; Dommisse, R. A.; Kolodziej, H.;
Wray, V.; Domke, T.; Vlietinck, A. J . Phytochemistry 1996, 43, 265.
(9) Steynberg, J . P.; Brandt, E. V.; Ferreira, D.; Helfer, C. A.;
Mattice, W. L.; Gornik, D.; Hemingway, R. W. Magn. Reson. Chem.
1995, 33, 611.
J r. J . Am. Chem. Soc. 1999, 121, 12073. See here for
a general
introduction of plant polyphenols, especially condensed tannins.
(2) Forsyth, W. G. C.; Roberts, J . B. Biochem. J . 1960, 74, 374.
(3) Geissman, T. A.; Dittmar, H. K. F. Phytochemistry 1965, 4, 359.
(4) (a) Creasy, L. L.; Swain, T. Nature 1965, 208, 151. (b) Weinges,
K.; Perner, J .; Marx, H.-D. Chem. Ber. 1970, 103, 2344.
(5) Engel, D. W.; Hattingh, M.; Hundt, H. K. L.; Roux, D. G. J . Chem.
Soc., Chem. Commun. 1978, 695.
10.1021/jo000485+ CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/03/2000

5372 J . Org. Chem., Vol. 65, No. 17, 2000
Kozikowski et al.
Sch em e 1
F igu r e 1. Strategy for the defunctionalization and oxidative
degradation of a protected epicatechin dimer.
defunctionalizing and degrading the parent structure to
a simpler compound that is either known or can be
resynthesized from known compounds by stereochemi-
cally well-defined operations. For example, the absolute
configuration of the flavonoid natural product, (-)-
hesperetin, was established by ozonolysis, which resulted
in the formation of L-malic acid isolated as its dimethyl
ester (eq 1).12
the C-ring protons 3-H and 4-H. However, while large
(approximately 8-10 Hz) values are indicative of a
(nearly) trans-diaxial arrangement of the coupling pro-
tons, smaller values are inconclusive. Recourse to circular
dichroism is usually made in such cases.10 Circular
1
dichroism, however, like H-1H coupling, is influenced
by the C-ring conformation as well. It is thus not
surprising that assignments based on this method have
been documented that contradicted those obtained dif-
ferently.11 On a more fundamental level, any and all
assumptions of preferred conformations, whether arrived
at intuitively or through computation, must be rejected
if a rigorous proof of structure is to be obtained.
We have recently described1 the formation of a single
dimeric product 3a (Scheme 1) in the electrophilic
substitution reaction between 5,7,3′,4′-tetra-O-benzyl-
epicatechin (1a ) and the 4-alkoxylated building block 2,
which upon hydrogenolysis and acetylation yielded the
decaacetate 3c, identical with deca-O-acetylprocyanidin
B2 as described by several sources. While the 4,8-position
of the interflavan bond in the free polyphenol 3b has
previously been established by others,7c its stereochem-
istry could be only tentatively assigned as â in line with
the literature tradition and mechanistic reasoning. An
unequivocal determination of this point is long overdue,
and the present paper describes our solution of the
problem. As this investigation confirms the hitherto
conjectured 4â stereochemistry, all formulas are depicted
accordingly.
The case at hand is more complex since degradation
of all aromatic rings to carboxyl groups would destroy
the asymmetric center at C-4 of ring C1. Differentiation
of rings A1 and A2 was therefore required with the aim
of preserving ring A1, and the final oxidation step would
have to be conducted under carefully controlled condi-
tions to achieve the required selectivity. Moreover, the
presence of multiple ether functionality on all aromatic
rings, which activates them toward oxidation, was deemed
detrimental to the control of their oxidation and would
furthermore complicate the structure of the degradation
product. Last, initial removal of the C-3 alcoholic hy-
droxyls was also considered prudent, as these functions,
too, increase structural complexity and additionally
might participate in elimination reactions or undergo
oxidation themselves depending on the oxidant and
protecting group used. Deactivation of ring A1 was
envisaged by introduction of an electron-withdrawing
substituent,13 but it soon transpired in model studies
(vide infra) that exhaustive derivatization of a dimeric
procyanidin with substituents that permit a subsequent
deoxygenation was difficult. As an alternative, differen-
tiation between the two epicatechin units in the required
sense could be achieved by maintaining the phenolic
hydroxyl groups of the “bottom” unit in a suitably
protected form while deoxygenating rings A1 and B1. The
use of identical phenol protection for the A and B rings
within each epicatechin unit, while inconsequential for
Str a tegy
Our attempts at obtaining crystalline derivatives of
procyanidin B2 have uniformly met with failure.1 We
therefore made recourse to the traditional approach of
(12) Arakawa, H.; Nakazaki, M. Liebigs Ann. Chem. 1960, 636, 111.
For similar examples, see: Arakawa, H.; Nakazaki, M. Chem. Ind.
(London) 1959, 671; Nakazaki, M. Chem. Ind. (London) 1962, 1577;
Goering, H. L.; Kimoto, W. I. J . Am. Chem. Soc. 1965, 87, 1748.
(13) Bromination of 3a with 1 equiv of NBS (compare ref 1) yielded
mostly the monobromide bearing the substituent at C-8 of ring A1,
which could have been used as a starting material: Kozikowski, A.
P.; Tu¨ckmantel, W., unpublished.
(10) (a) Barrett, M. W.; Klyne, W.; Scopes, P. M.; Fletcher, A. C.;
Porter, L. J .; Haslam, E. J . Chem. Soc., Perkin Trans. 1 1979, 2375.
(b) Botha, J . J .; Young, D. A.; Ferreira, D.; Roux, D. G. J . Chem. Soc.,
Perkin Trans. 1 1981, 1213.
(11) Hundt, A. F.; Burger, J . F. W.; Steynberg, J . P.; Steenkamp, J .
A.; Ferreira, D. Tetrahedron Lett. 1990, 31, 5073 and references quoted
therein.

Polyphenol Chemistry and Bioactivity
J . Org. Chem., Vol. 65, No. 17, 2000 5373
Sch em e 2
the goal at hand, was dictated by practical considerations.
The resulting overall plan is summarized in Figure 1.
Resu lts
P r ep a r a tion of th e Sta r tin g Ma ter ia l. Based on
availability from the literature and our previous work,
the obvious choice of phenol protecting groups was benzyl
for the “top” and methyl for the “bottom” epicatechin
moiety. (+)-Catechin was O-methylated14 and the C-3
stereochemistry of the product 4 inverted by the same
oxidation-reduction sequence as previously described in
the O-benzyl series1 to produce 5,7,3′,4′-tetra-O-methyl-
epicatechin (1b) in good yield and purity (eq 2). Reaction
1,15
of 1b with 2 mediated by TiCl4
gave, after HPLC
separation, a 67% yield of the desired “dimer” 3d (Scheme
1) as a single isomer besides 14% of a “trimer”, in all
likelihood the isomer 5 with two 4â,8 interflavan link-
ages. We ascertained that 3d and 3b contain the same
type of interflavan linkage by hydrogenolysis of 3d to the
partially deprotected compound 3e which upon meth-
ylation furnished the same octa-O-methyl derivative 3f
as obtained by methylation of procyanidin B2 (3b).16
with NaH, CS2, and MeI in DMF, a solvent not commonly
employed for this purpose (Scheme 2). n-Bu3SnH/AIBN
reduction of 6 gave the deoxygenated product 7 in the
same yield as reported in the literature. However,
product purification was complicated by the formation
of several percent of an isomeric byproduct to which we
assign structure 8 on the basis of its spectroscopic
properties.16 In the case of the “dimer” 3d , thioacylation
was performed efficiently with PhOC(S)Cl/DMAP.20 The
resulting ester 9 furnished the deoxygenation product 10
on treatment with n-Bu3SnH/AIBN in only 54% yield.
The byproducts when resubjected to the same conditions
did not give additional 10 and may conceivably result
from the contraction of one or both C rings in analogy to
the formation of 8. A 70% yield of 10 was subsequently
realized by application of the more recently introduced
reagent system, H3PO2/Et3N/AIBN,21 and the tin hydride
variant was therefore not further examined. We note in
passing that treatment of the crude product with alkaline
H2O2 in the same manner as for the oxidation of organo-
boranes destroyed all foul-smelling S- or P-containing
impurities and ensured a smooth hydrogenolysis in the
subsequent step.
Rem ova l of th e C-3 Hyd r oxyl Gr ou p s. Deoxygen-
ation of 4 was effected by means of the Barton protocol17
via xanthate 6.18 Preparation of the xanthate through the
reported phase-transfer method (CS2, MeI, n-Bu4NBr,
50% aq NaOH/CH2Cl2, 7 °C to reflux)18,19 gave an unac-
ceptably low conversion. A far better yield was obtained
P h en ol Deoxygen a tion . Model studies for this step
were conducted on the silyl ether 14, obtained from 1a
using a standard protocol (Scheme 3).22 The benzyl
groups were removed hydrogenolytically, and the inter-
mediate 12 was carried on without purification. Complete
mesylation of 12 succeeded readily with MsCl in pyri-
(14) Mehta, P. P.; Whalley, W. B. J . Chem. Soc. 1963, 5327.
(15) Kawamoto, H.; Nakatsubo, F.; Murakami, K. J . Wood Chem.
Technol. 1989, 9, 35.
(16) See Supporting Information for additional information concern-
ing this point.
(17) Review: Hartwig, W. Tetrahedron 1983, 39, 2609.
(18) Weinges, K.; Schick, H.; Zo¨llner, P. Liebigs Ann. Chem. 1992,
293.
(20) Robins, M. J .; Wilson, J . S.; Hansske, F. J . Am. Chem. Soc.
1983, 105, 4059.
(21) Barton, D. H. R.; J ang, D. O.; J aszberenyi, J . Cs. J . Org. Chem.
1993, 58, 6838.
(19) This is apparently an adaptation of Lee, A. W. M.; Chan, W.
H.; Wong, H. C.; Wong, M. S. Synth. Commun. 1989, 19, 547.
(22) Corey, E. J .; Venkateswarlu, A. J . Am. Chem. Soc. 1972, 94,
6190.

5374 J . Org. Chem., Vol. 65, No. 17, 2000
Kozikowski et al.
Sch em e 3
Sch em e 4
dine.23 Hydrogenolysis of the tetramesylate 13 at 1 bar
over Pd/C in the presence of Et3N24 was, unfortunately,
extremely sluggish; use of the more active Pearlman’s
catalyst at 5 bar for 22 h resulted in the isolation of a
mere 8% of the completely defunctionalized flavan 15
besides larger amounts of two dimesylates and several
minor products.16 Reports of the efficient hydrogenolysis
of aryl nonaflates25 and triflates26 in the presence of a
base prompted us to prepare the tetratriflate 14. Initial
experiments employing Tf2O met with limited success.16
Aryl triflates have also been prepared by reaction with
N,N-bis(trifluoromethanesulfonyl)aniline of Li phenox-
ides27 or of phenols in the presence of DMAP/Et3N.28
Combination of 12 with PhNTf2 and DBU in DMF
resulted in an exothermic reaction; under controlled
conditions, 14 was isolated in 89% yield.16 Hydrogenolysis
of 14 in the presence of Et3N proceeded efficiently at
atmospheric pressure over Pearlman’s catalyst. The
product 15 underwent near-quantitative desilylation with
n-Bu4NF to the flavanol 16, a compound which has
previously been reported in racemic form.29 Interestingly,
protection of the alcoholic hydroxyl during sulfonylation
is not necessary as demonstrated by the successful
selective sulfonylation of the phenolic hydroxyls in (+)-
catechin.16 Application of these conditions to the bis-
(TBDMS) ether derived from 3a by silylation and hydro-
genolysis did, however, not result in the isolation of a
defined octatriflate. We equally failed to effect formation
of the octakis(1-phenyl-5-tetrazolyl) ether,16 another type
of derivative that has been employed for phenol deoxy-
genation.30
At this point, we switched to the less-demanding
strategy of deoxygenating only the “top” epicatechin
moiety. Thus, the tetraphenol 17 was generated in situ
from 10 and subjected to the preceding sulfonylation
conditions. The tetratriflate 18 was formed in 69% yield
(Scheme 4). This material upon hydrogenolysis as for 14
surprisingly yielded a mixture. By extending the reaction
time, compound 20 was obtained as the major product
in which, in addition to the intended deoxygenation, a
hydrogenolytic scission of the C1 ring has occurred.
