On the role of the monolignol γ-carbon functionality in lignin biopolymerization

Article abstract of DOI:10.1016/j.phytochem.2008.10.014

In order to investigate the importance of the monomeric γ-carbon chemistry in lignin biopolymerization and structure, synthetic lignins (dehydrogenation polymers; DHP) were made from monomers with different degrees of oxidation at the γ-carbon, i.e., carb

Full text of DOI:10.1016/j.phytochem.2008.10.014

Phytochemistry 70 (2009) 147–155
Contents lists available at ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
On the role of the monolignol
c
-carbon functionality in lignin biopolymerization
Anders Holmgren a, Magnus Norgren b,c, Liming Zhang a, Gunnar Henriksson a,
*
a Division of Wood Chemistry, Department of Fibre and Polymer Technology, Royal Institute of Technology, KTH, Teknikringen 56, Stockholm, Sweden
b Division of Fibre Technology, Department of Fibre and Polymer Technology, Royal Institute of Technology, KTH, Stockholm, Sweden
c Centre for Fibre Science and Communication Network, Mid Sweden University, Sundsvall, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
In order to investigate the importance of the monomeric c-carbon chemistry in lignin biopolymerization
and structure, synthetic lignins (dehydrogenation polymers; DHP) were made from monomers with dif-
Received 2 November 2007
Received in revised form 6 October 2008
Available online 4 December 2008
ferent degrees of oxidation at the -carbon, i.e., carboxylic acid, aldehyde and alcohol. All monomers
c
formed a polymeric material through enzymatic oxidation. The polymers displayed similar sizes by size
exclusion chromatography analyses, but also exhibited some physical and chemical differences. The DHP
made of coniferaldehyde had poorer solubility properties than the other DHPs, and through contact angle
of water measurement on spin-coated surfaces of the polymeric materials, the DHPs made of coniferal-
dehyde and carboxylic ferulic acid exhibited higher hydrophobicity than the coniferyl alcohol DHP. A
structural characterization with 13C NMR revealed major differences between the coniferyl alcohol-based
polymer and the coniferaldehyde/ferulic acid polymers, such as the predominance of aliphatic double
bonds and the lack of certain benzylic structures in the latter cases. The biological role of the reduction
Keywords:
Lignin
Monolignol biosynthesis
Lignin-polysaccharide networks
Plant cell wall
Biopolymer
at the
c
-carbon during monolignol biosynthesis with regard to lignin polymerization is discussed.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
Generally, the following chemical modifications are carried out
on the amino acid in the pathway to becoming a monolignol:
Lignin is a main constituent of the cell walls of vascular plants,
especially in woody tissues where it is present in high amounts. It
1. The amino group is removed; simultaneously, a double bond is
introduced between the
- and b-carbons.
acts as
a
binding agent, provides compressive strength and
a
bending stiffness (Sederoff and Chang, 1991), functions as a
physico-chemical barrier against pathogens (Baucher et al.,
1998), and conveys hydrophobicity to the cell wall, inhibiting
swelling and boosting water transport through the cell lumen
(Sederoff and Chang, 1991; Brett and Waldron, 1996). From a
chemical point of view, lignin is a branched polymer built mainly
with p-hydroxycinnamyl alcohols with different degrees of meth-
oxylation (also called monolignols: p-coumaryl alcohol, coniferyl
alcohol and sinapyl alcohol; Fig. 1a). Moreover, lignin is believed
to be shaped like a network that connects different cell wall poly-
mers (Lundquist et al., 1983; Xie et al., 2000; Lawoko et al., 2006).
The polymerization of monolignols into lignin involves an oxi-
dation step that is catalyzed by peroxidases, creating resonance-
stabilized radicals, and an uncatalyzed radical–radical coupling
step. The polymer grows as its free phenolic groups are oxidized
and couple new radicals (Fig. 1b). In plants, the monolignols are
biosynthesized from the amino acid phenylalanine through the
phenylpropranoid pathway (Boerjan et al., 2003) (Fig. 2); as one of
the building blocks of proteins it is ubiquitous in all life forms.
2. Carbons 3 and 5 may be methoxylated via a hydroxylation and
carbon 4 is hydroxylated only.
3. The carboxylic acid at the
c-carbon is reduced first to an alde-
hyde and then to an alcohol.
The reasons for the first two modifications are rather obvious;
by removing the amino group and introducing the double bond,
nitrogen, which is often a limiting nutrient for plants, is recovered
for other uses. In addition, the double bond gives possibilities for
radical stabilization and radical couplings, while the methoxyla-
tions are a way for the plant to regulate the coupling pattern by
sterically blocking formation of bonds to the 50 and 30 carbons of
the aromatic ring. However, the reason for reducing the monomers
at the c-carbon is less obvious, since this carbon is not directly in-
volved in the polymerization, except in stabilizing ether bonds in
e.g., the resinol structure. Nevertheless, this reduction probably
conveys an important function. For each biosynthesized monolig-
nol, two nicotinamide adenine dinucleotide phosphates (NADPHs)
and one adenosine trisphosphate (ATP) are consumed to reduce
the carboxylic acid to the aldehyde and the aldehyde to the alcohol
(Fig. 2) (Sjöström, 1981). Taking into account the fact that 15–36%
or more of the cell wall in woody plants consists of lignin, a
* Corresponding author. Tel.: +46 87906163; fax: +46 87906166.
E-mail address: ghenrik@polymer.kth.se (G. Henriksson).
