Structural modification of phenylpropanoid-derived compounds and the effects on their participation in redox processes

Article abstract of DOI:10.1016/j.bmc.2005.01.047

Oxidation and reduction processes are fundamental to many of the proposed mechanisms by which dietary phytochemicals are thought to exert protective effects against cardiovascular disease and some cancers. An understanding of the redox chemistry of these compounds is essential in assessing their potential to participate in these processes. Phenylpropanoid-derived compounds were selected and synthesised where required to represent many of the structural features found in this important group of compounds. Using electron paramagnetic resonance spectroscopy and computational chemistry a structure-redox activity relationship was obtained. Good correlation of computational and experimental results was observed for the mono-hydroxylated compounds. This demonstrated the value of computational chemistry in obtaining information about compounds, not readily available and the effect of electron delocalisation on parent radical stability. For compounds containing more than one hydroxyl, the relationship was found to be more complex. The importance of quinone formation in compounds containing more than one hydroxyl substituent was highlighted, as this was found to have a significant effect on stabilisation and therefore, their participation in redox processes.

Full text of DOI:10.1016/j.bmc.2005.01.047

Bioorganic & Medicinal Chemistry 13 (2005) 2537–2546
Structural modification of phenylpropanoid-derived compounds
and the effects on their participation in redox processes
Wendy R. Russell,* Lorraine Scobbie and Andrew Chesson
Rowett Research Institute, Greenburn Road, Bucksburn, Aberdeen AB21 9SB, UK
Received 12 October 2004;accepted 21 January 2005
Abstract—Oxidation and reduction processes are fundamental to many of the proposed mechanisms by which dietary phytochemi-
cals are thought to exert protective effects against cardiovascular disease and some cancers. An understanding of the redox chemistry
of these compounds is essential in assessing their potential to participate in these processes. Phenylpropanoid-derived compounds
were selected and synthesised where required to represent many of the structural features found in this important group of com-
pounds. Using electron paramagnetic resonance spectroscopy and computational chemistry a structure–redox activity relationship
was obtained. Good correlation of computational and experimental results was observed for the mono-hydroxylated compounds.
This demonstrated the value of computational chemistry in obtaining information about compounds, not readily available and the
effect of electron delocalisation on parent radical stability. For compounds containing more than one hydroxyl, the relationship was
found to be more complex. The importance of quinone formation in compounds containing more than one hydroxyl substituent was
highlighted, as this was found to have a significant effect on stabilisation and therefore, their participation in redox processes.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
Although there are a large variety of plant secondary
metabolites, approximately one fifth of all carbon fixed
by plants is channelled through the shikimate pathway.
This pathway is absent in mammals and therefore, the
plant metabolites produced by this pathway are an
essential component of our diet. The switch from
primary to secondary metabolism occurs with the E2
elimination of ammonia from the amino acid, L-
phenylalanine to form cinnamic acid. Cinnamic acid is
the first metabolite of the phenylpropanoid pathway
and the metabolites of this pathway produce the largest
range of natural products considered potentially protec-
tive of health (Fig. 1). Cinnamic acid substitution fol-
lows an ortho oxygenation and subsequent methylation
pattern and compounds representing both methylated
and unmethylated, mono, di and tri-hydroxylated
cinnamic acids are included in this study. Compounds
representing further transformation reactions of the
side-chain reactions have also been included. These
compounds, along with the cinnamic acids are the build-
ing blocks to lignans and the lignan-like dilignols also
selected for study. Substituted benzoic acids and their
derivatives are also included, because although these
compounds are formed directly from intermediates early
in the shikimate pathway, in plants they are more usu-
ally formed by degradation of cinnamic acid derivatives.
Also included are the cinnamic acid lactone derivatives;
Diet appears to be an important factor contributing to
cancer aetiology and in particular to colorectal cancer.1,2
Dietary phytochemicals have the potential to modulate
many of the stages involved in the development of can-
cer.3 As antioxidants they are postulated to protect bio-
molecules such as DNA, protein and lipids from oxidant
species, or by direct scavenging of carcinogens.4,5 Phyto-
chemicals may also exert anticancer effects by modulat-
ing the signal transduction pathways that control
proliferation and programmed cell death.6–8 Other
important targets are their ability to inhibit the mamma-
lian metalloenzymes involved in the arachidonic acid
cascade and other pro-oxidant enzymes such as xanthine
oxidase.9–11 Since oxidation and reduction are funda-
mental to many of the stages involved in carcinogenesis,
an understanding of how the structure of these plant-de-
rived compounds effects their redox chemistry is essen-
tial in assessing their potential to participate in these
processes.
Keywords: Redox chemistry;Phenylpropanoid;Colorectal cancer;
EPR.
*
Corresponding author. Tel.: +44 1224 712751;fax: +44 1224
716687;e-mail: wrr@rri.sari.ac.uk
0968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.01.047

2538
W.R.Russell et al./ Bioorg.Med.Chem.13 (2005) 2537–2546
the coumarins and the structurally diverse and biologi-
cally important family of compounds known collectively
as the flavonoids, which include representatives of the
flavanones, flavones, flavanols, dihydroflavanols and
catechins. In addition to all of these basic structural
types, effects of hydroxylation, methylation and glyco-
sylation have also been investigated.
this observation, it appears that compounds with
dihydroxyl substituents para to each other are more eas-
ily oxidised than those containing ortho substituents,
which are in turn more easily oxidised than those con-
taining meta substituents. The carboxylic acid side chain
had the effect of increasing the ease of oxidation, when it
was meta to the substituents. For the dihydroxylated
compounds, it was observed that all compounds with
a HOMO energy greater than ꢀ9.46 eV were oxidised.
Modification of the structure to contain one methoxyl
group at C3 1k resulted in the compound not being oxi-
dised, whereas a compound containing two methoxyl
groups at C3 and C5 1l was oxidised. Therefore, it ap-
pears that methoxyl substitution increases the ease of
oxidation, but not to the same extent as when additional
hydroxyl were substituents. To study the effect of vary-
ing the functional group at C1 in the C6C1 compounds,
a series of compounds in which the aromatic substitu-
tion pattern (4-OH, 3-OMe) remained constant were se-
lected (Table 1). For these compounds, only the alcohol
3 reduced the galvinoxyl radical. The acid 1k, and alde-
hyde 2a were not oxidised. When the pattern of substi-
tution was reversed (3-OH, 4-OMe), as was the case
for the iso-aldehyde 2b, the compound still was not
oxidised.
Electron paramagnetic resonance (EPR) spectroscopy is
the most powerful single technique for the detection and
characterisation of radicals. Using EPR we have deter-
mined the ease with which the series of phenylpropa-
noid-derived compounds donated an electron, by
measuring the extent and rate at which they reduced
the synthetic radical, galvinoxyl.12 Computational tech-
niques can be used to estimate particular properties
when compounds are not readily available and in this
paper we have compared calculations made at the
AM1 level of theory13 to the results obtained experimen-
tally. In an attempt to elucidate the more complex effects
of structural change on the redox chemistry, we then
compared results of spin density distributions for a
selected set of structurally related compounds.
