Food Hydrocolloids 77 (2018) 636e645

Contents lists available at ScienceDirect

Food Hydrocolloids
journal homepage: www.elsevier.com/locate/foodhyd

Gelation behaviors of denaturated pea albumin and globulin fractions

during transglutaminase treatment

mi Saurel a, *
Attaf Djoullah a, b, Florence Husson a, Re
a
Univ. Bourgogne Franche-Comt
e, AgroSup Dijon, PAM UMR A 02.102, F-21000 Dijon, France
b
D

epartement de technologie alimentaire, Unit

e de Recherche et D

eveloppement de l’Intendance, 16000 Alger, Algeria

a r t i c l e i n f o a b s t r a c t

Article history: The behavior of pea albumins (Alb) and globulins (Glob) in a denatured state toward microbial trans-
Received 29 June 2017 glutaminase (MTGase) treatment was studied by SDS-PAGE analysis, free amine group determination,
Received in revised form dynamic rheology and confocal laser scanning microscopy. The denaturation of pea proteins by chemical
1 November 2017
reduction with dithiothreitol (DTT) or by thermal treatment at 80
C enhanced the enzymatic cross-
Accepted 3 November 2017
Available online 6 November 2017
linking degree of both protein isolates with greater crosslinking for the Glob fraction. The chemical
denaturation affected preferentially the participation in crosslinking reaction of legumin acid subunits
(40 kDa) for Glob sample and albumin polypeptides PA2 (26 kDa) for Alb sample, whereas the heat
Keywords:
Pea protein
treatment led to complete polymerization of 55, 35 and 30 kDa vicilin polypeptides for pea globulins as
Albumins shown by SDS-PAGE analysis. Up to 10 wt% concentration, the Alb fraction was not able to form MTGase
Globulins crosslinked gels whatever the initial native or denaturated state. Compared to the native state, chemical
Transglutaminase and thermal denaturation of Glob fraction before enzymatic treatment led to the formation of weaker
Gelation and stronger viscoelastic gels respectively. These contradictory results indicated that the enzymatic
Denaturation crosslinking reaction is highly related to polypeptides composition and conformation of proteins and the
use of denaturation as a strategy to enhance gel forming properties by transglutaminase treatment has to
be used with caution in the case of plant proteins.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction
Transglutaminase (TGase; E.C.2.3.2.13) is an enzyme that cata-
lyzes acyl-transfer reactions through the transfer of g-carboxamide
groups (acyl donor) of glutamine (Gln) residues to free amino
groups (acyl acceptor). In proteins, the ε-amino groups of lysine
(Lys) residues act as acyl acceptors and lead to the formation of
intra- and inter-molecular ε-(g-glutamyl)-lysine (G-L) cross-links
(Djoullah et al., 2015b). Transglutaminase of microbial origin
(MTGase) is the more used source in the food industry as it is
available in commercial preparations (Kieliszek & Misiewicz, 2014).
As a consequence of TGase-induced crosslinking, the functional and
textural properties of food proteins, such as thermal stability, water
retention, solubility, interfacial and gelling properties, can be
improved to open a wide range of applications (Gaspar & de Go
es-
Favoni, 2015). The generation of G-L isopeptide bonds has a pre-
dominant role in the gel formation where stable protein network is
* Corresponding author.
E-mail address: remi.saurel@u-bourgogne.fr (R. Saurel).
required. Highly elastic and irreversible gels can be obtained by the
use of Tgase with different protein substrates, even at relatively low
concentrations (Motoki & Kumazawa, 2000). The most traditional
applications concern reconstituted meat or fish foodstuffs, dairy
gels, bread or baked pastry products and soy-derived products such
as tofu, where the crosslinking of proteins makes it possible to
favorably influence the texture of the final products (Kieliszek &
Misiewicz, 2014).
The main condition for a protein to be a good substrate for TGase
is to have enough accessible Gln and Lys residues (De Jong &
Koppelman, 2002). This depends, on the one hand, on the pri-
mary structure of the protein which defines quantitatively the
composition of the protein in these two amino acids and on the
other hand on the conformation of the protein (ternary and qua-
ternary structures) which represents a limiting factor to the
accessibility of TGase to these two reactive groups. Another
important factor influencing this enzymatic reaction is the neigh-
borhood distance between the two residues (Jaros, Partschefeld,
Henle, & Rohm, 2006). Indeed, if the two residues belong to the
same protein molecule, the transglutaminase can form