0031-9422/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2008.10.014

148
A. Holmgren et al. / Phytochemistry 70 (2009) 147–155
Fig. 1. (a) The most common lignin monomers, the monolignols. (b) Simplified pathway for the formation of lignin through radical coupling of a monolignol (here coniferyl
alcohol) to a phenolic end-group forming the most common lignin inter-unit bond, b-O-40. Monolignols or phenolic end-groups are oxidized, probably by peroxidases, to
resonance-stabilized radicals. Next, the radicals couple to each other in an uncatalyzed step. The positions that can form covalent bonds are marked with bold arrows.
Observe that the c-carbon is not involved in the radical–radical coupling and that the 3 and 5 positions will be available depending on the type of monolignol (see a). Radical
coupling can occur not only between monomer and oligomer but between oligomers.
Fig. 2. The monolignol biosynthesis pathway. The starting substance is the amino acid phenylalanine. This is subjected to three major modifications in order to form the
monolignols: first, the removal of the amino group with the introduction of a double bond (1); second, the introduction of methoxy groups via hydroxylations (2, 3); and third,
reduction at the c-carbon carboxylic acid to an aldehyde (4) and eventually to an alcohol (5). Observe that reactions 4 and 5 consume one NADPH each. One ATP is also
consumed in a ‘‘preparation stage” for reaction 4, where coenzyme A forms a temporary thioester to the c-carbon, thereby activating it for reduction (6). In dashed boxes are
the monomers used in this study.
considerable part of the reducing power and free energy produced
during photosynthesis is consumed in these reductions. Given that
some of the aldehyde and carboxylic acid products of the phenyl-
propanoid pathway can polymerize through the same basic steps
as the monolignols (dehydrogenative polymerization), are there
any advantages to an alcohol-based lignin?

A. Holmgren et al. / Phytochemistry 70 (2009) 147–155
149
Table 1
In this study, we generate dehydrogenation polymers (DHPs)
from coniferyl alcohol, coniferaldehyde and ferulic acid (Fig. 2),
and characterize the products in order to compare them and inves-
Molecular weight averages relative to polystyrene standards from SEC analysis of
acetylated DHP products.
Polymer
Mw
Mn
P
MP
tigate the biological reasons for the reduction at the c-carbon.
Ferulic acid DHP
Coniferaldehyde DHP
Coniferyl alcohol DHP
4489
2709
4464
3395
1898
3206
1.3
1.4
1.4
5017
2222
2725
2. Results
Mw, molecular weight average (g/mol); Mn, Number average molecular weight
(g/mol); P = Mw/Mn, polydispersity; MP, peak molecular weight (g/mol).
2.1. DHP synthesis
Dehydrogenated polymers (‘‘synthetic lignin”; DHP) from feru-
lic acid, coniferaldehyde and coniferyl alcohol (Fig. 2) were synthe-
sized by peroxidase oxidation. After extraction of low molecular
weight compounds, the yields were 58%, 47% and 35%, respectively.
The relatively low coniferyl alcohol DHP yield could be due to ten-
dency of coniferyl alcohol to precipitate at the beginning of the
reaction in the conditions used here. Interestingly, the ferulic
acid-based DHPs did not precipitate during the reaction, in con-
trast to the other two DHPs.
made of coniferyl alcohol to be more soluble in apolar solvents (Ta-
ble 2). Thus the DHP products displayed different solubilities; the
coniferaldehyde DHP in particular was resistant to complete disso-
lution in any of the solvents.
2.4. NMR analyses
All three DHPs were analyzed with 13C NMR, and the spectra are
shown in Fig. 4. A more complete and unambiguous structural
characterization, although relevant, is outside the scope of this
study. Instead, signal assignments are suggested for several sub-
structures (Table 3 and Fig. 5) based on both reported values for
synthesized model components and NMR studies on isolated lig-
nins (Ralph et al., 1994, 2004; Landucci et al., 1998; Kim et al.,
2003). Several observations were made from the comparison of
the three spectra:
2.2. Size exclusion chromatography (SEC)
Size exclusion chromatography analysis (Fig. 3) revealed that
the three DHPs contained dispersed polymeric material. The
coniferyl alcohol and coniferaldehyde DHPs displayed a similar
polydispersity (Table 1, Fig. 3), while the ferulic acid DHP showed
a somewhat lower polydispersity. The molecular weight average of
the coniferyl alcohol and the ferulic acid DHPs were similar, but
the molecular weight of the peak (MP) of the latter was almost
twice the MP of the other DHPs, indicating that the ferulic acid
polymerization generally produced large populations of larger
polymers. This could be partly explained by the polymers staying
in solution throughout the reaction, making the larger species
available for further polymerization, as opposed to the other DHPs,
which commonly precipitated. Moreover, the SEC analysis of the
coniferaldehyde DHP revealed that there was a significant amount
of low molecular weight species in the coniferaldehyde DHP,
meaning they might not have been washed out as effectively as
in the case of the two other DHPs.
ꢀ The different DHPs could generally form the same type of bond-
ing structures: b-O-40, b-50, b–b0 and 5–50.
ꢀ Although equivalent bonding structures were found in all DHPs,
chemical differences were obvious when comparing the dc 60–
90 ppm region of the three spectra. The coniferyl alcohol DHP
displayed a number of signals in this region corresponding to
saturated a, b and c carbons, while the ferulic acid and conifer-
aldehyde DHPs almost completely lacked any such signals. Sat-
urated side-chains are formed when coniferyl alcohol radical
coupling occurs through at least one b-carbon, i.e., b-O-40, b-50
and b–b0. Instead, although such coupling occurred in the DHPs
from ferulic acid and coniferaldehyde, respectively, the struc-
tures formed were predominantly unsaturated (Fig. 5) and their
signals appear in the 100–140 ppm region.