2. Results
2.2. C6C3 compounds
2.1. C6C1 compounds
To study the effect of increasing the chain length from
C1 to C3, the aromatic substitution pattern (4-OH, 3-
OMe) and the carboxylic acid functional group were
kept constant (Table 2). Only the C6C3 acid 6h reduced
the galvinoxyl radical and neither the C6C2 4 nor the
C6C1 1k compounds were oxidised. For the C6C3 com-
pounds the effects of both substitution and functional
group were studied. Despite the C6C3 compounds being
more easily oxidised than their C6C1 homologues, the
same effect of increasing ease of oxidation with increas-
Mono-hydroxylated acids (2-, 3- and 4-hydroxybenzoic
acids 1a–c) were not oxidised and only when certain pat-
terns of di- and trihydroxylation were present was the
galvinoxyl radical reduced (Table 1). The trihydroxy-
lated compound (3,4,5-trihydroxybenzoic acid 1j) was
the most easily oxidised, with the dihydroxyl substitu-
ents 1d–i showing the following order of ease of oxida-
tion: 2,5 > 2,3 > 3,4, > 2,6. Structures containing 2,4-,
and 3,5-dihydroxyl substituents were not oxidised. From
Table 1. Reaction stoichiometry for galvinoxyl radical reduction by C6C1 compounds and the frontier electron orbital energy values (eV) for the
parent compounds and their radicals
Compound
Reaction stoichiometry
Parent compound
Parent radical
SOMO
HOMO
LUMO
HOMO
LUMO
1a 2-Hydroxybenzoic acid
1b 3-Hydroxybenzoic acid
1c 4-Hydroxybenzoic acid
1d 2,3-Dihydroxybenzoic acid
Not reduced
Not reduced
Not reduced
1.41 0.03
ꢀ9.50954
ꢀ9.51455
ꢀ9.60871
ꢀ9.22076
ꢀ0.45878
ꢀ0.56945
ꢀ0.48173
ꢀ0.55975
ꢀ10.91551
ꢀ10.94828
ꢀ10.93326
ꢀ10.4717
ꢀ10.45479
ꢀ10.33417
ꢀ10.35394
ꢀ10.87503
ꢀ10.88799
ꢀ10.37278
ꢀ10.50312
ꢀ10.48061
ꢀ10.33745
ꢀ9.86016
ꢀ9.97177
ꢀ10.33483
ꢀ9.71723
ꢀ10.28841
ꢀ10.26674
ꢀ9.88132
ꢀ5.96784
ꢀ5.82593
ꢀ6.11756
ꢀ5.6075
ꢀ0.41643
ꢀ0.68828
ꢀ0.25208
ꢀ0.46471
ꢀ0.61731
ꢀ0.48044
ꢀ0.16392
ꢀ0.46485
ꢀ0.59617
ꢀ0.28892
ꢀ0.66763
ꢀ0.26935
ꢀ0.77985
ꢀ0.77693
ꢀ0.28485
ꢀ0.16121
ꢀ0.10019
ꢀ0.1956
ꢀ5.59048
ꢀ6.15759
ꢀ6.20571
ꢀ5.6193
1e 2,4-Dihydroxybenzoic acid
1f. 2,5-Dihydroxybenzoic acid
Not reduced
1.50 0.03
ꢀ9.50396
ꢀ9.07923
ꢀ0.45423
ꢀ0.59993
ꢀ5.47255
ꢀ6.09513
ꢀ5.50581
ꢀ5.74064
ꢀ6.00346
ꢀ5.73572
ꢀ5.59501
ꢀ5.61908
ꢀ5.39653
ꢀ5.54942
ꢀ5.34458
ꢀ4.96278
1g 2,6-Dihydroxybenzoic acid
1h 3,4-Dihydroxybenzoic acid
0.80 0.03
1.30 0.11
ꢀ9.46883
ꢀ9.22056
ꢀ0.22441
ꢀ0.58606
1i 3,5-Dihydroxybenzoic acid
1j 3,4,5-Trihydroxybenzoic acid
Not reduced
2.13 0.04
ꢀ9.4869
ꢀ0.65919
ꢀ0.66964
ꢀ9.22625
1k. 4-Hydroxy-3-methoxybenzoic acid
1l 4-Hydroxy-3,5-dimethoxybenzoic acid
2a 4-Hydroxy-3-methoxybenzaldehyde
2b 3-Hydroxy-4-methoxybenzaldehyde
3 4-Hydroxy-3-methoxybenzyl alcohol
Not reduced
1.17 0.02
Not reduced
Not reduced
0.60 0.04
ꢀ9.10837
ꢀ9.01434
ꢀ9.03822
ꢀ9.033
ꢀ0.51922
ꢀ0.52645
ꢀ0.48898
ꢀ0.48716
0.39312
ꢀ0.58254
0.42639
ꢀ8.57475

W.R.Russell et al./ Bioorg.Med.Chem.13 (2005) 2537–2546
2539
O
O
OH
R1
HO
O
OH
R1
R5
R4
R1
R2
R5
R4
R5
R4
R5
R4
R1
R2
R5
R4
R1
R2
R2
R2
R3
R3
R3
R3
R3
2
3
4
5
1
R6
R6
O
O
O
O
HO
HO
HO
HO
R5
R1
R5
R4
R1
R2
OH
OH
R4
R2
OH
OH
R3
R3
O
O
O
O
6
7
8
9
HO
O
HO
R1
O
R3
HO
HO
O
O
O
O
R2
O
O
OH
OH
10
O
OH
HO
O
11
12
R1
R2
R3
OH
O
O
R5
O
O
O
HO
R4
OH
HO
R6
O
OH
14
15
13
OH
OH
OH
OH
OH
OH
OH
OH
HO
O
HO
O
O
HO
O
HO
O
OH
OH
OH
O
OH
OH
O
OH
16
17
18
19
Figure 1. R1 = OH, R2–R5 = H (1a);R2 = OH, R1, R3–R5 = H ( 1b);R3 = OH, R1, R2, R4 and R5 = H ( 1c);R1 and R2 = OH, R3–R5 = H ( 1d);
R1 and R3 = OH, R2, R4 and R5 = H (1e);R1 and R4 = OH, R2, R3 and R5 = H ( 1f);R1 and R5 = OH, R2–R4 = H ( 1g);R2 and R3 = OH, R1,
R4 and R5 = H (1h);R2 and R4 = OH, R1, R3 and R5 = H ( 1i);R2–R4 = OH, R1 and R5 = H ( 1j);R3 = OH, R2 = OMe, R1, R4 and R5 = H ( 1k);
R3 = OH, R2 and R4 = OMe, R1 and R5 = H (1l);R3 = OH, R2 = OMe, R1, R4 and R5 = H ( 2a);R4 = OH, R3 = OMe, R1, R4 and R5 = H ( 2b);
R3 = OH, R2 = OMe, R1, R4 and R5 = H (3);R3 = OH, R2 = OMe, R1, R4 and R5 = H ( 4);R3 = OH, R2 = OMe, R1, R4 and R5 = H ( 5);
R2 = OH, R1, R3–R5 = H, R6 = CO2H (6a);R1 = OH, R2–R5 = H, R6 = CO 2H (6b);R3 = OH, R1, R2, R4 and R5 = H, R6 = CO 2H (6c);
R3 = OH, R1, R2, R4 and R5 = H, R6 = CO2Et (6d);R3 = OH, R2 = OMe, R1, R4 and R5 = H, R6 = CO 2Et (6e);R3 = OH, R1, R2, R4 and
R5 = H, R6 = CH2OH (6f);R3 = OH, R2 and R4 = OMe, R1 and R5 = H, R6 = CO 2Et (6g);R3 = OH, R2 = OMe, R1, R4 and R5 = H,
R6 = CO2H (6h);R3 = OH, R2 = OMe, R1, R4 and R5 = H, R6 = CH 2OH (6i);R3 = OH, R2 and R4 = OMe, R1 and R5 = H, R6 = CO 2H (6j);
