NMR analysis of lignins in CAD-deficient plants. Part 1. Incorporation of hydroxycinnamaldehydes and hydroxybenzaldehydes into lignins

Article abstract of DOI:10.1039/b209686b

Peroxidase/H₂O₂-mediated radical coupling of 4-hydroxycinnamaldehydes produces 8-O-4-, 8-5-, and 8-8-coupled dehydrodimers as has been documented earlier, as well as the 5-5-coupled dehydrodimer. The 8-5-dehydrodimer is however produced kinetically in its cyclic phenylcoumaran form at neutral pH. Synthetic polymers produced from mixtures of hydroxycinnamaldehydes and normal monolignols provide the next level of complexity. Spectral data from dimers, oligomers, and synthetic polymers have allowed a more substantive assignment of aldehyde components in lignins isolated from a CAD-deficient pine mutant and an antisense-CAD-downregulated transgenic tobacco. CAD-deficient pine lignin shows enhanced levels of the typical benzaldehyde and cinnamaldehyde end-groups, along with evidence for two types of 8-O-4-coupled coniferaldehyde units. The CAD-downregulated tobacco also has higher levels of hydroxycinnamaldehyde and hydroxybenzaldehyde (mainly syringaldehyde) incorporation, but the analogous two types of 8-O-4-coupled products are the dominant features. 8-8-Coupled units are also clearly evident. There is clear evidence for coupling of hydroxycinnamaldehydes to each other and then incorporation into the lignin, as well as for the incorporation of hydroxycinnamaldehyde monomers into the growing lignin polymer. Coniferaldehyde and sinapaldehyde (as well as vanillin and syringaldehyde) co-polymerize with the traditional monolignols into lignins and do so at enhanced levels when CAD-deficiency has an impact on the normal monolignol production. The implication is that, particularly in angiosperms, the aldehydes behave like the traditional monolignols and should probably be regarded as authentic lignin monomers in normal and CAD-deficient plants.

Full text of DOI:10.1039/b209686b

View Article Online / Journal Homepage / Table of Contents for this issue
NMR analysis of lignins in CAD-deficient plants. Part 1.
Incorporation of hydroxycinnamaldehydes and
hydroxybenzaldehydes into lignins
Hoon Kim,a,b John Ralph,*a,b Fachuang Lu,a,b Sally A. Ralph,c Alain-M. Boudet,d
John J. MacKay,e Ronald R. Sederoff,f Takashi Ito,g Shingo Kawai,g Hideo Ohashig and
Takayoshi Higuchih
a US Dairy Forage Research Center, USDA Agricultural Research Service, Madison WI 53706,
USA
b Department of Forestry, University of Wisconsin, Madison, WI 53706-1598, USA
c US Forest Product Laboratory, USDA-Forest Service, Madison, WI 53705, USA
d UMR 5546 Pôle de Biotechnologie Végétale, 24 Chemin de Borde Rouge, BP17 Auzeville,
F31326 Castanet Tolosan, France
e Centre de Recherche en Biologie Forestière, Université Laval, Québec, Québec G1K 7P4,
Canada
f Department of Genetics, North Carolina State University, Raleigh, NC 27695, USA
g Department of Applied Bioorganic Chemistry, Faculty of Agriculture, Gifu University,
1-1 Yanagido, Gifu 501-1193, Japan
h Professor Emeritus, Kyoto University, Kyoto 611-0011, Japan
Received 4th October 2002, Accepted 7th November 2002
First published as an Advance Article on the web 11th December 2002
Peroxidase/H2O2-mediated radical coupling of 4-hydroxycinnamaldehydes produces 8–O–4-, 8–5-, and 8–8-coupled
dehydrodimers as has been documented earlier, as well as the 5–5-coupled dehydrodimer. The 8–5-dehydrodimer is
however produced kinetically in its cyclic phenylcoumaran form at neutral pH. Synthetic polymers produced from
mixtures of hydroxycinnamaldehydes and normal monolignols provide the next level of complexity. Spectral data
from dimers, oligomers, and synthetic polymers have allowed a more substantive assignment of aldehyde components
in lignins isolated from a CAD-deficient pine mutant and an antisense-CAD-downregulated transgenic tobacco.
CAD-deficient pine lignin shows enhanced levels of the typical benzaldehyde and cinnamaldehyde end-groups, along
with evidence for two types of 8–O–4-coupled coniferaldehyde units. The CAD-downregulated tobacco also has
higher levels of hydroxycinnamaldehyde and hydroxybenzaldehyde (mainly syringaldehyde) incorporation, but the
analogous two types of 8–O–4-coupled products are the dominant features. 8–8-Coupled units are also clearly
evident. There is clear evidence for coupling of hydroxycinnamaldehydes to each other and then incorporation
into the lignin, as well as for the incorporation of hydroxycinnamaldehyde monomers into the growing lignin
polymer. Coniferaldehyde and sinapaldehyde (as well as vanillin and syringaldehyde) co-polymerize with the
traditional monolignols into lignins and do so at enhanced levels when CAD-deficiency has an impact on the
normal monolignol production. The implication is that, particularly in angiosperms, the aldehydes behave like
the traditional monolignols and should probably be regarded as authentic lignin monomers in normal and
CAD-deficient plants.
coniferaldehyde dehydrodimers 5, but insufficient NMR data
for current structural studies were available.
Introduction
Aldehydes are well-known components in lignins,1–10 and are
Whether aldehydes are true components of lignin from
responsible for the characteristic phloroglucinol staining of
co-polymerization (radical cross-coupling) with monolignols/
lignified tissues.11 Relatively recent studies on plants deficient
in enzymes in the monolignol biosynthetic pathways have
oligolignols has recently become important to elucidate as the
lignins from various plants deficient in cinnamyl alcohol de-
renewed interest in the possible incorporation of 4-hydroxy-
hydrogenase (CAD; E.C. 1.1.1.195) are examined. In gymno-
cinnamaldehydes 1 (Fig. 1†) and their derived hydroxybenz-
aldehydes 2 into lignins, perhaps replacing some of the normal
monolignol 3 component.
sperms, in which the lignins are predominantly composed of
guaiacyl units 4G, CAD catalyses the final biosynthetic step
from coniferaldehyde 1G to coniferyl alcohol 3G. CAD has also
It has long been recognized that coniferaldehyde 1G is a
been detected and purified from various angiosperms,14–19
viable substrate for free-radical coupling reactions analogous to
where there are several isoform groups, named CAD1, CAD2,
those that occur with the standard monolignols 3.5,12 It is pos-
CAD1P, and CAD2P. These isozymes differ in amino acid
sible to make coniferaldehyde synthetic lignins (DHPs) for
sequence, molecular weight, and substrate specificity,20 and
example.5,13 Connors et al.12 prepared and purified some of the
each group has isozymes that have different sub-units.16,21 The
most studied, CAD2, appears to be the main CAD enzyme in
the lignin biosynthetic pathway. It was believed that the same
† The IUPAC name for guaiacyl is 4-hydroxy-3-methoxyphenyl,
CAD2 existed in both gymnosperms and angiosperms, and that
syringyl is 4-hydroxy-3,5-dimethoxyphenyl, sinapaldehyde is 3,5-di-
CAD in gymnosperms was more specific for coniferaldehyde 1G
methoxy-4-hydroxycinnamaldehyde, coniferaldehyde is 4-hydroxy-3-
methoxycinnamaldehyde and coumaran is 2,3-dihydrobenzofuran.
whereas CAD in angiosperms had equal activities towards both
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
268

View Article Online
Fig. 1 A) Structures of monomeric, dimeric, oligomeric and polymeric units. The numbering system follows the convention for lignins. The A- and
B-rings are arbitrarily defined to allow the units to be discussed in the text. B) Units in lignins with undefined groups attached via possible coupling
reactions at their 4–O- and 5-positions. Units U and V, as well as A (for 8–O–4), B (for 8–5), C (for 8–8), and D (for 5–5) are taken from the unit labels
established in the recent review “Solution-state NMR of Lignins”.40 Smaller qualifiers are: G = guaiacyl, S = syringyl, P = p-hydroxyphenyl, c = cyclic
(phenylcoumaran) form, o = open (non-cyclic) form; in dimeric units, the “A-ring” unit is first —thus 7ASG, for example, is from sinapaldehyde
coupled at its 8-position to a guaiacyl unit coupled at its 4–O-position.†
coniferaldehyde 1G and sinapaldehyde 1S.22,23 However, Chiang
and co-workers have recently discovered a sinapyl alcohol de-
hydrogenase (SAD) from aspen which has 60 times greater
enzymatic efficiency for sinapaldehyde 1S than the known CAD
enzyme.24 We will continue to use CAD here as the generic term
for cinnamyl alcohol dehydrogenase.
Although it has been suggested from flux studies using
suspension-cultured Pinus taeda that CAD should not be
rate limiting,25 down-regulation of CAD in a variety of
mutants and transgenics clearly leads to an accumulation of
hydroxycinnamaldehydes 1 (at the apparent expense of the
monolignols 3) and an apparent buildup of their content in
resultant lignins,1–5,26–31 although confirming the association of
these aldehydes with the polymeric lignin component has
sometimes been difficult;29,32 much of the aldehyde compon-
ent can remain as low molecular mass extractable compounds
whose location is uncertain.
After a tobacco-CAD enzyme was purified and characterized
and the gene obtained,16 tobacco CAD-downregulated trans-
genic plants were produced.3,28,33 The lignin levels of CAD-
downregulated transgenics were similar to those in the normal
tobacco, but the xylem had a red–brown coloration, a common
phenotype of CAD-deficiency. Lignin isolated from tobacco
CAD-downregulated transgenic plants with 8% residual CAD
activity was preliminarily characterized by NMR methods; the
appearance of aldehyde peaks corresponding to new aldehydic
structures in the lignin has been noted.27 More detailed NMR
studies, coupled with data derived from synthetic model com-
pounds, allowed details of the cross-coupling propensities of
coniferaldehyde and sinapaldehyde to be elucidated in vivo.34,35
The CAD enzyme in loblolly pine was also purified and charac-
terized,17 and a CAD-deficient mutant pine was discovered.36
This mutant had a similar Klason lignin level to normal pine,
despite CAD levels being less than 1% of those in the normal
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
269