Although the crude reaction mixture contained several
unidentified trace impurities, no significant amount of a
product was detected in which the C1 ring has survived
and the C2 ring has been opened. A minor product,
however, was identified as 21 in which both C rings have
been opened. While the C1 and C2 rings are a priori
different, and the observed regioselective cleavage could
thus result from differential access of the benzylic O-C2
bonds to the catalyst surface as a consequence of the
overall geometry of the molecule, we also note that some
degree of selectivity in favor of the hydrogenolytic cleav-
age of benzyl vs 4-methoxybenzyl and 3,4-dimethoxyben-
zyl ethers has previously been observed with Pd/C
catalysts, and a good selectivity with Raney-Ni.31 In a
recent study, a more complex picture emerged in which,
even though electron-donating substituents on a benzyl
group increase and electron-withdrawing ones decrease
the rate of hydrogenolysis of related ethers, a series of
ethanediol dibenzyl ethers bearing either type of sub-
stituent on only one of the aromatic rings indiscrimi-
nately lose the unsubstituted benzyl group first,32 a result
(23) Freudenberg, K.; Rein, H. G.; Porter, J . Liebigs Ann. Chem.
1957, 603, 177.
(24) Clauss, K.; J ensen, H. Angew. Chem., Int. Ed. Engl. 1973, 12,
918.
(25) Subramanian, L. R.; Garcia Martinez, A.; Herrera Fernandez,
A.; Martinez Alvarez, R. Synthesis 1984, 481.
(26) (a) Yagupol’skii, L. M.; Nazaretyan, V. P. J . Org. Chem. USSR
1971, 7, 1016. (b) Saa´, J . M.; Dopico, M.; Martorell, G.; Garc´ıa-Raso,
A. J . Org. Chem. 1990, 55, 991.
(27) Peterson, G. A.; Kunng, F.-A.; McCallum, J . S.; Wulff, W. D.
Tetrahedron Lett. 1987, 28, 1381.
(28) Corey, E. J .; Gin, D. Y.; Kania, R. S. J . Am. Chem. Soc. 1996,
118, 9202.
(29) (a) Brown, B. R.; Shaw, M. R. J . Chem. Soc., Perkin Trans. 1
1974, 2036. (b) Bird, T. G. C.; Brown, B. R.; Stuart, I. A.; Tyrrell, A.
W. R. J . Chem. Soc., Perkin Trans. 1 1983, 1831.
(30) (a) Musliner, W. J .; Gates, J . W. J r. J . Am. Chem. Soc. 1966,
88, 4271. (b) Musliner, W. J .; Gates, J . W., J r. Organic Syntheses;
Wiley: New York, 1973; Collect. Vol. VI, pp 150-152. (c) Hussey, B.
J .; J ohnstone, R. A. W.; Entwistle, I. D. Tetrahedron 1982, 38, 3775.
(31) Horita, K.; Yoshioka, T.; Tanaka, T.; Oikawa, Y.; Yonemitsu,
O. Tetrahedron 1986, 42, 3021.
(32) Gaunt, M. J .; Yu, J .; Spencer, J . B. J . Org. Chem. 1998, 63,
4172.

Polyphenol Chemistry and Bioactivity
J . Org. Chem., Vol. 65, No. 17, 2000 5375
Sch em e 5
Sch em e 6
in turn was obtained quantitatively by NBS bromination1
of 7 (Scheme 5). Halogen-lithium exchange followed by
addition of chalcone gave a mixture of addition products
the regio- and stereochemical nature of which we did not
attempt to establish since either of the 1,2- and 1,4-
adducts, 25 and 26, would yield the same two epimeric
products upon hydrogenolysis. This expectation was
borne out; the product is a 2:3 mixture of 23 and 27 in
which the “wrong” isomer 27 predominates. As a corol-
lary, we note that this result constitutes independent
evidence for the 4,8-regiochemistry in 3, as long as one
accepts the notion that deletion of the 3-hydroxyl in going
from 4 to 7 will not change the regiochemistry of
bromination from at least 98% 8-substitution to at least
98% 6-substitution.
The oxidation of phenyl to carboxyl is currently best
performed with a large excess of NaIO4 and a catalytic
amount of RuCl3 hydrate or RuO2 hydrate in a two-phase
solvent system composed of CCl4, CH3CN, and H2O,37 or
with a related, ligand-containing catalyst in CH3CN.38
Under the latter conditions, p-methoxyphenyl groups
have been reported to react more readily than phenyl
groups. The oxidation of a phenyl group using RuCl3
requires typically 1-2 d at room temperature. Compound
23 was subjected to the RuCl3 reagent system and the
oxidation stopped after 2 h at room temperature (Scheme
6). A product mixture was obtained which for the purpose
of separation was esterified with diphenyldiazomethane.39
Repeated chromatographic steps allowed the isolation of
a small sample of levorotatory benzhydryl 2,4-diphenyl-
butyrate (29) besides 1,3-diphenyl-1-propanone and ben-
zoic acid (isolated as its ester 30).16 Subjection of the
independently synthesized 2:3 mixture of 23 and 27 to
the same reaction conditions resulted in the isolation of
a sample of benzhydryl 2,4-diphenylbutyrate which was
enriched in the opposite enantiomer. This finding defi-
nitely precludes the remote possibility that epimerization
might have occurred during the oxidation reaction under
control by the additional asymmetric center in the still
intact “bottom” epicatechin moiety, a notion which is a
which was explained by preferred adsorption of the
unsubstituted benzyl group and which parallels our
present result. The fact that mono- or oligomeric 3-fla-
vanols with their aromatic hydroxyl or alkoxyl groups
in place do not readily suffer C-ring hydrogenolysis is the
very foundation for the use of benzyl as a protecting
group for this type of compounds. The absence of this
reaction during the preparation of 15 suggests that the
oxygen attached to C-3 also exerts a protecting effect on
the O-C2 bond. O-C2 cleavage in flavanoids has previ-
ously been reported to occur by treatment with Na in NH3
or EtOH33 and with NaBH3CN/CF3COOH.34 Under forc-
ing conditions, oligomeric proanthocyanidins undergo
more readily hydrogenolysis of the interflavan bond than
of the O-C2 bond.35
It is likely that C ring hydrogenolysis could have been
avoided under conditions of homogeneous catalysis.26,27,36
From a practical point of view, however, 20 is more
desirable for our purpose than compound 19 with an
intact C1 ring, since it contains solely the C-4 asymmetric
center under scrutiny and is therefore immune from the
objection affecting compound 19 that its degradation
product may have epimerized under control of the C-2
asymmetric center during the reaction. The same fact
reduces the determination of C-4 stereochemistry to a
mere reading of an optical rotation. To get ready for the
degradation step, we only needed to remove the hydroxyl
group from 20 to form compound 23, an operation which
succeeded uneventfully using the same method as before
via triflate 22.
Oxid a tive Degr a d a tion . To procure sufficient mate-
rial for the exploration of suitable oxidation conditions,
the degradation substrate 23 was synthesized as a
mixture with its epimer 27 from the bromide 24 which
(33) (a) Clark-Lewis, J . W.; Ramsay, G. C. Austral. J . Chem. 1965,
18, 389. (b) Mayer, W.; Bauni, G.; Stolp, F. Liebigs Ann. Chem. 1960,
630, 19 and references quoted therein.
(34) Lewin, G.; Bert, M.; Dauguet, J .-C.; Schaeffer, C.; Guinamant,
J .-L.; Volland, J .-P. Tetrahedron Lett. 1989, 30, 7049.
(35) (a) Nisi, D.; Panizzi, L. Gazz. Chim. Ital. 1966, 96, 803-819.
(b) J acques, D.; Haslam, E.; Bedford, G. R.; Greatbanks, D. J . Chem.
Soc., Perkin Trans. 1 1974, 2663. (c) Foo, L. Y. Phytochemistry 1982,
21, 1741.
(37) Carlsen, P. H. J .; Katsuki, T.; Martin, V. S.; Sharpless, K. B.
J . Org. Chem. 1981, 46, 3936.
(38) (a) Chakraborti, A. K.; Ghatak, U. R. Synthesis 1983, 746. (b)
Chakraborti, A. K.; Ghatak, U. R. J . Chem. Soc., Perkin Trans. 1 1985,
2605.
(39) Smith, L. I.; Howard, K. L. Organic Syntheses; Wiley: New
York, 1955; Collect. Vol. III, pp 351-352. A preliminary experiment
indicated that the methyl ester of 28 is too volatile to be isolated
without serious losses from the large solvent volumes present after
HPLC purification.
(36) Cacchi, S.; Ciattini, P. G.; Morera, E.; Ortar, G. Tetrahedron
Lett. 1986, 27, 5541.

5376 J . Org. Chem., Vol. 65, No. 17, 2000
Kozikowski et al.
Partisil 10, 500 × 9.4 mm, flow rate 5 mL/min; column 2:
Whatman Partisil 10, 500 × 22 mm, flow rate 25 mL/min; UV
detection at 280 nm for phenol and phenol ether chromophores,
260 nm for phenyl chromophores.
priori unlikely since the asymmetric center of interest
becomes subjected to potential enolization only after the
A2 ring has been destroyed and thereby the connection
between the two centers severed. Furthermore, if enoli-
zation occurs, the highly reactive enol immediately
undergoes further oxidation, resulting in the above-
mentioned byproducts.
While the racemate resolution of 28 has been reported
in the literature,40 the absolute configuration of its
enantiomers has not. We therefore prepared the nitrile
3141 by alkylation42 of PhCH2CN, and subsequent hy-
drolysis furnished the racemic acid 28.43 Recrystallization
of the (R)-(+)-R-methylbenzylamine salt 32 effected race-
mate resolution and at the same time provided crystals
suitable for X-ray structure analysis44 which established
that the (-)-acid has R configuration. Esterification of
the free (R)-(-)-acid with Ph2CN2 gave a levorotatory
benzhydryl ester 29, as obtained above by degradation
of 23. It follows that compound 23 also has R configura-
tion, and that the interflavan linkage in 3 has â stereo-
chemistry.
Con clu sion a n d P er sp ect ives. The present paper
provides the definitive proof that procyanidin B2 pos-
sesses a 4â,8 interflavan bond. It also provides a case
study demonstrating that traditional synthetic and deg-
radative techniques, augmented by modern methodology,
continue to have a role in the solution of analytical
problems. In the course of this work, a procedure was
developed which permits the exhaustive deoxygenation
of a single flavanoid building block in a partially pro-
tected dimeric structure. Future efforts may lead to a
more efficient technique which would also be applicable
to free proanthocyanidin dimers or even larger oligomers,
thus reducing the multitude of naturally occurring
structures to a smaller number of reference compounds,
the regio- and stereochemistry of which could be estab-
lished in analogy to the present work.