2.3. Solubility panel
The DHPs were characterized with a solvent panel (Table 2).
Generally, the coniferaldehyde DHP had the lowest solubility,
while the ferulic acid and coniferyl alcohol DHPs showed rather
similar patterns, although there was a tendency for the DHP made
of ferulic acid to be more soluble in polar solvents, and the DHP
ꢀ There is a general lack of cyclic structures (e.g., b-50 and b–b0) in
both the ferulic acid and coniferaldehyde DHPs compared to the
coniferyl alcohol DHP. The ferulic acid DHP did display some
signals, although weak, for possibly cyclic unsaturated b-50 and
b–b0.
ꢀ The ferulic acid and coniferaldehyde DHPs lacked benzyl alcohol
groups, i.e., OH-groups on the
a-carbon, and non-cyclic benzyl
aryl ether structures, both of which are typical of coniferyl alco-
hol DHPs (Freudenberg, 1959; Adler, 1977). In lignins, benzyl
alcohols are very common as they are formed during polymeri-
zation from the addition of water to the quinone methide inter-
mediate of the most abundant inter-monomeric structure: the
b-O-40 bond. Non-cyclic benzyl aryl ethers are also present in lig-
nins but generally in lower quantities compared to DHPs (Ede
and Kilpeläinen, 1995).
2.5. Contact angle measurements on DHP model surfaces
Fig. 6 shows AFM images of very smooth spin-coated DHP mod-
el surfaces. As a relative measure of the hydrophobicity of the dif-
ferent polymeric materials, the contact angle of water was
determined on the DHP model surfaces. The results (Table 4) indi-
cated that the coniferyl alcohol DHP was less hydrophobic than the
Fig. 3. SEC chromatogram of acetylated DHP products. See Table 1 for molecular
weight approximations.

150
A. Holmgren et al. / Phytochemistry 70 (2009) 147–155
Table 2
Ocular inspection of the solubility of the different unmodified DHP materials in organic and aqueous solvents with increasing dielectric constants (excluding dioxane:water and
acetone:water, which are mixes of solvents at 1:1 ratio). +/ꢁ means partly soluble. EtoAc is ethyl acetate.
Polymer
Dioxane
EtoAc
THF
CH2CL2
Pyridine
Acetone
DMF
DMSO
Water
Dioxane:water
Acetone:water
Ferulic acid DHP
Coniferaldehyde DHP
Coniferyl alcohol DHP
No
No
+/ꢁ
No
No
No
Yes
No
No
No
No
No
Yes
+/ꢁ
Yes
+/ꢁ
No
No
Yes
+/ꢁ
Yes
Yes
Yes
Yes
No
No
No
No
No
Yes
Yes
No
Yes
Fig. 4. 13C NMR spectra of (a) ferulic acid DHP; (b) coniferaldehyde DHP and (c) coniferyl alcohol DHP in DMSO-d6. Suggestions for the signal assignments are found in Table 3
and those for the types of inter-unit structures in Fig. 5.
coniferaldehyde and the ferulic acid DHPs. All three DHPs had a
much higher hydrophobicity than was previously determined for
kraft lignin (Norgren et al., 2006), but this was expected since
the latter is a lignin derivative where both free phenolic groups
and carboxyl groups are present. There are reasons to believe that
the hydrophobicity of the coniferaldehyde DHP is somewhat
underestimated, since this material was incompletely dissolved
in pyridine, the solvent used for surface preparation. Therefore,
the fraction of the DHP that was dissolved by the pyridine and used
for the model surface preparation most likely represented a some-
what lower molecular weight DHP that was less hydrophobic.
The DHP made of ferulic acid seemed even to form somewhat
larger polymers than the DHP from coniferyl alcohol, a common
monolignol (Fig. 1), most likely due to the charges on the carbox-
ylic acid groups, which increase the solubility (and thereby the
reactivity) of lignin oligomers. Thus, there is no reason to believe
that reduction of the monolignol precursors at the c-carbon is nec-
essary to obtain a polymer. We investigated the hydrophobicity/
hydrophilicity of the polymers by a solvent panel and by contact
angle measurement of DHP surfaces. The results indicated that
coniferyl alcohol DHP was the least hydrophobic. Therefore, the
reduction at the
c-carbon could be a way to make the lignin
slightly more hydrophilic. This is surprising for two reasons. First,
one of the biological roles of lignin is to make the cell wall hydro-
phobic and waterproof, which is probably a prerequisite for the
development of water conducting tissues in plants. However, a lig-
nin structure that is too hydrophobic may interact poorly with cel-
lulose and hemicelluloses, and therefore not fulfill its biological
functions efficiently. Second, in contrast to the primary alcohol of
coniferyl alcohol, the carboxylic acid in ferulic acid is a charged
structure at physiological pH. Thus, it was expected that the DHP
made of ferulic acid would be considerably more hydrophilic than
the DHP made from coniferyl alcohol, not the opposite as was seen
in our results.
