Chemical cross-linking gelatin with natural phenolic compounds as studied by high-resolution NMR spectroscopy

Article abstract of DOI:10.1021/bm1001284

Cross-linking gelatin with natural phenolic compound caffeic acid (CA) or tannic acid (TA) above pH 9 resulted in formation of insoluble hydrogels. The cross-linking reactivity was controlled by variation of pH, the concentration of the gelatin solution, or the amount of CA or TA used in the reaction. The cross-linking chemistry was studied by high-resolution NMR technique in both solution and solid state via investigation on small molecular model systems or using ¹³C enriched caffeic acid (LCA) in the reaction with gelatin. Direct evidence was obtained to confirm the chemical reactions occurring between the phenolic reactive sites of the phenolic compounds and the amino groups in gelatin to form C-N covalent bonds as cross-linking linkages in gelatin matrix. The cross-linked network was homogeneous on a scale of 2-3 nm. The cross-linking resulted in a significant decrease in the molecular mobility of the hydrogels, while the modulus of the films remained at high values at high temperatures.

Full text of DOI:10.1021/bm1001284

Biomacromolecules 2010, 11, 1125–1132
1125
Chemical Cross-Linking Gelatin with Natural Phenolic
Compounds as Studied by High-Resolution NMR Spectroscopy
Xiaoqing Zhang,*,† My Dieu Do,† Philip Casey,† Adrian Sulistio,‡ Greg G. Qiao,‡
Leif Lundin,§ Peter Lillford,§ and Shansha Kosaraju§
CSIRO Materials Science and Engineering, Private Bag 33, Clayton South MDC, Clayton South,
Victoria 3169, Australia, Polymer Science Group, Department of Chemical and Biomolecular Engineering,
The University of Melbourne, Victoria 3010, Australia, and CSIRO Food and Nutritional Science,
671 Sneydes Road, Werribee, Victoria 3030, Australia
Received February 2, 2010; Revised Manuscript Received February 26, 2010
Cross-linking gelatin with natural phenolic compound caffeic acid (CA) or tannic acid (TA) above pH 9 resulted
in formation of insoluble hydrogels. The cross-linking reactivity was controlled by variation of pH, the concentration
of the gelatin solution, or the amount of CA or TA used in the reaction. The cross-linking chemistry was studied
by high-resolution NMR technique in both solution and solid state via investigation on small molecular model
systems or using 13C enriched caffeic acid (LCA) in the reaction with gelatin. Direct evidence was obtained to
confirm the chemical reactions occurring between the phenolic reactive sites of the phenolic compounds and the
amino groups in gelatin to form C-N covalent bonds as cross-linking linkages in gelatin matrix. The cross-
linked network was homogeneous on a scale of 2-3 nm. The cross-linking resulted in a significant decrease in
the molecular mobility of the hydrogels, while the modulus of the films remained at high values at high temperatures.
materials.20–27 It is generally accepted that four potential in-
teractions may be involved, such as hydrogen bonding, ionic,
hydrophobic interactions, and covalent bonding.20,24,28–32 Co-
valent bonds between phenolic compounds and proteins are
more rigid and thermally stable than other interactions. The
postulated chemical pathway involves initial oxidization of
phenolic structures to form, under alkaline conditions, quinone
intermediates that can readily react with nucleophiles from
reactive amino acid groups in protein chains such as the
sulfuhydryl group of cystine, the amine group of lysine and
arginine, the amide group of asparagine and glutamine, the
indole ring of tryptophan, and the imidazole ring from
histidine.25,31,33 However, these reactions are complicated, and
the interactions between phenolic compounds and proteins in
forming cross-linked materials are poorly understood, particu-
larly in regard to the chemical structures of intermediates and
the chemistry of covalent cross-linking.
Introduction
Chemical cross-linking of natural proteins has a long history
in the development of protein-based polymer materials for food
processing, packaging, coating formulations, and pharmaceuti-
cal/medical applications. Gelatin, as a water-soluble protein
produced from hydrolysis of animal collagen, has been widely
used in these applications.1–6 Because gelatin can be obtained
from inexpensive sources derived from waste or byproduct of
manufacturing processes comprising tannery, pharmaceutical,
and food segments, it can be taken as a renewable and biode-
gradable material.7 However, its poor mechanical properties
especially when exposing to wet and humid conditions limit its
application in many areas. Structure modifications are required to
enhance its mechanical strength and water-resistant properties. The
traditional aldehyde-based cross-linkers (including formaldehyde,
glyoxal, and glutaraldehyde) have shown great efficiency in protein
modifications in producing biomaterials,8–12 the usage of these
additives has also caused a series of OH&S issues in product
manufacturing. The toxic nature of these compounds during mate-
rial biodegradation process also prohibits their application in food
and medical applications.13 Alternative cross-linkers such as
carbodiimides, epoxy, genipin, oxidized alginate, and transglutami-
nase have also been studied for protein cross-linking in recent
years.14–18
In this paper, gelatin cross-linking reactions were conducted
using natural phenolic compounds caffeic acid (CA) and tannic
acid (TA) as cross-linking agents. The chemical structure
changes occurring during the reaction were examined by high-
resolution NMR spectroscopy. Because only a small amount
of cross-linking agent is normally used in the reaction and the
signals derived from the cross-linker are usually too weak to
be observed, a 13C-labeled caffeic acid (LCA) was used in
parallel with unlabeled CA to enhance the signals derived from
LCA. This makes it possible to detect the behavior of the cross-
linker in the cross-linked gelatin network by high-resolution
solution- and solid-state NMR. Additionally, small molecular
model systems were studied using 2,4-dimethyl phenol (24DMP)
and 2,6-dimethyl phenol (26DMP) as model phenolic com-
pounds reacting with lysine as the model amino group in gelatin.