5,7,3′,4′-Tetr a -O-m eth ylca tech in (4).14 A 24.7 g (85.1
mmol) sample of (+)-catechin was dried overnight in an oil
pump vacuum at 90 °C, dissolved in 450 mL of acetone which
had been freshly distilled over B2O3,45 and added to 73 g (0.53
mol) of anhydrous K2CO3. After addition of 33 mL (0.35 mol)
of Me2SO4, the mixture was agitated with an overhead stirrer
and heated to reflux under N2 for 6 h. After cooling, solids were
removed by suction filtration over Celite and washed with 3
× 50 mL of acetone. The combined solutions were concentrated
to approximately 250 mL, and then 500 mL of 1 N aq NaOH
was added. Crystallization occurred readily. The precipitate
was recovered by suction filtration, washed with 50 mL of H2O,
pressed dry, and recrystallized from 250 mL of MeOH (reflux
to room temperature to 0 °C). Suction filtration, washing with
15 + 30 mL of cold (-20 °C) MeOH, and drying in vacuo gave
19.5 g of colorless needles. The acetone/NaOH aqueous phase
of the initial crystallization was partially evaporated to remove
acetone whereon additional product precipitated. This material
was isolated by suction filtration, washed with plenty of water,
and together with the mother liquor of the first recrystalliza-
tion recrystallized from MeOH to yield another 3.1 g of 4.
Finally, CC of the evaporated mother liquor on SiO2 with
EtOAc/hexane 1:1 furnished another 0.8 g of 4 (total yield:
23.4 g, 79%): 1H NMR δ 7.03-6.95 (m, 2 H), 6.89 (d, 1 H, J )
8 Hz), 6.14, 6.11 (ABq, 2 H, J ) 2 Hz), 4.65 (d, 1 H, J ) 8.5
Hz), 4.05 (m, 1 H), 3.89 (s, 6 H), 3.80 (m, 3 H), 3.75 (m, 3 H),
3.06, 2.58 (ABq, 2 H, J ) 16.5 Hz, both parts d with J ) 6 and
9 Hz, respectively), 1.80 (d, 1 H, J ) 3.5 Hz); 13C NMR (CDCl3,
TMS) δ 159.67, 158.69, 155.25, 149.30, 130.22, 119.91, 111.17,
109.91, 101.63, 92.96, 91.87, 81.75, 68.22, 55.93, 55.87, 55.46,
55.32, 27.62.
(2R)-5,7,3′,4′-Tetr a m eth oxyfla va n -3-on e. To 18.9 g (54.6
mmol) of 4 in 270 mL of anhydrous CH2Cl2 was added at room
temperature all at once 27.8 g (65.5 mmol, 1.2 equiv) of Dess-
Martin periodinane.46 Approximately 20 mL of water-saturated
CH2Cl2 was added dropwise within 70 min. TLC control (SiO2,
EtOAc/hexane 1:1) showed the absence of starting material.
After another 10 min, 350 mL of saturated NaHCO3 solution
was added, followed by a solution of 14 g of Na2S2O3‚5H2O in
140 mL of water. The phases were separated, and the aqueous
phase was extracted with 20 mL of CH2Cl2. The combined
organic phases were, without further treatment, filtered over
SiO2 with CHCl3/EtOAc 9:1. The eluate was evaporated and
dried in vacuo to obtain 16.2 g (86%) of the ketone as a cream-
colored solid sufficiently pure to be used directly in the
following step. An aliquot was further purified by recrystal-
lization from CH2Cl2/hexane: mp 108 °C; [R]D +27.7°, [R]546
Exp er im en ta l Section
Gen er a l P r oced u r es. Chemicals: Anhydrous K2CO3 refers
to material finely ground in a mortar and dried in an oil pump
vacuum at 150 °C for 2-3 h. Pearlman’s catalyst (20%
Pd(OH)2/C) was obtained from Aldrich and contained up to
1
50% H2O. For others, see ref 1. H, 13C, and 19F NMR spectra
were acquired at nominal frequencies of 300, 75, and 282 MHz,
respectively, in CDCl3 unless specified otherwise. 1H NMR
spectra are referenced to internal TMS; 13C NMR spectra to
internal TMS if so marked, otherwise to the CDCl3 signal (δ
77.00); 19F NMR spectra to internal CFCl3. Combustion
analyses: Micro-Analysis, Inc. (Wilmington, DE). Column
chromatography (CC): Merck silica gel 60 (No. 7734-7),
particle size 63-200 µm. TLC: Merck silica gel 60 F254 (No.
7734-7), layer thickness 250 µm; visualization by UV light or
with alkaline KMnO4 solution. HPLC: column 1: Whatman
1
+37.0° (EtOAc, c 11.5 g L-1); H NMR δ 6.93-6.87 (m, 2 H),
6.84 (d, 1 H, J ) 8 Hz), 6.30, 6.18 (ABq, 2 H, J ) 2 Hz), 5.28
(s, 1 H), 3.86 (s, 3 H), 3.85 (s, 3 H), 3.79 (s, 6 H), 3.60, 3.46
(ABq, 2 H, J ) 21.5 Hz); 13C NMR (CDCl3, TMS) δ 205.22,
160.36, 157.96, 154.62, 149.18, 149.04, 127.46, 119.28, 111.02,
109.74, 101.45, 94.40, 93.10, 83.19, 55.84, 55.81, 55.50, 55.43,
33.59. Anal. Calcd for C19H20O6: C, 66.27; H, 5.85. Found: C,
66.10; H, 5.88.
5,7,3′,4′-Tetr a -O-m eth ylep ica tech in (1b). In a 1 L three-
necked flask equipped with a magnetic stirrer, two dropping
funnels, and a N2 balloon, 21.5 g (248 mmol, 5.3 equiv) of
anhydrous LiBr (dried at 150 °C in an oil pump vacuum for
1.8 h) was dissolved in 60 mL of anhydrous THF. The solution
was cooled in an ice bath, and 61 mL (1.3 equiv) of a 1 M
solution of lithium tri-sec-butylborohydride (L-Selectride) was
added. The mixture was cooled in an acetone/CO2 bath, and
16.15 g (46.9 mmol) of crude (2R)-5,7,3′,4′-tetramethoxyflavan-
3-one in 240 mL of anhydrous THF was added dropwise within
80 min. Stirring was continued at -78 °C for 80 min. After
removal of the cold bath, 220 mL of 2.5 M aqueous NaOH was
(40) (a) Schaltegger, H.; Moser, K.; Hunkeler, W.; Wenger, T. Chimia
1967, 21, 91. (b) See also: Baker, R. H.; J enkins, S. H., J r. J . Am.
Chem. Soc. 1949, 71, 3969.
(41) Philpott, P. G.; Barltrop, J . A. J . Chem. Soc. 1956, 691.
(42) Butler, D. E.; Pollatz, J . C. J . Org. Chem. 1971, 36, 1308.
(43) Newman, M. S. J . Am. Chem. Soc. 1940, 62, 870.
(44) C24H27NO2; Mr 361.47; colorless monoclinic rods; crystal size
0.54 × 0.16 × 0.09 mm3; space group C2; T 293(2) K; a ) 30.288(2) Å,
b ) 6.156(1) Å, c ) 11.686(1) Å, â ) 108.21(1)°; V ) 2065.9(2) Å3; Z )
4; Dcalcd 1.162 gcm-3; absorption coefficient 0.572 mm-1; F(000) ) 776;
data collection: 3.08° e Θ e 57.50°; 1884 independent reflections; R1
) Σ[|Fo| - |Fc|]/Σ|Fo| ) 0.0316, wR2 ) [Σ[w(Fo - Fc2)2]/Σ[w(Fo2)2]]1/2
)
2
0.0819, S ) {Σ[w(Fo2 - Fc2)2]/(n - p)}1/2 ) 1.06 [I > 2σ(I)]; R1 ) 0.0347,
wR2 ) 0.0844, S ) 1.06 (all data); GOF ) 1.058; 254 refined
parameters; maximum peak/hole in final ∆F map 0.089/-0.105 eÅ3.
See Supporting Information for further details.
(45) Burfield, D. R.; Smithers, R. H. J . Org. Chem. 1978, 43, 3966.
(46) (a) Dess, D. B.; Martin, J . C. J . Org. Chem. 1983, 48, 4155. (b)
Ireland, R. E.; Liu, L. J . Org. Chem. 1993, 58, 2899.

Polyphenol Chemistry and Bioactivity
J . Org. Chem., Vol. 65, No. 17, 2000 5377
added (slowly at first until H2 evolution abated). The flask was
placed in a room temperature water bath, and a mixture of
55 mL of 35% aqueous H2O2 and 165 mL of EtOH was added
dropwise within 70 min. Stirring in the water bath was
continued overnight. A 300 mL amount of toluene was added,
the phases were separated, and the organic phase was washed
with 2 × 100 mL of brine. The combined aqueous phases were
saturated with NaCl and extracted with 100 mL of EtOAc.
The combined organic phases were dried over MgSO4 and
evaporated to leave a gellike mass. This material was dissolved
in 100 mL of boiling MeOH and allowed to crystallize at room
temperature and then in the freezer. The product was isolated
by suction filtration, washed with 2 × 20 mL of cold (-20 °C)
MeOH, and dried in vacuo to obtain 11.7 g (72%) of yellow
crystals. The evaporated mother liquor was filtered over SiO2
with EtOAc/hexane 1:1, the eluate evaporated, and the residue
recrystallized from boiling EtOAc (rt, then freezer), to obtain
an additional 1.4 g (9%) of 1b as off-white crystals: [R]D -62.8°,
[R]546 -75.8° (acetone, c 15.5 g L-1) (lit.4b [R]578 -61.0° (CHCl3,
c 20 g L-1)); 1H NMR47 δ 7.08 (s, 1 H), 7.05, 6.92 (ABq, 2 H, J
) 8.5 Hz, A part d with J ) 1.5 Hz), 6.20, 6.12 (ABq, 2 H, J )
2 Hz), 4.96 (s, 1 H), 4.28 (br s, 1 H), 3.92 (s, 3 H), 3.90 (s, 3 H),
3.80 (s, 3 H), 3.78 (s, 3 H), 2.95, 2.89 (ABq, 2 H, J ) 17.5 Hz,
both parts d with J ) 2 and 4 Hz, respectively), 1.76 (d, 1 H,
J ) 6 Hz); 13C NMR (CDCl3, TMS) δ 159.64, 159.23, 155.17,
149.09, 148.81, 130.81, 118.60, 111.15, 109.57, 100.22, 93.29,
92.16, 78.41, 66.41, 55.93, 55.43, 55.35, 28.09.
5,7,3′,4′-Tetr a -O-ben zylep ica tech in -4â,8-(5,7,3′,4′-tetr a -
O-m eth ylep ica tech in ) (3d ) a n d 5,7,3′,4′-Tetr a -O-ben zyl-
epicatech in -4â,8-(5,7,3′,4′-tetr a-O-ben zylepicatech in )-4â,8-
(5,7,3′,4′-tetr a -O-m eth ylep ica tech in ) (5). A 1.88 g (2.64
mmol) amount of 21 and 4.81 g (13.2 mmol, 5 equiv) of 1b were
dissolved in 24 mL of anhydrous THF and 30 mL of anhydrous
CH2Cl2 and cooled in an ice bath. To this mixture was added
dropwise in 5 min 5.3 mL of TiCl4 (1 M in CH2Cl2; 2 equiv).