View Article Online
pine. The CAD-downregulated pine mutant also had red–
brown colored xylem (see later in Fig. 4), and lignins were more
easily extractable in alkaline solution than from the control.
This pine mutant lignin was characterized by NMR studies26
prior to the NMR study of the tobacco CAD-downregulated
transgenics. Significantly enhanced DHCA (dihydroconiferyl
alcohol) levels were strikingly revealed; DHCA seemed to par-
tially offset the reduced coniferyl alcohol levels in the mutant’s
lignin. The original claims in that initial paper relating to this
component were challenged,37 but have since been validated as
more diverse data has been accumulated.35,38–40 The paper deal-
ing mainly with the elevated DHCA levels also began to discern
the nature of the more minor aldehyde components.26 However,
the aldehyde region in that pine still needed detailed structural
elucidation to clarify the aldehyde coupling and cross-coupling
reactions involved. Minor aldehyde peaks in the spectra were
originally assigned speculatively, but were definitively identi-
fied later.27,35,40 One of the main deductions from the initial
studies26,41 is that plants with curtailed abilities to produce
traditional monolignols appeared to be able to utilize other
phenolics at their disposal to produce polymeric “lignin” com-
ponents. This concept was controversial,42,43 but has garnered
increasing support with each new lignin-biosynthetic-pathway
mutant or transgenic that has been structurally examined.
Here we examine by NMR methods the dimerization and
oligomerization of coniferaldehyde 1G and sinapaldehyde 1S,
and the copolymerization of coniferaldehyde 1G with coniferyl
alcohol 3G to produce synthetic lignins (DHPs). Spectral data
from these model reactions are used to provide a more substan-
tive elucidation of the structures of aldehyde moieties in lignins
from CAD-deficient plants as illustrated using a gymnosperm
(the mutant pine) and an angiosperm (an antisense-CAD
transgenic tobacco). The result offers considerable new under-
standing of how hydroxycinnamaldehydes and hydroxybenz-
aldehydes are incorporated into various structures during
lignification.
8–β-cross-coupling of coniferaldehyde and coniferyl alcohol
radicals). As will be discussed below, the actual course of the
dehydrodimerization and oligomerization reactions was better
studied by examining crude reaction products (using 2D NMR)
of [9-13C]-labeled coniferaldehyde or sinapaldehyde. The next
level of detail comes from spectra of in vitro dehydropolymeri-
zation reactions. Finally, with the NMR data from all of these
model reactions, spectra from the two CAD-deficient plants, a
gymnosperm and an angiosperm, can be more fully interpreted.
The result provides a better understanding of the incorporation
of both hydroxycinnamaldehydes and hydroxybenzaldehydes
into guaiacyl and guaiacyl–syringyl lignins, and some insight
into the nature of the CAD-deficiency.
Hydroxycinnamaldehyde and hydroxybenzaldehyde end-groups
The only aldehyde structures in lignins that had been observed
in NMR spectra until recent studies on mutants and trans-
genics were the hydroxybenzaldehyde
U and hydroxy-
cinnamaldehyde V end-groups.6,10,45,46 These are therefore
considered to be the traditional aldehyde groups in lignins.
Hydroxycinnamaldehyde end-groups V (Fig. 1B) can arise from
the incorporation of an hydroxycinnamaldehyde monomer
1 into lignin by coupling at its 4–O- or 5-positions (generally
with a monolignol), or remain (as the “B-moieties”) follow-
ing homo-coupling of hydroxycinnamaldehydes. Hydroxy-
benzaldehyde units (producing benzaldehyde end-groups U in
Fig. 1B) result from hydroxybenzaldehydes 2.47 It is not known
whether they result directly from incorporation of hydroxy-
benzaldehyde monomers
2 (vanillin and syringaldehyde)
into lignin, or are produced post-lignification from hydroxy-
cinnamaldehyde V (Fig. 1B) units in lignin; the former is
strongly implicated in the current work, but is not necessarily
exclusive.
The chemical shifts of several key peaks of hydroxy-
cinnamaldehyde 1 and hydroxybenzaldehyde 2 monomers and
dehydrodimers are important to identify aldehyde end-group
structures in lignins (Tables 1 and 2). Coniferaldehyde 1G and
sinapaldehyde 1S have coincident 9-carbon peaks at ∼194 ppm;
the 7-proton chemical shifts are also similar (7.57 and 7.55
ppm). Since sinapaldehyde is symmetrical, the 2- and 6-carbons
are coincident at 107.4 ppm, but coniferaldehyde has different
chemical shifts for carbons 2 (112.0 ppm) and 6 (124.7 ppm). As
seen in Tables 1 and 2, these coniferaldehyde chemical shifts are
similar in 4–O–β and 5–β/4–O–α (phenylcoumaran) etherified
models for cross-coupled coniferaldehyde units in dimeric
models, and in the coniferaldehyde dehydrodimers 5AGG, 5BGG,
and 5DGG (the underline represents the terminal, or “B-unit” of
the dehydrodimer, see Tables 1 and 2). We don’t have similarly
etherified models for sinapaldehyde 1S, but suspect that the
crucial 9, 7, 2 and 6 shifts (that are seen in diagnostic 13C–1H
correlations, see later) also do not vary significantly. Vanillin 2G
and syringaldehyde 2S also have coincident 7-carbon peaks at
191.1 ppm, but syringaldehyde has its 2/6 proton peak at 7.15
ppm, whereas vanillin shows marginally different proton peaks
for protons 2 and 6 (7.41 and 7.44 ppm). Vanillin end-units UG
have more diagnostic shifts when 5-substituted (in 5–5, 5–O–4,
or presumably 5–β structures), so these types of units can
potentially be differentiated. However, the lack of separation
between syringyl vs guaiacyl hydroxycinnamaldehyde or
hydroxybenzaldehyde carbon shifts makes it clear that 1D
13C-NMR spectra will not be sufficient for distinguishing
between these components; 2D NMR methods exploiting other
(correlatable) differences are required.
Results and discussion
NMR data on aldehyde components were gathered from several
sources in order to provide the necessary database for assigning
the various structures in isolated lignins. It is important to
emphasize, as pointed out in a recent review on “Solution-state
NMR of Lignins”,40 that 13C-NMR data from “sufficiently
good” model compounds will be the same as the corresponding
data for such units in lignins (in the same solvent). This is
because 13C-NMR chemical shifts are only influenced by
through-bond, and not through-space interactions except in the
1
severe case of steric compression. Conversely, H-NMR chem-
ical shifts can be strongly influenced by through-space as well as
through-bond interactions. Proton chemical shifts therefore
need not be the same in simple models as they are in the poly-
mer. For most lignin units, the degree of agreement has been
remarkably good, implying that the shape of the molecule in
the models is similar in the polymer.40,44 However 1D proton
shifts should be augmented with 2D correlation data to validate
structural assignments.
Simple monomers (Tables 1 and 2) already provide the
required data for modeling hydroxycinnamaldehyde and
hydroxybenzaldehyde end-groups. The dehydrodimers provide
better end-group models (the “B-ring” in dehydrodimers
5A–5B, as well as 5D, Fig. 1) for the 4–O- and/or 5-linkages
modeled. The dehydrodimers also begin to provide the requisite
detail of particularly the 8-linked aldehydes. In many ways,
these are the most important models, since lignification is prin-
cipally concerned with monomers reacting predominantly at
their 8- or β-positions with the 4–O- or 5-positions of the grow-
ing lignin polymer. The 8–8-linked dehydrodimer 5C is a model
only for homocoupling reactions, although it might be expected
to adequately model the cross-coupled unit 7CЈ (e.g. from the
Hydroxycinnamaldehyde dimers (and pathways from quinone
methide intermediates)
Structures 5 (Fig. 1A) resulted from homo-coupling of
hydroxycinnamaldehydes. The coniferaldehyde dehydrodimers
except structures 5D, 5BcGG (see below) and the trimer 6GGG
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
270

View Article Online
Table 1 1H-NMR data for (un-acetylated) hydroxybenzaldehyde and hydroxycinnamaldehyde monomers (acetone-d6)
δH
Modela
DB#b
7
8
9
2
5
6
OMe
ArOH
1P, p-Coumaraldehyde
152
7.58
(d, 15.8)
7.60
7.57
(d, 15.8)
7.58
6.62
(dd, 15.8, 7.7)
6.65
6.67
(dd, 15.8, 7.8)
6.79
9.64
(d, 7.7)
9.66
9.64
(d, 7.8)
9.65
9.66
9.63
9.65
9.63
6.94
(m)
7.02
7.38
(d, 2.0)
7.38
7.40
7.29
7.36
7.61
(H-3,5, m)
7.68
6.92
(d, 8.2)
7.11
7.15
—
—
6.94
—
9.00
(br s)
—
8.23
(br s)
—
(m)
1P–(4-OMe)
1G, Coniferaldehyde
215
144
7.02
7.21
(dd, 8.2, 2.0)
7.27
—
3.93
(s)
3.90
3.91
3.91
3.95
3.90
(s)
1G–(4–O–β)e
1G–(4–O–β)t
1G–(5–β/4–O–α)
1G–(5–5)
3011
3010
2021
3033
153
7.59
7.59
7.59
7.55
6.70
6.65
6.68
6.69
7.25
7.32
7.27
7.08
—
—
c
?
1S, Sinapaldehyde
7.08
8.00
(br s)
9.40
(br s)
—
8.30
(br s)
—
(d, 15.8)
9.84
(dd, 15.8, 7.7)
(d, 7.7)
(s)
(s)
2P, p-OH-Benzaldehyde
14
7.79
(d, 8.7)
7.85
7.46
(d, 1.8)
7.46
7.42
7.58
7.40
7.26
7.00
(d, 8.7)
7.08
7.00
(d, 8.6)
7.15
—
7.79
(d, 8.7)
7.85
(s)
2P–(4–OMe)
2G Vanillin
203
15
9.86
9.81
—
—
—
—
—
3.91
(s)
7.43
(s)
(m)
2G–(4–O–α)
258
3061
71
2049
2049
3062
3062
41
9.81
9.82
9.96
9.86
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
7.39
7.53
7.44
6.88
7.01
7.74
7.70
7.21
3.97
3.92
4.01
3.90
3.94
4.03
3.91
3.90
(s)
3.95
3.99
(s)
3.92
(s)
3.96
(s)
3.95
(s)
3.75
(s)
3.82
(s)
3.47
(s)
3.70
(s)
3.72
(s)
3.69
(s)
3.83
(s)
3.98
(s)
2G–(5–β/4–O–α)
2G–(4–O–Me/5–5)
2G–(4–O–5/5–O–4)
2G–(5–O–4)
—
—
—
—
—
—
—
—
—
c
9.74
?
2G–(5–5/4–O–β)
2G–(5–5/4–O–α)
2S, Syringaldehyde
10.06
10.02
9.81
7.65
7.56
—
—
8.23
(br s)
—
7.21
(s)
(s)
(s)
2S–(4–O–β)
5AGG {1G–(4–O–8)}
228
3031
9.90
7.58
(d, 16.0)
7.61
(d, 15.8)
7.57
(d, 15.8)
7.59
(d, 15.8)
7.32
—
6.70
(dd, 16.0, 7.7)
6.70
(dd, 15.8, 7.7)
6.66
(dd, 15.8, 7.8)
6.68
—
9.65
(d, 7.7)
9.64
(d, 7.7)
9.61
(d, 7.8)
9.63
(d, 7.8)
9.52
(s)
9.94
(s)
9.69
(s)
9.66
(s)
9.67
(s)
7.25
7.49
(d, 2.0)
7.38
—
6.82
(d, 8.3)
—
7.25
7.15
(dd, 8.3, 2.0)
7.43
—
5BcGG {1G–(5–8/4–O–7)}
5BoGG {1G–(5–8)}
5DGG {1G–(5–5)}
5AGG (8–O–4)
3059
3030
3033
3031
3059
3030
3032
3060
3058
3058
3058
—
—
—
(s)
(s)
7.44
—
—
7.01
(d, 2.0)
7.27
(d, 2.0)
7.33
(dd, 8.3, 2.0)
6.90
(dd, 8.2, 1.8)
6.98
(dd, 8.3, 2.0)
7.21
(dd, 8.3, 2.0)
7.02
(d, 2.0)
7.36
(d, 2.0)
7.58
(d, 2.0)
7.06
(d, 1.8)
6.87
(d, 2.0)
7.28
(d, 2.1)
7.02
(dd, 15.8, 7.8)
—
6.88
(d, 8.3)
6.84
(d, 8.2)
6.77
(d, 8.3)
6.82
(d, 8.3)
—
8.32
(br s)
8.29
(br s)
8.43
(br s)
8.29
(br s)
7.94
(br s)
8.37
(br s)
—
(s)
5BcGG (8–5/7–O–4)
5BoGG (8–5 open)
5CGG (8–8)
6.20
(d, 6.5)
7.57
4.50
(d, 6.5)
—
(s)
7.78
—
—
—
—
(s)
5CSS (8–8)
7.79
(s)
(s)
(s)
6GGG {5AGG–(8–O–4)}
6GGG {(4–O–8)–5AGG}
6GGG {1G–(4–O–8)ϩ}
7.30
9.48
(s)
9.53
(s)
9.65
(d, 7.7)
7.52
6.85
(d, 8.3)
6.79
(d, 8.3)
6.82
7.30
(dd, 8.3, 2.0)
7.28
(dd, 8.3, 2.0)
7.15
(dd, 8.3, 2.0)
(s)
(d, 2.0)
7.67
(d, 2.0)
7.48
7.33
(s)
7.58
(d, 15.9)
6.70
(dd, 15.9, 7.7)
—
(d, 2.0)
(d, 8.3)
a See Fig. 1 for compound numbers and P, G, S designation. In dimers or trimers, ring designations are left to right as drawn; the underlined unit is
the one with the data shown—e.g. 6GGG is data for the first (free-phenolic) moiety of trimer 6. b Data are from, or have been entered into, the “NMR
Database of Lignin and Cell Wall Model Compounds.”66 Compound sources are given there. c Data should exist but are not seen and are not
reported in the Database.
were previously isolated by Connors et al.,12 but insuffi-
cient NMR data were provided. Quinone methide inter-
mediates have important roles in determining the final
products of hydroxycinnamaldehyde coupling, just as they
are important intermediates in the biosynthesis and degrad-
ation of lignin.48–52 When coniferaldehyde 1G dimerizes with
one of the radicals coupling at its 8-position, the resulting
intermediate product is a quinone methide; this is analogous
to the dimerization of coniferyl alcohol with one of the radicals
coupling at its β-position.48 Addition of water to quinone
methides is involved when there are no favorable internal
trapping mechanisms, or there is no acidic 8-proton; water
therefore adds to the quinone methide β–O–4-coupling product
between two hydroxycinnamyl alcohols 3 or an hydroxy-
cinnamyl alcohol and a lignin dimer or higher oligomer. In the
case of the hydroxycinnamaldehydes 1, however, the quinone
methide has new options. The resultant 8-proton is par-
ticularly acidic (even more acidic than in ferulate analogs)
because of the aldehyde group; elimination of the 8-proton
allows re-aromatization. From coniferaldehyde 1G therefore,
the 8–O–4-dimer produced is 5AGG (Fig. 1A),12 without
addition of water at the 7-position of the quinone methide
intermediate—the addition of water to the quinone methide
cannot compete with the faster 8-proton elimination. An
8–O–4/8–O–4-trimer 6GGG was also isolated here, confirming
that 8–O–4-coupling to another monomer or to a preformed
dimer can occur. The 8–8-dimer 5C similarly regains 7,8-
unsaturation by 8-proton elimination from the two quinone
methide moieties in the same manner as the 8–O–4-dimer. The
products are both analogous to those produced by ferulate
where the intermediate quinone methides also have acidic
8-protons.53
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
271