Because only a single phenolic reactive site is available in these
model compounds (ortho-reactive site in 24DMP and para-
reactive site in 26DMP), no cross-linked materials would be
formed in such reactions, allowing product structures to be
Natural phenolic compounds commonly found in plants and
fruits19 as antioxidant reagents have also been widely used in
food processing. The study of natural phenol-protein interac-
tions has been carried out for more than half a century. A
number of natural phenolic compounds derived from plants have
been reported to be interactive or reactive with proteins and
resulted in improved gel or film properties for gelatin-based
* To whom correspondence should be addressed. E-mail: xiaoqing.zhang@
csiro.au.
† CSIRO Materials Science and Engineering.
‡ The University of Melbourne.
§ CSIRO Food and Nutritional Science.
10.1021/bm1001284
 2010 American Chemical Society
Published on Web 03/17/2010

1126 Biomacromolecules, Vol. 11, No. 4, 2010
Zhang et al.
determined by solution NMR. Reactions using CA or TA as
the cross-linker were studied systematically by varying pH,
gelatin concentration, and the amount of cross-linking agent.
The physical properties and phase structures of the modified
gelatin materials were also examined using dynamic mechanical
analysis (DMA) and solid-state NMR techniques.
Experimental Section
Raw Materials. All raw materials used in this work, including
gelatin (type B obtained from bovine skin), caffeic acid (CA), tannic
acid (TA), L-lysine (Ly), 2,4 dimethyl phenol (24DMP), and 2,6-
dimethyl phenol (26DMP) were obtained from Sigma-Aldrich without
further purification. The 97% 13C-labeled caffeic acid (LCA, with all
the carbons in the compound were enriched) was synthesized by Sigma-
Aldrich as requested. All reactions were conducted in aqueous solution
using distilled water.
Figure 1. 13C NMR solution spectra of CA before oxidation, oxidized
at pH ) 9 for 1 h, and at pH ) 10 for 2 h in aqueous solution.
measured through the decay of 13C magnetization prepared by CP after
varied H spin-locking time, as reported previously.
Oxidation of CA and Ly. CA solution (20 wt %) was prepared by
dissolving CA in water under magnetic stirring. The temperature was
controlled using water bath set at 50 °C, and the pH was adjusted to a
target value (pH 9-10) by the addition of NaOH solution (46 wt %).
The solution was bubbled with oxygen during the reaction (50 °C for
1-3 h). Oxidization of Ly was conducted under the same conditions.
Reactions of Small Molecular Model Systems. Reactions in small
molecular model systems (24DMP/Ly and 26DMP/Ly) were conducted
using sodium periodate (NaIO4) as the oxidizing agent. The molar ratio
of Ly to 24DMP or 26DPM was 1:2; the amino functionality of Ly is
2, while the phenolic reactive site in 24DMP or 26DMP is only 1. The
reactions were conducted at 50 °C for 2 h. Solution concentration was
20 wt % and the pH was 9, which was adjusted by the addition of
NaOH solution (46 wt %).
Reactions of Gelatin and CA or TA. Gelatin (10 wt %) and CA
or TA solutions were prepared separately, and the pH of each solution
was adjusted to a target value (9-10) by the addition of NaOH (46%)
solution. The gelatin solution was heated to 45 °C to ensure complete
dissolution. CA or TA solution was then slowly added into the gelatin
solution at 58 ( 1 °C under stirring to a predetermined ratio and allowed
to react for a varied period of time. Solution pH (9-10) was monitored
using a pH meter, and a few drops of NaOH solution (1 M) was required
to maintain the desired pH during the reaction. Oxygen was continu-
ously bubbled throughout the reaction at a rate of 47 mL/min. As this
study was focused on cross-linking chemistry rather than reaction
dynamics, the effects from variations such as the oxygen bubbling rate,
bubble size, and stirring rate were not examined.
34,35
1
Dynamic mechanical analysis (DMA) experiments were operated
on a PerkinElmer PYRIS Diamond DMA in tension mode at a
frequency of 1 Hz, similar to what reported in literature.36 The
temperature range was set at -100 to 160 °C with a heating rate of 2
°C/min. The storage modulus (E′), the loss modulus (E′′), and tan δ
were recorded as a function of temperature throughout the experiment.
Results and Discussion
Oxidation of CA and Reactions in Model Phenol/Amino
Acid Systems. Reactions between natural phenolic compounds
and proteins have been postulated to be initiated by oxidation
of phenolic structures to form quinone intermediates under
alkaline conditions.37 As quinone structures are very reactive,
phenolic oligomers would be expected to form under such
conditions. In our studies, the CA solution changed color from
pale yellow to dark brown during the oxidation, a color typical-
ly associated with phenol oxidation. Figure 1 shows the 13C
solution NMR spectra of CA after the oxidation at pH 9 for 1 h
and at pH 10 for 2 h. It was noted that oxidation at pH 9/1 h
resulted in a series of new resonances. As separation and
purification of each product was beyond the scope of the current
study, it is not possible to assign all of these new resonances.