R3 = OH, R2 and R4 = OMe, R1 and R5 = H, R6 = CH2OH (6k);R2 and R3 = OH, R1, R4 and R5 = H, R6 = CO 2H (6l);R3 = OH, R2 = OMe,
R1, R4 and R5 = H, R6 = CO2H (7a);R2 and R3 = OH, R1, R4 and R5 = H, R6 = CO 2H (7b);R3 = OH, R1, R2, R4 and R5 = H, R6 = CO 2H (7c);
R2 = OH, R1, R3–R5 = H, R6 = CO2H (7d);R2 = OH, R1 and R3 = H ( 10a);R1 = OH, R2 and R3 = H ( 10b);R2 and R3 = OH, R1 = H ( 10c);
R1 = Me, R2 = OH and R3 = H (10d);R2 = Me, R1 = OH and R3 = H ( 10e);R2 = OH, R3 = OMe and R1 = H ( 10f);R2 = OH, R3 = O-glucose
and R1 = H (10g);R2, R4–R6 = OH R1 and R3 = H, ( 15a);R1, R2, R4–R6 = OH, R3 = H, ( 15b);R1–R6 = OH ( 15c);R1, R2, R4 and R6 = OH,
R3 = H, R5 = OMe (15d);R2, R4–R6 = OH, R3 = H, R1 = OMe ( 15e);R1, R4–R6 = OH, R3 = H, R2 = OMe ( 15f) and R1, R2, R5, R6 = OH,
R3 = H, R4 = O-glucoside (15g).
ing ring hydroxylation and methoxylation was observed.
For all substitution patterns, modification of the side
chain from the carboxylic acid to the alcohol resulted
in increasing the ease of oxidation, as seen for the
C6C1 compounds and esterification decreased the com-
pound ability to reduce the galvinoxyl radical. Again
replacing the 3-methoxyl group with a hydroxyl group
to give 3,4-dihydroxycinnamic acid 6l resulted in
increasing the ease of oxidation. With the standard 4-
hydroxyl, 3-methoxyl and 3,4-dihydroxyl substitution

2540
W.R.Russell et al./ Bioorg.Med.Chem.13 (2005) 2537–2546
Table 2. Reaction stoichiometry for galvinoxyl radical reduction by C6C2 and C6C3 compounds and the frontier electron orbital energy values (eV)
for the parent compounds and their radicals
Compound
Reaction stoichiometry
Parent compound
HOMO LUMO
Parent radical
SOMO
HOMO
LUMO
4 Homovanilic acid
5 Eugenol
Not reduced
0.79 0.06
Not reduced
Not reduced
Not reduced
Not reduced
0.14 0.05
0.74 0.02
0.87 0.17
1.37 0.05
1.49 0.10
1.67 0.06
1.78 0.06
1.69 0.02
ꢀ8.88491
ꢀ8.5759
0.13174 ꢀ10.05663 ꢀ5.17296
0.36403
ꢀ9.88842 ꢀ4.96512
0.25485
0.40289
6a 3-Hydroxycinnamic acid
6b 2-Hydroxycinnamic acid
ꢀ9.32604 ꢀ0.86748 ꢀ10.22476 ꢀ5.77831 ꢀ0.99105
ꢀ9.85363 ꢀ0.80141 ꢀ10.08041 ꢀ5.77815 ꢀ0.75452
ꢀ9.11988 ꢀ0.78821 ꢀ10.87332 ꢀ5.80599 ꢀ0.8232
ꢀ9.16465 ꢀ0.77723 ꢀ10.89863 ꢀ5.83169 ꢀ0.77034
ꢀ8.95258 ꢀ0.82792 ꢀ10.28084 ꢀ5.47885 ꢀ0.68911
ꢀ8.76362 ꢀ0.12819 ꢀ10.38905 ꢀ5.39648 ꢀ0.26929
6c 4-Hydroxycinnamic acid
6d Ethyl 4-hydroxycinnamate
6e Ethyl 4-hydroxy-3-methoxycinnamate
6f 4-Hydroxycinnamyl alcohol
6g Ethyl 4-hydroxy-3,5-dimethoxycinnamate
6h 4-Hydroxy-3-methoxycinnamic acid
6i 4-Hydroxy-3-methoxycinnamyl alcohol
6j 4-Hydroxy-3,5-dimethoxycinnamic acid
6k 4-Hydroxy-3,5-dimethoxycinnamyl alcohol
6l 3,4-Dihydroxycinnamic acid
ꢀ8.80779 ꢀ0.83035
ꢀ8.78757 ꢀ5.26116 ꢀ0.62532
ꢀ8.84913 ꢀ0.82485 ꢀ10.23216 ꢀ5.43796 ꢀ0.7103
ꢀ8.45336 ꢀ0.11646
ꢀ8.76742 ꢀ0.83769
ꢀ8.38945 ꢀ0.07973
ꢀ9.72197 ꢀ5.00762 ꢀ0.06633
ꢀ9.77507 ꢀ5.24628 ꢀ0.68652
ꢀ9.38175 ꢀ4.79961 ꢀ0.02136
ꢀ8.9397
ꢀ0.88366 ꢀ10.57365 ꢀ5.54341 ꢀ0.93819
ꢀ9.98376
5.43439 ꢀ0.93819
7a 3-(4-Hydroxy-3-methoxy phenyl)propionic acid 0.17 0.06
7b 3-(3,4-Dihydroxyphenyl)propionic acid
ꢀ8.79374
ꢀ8.93147
0.13826 ꢀ10.07926 ꢀ5.1014
0.09917 ꢀ10.17905 ꢀ5.34007
ꢀ10.09473 ꢀ5.26871
0.22796
0.1038
2.17 0.10
0.02141
0.17081
0.0472
7c 3-(4-Hydroxyphenyl)propionic acid
7d 3-(3-Hydroxyphenyl)propionic acid
Not reduced
Not reduced
ꢀ9.104
0.2111
ꢀ10.67721 ꢀ5.51714
ꢀ9.22479
0.16044 ꢀ10.43876 ꢀ5.61166
patterns, the effect of a fully saturated side chain was
investigated. For the methoxylated compound 7a
this resulted in a decrease in the ease of oxidation,
suggesting that for this pattern of substitution, the pres-
ence of the conjugated side chain, has an important role
in stabilising the parent radical. However, for the
dihydroxylated compound 6l an increased ease of oxida-
tion was achieved by reduction of the olefinic side chain
7b, demonstrating that for this structure the ability of
the compound to reduce the galvinoxyl radical was in-
creased, not by extended delocalisation of the parent
radical, but more likely by formation of a quinone. This
stabilising property of the olefinic double bond was fur-
ther investigated for a series of compounds, which
showed an increased ease of oxidation with increased
methoxyl substitution and also an increased ease of oxi-
dation as the functional group was modified. The alco-
hol was found to be more easily oxidised than the
acid, which in turn was more easily oxidised than the
ester. For these compounds the distribution of the un-
paired electron was calculated and compared (Fig. 2).