https://doi.org/10.1016/j.foodhyd.2017.11.005
0268-005X/© 2017 Elsevier Ltd. All rights reserved.
 A. Djoullah et al. / Food Hydrocolloids 77 (2018) 636e645
intramolecular G-L bond if the Gln and the Lys are sufficiently close
to each other. In the case of intermolecular bonds, i.e. binding be-
tween two distinct proteins, steric hindrance or repulsive in-
teractions between proteins may limit the proximity of the two
residues thus preventing the enzymatic reaction.
Among protein sources, casein, myosin and soybean glycinin
were evaluated as excellent candidates for MTGase because of their
richness in Gln and Lys residues (De Jong & Koppelman, 2002; De
Jong, Wijngaards, Boumans, Koppelman, & Hessing, 2001; Jaros
et al., 2006). Their Gln and Lys residues, in contrast to those of b-
lactoglobulin or bovine serum albumin, appeared to be easily
accessible for MTGase and sufficiently close to each other for
enzymatic catalysis. It is possible to improve the susceptibility of
proteins to the enzymatic treatment. The principle is based on the
destabilization of the protein in order to unfold the structure thus
allowing a better exposure of the Gln and Lys residues to the TGase
to catalyze the reaction. Several strategies have been exploited in
the literature to improve crosslinking efficiency: e.g. pH variation,
use of reducing agents, enzymatic hydrolysis by protease or ther-
mal denaturation.
In the case of whey proteins, variation in pH may affect the
structure of the protein and make it partially unfolded and thus
more susceptible to enzymatic treatment (Eissa, Bisram, & Khan,
2004). The reducing electrophoretic profile of the whey proteins
at pH 8 treated with MTGase showed that the migration bands
were less intense than those treated at pH 6 and 7, in particular for
b-lactoglobulin; this accounted for the formation of a greater
number of high molecular weight aggregates by covalent bonds
initiated by MTGase.
In its native state, the folded structure of b-lactoglobulin is
stabilized by two disulfide bridges, which limits the crosslinking
between protein molecules by MTGase (Eissa & Khan, 2006). It has
thus been shown that in the presence of a reducing chemical agent
such as dithiothreitol (DTT) (10 mM) whose role is to break the
protein-stabilizing disulfide bridges, the enzymatic reaction was
markedly improved compared to the native sample (De Jong &
Koppelman, 2002). However, since DTT is not allowed in the food
industry, other reducing agents have been tested. Sodium sulphite
and cysteine improved enzymatic cross-linking of b-lactoglobulin
by MTGase, but to a lesser degree than DTT.
Another means used to improve the susceptibility of proteins to
enzymatic treatment is the use of chelating agents such as ethyl-
enediaminetetraacetic acid (EDTA). The latter was used to improve
the crosslinking of a-lactalbumin by eliminating the calcium (cal-
cium bridge) stabilizing the structure of this protein (Jaros et al.,
2006).
Partial hydrolysis of proteins by proteolytic enzymes caused an
increase in the degree of crosslinking as shown for soy proteins
(Walsh, Cleary, McCarthy, Murphy, & FitzGerald, 2003). The partial
cleavage of the amino acids following proteolysis makes it possible
to alter the compact structure of the proteins and thus to better
expose the Gln and Lys residues to the MTGase.
Finally, an alternative to the chemical modification of structured
proteins is thermal denaturation. Indeed, heat treatment de-
stabilizes the structure of b-lactoglobulin and thus makes the Gln
and Lys residues more accessible to MTGase (De Jong & Koppelman,
2002).
To our knowledge a few studies reported the improvement of
TGase-induced crosslinking by the application of chemical or
thermal pretreatments on plant proteins in relation to their gelling
properties. The gelation induced by the enzymatic treatment of soy
proteins is generally carried out on freshly extracted proteins
without major denaturation or alternatively on commercial isolates
whose denaturation state is not controlled (Dube, Scha €fer,
Neidhart, & Carle, 2007; Gaspar & de Go
es-Favoni, 2015). This is
637
also true for pea proteins which are interesting alternative candi-
dates to soybean proteins for such application Recent studies have
reported the effect of transglutaminase on pea protein gel forma-
tion from native and/or commercial pea protein isolates (Schafer,
Zacherl, Engel, Neidhart, & Carle, 2007; Shand, Ya, Pietrasik, &
Wanasundara, 2008; Sun & Arntfield, 2011). In these studies the