However, the formation of aliphatic resonances at 45-52 ppm
(corresponding to the >CH- and >C< structures) and 75-90
ppm (the >CH-O- structure) are consistent with the formation
of the dimeric structures (B) and (D) in Scheme 1 via different
mechanisms, as suggested in the literature,37–39 while resonances
around 128-130 ppm are consistent with the formation of
structure (A) in Scheme 1. A formic acid signal was also
detected at 165 ppm, likely resulting from dissociation of the
-COOH groups from CA oligomers and proton transfer during
the oxidation.37 At pH 10/2 h, all of the resonances became
broad, indicating formation of polymeric products. No observ-
able oxidation was detected for Ly under the same conditions.
Oxidation of phenolic compounds is prerequisite for their
reactive cross-linking with proteins. Such oxidation reactions
consume phenolic reactive sites and produce large phenolic
molecular species, but a sufficient number of remaining unre-
acted phenolic sites should still be present and therefore
available in these CA oligomers (Scheme 1) for cross-linking
reactions with proteins.
Gelatin films (thickness of around 0.1-0.2 mm) were produced by
solution casting at room temperature for 2 weeks after the reaction
without any heating or vacuum. Samples using LCA were prepared
under the same conditions. All film samples were stored in a
conditioning tank at a relative humidity (RH) of 52% and a temperature
of 22 °C for 7 days before characterization. The moisture content of
these films was around 17-19%, as determined by drying the samples
at 105 °C for 6 h.
Characterization. NMR experiments were conducted at room
temperature using a Varian NMR300 System at resonance frequencies
of 75 MHz for 13C and 300 MHz for 1H. High-resolution solution NMR
spectra were measured using a 10 mm solution probe. High resolution
solid-state 13C NMR spectra were observed under cross-polarization,
magic angle spinning, and high power dipolar decoupling (CP/MAS/
1
DD) conditions. The 90° pulse was 6.0 µs for H and 13C, while the
spinning rate of MAS was set at a value in the range of 7-9 kHz.
Different CP contact time (either 1 ms or 20 µs) was used to measure
the 13C CP/MAS spectra. The chemical shift of 13C CP/MAS spectra
was determined by taking the carbonyl carbon of solid glycine (176.3
ppm) as an external reference standard. 1H spin-spin relaxation times
(T2) for the gelatin hydrogels were measured by CPMG pulse
sequence34 using CP/MAS probe without sample spinning. The proton
spin-lattice relaxation times in the rotating frame (1H T1F) were
The formation of polymeric phenolic products and cross-
linked gelatin networks makes it difficult to study the reaction
pathway and chemical structures formed during reaction. To
overcome this difficulty, small molecular model phenol-amino

Cross-Linking Gelatin with Natural Phenolic Compounds
Biomacromolecules, Vol. 11, No. 4, 2010 1127
Scheme 1. Representative CA Dimeric Structures Proposed in
the Literature
Figure 2. 13C NMR solution spectra of 24DMP, Ly, 26DMP, 24DMP/
Ly, and 26DMP/Ly reactions at pH ) 9.
in parallel with unlabeled CA. The 13C solution NMR spectra
of CA and LCA are shown in Figure 3. A coupling split was
observed for all carbon resonances of LCA, because the coupling
interaction between 1H and 13C was very strong when all carbons
of the compound were 13C-enriched. The minor resonance
appearing at 76 ppm indicates the existence, as an impurity, of
dimer-D shown in Scheme 1. The 13C CP/MAS spectrum of
the solid LCA (top-right) displayed a significant line width
acid systems were designed to aid such an investigation. Two
compounds, 24DMP and 26DMP, were chosen as model phenols
because both compounds have only one (but different) reactive
site (ortho- or para-reactive site for 24DMP or 26DMP,
respectively). Ly was chosen as a model amino acid. Although
Ly contains two amino groups, 24DMP/Ly and 26DPM/Ly
systems are not expected to produce high molecular weight
cross-linked products and thus avoid the major difficulties
encountered in gelatin systems.
1
broadening due to strong dipolar interactions between H and
enriched 13C.
Figure 2 shows 13C solution NMR spectra of 24DMP/Ly and
26DMP/Ly systems after reaction at pH 9. Methanol-d4 was
used as solvent for measuring the spectra of 24- or 26DMP,
but the spectra of rest samples were detected in aqueous solution.
In the 24DMP/Ly system, a number of new resonances were
formed after reaction and the ortho-reactive site of 24DMP at
115 ppm disappeared. Two significant peaks were observed at
67 and 54 ppm, corresponding to a and b carbons in the
representative product structure shown in Scheme 2-1 having
C-N linkages between 24DMP and Ly. The corresponding
phenolic carbon resonances (c and d) of the structure appeared
at 145-147 ppm. Most other resonances also changed their
chemical shift due to formation of this 24DMP-Ly product. In
the 26DMP/Ly system, the situation was more complicated.
However, there was evidence that a similar 26DMP-Ly product
was formed from resonances at 77, 54 (a′ and b′ carbons), and
146 ppm (c′ and d′ carbons), corresponding to the product
structure shown in Scheme 2-2. 26DMP dimers and quinone
structures (>CdO at 205-215 ppm) were simultaneously
formed. Nevertheless, these small molecule model system
studies provided direct evidence of chemical reactions between
phenolic compounds and protein amino groups to form struc-
tures with C-N covalent linkages.