3. Discussion
In summary, our investigation has demonstrated that mono-
mers with carbonyl c-carbons can form dehydrogenated polymers
with generally the same types of inter-unit bonds as natural lig-
nins. Although synthetic polymers from coniferaldehyde and feru-
lic acid have been produced through dehydrogenation in various
works (Karmanov et al., 1991; Higuchi et al., 1994; Ward et al.,
2001), no comparison of physical and chemical properties of the
DHP products with the more common coniferyl alcohol DHP has
been presented to our knowledge. In this work, we raise the ques-
tion of why the alcohol-type monomers are seemingly preferred in
lignin biosynthesis and suggest some answers based on the com-
parison of the different DHPs obtained.
The NMR characterization demonstrated that, in addition to the
expected differences from the
c-carbon signals (carboxylic acids
for the ferulic acid DHP, aldehydes for the coniferaldehyde DHP

A. Holmgren et al. / Phytochemistry 70 (2009) 147–155
151
Table 3
Signal assignments from 13C NMR spectra of ferulic acid (FA), coniferaldehyde (CAld) and coniferyl alcohol (CA) DHPs in DMSO-d6.
Signal
d
Assignments
Peaks in 13C NMR, ferulic acid DHP
1
2
3
4
5
6
7
171.7
170.4
167.8
150.0
149.1
147.7
147.1
145.4
144.2
143.8
138.7
129.5
128.4
126.7
124.2
121.8
118.1
116.9
115.6
115.2
112.4
111.5
110.3
86.7
C
c
in b-50
Carbonyl carbon from ethyl acetate (solvent)
c0 (FA) in b-O-40, C in b–b0 nc; C and Cc0 (FA) in b-50
C3 in 4-O-50
C30 and C3 in 4-O-50 and C40 in b-50 c; C30 and C4 in b-O-40
c
C
c
c
c
C3, C40 in b-O-40; C4 and C30 in b-50 nc; C3 in b-50; C3 in 5–50; C4 in 4-O-50
C3 and C40 in b-O-40; C4 in b–b0 c; C3 in b–b0 nc; C40 in b-50 nc; C4 in b-50
C3 in b-50 nc
c
8
9
C
C
C
a0 (FA) in b-50 nc; C
a
in 5–50; Ca0 (FA) and C30 in b-50
c
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
a0 (FA) in b-O-40; Ca0 (FA) and C30 in b-50 c, C
a
and Ca0 (FA) 4-O-50
a
in b-50 nc; Cb in b-O-40; C1 in b-5 c
C1 in bꢁb0
C10 in b-O-40; C1 in b–b0
c
Ca
in b-O-40; Cb in b-50 nc
C6 in b-O-40; C6 in b–b0 nc; C50 and C60 in b-50 nc
C60 in b-O-40; C6 in 4-O-50
C60 in b-50
c
Cb0 (FA) in b-50 c; Cb0 (FA) in 4-O-50; Cb0 (FA) in b-O-40
C5 in b-50 nc
C5 in b-O-40; C5 in b–b0 c; C5 in b–b0 nc; C5 in b-50
c
C2 in b-O-40; C2 in b-50 nc; C20 in b-50 c; C2 in 4-O-50
C20 in b-O-40; C2 in 4-O-50
C2 in b–b0 c; C20 in b-50 nc; C2 in b-50
c
C
C
a
a
in b-50
in b–b0
c
c
80.3
73.5
59.8
55.9
52.7
39.4
Cb in b-O-40 trimer (see b1 in Fig. 6)
CH2–C@O ethyl acetate (solvent)
Methoxy carbons
Cb in b-50 c; b in b–b0
Methyl carbon, DMSO
c
Peaks in 13C NMR, coniferaldehyde DHP
1
2
3
4
5
6
7
194.3
193.1
163.3
153.2
152.9
150.0
147.6
134.2
132.6
128.8
126.5
125.8
125.1
123.2
115.9
113.6
111.5
109.5
55.6
C
C
c0 (CAld) in b-O-40; C
c0 (CAld) in b-50 c; C
c
c
and Cc0 (CAld) in b-50 nc; Cc0 (CAld) in b-50
in b–b0
c
Unknown
C
C
a0 (CAld) in b-O-40; Ca0 (CAld) in b-50 c; C
a0 (CAld) in b-O-40; C in b–b0
a
in b–b0
a
C3 in b-O-40; C40 in b-50 nc; C4 in b–b0; C5 in 5–50
Cb in b-O-40; C3 in b-50 nc; C3 and C30 in b-50
Cb in bꢁb0
c
8
9
C1 in b-50
c
10
11
12
13
14
15
16
17
18
19
20
Cb0 (CAld) in b-O-40
C6 in b-O-40; C6 and C60 in b-50 nc; C6 in b–b0, b in 5–50
C1 in b-O-40; C60 in b-50 nc; C50 in b-50 c; C6 in b–b0, C4 in 5–50
C1 in b-O-40
Cb in b-50 nc; C60 in b-O40
C5 in b-O-40; C5 in b-50 nc; C5 in b-50 c; C5 in b–b0
C2 in b-O-40; C2 in b-50 nc; C2 in b–b0
C20 in b-O-40; C20 in b-50 nc; C2 in b-50
c
C2 in 5–50
Methoxy carbons
Methyl carbon, DMSO
39.2
Peaks in 13C NMR, coniferyl alcohol DHP
1
2
3
4
5
6
7
8
194.2
152.9
151.8
150.7
149.8
147.5
147.1
146.1
143.7
143.1
138.1
136.5
134.6
132.2
130.6
129.0
128.6
128.1
125.4
120.4
119.0
C
c
, coniferaldehyde
C30 in b-O-40; phenolic C3 in 5–50
C40 in b-O-40, C30 in -O-40
in coniferaldehyde
C40 in b-50; CA C30 in b-O-40; C30 in 4-O-50
C40 in b-O-40; C40 in -O-40; C3 in b–b0
C3 in b-O-40 -O-40 b-O-50
C4 in
-O-40 in b-50; C4 in b–b0
a
Ca
a
a
a a
-O-40; C4 in
9
C30 in b-50; C4 in 4-O-50
Unknown
10
11
12
13
14
15
16
17
18
19
20
21
C4 in 5–50
C1 and C10 in b-O-40, b-50; C1 in 5–50
C
a0 (CA) in b-50
C1 in b-O-40, b-50 and b–b0; C5 in 5–50
C10 in b-O-40 and b-50; C1 in -O-40
a0 (CA) in b-50; C50 (CA) in b-50
a
C
C
a0 (CA) in b-O-40 and
a
a
-O-40
-O-40
Cb0 (CA) in b-O-40 and
C1 in 5–50
C6 in b-O-40
C6 in b–b0
(continued on next page)

152