link between pea globulin denaturation and gel forming properties
upon transglutaminase treatment had never been established. So in
the present paper, we propose to elucidate this relation by taking
into account the effect of enzymatic crosslinking on constitutive
polypeptides of pea proteins after their chemical reduction or heat
denaturation.
Pea (Pisum sativum L.) proteins representing 20e30% of total dry
pea seeds, are mainly composed of 50e60% globulins and 15e25%
albumins (Boye, Zare, & Pletch, 2010). Pea globulins (Glob) are salt-
extractible proteins and are composed of the two major groups of
legumin-type (11S) and vicilin-type (7S) families. Legumin has
been described as a hexameric protein of between 360 and
400 kDa. Each monomer consists of a 40 kDa acidic subunit and a
20 kDa basic subunit linked by a disulfide bond (Croy, Gatehouse,
Evans, & Boulter, 1980a). Vicilin is a trimer of approximately
160e200 kDa. Vicilins consist of ~50 kDa primarily subunit and
different polypeptidic subunits of 30e35 and 13e20 kDa resulting
from post-transductional processing (Croy, Gatehouse, Evans, &
Boulter, 1980b). A third major storage protein, convicilin (7S), has
a subunit of ~71 kDa and a molecular weight of 210e290 kDa in its
trimeric native form (Croy, Gatehouse, Tyler, & Boulter, 1980). The
albumin fractions (Alb; 2S) are water-soluble proteins including
two major components richer in sulfur amino acids and lysine than
pea Glob (Croy, Hoque, Gatehouse, & Boulter, 1984). The most
abundant component is the albumin protein (PA2), a homodimer of
two polypeptides of approximately 25e26 kDa linked with non
specific interactions. The second component is the low molecular
weight albumin protein (PA1) containing two polypeptides of
approximately 4 and 6 kDa. Lipoxygenases, glycosidases, protease
inhibitors, and lectins present in less quantity also belong to the
albumin fraction (Be
rot, Le Goff, Foucault, & Quillien, 2007).
In a previous paper we evaluated the influence of the operating
conditions on the cross-linking yield by the MTGase of the Alb and
Glob fractions of pea protein in the native state (Djoullah,
Djemaoune, Husson, & Saurel, 2015a). Alb fractions had never
been considered in previous studies concerning enzymatic cross-
linking of pea protein. It was determined that a temperature of
40
C, pH 7, activity of 20 MTGase units/g protein, 20 mM NaCl ionic
strength and 2 h incubation time constituted parameters offering a
good compromise between high crosslinking level and stability of
enzymatic activity. Compared to the behavior of soy protein (Gan,
Cheng, & Easa, 2009), it was concluded that native albumin and
pea globulin are classified as poor and medium substrates respec-
tively for MTGase, despite their amino acid composition rich in
lysine and glutamine. It was concluded that the structure and
conformation of the proteins play a decisive role and their dena-
turation would allow a better accessibility of the enzyme to the
lysine and glutamine residues, thus improving the yield of
crosslinking.
In this new work, our objective was to understand the effect of
chemical (reducing treatment by DTT) and physical (heat treat-
ment) denaturation of the Alb and Glob fractions of pea proteins on
the enzymatic reaction by MTGase involving the different poly-
peptides constituting these fractions. This enables us to establish as
novel findings a link between enzymatic crosslinking effect at the
molecular level and gel forming properties of denaturated pea
proteins. For this reason, the resulting rheological and structural
properties of the formed protein networks were also investigated
by comparison to the native fractions previously studied.
638 A. Djoullah et al. / Food Hydrocolloids 77 (2018) 636e645
2. Material and methods
2.1. Materials
Pea flour from smooth yellow peas (P. sativum L.) was provided
by Roquette SA (Lestrem, France). A Ca2þ-independent microbial
transglutaminase (MTGase) derived from Streptoverticillium
mobaraense was provided by Ajinomoto Foods Europe SAS, France.
The enzyme preparation (Activia® WM) consists in 99% maltodex-
trin and 1% active enzyme with a MTGase activity of 100 units per
gram. All chemicals of analytical grade were obtained from VWR
International SAS, Fontenay-sous-Bois, France and SigmaeAldrich.
2.2. Methods
2.2.1. Protein extraction and purification
Pea albumin (Alb) and globulin (Glob) fractions were extracted
from pea flour at a laboratory scale with the same procedure
adapted from Cre
vieu, Be