The reactions between gelatin and CA at pH 9-10 generated
hydrogel materials. This gel formation was a significant
phenomenon in the reaction. After a certain period of reaction
time, the viscosity of the gelatin solution suddenly started to
increase, and the clear solution became a gel material within a
few seconds. Figure 4 shows the 13C solution NMR spectra of
gelatin, CA, LCA, and gelatin/CA or gelatin/LCA after reaction
at pH 9, 58 °C, for 10 min (A), 20 min (B, just before gel
formation), and 30 min (C, extra 10 min heating after gel
formation). All of these spectra were taken after cooling samples
down to room temperature and stabilizing at room temperature
for 3-6 h for achieving equilibrium within the systems. The
signals derived from unlabeled CA (1.5 wt % to dry gelatin)
were barely detectable. But for those using LCA, because of
greatly enhanced signals, LCA resonances were observed clearly
providing direct information on its structural changes occurred
during the reaction.
After 10 min of reaction (A), LCA resonances were ef-
fectively unchanged and the system appeared as a homogeneous
mixture of pure gelatin and LCA. After 20 min of reaction (B),
significant LCA structural changes occurred such that the
resonance at 115 ppm (the ortho-reactive site of LCA) and 124
ppm (the para-reactive site of LCA) disappeared and strong
peaks at 130, 138, and 145 ppm were observed. This evidence
directly shows that reactions occur at ortho- and para-unsub-
Chemistry and Reactivity of Gelatin Reactions with CA
or TA. To enhance the signals derived from CA in the gelatin
cross-linking reaction, LCA was used in experiments running

1128 Biomacromolecules, Vol. 11, No. 4, 2010
Zhang et al.
Scheme 2. Representative Products Formed from Reactions between Ly and 24- or 26DMP
stituted phenolic reactive sites and that the C-N bonds between
CA and gelatin, as described in Scheme 3, are indeed formed.
As discussed above, CA dimers can also be formed during the
reaction. The sharp resonance detected at 165 ppm likely
corresponds to the formation of formic acid (HCOOH) through
oxidative dissociation from the CA. The double bond of CA
(C-7 and C-8 in Figure 3) may also be converted (according to
the changes in chemical shift) into other structures as suggested
in Scheme 1. The phenolic C-OH resonances (C-1 and C-2 in
Figure 3) also shifted to around 148 ppm after the reaction. At
longer reaction times, the gelatin solution converted to a gel
and the line width of the gel sample (C) significantly broadened
due to formation of cross-linked gelatin networks. Interestingly,
no new peaks were formed (compare to sample B), but the
resonance at 145 ppm (C-N linkages between LCA and gelatin)
was enhanced during the extra 10 min reaction.
Samples A, B, and C were dried at room temperature for 2
weeks (the same as solution casting) and then conditioned at
RH ) 52% for 7 days. The 13C CP/MAS NMR spectra of these
film samples are shown in Figure 5 together with that of pure
gelatin obtained under the same conditions. Although line widths
of the spectra became broader for all samples, resonances
derived from LCA observed in the range of 115-150 ppm were
significant. It is noted that the spectrum pattern of sample A
changed on drying and became similar to those obtained at
longer reaction times (samples B and C). This suggests that
cross-linking reactions also occurred during the drying and
conditioning process albeit to a lesser degree. For samples B
and C, the resonances at 130 and 145 ppm were much stronger
than those for sample A, indicating the number of the C-N
linkages in sample A was lower than those in samples B and
C.
To implicitly determine whether the formation of C-N
covalent linkages between LCA and gelatin had occurred, the
carbon resonances of the gelatin/LCA systems were examined
by high-resolution solid-state NMR using differing pulse
sequences in order to detect different carbon species. Figure 6
shows the 13C CP/MAS NMR spectra of A, B, and C film
samples measured by normal CP pulse sequence with a contact
time of 1 ms or 20 µs, or by a CP pulse sequence with 10 ms
spin-locking time in 13C channel before acquisition.35 With a
CP contact time of 1 ms, most resonances are detected for the
gelatin/LCA samples shown in Figures 5 and 6 (A-1, B-1, and
C-1). However, only carbons that are directly bonded to
hydrogens can be observed at short contact times (20 µs; Figure
6, A-2, B-2, and C-2). Here, the strong CdO peak at 175 ppm
as well as resonances at 130, 145-148 either disappeared or
their intensities were significantly reduced. This confirms that
resonances at 130 and 145 ppm had no bonded hydrogen, the
same as the CdO in gelatin and phenolic C-OH carbons of
LCA, corresponding to the implicit formation of C-N linkages
between LCA and gelatin. In contrast, using 10 ms spin-locking
time in 13C channel before acquisition detects only carbons with
longer 13C T1F, that is, normally carbons having no bonded
hydrogen(s) in amorphous polymers. Moreover, the spectra
shown in Figure 6 (A-3, B-3, and C-3) confirm that the
resonances at 130 and 145 ppm have longer 13C T1F values and
Figure 3. 13C NMR solution spectra of unlabeled CA, 97% 13C-labeled
CA (LCA) in aqueous solution and the CP/MAS spectrum of solid
LCA (top) where “ssb” represents spinning sideband.

Cross-Linking Gelatin with Natural Phenolic Compounds
Biomacromolecules, Vol. 11, No. 4, 2010 1129
Figure 4. 13C NMR solution spectra in aqueous solution for gelatin cross-linked with 1.5 wt % of unlabeled CA (left) or 97% 13C-labeled CA
(LCA, right) after reaction for 10 min (A), 20 min (B), and 30 min forming gel (C); *resonances were derived from LCA.