View Article Online
Table 2 13C-NMR data for (un-acetylated) hydroxybenzaldehyde and hydroxycinnamaldehyde monomers (acetone-d6)
δC
Modela
DB#b
1
2
3
5
5
6
7
8
9
OMe
1P, p-Coumaraldehyde
1P–(4–OMe)
1G, Coniferaldehyde
1G–(4–O–β)e
1G–(4–O–β)t
1G–(5–β/4–O–α)
1G–(5–5)
1S, Sinapaldehyde
2P, p-OH-Benzaldehyde
2P–(4–OMe)
152
215
144
3011
3010
2021
3033
153
14
203
15
258
3061
71
2049
2049
3062
3062
41
126.7
127.9
127.5
130.2
129.9
131.2
126.1
126.2
130.3
131.1
130.6
132.1
132.4
133.2
134.8
128.7
134.4
134.6
129.0
133.3
130.2
129.7
127.4
126.1
125.3
132.2
127.4
127.5
126.1
125.2
128.0
130.4
131.5
131.3
111.7
112.9
112.6
113.6
110.0
107.4
132.8
132.4
110.8
111.4
113.4
111.9
108.0
108.4
111.7
111.9
107.7
107.2
112.6
114.3
110.6
110.0
114.0
110.7
113.6
113.7
109.1
114.0
114.9
112.6
116.8
115.4
150.8
152.1
151.7
145.7
149.4
149.0
116.6
115.2
148.9
151.6
145.7
153.0
155.0
149.7
154.2
154.4
148.9
154.4
150.4
146.0
149.3
149.4
148.5
147.9
148.1
148.4
148.7
148.5
149.9
150.4
161.2
163.1
148.8
151.2
151.7
152.4
126.1
140.3
163.8
165.4
153.5
153.4
154.9
154.3
138.3
143.0
152.5
152.8
142.9
141.8
148.9
151.8
148.6
126.1
150.6
148.6
150.2
150.5
140.1
150.6
148.5
148.7
116.8
115.4
116.2
118.5
117.9
129.0
149.6
149.0
116.6
115.2
115.9
115.9
131.2
133.0
153.7
147.2
133.3
133.7
148.9
154.4
115.3
125.8
136.3
149.6
116.3
115.9
116.0
116.2
148.7
116.3
115.2
115.4
131.5
131.3
124.7
123.9
123.8
119.6
127.1
107.4
132.8
132.4
127.0
126.0
121.4
127.2
109.9
110.2
125.3
125.2
107.7
107.2
123.6
119.5
126.4
127.1
126.8
119.8
126.9
126.3
109.1
126.8
125.6
123.6
153.6
153.2
153.9
153.5
153.3
154.1
154.4
154.2
191.0
191.1
191.1
191.3
190.9
191.5
191.7
191.0
191.7
191.7
191.1
191.7
153.2
153.5
153.8
154.4
138.2
85.2
127.0
127.4
127.0
128.5
128.2
127.1
126.8
127.3
—
—
—
—
—
—
—
—
—
—
—
—
128.3
127.7
127.3
126.8
147.7
62.9
123.2
134.4
134.8
147.5
148.5
128.4
193.8
193.8
193.8
194.1
193.6
193.8
193.9
193.7
—
—
—
—
—
—
—
—
—
—
—
—
193.9
193.8
194.0
193.9
187.5
197.6
193.4
192.7
192.7
187.4
187.7
193.9
—
—
56.4
56.4
56.4
56.5
56.5
56.7
—
56.0
56.3
56.3
56.4
56.4
56.9
56.8
56.6
56.3
56.6
56.6
56.5
56.5
56.6
56.5
56.0
56.3
55.6
56.1
56.5
55.9
56.1
56.5
2G, Vanillin
2G–(4–O–α)
2G–(5–β/4–O–α)
2G–(4–O–Me/5–5)
2G–(4–O–5/5–O–4)
2G–(5–O–4)
2G–(5–5/4–O–β)
2G–(5–5/4–O–α)
2S, Syringaldehyde
2S–(4–O–β)
5AGG {1G–(4–O–8)}
5BcGG {1G–(5–8)}
5BoGG {1G–(5–8)}
5DGG {1G–(5–5)}
5AGG (8–O–4)
228
3031
3059
3030
3033
3031
3059
3030
3032
3060
3058
3058
3058
5BcGG (8–5/7–O–4)
5BoGG (8–5 open)
5CGG (8–8)
151.4
152.8
153.0
138.3
137.2
153.1
5CSS (8–8)
6GGG {5AGG–(8–O–4)}
6GGG {(4–O–8)–5AGG}
6GGG {1G–(4–O–8)ϩ}
a See Fig. 1 for compound numbers and P, G, S designation. In dimers or trimers, ring designations are left to right as drawn; the underlined unit is
the one with the data shown—e.g. 6GGG is data for the first (free-phenolic) moiety of trimer 6. b Data is from, or has been entered into, the “NMR
Database of Lignin and Cell Wall Model Compounds.”66 Compound sources are given there.
Useful insight into the relative amounts and the nature of the
dimers was gained from small-scale (5–10 mg) experiments
using synthesized [9-13C]-labeled coniferaldehyde or sinapalde-
hyde. Assignments for well-characterized peaks could be readily
made from 1D 13C-NMR of the mixture which, because of the
[9-13C]-labeling, could be acquired in minutes, and from 2D
HMQC and HMBC experiments where the aldehyde carbonyl
region could be selectively and quickly acquired.
at ∼6.3 and ∼4.5 ppm, a diagnostic fingerprint for the phenyl-
coumaran structure (and agreeing with the data for the 5BcGG
dimer in Tables 1 and 2). Fig. 3B shows a more complete
HMBC spectrum of the crude product in which all of the corre-
lations for 5Bc/7Bc are highlighted. Extensive long-range
correlations between proton A7 and carbons A9, B4, A1, B5,
A6, A2, and A8 proves the cyclic nature of the product; such
long-range correlations are over either 2- or 3-bonds so the
correlations between proton A7 and carbons B4 and B5 are
diagnostic for the 5-membered phenylcoumaran ring. It is
therefore well established that, under the dehydrogenation con-
ditions used here, it is the cyclic form of the 8–5-dehydrodimer
5Bc that forms from coniferaldehyde.
This cyclic phenylcoumaran structure is, however, not as
stable as other dimers. Attempted isolation of 5Bc from the
mixture by using routine TLC produced mainly the ring-opened
elimination product, dimer 5Bo. Isolation by flash column did
provide a product rich in the cyclic product for NMR, but
certainly not free of other components. Ensuring that the crude
coniferaldehyde dehydrodimers were not excessively dried
helped maintain the phenylcoumaran structure 5Bc and it
could be successfully isolated in substantial portions even by
TLC. Acetylation of the dehydrodimer mixture also converted
the phenylcoumaran 5Bc to the (acetate of ) acyclic dimer 5Bo
by the mechanism assumed (again via the quinone methide) in
Fig. 2 (also see Fig. 4J where the cyclic component in Fig. 4E
has disappeared). HMBC spectra show the disappearance of
the easily recognizable correlations of the phenylcoumaran
(Fig. 3) after acetylation (not shown). Even dissolving the crude
mixture in acetone–water for the spectrum in Fig. 4F appears to
have destroyed some of the cyclic component (although this is
The 8–5-coupled dehydrodimer isolated by Connors et al.
was the non-cyclic dimer 5BoGG.12 It has therefore since been
considered that the 8-proton elimination (pathway a, Fig. 2)
was faster than the internal trapping of the quinone methide
(pathway b). We have discovered by running NMR spectra of
the crude reaction products from [9-13C]-labeled coniferalde-
hyde that, in biomimetic peroxidase/H2O2 reactions at neutral
pH, the 8–5-cyclic phenylcoumaran 5Bc is the kinetic product,
not the open 8–5-dimer 5Bo. This cyclic phenylcoumaran prod-
uct 5Bc is analogous to those formed from coniferyl alcohol or
ferulate.53 Rearomatization of the B-moiety is therefore faster
than 8-proton elimination of the A-moiety, so the result is the
intramolecular nucleophilic quinone methide trapping reaction.
Proof of the structural assignment for the phenylcoumaran
8–5-product 5Bc is shown in Fig. 3. Fig. 3A is a partial HMBC
spectrum, showing just the aldehyde carbonyl carbon region, of
crude products (dimers 5 plus possibly oligomers 7) from a per-
oxidase/H2O2 reaction of coniferaldehyde 1G. It shows nicely
resolved aldehyde 13C resonances (also seen from the data
in Table 2) with diagnostic 13C–1H long-range correlations
whereby the various dehydrodimeric units 7 and V were readily
assigned. Unit 7BcGG shows clear correlations from the alde-
hyde carbonyl carbon at 197.6 ppm with two aliphatic protons
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
272

