Radical formation and coupling of hydroxycinnamic acids containing 1,2-dihydroxy substituents

Article abstract of DOI:10.1016/S0045-2068(03)00042-7

Hydroxycinnamic acids involved in the deposition and cross-linking of plant cell-wall polymers do not usually contain 1,2-dihydroxy substituents, despite the presence of both 3,4-dihydroxycinnamic acid and 4,5-dihydroxy-3-methoxycinnamic acid as intermediates in the biogenesis of lignin. Since the O-methyl transferases, enzymes catalysing methylation, are targets for the genetic manipulation of lignin biosynthesis, the potential incorporation of these 1,2-dihydroxated substrates becomes increasingly significant. Using EPR spectroscopy, it was observed that 1,2-dihydroxy substituents did not have an inhibitory effect on radical formation. Increasing the extent of hydroxylation and methoxylation, resulted in an increased ease of substrate oxidation. Despite formation of the parent radicals, coupling did not proceed, under conditions that generally result in phenylpropanoid polymerisation. It is postulated that intermolecular radical-coupling reactions are inhibited due to rapid conversion to the o-quinone. In contrast, when methoxylated at C3, as in 4,5-dihydroxy-3-methoxycinnamic acid, radical coupling proceeds with the major product resulting from 8-O-3 radical coupling and formation of a substituted 2,3-dihydro-1,4-dioxin ring.

Full text of DOI:10.1016/S0045-2068(03)00042-7

BIOORGANIC
CHEMISTRY
Bioorganic Chemistry 31 (2003) 206–215
www.elsevier.com/locate/bioorg
Radical formation and coupling
of hydroxycinnamic acids containing
1,2-dihydroxy substituents
Wendy R. Russell,a,* Mark J. Burkitt,b Lorraine Scobbie,a
and Andrew Chessona
a
Rowett Research Institute, Greenburn Road, Bucksburn, Aberdeen AB21 9SB, UK
b
Molecular Mechanisms Group, GrayLaboratoryCancer, Research Trust, PO Box 100,
Mount Vernon Hospital, Northwood, Middlesex, HA6 2JR, UK
Received 12 December 2001
Abstract
Hydroxycinnamic acids involved in the deposition and cross-linking of plant cell-wall poly-
mers do not usually contain 1,2-dihydroxy substituents, despite the presence of both 3,4-
dihydroxycinnamic acid and 4,5-dihydroxy-3-methoxycinnamic acid as intermediates in the
biogenesis of lignin. Since the O-methyl transferases, enzymes catalysing methylation, are tar-
gets for the genetic manipulation of lignin biosynthesis, the potential incorporation of these
1,2-dihydroxated substrates becomes increasingly significant. Using EPR spectroscopy, it
was observed that 1,2-dihydroxy substituents did not have an inhibitory effect on radical for-
mation. Increasing the extent of hydroxylation and methoxylation, resulted in an increased
ease of substrate oxidation. Despite formation of the parent radicals, coupling did not pro-
ceed, under conditions that generally result in phenylpropanoid polymerisation. It is postu-
lated that intermolecular radical-coupling reactions are inhibited due to rapid conversion to
the o-quinone. In contrast, when methoxylated at C3, as in 4,5-dihydroxy-3-methoxycinnamic
acid, radical coupling proceeds with the major product resulting from 8-O-3 radical coupling
and formation of a substituted 2,3-dihydro-1,4-dioxin ring.
Ó 2003 Elsevier Science (USA). All rights reserved.
Keywords: Hydroxycinnamic acid; EPR; O-Methyl transferase; Lignin biosynthesis; Caffeic acid; Catechol
* Corresponding author. Fax: +44-01224-716687.
E-mail address: wrr@rri.sari.ac.uk (W.R. Russell).
0045-2068/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved.
doi:10.1016/S0045-2068(03)00042-7