Scheme 3. Representative Products Formed in Reactions
between Gelatin and Caffeic Acid
Figure 5. 13C CP/MAS NMR spectra of solution cast gelatin samples
after cross-linking reactions with 97% 13C-labeled CA (LCA) for 10
min (A), 20 min (B), and 30 min forming gel (C); “ssb” represents
spinning sideband and “*” peaks were derived from LCA.
of reactant, that is, CA and TA and this may possibly be
explained by formation of a high degree of CA oligomers during
thus provide further evidence of the formation C-N covalent
linkages between LCA and gelatin.
oxidation and these CA oligomers would behave similarly to
TA in reactions with gelatin. The only data point where gel
times are significantly shorter for CA than TA is at the low
reactant (1.5 wt %) and solution concentration (10%). These
results demonstrated that the reactivity could be controlled by
varying the pH, the solution concentration and the amount or
the species of cross-link agents used in the cross-linking reac-
tions.
Properties of the Gelatin/CA and Gelatin/TA Gel and
Film Materials. The hydrogel materials obtained in the reaction
did not melt upon heating to 100 °C, being consistent with the
formation of stable covalent bonds within the gelatin matrix.
The effect of pH and CA or TA concentration on gel times
of gelatin solution (recorded when the conversion from clear
solution to gel occurred during the reaction) are presented in
Figures 7 and 8. Under reaction conditions of pH 9, 9.5, and
10 using 1.5 wt % of CA or TA (mass to dry gelatin) in a 10%
gelatin solution, the gel time of the gelatin/CA system decreased
abruptly as pH increased while that of gelatin/TA gradually
reduced as pH increased (see Figure 7). Varying the proportional
or reactant concentration also changed the gel time at pH 9 (see
Figure 8). The effect appeared not to be influenced by the choice

1130 Biomacromolecules, Vol. 11, No. 4, 2010
Zhang et al.
Figure 6. 13C NMR spectra of solution cast gelatin samples after cross-linking reactions with 97% 13C-labeled CA (LCA) for (A) 10 min, (B) 20
min, and (C) 30 min forming gel. (1) CP/MAS spectra with contact time of 1 ms. (2) CP/MAS spectra with contact time of 20 µs. (3) CP/MAS
spectra with 10 ms spin locking time in 13C channel before CP. “ssb” represents spinning sideband.
Table 1. 1H T2 Data (ms) at Room Temperature of Gelatin Gel
(10 wt %) Samples after Reactions at pH ) 9
samples
reaction conditions
1H T2 of the gel at RT
gelatin
30 min, no O2
218 ( 16
139 ( 5
133 ( 4
179 ( 23
90 ( 2
217 ( 14
92 ( 5
CA-1.5%
CA-3.0%
TA-1.5%
TA-1.5%
TA-3.0%
TA-3.0%
forming gel with O2
forming gel with O2
30 min, no O2
forming gel with O2
30 min, no O2
forming gel with O2
gelatin/CA gels was lower at 130-140 ms, suggesting a
Figure 7. Gel times at pH 9-10 when reacting gelatin with 1.5 wt %
of CA or TA in 10 wt % solutions.
significant restriction of the molecular motions due to the cross-
1
linking effect. For gelatin/TA gel, the H T2 value was even
1
lower ∼ 90 ms. The H T2 value of systems where no oxygen
was bubbled into the reaction solutions was universally higher
and confirmed that oxygen was necessary for the reaction.
Dynamic mechanical analysis (DMA) was conducted in
tension mode for these gelatin films to observe any changes in
viscoelastic response subsequent to reaction and these results
are shown in Figure 9. In general, all gelatin films had adequate
strength (storage modulus E′ of 6-9 GPa) at room temperature
and showed behavior typical of plasticized protein materials in
that their storage modulus slowly decreased as temperature
increased from -100 to 60 °C. Between -100 °C and ambient
temperature, CA or TA modified gelatin films displayed slightly
Figure 8. Gel times at pH 9 when reacting gelatin with 1.5 or 3.0 wt
% of CA or TA in 10 or 20 wt % solutions.
This is in contrast to the gelatin gel formed during cooling which
involving no chemical bonding. Strong intermolecular interac-
tions between macromolecular chains and chemical cross-linking
within the protein matrix will significantly restrict molecular
motions and enhance dipolar interactions within the gel system.
The 1H spin-spin relaxation time (T2) is sensitive to the
molecular motions of protein hydrogels. If strong intermolecular
interactions or chemical cross-linking has occurred, it will result
in decreased 1H T2 values of the gel samples. The 1H T2 data of
gelatin/CA and gelatin/TA gels obtained after reactions in 10%
solution at pH 9 are listed in Table 1. All of these samples were
aged at 5 °C for 24 h and then allowed to warm naturally to
room temperature before 1H T2 measurement. After the reaction
the gelatin-only gel had a 1H T2 value of 218 ms, while that of
Figure 9. E′ storage-modulus data measured by DMA: tension mode
for gelatin and modified gelatin films.