rot, and Gueguen (1996) as described in
our previous paper (Djoullah et al., 2015a). Defatted and desalted
protein powders were obtained after freeze-drying. The low
denaturated Alb and Glob protein powders contained approxi-
mately 3.5% water, and 90 and 93% protein on a dry weight basis
respectively with a low content of lipid (<0.8%) (Djoullah et al.,
2015a). Pea Glob in this study was composed of 25% legumin, 65%
vicilins and 10% convicilin as previously determined by densitom-
etry after electrophoresis of the native Glob sample (Djoullah et al.,
2015a).
2.2.2. Thermal and chemical pre-denaturations of Alb and Glob
fractions
Glob and Alb solutions were prepared in phosphate buffer
(50 mM, pH 7) at the desired concentrations. After one hour stirring
at ambient temperature, the protein solutions were separated into
three distinct samples. The first sample was thermally denatured
(heating 80
C for 1 h with a heating and cooling ramp of 1
C/min).
The second part was chemically denatured by incubation with
20 mM DTT overnight at room temperature. The third sample was
the native control without pretreatment. All mixtures were pre-
pared at low ionic strength (20 mM NaCl) so as to avoid possible
overaggregation of proteins by charge screening.
2.2.3. MTGase treatment and crosslinking quantification
Native and chemically or thermally denatured protein solutions
from the Alb and Glob fractions were incubated with 20 units of
MTGase/g protein at 40
C and pH 7 at different incubation times.
At time intervals from 10 to 300 min, the reaction was stopped by
heating the solutions at 90
C for 5 min. In parallel for each
experiment, control samples were submitted to the same treat-
ments in the absence of MTGase. After cooling at room tempera-
ture, the samples were frozen, freeze-dryed and then analyzed by
SDS-PAGE under reducing conditions and by the O-phthaldialde-
hyde (OPA) spectrophotometric method as described in our pre-
vious paper (Djoullah et al., 2015a). The SDS-PAGE profiles were
analyzed in duplicate to determine qualitatively which poly-
peptides derived from the protein fractions provided substrates for
enzymatic crosslinking. The OPA method allowed determination of
the free amino group content of the samples over time and calcu-
lation of the degree of enzymatic crosslinking expressed in %.
2.2.4. Minimum gelling concentration and gel rheology
The determination of the minimum gelling concentration by
MTGase (MGCEz) was performed for the native and denatured pea
protein samples (Alb, Glob) by a test-tube-inversion method as
described in a previous work (Djoullah et al., 2015a). The protein
samples in the concentration range from 1 to 10 wt% were treated
in duplicate using 20 units of MTGase/g protein at pH 7 and an
incubation time of 10 h at 40
C. The lowest concentration at which
the sample was stable (not flowing) to tube inversion was consid-
ered as the minimum gelling concentration.
Changes in viscoelastic properties under isothermal conditions
(40
C) of the protein samples upon enzymatic treatment were
evaluated with a controlled-stress rheometer (SR-5 Rheometric
Scientific - Piscataway, NJ, USA), using parallel plate-plate geometry
(25 mm diameter) and a Peltier temperature unit. The protein
samples (1.5 ml) at 10 wt% concentration were introduced with an
amount of enzyme preparation corresponding to 20 units of
MTGase/g protein into Ependorf tubes (2 ml). Each mixture was
rapidly vortexed (1 min) to disperse MTGase, centrifuged
(1000  g, 1 min) to remove air bubbles, and then 1 ml was
deposited on the lower plate of the rheometer and then the upper
plate was lowered till the gap width was 1 mm. To prevent any
evaporation during experiment, mineral oil was deposited gently
on the edges of the sample to minimize water evaporation during
measurements. The measurements were carried out at imposed
oscillatory stress (0.5e2 Pa) at an oscillation frequency of 1 rad/s.
These parameters allowed measurements within the linear visco-
elastic region, as determined from preliminary stress-sweep tests
(0.06e12 Pa) on protein gels. The evolution of the storage (G0 ) and
loss (G00 ) moduli was recorded as a function of time (kinetics) and
frequency (mechanical spectrum) using the RSI Orchestrator soft-
ware Version (V.6.5.8).
2.2.5. Gel microstructure by confocal laser scanning microscopy
(CLSM)
Solutions at 10 wt% of native and denatured pea protein frac-
tions prepared as above described were non-covalently labeled
prior to enzymatic incubation. Twenty ml of an aqueous solution of
rhodamine B isothiocyanate (RITC, 1 mg/ml) was added to 2 ml
protein solution followed by stirring in dark at 4
C for at least 1 h.
RITC binds to free amine groups of protein through hydrophobic
interactions. An amount of enzyme corresponding to 20 units of
MTGase/g protein, was added and dispersed in the labeled protein
solutions. The mixture was centrifuged (1000  g, 1 min) to remove
air bubbles and incubated at 40
C for a 12 h period. The formed gels
were then unmolded and a fine cut (<1 mm) was deposited on a
slide for microscopic observation. The apparatus used to examine
the microstructure of the formed gels was a confocal laser scanning
microscope (Nikon, Eclipse TE-2000) in fluorescence mode. The
light source was a neon-argon laser beam and the objective lens
used had a magnification of 20. The emission and excitation
wavelengths were 543 nm and between 560 and 590 nm, respec-
tively. The acquisition of the images was done in “.JPEG” format
with a resolution of 512  512 pixels.
3. Results and discussion
3.1. Effect of denaturation on protein crosslinking by MTGase
Fig. 1 shows the SDS-PAGE profiles of the globulin samples un-
der reducing conditions: (a) native Glob fraction (Globnat), (b) de-
natured Glob fraction with DTT (Globdtt) and (c) thermally
denatured Glob fraction (Globther), treated with 20 units of
MTGase/g protein at 40
C and pH 7 at different incubation times.
All the electrophoretic patterns of the MTGase-treated samples
showed non migrating protein fractions (>200 kDa) at the top of
the gel with respect to their control (lanes T). This is due to the
presence of high molecular weight aggregates formed as a result of
intermolecular bonds catalyzed by MTGase. Some migrating bands
faded over time indicating their progressive participation in the
 A. Djoullah et al. / Food Hydrocolloids 77 (2018) 636e645 639