W.R. Russell et al. / Bioorganic Chemistry31 (2003) 206–215
207
1. Introduction
Substituted hydroxycinnamic acids involved in the deposition and cross-linking of
plant cell-wall polymers are produced by the phenylpropanoid pathway. Cinnamic
acid (Fig. 1, 1) undergoes a succession of hydroxylation and methylation reactions
to provide a series of intermediate acids (Fig. 1, 2–6). In theory, any of these acids
may be reduced via the aldehyde (Fig. 1, 8a–c) to the corresponding alcohol (Fig.
1, 9a–c) but, in vivo, this appears to be restricted to acids which do not contain
1,2-dihydroxy substituents (Fig. 1, 2, 4, and 6).
The O-methyl transferases (OMT), the enzymes responsible for methylation of
the hydroxyl groups at C3 and C5, are targets for genetic manipulation of lignin
biosynthesis. Genes encoding OMT have been characterised from a number of
species and RNA antisense and sense strategies have been successfully used to
modulate OMT expression [1–4]. Significant reduction in lignin content without
apparent detrimental consequences for the plant have been observed in several
Fig. 1. Phenylpropanoid pathway; cinnamic acid (1), 4-hydroxycinnamic acid (2), 3,4-dihydroxycinnamic
acid (3), 4-hydroxy-3-methoxycinnamic acid (4), 4,5-dihydroxy-3-methoxycinnamic acid (5), 3,5-dimeth-
oxy-4-hydroxycinnamic acid (6), R1 and R2 ¼ H CoA thioester of 4-hydroxycinnamate (7a), R1 ¼ H
and R2 ¼ OCH3 CoA thioester of 4-hydroxy-3-methoxycinnamate (7b), R1 and R2 ¼ OCH3 CoA thio-
ester of 3,5-dimethoxy-4-hydroxycinnamate (7c), R1 and R2 ¼ H 4-hydroxycinnamyl aldehyde (8a),
R1 ¼ H and R2 ¼ OCH3 4-hydroxy-3-methoxycinnamyl aldehyde (8b), R1 and R2 ¼ OCH3 3,5-dimeth-
oxy-4-hydroxycinnamyl aldehyde (8c), R1 and R2 ¼ H 4-hydroxycinnamyl alcohol (9a), R1 ¼ H and
R2 ¼ OCH3 4-hydroxy-3-methoxycinnamyl alcohol (9b), and R1 and R2 ¼ OCH3 3,5-dimethoxy-4-
hydroxycinnamyl alcohol (9c).

208
W.R. Russell et al. / Bioorganic Chemistry31 (2003) 206–215
constructs [5]. The presence of a 4,5-dihydroxy-3-methoxylated unit was identified
by thioacidolysis for lignins from both transgenic poplar and tobacco, although
no accumulation of the 3,4-dihydroxylated compound was observed [6]. These re-
sults apparently are not consistent with the view that angiosperm OMT are non-
specific with regard to ortho methylation of the 3,4-dihydroxylated substrate and
its 4,5-dihydroxy-3-methoxy analogue. Whether this observation is actually due to
a difference in enzyme specificity, the result of an inability to detect the 3,4-di-
hydroxylated compound, or to the selective incorporation of these compounds
into the polymer, is uncertain. In this paper we address the likelihood of par-
ent phenylpropanoid radicals containing 1,2-dihydroxy substituents participating
in polymer cross-linking reactions and being incorporated into the lignin macro-
molecule.
Parent radical generation is catalysed by iron-containing peroxidases and possibly
monooxygenases [7–9]. The redox cycling capacity of peroxidase can be monitored
by electron paramagnetic resonance (EPR) spectroscopy. Addition of hydrogen per-
oxide to peroxidase results in formation of compound I, and the concomitant loss of
the signal at g ¼ 6:277 as FeIII is oxidised to FeIV. Therefore, the ability of substrates
to reduce peroxidase compounds I and II back to the resting state of peroxidase
can be used as a guide to their ease of oxidation and hence their potential for further
participation in coupling reactions.
Fig. 2. Low temperature EPR spectra of peroxidase, compound I in the presence of 4-hydroxycinnamic
acid (A), 3,4-dihydroxycinnamic acid (B), 3,4,5-trihydroxycinnamic acid (C), 4-hydroxy-3-methoxycin-
namic acid (D), and 3,4-dihydroxy-5-methoxycinnamic acid (E). Spectral width ¼ 3500 Gauss.