A. Holmgren et al. / Phytochemistry 70 (2009) 147–155
Table 3 (continued)
Signal
d
Assignments
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
118.6
116.1
115.0
111.9
110.4
109.9
105.1
103.6
86.9
85.2
82.6
79.0
71.0
67.0
63.0
61.7
59.8
55.7
53.6
53.1
39.8
C6 in b-50
C50 in b-O-40 and -O-40
a
C50 and C5 in b–b0; C60 in b-50
C2 in b-O-40
C2 and C20 in b-50 and b–b0
C20 in b-O-40 and
Unknown
a
-O-40
Unknown
C
C
a0 (CA) in b-50
a0 (CA) in b–b0
Cb0 (CA) in b-O-40; Cb in
a
-O-40
C
C
a
a
in
a
-O-40
in b-O-40;
c
in b–b0
Unknown
in b-50
Cb0 (b carbon in CA endgroup)
in b-O-40 and -O-40
Cc
C
c
a
Methoxyl C
Cb in b–b0
Cb in b-50
Methyl carbon, DMSO
c, cyclic; nc, non-cyclic. (FA) (CAld) and (CA) are used when the signals come ferulic acid, coniferaldehyde and coniferyl alcohol, respectively, as end-groups, with unsaturated
and b carbons.
a
Fig. 5. Types of inter-monomeric structures found in the three different DHPs as deduced from the 13C NMR spectra. The structures displaying unsaturated
a–b bonds were
prevalent in the coniferaldehyde and ferulic acid DHPs.
Fig. 6. AFM tapping mode height images (1 ꢂ 1
l
m2) and rms roughness determinations of the different spin-coated DHP model surfaces that were used in the
measurements of the contact angle of water. (a) DHP from coniferyl alcohol; rms roughness 0.349 nm. (b) DHP from coniferaldehyde; rms roughness 0.411 nm. (c) DHP from
ferulic acid; rms roughness 0.404 nm.
and primary alcohol for the coniferyl alcohol lignin), there was one
main difference between the coniferyl alcohol DHP and ferulic
acid/coniferaldehyde DHPs: the latter DHPs displayed nearly only
–b unsaturated side-chains, whereas the coniferyl alcohol DHP
a

A. Holmgren et al. / Phytochemistry 70 (2009) 147–155
153
Table 4
Initial equilibrium contact angles of water on DHP model surfaces.
Sample
Contact angle (°)
DHP from ferulic acid
63 0.3
60 0.2
58 0.5
46*
DHP from coniferaldehyde
DHP from coniferyl alcohol
Kraft lignin (spruce)
*
From Norgren et al. (2006).
displayed both unsaturated and saturated
a–b bonds. In coniferyl
alcohol DHPs, unsaturated –b side-chains commonly exist in
a
coniferyl alcohol end-groups, whereas the saturated structures
are formed through b-coupling. Additionally, benzyl alcohols, a fre-
quent product of b-coupling of coniferyl alcohol DHP (b-O-40;
Fig. 5) may partially explain the lower hydrophobicity. Neverthe-
less, it is puzzling that the DHP made of ferulic acid is as hydropho-
bic as it is. One possible explanation is that it has undergone
substantial decarboxylation, and thus loss of charge. Apparently,
there is a tendency for decarboxylation of certain structures in
the radical polymerization of ferulic acid (Ralph et al., 1994,
2007; Ward et al., 2001), which can lead to new structures through
further oxidation and coupling (Fig. 7). If and to what extent decar-
boxylation has occurred in our polymerisate is difficult to ascertain
from the NMR measurements, since signals from such structures
were difficult to distinguish from the non-decarboxylated ones.
Why do the coniferaldehyde and ferulic acid DHPs display these
differences compared with the coniferyl alcohol DHP, when the
Fig. 7. a1 and a2 are products of decarboxylation in ferulic acid dehydrogenative
polymerization; b1 and b2 are products of further oxidative coupling of a1 and a2,
respectively.
coupling reactions do not directly involve the-carbon? An explana-
tion might be the acidity of the b proton that is released upon the
rearomatization of the quinone methide intermediates in the case
of the coniferaldehyde and ferulic acid DHPs (Fig. 8). In the latter
Fig. 8. Hypothetical mechanisms for reactions of quinone methides formed by radical couplings of the monolignols used in this work. (a) The quinomethide generated from
coniferyl alcohol is relatively stable, and water can perform a nucleophilic attack on the -carbon. (b) The quinone methide generated from coniferaldehyde is stabilized by
a
deprotonation of the b-proton, which is relatively acidic due to the neighbouring carbons with d+. (c) The quinone methide generated from ferulic acid may undergo
decarboxylation caused by the relatively electron abstracting b-carbon, leading to a rearomatized product (Ralph et al., 2007). Alternatively, the quinone methide generated
from ferulic acid system might also react similarly to that of case (b).