View Article Online
Fig. 2 Pathways to 8–5-coupled dehydrodimers 5B. Pathway a is the previously accepted pathway, with 8-proton elimination being faster than re-
aromatization of the B-ring and the internal phenolate trapping of the quinone methide, producing the acyclic product 5Bo. However, the cyclic
phenylcoumaran 8–5-dimer 5Bc, is the kinetic product observed here at neutral pHs from coniferaldehyde via peroxidase/H2O2 indicating that
rearomatization and quinone methide trapping is faster than 8-proton elimination pathway b. Dehydrodimer 5Bc however undergoes ring-opening
and 8-proton elimination under basic acetylation conditions (pyridine ϩ acetic anhydride), presumably via the quinone methide as shown here, to
produce the acyclic 8–5-dehydrodimer 5Bo. Analogous ring-opening is likely to occur under acidic conditions. Only if the dehydrodimer is etherified
before isomerizing will the phenylcoumaran structure be preserved in the polymer. There is some evidence from the spectra in Fig. 4 that the cyclic
form may exist in isolated (and presumably in situ) lignins.
from a different reaction than Fig. 4E). The point to be made
here is that the course of the coupling and rearomatization
reactions favors formation of the phenylcoumaran structure
(as the kinetic product), whereas subsequent reactivity causes
the structure to open (to give the thermodynamically more
stable product). It is assumed that if the cyclic phenylcoumaran
product becomes etherified as such during further lignification
reactions, the cyclic form would then be stable. There is ten-
tative evidence for the cyclic coniferaldehyde-8–5-structure in
plant isolates (see discussion below under Lignin Spectra).
Unlike with coniferyl alcohol where the overwhelmingly
major products are produced by coupling reactions at the
β-position, coniferaldehyde (like its ferulate analog) will
undergo 5-coupling reactions. Thus the 5–5-dehydrodimer 5D is
observed in the product mixtures (∼13% of the purified prod-
ucts). Presumably cross-coupling reactions between conifer-
aldehyde 1G and guaiacyl units 4G can also produce 5–5-units.
As is normal with such units, they are integrated into the poly-
mer via further coupling reactions. Coupling is still possible at
the 8-position to form 8–O–4-ethers. 4–O-Coupling is also
possible, either from the cinnamaldehyde directly or follow-
ing 8–O–4-ether formation, to produce dibenzodioxocins,40,54
for example. It is not yet known to what extent 5-coupled con-
iferaldehyde units retain their double bond and therefore show
up as structure V in lignin spectra.
mers to allow acquisition of the 1D 13C-NMR spectra shown in
Figs. 4G and K (as well as for 2D experiments, not shown).
Due to severe matrix and solvent-dependent shifts, the model
data (in acetone-d6) and the lignin spectra do not appear to
coincide—see caption to Fig. 4. The direction of the shifts upon
adding water (necessary for the solution of unacetylated
lignins) is illustrated with the crude dehydrodimeric products
(Figs. 4E–F) and the copolymer DHP (Fig. 4G). Dotted
assignment lines in Fig. 4 have been authenticated by further
2D correlation experiments. Once the models and the lignins
are acetylated, the data coincide much more closely (right-hand
plots in Fig. 4). The aldehyde-carbon regions of spectra from
lignin isolates are also shown in Fig. 4.
In synthetic cross-coupling reactions, as seen in the DHP
example, Fig. 4G, that copolymerizes [9-13C]-coniferaldehyde
1G (10%) with coniferyl alcohol 3G (90%), one feature of the
synthetic lignification is notable: most of the coniferaldehyde
appears not to have coupled at its 8-position. The dominant
aldehyde end-group peak V indicates that any cross-coupling
entered into by coniferaldehyde was predominantly at the ring
5- or 4–O-positions. As will be discussed below, this is not
entirely reflected in the isolated lignins, but similar details could
be observed in the CAD-deficient pine lignin. The discrepancy
likely results from the considerably more “bulk” nature of syn-
thetic lignification. With excess coniferyl alcohol in the system,
the likely cross-coupling reaction is between coniferaldehyde 1G
and coniferyl alcohol 3G monomer radicals. As has been seen
countless times, coniferyl alcohol overwhelmingly couples at its
β-position in cross-coupling reactions. Although coniferalde-
hyde may also couple at its 8-position, as seen from the
(unauthenticated) 7CЈ peak, it more commonly couples at the
4–O- or 5-position. As a result, most coniferaldehyde remained
as end-groups as evidenced by the large V peak in the synthetic
lignin, Figs. 4G and K.
Dehydrogenation polymers (DHPs)
The next step in modeling lignins is the introduction of alde-
hydes into polymeric systems, dehydrogenation polymers
(DHPs), by methods which somewhat mimic lignification. Even
with the slow addition of monomers, the polymerization is
largely a bulk polymerization with too many dimerization
events rather than the cross-coupling of monomers with the
growing polymer.40,47,55 Nevertheless, all the correct coupling
and cross-coupling products that are found in lignins are usu-
ally produced, just in different ratios.40,56 To gather the required
NMR data, it is not usually necessary to resort to more time-
consuming methods to enhance cross-coupling reactions such
as diffusion of monomers through dialysis membranes.55,57
Simple in vitro DHP methods were used here to produce poly-
A new product remains unauthenticated at present but is
thought to be the 8–βЈ-cross-product 7CЈ (Fig. 1). If so, the
minor change in its chemical shift indicates that, like the 8–8-
coupled dimer 7C, the aldehyde moiety remains unsaturated.
This means that the intermediate quinone methide produced
following radical coupling was not efficiently trapped internally
by the γ–OH. Thus, unlike the 8–5-coupling product which is
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
273

View Article Online
Fig. 3 A) NMR evidence for the dimer/oligomer assignments of aldehyde 13C NMR peaks to structures 7A–C and V (Fig. 1), and proof that the
8–5-dimer is in the cyclic phenylcoumaran form following peroxidase-mediated radical coupling at neutral pH. The sample is the crude dimer
fraction formed using [9-13C]-labeled coniferaldehyde. The 13C–1H long-range correlation experiment (gradient-edited HMBC, solvent: acetone-d6)
correlates each aldehyde carbonyl carbon with its directly attached aldehyde proton (split by the 1-bond 13C–1H coupling constant; the proton
chemical shift is halfway between the pair of correlation peaks) and other protons 2- or 3-bonds away (in this case, protons 7 and 8 on the sidechain).
The 8–O–4- 7A and 8–8-units 7C are clearly unsaturated as shown by their correlation with single 7-protons. End-units V (which also derive from the
B-moieties of dehydrodimers 5A and 5B, and from 5D) show correlations to both unsaturated sidechain protons (7 and 8). The 5–5-dimer 5D has the
same correlations as the end-units V. The coincident chemical shifts of both carbons and protons make it virtually impossible to differentiate
the 5–5-structure from other end-units. The 8–5-unit 7Bc shows typical correlations of the (cyclic) phenylcoumaran structure. B) Further proof for
the cyclic form of the 8–5-coupled dehydrodimer in the crude mixture of dehydrodimers from the extensive full sidechain HMBC correlations.
The correlations between proton A7 and carbons B4 and B5 in particular prove the cyclic phenylcoumaran structure.
internally trapped by the phenolic-OH, 8-proton elimination is
Lignin spectra
faster here than the internal trapping by the primary alcohol.
The copolymer synthetic lignin (Figs. 4G and K) also con-
tains 8–O–4-coupled aldehyde dehydrodimeric units 7A, as well
as the 8–O–4-cross-product 7AЈ that was not seen in the conifer-
aldehyde-only reactions (Figs. 4F and J). Units 7AЈ are from
cross-coupling of an hydroxycinnamaldehyde radical with a
radical from a preformed lignin oligomer (B-moiety) which is
likely to be 5–5-or 4–O–5-structures in this DHP (as described
in the following section on lignins).
CAD-downregulated tobacco transgenic. Well-dispersed alde-
hyde peaks at 180–200 ppm in the 13C-NMR spectra show con-
siderable differences between lignins from the CAD-deficient
transgenic (Figs. 4I and M) and the control (not shown) as has
been well documented.27,35,40 The control lignin had the normal
cinnamaldehyde V and benzaldehyde U end-group structures,
but peaks labeled 7C, 7A, and 7AЈ (Figs. 4I and M) were new in
the transgenic’s lignin. The basic assignments for the acetylated
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
274

View Article Online
Fig. 4 Aldehyde regions (180–200 ppm) of 13C-NMR spectra of synthetic dehydrodimers, oligomers and lignins, and isolated lignins from a CAD-
deficient transgenic tobacco and a mutant pine. Left: aldehyde sub-regions of 13C-NMR spectra of: A)-D) various coniferaldehyde dimers (relating to
structures in Fig. 1) run in acetone-d6; crude mixtures of dimers resulting from low-extent oligomerization of [9-13C]coniferaldehyde with horse-
radish peroxidase and H2O2 in E) acetone-d6 and F) acetone-d6–D2O, 6 : 1 (note: this was from a different reaction than the one above it); G) a
synthetic lignin (DHP) prepared from [9-13C]coniferaldehyde (10%) and coniferyl alcohol (90%) in 6 : 1 acetone-d6–D2O; H) an isolated lignin from a
CAD-deficient pine mutant; and I) an isolated lignin from uniformly 13C-enriched (∼13%) antisense-CAD-downregulated tobacco. Note that radical
coupling of coniferaldehyde via peroxidase/H2O2 clearly produces the cyclic (phenylcoumaran) 8–5-dimer corresponding to structure 5Bc/7Bc. We do
not currently have a model compound for the 8-βЈ-cross-product 7CЈ seen in the copolymer DHP in G, so this peak remains unauthenticated. If it is
the 8-βЈ-cross-product, it appears to be in the open form as shown in Fig. 1, in which the intermediate quinone methide on the aldehyde moiety is not
internally trapped by the 9-OH (as occurs in 8–8-coupling of coniferyl alcohol to pinoresinol). Substantial solvent and matrix-dependent shifts are
noted for the unacetylated dimers, oligomers, and polymers (which were run in ∼6 : 1 acetone-d6–D2O). The assignment lines linking the various
peaks in the two lignin spectra have been authenticated by further correlation experiments (not shown) on the lignin samples. Right: corresponding
spectra of acetylated oligomeric products and lignin isolates. Solvent shifts are not a problem in this case since all samples dissolved in acetone-d6.
Thus the assignments of 8–O–4- and 8–8-coupled products are verified (and further authenticated by diagnostic correlations in various 2D NMR
experiments). Note that the cyclic phenylcoumaran form of the 8–5-dimer 5Bc converts to the opened product 5Bo upon acetylation (see Fig. 2)
except, presumably, when it becomes etherified during lignification—traces appear to remain in the DHP and the lignins.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
275