W.R. Russell et al. / Bioorganic Chemistry31 (2003) 206–215
209
2. Materials and methods
2.1. Synthesis of 4,5-dihydroxy-3-methoxycinnamic acid (3-(4,5-dihydroxy-3-meth-
oxyphenyl) prop-2-enoic acid)
A solution containing 3,4,5-trihydroxybenzaldehyde (500 mg, 3.2 mmol) and ma-
lonic acid (2 g) dissolved in pyridine (10 mL) was heated to 55 °C until complete dis-
solution occurred. Aniline and piperidine (both freshly distilled; 0.1 mL) were added
to this solution and the reaction stirred at room temperature for 1 h. The mixture
was left to stand overnight and then neutralised with HCl (3 M). The pyridine was
removed by co-distillation with toluene and the oily residue was then redissolved
in ethyl acetate. The organic layer was washed with saturated aq. NaCl, left to stand
over Na2SO4 (anhydrous), filtered, and the solvent evaporated to afford 3,4,5-
trihydroxycinnamic acid as a yellow solid. Yield 445 mg (70%); NMR dH
ððCD3Þ SOÞ 6.11 (1H, d, J 15, C(7)H), 6.59 (2H, s, C(2,6)H), 7.35 (1H, d, J 15,
2
C(8)H) ppm and dC ððCD3Þ SO4Þ 107.5 (C2,6), 115.2 (C8), 124.7 (C1), 136.2 (C4),
2
145.0 (C3,5), 146.2 (C7), 167.8 (C9) ppm. Subsequently, 3,4,5-trihydroxycinnamic
acid (105 mg, 0.5 mmol) was dissolved in aqueous sodium tetraborate (5% w/v;
80 mL) and stirred under nitrogen at room temperature. Dimethyl sulphate (3 mL)
and NaOH (5 mL, 7.5 M) were added over a period of ca. 3 h. The solution was then
left to stand at room temperature overnight. The solution was acidified to pH 1 with
H2SO4 (20% v/v) and partitioned between ethyl acetate and saturated aq. NaCl. The
organic layer was left to stand over Na2SO4 (anhydrous), filtered, and the solvent
evaporated under reduced pressure. The product was then purified by flash chroma-
tography on silica [Merck 109385], eluting with ethyl acetate/hexane 2:3 (33 Â 7 mL),
ethyl acetate (11 Â 7 mL), and methanol (33 Â 7 mL). Yield 31.5 mg (30%); NMR
dH ðCD3Þ SO4Þ 3.79 (3H, s, OCH3) 6.24 (1H, d, J 15.9, C(7)H), 6.69 (1H, d, J 1.7
2
C(6)H), 6.81 (1H, d, J 1.7, C(2)H), 7.39 (1H, d, J 15.9, C(8)H) ppm and
dC ððCD3Þ SO4Þ 56.0 (OCH3), 104.0 (C2), 109.7 (C6), 115.8 (C8), 124.7 (C1),
2
137.0 (C4), 144.8 (C7), 145.8 (C5), 148.6 (C3), 167.9 (C9) ppm. The correlation of
methoxyl protons to C3 was confirmed by an heteronuclear multiple bond connec-
tivity (HMBC) NMR experiment.
Alternatively, 4,5-dihydroxy-3-methoxycinnamic acid can be prepared by the fol-
lowing method. 3,4-Dihydroxybenzaldehyde (2.25 g, 16.3 mmol) was dissolved in
acetic acid (60 mL) and bromine (0.7 mL) was added over a period of 2 h. The solu-
bility increased initially with the addition of bromine, but began to precipitate as re-
action neared completion. The reaction mixture was then poured into an ice-water
mixture (35 mL) and the light brown precipitate collected and dried. Yield 2.4 g
(67%). Sodium (freshly cut; 0.9 g) was dissolved in methanol (anhydrous; 15 mL)
and the volume of methanol reduced to 5 mL. 5-Bromo-3,4-dihydroxybenzaldehyde
(1.3 g, 6 mmol) and copper(II) chloride (0.32 g) were dissolved in N,N-dimethylfor-
mamide (anhydrous; 6 cm3) and this solution was added to the sodium methoxide
solution and refluxed for 2 h. After most of the methanol was removed, water
(15 mL) was added and the solution acidified with HCl (3 M). This solution was then
extracted into ethyl acetate and washed with sodium hydrogen sulphite (10% w/v).