For all compounds, it was observed that for the com-
pounds which reduced the galvinoxyl radical, a greater
proportion of the positive spin density was found at
EPR-silent positions (i.e., >0.53). This value is most
likely to reflect the amount of spin density residing on
the phenolic oxygen and is in good correlation with
the graphical distributions calculated computationally.
This suggests that for the mono-hydroxylated com-
pounds, extending the conjugation stabilises the parent
radical formed and that the distribution patterns of
the unpaired electron appears to also have an important
effect. For all of the mono-hydroxylated compounds
(with the exception homovanillic acid), there was good
correlation of the HOMO energies with the ease of oxi-
dation. Compounds with HOMO energies greater than
ꢀ9.01 eV and SOMO energies of the parent radicals
greater than ꢀ5.48 eV were oxidised, whereas the com-
pounds with lower energy orbitals were not.
2.3. C6C3 coumarins and (C6C3)2 compounds
For the coumarins, it is possible to delocalise the radical
between both rings and so a series of structures, which
represented both the standard 4-hydroxyl, 3-methoxyl
and 3,4-dihydroxyl substitution patterns were selected
(Table 3). Structures with hydroxyls on each separate
ring, with 10d,e and without methylation 10a,b and
where the hydroxyl at C3 was methoxylated 10f and gly-
cosylated 10g were also included. Only the 3,4-dihydr-
oxylated compound 10c reduced the galvinoxyl radical.
As for all the simple single ring phenols, monohydroxy-
lated compounds were not oxidised. However, unlike the
4-hydroxy,3-methoxycinnamic acid 6h, the 4-hydroxyl,
3-methoxyl pattern of substitution for the coumarin
structure 10f did not reduce the galvinoxyl radical. This
demonstrates that formation of a second ring of this
structure had an inhibitory effect on the ease of oxida-
tion. To examine this further, compounds were synthes-
ised to represent potential delocalisation of the radical
from 4-hydroxy,3-methoxycinnamic acid into further ring
systems by coupling at positions C5 and C8 (Table 3).
Coupling at C5 8 resulted in an increase in the ease
of oxidation, whereas coupling at C8 to C5 11 or C8 to
C8 12 both resulted in a decrease in the ease of oxidation,
as did full saturation of the olefinic double bond 9. It ap-
peared that extended conjugation through the side chain
caused a reduction in the ability of the compound to re-
duce the galvinoxyl radical whereas, when the compound
was coupled through the aromatic rings alone, an
increase in the ability of this compound to reduce the
galvinoxyl radical was observed. To test whether the
increased ease of oxidation with C5–C5 coupling was a
result of the presence of an additional hydroxyl group

W.R.Russell et al./ Bioorg.Med.Chem.13 (2005) 2537–2546
2541
Table 3. Reaction stoichiometry for galvinoxyl radical reduction by
compounds containing more than one ring. Standard deviations are
calculated for n = 3 (a), n = 6 (b) and n = 5 (c) values
Compound
Reaction
stoichiometry
8 4-Hydroxy-3-methoxycinnamic acid, 5-5 linked
9 3-(4-Hydroxy-3-methoxyphenyl)propionic acid,
5-5 linked
1.72 0.03a
0.22 0.01a
10a 7-Hydroxycoumarin
10b 4-Hydroxycoumarin
10c 6,7-Dihydroxycoumarin
Not reduced
Not reduced
3.77 0.01a
Not reduced
Not reduced
Not reduced
Not reduced
10d 7-Hydroxy-4-methylcoumarin
10e 4-Hydroxy-7-methylcoumarin
10f 7-Hydroxy-6-methoxycoumarin
10g 7-Hydroxycoumarin-6-glucoside
11 4-Hydroxy-3-methoxycinnamic acid, 8-5 linked 0.97 0.03a
12 4-Hydroxy-3-methoxycinnamic acid, 8-8 linked Not reduced
13 Resveratol
14 Curcumin
0.55 0.12b
1.75 0.09b
2.68 0.15c
3.34 0.40c
4.80 0.26c
2.52 0.22b
2.26 0.21b
1.07 0.03b
1.35 0.05b
2.92 0.18c
3.53 0.05c
2.98 0.12c
0.90 0.17c
15a Kaempferol
15b Quercetin
15c Myricetin
15d Rhamnetin
15e Isorhamnetin
15f Tamarixetin
15g Isoquercetin
16 Epicatachin
17 Luteolin
18 Taxifolin
19 Eriodictyol
pound 15a. To determine whether the A-ring or B-ring
had the most significant effect on ease of oxidation, a
series of compounds, which were selectively, methox-
ylated on the A and B rings were examined 15d–f.
Methoxylation on both rings resulted in a decrease in
the ease of oxidation and this followed the pattern:
B4 > B3 > A7, showing the B-ring to have the most sig-
nificant effect. However, when hydroxyl formation on
the C-ring at position 3 was blocked by glycosylation
15g, this had the most significant effect on the ability
of this compound to reduce the galvinoxyl radical. Mod-
ifying the C-ring in this way not only removed the addi-
tional hydroxyl group from the structure, but will also
have an effect on the delocalisation of the parent radical
throughout the entire molecule. Keeping the 5,7-hydrox-
ylation on the A-ring and 3,4-hydroxylation on the
B-ring, the effect of C-ring structure modification was
determined. When an olefinic double bond was present
in the C-ring 15,17, this resulted in an overall increased
ease of oxidation, but this was slightly reduced if posi-
tion 3 was hydroxylated 15. This suggests that for these
compounds the extended conjugation from the B-ring
into the C-ring is important in stabilising the parent rad-
ical, as the presence of a substituent at C3 in the C-ring
will block the formation of a radical at this site, there-
fore decreasing the stability of the parent radical
formed. This is further confirmed by the structure con-
taining a carbonyl group at position 4 and absence of
an olefinic double bond in the C-ring 18, which showed
a dramatic decrease in the ability of this compound to
reduce the galvinoxyl radical. For this structure, intro-
duction of the hydroxyl at C3 19 showed an increase
in the ease of oxidation in contrast to when extended
Figure 2. Graphical distribution of the unpaired electron for geomet-
rically optimised structures calculated computationally with spin
density distributions at the carbon nuclei calculated from EPR
hyperfine constants (see experimental methods). Positive spin densities
are shown in green and the calculated EPR-silent values are given in
parentheses.
at C4, a compound representing coupling through the
side chain without involvement and extension of conju-
gation of the olefinic double bond was tested 14. This
compound was also more easily oxidised and showed a
very similar level of oxidation to the C5–C5 coupled
compound, suggesting that the presence of the second
hydroxyl substituent is most likely to be responsible.