The resulting dark red solution was stirred in the ice bath for
5 min and then at room temperature for 2.5 h. The reaction
was terminated by addition of 40 mL of saturated aq NaHCO3.
Extraction with 2 × 20 mL of CH2Cl2 (filtration over Celite
was necessary to achieve phase separation during the second
extraction) was followed by drying over MgSO4 and evapora-
tion. The residue was chromatographed on SiO2 with toluene/
EtOAc mixtures as the eluent. With an initial solvent ratio of
5:2, 2.40 g of oligomers were eluted, and then with a ratio of
5:3, 3.72 g of unreacted 1b. The “dimer” and “trimer” were
separated by preparative HPLC (column 2, EtOAc/hexane 2:3,
26 mL/min) to give 1.75 g (67%) of 3d (tR 26.7 min) and 308
mg (14%) of 5 (tR 32.7 min). Compound 3d : colorless foam;
[R]D +30.4°, [R]546 +37.1° (EtOAc, c 9.1 gL-1); 1H NMR
(selection; MR ) major rotamer, mr ) minor rotamer) δ 6.82
(d, J ) 8.5 Hz, mr), 6.76 (s, 1 H, MR), 6.70, 6.52 (ABq, 2 H,
MR, J ) 8.5 Hz, B part somewhat br), 6.36 (d, 1 H, J ) 2 Hz,
mr), 6.24 (narrow m, 1 H, MR + mr), 6.11 (s, 1 H, mr), 6.00
(d, 1 H, MR, J ) 2 Hz), 5.62 (d, 1 H, J ) 1.5 Hz), 5.48 (s, 1 H,
MR), 5.26 (s, 1 H, mr), 5.15 (s), 5.13 (s), 5.06 (s, 2 H, mr), 4.99
(s, 1 H, mr), 4.94 (s, 1 H, mr), 4.87 (s, 2 H, mr), 4.80 (s, 2 H,
MR), 4.75 (s, 1 H, MR), 4.62, 4.40 (ABq, 2 H, mr, J ) 11 Hz),
4.11 (1 H, MR), 4.02 (br d, 1 H, MR, J ) 5 Hz), 3.98-3.88 (m),
3.88 (s, 3 H, MR), 3.83 (s, 6 H, mr), 3.81 (s, 3 H, MR), 3.78 (s,
3 H, mr), 3.73 (s, 3 H, MR), 3.63 (s, 3 H, MR), 3.25 (s, 3 H,
mr), 3.06-2.80 (m, 2 H, MR + mr), 1.83 (d, 1 H, MR, J ) 6
Hz), 1.75 (d, 1 H, mr, J ) 5.5 Hz), 1.64 (d, 1 H, MR, J ) 4.5
Hz), 1.38 (d, 1 H, mr, J ) 5 Hz). Anal. Calcd for C62H58O12: C,
74.83; H, 5.87. Found: C, 74.84; H, 5.74. Compound 5:
colorless foam; [R]D +75.2°, [R]546 +91.9° (EtOAc, c 7.1 gL-1);
1H NMR (OCH3 signals only) δ 3.87, 3.83, 3.78, 3.77, 3.18, 2.94.
Anal. Calcd for C105H94O18: C, 76.72; H, 5.76. Found: C, 77.00;
H, 5.77.
over 22 mg of 20% Pd(OH)2/C for 10 min. TLC (SiO2, CH2Cl2/
MeOH 11:1; Rf approximately 0.3) indicated the formation of
a single product. The catalyst was removed by filtration over
cotton and the solution evaporated. The residue was taken up
in 1 mL of 3-pentanone, the solution was added to 127 mg (0.92
mmol, 43 equiv) of anhydrous K2CO3, and the first of four 22
µL portions (total: 0.93 mmol) of Me2SO4 was added. The
mixture was stirred at 60 °C under N2, and the remaining
portions of Me2SO4 were added in 1 h intervals. After a final
1 h at 60 °C and cooling to room temperature, 0.2 mL of concd.
aq NH3 was added, and the mixture was stirred at room
temperature for 30 min. Direct filtration over SiO2 with EtOAc
followed by evaporation and drying in vacuo gave 16 mg of
crude product which was purified by preparative TLC (SiO2,
200 × 200 × 0.25 mm, EtOAc/hexane 7:3) to yield 10.7 mg
(72%) of 3f.
(b) From 3b by methylation with Me2SO4/K2CO3: A solution
of 16.3 mg (28.2 µmol) of procyanidin B2 (3b)1 in 1 mL of
3-pentanone was added to 167 mg (1.2 mmol, 43 equiv) of
anhydrous K2CO3, and the first of four 28.5 µL portions
(total: 1.2 mmol) of Me2SO4 was added. The mixture was
stirred at 60 °C under N2, and the remaining portions of
Me2SO4 were added in 1 h intervals. After a final 1 h at 60 °C
and cooling to room temperature, 0.3 mL of concd aq NH3 was
added, and the mixture was stirred at room temperature for
30 min. Direct filtration over SiO2 with EtOAc followed by
evaporation and drying in vacuo gave 21 mg of crude product
which was purified by preparative HPLC (column 1, EtOAc/
hexane 7:3; tR 21.7 min) to yield 7.9 mg (41%) of 3f as a
colorless film: 1H NMR ( two rotamers in an approximately
1:1 ratio) δ 7.23 (s, 0.5 H), 7.13-6.97 (m, 2.5 H), 6.93-6.83
(m, 2 H), 6.77, 6.58 (ABq, 1 H, J ) 8 Hz), 6.28, 6.07, 5.82, 5.53
(each d, 0.5 H, J ) 2 Hz), 6.25, 6.09, 5.47, 5.36, 5.12, 4.91,
4.66, 4.27 (each s, 0.5 H), 4.47 (br s, 0.5 H), 4.03-3.95 (m, 1.5
H), 3.87 (s, 10.5 H), 3.93, 3.90, 3.82, 3.81, 3.72, 3.61, 3.54, 3.40,
3.30 (each s, 1.5 H), 3.05, 2.95 (ABq, 1 H, J ) 16 Hz, B part d
with J ) 3.5 Hz), 2.90 (m, 1 H), 1.82 (d, J ) 5.5 Hz), 1.77-
1.72 (m, 2 H), 1.65 (d, J ) 4 Hz).
5,7,3′,4′-Tetr a-O-m eth yl-3-O-[[(m eth ylth io)th iocar bon yl]-
oxy]ca tech in (6).18 To 1.84 g (5.31 mmol) of 4 in 5 mL of DMF
was added all at once 0.25 g (6.4 mmol) of NaH (60% in oil).
The mixture was stirred with a powerful magnetic stirrer at
room temperature (mild exotherm) for 10 min while H2
evolved, and then it was immersed in a +10 °C water bath,
0.48 mL (8.0 mmol) of CS2 was added dropwise in 5 min, and
stirring at +10 °C was continued for 10 min. To the resulting
yellow to amber suspension was added dropwise in 5 min 0.50
mL (8.0 mmol) of MeI. Halfway through the addition, the
stirrer stopped, and the mildly exothermic addition was
continued without cooling to reduce viscosity. The reaction
mixture remained pasty and soon hardened; it was left at room
temperature for 20 min and then dissolved in 40 mL of H2O
and 60 mL of toluene. The phases were separated, and the
aqueous phase was extracted with 20 mL of toluene. The
combined organic phases were washed with 40 mL of H2O and
concentrated to 30 mL, and the residue was chromatographed
on SiO2 first with CH2Cl2/hexane 1:1 to remove a forerun and
then with CH2Cl2/EtOAc 19:1 to elute the product. Mixed
fractions with a nonpolar contaminant were resubjected to CC
(SiO2, EtOAc/CH2Cl2/hexane 1:7:12 for forerun, 1:19:0 for
product), and all product-containing fractions were combined,
evaporated, and dried in vacuo to yield 2.04 g (88%) of 6 as a
yellowish solid: 1H NMR δ 6.97-6.90 (m, 2 H), 6.83 (d, 1 H,
J ) 8 Hz), 6.19, 6.10 (ABq, 2 H, J ) 2.5 Hz), approximately
6.16 (m, 1 H, overlapping), 5.26 (d, 1 H, J ) 6.5 Hz), 3.86 (s,
3 H), 3.84 (s, 3 H), 3.77 (s, 6 H), 3.04, 2.80 (ABq, 2 H, J ) 17
Hz, both parts d with J ) 5.5 and 6.5 Hz, respectively), 2.45
(s, 3 H); 13C NMR δ 214.96, 159.82, 158.47, 154.66, 148.87,
5,7,3′,4′-Tetr a -O-m eth ylep ica tech in -4â,8-(5,7,3′,4′-tetr a -
O-m eth ylep ica tech in ) (3f).48 (a) From 3d : A solution of 21.3
mg (21.4 µmol) of 3d in a mixture of 1 mL of EtOAc and 1 mL
of MeOH was hydrogenated at 1 bar and room temperature
(48) This compound has been prepared previously but was charac-
terized not as such but rather as its diacetate: Kolodziej, H. Phyto-
chemistry 1986, 25, 1209. We omitted the acetylation as it offers no
advantage in establishing the identity of the samples derived from 3b
and 3d .
(47) Clark-Lewis, J . W.; J ackman, L. M.; Spotswood, T. M. Austr.
J . Chem. 1964, 17, 632.

5378 J . Org. Chem., Vol. 65, No. 17, 2000
Kozikowski et al.
129.81, 119.02, 110.93, 109.37, 100.22, 92.90, 91.79, 77.69,
77.48, 55.78, 55.36, 55.29, 23.24, 18.88.
3.61 (s), 3.26 (s), 2.91-2.75 (m), 2.75-2.56 (m), 2.30-1.82 (m),
1.80-1.63 (m); 13C NMR (CDCl3, TMS, excluding δ 129-126
region) δ 158.02, 157.73, 157.02, 156.87, 156.37, 156.25, 155.36,
153.39, 149.07, 149.01, 148.86, 148.42, 148.29, 148.23, 148.10,
137.61, 137.40, 137.32, 137.23, 136.18, 135.96, 135.00, 134.43,
119.66, 119.55, 118.03, 117.82, 115.62, 115.12, 114.48, 113.48,
113.34, 110.75, 110.24, 109.31, 108.84, 107.87, 107.50, 103.99,
103.95, 94.52, 93.75, 92.63, 92.31, 90.66, 88.02, 77.98, 74.84,
74.50, 71.40, 71.36, 70.06, 69.57, 69.51, 69.26, 56.89, 55.85,
55.74, 55.58, 55.49, 55.36, 38.97, 37.79, 30.56, 29.99, 26.60,
26.37, 20.33, 19.82. Anal. Calcd for C62H58O10: C, 77.33; H,
6.07. Found: C, 77.33; H, 6.09.
(2S)-5,7,3′,4′-Tetr a m eth oxyfla va n (7).18 A stirred solution
of 4.17 g (9.55 mmol) of intermediate 6 in 150 mL of anhydrous
toluene was heated to reflux under N2, and a solution of 10.3
mL (38.2 mmol) of nBu3SnH and 0.31 g (1.9 mmol) of AIBN in
80 mL of anhydrous toluene was added dropwise through the
reflux condenser over a period of 4.25 h. Reflux was continued
for 4 h. The cooled reaction mixture was directly chromato-
graphed on SiO2. A forerun was eluted using toluene, and then
a mixture of 7 and a slightly more polar impurity 8 using
toluene/EtOAc (12:1). A late fraction containing mostly 8 was
set aside for the isolation of this impurity (see Supporting
Information); from the remaining mixture, 7 was isolated by
further CC on SiO2 with EtOAc/CHCl3/hexane 1:7:12. Evapo-
ration and drying in vacuo yielded 2.52 g (80%) of 7 as a
colorless glass: 1H NMR δ 7.00-6.85 (m, 3 H), 6.14, 6.08 (ABq,
2 H, J ) 2.5 Hz), 4.91 (dd, 1 H, J ) 2, 10.5 Hz), 3.91 (s, 3 H),
3.89 (s, 3 H), 3.80 (s, 3 H), 3.76 (s, 3 H), 2.77, 2.64 (ABq, 2 H,
J ) 16.5 Hz, both parts dd with J ) 2.5, 5.5 Hz and 6.5, 11.5
Hz, respectively), 2.17, 2.02 (ABq, 2 H, J ) 13.5 Hz, both parts
dt with J ) 2.5 Hz (t), 6.5 Hz (d) and 5.5 Hz (d), 11 Hz (t),
respectively); 13C NMR δ 159.24, 158.47, 156.26, 148.96,
148.62, 134.16, 118.46, 110.97, 109.28, 103.22, 93.30, 91.28,
77.72, 55.85, 55.78, 55.32, 55.21, 29.45, 19.39.