210
W.R. Russell et al. / Bioorganic Chemistry31 (2003) 206–215
The aqueous layer was then acidified with sulphuric acid (3 M) and extracted into
ethyl acetate. The organic layer was left to stand over anhydrous Na2SO4, filtered,
and evaporated under reduced pressure to afford 4,5-dihydroxy-3-methoxy-benzal-
dehyde. Yield 0.72 g (71%). Condensation with malonic acid was achieved by treat-
ment of 4,5-dihydroxy-3-methoxy-benzaldehyde (0.54 mg, 3.2 mmol) in the same
way as indicated above for 3,4,5-trimethoxybenzaldehyde. The product was then pu-
rified by dry-column chromatography on silica [Merck 1.07736], eluting with ethyl
acetate/hexane 3:2 + 1% acetic acid (11 Â 7 mL) and ethyl acetate (11 Â 7 mL). Yield
0.63 g (94%); The NMR spectrum corresponded to that reported above.
2.2. EPR spectroscopy
Solutions were prepared to contain the substrate (5.0 lmol), peroxidase
(0.25 mmol), and hydrogen peroxide (1.0 mmol) in zinc acetate buffer (pH 5.5;
0.2 M; 2 mL). These solutions were applied by autosampler and the spectra (X-band)
recorded at 298 K in a Bruker EPR cell positioned in the cavity of a Bruker E106
spectrometer using the following instrument settings: modulation frequency,
100 kHz; centre field, 3471.5 Gauss; sweep width, 50 Gauss; time constant,
40.96 ms; sweep time 81.92 s and power, 20 mW. Continuous-flow EPR spectroscopy
was performed as reported previously [10]. Computer simulations of spectra, giving
the hydrogen hyperfine coupling constants (aH), were performed using the SIMEPR
program [11], which sequentially varies all the parameters for each radical species
until a minimum in the error surface is located. The best fit was determined by visual
comparison and as a minimum in the sum of the squared residuals.
2.3. Treatment of substrates with peroxidase and hydrogen peroxide
3,4-Dihydroxycinnamic acid (100.8 mg, 0.56 mmol) was dissolved in phosphate
buffer (88 mL; pH 6.5; 25 mM) to give solution A. Solution Bwas prepared by dis-
solving hydrogen peroxide (0.076 mL, 0.67 mmol, 1.2 eq) in phosphate buffer (88 mL;
pH 6.5; 25 mM). Both A and B were added to a solution of horseradish peroxidase
(275 U, 2.4 mg) in phosphate buffer (24 mL; pH 6.5; 25 mM) at a rate of 2 mL hÀ1 un-
der nitrogen with continuous stirring. After 24 h a further aliquot of peroxidase was
added (125 U, 1.1 mg). The pH was reduced with HCl (3 M) and partitioned between
ethyl acetate and saturated aq. NaCl. The organic layer was left to stand over
Na2SO4, filtered, and evaporated to dryness. Yield based on molecular weight of
starting material, 93.7 mg (93%). 4,5-Dihydroxy-3-methoxycinnamic acid was trea-
ted in the same way. Yield based on molecular weight of starting material,
108.2 mg (92%). NMR (major product) dH ððCD3Þ SO4Þ 3.71 (3H, s, AOCH3),
2
3.81 (3H, s, BOCH3) 5.06 (1H, d J 5.5, A(7)H), 5.22 (1H, d, J 5.5 A(8)H), 6.41
(1H, d, J 15.6, B(8)H), 6.45 (1H, d, J 2.3, A(2)H), 6.54 (1H, d, J 2.3, A(6)H), 6.90
(1H, d, J 1.8, B(6)H), 6.95 (1H, d, J 1.8, B(2)H), and 7.45 (1H, d, J 15.6, B(7)H)
ppm and dC ððCD3Þ SO4Þ 55.74 (AOCH3), 55.86 (BOCH3), 74.67 (A8), 74.86 (A7),
2
102.87 (A2), 104.58 (B2), 107.98 (B6), 110.27 (A6), 117.58 (B8), 125.80 (A1),
126,86 (B1), 133.78 (B4), 134.42 (A4), 143.79 (B7), 145.49 (A5), 145.53 (A3),