Fig. 1. SDS-PAGE profiles of Glob fraction samples at concentration of 1 wt% incubated with 20 units of MTGase/g protein at 40
C and pH 7 at different incubation times (lanes 10 to
300 in min): (a) native Glob fraction: Globnat, (b) DTT treated Glob fraction: Globdtt and (c) thermally denaturated Glob fraction: Globther. M: molecular weight markers (kDa); T:
control samples incubated 300 min without MTGase. CV: convicilin; V: polypeptides of vicilin; La: acid polypeptide of legumin acid; Lb: basic polypeptide of legumin.

network formed by enzymatic crosslinking. (a)

The SDS-PAGE profile of the Globnat sample (Fig. 1a) showed that
the bands most concerned by cross-linking corresponded mainly to
Glob nat Glob dƩ Glob ther
subunits of convicilin (71 kDa), acidic legumin (La, 40 kDa) and to a
100
lesser extent vicilins of 55, 50 and 35 kDa. When the disulphide
90
bridges in the globulin fraction have been reduced by DTT (Globdtt)
80
as pretreatment, the electrophoretic profile was close to that of
Crosslinking yield (%)

70
Globnat sample except for the La band (40 kDa) which disappeared
60
almost totally beyond 1 h of treatment (Fig. 1b). It may be hy-
50
pothesized that the loss of protein structure by the chemical
40
treatment gave MTGase more access to dissociated reactive sub-
30
units like La. On the other hand, the thermal denaturation signifi-
20
cantly improved the crosslinking of the globulin fraction (Globther),
10
which has essentially resulted in the almost total disappearance of
0
the 55, 35 and 30 kDa vicilin polypeptide bands within the first 0(b) 60 120 180 240 300
hour of incubation (Fig. 1c). Time (min)
Under these chemical and thermal denaturation conditions, the
enzymatic treatment appeared to have little or no effect on the
33 kDa vicilin and Lb legumin polypeptides, which was also (b)
observed for the native globulin sample. Such behavior had already
been evidenced for pea 11S legumin (Larre, Chiarello, Dudek,
Chenu, & Gueguen, 1993). Alb nat Alb dƩ Alb ther
In order to better account for the effect of chemical or thermal 100
denaturation on the enzymatic reaction, the crosslinking yields for 90
the Globdtt and Globther samples were calculated on the basis of the 80
Crosslinking yield (%)

free amines evolution during 5 h of enzymatic reaction, assayed by 70

the OPA method. These yields are shown in Fig. 2a. The crosslinking 60
kinetics for the Globnat sample already recorded in our previous 50
work was added on the graph for comparison (Djoullah et al., 40
2015a). 30
20
The degree of crosslinking increased strongly during the first
10
hour of incubation for the three samples and plateaued after 2 h of
0
treatment. The cross-linking percentage reached final values of 27,
0 60 120 180 240 300
37 and 74% for the Globnat, Globdtt and Globther samples respec- Time (min)
tively. Otherwise the crosslinking rate for the Globther sample was
significantly higher than that for the Globnat and Globdtt samples Fig. 2. Evolution of crosslinking yields of native, DTT or thermally treated Glob (a) and
(Fig. 2a). Indeed the percentage of crosslinking for the Globther Alb (b) samples incubated with 20 units of MTGase/g protein at 40
C and pH 7 during
5 h. The protein concentration was 1 wt% in all samples. Globnat and Albnat data were
sample reached 50% at 10 min treatment versus 7 and 11% for the
reproduced from Djoullah et al. (2015a,b).
Globnat and Globdtt samples respectively. This effect can be related
to the early participation of vicilin polypeptides (55, 35 and 30 kDa)
to the protein crosslinking as observed on SDS PAGE profiles when The effect of chemical and thermal denaturation on the evo-
thermal denaturation was applied (Fig. 1c). lution of the SDS-PAGE profile of the Alb fraction during the
640 A. Djoullah et al. / Food Hydrocolloids 77 (2018) 636e645