5,7,3′,4′-Tet r a -O-b en zyl-3-O-(ter t-b u t yld im et h ylsilyl)-
ep ica tech in (11). A solution of 4.37 g (6.72 mmol) of 1a , 0.69
g (10.1 mmol, 1.5 equiv) of imidazole, and 1.42 g (9.4 mmol,
1.4 equiv) of TBDMS-Cl in 6 mL of anhydrous DMF was stirred
at room temperature in a closed flask for 19.5 h. Direct CC on
SiO2 with EtOAc/hexane 1:5, followed by evaporation and
drying in vacuo gave 5.02 g (98%) of the silyl ether as a
yellowish glass: 1H NMR δ 7.47-7.25 (m, 20 H), 7.11 (s, 1 H),
6.94, 6.90 (ABq, 2 H, J ) 1 Hz), 6.24, 6.22 (ABq, 2 H, J ) 2
Hz), 5.15 (s, 2 H), 5.14 (narrow ABq, 2 H), 5.03 (s, 2 H), 5.02,
4.98 (ABq, 2 H, J ) 11.5 Hz), 4.93 (s, 1 H), 4.18 (narrow m, 1
H), 2.86, 2.77 (ABq, 2 H, J ) 17 Hz, both parts d with J ) 4
Hz), 0.76 (s, 9 H), -0.15 (s, 3 H), -0.30 (s, 3 H); 13C NMR
(CDCl3, TMS) δ 158.49, 157.85, 155.56, 148.69, 148.29, 137.39,
137.36, 137.25, 137.01, 132.75, 128.55, 128.49, 128.41, 127.93,
127.74, 127.69, 127.63, 127.47, 127.31, 127.07, 120.07, 115.03,
114.21, 101.66, 94.49, 93.45, 78.93, 71.47, 71.36, 70.08, 69.88,
67.48, 28.38, 25.78, 18.07, -5.09, -5.12. Anal. Calcd for
(2S )-5,7,3′,4′-Te t r a k is(b e n zyloxy)fla va n -4â,8-[(2S )-
5,7,3′,4′-tetr a m eth oxyfla va n ] (10). To a solution of 519 mg
(520 µmol) of 3d and 254 mg (2.08 mmol, 4 equiv) of DMAP in
2 mL of 1,2-dichloroethane (HPLC grade, 0.002% H2O) was
added dropwise in 7 min with water cooling 252 µL (1.82 mmol,
C
49H52O6Si: C, 76.93; H, 6.85. Found: C, 77.17; H, 6.62.
3.5 equiv) of PhOC(S)Cl. The mixture was stirred in
a
resealable tube at room temperature for 1 h and then at 50
°C for 48 h. After addition of 20 mL of saturated aqueous
NaHCO3 and 50 mL of H2O, the mixture was extracted with
50 + 10 mL of CH2Cl2, and the extract was dried over MgSO4
and evaporated. CC (SiO2, EtOAc/hexane 1:2; Rf approximately
0.35) gave, after evaporation and drying in vacuo, 494 mg
(75%) of the bis[(phenoxy)thiocarbonyl] derivative 9 as a
colorless glass: 1H NMR (major rotamer only) δ 7.48-7.10 (m),
7.10-6.82 (m), 6.76 (d, 1 H, J ) 8.5 Hz), 6.67, 6.51 (ABq, 2 H,
J ) 8.5 Hz, B part br), 6.29 (s, 1 H), 6.07 (s, 1 H), 5.88 (s, 1 H),
5.81 (s, 1 H), 5.64 (d, 1 H, J ) 3.5 Hz), 5.52 (s, 1 H), 5.12 (s, 1
H), 5.07 (s, 2 H), 5.02 (s, 2 H), 4.88 (s, 2 H), 4.64, 4.45 (ABq,
2 H, J ) 11.5 Hz), 4.24 (s, 1 H), 3.92 (s, 3 H), 3.81 (s, 3 H),
3.75 (s, 3 H), 3.52 (s, 3 H), 3.24, 3.02 (ABq, 2 H, J ) 18.5 Hz,
B part d with J ) 4.5 Hz). To a solution of 532 mg (420 µmol)
of 9 in 8.4 mL of anhydrous 1,4-dioxane (distilled over Na/
benzophenone) were added at room temperature 1.46 mL (10.5
mmol, 25 equiv) of Et3N and 0.88 mL (8.4 mmol, 20 equiv) of
50% (approximately 9.6 M) aq H3PO2. The mixture was stirred
and heated under N2 to gentle reflux, and a solution of 138
mg (0.84 mmol, 2 equiv) of AIBN in 2.1 mL of anhydrous 1,4-
dioxane was added in 7 equal portions in 20 min intervals.
Reflux was continued for 1 h, and then the mixture was cooled
and without further treatment filtered over a short SiO2
column with EtOAc/hexane 1:2. The evaporation residue (0.45
g) was taken up in 25 mL of THF, and to this solution were
added sequentially 15 mL of EtOH, 10 mL of 2.5 M aq NaOH,
and 2.5 mL of 35% aq H2O2. After stirring in an room
temperature water bath for 4 h, another 2.5 mL of 35% aq
H2O2 was added, and stirring was continued overnight. Partial
evaporation was followed by addition of 10 mL of H2O,
extraction with 3 × 20 mL of EtOAc, and drying over MgSO4.
After evaporation, CC on SiO2 with EtOAc/CHCl3/hexane
1:12:7 yielded 286 mg (70%) of the product 10 which was
sufficiently pure to be used in the subsequent step. The
analytical sample was obtained by preparative HPLC (column
1, EtOAc/hexane 1:3; tR 17.1 min): [R]D +109°, [R]546 +132°
(EtOAc, c 4.6 g L-1); 1H NMR δ 7.50-7.12 (m), 7.07-6.80 (m),
6.72 (d, J ) 8.5 Hz), 6.67, 6.52 (ABq, J ) 8.5 Hz), 6.60 (s),
6.30 (s), 6.21 (s), 6.18 (s), 6.09 (s), 6.00 (s), 5.51 (s), 5.33 (d, J
) 11 Hz), 5.21 (d, J ) 12 Hz), 5.17-5.07 (m), 5.04 (s), 4.91
(presumably one line of a partially concealed d), 4.87 (s), 4.81
(s), 4.69 (d, J ) 5 Hz), 4.63, 4.49 (ABq, J ) 11.5 Hz), 4.04 (d,
J ) 8 Hz), 3.87 (s), 3.83 (s), 3.81 (s), 3.75 (s), 3.73 (s), 3.64 (s),
3-O-(ter t-Bu t yld im et h ylsilyl)-5,7,3′,4′-t et r a -O-(m et h -
a n esu lfon yl)ep ica tech in (13). A solution of 188 mg (246
µmol) of 11 in 4 mL of EtOAc was hydrogenated at 1 bar over
34 mg of 20% Pd(OH)2/C for 1 h. Filtration over cotton,
evaporation, and drying in vacuo gave 109 mg of 12 as a
colorless foam which was pure except for solvent residues: 1H
NMR (CDCl3/CD3OD 9:1, TMS) δ 8.69 (br s, 1 H), 8.63 (s, 1
H), 8.10 (br s, 1 H), 7.67 (br s, 1 H), 6.93 (d, 1 H, J ) 1.5 Hz),
6.80, 6.76 (ABq, 2 H, J ) 8 Hz, B part d with J ) 1.5 Hz),
5.98, 5.94 (ABq, 2 H, J ) 2 Hz), 4.88 (s, 1 H), 4.20 (narrow m,
1 H), 2.82, 2.69 (ABq, 2 H, J ) 16 Hz, both parts d with J )
4.5 and 4 Hz, respectively), 0.77 (s, 9 H), -0.15 (s, 3 H), -0.20
(s, 3 H). The crude intermediate 12 was dissolved in 1 mL of
anhydrous pyridine, and a solution of 0.15 mL (2.0 mmol) of
MeSO2Cl in 0.5 mL of anhydrous pyridine was added dropwise
with ice cooling and exclusion of moisture in 4 min. After
standing at 0 °C for 71 h, 0.2 mL of water was added, and the
mixture was allowed to stand for 10 min. Twenty milliliters
of CH2Cl2 and 40 mL of 0.5 M aqueous H3PO4 were added,
the phases were separated, and the aqueous phase was
extracted with 10 mL of CH2Cl2. The combined organic phases
were washed with 30 mL of water and 15 mL of saturated
aqueous NaHCO3, dried over MgSO4, and evaporated. CC of
the residue (SiO2, EtOAc/hexane 3:2), evaporation, and drying
in vacuo gave 163.5 mg (93%) of 13 as a colorless film: [R]D
1
-9.3°, [R]546 -12.2° (EtOAc, c 6.6 g L-1); H NMR δ 7.53 (s, 1
H), 7.51, 7.46 (ABq, 2 H, J ) 8.5 Hz, B part d with J ) 1.5
Hz), 6.92 (s, 2 H), 5.09 (s, 1 H), 4.29 (br s, 1 H), 3.253 (s, 3 H),
3.246 (s, 3 H), 3.238 (s, 3 H), 3.19 (s, 3 H), 3.10, 2.97 (ABq, 2
H, J ) 17 Hz, both parts d with J ) 3.5 and 3 Hz, respectively),
0.68 (s, 9 H), -0.10 (s, 3 H), -0.39 (s, 3 H); 13C NMR (CDCl3,
TMS) δ 155.73, 148.20, 147.74, 140.85, 140.56, 139.27, 126.54,
123.91, 122.57, 113.75, 109.35, 108.97, 78.49, 66.00, 38.62,
38.46, 38.33, 37.60, 29.56, 25.51, 17.83, -5.11, -5.55. Anal.
Calcd for C25H36O14S4Si: C, 41.89; H, 5.06. Found: C, 42.15;
H, 4.90.
