1430 J. Agric. Food Chem., Vol. 51, No. 5, 2003
Bunzel et al.
minutes. The reaction mixture was stirred for a further 5 min. Na2SO3
(solid, 120 mg) was added, and stirring of the mixture was continued
for 10 min. To acetylate the phenols acetic anhydride (10 mL) was
added. The mixture was stirred for 2 h. After evaporation of solvents
on a rotary evaporator, the mixture was partitioned between CH2Cl2
and saturated aqueous NaCl. The organic layer was washed consecu-
tively with 2 M HCl, 0.4 M NaHCO3, and saturated aqueous NaCl,
then dried over MgSO4 and evaporated to dryness to provide 198 mg
of product (yield 85%), which was crystallized from acetone/petroleum
ether as red needles: mp 104 °C. Diacetate of compound 4C1SS [X )
Et, diethyl 2,3-bis(4-acetoxy-3,5-dimethoxybenzylidene)succinate]: 1H
NMR δ 1.11 (6H, t, J ) 7.1, CH3-CH2-O), 2.20 (6H, s, OCOCH3),
3.73 (12H, s, OMe), 4.14 (4H, q, J ) 7.1, CH3-CH2-O), 6.90 (4H, s,
H-2/6), 7.85 (2H, s, H-7); 13C NMR δ 14.4 (CH3-CH2-O), 20.2
(OCOCH3), 56.4 (OMe), 61.7 (CH3-CH2-O), 107.3 (C-2/6), 129.1 (C-
8), 130.8 (C-4), 133.9 (C-1), 142.5 (C-7), 153.2 (C-3/5), 167.0 (C-9),
168.3 (OCOCH3).
123.3 (A C-1), 123.9 (A C-8), 124.6 (A C-6), 134.2 (B C-1), 135.8 (B
C-4), 138.4 (A C-7), 142.8 (A C-4), 146.4 (A C-5), 148.49 (B C-3/5),
148.59 (A C-3), 167.5 (A C-9), 172.8 (B C-9).
Compound 4C3SS (X ) Me) (100 mg) was dissolved in 2 M NaOH
(20 mL) and hydrolyzed at room temperature for ∼16 h. The solution
was acidified with 2 M HCl to pH 2 and extracted into EtOAc. The
organic layer was dried over MgSO4, evaporated, and submitted to
preparative silica TLC in EtOAc/CHCl3/formic acid (2:1:0.1). EtOAc
was used to extract products from the scraped-off TLC silica. The
slower moving material was identified as compound 5C3SS (56 mg,
yield 60%). Compound 5C3SS was crystallized from acetone/petroleum
ether as gray granules, mp 230 °C. Compound 5C3SS: 1H NMR δ
3.40 (1H, d, J ) 1.3, B H-8), 3.62 (3H, s, A5-OMe), 3.69 (6H, s,
B3/5-OMe), 3.88 (3H, s, A3-OMe), 5.05 (1H, bs, B H-7), 6.40 (2H, s,
B H-2/6), 6.95 (1H, s, A H-2), 7.68 (1H, s, A H-7); 13C NMR δ 40.3
(B C-7), 47.3 (B C-8), 56.58 (A3-OMe), 56.64 (B3/5-OMe), 60.4 (A5-
OMe), 106.1 (B C-2/6), 108.9 (A C-2), 124.07 (A C-1), 124.11 (A
C-8), 125.0 (A C-6), 134.6 (B C-1), 135.7 (B C-4), 138.3 (A C-7),
142.7 (A C-4), 146.4 (A C-5), 148.45 (B C-3/5), 148.52 (A C-3), 168.6
(A C-9), 173.4 (B C-9); high-resolution MS, found M+ 446.1219,
C22H22O10 requires M 446.1213.
The faster moving fraction (14 mg, yield 27%) was identified by
NMR and GC-MS as compound 5C3′S (6-hydroxy-5,7-dimethoxy-
naphthalene-2-carboxylic acid). Compound 5C3′S: 1H NMR δ 3.99
(3H, s, A5-OMe), 4.01 (3H, s, A3-OMe), 7.30 (1H, s, A H-2), 7.91
(1H, dd, J ) 8.7, 1.6, B H-8), 8.00 (1H, d, J ) 8.7, B H-7), 8.48 (1H,
d, J ) 1.6, A H-7); 13C NMR δ 56.4 (A3-OMe), 60.9 (A5-OMe), 104.1
(A C-2), 121.6 (B C-7), 124.2 (B C-8), 126.8 (A C-4), 127.8 (A C-6),
128.2 (A C-8), 130.6 (A C-7), 141.05 (A C-5), 141.11 (A C-1), 150.8
(A C-3), 168.1 (A C-9); high-resolution MS, found M+ 248.0692,
C13H12O5 requires M 248.0685. Methods for the synthesis of compound
5C3′S described in the literature (27, 28) are time-consuming, requiring
five or six steps. To produce small amounts, the procedure described
here seems to be more convenient.
The diacetate of 4C1SS (X ) Et) (80 mg) was dissolved in dioxane
(5 mL), and degassed 2 M NaOH (20 mL) and additional dioxane (5
mL) were added. Deacetylation was carried out at room temperature
for 16 h. The solution was acidified (pH 2) with 2 M HCl and parti-
tioned between EtOAc and saturated aqueous NaCl. The organic layer
was dried over MgSO4 and evaporated. After degassed 2 M NaOH
(20 mL) had been added, ester hydrolysis was carried out for 24 h at
room temperature in the dark. This two-step saponification was neces-
sary because the acetylated derivative of diethyl ester 4C1SS (X )
Et) was not completely dissolved in NaOH without dioxane and
dissolved in dioxane/NaOH, but these conditions removed only the
acetate groups. Following the saponification process, the solution was
acidified with 2 M HCl and extracted into EtOAc. The organic layer
was dried over MgSO4, evaporated, and submitted to preparative silica-
TLC in EtOAc/CHCl3/formic acid (1:1:0.1) to afford compound 5C1SS
as a red solid (19 mg, yield 30%). The residual material still contained
the ester, which was therefore surprisingly difficult to saponify.