View Article Online
(Fig. 4M) and unacetylated (Fig. 4I) lignins were readily made
by comparison with model and DHP data in Fig. 4, and via
HMBC spectra with characteristics analogous to those in
Fig. 3A for the DHP. The major types of units are distinguish-
able, with long-range correlations yielding information about
their nature. Unfortunately, the 8–5-units have overlapping or
ambiguous correlations in all spectra and cannot therefore be
unambiguously identified here.
8–O–4-Coupled units, 7A and 7AЈ. Two significant 8–O–4-
peaks 7A and 7AЈ at 188.1 ppm and 186.8 ppm in the 13C-NMR
were the primary new features identified in the transgenic
tobacco plant (Figs. 4M and 5A). When the B-ring is a normal
guaiacyl unit (arising from dehydrodimerization of conifer-
aldehyde, or from cross-coupling of either coniferaldehyde or
sinapaldehyde with a lignin guaiacyl unit) the carbon is at
∼188.1 ppm, with the correlated 7-proton centered at 7.28 ppm,
as in the red contour labeled 7A in Fig. 5A. When the B-ring is a
syringyl unit, or possibly a 5-substituted guaiacyl unit, as we
hope to establish later, the carbon is at ∼186.8 ppm, with the
correlated 7-proton centered around 6.7 ppm, as in the magenta
and dark red contours labeled 7AЈ in Fig. 5A. As has been
communicated more fully elsewhere,34 the 8–O–4-syringyl com-
ponent 7AЈ, is derived from coupling reactions involving both
coniferaldehyde 1G (magenta G2 and G6 carbon correlations
at 115.2 and 124.0 ppm) and sinapaldehyde 1S (dark red S2/6
carbon correlations at 108.1 ppm). The clear implication is that
both coniferaldehyde 1G and sinapaldehyde 1S will cross-couple
with syringyl units 4S in the growing polymer to form 8–O–4-
structures. On the other hand, the 8–O–4-guaiacyl component
7A derives only from a coupling reaction involving sinapalde-
hyde (see red S2/6-carbon correlations at 108.1 ppm) and not
coniferaldehyde (see the red crosses where correlations would
be present, as seen later in the CAD-deficient pine of Fig. 5B).
Here the implication is that sinapaldehyde 1S but not conifer-
aldehyde 1G will cross-couple with guaiacyl units 4G in the
growing polymer. And as noted previously,34,35 this apparent
selectivity is explained by nothing more than simple chemical
cross-coupling propensities—coniferaldehyde will not readily
cross-couple with guaiacyl units in vitro either. It will however
dehydrodimerize to 5AGG. Presumably this is the source of the
7A peaks in the crude coniferaldehyde dimers (Fig. 4J) and in
the copolymer DHP (Fig. 4K). The absence of coniferaldehyde-
8–O–4-guaiacyl peaks 7A in this tobacco lignin however
suggests that dehydrodimerization of coniferaldehyde is not a
significant reaction occurring during lignification, and that
most of the aldehydes incorporated into 8–O–4-structures are
therefore the result of endwise cross-coupling reactions of a
monomer with the growing polymer. Such reactions of mono-
lignols are the foundations of lignification and indicate that the
aldehyde monomers are incorporating into these lignins analo-
gously to the traditional monolignols and are therefore likely to
be true “lignin” monomers.
Cinnamaldehyde end-groups, V. These end-groups can arise
from two sources. They may arise from preliminary dehydro-
dimerization reactions of coniferaldehyde or sinapaldehyde,
forming dehydrodimers 5A, 5B or 5D which retain the unsatur-
ated cinnamaldehyde sidechain. In the case of 5A and 5B, these
end-group units are now etherified and will not therefore be
able to undergo further coupling reactions at their 8-positions.
Dimer 5D is unique in that it can still undergo 8-coupling reac-
tions, so end-groups V would only result from dimer 5D if the
subsequent polymerization reaction was cross-coupling with a
monolignol to form a dibenzodioxocin (etherified) struc-
ture. Alternatively, end-groups V result as the B-rings from
cross-coupling reactions of hydroxycinnamaldehydes 1 at their
4–O- or 5-positions with monolignols (at their β-positions),
or possibly with a preformed lignin oligomer at its 4–O- or
5-positions. Not all of these possibilities can be distinguished
with the NMR data available. The long-range correlations in
dark blue starting from the aldehyde carbonyl carbon in units V
identify proton-7 at ∼ 7.6 ppm, Fig. 5A. This proton correlates
with carbons in the aromatic ring, allowing a distinction
between syringyl and guaiacyl units. The observations of a
strong S2/6 carbon correlation (at 107.4 ppm) and the guaiacyl
G2 (111.7 ppm) and G6 (124.7 ppm) carbon correlations indi-
cate that both coniferaldehyde and sinapaldehyde have been
incorporated into this transgenic’s lignin in end-group units V.
Benzaldehyde end-groups, U. As noted above from spectra of
the monomers, the 2- and 6-protons correlating to the aldehyde
carbonyl, carbon-7, in benzaldehydes U, are resolved with the
syringyl protons having lower chemical shifts. It is readily
apparent from the HMBC spectrum in Fig. 5A that most of the
benzaldehydes U in the CAD-deficient tobacco lignin arise
from syringaldehyde. The huge HMBC correlation (centered at
191.7/7.21 ppm) between the aldehyde carbon and the S2/6 pro-
ton of syringaldehyde units U was enhanced in the transgenic
over the normal tobacco lignin (not shown). The green con-
tours labeled U show expected correlations to S2/6 protons at
∼108 ppm and no obvious guaiacyl G2 or G6 correlations, con-
firming the syringyl nature of those aldehydes. (Strictly, 5–O–4-
linked vanillin units will have resonances masked by the
syringyl units and appear in the green colored peaks, so these
may not be resolved.) A smaller guaiacyl benzaldehyde unit U
is apparent with its correlation at ∼191.1/7.56 ppm; the proton
to ring-carbon correlations overlap with those of the G2 and
G6 carbons in unit V described above (and colored dark blue).
It appears that sinapaldehyde 1S has been rather substantially
converted to syringaldehyde 2S and incorporated into this
lignin fraction, whereas coniferaldehyde 1G has mainly been
incorporated as such and less effectively converted to, or
incorporated as, its vanillin 2G counterpart. That vanillin units
will incorporate into lignins if available is evidenced by its sig-
nificant incorporation in the pine lignin, Fig. 5B (see below).
8–8-Coupled units, 7C. Sinapaldehyde 1S, like sinapyl alcohol
3S, favors 8–8-coupling, in part because there are fewer options
than for coniferaldehyde 1G (which has an additional 5-position
available for radical coupling). From the HMBC spectrum
(Fig. 5A), only 8–8-sinapaldehyde structures were detected in
the CAD-deficient transgenic tobacco lignin. The cyan-colored
correlation labeled 7C between the 9-carbon and 7-proton was
weak but the correlation between the 2/6-carbon and 7-proton
was as pronounced as others. We could detect no traces of con-
iferaldehyde 8–8-structures (Fig. 4A, cyan crosses) as detected
in synthetic lignins (Fig. 4K).
CAD-deficient pine mutant. Despite having a residual CAD
activity of less than 1% of normal levels, incorporation of
hydroxycinnamaldehydes into the CAD-deficient pine mutant’s
lignin was relatively low compared to the CAD-downregulated
transgenic tobacco lignin. Most of the aldehydes remained as
simple extractable, non-polymeric components.36,58 The major
unanticipated incorporation of dihydroconiferyl alcohol
(DHCA)26 and its guaiacylpropane-1,3-diol (GPD) derivative59
complicates this analysis.
Aldehyde levels in the mutant pine’s soluble lignin fraction
were enhanced over the control.26 The increase was mainly due
to increased hydroxycinnamaldehyde V and hydroxybenzalde-
hyde U end-group units. Smaller peaks labeled 7A and 7AЈ,
analogous to those found in tobacco were also evident,
Figs. 4H, 4L and 5B; here they appear enhanced over those in
the spectra originally published26 as a result of more extensive
EDTA-washing (and possible fractionation).
Cinnamaldehyde end-groups, V. Typical hydroxycinnamalde-
hyde end-groups V were observed to be present due to the
peak at 194.0 ppm in the 13C NMR spectra of acetylated lignin
(Figs. 4L, 5B). In this case, pine being a softwood, the units do
not derive from sinapaldehyde (as seen by the absence of S2/6
carbon correlations in Fig. 5B, and as noted in the tobacco
spectrum of Fig. 5A). The guaiacyl correlations are more com-
plex (and overlap with the benzaldehyde U correlations) but
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
276

View Article Online
Fig. 5 Partial HMBC spectra of lignins from the CAD-deficient A) tobacco transgenic and B) mutant pine, allowing the syringyl/guaiacyl nature of
the various dehydrodimers and cross-products to be assigned. See the text for details. Subscripts 5 indicate a 5-linkage on the unit. G2 and G6
assignments for the UG and V units in the pine lignin, Fig. 5B, overlap too severely to allow accurate color-coding of these peaks.
appear to be assignable mainly to G2 and G6 carbons in both
4–O-etherified and 5–β/4–O–α- (phenylcoumaran) structures as
indicated in Fig. 5B (from the data in Tables 1 and 2).
idea expressed above that the main 8–8-products in the tobacco
lignin derive from sinapaldehyde coupling, and that conifer-
aldehyde is less involved in 8–8-coupling reactions. Analogously
enhanced 8–8-coupling is seen with sinapyl alcohol over
coniferyl alcohol in normal hardwood vs. softwood lignins.60
However, the analogous ferulates undergo extensive 8–8-
homocoupling.53 The 8–8-dehydrodimer of coniferaldehyde
represents only about 1% of the dimer fraction in the model
coupling reactions. The absence of 8–8-coupled products
in the pine lignin can therefore be taken as an indication
that a similarly low 8–8-coupling propensity occurs in vivo
or, as in the tobacco, the aldehydes are incorporated mainly
by cross-coupling rather that homo-dehydrodimerization
reactions.
Benzaldehyde end-groups, U. The benzaldehyde U aldehydic
7-carbon was at 191.7 ppm, and correlates strongly with the
ring 2/6 protons over a broad range centered at about 7.6 ppm
(and therefore overlapping in the proton dimension with
V-units). These can be mainly attributed to vanillin units U by
examination of the HMBC spectrum, Fig. 5B. Thus vanillin
2G contributes substantially to the aldehyde component in
this CAD-deficient pine lignin, whereas it had only a minor
contribution in the tobacco lignin (Fig. 5A).
8–8-Coupled Units, 7C. There were no 8–8-structure peaks
visible in the mutant pine’s lignin spectra. This reinforces the
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
277