Cross-Linking Gelatin with Natural Phenolic Compounds
Biomacromolecules, Vol. 11, No. 4, 2010 1131
Table 2. 1H T1F Data (ms) of Gelatin/LCA Films
LCA component
gelatin component
samples
145 ppm
128 ppm
116 ppm
60 ppm
44 ppm
27 ppm
gelatin
A
B
C
4.7 ( 0.3
3.9 ( 0.1
4.2 ( 0.3
4.4 ( 0.2
4.9 ( 0.2
4.0 ( 0.1
4.2 ( 0.3
4.3 ( 0.2
4.9 ( 0.2
3.9 ( 0.1
4.2 ( 0.2
4.2 ( 0.2
3.9 ( 0.1
4.2 ( 0.6
4.6 ( 0.6
32.9 ( 1.4
4.1 ( 0.1
4.2 ( 0.3
4.4 ( 0.2
LCA
32.7 ( 1.1
lower E′ values than their unmodified counterpart suggesting
the protein aggregative structure of gelatin was modified. On
the other hand, the difference in moisture content of these
samples might also play a role in determining modulus behavior.
Although the moisture content was around 17-19 wt % for all
samples, the value for modified samples (18-19%) was 1-2%
higher than gelatin film (∼17%). As moisture is an efficient
gelatin plasticizer, a small difference in moisture content could
cause a change in mechanical properties. The E′ of CA or TA
modified gelatin films was constant from ∼60 to 160 °C, while
in unmodified gelatin film there was a significant drop in E′
between 100-160 °C due to glass transition (Tg) and melting
of the sample. For the CA or TA cross-linked gelatin materials,
their behavior is consistent with the formation of stable cross-
linked protein networks. Small differences between gelatin/CA
and gelatin/TA film performance are likely attributed to the
character of the different cross-link structures (cross-link density,
cross-link segment length, etc.) in the respective systems.
High resolution solid-state NMR also provides a powerful
technique to explore the intermolecular interactions, molecular
motions, and phase structures of protein materials, as demon-
strated in our previous publications.40–45 To assess the molecular
motions and intermolecular interactions between gelatin and
samples was shorter than that of unmodified gelatin possibly
due to a weak cross-linking effect (low cross-link density) and
slightly higher moisture content (plasticization effect) in the
gelatin/LCA systems. Modification of aggregated gelatin struc-
tures may also influence the H T1F values. However, the H
T1F values for samples B and C are higher than that of sample
A, which is consistent with an enhanced cross-linking effect in
B and C samples.
1
1
Conclusion
Cross-linking gelatin at a pH above 9 with natural phenolic
compounds CA and TA resulted in formation of insoluble
hydrogels. The mobility of these modified gelatin gels was lower
than the original gelatin gel. The cross-linking reactivity could
be controlled by varying pH, the concentration of the gelatin
solution or the amount of CA or TA used in the cross-linking
reaction. Oxidation mechanism played a key role in the cross-
linking reaction and oxygen acted as an efficient and green
oxidizer in the reaction. However, the oxygen bubbling might
not be controlled constantly, and some bubbles might be trapped
in the gel. The cross-linking chemistry and structures of the
modified gelatin gel and cast film samples were studied by both
solution and solid state high-resolution NMR technique to
investigate small molecular model systems or using 13C labeled
caffeic acid (LCA) in the reaction with gelatin. Direct evidence
was obtained confirming that chemical reactions occurred
between phenolic reactive sites of the phenolic compounds and
the amino groups in gelatin to form C-N covalent bonds as
the cross-linking linkages in gelatin matrix. The phenolic cross-
linker and the C-N cross-linking linkages were found to be
homogeneously distributed within the gelatin matrix and the
entire cross-linked network was homogeneous on a scale of 2-3
nm. This cross-linking resulted in the modified gelatin films
showing no Tg/melting behavior, thus, the films maintain high
modulus data at high temperatures.
1
LCA, H T1F data were measured through high-resolution 13C
resonances, as listed in Table 2. This parameter is sensitive to
the molecular motions of polymer chains in the regions of tens
of kHz. Only a single exponential T1F decay was obtained from
all 13C resonances in all samples. As seen in Figure 7, the
resonances derived from LCA appear at a range of 110-150
ppm and have no overlap with the signals of gelatin at 20-60
ppm. Thus, the 1H T1F data obtained at 110-150 ppm reflected
the behavior of LCA component, while those observed at 20-60
ppm reflect the gelatin behavior in gelatin/LCA samples. The
1H T1F data for gelatin obtained at different resonances were
essentially identical (4.7-4.9 ms), indicating strong spin-
diffusion interactions within the gelatin matrix and the material
is homogeneous at the scale of the spin-diffusion path length
within the T1F time, about 2-3 nm.34,35,46 In the gelatin/LCA
samples (A-C), the 1H T1F of pure LCA (∼33 ms) significantly
decreased and became similar to that of the gelatin component
(∼3.9-4.2 ms). This suggests that LCA is homogeneously
distributed in the gelatin matrix at the scale of 2-3 nm and
that the whole gelatin/LCA materials are homogeneous with
the domain sizes of phases (such as cross-linked regions,
plasticized phases, high molecular LCA species, etc.) smaller
than the scale of 2-3 nm. As only 1.5 wt % of LCA was used
References and Notes
(1) Veis, A. The Macromolecules Chemistry of Gelatin New York;
Academic Press: London, 1964.
(2) Hastings, G. W.; Ducheyne, P. Macromolecular Biomaterials; CRC
Press: Boca Raton, FL, 1984.
(3) Esposito, E.; Cortesi, R.; Nastruzzi, C. Gelatin microspheres: influence
of preparation parameters and thermal treatment on chemico-physical
and biopharmaceutical properties. Biomaterials 1995, 20, 2009–2020.