Compound 5C1SS: 1H NMR δ 3.75 (12H, s, OMe), 7.01 (4H, s, H-2/
6), 7.85 (2H, s, H-7); 13C NMR δ 56.5 (OMe), 108.8 (C-2/6), 126.4
(C-8), 126.6 (C-1), 138.8 (C-4), 142.6 (C-7), 148.5 (C-3/5), 168.6 (C-
9); high-resolution MS, M+ not sufficiently strong, seen in low-
resolution spectrum; M+ - H2O 428.1103, C22H20O9 requires M
428.1107; M+ - H2O - CO 400.1143, C21H20O8 requires M 400.1158.
Preparation of 5C3SS [Cyclic Form of 8-8-Coupled Dehydrodi-
sinapic Acid, Thomasidioic Acid, trans-7-Hydroxy-1-(4-hydroxy-3,5-
dimethoxyphenyl)-6,8-dimethoxy-1,2-dihydronaphthalene-2,3-dicarbox-
ylic Acid] (Figure 1). Ahmed et al. (24) described the synthesis of
thomasidioic acid in two steps starting from sinapic acid. Oxidation of
methyl sinapate 1S (X ) Me) with ferric chloride in aqueous acetone
gave the dimethyl ester of thomasidioic acid (4C3SS, X ) Me) as a
byproduct (25), whereas oxidative coupling of methyl sinapate in
aqueous acetone at pH 4 with H2O2 and horseradish peroxidase gave
the dimethyl ester of thomasidioic acid in 41% yield (26). Changing
the oxidation conditions of the methyl sinapate reaction with ferric
chloride gave the dimethyl ester of thomasidioic acid in good yields.
To sinapic acid (5 g) in methanol (150 mL) was slowly added acetyl
chloride (15 mL) over 15 min, and the mixture was stirred for 16 h,
producing methyl sinapate as determined by TLC. Water (30 mL) and
FeCl3‚6H2O (45 g) were added, and the mixture was stirred continuously
for another 48 h, by which time the methyl sinapate had been consumed,
as indicated by TLC. After methanol was removed by evaporation under
reduced pressure, the mixture was poured into EtOAc and washed with
3% HCl and saturated aqueous NH4Cl. The EtOAc fraction was dried
over MgSO4 and evaporated. Crude products were applied to flash
chromatography on silica gel in CHCl3/EtOAc (5:1). The dimethyl ester,
compound 4C3SS (X ) Me), was obtained in 52% yield along with
∼5% of a minor isomer, which had a cis configuration between carbons
B7 and B8. Compound 4C3SS (X ) Me, thomasidioic acid dimethyl
ester): 1H NMR δ 3.60 (3H, s, B-COOCH3), 3.62 (3H, s, A5-OMe),
3.68 (6H, s, B3/5-OMe), 3.68 (3H, s, A-COOCH3), 3.88 (3H, s, A3-
OMe), 4.02 (1H, s, B H-8), 5.00 (1H, s, B H-7), 6.36 (2H, s, B H-2/6),
6.96 (1H, s, A H-2), 7.66 (1H, s, A H-7); 13C NMR δ 40.3 (B C-7),
47.5 (B C-8), 51.9 (A-COOCH3), 52.4 (B-COOCH3), 56.58 (A3-OMe),
56.64 (B3/5-OMe), 60.4 (A5-OMe), 106.1 (B C-2/6), 109.0 (A C-2),
RESULTS AND DISCUSSION
Possible Pathways for Dimerization of Sinapate Esters.
Along with ferulic acid, the main hydroxycinnamic acid in the
plant cell wall, alkali-extractable p-coumaric acid and sinapic
acid [1S (X ) H), Figure 1] have been identified from grasses
(1-4, 18). p-Coumaric acid may cross-link cell-wall polymers
by [2+2]-cyclodimerization; radical coupling products have not
been identified (29). Radical dehydrodimerization of sinapate
esters 1S can produce only two primary products, 8-O-4- and
8-8-dehydrodimers (Figure 1). Ferulates additionally produce
8-5-, 5-5-, and 4-O-5-coupled products (5). Following the
radical coupling step, which produces intermediates 3 (quinone
methides for the 8-coupled dimers 3A-3C), re-aromatization
is achieved during subsequent reactions. For the 8-O-4-
coupled intermediate 3A re-aromatization is via elimination of
the acidic 8-proton (5). 8-5-Coupling leads to the formation
of phenylcoumaran structures 4B; with no 5-position available
for coupling, there is obviously no disinapate 8-5-dehydro-
dimer, although sinapate-ferulate cross-coupling can give 8-5-
coupled products. Saponification of the phenylcoumaran ester
4B gives variable partitioning between the phenylcoumaran acid
5B1 and the eliminated open form 5B2. As with ferulate
dimerization, the pathway from the 8-8-coupled product is less
clear. Elimination of both 8-protons from quinone methide 3C
can occur to produce the conjugated noncyclic product 4C1.
However, this cannot be the only path occurring in vivo; such
a product cannot give rise to the observed cyclic 8-8-product
5C3 or the more recently discovered diferulate tetrahydrofuran
5C2FF involving water addition to the quinone methide (23).
The finding of the cyclic product 5C3SS, complementing its
ferulate analogue 5C3FF, implies that either 4C2 or 4C3 (or
both, or possibly an alternate intermediate) are the in vivo
products. As is the case with ferulates, the nature of the 8-8-