View Article Online
8–O–4-Coupled units, 7A and 7AЈ. The important but weak
carbon peaks at 187.9 and 186.8 ppm (Figs. 4L and 5B) appear
analogous to those from 8–O–4-structures in the transgenic
tobacco. Indeed peak 7A at 187.9 ppm has strong but disperse
correlations (red, Fig. 5B) suggesting the presence of conifer-
aldehyde-8–O–4-guaiacyl structures. The correlations suggest
that the unit appears in both 4–O-etherified and 5-linked units
(see the correlation labeled G65). We noted above and previ-
ously,34 however, that coniferaldehyde did not readily cross-
couple with guaiacyl units in lignins. Either that is in fact what
is happening here when there are fewer options (including no
possibility of the favorable coupling reactions with syringyl
units), or these units all arise from dehydrodimers 5A which
become incorporated into the lignin by further coupling reac-
tions. Unfortunately that issue remains unresolved here as the
NMR data do not allow cross-coupling to be distinguished
from dehydrodimerization in these 8–O–4-structures 7A. The
peak labeled 7AЈ at 186.8 ppm is more complex to interpret. In
the tobacco spectrum it was noted that this was derived from
hydroxycinnamaldehydes 8–O–4-coupled with syringyl units
4S. Obviously, there are no syringyl units in this pine, or in the
synthetic DHP which also displays a 7AЈ peak. The weak
correlations (in magenta) suggest a guaiacyl ring (i.e. derived
from coniferaldehyde 1G) as expected, but reveal nothing of the
8-coupled unit. Its elucidation and identification has been
developed beyond the scope of this study and remains a focus
of current research, but it is perhaps obvious to state here that
5-linked guaiacyl “B-rings” (see structure for 7AЈ in Fig. 5) may
sufficiently mimic syringyl units in their NMR behavior. GPD
and DHCA were massively incorporated into this lignin.26
These monomers do not have unsaturated side chains, so are
limited to coupling at the aromatic ring 4–O- and 5- positions,
and are expected to be mostly involved in the terminal units in
lignin. These 5–5- or 4–O–5-dehydrodimers of homo- or cross-
coupled GPD and DHCA are possible candidates for the
syringyl-like 8–O–4-cross-coupled structures in the pine
mutant’s lignin; certainly the elevated 5–5-level is the explan-
ation for the elevated dibenzodioxocin levels.40 Peak 7AЈ is
therefore simply assigned here as arising via coniferaldehyde 1G
coupling 8–O–4 to a 5-substituted guaiacyl end-unit in the
growing polymer, and provides further evidence for endwise
coupling reactions incorporating coniferaldehyde into this
polymer, albeit on a far smaller scale than was evident in the
angiosperm, tobacco.
ents of the transgenic and mutant lignins are the traditional
monolignols. However, in the gymnosperm, it appears from
the build-up of extractable hydroxycinnamaldehyde, hydroxy-
benzaldehyde, and DHCA monomers in the plant stems that
producing the polymer from these components is not straight-
forward. In synthetic polymer systems too, coniferaldehyde is
not as readily incorporated as coniferyl alcohol. Nevertheless,
the hydroxycinnamaldehydes 1, and their derived hydroxy-
benzaldehydes 2, become a significant part of the polymer
fraction.
There has been debate about whether these phenylpropanoid
polymers derived from monomers other than the three trad-
itional monolignols 3 are in fact lignin, and whether they are
functioning as lignin in the plant.42,43 The important first ques-
tion answered from the observations reported here is: are alde-
hydes incorporated, as monomers, into lignins or are they sim-
ply post-lignification artifacts of oxidation? Peroxidase/H2O2 is
capable of producing aldehyde monomers from monolignols, so
aldehydes may be expected in the lignifying zone even in the
absence of CAD-deficiency. The high levels of aldehydes in
lignins from CAD-deficient mutants and transgenics and the
incorporation of monomers by homo- and hetero-coupling
reactions into polymeric fractions suggests that they can indeed
enter into the phenylpropanoid polymer fractions by the mech-
anisms characteristic of lignification. For most researchers,
there is little surprise in this. If phenols are present in the cell
wall during lignification, and if lignification is not carefully
enzymatically controlled,43,61 it is logical that they will be
incorporated into the polymer, depending on their abilities to
form radicals and their cross-coupling propensities under the
conditions of lignification. Many monomers (e.g. ferulates,
acylated monolignols, 5-hydroxyconiferyl alcohol) other than
the traditional monolignols have been shown to be components
of lignins, so there is ample precedence for the incorporation of
non-traditional monomers into lignins.35,40,43 There is also par-
ticularly good evidence from studies on COMT-deficient angio-
sperms that substitution of a traditional monolignol with a
novel phenolic component can be accommodated in lignifi-
cation—5-hydroxyconiferyl alcohol is readily incorporated in
the place of sinapyl alcohol into such lignins.35,62 Exclusion of,
for example, the hydroxycinnamaldehydes from lignin mono-
mer definitions would result in the troubling implication that
CAD-deficient plants contain little lignin since most of the
lignin molecules probably contain incorporated aldehyde units;
this is likely even in normal plants.
Implications
Conclusions
The phenolic polymers isolated from the mutant pine and the
transgenic tobacco contain substantial aldehyde components,
copolymerized by radical coupling reactions that typify lignifi-
cation. The polymers contain cross-coupling products of con-
iferaldehyde or sinapaldehyde with lignin oligomers (e.g. 8–O–
4-structures 7A and 7AЈ; Figs. 4 and 5) as well as possible
homo-coupling products derived from dehydrodimers 5A and
5C. The hydroxycinnamaldehydes therefore appear to be true
components that are polymerized by radical coupling mechan-
isms into phenylpropanoid polymers that may function as
lignins. Obviously this is not the same polymer that would be
produced (and can be isolated) when the plant does not have a
CAD-deficiency. It appears, as originally proposed,26,41 that the
plants utilize more of the hydroxycinnamaldehyde precursors
of the normal monolignols as monomers when the plant’s abil-
ity to complete the biosynthesis of monolignols is impeded.
Further evidence has recently come from the identification of
molecular marker compounds of hydroxycinnamaldehyde-8–
O–4-coupled units in transgenic poplar lignins.32 Hydroxy-
cinnamaldehydes are logically anticipated to build up if the flux
through the final reduction step, catalyzed by CAD, is reduced.
The total levels of the phenylpropanoid polymers are close
to those in the normal plants and still the major compon-
Hydroxycinnamaldehyde monomers appear to be incorporated
reasonably well into synthetic and natural lignins as anticipated
from the currently accepted lignification mechanism. The radi-
cals couple in a variety of anticipated ways. Lignins isolated
from plants with CAD-deficiencies have elevated levels of
aldehydes with bonding patterns discernable from NMR that
provide evidence suggesting the incorporation of monomeric
hydroxycinnamaldehydes (e.g. 8–O–4-cross-coupled structures)
by endwise coupling onto the phenolic end of the growing
polymer. Work is still required to determine whether the plant is
really producing a modified lignin by the incorporation of alde-
hyde monolignol precursors to ensure the viability of the plant,
or produces this polymer as some kind of a wound response.
Either way, if such transgenic plants are to be utilized as forages
for ruminants, or for chemical pulping, these polymers contrib-
ute to the non-polysaccharide portion and it is logical to
classify them broadly as lignins. It has already been shown
that CAD-deficient angiosperms are more easily delignified63
(attributable to the greater phenolic content and greater solubil-
ity of the lignin oligomers in the base) and may be more digest-
ible,64 even though it has been shown that aldehyde components
per se reduce digestibility of the cell walls that incorporate
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
278