3-O-(ter t-Bu tyld im eth ylsilyl)-5,7,3′,4′-tetr a -O-(tr iflu o-
r om eth a n esu lfon yl)ep ica tech in (14). A solution of 67.6 mg
(88.4 µmol) of 11 in 2 mL of EtOAc was hydrogenated at 1 bar
over 20 mg of 20% Pd(OH)2/C for 40 min. Filtration over cotton,
evaporation, and drying in vacuo gave a residue of crude 12
which was dissolved in 0.5 mL of anhydrous DMF. After
cooling to -45 °C and addition of 73 µL (0.49 mmol, 5.5 equiv)
of DBU, a solution of 189 mg (0.53 mmol, 6 equiv) of N,N-bis-

Polyphenol Chemistry and Bioactivity
J . Org. Chem., Vol. 65, No. 17, 2000 5379
(trifluoromethanesulfonyl)aniline was added dropwise in 2 min
under N2. The mixture was thawed to +5 °C within 80 min
and then stirred at room temperature for 3.5 h. Direct CC on
SiO2 with EtOAc/hexane 1:5 followed by HPLC (column 2,
EtOAc/hexane 1:9; tR 19.6 min) gave 71.0 mg (89%) of 14 as a
colorless oil: [R]D -15.8°, [R]546 -19.7° (EtOAc, c 11.5 g L-1);
1H NMR δ 7.58 (narrow m, 3 H), 7.00, 6.90 (ABq, 2 H, J ) 2
Hz), 5.19 (s, 1 H), 4.32 (br s, 1 H), 3.12, 3.00 (ABq, 2 H, J ) 17
Hz, both parts d with J ) 3.5 and 2.5 Hz, respectively), 0.66
(s, 3 H), -0.08 (s, 3 H), -0.44 (s, 3 H); 13C NMR (CDCl3, TMS)
δ 155.86, 148.33, 148.00, 140.43, 140.15, 127.67, 123.59,
122.11, 118.76 (q, J ) 320.5 Hz), 118.69 (q, J ) 320.5 Hz),
116.62 (q, J ) 320 Hz), 114.35, 110.10, 107.98, 78.73, 65.41,
29.51, 25.30, 17.72, -5.21, -5.83; 19F NMR: δ 104.22, 103.70,
103.62, 103.46. Anal. Calcd for C25H24O14F12S4Si: C, 32.19; H,
2.59. Found: C, 32.44; H, 2.41.
evaporation, 299 mg of crude triflate. This material was
subjected to preparative HPLC (column 2, EtOAc/hexane 1:4;
tR 17.3 min), and the eluate was evaporated and dried in vacuo
to yield 244 mg (69% over two steps) of 18 as a colorless glass:
[R]D +94.4°, [R]546 +115° (EtOAc, c 7.7 g L-1); 1H NMR δ 7.50
(s), 7.48 (s), 7.15-6.4 (br m), 6.25-5.95 (br m), 5.55-5.3 (br
m), 5.0-4.75 (br m), 4.19 (br s), 3.89 (s), 3.87 (s), 3.62 (br s),
3.33 (br s), 2.95-2.75 (br m), 2.75-2.55 (br m), 2.27-2.07 (m),
1.97 (br s); 13C NMR (CDCl3, TMS) δ 157.92, 157.1 (br), 156.3
(br), 154.7 (br), 149.0 (br), 147.0 (br), 143.30, 140.40, 139.71,
134.7 (br), 133.1 (br), 127.08, 123.82, 121.52, 119.3 (br), 118.63
(q, 3 C, J C-F ) 319 Hz), 118.5 (br), 118.44 (q, 1 C, J C-F ) 319
Hz), 110.5 (br), 109.7 (br), 109.0 (br), 108.8 (br), 106.3 (br),
105.7 (br), 104.0 (br), 89.0 (br), 88.0 (br), 78.76, 74.0 (br), 55.88,
55.75, 55.46, 37.18, 30.0 (br), 29.5 (br), 26.70, 20.14; 19F NMR
δ 104.34 (br), 103.84 (br, sh), 103.80, 103.03 (br, sh), 102.95
(br).
(2R,3R)-cis-3-[(ter t-Bu tyld im eth ylsilyl)oxy]fla va n (15).
A solution of 1.75 g (1.88 mmol) of 14 and 1.3 mL (9.4 mmol,
5 equiv) of Et3N in 10 mL of EtOAc and 30 mL of MeOH was
hydrogenated at 1 bar over 131 mg of 20% Pd(OH)2/C for 1 h.
The solution was evaporated and the residue filtered over SiO2
with EtOAc/hexane 1:15. Evaporation and drying in vacuo
gave 563 mg (88%) of 15 as a colorless oil: [R]D -46.0°, [R]546
(2S)-Flavan -4â,8-[(2S)-5,7,3′,4′-tetr am eth oxyflavan ] (19),
(2S)-8-[(1R)-1-(2-Hydr oxyph en yl)-3-ph en ylpr opyl]-5,7,3′,4′-
tetr a m eth oxyfla va n (20), a n d (R)-3,5-Dim eth oxy-2-[3-
(3,4-d im eth oxyp h en yl)p r op yl]-6-[1-(2-h yd r oxyp h en yl)-3-
p h en ylp r op yl]p h en ol (21). A solution of 228 mg (202 mmol)
of 18 and 141 µL (1.01 mmol, 5 equiv) of Et3N in 25 mL of
EtOAc was hydrogenated over 150 mg of 5% Pd/C at room
temperature and 1 bar for 2.5 h. TLC (SiO2, EtOAc/hexane
1:3) showed spots for products 19 and 20 (Rf approximately
0.41 and 0.33, respectively) in approximately equal intensities.
Another 150 mg of the catalyst was added, and the hydrogena-
tion was continued for 6.5 h. At this point, TLC (as above)
indicated that the product mixture consisted mostly of 20
besides small amounts of 19 and 21 (Rf approximately 0.22).
The catalyst was filtered off over cotton, the solution was
evaporated, and the residue was prepurified by filtration over
SiO2 with EtOAc/hexane 1:1 to yield 104 mg of a colorless film.
This mixture was separated by preparative HPLC (column 2,
EtOAc/hexane 1:3; tR for 19, 20, and 21: 17.5, 22.2, and 27.2
min, respectively). Evaporation of the fractions and drying in
vacuo yielded 2.2 mg (3%) of 19, 74.9 mg (69%) of 20, and 10.7
mg (10%) of 21 as colorless oils. Compound 19: 1H NMR δ
7.36-7.18 (m, 6 H), 6.97 (t, 1 H, J ) 7 Hz), 6.85-6.66 (m, 5
H), 6.13 (s, 1 H), 5.43 (dd, 1 H, J ) 3.5, 7.5 Hz), 4.64 (br, 1 H),
4.59 (t, 1 H, J ) 6 Hz), 3.87 (s, 3 H), 3.84 (s, 3 H), 3.77 (s, 3
H), 3.56 (br, 3 H), 2.82, 2.66 (ABq, 2 H, J ) 17 Hz, both parts
dd with J ) 2.5, 6 Hz and 6.5, 11.5 Hz, respectively), 2.50,
2.31 (ABq, 2 H, J ) 13.5 Hz, A part dd with J ) 3.5, 6.5 Hz,
B part t with J ) 7 Hz), 2.18, 1.94 (ABq, 2 H, J ) 13.5 Hz, A
part m, B part dt with J ) 6.5 Hz (d), 11.5 Hz (t)). Compound
20: [R]D +48.2°, [R]546 +56.8° (EtOAc, c 9.6 g L-1); 1H NMR δ
7.27 (br, 1 H), 7.24-7.06 (m, 5 H), 7.03-6.86 (m, 4 H), 6.75
(dd, 1 H, J ) 1, 8 Hz), 6.66 (br, 1 H), 6.12 (s, 1 H), 4.73 (br, 1
H), 4.43 (dd, 1 H, J ) 6.5, 8.5 Hz), 3.92 (s, 3 H), 3.83 (br s, 3
H), 3.81 (s, 6 H), 2.87-2.36 (m, 6 H), 2.18-1.94 (m, 2 H); 13C
NMR δ 156.84, 154.64, 149.05, 148.80, 142.63, 128.91, 128.76,
128.42, 128.11, 126.73, 125.52, 119.23, 118.89, 115.49, 110.83,
110.44, 109.35, 55.89, 55.77, 55.38, 34.60, 33.15, 32.99, 29.45,
20.02. Compound 21: [R]D +53.4°, [R]546 +64.1° (EtOAc, c 5.2
g L-1); 1H NMR δ 7.54 (dd, 1 H, J ) 1.5, 8 Hz), 7.28-7.06 (m,
6 H), 6.91 (dt, 1 H, J ) 1 Hz (d), 7.5 Hz (t)), 6.83-6.70 (m, 4
H), 6.5 (br, 1 H), 6.12 (s, 1 H), 5.3 (br, 1 H), 6.50 (dd, 1 H, J )
5.5, 8.5 Hz), 3.85 (s, 3 H), 3.84 (s, 2 H), 3.78 (s, 3 H), 2.73-
2.42 (m, 8 H), 1.77 (quint, 2 H, J ) 7.5 Hz); 13C NMR δ 157.34,
155.90, 154.45, 153.19, 148.75, 147.04, 142.37, 135.03, 128.66,
128.47, 128.22, 127.50 (with high-field shoulder), 125.68,
120.26, 120.17, 116.05, 111.64, 111.16, 110.78, 109.44, 88.72,
55.92, 55.81, 55.55, 35.16, 34.42, 33.45, 32.74, 30.66, 29.71,
22.45.
1
-55.9° (EtOAc, c 10.0 g L-1); H NMR δ 7.46-7.25 (m, 5 H),
7.15 (t, 1 H, J ) 7.5 Hz), 7.05 (d, 1 H, J ) 7.5 Hz), 6.94 (d, 1
H, J ) 8 Hz), 6.88 (t, 1 H, J ) 7.5 Hz), 5.09 (s, 1 H), 4.26
(narrow m, 1 H), 3.14, 2.77 (ABq, 2 H, J ) 3.5 and 4 Hz,
respectively), 0.72 (s, 9 H), -0.16 (s, 3 H), -0.36 (s, 3 H); 13C
NMR δ 154.39, 139.17, 129.87, 127.83, 127.52, 127.34, 126.93,
120.39, 119.44, 116.23, 79.32, 67.55, 34.09, 25.64, 17.95, -5.23,
-5.43.
(2R,3R)-cis-F la va n -3-ol (16). To a solution of 563 mg (1.65
mmol) of 15 in 5 mL of anhydrous THF was added 1.8 mL of
n-Bu4NF solution (1 M in THF). After 2 h at room temperature,
5 mL of water was added, the THF was evaporated in vacuo,
and the product was extracted into 2 × 20 mL of ether. The
combined organic phases were dried over MgSO4 and evapo-
rated, and the residue was chromatographed on SiO2 with
EtOAc/hexane 1:4. Finally, bulb-to-bulb distillation (oven 170
°C/0.65 Torr) gave 361 mg (97%) of 16 as a colorless, viscous
oil: [R]D -86.2°, [R]546 -104° (EtOAc, c 33.0 g L-1); 1H NMR δ
7.54-7.31 (m, 5 H), 7.15 (t, 1 H, J ) 7.5 Hz), 7.10 (d, 1 H, J )
8 Hz), 6.96 (d, 1 H, J ) 8 Hz), 6.92 (t, 1 H, J ) 7 Hz), 5.09 (s,
1 H), 4.30 (br s, 1 H), 3.27, 2.97 (ABq, 2 H, J ) 17 Hz, both
parts d with J ) 4 and 2 Hz, respectively), 1.79 (d, 1 H, J )
6.5 Hz); 13C NMR δ 153.93, 138.17, 130.28, 128.52, 128.03,
127.56, 126.16, 121.25, 118.79, 116.66, 78.49, 66.61, 33.34.