(4) Arvanitoyannis, I.; Psomiadou, E.; Nakayama, A.; Biba, S.; Yamamoto,
N. Edible films made from gelatin, soluble starch and polyols. Food
Chem. 1997, 60, 593–604.
(5) Sobral, P. J. A.; Menegalli, F. C.; Hubinger, M. D.; Roques, M. A.
Mechanical, water vapor barrier and thermal properties of gelatin based
ediable films. Food Hydrocolloids 2001, 15, 423–432.
(6) Yakimets, I.; Wellner, N.; Smith, A. C.; Wilson, R. H.; Farhat, I.;
Mitchell, J. Mechanical properties with respect to water content of
gelatin films in glassy state. Polymer 2005, 46, 12577–12585.
(7) Chiellini, E.; Cinelli, P.; Corti, A.; Kenawy, E. R. Composite films
based on waste gelatin: thermal-mechnical properties and biodegrada-
tion testing. Polym. Degrad. Stab. 2001, 73, 549–555.
1
in the system, its H T1F value was averaged out and became
similar to those of gelatin in the gelatin/LCA samples, with
gelatin as the majority component.
For gelatin/LCA samples at room temperature, significant
cross-linking would result in a rigid polymer network and cause
1H T1F data of the samples to increase. In contrast, plasticization
effect derived from either water or LCA would mobilize the
gelatin matrix and reduce the T1F value. It was noted that the
1H T1F value of the gelatin component in the gelatin/LCA

1132 Biomacromolecules, Vol. 11, No. 4, 2010
Zhang et al.
(8) Gerrard, J. A. Protein-protein crosslinking in food: methods, con-
sequences, applications. Trends Food Sci. Technol. 2002, 13, 389–
397.
(27) Cao, N.; Fu, Y.; He, J. Mechnical properties of gelatin films
crosslinked, respectively, by ferulic acid and tannin acid. Food
Hydrocolloids 2007, 21, 575–584.
(28) Spencer, M. C.; Cai, Y.; Martin, R.; Gaffney, S. H.; Goulding, P. N.;
Magnolato, D.; Lilley, T. H.; Haslam, E. Polyphenol complexation:
some thoughts and observations. Phytochemistry 1988, 27, 2397–2409.
(29) Siebert, K. J. Effects of protein-polyphenol interactions on beverage haze,
stabilization, and analysis. J. Agric. Food. Chem. 1999, 47, 353–362.
(30) Charlton, A. J.; Baxter, N. J.; Khan, M. L.; Moir, A. J. G.; Haslam,
E.; Davies, A. P.; Williamson, M. P. Polyphenol/peptide binding and
precipitation. J. Agric. Food Chem. 2002, 50, 1593–1601.
(31) Singh, H. Modification of food proteins by covalent crosslinking.
Trends Food Sci. Technol. 1991, 2, 196–200.
(9) Bigi, A.; Cojazzi, G.; Panzavolta, S.; Rubini, K.; Roveri, N. Mechanical
and thermal properties of gelatin films at different degrees of
glutaraldehyde crosslinking. Biomaterials 2001, 22, 763–768.
(10) Gerrard, J. A.; Brown, P. K.; Faye, S. E. Maillard crosslinking of
food proteins I: the reaction of glutaraldehyde, formaldehyde and
glyceraldehyde with ribonuclease. Food Chem. 2002, 79, 343–349.
(11) Gerrard, J. A.; Brown, P. K.; Faye, S. E. Maillard crosslinking of
food proteins II: the reactions of glutaraldehyde, formaldehyde and
glyceraldehyde with wheat proteins in vitro and in situ. Food Chem.
2003, 80, 35–43.
(32) Robbins, R. J. Phenolic acids in food: an overview of analytical
methodology. J. Agric. Food Chem. 2003, 51, 2866–2887.
(33) Bittner, S. When quinones meet amino acids: chemical, physical and
biological consequences. Amino Acids 2005, 30, 205–224.
(34) McBrierty, V.; Packer, K. Nuclear Magnetic Resonance in Solid
Polymers; Cambridge University Press: Cambridge, U.K., 1993.
(35) Alexlson, D. E.; Russell, K. Characterization of polymers by means
of 13C NMR spectroscopy. Prog. Polym. Sci. 1985, 11, 221–282.
(36) Mendieta-Taboada, O.; Sobral, P. J. d. A.; Carvalho, R. A.; Habitante,
A. M. Thermomechanical properties of biodegradable films based on
blends of gelatin and poly(vinyl alcohol). Food Hydrocolloids 2008,
22, 1485–1492.
(37) Pati, S.; Losito, I.; Palmisano, F.; Zambonin, P. G. Characterization
of caffeic acid enzymatic oxidation by-products by liquid chroma-
tography coupled to electrospray ionization tandem mass spectrometry.
J. Chromatogr., A 2006, 1102, 184–192.
(38) Amiot, M. J.; Tacchini, M.; Aubert, S.; Nicolas, J. Phenolic composi-
tion and browning susceptibility of various apple cultivars at maturity.
J. Food Sci. 1992, 57, 958–962.
(39) Tazaki, H.; Taguchi, D.; Hayashida, T.; Nabeta, K. Stable isotope
labelling on oxidative coupling of caffeic acid via o-quinone. Biosci.,
Biotechnol., Biochem. 2001, 65, 2613–2621.