enzymatic crosslinking is shown in Fig. 3. The enzymatic treat-
ment applied to the albumin fraction at the native state (Albnat),
treated by DTT (Albdtt) and thermally denatured (Albther) resulted
in the formation of high molecular weight molecules that
remained trapped at the top of the gel with the appearance of
smearing, which intensified with the incubation time indicating
the development of the enzymatic reaction. For the three
considered Alb samples (Fig. 3), the traces of globulins contami-
nating the total albumin fraction followed the same behavior as in
the case of Glob fraction alone (Fig. 1) excepted for the La subunit
(40 kDa), which disappeared immediately after the first minutes
of enzymatic treatment. The major polypeptide of albumin (PA2),
representing about 70% total protein in Alb fraction (Djoullah
et al., 2015a), remained less sensitive to MTGase as none or
incomplete disappearance of the characteristic band at 26 kDa was
observed. Nevertheless chemical denaturation by DTT or thermal
denaturation seemed to increase the effect of MTGase on this
fraction.
The crosslinking yields corresponding to the three samples of
albumins are shown in Fig. 2b. The crosslinking degree values
plateaued at 12.5, 40 and 53% after 60 min of treatment for the
Albnat, Albdtt and Albther samples respectively. The denaturation of
the albumin fraction improved the crosslinking efficiency by about
3- and 4-folds with the DTT treatment and the heat treatment
respectively.
3.2. Effect of pretreatment on protein gelling properties
The determination of the minimum gelling concentration by
MTGase crosslinking (MGCEz) for the chemically or thermally de-
natured Glob and Alb fractions is shown in Fig. 4.
The albumin fraction was not able to form gel after thermal or
chemical denaturation in the applied concentration range and
under current enzymatic treatment conditions (Fig. 4b and d). This
was also observed in our previous work for the native protein
counterpart (Djoullah et al., 2015a). The gels formed observed in
Fig. 4d from 4% concentration resulted directly from the heat pre-
treatment, without the action of MTGase as no enzyme quantity
could be practically added in this case. Despite the increase in re-
action yield even after chemical denaturation with DTT, the for-
mation of a continuous gel was not observed for the Alb fraction. In
the current experimental conditions it could be assumed that the
formation of intramolecular G-L bonds would occur preferentially
at the expense of intermolecular bonds as reported elsewhere for
the major Alb fraction (Djoullah, Krechiche, Husson, & Saurel,
2016).

Fig. 3. SDS-PAGE profiles of Alb fraction samples at concentration of 1 wt% incubated with 20 units of MTGase/g protein at 40
C and pH 7 at different incubation times (lanes 10 to
300 in min): (a) native Alb fraction: Albnat, (b) DTT treated Alb fraction: Albdtt and (c) thermally denaturated Alb fraction: Albther. M: molecular weight markers (kDa); T: control
samples incubated 300 min without MTGase. LP: lipoxygenase; CV: convicilin; V: vicilin polypeptide; La: acid polypeptide of legumin; PA2: major polypeptide of albumin fraction.