Anal. Calcd for C15H14O2: C, 79.62; H, 6.24. Found: C, 79.38;
H, 6.22.
(2S)-5,7,3′,4′-Tetr a k is[(tr iflu or om eth a n esu lfon yl)oxy]-
fla va n -4â,8-[(2S)-5,7,3′,4′-tetr a m eth oxyfla va n ] (18). A so-
lution of 303 mg (314 µmol) of 10 in 14 mL each of EtOAc and
MeOH was hydrogenated over 165 mg of 20% Pd(OH)2/C at
room temperature and 1 bar for 75 min. The catalyst was
filtered off over cotton and rinsed with EtOAc. The solution
was evaporated, taken up in 10 mL of toluene and again
evaporated, and dried in vacuo to give 231 mg of crude 17,
which was characterized only by its 1H NMR spectrum
(CD3OD, TMS): δ 7.05-6.75 (br, 3 H), 6.77 (d, 1 H, J ) 1.5
Hz), 6.69, 6.60 (ABq, 2 H, J ) 8 Hz, B part d with J ) 2 Hz),
6.20 (br s, 1 H), 5.77 (br, 1 H), 5.12 (br d, 1 H, J ) 9.5 Hz),
4.61 (br, 1 H), 3.80 (s, 6 H), 3.75 (s, 3 H), approximately 3.8-
3.3 (very br, presumably containing one Me group and one
C-ring H), 2.76-2.54 (m, 2 H), 2.16-2.00 (m, 2 H), 2.00-1.80
(m, 2 H); ore aromatic H too broad for detection or undergoing
rapid H-D exchange. Intermediate 17 was taken up in 1.5
mL of anhydrous DMF, and the solution was cooled in an ice
bath whereon 235 µL (1.57 mmol, 5 equiv) of DBU was added
dropwise in 5 min. After cooling to -50 °C, a solution of 505
mg (1.41 mmol, 4.5 equiv) of PhNTf2 in 2.5 mL of anhydrous
DMF was added dropwise within 12 min at -45 to -50 °C.
The temperature was allowed to rise to 0 °C over a period of
40 min. Stirring was further continued at 0 °C for 30 min and
at room temperature for 70 min. The mixture was then directly
filtered over SiO2 with EtOAc/hexane 1:3 to give, after
(2S)-8-[(1R)-1,3-Dip h en ylp r op yl]-5,7,3′,4′-t et r a m et h -
oxyfla va n (23). To a solution of 74.9 mg (139 µmol) of
intermediate 20 in 0.4 mL of anhydrous DMF was added with
stirring at -40 °C under exclusion of moisture 42 µL (0.28
mmol) of DBU and then dropwise in 7 min a solution of 74
mg (0.21 mmol) of PhNTf2 in 0.4 mL of DMF. The reaction
mixture was allowed to reach 0 °C within 25 min and then
stirred in an ice bath for 35 min and at room temperature for
1 h. After addition of 20 mL of half-saturated aq Na2CO3

5380 J . Org. Chem., Vol. 65, No. 17, 2000
Kozikowski et al.
solution, the mixture was extracted with 3 × 10 mL of EtOAc/
hexane 1:1, and the combined organic phases were dried over
MgSO4. Filtration over SiO2 with EtOAc/hexane 1:2, evapora-
tion, and drying in vacuo gave 87.7 mg (94%) of the triflate
22 as a colorless glass which was used without further
purification in the following step: 1H NMR δ 7.45 (dd, 1 H, J
) 1.5, 8 Hz), 7.24-7.06 (m, 7 H), 6.93 (dt, 1 H, J ) 1 Hz (d),
7.5 Hz (t)), 6.87 (s, 1 H), 6.82 (s, 2 H), 6.12 (s, 1 H), 4.94 (dd,
1 H, J ) 5.5, 8 Hz), 4.60 (dd, 1 H, J ) 2, 10 Hz), 3.89 (s, 3 H),
3.82 (s, 3 H), 3.78 (s, 3 H), 3.66 (s, 3 H), 2.79 (A part of an
ABq, 1 H, J AB ) 17 Hz, dd with J ) 2.5, 5.5 Hz), 2.67-2.50
(m, 4 H), 2.40-2.25 (m, 1 H), 2.12-1.90 (m, 2 H); 13C NMR δ
156.98, 156.85, 154.94, 149.93, 148.53, 148.49, 142.78, 137.22,
134.46, 131.55, 128.35, 128.13, 126.98, 126.73, 125.51, 120.00,
118.69 (q [only 2 center lines observed], J ) 319 Hz), 118.60,
110.72, 109.83, 109.21, 103.42, 87.63, 77.73, 55.89, 55.64,
55.54, 55.30, 34.44, 34.09, 29.65, 19.89. A solution of 87.7 mg
(130 µmol) of 22 and 54 µL (0.39 mmol) of Et3N in 9 mL of
EtOAc was hydrogenated at 1 bar and room temperature over
72 mg of 5% Pd/C for 3 h. Completion of the reaction was
established by HPLC (column 1, EtOAc/hexane 1:3; tR 14.2 min
(23), 15.4 min (22)). After filtration over a cotton plug and
evaporation, the residue was filtered over SiO2 with EtOAc/
hexane 1:3 to furnish 64.1 mg (94%) of 23 as a colorless glass.
The analytical sample was prepared by HPLC under the above
conditions: [R]D +21.1°, [R]546 +24.9° (EtOAc, c 6.5 g L-1); 1H
NMR δ 7.30 (d, 2 H, J ) 7.5 Hz), 7.23-7.01 (m, 8 H), 6.92-
6.81 (m, 3 H), 6.12 (s, 1 H), 4.76 (d, 1 H, J ) 10 Hz), 4.65 (dd,
1 H, J ) 6, 8.5 Hz), 3.90 (s, 3 H), 3.82 (s, 3 H), 3.77 (s, 3 H),
3.70 (s, 3 H), 2.80 (A part of an ABq, 1 H, J AB ) 17 Hz, dd
with J ) 1, 6 Hz), 2.71-2.49 (m, 4 H), 2.45-2.32 (m, 1 H),
2.20-2.09 (m, 1 H), 1.93 (ddt, 1 H, J ) 6 Hz (d), 11.5 Hz (t),
13 Hz (d)); 13C NMR δ 157.04, 156.45, 154.40, 148.86, 148.33,
145.69, 143.22, 134.74, 128.41, 128.14, 128.03, 127.55, 125.34,
125.05, 118.17, 112.74, 110.71, 109.11, 103.65, 88.33, 55.93,
55.89, 55.66, 55.30, 39.54, 34.84, 34.38, 29.73, 19.87.
The cold bath was removed, 50 mL of water was added, and
the THF was removed by partial evaporation. After extraction
of the aqueous residue with 50 + 30 mL of EtOAc, the
combined organic phases were dried over MgSO4 and evapo-
rated. The residue was chromatographed on SiO2 with EtOAc/
hexane mixtures (3:8 for unreacted chalcone and 7, 2:3 for the
desired intermediates). Evaporation of the respective fractions
yielded 0.86 g (67%) of 7 and 0.57 g (28%) of the addition
product (25 and/or 26) as a mixture of isomers. The crude
intermediate was taken up in 24 mL of EtOAc/MeOH 1:1 and
hydrogenated at 1 bar and room temperature over 0.29 g of
20% Pd(OH)2/C for 6 h. After filtration over cotton and
evaporation, the residue was chromatographed on SiO2 with
toluene/EtOAc 24:1, isolating the major, less-polar component
of the product mixture. Evaporation and drying in vacuo
yielded 367 mg (66%) of a mixture of 23 and 27 as a colorless
glass in an isomer ratio of 2:3. The analytical sample was
obtained by preparative HPLC (column 1, EtOAc/hexane 1:4;
tR 17.2 min, no separation of isomers under these conditions):
1H NMR (selected signals of 27 only) δ 4.89 (dd, 1 H, J ) 2,
10.5 Hz), 4.68 (dd, 1 H, J ) 5.5, 8.5 Hz), 3.87 (s, 3 H), 3.81 (s,
3 H), 3.733 (s, 3 H), 3.729 (s, 3 H). Anal. Calcd for C34H36O5:
C, 77.84; H, 6.92. Found: C, 78.01; H, 6.85.
Oxid a tive Degr a d a tion of 23. To 64.1 mg (122 µmol) of
23 in 1.2 mL of anhydrous CCl4, 1.2 mL of HPLC grade
CH3CN, and 1.8 mL of HPLC grade H2O were added 0.52 g
(2.4 mmol) of solid NaIO4 and 5 mg (24 µmol if dry, but water
content unknown) of RuCl3‚xH2O. The mixture was stirred
vigorously at room temperature for 2 h. After addition of 5
mL of 5% HCl, the products were extracted into 3 × 10 mL of
CH2Cl2, and the combined organic phases were dried over
MgSO4 and evaporated. The residue was taken up in 20 mL
of ether and set aside at room temperature for 15 min to allow
finely dispersed residual Ru compounds to coagulate, which
were then removed by filtration over a cotton plug. After
evaporation, the residue was taken up in 0.3 mL of THF, and
0.9 mL of MeOH was added followed by a solution of 63 mg
(0.22 mmol) of diphenyldiazomethane in 0.3 mL of THF. This
mixture was stirred at room temperature in the dark for 80
min, after which period the red color of the reagent was found
to be completely discharged. Another 90 mg (0.46 mmol) of
Ph2CN2 in 0.3 mL of THF was added, and the mixture was
kept at room temperature in the dark for 64 h. The solvent
was evaporated and the residue chromatographed on SiO2 with
CH2Cl2/hexane 1:1. The required degradation product eluted
after several minor nonpolar byproducts as part of a band of
Rf approximately 0.5-0.4 which appeared on TLC as a group
of two closely spaced spots. This material (39.5 mg) was further
fractionated by preparative TLC (SiO2, two plates of dimen-
sions 200 × 200 × 0.25 mm, EtOAc/hexane 1:15). The slower-
moving of the two major bands was isolated (8.2 mg) and
subjected to preparative HPLC (column 1; gradient of CH2Cl2
in hexane: 0-43 min, 25 to 50% CH2Cl2; 5.2 mL/min). A peak
eluting at tR 29.1 min on evaporation yielded 5.1 mg (10%) of
(2S)-8-Br om o-5,7,3′,4′-tetr a m eth oxyfla va n (24). To a
solution of 2.52 g (7.63 mmol) of 7 in 50 mL of anhydrous
CH2Cl2 was added with stirring at -70 °C all at once 1.37 g
(7.70 mmol) of recrystallized NBS. The mixture was stirred
in the slowly thawing bath for 110 min, at which point the
bath temperature was -16 °C. The cold bath was removed,
and the mixture was stirred vigorously for 25 min with a
solution of 0.5 g of Na2S2O3•5H2O in 20 mL H2O. The phases
were separated, the aqueous phase was extracted with 5 mL
of CH2Cl2, and the combined organic phases were dried over
MgSO4. After concentration, the residue was filtered over SiO2
with EtOAc/hexane 1:2. Evaporation and drying in vacuo gave
2.73 g (87%) of 24 as a colorless solid. This material was readily
recrystallized from EtOAc/hexane but tenaciously held re-
sidual EtOAc even after drying in vacuo overnight at 70 °C,
under which conditions the material turned pink: mp 145-
148 °C (to form a red liquid; capillary inserted into bath at
1
136 °C; slow heating results in broader melting interval); H
1
NMR δ 7.07 (d, 1 H, J ) 2 Hz), 6.97-6.87 (ABq, 2 H, J ) 8.5
Hz, A part dd with J ) 0.8, 2 Hz), 6.15 (s, 1 H), 5.11 (dd, 1 H,
J ) 2.5, 9.5 Hz), 3.90 (s, 6 H), 3.88 (s, 3 H), 3.83 (s, 3 H), 2.74,
2.65 (ABq, 2 H, J ) 17 Hz, both parts dd with J ) 4, 6.5 Hz
and 5.5, 10 Hz, respectively), 2.27, 1.99 (ABq, 2 H, J ) 13.5
Hz, A part ddd with J ) 2.5, 4, 6 Hz, B part dt with J ) 6 Hz
(d), 10 Hz (t)); 13C NMR δ 157.17, 155.39, 152.42, 148.91,
148.32, 133.74, 117.69, 110.95, 109.07, 105.04, 91.75, 88.61,
77.49, 56.46, 55.89, 55.84, 55.55.