154
A. Holmgren et al. / Phytochemistry 70 (2009) 147–155
cases, a decarboxylation is an additional way of rearomatizing a
quinone methide. These additional possibilities for the rearomati-
zation of the quinone methide intermediate during ferulic acid
and coniferaldehyde dehydrogenative polymerization apparently
limit the nucleophilic attacks by water or a phenol, as compared
to coniferyl alcohol polymerization. Due to the carbonyl groups
4. Conclusions
ꢀ Coniferaldehyde and ferulic acid form polymers equivalent to
coniferyl alcohol polymers by dehydrogenative polymerization.
ꢀ The DHPs of coniferaldehyde and ferulic acid were more hydro-
phobic than the DHP made of coniferyl alcohol.
at the
c position, the formation of a Ca–Cb double bond is predom-
ꢀ The DHPs of coniferaldehyde and ferulic acid displayed mostly
unsaturated inter-unit structures and only small amounts of
cyclic inter-unit structures for the ferulic acid DHP, whereas
coniferyl alcohol DHP displayed a variety of saturated and cyclic
inter-unit structures.
inant during coupling in both cases. In Fig. 8, the hypothetical
mechanisms are summarized for the creation of the most common
intermonolignol bond in lignin, the b-O-40 bond. Another way of
rearomatizing the quinone methide intermediate is by an intramo-
lecular nucleophilic attack leading to a cyclic structure. Although
the possibility of forming cyclic structures (i.e., b-50 and b–b0 for
the ferulic acid and b-50 for the coniferaldehyde DHP), producing
ꢀ Lignin made of coniferyl alcohol may be able to form LCC easier
than lignin made of coniferaldehyde or ferulic acid.
saturated
a–b side-chains, has been shown elsewhere (Ward
et al., 2001; Kim et al., 2003), in this work, only weak signals from
possible cyclic b-50 and b–b0 structures were produced in the ferulic
acid DHPs. Other reports of coniferaldehyde dehydrogenative poly-
merization support the lack of cyclic b-50 structure (Connors et al.,
1970; Higuchi et al., 1994). However, it has been proposed that this
cyclic b-50 is formed first, but that it is not thermodynamically sta-
ble and is subsequently transformed into the non-cyclic form (Kim
et al., 2003). In the case of the ferulic acid DHP, a suggestion has
been made that cyclic and non-cyclic dimers of b-50 and b–b0 struc-
tures could both be present and in some sort of equilibrium (Ward
et al., 2001; Arrieta-Baez and Stark, 2006).
5. Experimental
5.1. Materials
Coniferaldehyde
(4-hydroxy,3-methoxy-cinnamaldehyde),
ferulic acid (4-hydroxy,3-methoxy-cinnamic acid), and horseradish
peroxidase (HRP) type VI were purchased from Sigma–Aldrich
Sweden AB, Stockholm, Sweden. Coniferyl alcohol was obtained
from the reduction of coniferaldehyde with NaBH4 (Ludley and
Ralph, 1996). All other chemicals were of analytical grade.
Can these differences in the structure provide an explanation
for the large investment by plants in a reduction of the mono-
5.2. Making of synthetic lignin (DHP)
mers at the
c-carbon? Apart from an increased hydrophobicity,
Coniferaldehyde-based DHP and coniferyl alcohol-based DHP
were synthesized using the following method: 1 g of monomer
was dissolved in 10 ml acetone and then mixed together with
HRP (5 mg, 1450 U) and 190 ml 50 mM K2HPO4 buffer (pH 6.5). A
solution of hydrogen peroxide (74 mM) in 200 ml phosphate buffer
was dropped at 10 ml ꢃ hꢁ1 into the monolignol solution. After
24 h, the reaction was stopped and the suspended polymerisate
was centrifuged at 10,000 rpm and the pellet washed with deion-
ized water 2–3 times. Finally, the pellet was suspended in water
to be freeze-dried. Low molecular weight compounds were ex-
tracted for 1 h with 50 ml CH2Cl2. The ferulic acid-based DHP
was similarly prepared except for the following modifications:
ferulic acid was dissolved in 10 ml of acetone:K2HPO4 buffer
(1:1) before adding to the rest of the buffer. The pH of the reaction
mixture was adjusted back to 6.5 with diluted NaOH after the
addition of ferulic acid. After the reaction, diluted H2SO4 was added
until the pH was 2–3 to precipitate the polymerisate. Low
molecular weight compounds were extracted with ethyl acetate
for 1 h.