Fig. 4. Minimum gelling concentration determination for Globdtt (a), Albdtt (b), Globther (c) and Albther (d) samples treated with 20 MTGase units/g protein at 40
C and pH 7.
 A. Djoullah et al. / Food Hydrocolloids 77 (2018) 636e645
The MGCEz values observed for Globdtt and Globther samples
were 8% and 3% respectively (Fig. 4a and c). A 6% MGCEz value has
been found in a previous work for the native globulin sample
prepared in the same experimental conditions (Djoullah et al.,
2015a). The decrease in the MGCEz value after thermal denatur-
ation of protein was consistent with the corresponding increase in
crosslinking efficiency assigned to the higher involvement of vicilin
polypeptides during the enzymatic reaction as mentioned above
(Figs. 1 and 2).
On the other hand, the increase in MGCEz after chemical dena-
turation is an unexpected result. Indeed, after denaturation of the
Glob fraction by DTT, the crosslinking efficiency was improved
compared to the native globulin fraction (Fig. 2a). This effect was
attributed to the dissociated 40 kDa acid legumin subunit which
was more consumed by the enzymatic reaction as observed in the
corresponding electrophoretic profiles (Fig. 1). Indeed it could be
expected that an increase in the degree of crosslinking would cause
a decrease in MGCEz. A characterization of the viscoelastic proper-
ties as well as the microstructure of the gels were necessary in
order to better understand and explain these contradictory results.
The gelation kinetics by enzymatic crosslinking of pea globulins
was monitored by low amplitude oscillatory rheology. In dynamic
rheology mode a sol-gel transition is signified by the crossover of
the G’ (storage) and G” (loss) moduli at the gel point (Gp) where
tan d ¼ G”/G’ ¼ 1, followed by a marked increase of G0 , indicative of
a reinforcement of the formed network. The material thus passes
from a liquid/viscous state with G” > G0 to a dominant elastic state
from Gp with G’ >> G’’. The system reaches an equilibrium when
the two moduli are practically stable, generally fluctuating within
±10% around an average value. This is recorded as the final elasticity
value of the gel.
Fig. 5 shows the evolution of the G’ and G” moduli during
enzymatic incubation (20 units of MTGase/g protein, pH 7, 40
C) of
Globnat, Globdtt and Globther samples at a 10 wt% concentration. It
Fig. 5. Evolution of storage (G0 ) and loss (G00 ) moduli with time for Globnat, Globdtt and Globther samples incubated in the presence of 20 units of MTGase/g protein at pH 7 and 40
C.
The protein concentration wass 10 wt% for all samples.
641
should be noted that the protein concentration was fixed above the
MGCEz value (8 wt%) of the DTT treated Glob sample and below the
critical thermal gelation concentrations for Globther sample deter-
mined at 12 wt% (data not shown).
All samples showed a gel formation profile, except for the
Globther control sample, incubated under the same conditions
without addition of MTGase, which showed practically constant
moduli throughout the analysis with G” > G’. This indicated that the
gelation of the studied samples was induced only by the trans-
glutaminase catalyzed enzymatic reaction. Thermal pre-
denaturation as well as incubation at 40
C did not participate
directly in the gelling process.
The gelation of Globnat with MTGase showed a gel point (Gp)
after 46 min of incubation (Fig. 5), followed by a rapid increase of G’
until a plateau was reached at a final average value of 1000 Pa after
6 h treatment.
The chemical or thermal denaturation of the Glob fraction,
considerably affected the gelation kinetics as well as the final gel
strength. Thermal denaturation (Globther) led to an acceleration of
the gelling kinetics with a stronger gel (3000 Pa) than that obtained
for Globnat (1000 Pa). No gel point could be observed in this case as
gelation was initiated very rapidly during sample preparation and
the storage modulus G0 was greater than the loss modulus G00 at the
beginning of rheological analysis. These results were in agreement
with the data obtained previously for electrophoresis, degree of
crosslinking as well as minimum gelling concentration indicating
higher gel forming capacity due to enhanced enzymatic protein
crosslinking.
Unlike thermal denaturation, the chemical change of protein
structure by DTT for the Glob fraction slowed down the gel kinetics
and reduced the final gel strength (400 Pa) with respects to the
other enzymatically treated samples. The Gp was observed at
68 min and the storage modulus value plateaued beyond 7 h of
enzymatic reaction, therefore later than for the Globnat and Globth
642 A. Djoullah et al. / Food Hydrocolloids 77 (2018) 636e645
G' Glob nat
10000 G'' Glob nat
1000
G', G'' (Pa)
100
10
0.01 0.1
Frequence (Hz)
Fig. 6. Mechanical spectra (G0 , G00 as a function of oscillatory frequency) of MTGase-induced gels from Globnat, Globdtt et Globther samples at the end of gelation at 40
C.
samples. This result reinforced the previous observation for CMGEz,
which increased by 6e8 wt % for the Globnat fraction and the Globdtt
fraction respectively (Fig. 4). This tendency to form gels with more
difficulty was contradictory with the increase in the reaction yield
(Fig. 2).