Mixtu r e of (2S)-8-[(1R)-1,3-Dip h en ylp r op yl]-5,7,3′,4′-
tetr a m eth oxyfla va n (23) a n d (2S)-8-[(1S)-1,3-Dip h en yl-
p r op yl]-5,7,3′,4′-tetr a m eth oxyfla va n (27). To a solution of
1.59 g (3.89 mmol) of 24 (which has previously been evaporated
with toluene to remove residual EtOAc) in 40 mL of anhydrous
THF was added at -78 °C under N2 with stirring over a period
of 1.5 min 6.5 mL (11.0 mmol) of t-BuLi (1.7 M in pentane).
The resulting yellow solution was stirred at -78 °C for 4 min,
and then a solution of 1.13 g (5.45 mmol) of chalcone in 10
mL of anhydrous THF was added within 4 min, causing an
initial color change to greenish-black and eventually to orange.
The mixture was thawed to -10 °C over a period of 100 min.
(-)-29 the H NMR spectrum of which was identical with that
of the authentic material prepared below; [R]D -22°, [R]546
-27°, [R]435 -51° (c 2.4 g L-1, EtOAc).
r a c-2,4-Dip h en ylbu tyr on itr ile (31).41 To a suspension of
2.12 g (53 mmol) of NaH (60% in oil) in 50 mL of anhydrous
DMSO under N2 was added dropwise with stirring and cooling
(rt water bath) 5.8 mL (50 mmol) of PhCH2CN at such a rate
that the internal temperature remained in the 25-30 °C
range. The addition required 15 min. After another 10 min,
6.8 mL (50 mmol) of PhCH2CH2Br was added at an internal
temperature of 27-32 °C. This operation required 20 min.
After another 30 min at room temperature, the solution was
poured onto 90 g of ice, 200 mL of water was added, and the
products were extracted into 3 × 100 mL of ether. The
combined organic phases were washed with 50 mL of brine,
dried over MgSO4, and evaporated, and the residue was
chromatographed on SiO2 with EtOAc/hexane 1:25. The first-
eluted product (0.88 g, 5%) was identified as 2-phenethyl-2,4-
diphenylbutyronitrile based on its spectroscopic data. 1H NMR
δ 7.56-7.06 (m, 15 H), 2.87-2.73 (m, 2 H), 2.48-2.14 (m, 6
H); IR (film) 2236, 1603, 1495, 1454, 757, 701 cm-1; MS m/z

Polyphenol Chemistry and Bioactivity
J . Org. Chem., Vol. 65, No. 17, 2000 5381
325 (M+, 6%), 234, 221, 105 (100%), 91. The major fraction
was eluted subsequently and, after evaporation and drying in
vacuo, yielded 6.41 g (58%) of 31 as an oil: 1H NMR δ 7.44-
7.16 (m, 10 H), 3.74 (dd, 1 H, J ) 6, 9 Hz), 2.90-2.73 (m, 2
H), 2.35-2.10 (m, 2 H).
washed with 5 mL of 5% aq HCl, and the aqueous phase was
back-extracted with 5 mL of EtOAc. Drying over MgSO4,
evaporation, and distillation as above yielded (-)-28 as a
colorless oil: [R]D -56.8°, [R]546 -68.7° (c 24.9 g L-1, CHCl3)
(lit.:41 [R]D +57° for an enantiomer of unknown absolute
configuration (c “10%”, CHCl3)).
r a c-2,4-Dip h en ylbu tyr ic Acid ((()-28).43 A mixture of
5.01 g (22.6 mmol) of 31, 18 mL of AcOH, and 3 mL of 70% aq
H2SO4 was refluxed with exclusion of moisture for 3 d.
Volatiles were evaporated, 100 mL of ice-water was added,
and the product was extracted into 3 × 30 mL of ether. After
washing of the combined ether phases with 30 mL of ice-
water, the product was extracted into a cold solution of 3.2 g
(80 mmol) of NaOH in 60 mL of H2O. The aqueous phase was
washed with 30 mL of ether and then cooled by addition of ice
and acidified with 7 mL of concd HCl. The product was
extracted into 3 × 30 mL of ether, and the combined organic
phases were dried over MgSO4, evaporated, and briefly dried
in vacuo to yield 5.49 g (101%) of rac-28 as an amber oil. Bulb-
to-bulb distillation (oven 165 °C/oil pump vacuum) gave a
colorless oil which gradually solidified on standing: 1H NMR
δ 7.40-7.10 (m, 10 H), 3.56 (t, 1 H, J ) 7.5 Hz), 2.58 (t, 2 H,
J ) 7.5 Hz), 2.49-2.35 (m, 1 H), 2.18-2.04 (m, 1 H).
Ben zh yd r yl (R)-2,4-Dip h en ylbu tyr a te ((-)-29). To a
solution of 22.6 mg (94 µmol) of (-)-28 in 0.3 mL of THF and
0.6 mL of MeOH was added a solution of 36.5 mg (188 µmol)
of diphenyldiazomethane in 0.3 mL of THF. The mixture was
stirred in a loosely stoppered flask at room temperature in
the dark for 3 h. Initial purification was achieved by prepara-
tive TLC (SiO2, two plates, 200 × 200 × 0.25 mm, EtOAc/
hexane 1:15) which removed a close nonpolar zone. Subse-
quently, the product (39.2 mg) was subjected to preparative
HPLC (column 1, EtOAc/hexane 1:19; tR 11.4 min (impurity),
12.3 min (29)) to yield 29.1 mg of a colorless oil which,
according to GC-MS, remained contaminated with Ph2CO.
Suitable HPLC conditions were later found for the removal of
this contaminant (see above); in the present experiment, the
material was taken up in 0.4 mL of anhydrous THF and
reacted with 0.1 mL of BH3‚THF (1 M in THF) under N2 at
room temperature for 30 min. The reagent was then quenched
by dropwise addition of a mixture of 0.1 mL of MeOH and 0.5
mL of THF, the mixture was evaporated, and the residue was
chromatographed on SiO2 with EtOAc/hexane 1:12 to furnish
22.0 mg (57%) of the pure oily ester: [R]D -23.7°, [R]546 -28.9°,
[R]435 -54.2° (c 5.2 g L-1, EtOAc); 1H NMR δ 7.37-7.13 (m, 16
H), 7.11-7.04 (m, 4 H), 6.83 (s, 1 H), 3.68 (t, 1 H, J ) 7.5 Hz),
2.54 (t, 2 H, J ) 7.5 Hz), 2.50-2.36 (m, 1 H), 2.19-2.06 (m, 1
H). Anal. Calcd for C29H26O2: C, 85.68; H, 6.45. Found: C,
85.62; H, 6.30.
(R)-r-Met h ylb en zyla m m on iu m (R)-2,4-Dip h en ylb u -
tyr a te (32).40a To a boiling solution of 1.94 g (8.07 mmol) of
rac-28 in 20 mL of MeOH was added 565 µL (4.44 mmol) of
(R)-(+)-R-methylbenzylamine (Fluka, 99+%). No crystalliza-
tion occurred at room temperature overnight; after 2 d at -20
°C, the mixture was found to have set to a compact solid. This
material partially reliquefied on thawing, and subsequent
suction filtration, washing with 5 mL of MeOH, and drying in
vacuo yielded 607 mg of a colorless solid exhibiting [R]546
-12.2° (c 17.7 g L-1, MeOH). A second crop of 513 mg was
obtained by concentrating the mother liquor. Both crops
together were recrystallized from 20 mL of MeOH (reflux to
room temperature) to give 523 mg of a solid ([R]546 -14.5° (c
16.9 g L-1, MeOH)). Another recrystallization from 12 mL of
MeOH (reflux to room temperature) yielded 188 mg of a solid
([R]546 -14.3° (c 17.5 g L-1, MeOH)). The constant optical
rotation suggested that stereochemical purity had been
achieved. However, when the mother liquor of this recrystal-
lization was allowed to stand at room temperature for ap-
proximately 1 month during which period part of the solvent
evaporated, long needles of substantial diameter were formed
which exhibited [R]D -13.2°, [R]546 -15.3° (c 17.2 g L-1, MeOH).
This material was used for X-ray structure analysis and for
the preparation of (-)-28. 1H NMR (CDCl3/CD3OD approxi-
mately 8:1) δ 7.40-7.12 (m, 15 H), 4.14 (q, 1 H, J ) 6.5 Hz),
3.48 (t, 1 H, J ) 7.5 Hz), 2.67-2.50 (m, 2 H), 2.44-2.31 (m, 1
H), 2.14-2.00 (m, 1 H), 1.46 (d, 3 H, J ) 6.5 Hz).
Ack n ow led gm en t. We are grateful to MARS, In-
corporated (A.P.K., W.T.) and to ONR and NIDA (C.G.)
for funding of this work.
Su p p or tin g In for m a tion Ava ila ble: Methylation of 3b;
byproduct 8; hydrogenolysis of mesylate 13; formation and
partial hydrogenolysis of triflate 14; trifluoromethanesulfon-
ylation of (+)-catechin; deoxygenation of 12 through its
1-phenyl-5-tetrazolyl derivative; byproducts isolated from the
oxidative degradation of 23/27; preparation of a bisgallate
derived from 3e; additional Experimental Section including
(a) selected IR bands of (2R)-5,7,3′,4′-tetramethoxyflavan-3-
one and compounds 3d , 5-7, 9-11, 13-16, 18-23, (-)-29,
and 31; (b) selected MS peaks of (2R)-5,7,3′,4′-tetramethoxy-
flavan-3-one and compounds 1b, 3f, 7, 12, 15, 16, 19-24, (-)-
29, and 31; (c) tables of atom coordinates, displacement
parameters, bond lengths and angles, torsion angles, and
hydrogen bonds for compound 32 (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
(R)-(-)-2,4-Dip h en ylbu tyr ic Acid ((-)-28).40a Part of the
crystals of [R]546 -15.3° obtained above were shaken with 5
mL of 5% aq HCl and 5 mL of EtOAc, the phases were
separated, and the aqueous phase was extracted with two more
5 mL portions of EtOAc. The combined organic phases were
J O000485+