View Article Online
them.65 Processing problems or advantages of such modified
plants are likely to be significant issues in the near future.
(cyclic) 5Bc (97 mg) coniferaldehyde dehydrodimers and small
amounts of an 8–O–4-trimer 6 (2 mg) were obtained.
3-(4-Hydroxy-3-methoxy-phenyl)-2-[2-methoxy-4-(3-oxo-
propenyl)phenoxy]propenal (coniferaldehyde 8–O–4-dehydro-
dimer) 5AGG. 1H-NMR, Table 1; 13C-NMR Table 2.
Experimental
General
2-(4-Hydroxy-3-methoxyphenyl)-7-methoxy-5-(3-oxo-
propenyl)-2,3-dihydrobenzofuran-3-carbaldehyde (coniferalde-
hyde cyclic 8–5-dehydrodimer) 5BcGG. 1H-NMR, Table 1;
13C-NMR Table 2.
All chemicals, enzymes and solvents were purchased from
Aldrich and Sigma (Milwaukee, WI, USA) unless otherwise
noted. Flash column chromatography was performed on an
ISCO/Biotage system using FLASH 40 pre-packed silica-gel
cartridges (Biotage) with the peaks detected using a UA-6 UV/
VIS detector (ISCO. Inc). Preparative TLC was carried out on
pre-coated 2.0 mm Sil-G-200/UV254 plates (Macherey-Nagel).
2-[2-Hydroxy-3-methoxy-5-(3-oxopropenyl)phenyl]-3-[4-
hydroxy-3-methoxyphenyl)propenal (coniferaldehyde open 8–5-
dehydrodimer) 5BoGG. 1H-NMR, Table 1; 13C-NMR Table 2.
NMR methods
2,3-Bis(4-hydroxy-3-methoxybenzylidene)succinaldehyde
1
A Bruker DRX-360 fitted with a 5 mm 1H/broadband gradient
probe with inverse (1H-detected) geometry was used for all
NMR experiments. Acetylated lignins or model compounds
were dissolved in 0.5 mL acetone-d6; unacetylated lignins were
dissolved in acetone-d6 (0.42 mL) and deuterium oxide (D2O,
0.07 mL). The central acetone solvent peak was used as the
internal reference (δC 29.80, δH 2.04). 1D NMR (1H, 13C and
DEPT-135) and 2D NMR (HMQC and HMBC) experiments
used standard Bruker pulse programs. HMQC (inv4gstp)
spectra were used to elucidate structures of model compounds.
The one-bond coupling constant (1JCH) was set at 145 Hz for
HMQC experiments; HMBC (inv4lplrnd) spectra were opti-
mized for aldehyde correlations in both lignin samples and
model compounds—170.5 Hz for the one-bond 13C9–1H9 coup-
ling constant (1JCH) and 80 ms for the long-range coupling delay
(0.5/nJCH) corresponding to a coupling constant of 6.25 Hz.
Spectra for the dimeric models and lignins were recorded and
compared using the same plotted regions.
(coniferaldehyde 8–8-dehydrodimer) 5CGG. H-NMR, Table 1;
13C-NMR Table 2.
3-[6,2Ј-Dihydroxy-5,3Ј-dimethoxy-5Ј-(3-oxopropenyl)-
biphenyl-3-yl]propenal (coniferaldehyde 5–5-dehydrodimer)
5DGG. 1H-NMR, Table 1; 13C-NMR Table 2.
3-(4-Hydroxy-3-methoxyphenyl)-2-(2-methoxy-4-{2-[2-
methoxy-4-(3-oxopropenyl)phenoxy]-3-oxopropenyl}phenoxy)-
propenal (coniferaldehyde 8–O–4/8–O–4-dehydrotrimer) 6GGG.
1H-NMR, Table 1; 13C-NMR Table 2.
Sinapaldehyde dehydrodimers (Fig. 2)
Method I. The method of synthesizing coniferaldehyde de-
hydrodimers was also applied to sinapaldehyde. Sinapaldehyde
(1 g, 4.81 mmol), acetone : water (20 mL : 980 mL), horseradish
peroxidase (5 mg) and hydrogen peroxide (1%, 20 mL) were
used. Only the 8–8-dehydrodimer 5CSS was isolated.
Model compounds
Method II. Sinapaldehyde (530 mg, 2.59 mmol) was dissolved
in EtOAc (30 mL). Silver() oxide (720.7 mg, 3.11 mmol) was
added and the reaction mixture stirred overnight at room tem-
perature. The mixture was filtered through a fine sintered glass
filter to remove Ag, and evaporated to give a red solid. Separ-
ation of the crude products was performed on preparative TLC
plates (CHCl3 : EtOAc, 1 : 1, v/v), until relatively pure com-
pounds were obtained. Only the 8–8-dehydrodimer 5CSS was
obtained as in method I.
Dehydrodimers 5 from coniferaldehyde (1G, 4-hydroxy-3-
methoxycinnamaldehyde) and sinapaldehyde (1S, 3,5-di-
methoxy-4-hydroxycinnamaldehyde) were synthesized and
purified by method of Connor et al.,12 which uses horseradish
peroxidase (type II, 150–200 units per mg solid) and H2O2. The
synthesized dehydrodimers were used for full NMR spectral
assignments. All other models reported in Tables 1 and 2 are
from data collected over many years in the DFRC and USFPL
laboratories, as documented by their entries in the NMR
database.66
2,3-Bis(4-hydroxy-3,5-dimethoxybenzylidene)succinaldehyde
(sinapaldehyde 8–8-dehydrodimer) 5CSS.. 1H-NMR, Table 1;
13C-NMR Table 2.
Coniferaldehyde dehydrodimers (5, Fig. 1)
Coniferaldehyde 1G (1.1 g, 6.15 mmol) was dissolved in acetone
: water (100 mL : 1.5 L). Horseradish peroxidase (11 mg; EC
1.11.1.7, 158 purpurogallin units per mg solid, type II) in water
was added to the solution which was then stirred. Hydrogen
peroxide (1%, 10.45 mL, 3.1 mmol) was added dropwise into
the reaction solution over 30 min (pH was checked frequently
and maintained at 5–6 by controlling the addition of peroxide)
while the solution was stirred. The yellow solution changed to
red during the addition. The solution was stirred for an addi-
tional 1 h. The mixture was extracted with EtOAc and washed
with water. The EtOAc layer was NOT dried (as would be typi-
cal) to avoid damaging the cyclic phenylcoumaran 8–5-struc-
tures. Mild evaporation gave a red solid. Pre-separation of the
crude products was performed by flash column chromato-
graphy (CHCl3 : EtOAc, 1 : 1, v/v), and each fraction was
further purified on preparative TLC plates using combination
of various solvent systems such as CHCl3 : MeOH (99 : 1, v/v)
and EtOAc : hexane (2 : 1, v/v) until relatively pure compounds
were obtained. Yields of isolated products were variable and
low. From 1.1 g of coniferaldehyde, 8–8 5C (2 mg), 8–O–4 5A
(25 mg), 5–5 5D (25 mg), 8–5 (open) 5Bo (39 mg), and 8–5
Synthesis of [9-13C]coniferaldehyde
The labeled coniferaldehyde was prepared from vanillin via
ethyl ferulate and coniferyl alcohol as follows. Acetylated ethyl
[9-13C]ferulate was prepared as described previously,67 except
that an acetate protecting group was used (i.e. acetylated vanil-
lin was the starting material for the Wittig–Horner reaction
with triethyl [1-13C]phosphonoacetate (99 atom% 13C). The
acetylated ethyl [9-13C]-ferulate was obtained as a pale yellow
solid in 93% yield. NMR δH 1.27 (3H, t, J = 7.1 Hz, CH3CH2),
2.24 (3H, s, Ac Me), 3.89 (3H, s, OMe), 4.20 (2H, qd, J = 7.1,
3.0 Hz, CH3CH2), 6.55 (1H, dd, J = 16.0, 2.4 Hz, 8-H), 7.09
(1H, d, J = 8.1 Hz, 5-H), 7.23 (1H, dd, J = 8.1, 1.8 Hz, 6-H),
7.41 (1H, d, J = 1.8 Hz, 2-H), 7.64 (1H, dd, J = 16.0, 6.8 Hz,
7-H). Note: protons (CH3CH2, 7-H, and 8-H) are long-range
coupled to the labeled 9-13C with coupling constants of 3.0, 6.8,
and 2.4 Hz). [9-13C]Coniferyl alcohol was prepared using
diisobutylaluminium hydride (DIBAL-H, 10 eq.), as described
previously,68 to give [9-13C]3G as a yellow–white solid in essen-
tially quantitative yield; concomitant phenolic deacetylation
occurs.69 NMR δH 3.85 (3H, s, OMe), 4.17 (2H, dtd, J = 140.4,
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
279

View Article Online
5.5, 1.7 Hz, 9-H), 6.20 (1H, dtd, J = 15.9, 5.5, 4.1 Hz, 8-H), 6.48
(1H, ddt, J = 15.9, 7.1, 1.7 Hz, 7-H), 6.76 (1H, d, J = 8.1 Hz,
5-H), 6.84 (1H, dd, J = 8.1, 1.9 Hz, 6-H), 7.04 (1H, d, J = 1.9
Hz, 2-H).
32% for the CAD mutant pine. Lignins were isolated as previ-
ously described;26 the final yield of lignin was 17% after removal
of water-soluble components and metal ions by washing with
aqueous EDTA (6 mM, pH 8). Acetylation of samples of each
of the lignins was achieved using acetic anhydride and pyridine.
The acetylated lignins were extracted into freshly distilled ethyl
acetate and washed extensively with aqueous EDTA (6 mM,
pH 8) to remove trace metal contaminants prior to NMR study
for better resolution and signal-to-noise levels. The aldehyde
region of the spectra is more highly resolved than in previously
published spectra of this mutant.26,35,40
The aldehyde was prepared by oxidation of the alcohol using
dichloro-5,6-dicyanobenzoquinone (DDQ).70 [9-13C]Coniferyl
alcohol 3G (774 mg, 4.28 mmol) was dissolved in THF (50 mL).
DDQ (1.167 g, 5.14 mmol) was added and stirred overnight.
The resulting solution was dried on an evaporator. The product
was separated from DDQ and the trace of starting material by
flash column chromatography (CHCl3–EtOAc, 1 : 1). [9-13C]-
Coniferaldehyde [9-13C]1G (208 mg, 1.16 mmol, 27%) was
obtained as a brown solid after drying. NMR δH 3.90 (3H,
s, OMe), 6.64 (1H, ddd, J = 15.9, 7.8,1.3 Hz, 8-H), 6.91 (1H, d,
J = 8.3 Hz, 5-H), 7.17 (1H, dd, J = 8.3, 2.0 Hz, 6-H), 7.33 (1H,
d, J = 2.0 Hz, 2-H), 7.52 (1H, dd, J = 16.0, 9.2 Hz, 7-H), 9.61
(1H, dd, J = 170.6, 7.8 Hz, 9-H).
Acknowledgements
This research was supported in part by grant no. DE-AI02–
00ER15067 from the Division of Energy Biosciences, US
Department of Energy, and grant no. 99–02351 from the
US Department of Agriculture (USDA)-National Research
Initiative Competitive Grants Program.
Synthesis of [9-13C]sinapaldehyde
[9-13C]Sinapaldehyde was synthesized by the same methods as
for [9-13C]coniferaldehyde. Acetylated ethyl [9-13C]sinapate was
obtained as yellow crystals in essentially quantitative yield from
acetylated syringaldehyde. NMR δH 1.28 (3H, t, J = 7.1 Hz,
CH3CH2), 2.25 (3H, s, Ac Me), 3.86 (6H, s, OMe), 4.20 (2H, qd,
J = 7.1, 3.2 Hz, CH3CH2), 6.55 (1H, dd, J = 16.0, 2.4 Hz,
8-H), 7.05 (2H, s, 2-, 6-H), 7.61 (1H, dd, J = 16.0, 6.7 Hz, 7-H).
[9-13C]Sinapyl alcohol [9-13C]3S was obtained as a yellow–white
solid in essentially quantitative yield following DIBAL-H
reduction. NMR δH 3.81 (6H, s, OMe), 4.24 (2H, br d, J = 141.5
Hz, 9-H), 6.26 (1H, dtd, J = 15.8, 5.6, 4.1 Hz, 8-H), 6.48 (1H,
ddt, J = 15.8, 6.8, 1.6 Hz, 7-H), 6.71 (2H, s, 2-,6-H), 7.36 (1H, s,
4-OH). DDQ oxidation produced [9-13C]sinapaldehyde 1S as a
brown solid in 56% yield. NMR δH 3.88 (6H, s, OMe), 6.67 (1H,
ddd, J = 15.8, 7.8,1.3 Hz, 8-H), 7.06 (2H, s, 2-,6-H), 7.53 (1H,
dd, J = 15.8, 9.4 Hz, 7-H), 7.90 (1H, s, OH), 9.63 (1H, dd,
J = 170.35, 7.7 Hz, 9-H).
References
1 D. Stewart, N. Yahiaoui, G. J. McDougall, K. Myton, C. Marque,
A. M. Boudet and J. Haigh, Planta, 1997, 201, 311.
2 A. M. Boudet and J. Grima-Pettenati, Mol. Breed., 1996, 2, 25.
3 C. Halpin, M. E. Knight, G. A. Foxon, M. M. Campbell,
A.-M. Boudet, J. J. Boon, B. Chabbert, M.-T. Tollier and W. Schuch,
Plant J., 1994, 6, 339.
4 T. Hibino, K. Takabe, T. Kawazu, D. Shibata and T. Higuchi,
Biosci., Biotechnol., Biochem., 1995, 59, 929.
5 T. Higuchi, T. Ito, T. Umezawa, T. Hibino and D. Shibata,
J. Biotechnol., 1994, 37, 151.
6 H.-D. Lüdemann and H. Nimz, Makromol. Chem., 1974, 175, 2409.
7 H. H. Nimz and H.-D. Lüdemann, Holzforschung, 1976, 30, 33.
8 H. H. Nimz, D. Robert, O. Faix and M. Nemr, Holzforschung, 1981,
35, 16.
9 H. H. Nimz, U. Tschirner, M. Stähle, R. Lehmann and M. Schlosser,
J. Wood Chem. Technol., 1984, 4, 265.
10 K. Lundquist and T. Olsson, Acta Chem. Scand., Ser. B, 1977, 31,
788.
11 E. Adler, K. J. Björkquist and S. Häggroth, Acta Chem. Scand.,
1948, 2, 93.
Preparation of dehydration polymers (DHPs)
12 W. J. Connors, C.-L. Chen and J. C. Pew, J. Org. Chem., 1970, 35,
1920.
13 T. Ito, R. Hayase, S. Kawai, H. Ohashi and T. Higuchi, J. Wood Sci.,
2002, 48, 216.
14 T. Luederitz and H. Grisebach, Eur. J. Biochem., 1981, 119, 115.
15 C. Pillonel, P. Hunziker and A. Binder, J. Exp. Bot., 1992, 43, 299.
16 C. Halpin, M. E. Knight, J. Grima Pettenati, D. Goffner, A. Boudet
and W. Schuch, Plant Physiol., 1992, 98, 12.
17 D. M. O’Malley, S. Porter and R. R. Sederoff, Plant Physiol., 1992,
98, 1364.
18 D. Goffner, I. Joffroy, J. Grima-Pettenati, C. Halpin, M. E. Knight,
W. Schuch and A. M. Boudet, Planta, 1992, 188, 48.
19 J. Grima-Pettenati, C. Campargue, A. Boudet and A. M. Boudet,
Phytochem., 1994, 37, 941.
20 A. M. Boudet, C. Lapierre and J. Grima-Pettenati, New Phytol.,
1995, 129, 203.
21 S. W. Hawkins and A. M. Boudet, Plant Physiol., 1994, 104, 75.
22 H. Kutsuki, M. Shimada and T. Higuchi, Phytochemistry, 1982, 21,
19.
Three different solutions were prepared for the DHP synthesis.
(I ) Coniferaldehyde (34.1 mg) and coniferyl alcohol (306.0 mg)
were dissolved in acetone (5 mL) and the solution was added
to phosphate buffer (140 mL, 10 mM, pH = 6.5). (II) 30%
H2O2 (225 µL) was diluted with phosphate buffer (140 mL).
(III) Horse radish peroxidase (4 mg) was dissolved in phosphate
buffer (70 mL). Solutions (I ) and (II) were added to the solu-
tion (III) using a peristaltic pump at a flow rate of 6 mL hϪ1
.
The resultant solution was stirred for a further 3 days. Then the
DHP mixture was transferred into a plastic centrifuge bottle
and centrifuged for 20 min at 1000g. The precipitates were
carefully collected and washed with water using a 6 µm nylon
membrane filter and freeze-dried. The resultant DHP (94.1 mg)
was obtained as a red–brown powder.
Plant materials and lignin isolation
23 M. Baucher, B. Monties, M. Van Montagu and W. Boerjan,
Crit. Rev. Plant Sci., 1998, 17, 125.
24 L. G. Li, X. F. Cheng, J. Leshkevich, T. Umezawa, S. A. Harding and
V. L. Chiang, Plant Cell, 2001, 13, 1567.
25 A. M. Anterola, J.-H. Jeon, L. B. Davin and N. G. Lewis,
J. Biol. Chem., 2002, 277, 18272.
26 J. Ralph, J. J. MacKay, R. D. Hatfield, D. M. O’Malley,
R. W. Whetten and R. R. Sederoff, Science, 1997, 277, 235.
27 J. Ralph, R. D. Hatfield, J. Piquemal, N. Yahiaoui, M. Pean,
C. Lapierre and A.-M. Boudet, Proc. Nat. Acad. Sci. U.S.A., 1998,
95, 12803.
Antisense CAD transgenic tobacco. Production of 13C-
enriched antisense CAD tobacco plants and the lignin isolation
have been described previously.27 The tobacco transgenic had a
cell wall that was 15% Klason lignin, and the final yield of
soluble lignin was 17%. Acetylation was achieved using acetic
anhydride and pyridine. The acetylated lignins were extracted
into freshly distilled ethyl acetate and washed with aqueous
EDTA to remove trace metal contaminants prior to NMR
study.40
28 N. Yahiaoui, C. Marque, K. E. Myton, J. Negrel and A.-M. Boudet,
Planta, 1998, 204, 8.
29 G. J. Provan, L. Scobbie and A. Chesson, J. Sci. Food Agric., 1997,
73, 133.
30 M. A. B. Vailhe, J. M. Besle, M. P. Maillot, A. Cornu, C. Halpin and
M. Knight, J. Sci. Food Agric., 1998, 76, 505.
CAD-null mutant loblolly pine (Pinus teada L.). Identification
and characterization of heterozygous parent and the homo-
zygous daughter trees have been previously described.36,39 Cell
wall Klason lignin levels were 31% w/w for the normal pine and
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
280