W.R. Russell et al. / Bioorganic Chemistry31 (2003) 206–215
211
148.01 (B3), 148.38 (B5), 167.71 (B9), and 168.67 (A9) ppm. All correlations were
confirmed by heteronuclear multiple quantum coherence (HMQC) and heteronu-
clear multiple bond connectivity (HMBC) experiments.
3. Results and discussion
The presence of the high-spin iron (III) signal ðg ꢀ 6Þ indicates the regeneration
of the resting state of the enzyme following the oxidation of the substrates by its
H2O2 activated forms, compound I and compound II, both of which contain iron
(IV). The signal observed at g ¼ 2:003 was similar to that observed at 4 K (data
not shown). At the lower temperature this signal is comprised of two components,
one of which is the porphyrin p-cation radical of compound I [12]. The other com-
ponent is believed to be a free radical located on the protein. At this temperature
(125 K), this signal is considered more likely to be solely from the latter component.
It is, however, interesting to note that this signal decreases with increasing extent of
hydroxylation and methoxylation and is representative of reduction by compounds
with increasing oxidation potential. For all substrates studied, the iron (III) signal
was regenerated. This demonstrated that replacement of a hydroxy substituent with
a methoxyl group had little effect on the ability of parent radical formation. From
this observation it appeared that the catalytic action of peroxidase to oxidise phe-
nylpropanoid substrates was not impaired by the presence of a catechol functional
group.
Since it appeared that radical formation was not inhibited by the presence of 1,2-
dihydroxy substituents, it was necessary to characterise the electron-spin distribu-
tion, as this has been shown previously to determine the product of phenylpropanoid
coupling [10]. Using CeIV as an oxidant, in the continuous-flow system used previ-
ously for similar substrates [10], failed to generate a stable radical from 3,4-dihydr-
oxycinnamic acid and no EPR spectrum was observed. However, incubation of this
substrate with peroxidase and hydrogen peroxide in the presence of ZnII ions permit-
ted the observation of an EPR spectrum showing good signal-to-noise ratio (Fig. 3).
In this system, ZnII serves a spin stabilisation agent [13]. In order to simulate the
spectrum, two radical species were required. Species A, which made up 27% of the
signal and species B, which represented 73% of the signal. This gave an r2 value of
0.988 and from this simulation, hyperfine coupling constants were obtained (Table
1). Due to the presence of two radical species, distribution of the unpaired electron
was difficult to assign and therefore structures generated computationally to repre-
sent the unpaired electron distribution were compared (Fig. 4).
The structure representing the radical generated on the hydroxyl at C4 showed
potential coupling positions at C5, C8 and through the oxygen at C4, as indicated
in green. When the radical is generated on the hydroxyl at C3 the potential coupling
positions become C6, C7 and through the oxygen at C3. The same patterns are ob-
served when the compound is methoxylated at C3. Therefore, to determine the actual
coupling regiochemistry both compounds were incubated with peroxidase and hy-
drogen peroxide.

212
W.R. Russell et al. / Bioorganic Chemistry31 (2003) 206–215
Fig. 3. Zinc-stabilised, EPR spectrum (A) and simulation (B) of 3,4-dihydroxycinnamic acid in the pres-
ence of peroxidase and hydrogen peroxide. The centre field is at 3479 Gauss with a spectral width of
40 Gauss.
Table 1
Hyperfine coupling constants (aH) derived from the computer-simulated zinc stabilised EPR spectrum of
3,4-dihydroxycinnamic acid for the two radical species A and B
aH
aH
aH
aH
A
B2.348
3.214
3.702
0.628
0.555
2.275
0.018
0.018
Fig. 4. Graphical distribution of unpaired-electron density calculated for geometrically optimised struc-
tures calculated by AM1 for 3,4-dihydroxy cinnamic acid with the radical generated on the hydroxyl at
C4 (A) and with the radical generated on the hydroxyl at C3 (B). Positive spin densities are represented
in green.