(40) Zhang, X.; Hoobin, P.; Burgar, I.; Do, M. The pH effect on the
mechanical performance and phase mobility of thermally processed
wheat gluten-based natural polymer materials. Biomacromolecules
2006, 7, 3466–3473.
(41) Zhang, X.; Hoobin, P.; Burgar, I.; Do, M. Chemical modification of
wheat protein-based natural polymers: cross-linking effect on me-
chanical properties and phase structures. J. Agric. Food Chem. 2006,
54, 9858–9865.
(42) Zhang, X.; Do, M.; Bilyk, A. Chemical modification of wheat protein-
based natural polymers: formation of polymer networks with alkox-
ysilanes to modify molecular motions and enhance the material
performance. Biomacromolecules 2007, 8, 1881–1889.
(43) Zhang, X.; Do, M. Plasticization and crosslinking effects of acetone
-formaldehyde and tannin resins on wheat protein-based natural
polymers. Carbohydr. Res. 2009, 344, 1180–1189.
(44) Zhang, X.; Burgar, I.; Do, M.; Lourbakos, E. Intermolecular interac-
tions and phase structures of plasticized wheat proteins materials.
Biomacromolecules 2005, 6, 1661–1671.
(45) Zhang, X.; Do, M.; Hoobin, P.; Burgar, I. The phase composition and
molecular motions of plasticized wheat gluten-based biodegradable
polymer materials studied by solid-state NMR spectroscopy. Polymer
2006, 47, 5888.
(46) Zhang, X.; Takrgoshi, K.; Hikichi, K. Composition dependence of
the miscibility and phase structure of amorphous/crystalline polymer
blends as studied by high-resolution solid-state 13C NMR spectroscopy.
Macromolecules 1992, 25, 2336–2340.
(12) Martucci, J. F.; Ruseckaite, R. A.; Vazquez, A. Creep of glutaralde-
hyde-crosslinked gelatin films. Mater. Sci. Eng., A 2006, 435-436,
681–686.
(13) Olde, D.; Dijkstra, P. J.; van Luyn, M. J. A.; van Wachem, P. B.;
Nieuwenhuis, P.; Feijen, J. Glutaraldehyde as a crosslinking agent for
collagen-based biomaterilas. J. Mater. Sci.: Mater. Med. 1995, 6, 460–
472.
(14) Kuijpers, A. J.; Engbers, G. H. M.; De Smedt, S. C.; Meyvis, T. K. L.;
Demeester, J.; Krijgsveld, J.; Zaat, S. A. J.; Dankert, J. Characterization
of the netword structure of carbodiimide cross-linked gelatin gels.
Macromolecules 1999, 32, 3325–3333.
(15) Sung, H. W.; Huang, D. M.; Chang, W. H.; Huang, R. N.; Hsu, J. C.
Evalution of gelatin hydrogel crosslinked with various crosslinking
agents as bioadhesives: in vitro study. J. Biomed. Mater. Res. 1999,
46, 520–530.
(16) van Wachem, P. B.; Zeeman, R.; Dijkstra, P. J.; Feijen, J.; Hendricks,
M.; Cahalan, P. T.; can Luyn, M. J. A. Characterization and
biocompatibility of epoxy-crosslinked dermal sheep collagens.
J. Biomed. Mater. Res. 1999, 47, 270–277.
(17) Kolodziejska, I.; Piotrowska, B. The water vapour permeability,
mechanical properties and solubility of fish gelatin-chitosan films
modified with transglutaminase or 1-ethyl-3(3-dimethylaminopropyl)
carbodiimide (EDC) and plasticied with glycerol. Food Chem. 2007,
103, 295–300.
(18) Boanini, E.; Rubini, K.; Panzavolta, S.; Bigi, A. Chemico-physical
characterization of gelatin films modified with oxidized alginate. Acta
Biomater. 2010, 6, 383–388.
(19) Spanos, G.; Wrolsted, R. E. Phenolics of apple, pear, and white grape
juices and their changes with processing and storage. A review. J.
Agric. Food Chem. 1992, 40, 1478–1487.
(20) Matheis, G.; Whitaker, J. R. Modification of proteins by polyphenol
oxidase and peroxidase and their products. J. Food Biochem. 1984,
8, 137–162.
(21) Asquith, T. N.; Butler, L. G. Interactions of condensed tannins with
selected proteins. Phytochemistry 1986, 25, 1591–1593.
(22) Wu, J.; Chiu, S.-C.; Pearce, E. M.; Kwei, T. K. Effect of phenolic
compounds on gelatin behavior of gelatin gels. J. Polym. Sci., Part
A: Polym. Chem. 2001, 39, 224–231.
(23) Chatterjee, S.; Chowdhury, R.; Bhattacharya, P. A mathematical
analysis of the effect of pH and protein concentration on adsorption
of proteins on immobilised tannin. Biochem. Eng. J. 2000, 5, 77–82.
(24) Rawel, H. M.; Kroll, J.; Hohl, U. C. Model studies on reactions of
plant phenols with whey proteins. Nahrung/Food 2001, 2, 72–81.
(25) Strauss, G.; Gibson, S. M. Plant phenolics as cross-linkers of gelatin
gels and gelatin-based coacervates for use as food ingredients. Food
Hydrocolloids 2004, 18, 81–89.
(26) Xing, F.; Cheng, G.; Yi, K.; Ma, L. Nanoencapsulation of capsaicin
by complex coacervation of gelatin, acacia and tannins. J. Appl. Polym.
Sci. 2005, 96, 2225–2229.
BM1001284