2.4. C6C3C6 compounds
Quercetin 15b was chosen as the standard skeleton-type
representative of the flavonoids, as this compound has
the 5,7-dihydroxylation pattern on the A-ring most
commonly observed in nature and the 3,4-dihydroxyl-
ation pattern on the B-ring investigated thoroughly
for the simpler phenols. As for the simple phenols, when
the extent of hydroxylation on the B-ring was varied, the
results followed the same pattern with the trihydroxy-
lated compound 15c being more easily oxidised than
the dihydroxylated compound 15b, which in turn was
more easily oxidised that the monohydroxylated com-

2542
W.R.Russell et al./ Bioorg.Med.Chem.13 (2005) 2537–2546
conjugation was present. There is very little change in
the ability of this compound to reduce the galvinoxyl
radical whether the carbonyl group at C4 is present 19
or not 16, suggesting that this effect is predominantly
due to the presence of the additional hydroxyl group
alone. Overall this suggests that, although the presence
of an additional hydroxyl group increases the ease of
oxidation, this has a less significant effect than the
increased stability obtained through extended
conjugation.
3. Discussion
For all compounds studied, the ease of phenolic oxida-
tion was increased by introduction of methoxyl or
hydroxyl substituents into the aromatic ring, hydroxyl-
ation having the greater effect. Keeping the aromatic
substitution pattern constant and increasing the chain
length also increased the ease of oxidation and this was
additionally increased by introduction of an olefinic dou-
ble bond in the C6C3 cinnamic acids. However, this effect
can not be simply attributed to the effect of extending the
conjugation of the system, as the ease of oxidation for the
3,4-dihydroxylated cinnamic acid 6l increased on full sat-
uration of the olefinic double bond 7b. For the mono-
hydroxylated phenols, this is due to the distribution of
the unpaired electron within the parent radicals, as dem-
onstrated for a series of phenylpropanoids varying in
their ability to oxidise the galvinoxyl radical. However,
for compounds containing 3,4-dihydroxylated substitu-
ents this is more likely to be attributed to quinone forma-
tion and for the simpler phenols this has been addressed
in a separate study.14
Stoichiometric studies are ideal for comparing the
reducing capabilities of the various phenolic compounds
studied. However, in a biological system kinetic factors
may also have an important role in controlling the redox
chemistry. This is likely to be particularly relevant in
poly-hydroxylated compounds such as the flavanoids
and for this group of compounds, studies have already
demonstrated differences in reaction stoichiometry and
kinetics.12 To determine the kinetic effect of the C6C1
and C6C3 phenolic acids on the reduction of the galvin-
oxyl radical, the decay curves of this radical was ob-
served in the presence of a selected subset of phenolic
compounds (Fig. 3). For the hydroxylated C6C1 com-
pounds (Fig. 3A), the 3,4,5-trihydroxylated compound,
which had the highest effect in the stoichiometric mea-
surements also had the greatest effect on the in situ
reduction of the galvinoxyl radical. The 3,4-dihydrox-
ylated and monhydroxylated compounds were all less
effective and followed the order 3-OH > 2-OH > 3,4-
diOH > 4-OH. The effect of methoxylation on the
C6C3 hydroxycinnamyl alcohols also showed the same
pattern as the stoichiometric measurements with the
dimethoxylated compound having the greatest effect
and the mono-methoxylated compound being more
effective than the non-methoxylated compound. The ef-
fect of functional group in the C6C3 compounds showed
that as for the stoichiometric measurements, that the
cinnamyl alcohol was the most effective. However, un-
like the stoichiometric results, the ester appeared to be
slightly more effective than the acid, but both of these
compounds were quite low in effect.
When the effect of ring formation was investigated,
again contrasting results were obtained for the 4-hy-
droxy-3-methoxyl and the 3,4-dihydroxyl patterns of
substitution. The coumarin structure containing the
4-hydroxy-3-methoxyl substituents 10f was not oxi-
dised, demonstrating that ring formation of this type
had an inhibitory effect, whereas the 3,4-dihydrox-
ylated coumarin 10c reduced the galvinoxyl radical
more easily than 3,4-dihydroxycinnamic acid 6l.
Again, this is likely to be due to the fact the dihydr-
oxylated does not give increased stabilisation through
extended delocalisation of the parent radical, but
through quinone formation. An inhibitory effect on
ease of oxidation was also observed when 4-hydr-
oxy-3-methoxycinnamic acid was coupled through C8
12. Again, this is most likely to be due to removal
of extended conjugation, as the C8 position is not free
to support a stable radical. This is further confirmed
by an increased and similar ease of oxidation observed
(A)
(B)
(C)
control
4-OH(1c)
2-OH (1a)
3,4-diOH (1h)
3-OH (1b)
control
control
acid (6h)
ester (6e)
4-OH (6f)
4-OH-3-OMe (6i)
4-OH-3,5-diOMe (6k)
3,4,5-triOH(1j)
alcohol (6i)
5
10
15
20
5
10
15
20
5
10
15
20
Time (seconds)
Figure 3. Decay curves of the galvinoxyl resonance obtained during in situ reduction by A. hydroxylated C6C1 compounds;B. 4-hydroxy-3-
methoxycinnamyl alcohol, 4-hydroxy-3-methoxycinnamic acid and ethyl 4-hydroxy-3-methoxycinnamate and C. substituted C6C3 hydroxycinnamyl
alcohols.

W.R.Russell et al./ Bioorg.Med.Chem.13 (2005) 2537–2546
2543
when 4-hydroxy-3-methoxycinnamic acid is coupled
4. Experimental procedures
through C5 8 and C9 14. The flavonoids exhibited
the same results with regard to hydroxylation patterns
as the simpler phenols. With the A-ring containing the
commonly found 5,7-dihydroxylation pattern, it was
observed that the ortho-hydroxyl substituents on the
B-ring had a greater contribution to the ability of
the compound to reduce the galvinoxyl radical. How-
ever, the most significant effect on ease of oxidation
was observed when the C-ring structure was modified,
as this will effect both stabilisation obtained through
increased delocalisation of the radical and through
quinone formation.
Compounds 1b, 1e–i, 2a, 2b, 3, 5, 6a–c, 6j, 10c, 10d and
10f were purchased from Aldrich, compounds 1a, 1c, 1j,
6l, 10a, 10b, 10e, 10g, 13, 15, 15a–c and 16 were pur-
chased from Sigma and compounds 1d, 4 and 6h were
purchased from Fluka. Compounds 15d–g, 17–19 were
purchased from Apin. Compound 14 was purchased
from Alexis. Compounds 6d–g, 6i and 6k were synthes-
ised as reported previously.19 Compound 8 was pre-
pared by the initial coupling of 4-hydroxy-3-meth-
oxy-benzaldehyde. The 4-hydroxyl substituent was then
protected by acetylation and the side chain extended
by a malonic acid condensation, as detailed below.