W.R. Russell et al. / Bioorganic Chemistry31 (2003) 206–215
213
Incubation of 3,4-dihydroxycinnamic acid with peroxidase and hydrogen perox-
ide, under the conditions generally used for the production of dehydrogenation poly-
mers [14], failed to give any coupling products and the starting material was
recovered unchanged. Although the catechol functional group did not prevent rad-
ical formation as shown by EPR spectroscopy, coupling of the parent radical did not
proceed. However, it should be noted that this observation occurs only under en-
zyme-catalysed oxidative coupling conditions, as this is in contrast to results
observed under acidic or pyrolysis conditions [15] and in systems modelling the
non-enzymic auto-oxidation of 3,4-diydoxycinnamic acid [16]. This may also explain
the possible presence of dimers in suspension-cultured cell walls [17]. The failure to
detect radicals from 3,4-dihydroxycinnamic acid in the continuous-flow EPR system,
under conditions suitable for detection of phenylpropanoids which do not contain
1,2-dihyroxyl substituents, suggests that any radicals formed are rapidly removed
from the system. One possible explanation is that the radicals are rapidly converted
into the o-quinone, as previously observed by electrochemical methods and pulse ra-
diolysis [18]. Recording of an EPR spectrum in the presence of zinc indicated forma-
tion of two different radical species. Whether the mechanism of o-quinone formation
is by two-step oxidation of 3,4-dihydroxycinnamic acid (Eq. (1)) or disproportion-
ation of the o-semiquinone (Eq. (2)) is uncertain and almost certainly dependent
on reaction conditions. Although this reaction is reversible, formation of the semi-
quinone radical is not thermodynamically favourable [19]. Therefore, intermolecular
radical-coupling reactions are unlikely to occur.
PhðOHÞ ! PhðOHÞOÅ ! PhðOÅÞ ! PhðOÞ
ð1Þ
ð2Þ
2
2
2
2PhðOHÞ ! 2PhðOHÞOÅ ! 2PhOÀOÅ ! PhðOÞ þ PhðOHÞ
2
2
2
EPR demonstrated that further substitution with hydroxyl groups increases the
ability of the compounds to reduce peroxidase compound I, as did additional meth-
oxyl substitution (Fig. 2). When 4,5-dihydroxy-3-methoxycinnamic acid was incu-
bated with peroxidase and hydrogen peroxide, radical coupling was observed. The
major product was a dimer, which resulted from 8-O-3 radical coupling and forma-
tion of a substituted 2,3-dihydro-1,4-dioxin ring (Fig. 5). This demonstrates that
coupling regiochemistry favours the pattern of positive spin density observed for
the parent radical generated on the hydroxyl at C4 (Fig. 4). Incorporation of these
units into lignin would result in formation of this linkage type and 13C NMR spec-
troscopy of lignin extracted from OMT downregulated plants, showed the presence
of unassigned signals in the same spectral region [20,21] and later in caffeic acid
OMT deficient plants [22]. The minor products were not identified. However, their
high solubility in ethyl acetate suggested they were not polymeric in nature. As well
as increasing the ease of oxidation, additional methoxyl substituents will increase the
stability of the parent radical. These features may account for the ability of 4,5-di-
hydroxy-3-methoxycinnamic acid to participate in radical coupling reactions. It also
explains the detection of this unit and not 3,4-dihydroxylated units in the cell walls of
plants in which OMT activity has been reduced. This indicates the significant effect
substrate structure, as well as mechanism of dimerisation, has on the incorporation

214
W.R. Russell et al. / Bioorganic Chemistry31 (2003) 206–215
Fig. 5. Structure of dimer resulting from 8-O-3 radical coupling of 4,5-dihydroxy-3-methoxycinnamic acid
and formation of a 2,3-dihydro-1,4-dioxin ring.
of phenylpropanoids into the plant cell wall. It also highlights the importance for
polymer formation, of hydroxyl methylation at C3 and C5 prior to reduction and
incorporation into lignin.
Acknowledgments
Financial support for this work was provided by the Scottish Executive, Rural
Affairs Department (formally the Scottish Office, Agriculture, Environment and
Fisheries Department).
References
[1] R.C. Bugos, V.L.C. Chiang, W.H. Campbell, cDNA cloning sequence analysis and seasonal
expression of lignin bispecific caffeic acid/5-hydroxyferulic acid O-methyl transferase of aspen, Plant
Mol. Biol. 17 (1991) 1203–1215.
[2] P. Collazo, L. Montoliu, P. Puigdomemech, J. Rigau, Structure and expression of the lignin
O-methyltransferase gene from Zea mays, Plant Mol. Biol. 20 (1992) 857–867.
[3] B. Dumas, J. Van Doorsselaere, J. Geilen, M. Legrand, B. Fritig, M. Van Montagu, D. Inze,
Nucleotide sequence of a complementary DNA encoding O-methyltransferase, Plant Physiol. 98
(1992) 796–797.
[4] O. Poeydomenge, A.M. Boudet, J.A. Grima-Pettenati, cDNA encoding S-adenosyl methionine
caffeic 3-O-methyltranferase from Eucalyptus, Plant Physiol. 105 (1994) 749–750.
[5] W.T. Ni, N.L. Paiva, R.A. Dixon, Reduced lignin in transgenic plants containing a caffeic acid
O-methyltransferase antisense gene, Transgenic Res. 3 (1994) 120–126.
ꢀ
ꢀ
[6] C. Lapierre, M.-T. Tollier, B. Monties, Mise en evidence dÕun noveau type dÕunite constitutive dans
les lignines dÕun mutant de Mais nm3, C. R. Acad. Sci. Paris 307 (III) (1988) 723–728.
[7] J.M. Harkin, J.R. Obst, Lignification in trees: Indication of exclusive peroxidase participation,
Science 180 (1973) 296–298.
[8] S. Sarkanen, R. Razal, T. Piccariello, E. Yamamoto, N.G. Lewis, Lignin peroxidase: Toward a
clarification of its role in vivo, J. Biol. Chem. 66 (1991) 3636–3643.
[9] R. Sterjiades, J.E.D. Dean, K.-E.L. Eriksson, Laccase from sycamore maple (Acer pseudoplatanus)
polymeriizes monolignols, Plant Physiol. 99 (1992) 1162–1168.