the polymerization of coniferaldehyde and ferulic acid (Fig. 8)
could be expected to lead to less covalent bonding between lignin
and cell wall polysaccharides. Such covalent bonds, which yield
lignin-carbohydrate complexes (LCC), may be created by nucleo-
philic attacks on the a-carbon of the quinone methide intermedi-
ate by alcohols or carboxylic acids on the polysaccharide, i.e., by
replacing water in the mechanism of Fig. 8a (Sarkanen and Lud-
wig, 1971; Eriksson et al., 1980). As the mechanisms in Figs. 8b
and c describe, and the results of this studies show, the possibil-
ities for extramolecular nucleophilic attacks might be limited and
consequently, limit the formation of this type of LCC. Recently, it
was demonstrated that LCC form networks in wood, where lignin
cross-links different kinds of polysaccharides (Lawoko et al.,
2006). These networks have been suggested to play an important
role in the mechanical properties of the wood. Thus, it might be
possible that one reason for the reduction at the
c-carbon is to
improve the formation of lignin-polysaccharide covalent net-
works. Interestingly, a mutant loblolly pine was discovered that
had deficient expression of the cinnamyl alcohol dehydrogenase
enzyme of the phenylpropanoid pathway (step 5, Fig. 2). The via-
ble plant exhibited a decreased lignin content, but also had
abnormally high amounts of aldehyde groups from coniferalde-
hyde end-groups, among others (Ralph et al., 1997). Coniferalde-
hyde was able to copolymerize with normal lignin monomers to
some degree and keep the lignin polymer somewhat functional.
However, this does not indicate that the reduction is unnecessary
for obtaining the fully functional lignin, since the lignin of this
mutant still contained mostly alcohol monomers, which can form
5.3. Solubility panel
The solubility of the three DHP materials was analyzed by dis-
solving 1–2 mg into 0.5 ml solvent and determining by simple ocu-
lar inspection if the material was completely, partly or poorly
dissolved.
5.4. Molecular mass determinations
the
a-carbon LCC.
Samples of each DHP product were acetylated in acetic anhy-
dride:pyridine (1:1) overnight and analyzed by size exclusion
chromatography (SEC), performed on a system of three cross-
linked (PS) columns in series (Styragel, Waters HR4 + HR2 + HR0.5)
using THF as the effluent at 0.8 ml ꢃ minꢁ1. Detection was per-
formed with a UV-light detector set at 280 nm. External calibration
was made with a kit of monodisperse polystyrene standards in the
Lignin has a unique ability among biopolymers of ‘‘metabolic
plasticity”, i.e., that non-conventional monomers can be incorpo-
rated into the polymer (Ralph et al., 1997; Pilate et al., 2002). How-
ever, this should not be understood as any potential monomer
being able to form a highly functional lignin. The standard mono-
mers of lignin have been very well chosen by nature to fulfill its
biological functions, as discussed here for the reduction at the
molar mass range of 580–3,053,000 g ꢃ molꢁ1
.
c-carbon.

A. Holmgren et al. / Phytochemistry 70 (2009) 147–155
155
Arrieta-Baez, D., Stark, R.E., 2006. Modeling suberization with peroxidase-catalyzed
polymerization of hydroxycinnamic acids: cross-coupling and dimerization
reactions. Phytochemistry 67, 743–753.
Baucher, M., Monties, B., Van Montagu, M., Boerjan, W., 1998. Biosynthesis and
genetic engineering of lignin. Crit. Rev. Plant Sci. 17, 125–197.
Boerjan, W., Ralph, J., Baucher, M., 2003. Lignin biosynthesis. Annu. Rev. Plant Biol.
54, 519–546.
Brett, C., Waldron, K., 1996. Physiology and Biochemistry of Plant Cell Wall.
Chapman and Hall, London.
Connors, W.J., Chen, C.-L., Pew, J.C., 1970. Enzymic dehydrogenation of the lignin
model coniferaldehyde. J. Org. Chem. 35, 1920–1924.
Ede, R.M., Kilpeläinen, I., 1995. Unambiguous structural probes for non-cyclic
benzyl aryl ethers in soluble lignin samples. In: 8th International Symposium
on Wood and Pulping Chemistry, Helsinki, June 6–9, 1995, pp. 487–494.
Eriksson, O., Goring, D.A.I., Lindgren, B.O., 1980. Structural studies on the chemical-
bonds between lignins and carbohydrates in spruce wood. Wood Science and
Technology 14, 267–279.
5.5. NMR
NMR experiments were performed on a Bruker Avance 400 MHz
instrument. 13C NMR was performed on the DHP polymeric frac-
tion by dissolving the dried material in DMSO-d6. A pulse program
with inverse-gated proton decoupling was applied to acquire 13C
NMR data. The measurements were acquired at 24 °C, with a spec-
tral window of 240 ppm centered at 110 ppm. The data points ac-
quired in the time domain were set to 32 K, which gave an
acquisition time of 0.68 s. A relaxation delay of 14 s was used to
avoid signal saturation. Five thousand and eight hundred scans
were collected for each measurement.
Freudenberg, K., 1959. Biosynthesis and constitution of lignin. Nature 183, 1152–
1155.
5.6. Preparation of DHP model surfaces
Higuchi, T., Ito, T., Umezawa, T., Hibino, T., Shibata, D., 1994. Red-brown color of
lignified tissues of transgenic plants with antisense CAD (cinnamyl alcohol
dehydrogenase) gene: wine-red lignin from coniferyl aldehyde. J. Biotechnol.
37, 151–158.
Karmanov, A.P., Belyaev, V.Y., Monakov, Y.B., 1991. Dehydropolymers from ferulic
acid. Koksnes Kimija, 73–75.
In the preparation of DHP model surfaces, a Chemat Technology
KW-4A spin-coater was used. Silicon wafers cut in pieces of
approximately 1 cm2 were used as substrates. The DHP samples
were dissolved in pyridine overnight to yield concentrations of
2.0 wt%. Thereafter, the DHP solutions were gently passed through
Kim, H., Ralph, J., Lu, F., Ralph, S.A., Boudet, A.-M., MacKay, J.J., Sederoff, R.R., Ito, T.,
Kawai, S., Ohashi, H., Higuchi, T., 2003. NMR analysis of lignins in CAD-deficient
plants. Part 1. Incorporation of hydroxycinnamaldehydes and hydroxy-
benzaldehydes into lignins. Org. Biomol. Chem. 1, 268–281.