Compound 9 was then prepared by hydrogenation
of compound 8, as were compounds 7a–d from their
corresponding cinnamic acids. Compound 11 was
prepared according to the procedure of Ralph et al.20
and compound 12 by the procedures of Cartwright
and Haworth21 and Ralph et al.20
The HOMO and SOMO energies calculated for the
mono- and dihydroxylated compounds and their radi-
cals at the AM1 level of theory correlate well with the
experimental data obtained and gave a good indica-
tion as to whether a compound reduce the galvinoxyl
radical. However, it was observed that the dihydroxy-
lated compounds, which were oxidised, had lower
HOMO and SOMO energies than the mono-hydroxyl-
ated compounds and this in part may be due to differ-
ent mechanism of stabilisation. Since the mechanisms
of stabilisation in the (C6C3)2 and C3C6C3 compounds
are more complex due to the extended conjugation,
multiple hydroxyl substituents and the ability of some
of these compounds to form quinones,15 no attempt
was made to correlate the frontier electron energies
with the experimental data. However, rigorous compu-
tational calculations have been shown to be extremely
useful in studies of particular flavonoids with regard
to their radical scavenging ability.16–18 Generally,
semi-empirical calculations gave a good indication of
the ease of oxidation within a series of compounds
with similar structural properties. To explore the
mechanisms of stabilisation further and to investigate
the redox properties of more complex molecules, fur-
ther work to improve the accuracy of these results
using ab initio calculations would be desirable. Over-
all, these results highlight the importance of electron
delocalisation and the effect this has on stability of
the parent radicals. It also draws attention to the sig-
nificance of quinone formation in compounds that
contain more than one hydroxyl, which for the com-
pounds of interest tend to be the more biologically
important in terms of their redox potential. The ki-
netic data obtained for the C6C1 and C6C3 compounds
showed similar effects for hydroxyl and methoxyl sub-
stitution to that of the stoichiometric data studied.
However, it is likely that in a biological system where
many of these compounds will be present as mixtures,
the reaction rates will be an important aspect.
Although, the data suggests that this may be less
important in the simpler phenolics, where the trends
are similar, it may be an overriding factor in the more
complex poly-hydroxylated C6C3C6 compounds. So,
although the redox chemistry of phytochemicals is
not the only factor to be considered in the interaction
of these compounds with metalloenzymes and other
redox systems, an understanding of these processes
can serve as an extremely useful aid in interpreting re-
sults from further studies in this area.
4.1. Synthesis: 5-50 dehydrodi-(4-hydroxy-3-methoxy-
cinnamic acid) 8
4-Hydroxy-3-methoxybenzaldehyde [Aldrich] (5.22 g;
6.6 mmol) was dissolved in citrate/phosphate buffer
3
(0.1 moldmꢀ3;pH 4.2). Hydrogen peroxide (2.82 cm
;
27.5% w/v) and horseradish peroxidase [Sigma] (2025
units) were added and the reaction left stirring at
36.5 ꢁC overnight.22 The precipitate was filtered and
washed with water and then chloroform to give the cou-
pled product. Yield 97%; dH ((CD3)2SO) 3.92 (6H, s,
OCH3), 7.44 (4H, s, C(2 and 6)H) and (2H, s, C(7)H)
ppm. dC ((CD3)2SO) 55.95 (OCH3), 109.20 (C2),
124.63 (C3), 127.61 (C6), 127.97 (C1), 148.20 (C4),
150.66 (C5), 190.95 (CHO) ppm. The product (2 g,
6.6 mmol) was dissolved in a mixture of acetic anhydride
(50 cm3) and sodium acetate (2 g) and the temperature
raised to 100 ꢁC until dissolved and then maintained at
80 ꢁC for 1 h. Diluted with ice water (100 cm3) and left
for 1 h. Extracted into chloroform (20 cm3 · 3) and
washed with sodium hydrogen carbonate (3% w/v).
The organic layer was left to stand over sodium sulfate
(anhydrous), filtered and the solvent removed in vacuo.
Yield 79%; dC ((CD3)2SO) 19.86 (COCH3), 56.34
(OCH3), 111.87 (C2), 124.50 (C5), 130.60 (C6), 134.45
(C1), 141.84 (C4), 151.88 (C3), 167.38 (COCH3),
191.67 (CHO) ppm. dH ((CD3)2SO), 2.06 (6H, s,
(COCH3)), 3.91 (6H, s, OCH3), 7.45 (2H, d, J 1.90,
C(2)H), 7.69 (2H, d, J 1.90, C(6)H), 9.99 (2H, s,
CHO) ppm. The product (0.5 g, 1.3 mmol) and malonic
acid [BDH], (1 g) were dissolved in pyridine (2 cm3).23
Piperidine and aniline (0.04 cm3 of each;both freshly
distilled) were added and the mixture warmed to 55 ꢁC
for 1 h under nitrogen. Left to stand for 16 h at room
temperature and then precipitated with HCl
(2 moldmꢀ3). The precipitate was filtered, washed with
water and crystallised from acetic acid to give a white
solid. Yield 68%. dC ((CD3)2SO) 19.93 (COCH3), 56.23
(OCH3), 111.44 (C2), 119.98 (C5), 122.70 (C8), 130.66
(C6), 132.62 (C1), 138.46 (C4), 142.86 (C7), 151.33
(C3), 167.32 (C9), 167.61 (COCH3) ppm. dH
((CD3)2SO), 2.02 (6H, s, COCH3), 3.91 (6H, s, OCH3),

2544
W.R.Russell et al./ Bioorg.Med.Chem.13 (2005) 2537–2546
6.59 (2H, d, J 16, C(8)H), 7.10 (2H, s, C(2)H), 7.52 (2H,
s, C(6)H), 7.57 (2H, d, J 16, C(7)H) ppm. The product
was dissolved in sodium hydroxide (5 cm3;20% w/v)
and left at 55 ꢁC for three hours. Crystallised from
methanol to give 5-50 dehydrodi-(4-hydroxy-3-methoxy-
cinnamic acid) as an off-white solid. Yield 98%. dC
((CD3)2SO) 56.00 (OCH3), 109.53 (C2), 115.84 (C5),
124.82 (C6), 125.16 (C8, C1), 144.37 (C7), 146.34 (C3),
147.88 (C4), 167.76 (C9) ppm. dH ((CD3)2SO), 3.88
(6H, s, OCH3), 6.38 (2H, d, J 15.8, C(8)H), 7.03 (2H,
d, J 2.0 C(2)H), 7.30 (2H, d, J 2.0 C(6)H), 7.50 (2H,
d, J 15.8, C(7)H) ppm.
NaCl. The organic layer was then left to stand over
Na2SO4 (anhydrous), filtered and the solvent removed
in vacuo. Yield 84%. Purified by flash chromatography
eluting with CHCl3/ethyl acetate (5:1) and ethyl acetate
to give 8-5 dehydrodi-(4-hydroxy-3-methoxycinnamic
acid) as a yellow solid. Yield 45%;NMR dH (DMSO)
3.35 (3H, s, AOCH3), 3.87 (3H, s, BOCH3), 6.24 (1H,
d, J 15.9 B(8)H), 6.57 (1H, d, J 2.1 B(2)H), 6.61 (1H,
d, J 8.4 B(5)H), 6.74 (1H, dd, J 2.1, 8.4 B(6)H), 6.88
(1H, d, J 1.9 B(6)H), 7.18 (1H, d, J 1.9 B(2)H), 7.52
(1H, d, J 15.9 B(7)H) and 7.74 (1H, s, A(7)H) ppm.