W.R. Russell et al. / Bioorganic Chemistry31 (2003) 206–215
215
[10] W.R. Russell, A.R. Forrester, A. Chesson, M.J. Burkitt, Oxidative coupling during lignin
polymerization is determined by unpaired electron delocalization within parent phenylpropanoid
radicals, Arch. Biochem. Biophys. 2 (1996) 357–366.
[11] D.R. Duling, Simulation of multiple isotropic spin-trap EPR spectra, J. Magn. Res. 104 (1992) 105–
110.
€
ꢁ
[12] R. Aasa, T. Vanngard, H.B. Dubford, EPR studies on compound I of horseradish peroxidase,
Biochim. Biophys. Acta 391 (1975) 259–264.
[13] B. Kalyanaraman, C.C. Felix, R.C. Sealy, Electron spin resonance stabilisation of semiquinones
produced during oxidation of epinephrine and its analogues, J. Biol. Chem. 259 (1984) 354–358.
€
[14] K. Freudenberg, K. Jones, H. Renner, Kunstliches lignin aus markiertem coniferyl alkohol, Berichte
der Deutschen Chemischen Gesellschaft 96 (1963) 1844–1848.
[15] H. Fulcrand, A. Cheminat, R. Brouillard, V. Cheynier, Characterisation of caffeic acid oxidation
ꢀ
products, in: R. Bouilliard, M. Jay, A. Scalbert (Eds.), Polyphenols 94 XVII Journees Internationales
ꢀ
Groupe Polyphenols. INRA Editions, Paris, 1995, 157–158.
[16] J.L. Cilliers, V.L. Singleton, Characterization of the products of nonenzymic autoxidative browning
reactions in a caffeic acid model system, J. Agric. Food Chem. 39 (1991) 1298–1303.
[17] J.-G. Yang, T. Uchiyama, Dehydrodimers of caffeic acid in the cell walls of suspension-cultured
Mentha, Biosci. Biotechnol. Biochem. 64 (2000) 862–864.
[18] P. Hapiot, A. Neudeck, J. Pinson, H. Fulcrand, P. Neta, C. Rolando, Oxidation of caffeic acid and
related hydroxycinnamic acids, J. Electroanal. Chem. 405 (1996) 169–176.
[19] E. Laviron, Electrochemical reactions with protonations at equilibrium. Part XII, J. Electroanal.
Chem. 169 (1984) 29–46.
[20] G.J. Provan, L. Scobbie, A. Chesson, Characterisation of lignin from CAD and OMT deficient Bm
mutants of maize, J. Sci. Food Agric. 73 (1997) 133–142.
[21] J. Marita, J. Ralph, R.D. Hatfield, C. Chapple, NMR characterisation of lignins in Arabidopsis
altered in activity of ferulate-5-hydroxylase, Proc. Natl. Acad. Sci. USA 96 (1999) 12328–12332.
[22] J. Ralph, C. Lapierre, J.M. Marita, H. Kim, F. Lu, R.D. Hatfield, S. Ralph, C. Chapple, R. Franke,
M.R. Hemm, J.V. Doorsselaere, R.R. Sederoff, D.M. OÕMalley, J.T. Scott, J.J. MacKay, N.
Yahiaoui, A.-M. Boudet, M. Pean, G. Pilate, L. Jouanin, W. Boerjan, Elucidation of new structures
in lignins of CAD and COMT-deficient plants by NMR, Phytochemistry 57 (2001) 993–1003.