Landucci, L.L., Ralph, S.A., Hammel, K.E., 1998.
a 0.45 l
m MillexÒ-HV Durapore PVDF syringe-driven filter unit to
remove possible dust and undissolved DHP particles. Immediately
before the spin-coating, the silicon wafers were exposed to 0.1 M
NaOH for twenty seconds to hydroxylate the SiO2 layer, rinsed with
plenty of Milli-Q water, and finally dried by a stream of nitrogen
gas. In the spin-coating, a rotation speed of 3500 rpm was applied
for 1 min.
13C NMR characterization of guaiacyl,
guaiacyl/syringyl, and syringyl dehydrogenation polymers. Holzforschung 52,
160–170.
Lawoko, M., Henriksson, G., Gellerstedt, G., 2006. Characterisation of lignin-
carbohydrate complexes (LCCs) of spruce wood (Picea abies L.) isolated with
two methods. Holzforschung 60, 156–161.
Ludley, F.H., Ralph, J., 1996. Improved preparation of coniferyl and sinapyl alcohols.
J. Agric. Food Chem. 44, 2942–2943.
Lundquist, K., Simonson, R., Tingsvik, K., 1983. Lignin carbohydrate linkages in
milled wood lignin preparations from spruce wood. Svensk Papperstidning 86,
R44–R47.
Norgren, M., Notley, S.M., Majtnerova, A., Gellerstedt, G., 2006. Smooth model
surfaces from lignin derivatives. I. Preparation and characterization. Langmuir
22, 1209–1214.
Pilate, G., Guiney, E., Holt, K., Petit-Conil, M., Lapierre, C., Leple, J.-C., Pollet, B., Mila,
I., Webster Elizabeth, A., Marstorp Hakan, G., Hopkins David, W., Jouanin, L.,
Boerjan, W., Schuch, W., Cornu, D., Halpin, C., 2002. Field and pulping
performances of transgenic trees with altered lignification. Nat. Biotechnol.
20, 607–612.
5.7. Contact angle measurements
The contact angle measurements were performed on a Fibro
DAT dynamic angle tester. This instrument uses image analysis
when monitoring the spreading of a liquid on a substrate as a func-
tion of time. In this work, water was used as the fluid. The initial
equilibrium contact angle between the water drop and the surface
was evaluated and reported as an average of 4 measurements.
Ralph, J., Kim, H., Lu, F., Grabber, J.H., Leple, J.-C., Berrio-Sierra, J., Derikvand, M.M.,
Jouanin, L., Boerjan, W., Lapierre, C., 2007. Identification of the structure and
origin of a thioacidolysis marker compound for ferulic acid incorporation into
angiosperm lignins (and an indicator for cinnamoyl-CoA reductase deficiency).
Plant J. 53, 368–379.
Ralph, J., MacKay, J.J., Hatfield, R.D., O’Malley, D.M., Whetten, R.W., Sederoff, R.R.,
1997. Abnormal lignin in a loblolly pine mutant. Science 277, 235–239.
Ralph, J., Quideau, S., Grabber, J.H., Hatfield, R.D., 1994. Identification and synthesis
of new ferulic acid dehydrodimers present in grass cell-walls. J. Chem. Soc.
Perkin Trans. 1, 3485–3498.
Ralph, S.A., Ralph, J., Landucci, L.L., 2004. NMR Database of Lignin and Cell Wall
Model Compounds. Retrieved November 2004, from <http://ars.usda.gov/
Services/docs.htm?docid=10491>.
Sarkanen, K.V., Ludwig, C.H. (Eds.), 1971. Lignins: Occurrence Formation, Structure
and Reactions. Wiley, New York.
5.8. AFM surface imaging and roughness determinations
A Veeco Picoforce scanning probe microscopy equipped with a
Veeco silica cantilever was used in tapping mode to image the
DHP model surfaces. The root-mean-square (rms) roughness val-
ues of the films were determined from the height images. All of
the images were recorded in ambient air conditions (25 °C and
50% relative humidity).
Acknowledgements
Sederoff, R.R., Chang, H.-M., 1991. Lignin biosynthesis. In: Menachem, L., Goldstein,
I.S. (Eds.), Wood Structure and Composition. Marcel Dekker Inc., New York, pp.
263–285.
Sjöström, E., 1981. Wood Chemistry: Fundamentals and Applications. Academic
Press, London.
Ward, G., Hadar, Y., Bilkis, I., Konstantinovsky, L., Dosoretz, C.G., 2001. Initial steps of
ferulic acid polymerization by lignin peroxidase. J. Biol. Chem. 276, 18734–
18741.
This work was financially supported by the Swedish Research
Council for Environmental, Agricultural Science and Spatial Plan-
ning (FORMAS), contract #229-2004-1240. Mr. Erik Johansson is
thanked for helpful assistance in the AFM characterisation.
References
Xie, Y., Yasuda, S., Wu, H., Liu, H., 2000. Analysis of the structure of lignin-
carbohydrate complexes by the specific13C tracer method. J. Wood Sci. 46, 130–
136.
Adler, E., 1977. Lignin chemistry – past, present and future. Wood Sci. Technol. 11,
169–218.