4.5. 8-80 Dehydrodi-(4-hydroxy-3-methoxycinnamic acid)
12
4.2. 5-50 Dehydrodi-3-(4-hydroxy-3-methoxyphenyl)pro-
pionic acid 9
Initial coupling of compound 6h (2 g) was as described
for compound 6e. Purified by flash chromatography
eluting with ethyl acetate/CHCl3 3:1. Yield 10%;NMR
dH (DMSO) 3.80 (6H, s, OCH3), 4.17 (2H, br s,
C(8)H), 5.83 (2H, br s, C(7)H), 6.1 (2H, d, J 9.1
C(5)H), 6.87 (2H, d, J 9.1 C(6)H) and 6.98 (2H, br s,
C(2)H) ppm. NMR dC (DMSO) 47.96 (C8), 55.76
(OCH3), 81.91 (C7), 110.64 (C2), 115.41 (C5), 119.07
(C6), 128.90 (C1), 147.31 (C4), 147.80 (C3) and 175.21
(C9) ppm. The product (100 mg) was dissolved in
NaOH (10 cm3;2 moldm ꢀ3) and left stirring overnight
under nitrogen. Acidified with HCl (2 moldmꢀ3) and
partitioned between ethyl acetate and saturated aq
NaCl. The organic layer was left to stand over Na2SO4,
filtered and removed solvent in vacuo. The product was
dissolved in methanol (10 cm3) and diazo(trimethyl-
silyl)methane (1.62 cm3 of 2 moldmꢀ3 in hexane) was
added in aliquots of 0.27 cm3. The solvent was removed
in vacuo, the product dissolved in CH2Cl2 (anhydrous,
10 cm3) and DBU (0.15 cm3) was added. Stirred at room
temperature for 4 h, diluted with CH2Cl2 and washed
with HCl (3% w/v) and saturated aq NaCl. The organic
layer was left to stand over Na2SO4 and removed sol-
vent in vacuo. The product was dissolved in 1,4-dioxane
(2 cm3) under nitrogen at room temperature and NaOH
Compound 8 was dissolved in ethyl acetate (5 cm3), pal-
ladium on activated carbon was added (5%;1 mg) and
the mixture was stirred at room temperature under
hydrogen for 2 h. The mixture was then filtered and
the solvent removed in vacuo to give 5-50 dehydrodi-
(4-hydroxy-3-methoxy- phenylpropionic acid) as a white
solid. Yield 100%. dC ((CD3)2SO) 29.74 (C8), 35.27 (C7),
55.52 (OCH3), 110.52 (C2), 122.29 (C5), 125.51 (C6),
130.63 (C1), 141.35 (C3), 147.27 (C4) and 173.46 (C9)
ppm. dH ((CD3)2SO), 2.50 (4H, t, J 7.2, C(8)H), 2.74
(4H, t, J 7.2, C(7)H), 3.79 (6H, s, OCH3), 6.57 (2H, d,
J 1.9 C(2)H) and 6.79 (2H, d, J 1.9 C(6)H) ppm.
4.3. 3-(4-Hydroxy-3-methoxyphenyl)propionic acid 7a,
3-(3,4-dihydroxyphenyl)propionic acid 7b, 3-(4-hydroxy-
phenlyl)propionic acid 7c and 3-(3-hydroxyphenyl)-
propionic acid 7d
Compounds 6h, 6l, 6c and 6a were hydrogenated as de-
scribed for compound 9 to give the corresponding
hydroxypropionic acids 7a–7d.
4.4. 8-5 Dehydrodi-(4-hydroxy-3-methoxycinnamic acid)
11
3
(2 moldmꢀ3; 5 cm ) was added. After 20 h, acidified
Compound 6e (2 g) was dissolved in acetate buffer
(2 moldmꢀ3;pH 4) by heating to 60 ꢁC and then cooled
to 40 ꢁC. Hydrogen peroxide (0.76 cm3) and horseradish
peroxidase [Sigma] (10 mg in 2 cm3 buffer) were added
and the precipitate collected by filtration after 10 min.
The filtrate was partitioned between ethyl acetate and
saturated aq NaCl. The organic layer was then left to
stand over Na2SO4 (anhydrous), filtered and the solvent
removed in vacuo. Yield 99%. This compound (83.5 mg,
0.189 mmol) was dissolved in CH2Cl2 (2 cm3), 1,8-di-
azabicyclo(5.4.0.)undec-7-ene (0.125 cm3, 0.836 mmol)
was added and the solution stirred at room temperature
for 4 h. Diluted with CH2Cl2 and washed with HCl (3%
v/v) and saturated aq NaCl. The organic layer was left
to stand over Mg2SO4, filtered and evaporated to a res-
idue. Purified by flash chromatography eluting with
CHCl3/EtOAc (5:1) and ethyl acetate to obtain a yellow
solid. Yield 85%. This compound was dissolved in 1,4-
dioxane (2 cm3) under nitrogen, KOH (40% w/v;
5 cm3) was added and the solution stirred at room tem-
perature for 2 h. Acidified with HCl (2 moldmꢀ3) and
partitioned between ethyl acetate and saturated aq
with HCl (2 moldmꢀ3) and partitioned between EtOAc
and saturated aq. NaCl. The organic layer was left to
stand over Na2SO4, filtered and the solvent removed
in vacuo. Purified by flash chromatography eluting with
ethyl acetate/CHCl3 3:1 and recrystallised from metha-
nol to give 8-80 dehydrodi-(4-hydroxy-3-methoxycin-
namic acid) as a yellow solid. Yield 76%;NMR dH
(DMSO) 3.80 (6H, s, OCH3), 6.75 (2H, d, J 9.1
C(5)H), 7.02 (2H, d,d J 9.1, 2.1 C(6)H), 7.19 (2H, d, J
2.1 C(2)H) and 7.83 (2H, s, C(7)H), ppm.
4.6. Electron paramagnetic resonance spectroscopy
Aliquots (3 cm3) of galvinoxyl [Aldrich] (0.5 mmoldmꢀ3
in methanol) were mixed with substrates (measured at
various concentrations to determine linearity, typically
0.1–0.5 mmoldmꢀ3 in methanol) and transferred to an
EPR quartz cell. Spectra (X-band) of unreacted galvin-
oxyl were recorded after 5 on a Bruker E106 spectrom-
eter, equipped with a TM110 cavity. The following
instrument settings were used: modulation frequency
100 kHz;centre field 3480.40 Gauss;sweep width

W.R.Russell et al./ Bioorg.Med.Chem.13 (2005) 2537–2546
2545
immune status and DNA damage. J.Nut.Biochem.
2003, 14(2), 90–98.
6. Johnson, I. T. Anticarcinogenic effects of diet-related
apoptosis in the colorectal mucosa. Food Chem.Toxicol.
2002, 40(8), 1171–1178.