View Article Online
31 C. Halpin, K. Holt, J. Chojecki, D. Oliver, B. Chabbert, B. Monties,
K. Edwards, A. Barakate and G. A. Foxon, Plant J., 1998, 14,
545.
32 H. Kim, J. Ralph, G. Pilate, J. C. Leplé, B. Pollet and C. Lapierre,
J. Biol. Chem., 2002, 277, 47412.
33 J. Piquemal, C. Lapierre, K. Myton, A. O’Connell, W. Schuch,
J. Grima-Pettenati and A.-M. Boudet, Plant J., 1998, 13, 71.
34 H. Kim, J. Ralph, N. Yahiaoui, M. Pean and A.-M. Boudet,
Org. Lett., 2000, 2, 2197.
35 J. Ralph, C. Lapierre, J. Marita, H. Kim, F. Lu, R. D. Hatfield,
S. A. Ralph, C. Chapple, R. Franke, M. R. Hemm, J. Van
Doorsselaere, R. R. Sederoff, D. M. O’Malley, J. T. Scott,
J. J. MacKay, N. Yahiaoui, A.-M. Boudet, M. Pean, G. Pilate,
L. Jouanin and W. Boerjan, Phytochem., 2001, 57, 993.
36 J. J. MacKay, D. M. O’Malley, T. Presnell, F. L. Booker, M. M.
Campbell, R. W. Whetten and R. R. Sederoff, Proc. Natl. Acad. Sci.
U. S. A., 1997, 94, 8255.
47 K. V. Sarkanen and C. H. Ludwig, Lignins, Occurrence, Formation,
Structure and Reactions, Wiley-Interscience, New York, 1971.
48 J. M. Harkin, in Oxidative Coupling of Phenols, eds. W. I. Taylor and
A. R. Battersby, Marcel Dekker, New York, 1967, p. 243.
49 K. Freudenberg and A. C. Neish, Constitution and Biosynthesis of
Lignin, Springer-Verlag, Berlin, Heidelberg, New York, 1968.
50 G. J. Leary, Wood Sci. Technol., 1980, 14, 21.
51 F. Nakatsubo, K. Sato and T. Higuchi, Mokuzai Gakkaishi, 1976,
22, 29.
52 T. Higuchi, Adv. Enzymol., 1971, 34, 207.
53 J. Ralph, S. Quideau, J. H. Grabber and R. D. Hatfield, J. Chem.
Soc., Perkin Trans. 1, 1994, 9.
54 P. Karhunen, P. Rummakko, J. Sipilä, G. Brunow and I. Kilpeläinen,
Tetrahedron Lett., 1995, 36, 4501.
55 K. Syrjanen and G. Brunow, J. Chem. Soc. Perkin Trans. 1, 2000,
183.
56 J. Ralph, Y. Zhang and R. M. Ede, J. Chem. Soc., Perkin Trans. 1,
1998, 2609.
57 M. Tanahashi and T. Higuchi, Wood Res., 1981, 67, 29.
58 J. J. MacKay, A mutation in lignin biosynthesis in loblolly pine:
genetic, molecular and biochemical analyses, Ph. D. Thesis,
North Carolina State University, 1997.
37 D. R. Gang, M. Fujita, L. D. Davin and N. G. Lewis, in Lignin and
Lignan Biosynthesis, eds. N. G. Lewis and S. Sarkanen, Am. Chem.
Soc., Washington, DC, 1998, p. 389.
38 J. J. MacKay, D. R. Dimmel and J. J. Boon, J. Wood Chem. Technol.,
2001, 21, 19.
39 C. Lapierre, B. Pollet, J. J. MacKay and R. R. Sederoff, J. Agric.
Food Chem., 2000, 48, 2326.
59 J. Ralph, H. Kim, J. Peng and F. Lu, Org. Lett., 1999, 1, 323.
60 E. Adler, Wood Sci. Technol., 1977, 11, 169.
61 J. Ralph, J. Peng, F. Lu and R. D. Hatfield, J. Agric. Food Chem.,
1999, 47, 2991.
62 J. M. Marita, J. Ralph, R. D. Hatfield, D. Guo, F. Chen and
R. A. Dixon, Phytochem., 2003, 62, 53.
63 C. Lapierre, B. Pollet, M. Petit-Conil, G. Pilate, C. Leple, W. Boerjan
and L. Jouanin, in Lignin: Historical, Biological, and Materials
Perspectives, American Chemical Society, Washington, 2000, Vol.
742, p. 145.
64 A. J. Gordon and T. S. Neudoerffer, J. Sci. Food Agric., 1973, 24,
565.
40 J. Ralph, J. M. Marita, S. A. Ralph, R. D. Hatfield, F. Lu, R. M.
Ede, J. Peng, S. Quideau, R. F. Helm, J. H. Grabber, H. Kim,
G. Jimenez-Monteon, Y. Zhang, H.-J. G. Jung, L. L. Landucci,
J. J. MacKay, R. R. Sederoff, C. Chapple, A. M. Boudet, in Advances
in Lignocellulosics Characterization, eds. D. S. Argyropoulos and
T. Rials, TAPPI Press, Atlanta, GA, 1999, p. 55.
41 J. Ralph, Recent advances in characterizing Јnon-traditionalЈ lignins
and lignin-polysaccharide cross-linking, 9th International
Symposium on Wood and Pulping Chemistry, CPPA (Canadian
Pulp and Paper Association), Montreal, Quebec, 1997, vol. 1,
p. PL2.
65 J. H. Grabber, J. Ralph and R. D. Hatfield, J. Sci. Food Agric., 1998,
78, 81.
66 S. A. Ralph, L. L. Landucci and J. Ralph, available on the Internet at
http://www.dfrc.ars.usda.gov/software.html.
42 N. G. Lewis, Curr. Opin. Plant Biol., 1999, 2, 153.
43 R. R. Sederoff, J. J. MacKay, J. Ralph and R. D. Hatfield,
Curr. Opin. Plant Biol., 1999, 2, 145.
44 R. M. Ede and G. Brunow, J. Org. Chem., 1992, 57, 1477.
45 Y. Z. Lai and K. V. Sarkanen, in Lignins, Occurrence, Formation,
Structure and Reactions, eds. K. V. Sarkanen and C. H. Ludwig,
Wiley-Interscience, New York, 1971, p. 165.
46 K. Lundquist, A. Paterson and L. Ramsey, Acta Chem. Scand.,
Ser. B, 1983, 37, 734.
67 J. Ralph, R. F. Helm, S. Quideau and R. D. Hatfield, J. Chem. Soc.,
Perkin Trans. 1, 1992, 9.
68 S. Quideau and J. Ralph, J. Agric. Food Chem., 1992, 40, 1108.
69 N. Terashima, S. A. Ralph and L. L. Landucci, Holzforschung, 1995,
50, 151.
70 H. D. Becker, A. Bjoerk and E. Adler, J. Org. Chem., 1980, 45, 1596.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 6 8 – 2 8 1
281