60 Gauss;time constant 40.96 ms;power, 1.01 mW and
a suitable receiver gain setting, typically 1 · 104. The
galvinoxyl concentrations remaining were calculated
by integration of the signal and comparison with the
control. From these concentrations, the stoichiometry
of the reaction was calculated. Continuous-flow EPR
spectroscopy was performed as reported previously.19
Computer simulations of spectra, giving the hydrogen
hyperfine coupling constants (aH), were performed using
the SIMEPR program,24 which sequentially varies all
the parameters for each radical species until a minimum
in the error surface is located. Goodness-of-fit was
determined by visual comparison and as a minimum in
the sum of the squared residuals. The density of the
unpaired electron was calculated from the hyperfine
coupling constants using the McConnell relationship25
in which the proportionality factor was 22.5 Gauss, the
best estimate assuming the value to be constant for all
C–H bonds.26 The unpaired electron density on the
EPR-silent positions was then calculated by difference.
Kinetic measurements were obtained using solutions of
7. Ahn, D.;Putt, D.;Kretsy, L.;Stoner, G. D.;Fromon, D.;
Hollenberg, P. F. The effects of dietary ellagic acid on
rat hepatic and esophageal mucosal cytochromes P450 and
phase II enzymes. Carcinogenesis 1995, 17, 821–828.
8. Diaz, G. D.;Li, Q. J.;Dashwood, R. H. Caspase-8 and
apoptosis-inducing factor mediate a cytochrome c-inde-
pendent pathway of apoptosis in human colon cancer cells
induced by the dietary phytochemical chlorophyllin.
Cancer Res. 2003, 63(6), 1254–1261.
9. MacCarrone, M.;Lorenzon, T.;Guerneri, P.;Argo, A. F.
Resveratrol prevents apoptosis in K562 cells by inhibiting
lipoxygenase and cyclooxygenase. Eur.J.Biochem. 1999,
265, 27–34.
10. Jones, R.;Adel-Alvarez, L. A.;Alvarez, O. R.;Broaddus,
R.;Das, S. Arachidonic acid and colorectal carcinogen-
esis. Mol.Cell.Biochem. 2003, 253(1–2), 141–149.
11. Chang, W. S.;Lee, Y. J.;Lu, F. J.;Chaing, H. C.
Inhibitory effects of flavanoids on xanthine oxidase.
Anticancer Res. 1993, 13, 2165–2170.
galvinoxyl and the selected phenolics (0.2 mmoldmꢀ3
)
12. McPhail, D. B.;Hartley, R. C.;Gardner, P. T.;Duthie, G.
G. Kinetic and Stoichiometric assessment of the antiox-
idant activity of flavonoids by electron spin resonance
in methanol and de-oxygenating in a stream of nitrogen.
Aliquots were transferred to gas-tight syringes, which
were rapidly evacuated into a two-stream EPR cell. De-
cay curves were obtained by operating in time sweep
mode with a static field set at the resonance maximum
for the galvinoxyl radical signal and using a microwave
power of 1.01 mW and time constants of 5.12 ms. Decay
curves were obtained for each phenolic in triplicate and
the mean curves are given in Figure 3.
spectroscopy. J.Agric.Food Chem.
1690.
2003, 51(6), 1684–
13. Aleman, C.;Vega, M. C.;Perez, J. J. On the reliability of
MINDO/3, MNDO and AM1 radical geometries for the
computation of INDO electron spin resonance hyperfine
coupling constants. J.Mol.Struct. 1993, 281, 39–44.
14. Russell, W. R.;Burkitt, M. J.;Scobbie, L.;Chesson, A.
Radical formation and coupling of hydroxycinnamic acids
containing 1,2-dihydroxy substituents. Bioorg.Chem.
2003, 31, 206–215.
4.7. Semi-empirical calculations
15. Bors, W.;Heller, W.;Michel, C.;Saran, M. Flavonoids as
antioxidants: determination of radical scavenging efficien-
cies. Meth.Enziomol. 1990, 186, 343–354.
16. Meyer, M. Ab initio study of flavonoids. Int.J.Quantum
Chem. 2000, 76, 724–732.
17. Mantas, A.;Deretey, E.;Ferretti, F. H.;Estrada, M.;
Csizmadia, I. G. Ab initio conformational analysis of
flavone and related compounds. J.Mol.Struct.(Theo-
chem) 2000, 504, 77–103.
Calculation of the frontier electron orbitals and graphi-
cal representations of the spin densities were made using
the AM1 theory on geometry optimised structures with
the MOPAC program running on an SGI Indigo 2
workstation.
Acknowledgements
18. Oganesyan, E. T.;Mal Õtsev, Y. A.;Tvorovskii, D. E.
Mechanism of reaction of flavone derivatives with
hydroxyl radical by semi-empirical methods. Russ.J.
Gen.Chem. 2001, 939–944.
19. Russell, W. R.;Burkitt, M. J.;Forrester, A. R.;Chesson,
A. Oxidative coupling during lignin polymerization is
determined by unpaired electron delocalization within
parent phenylpropanoid radicals. Arch.Biochem.Biophys.
1996, 332(2), 357–366.
Funding is gratefully acknowledged from the Scottish
Executive Environment and Rural Affairs Department
(SEERAD).
References and notes
1. Dove-Edwin, I.;Thomas, J. W. Review article: The
prevention of colorectal cancer. Alimentary Pharmacol.
Therapeut. 2001, 15, 323–336.
2. de Leon, M. P. Prevention and chemoprevention of colo-
rectal neoplasms. Digestive Liver Disease 2000, 34, 59–69.
3. Wattenberg, L. W. Inhibition of carcinogenesis by
minor dietary constituents. Cancer Res. 1992, 52, 2085S–
2091S.
4. Rice-Evans, C. A.;Miller, N. J. Antioxidant activities of
flavanoids as bioactive components of food. Biochem.Soc.
Trans. 1996, 24, 790–795.
5. Bub, A.;Watzl, B.;Blockhaus, M.;Briviba, K.;Liegibel,
U.;Muller, H.;Pool-Zobel, B. L.;Rechkemmer, G. Fruit
juice consumption modulates antioxidative status,
20. Ralph, J.;Quideau, S.;Grabber, J. H.;Hatfield, R. D.
Identification and synthesis of new ferulic acid dehydro-
dimers present in grass cell walls. J.Chem.Soc,. Perkin
Trans.1 1994, 3485–3497.
21. Cartwright, N. J.;Haworth, R. D. The constituents of
natural plant resins. Part XIX The oxidation of ferulic
acid. J.Chem.Soc. 1944, 66, 535–537.
22. Baumgartner, J.;Neukom, H. Enzymatische oxydation
von vanillia. Chimia 1972, 26, 366–368.
23. Richtzenhain, H. Enzymatische versuche zur entstehung
des lignins. Berichte der Deutschen Chemischen Gesell-
schaft 1949, 447–538.
24. Duling, D. R. Simulation of multiple isotropic spin-trap
epr spectra. J.Mag.Res. 1994, 104, 105–110.

2546
W.R.Russell et al./ Bioorg.Med.Chem.13 (2005) 2537–2546
25. McConnell, H. M. Molecular orbital approximation to
electron coupled interaction between nuclear spins. J.
Chem.Phys. 1956, 24, 460–467.
26. McLachlan, A. D. Self-consistent field theory of the
electron spin distribution in p-electron radicals. Mol.
Phys. 1960, 3(3), 233–252.