The mechanical spectra of the studied systems, i.e. variation
curves of G’ and G” as a function of the oscillatory frequency are
presented in Fig. 6. The experimental data were recorded at the end
of the enzymatic treatment (10 h) at 40
C.
Whatever the system, G0 was always greater than G00 and the
two moduli had little dependence to oscillatory frequency, with the
exception of the Globdtt sample where G0 and G00 were relatively
dependent at low frequencies. The Globther sample had the highest
G0 and G00 values, followed by the Globnat sample (Fig. 6). Similarly,
the tan d values of the Globnat and Globther samples were more
stable than that of the Globdtt sample (data not shown). The mean
values of tan d were 0.05, 0.12 and 0.03 for Globnat, Globdtt and
Globther samples respectively. In practice, tan d is a valuable indi-
cator of the boundary between the elastic and viscous domain. If
the value of tan d < 1 (tan d ¼ 1 for the gel point), the sample has a
more elastic than viscous behavior and the smaller tan d is (G’ >>
G00 ) the stronger the gel. According to Clark and Ross-Murphy
(1987), a “strong” or “true” gel is a gel having G0 and G” dynamic
moduli little dependent on the oscillatory frequency with
G’ > 10 G’’ (tan d < 0.1). Otherwise, the gel is considered as “weak”.
Based on the definition of Clark and Ross-Murphy (1987), the gels
for Globnat and Globther samples were considered as strong elastic
gels while a weak gel corresponded to the Globdtt sample. More-
over, as the viscoelastic parameters were independent from oscil-
latory frequency, the gels could be designated as chemical gels
characterized by prevalent covalent intermolecular bonds which is
consistent with the enzymatic crosslinking effect.
3.3. Gel microstructure by CLSM
In order to clarify differences in the gelation properties between
pea protein samples, the microstructure of gels formed by enzy-
matic treatment was observed by confocal laser scanning micro-
scopy (CLSM). The microscopic observations of the studied native
and denatured protein systems before and after enzymatic
G' Glob d G' Glob ther
G'' Glob d G'' Glob ther
1 10
treatment are shown in Figs. 7 and 8 for the Alb and Glob fractions
respectively. The enzymatic treatment conditions were 20 units of
MTGase/g protein at pH 7, 40
C and 10 wt% protein concentration
in presence of RITC.
Before enzymatic treatment, the protein solutions of Albnat and
Albdtt (controls) showed an uniformly distributed fluorescence in
the medium. No continuous protein network were formed at this
stage (Fig. 7). Thermal pre-denaturation, on the other hand, led to
the formation of a characteristic gel having a smooth structure with
the formation of a very thin and homogeneous network, which
prevented further enzymatic treatment. After the treatment with
MTGase, the Albnat fraction exhibited a heterogeneous distribution
of the fluorescence in the form of stranded protein particles or
aggregates. The non gelling property of the native albumin fraction
was also confirmed for the Albdtt sample which developed a more
granular discontinuous structure with larger sized aggregates than
for the Albnat sample. This result showed that even if no gel has
been formed for the two latter samples, intermolecular covalent
bonds resulting from enzymatic reaction in addition to supposedly
dominant intramolecular bonds could be responsible for the for-
mation of protein clusters.
The Globnat and Globdtt solutions (Fig. 8) had the same micro-
structure aspect as that of the Alb fraction in solution (Fig. 7). Ho-
mogenous protein suspensions could be observed prior to
enzymatic treatment. On the other hand, the micrograph for the
thermal denaturated Glob sample showed a dense aggregated state
and the incipient formation of a protein network.
Characteristic gel microstructure was observed for all the pea
globulin samples after MTGase treatment. The enzymatic reaction
of the Globnat fraction resulted in the formation of protein network
with a coarse structure, characterized by dark areas of heteroge-
neous sizes. The structure of the Globdtt sample after enzymatic
crosslinking revealed compact small protein fragments dispersed in
a relatively homogeneous continuous phase with reduced contrast.
The Globther sample after enzymatic treatment presented the most
gel homogeneous structure. The dark areas were reduced and the
connectivity between the particles was improved. This observation
was in agreement with the rheological data showing the
improvement of the gelling properties in the case of a prior heat
treatment of the globulin fractions.
 A. Djoullah et al. / Food Hydrocolloids 77 (2018) 636e645 643

Fig. 7. CLSM micrographs of Albnat, Albdtt et Albther samples without (control) and with MTGase treatment with 20 units of MTGase/g protein at pH 7 and 40
C (scale bar ¼ 100 mm).
The protein concentration was 10 wt% for all samples.

Fig. 8. CLSM micrographs of Globnat, Globdtt et Globther samples without (control) and with MTGase treatment with 20 units of MTGase/g protein at pH 7 and 40
C (scale
bar ¼ 100 mm). The protein concentration was 10 wt% for all samples.

3.4. Discussion stabilize some gobular proteins. The chemical denaturation of al-
bumin, a protein rich in sulfur-containing amino acids (Croy et al.,
DTT is a reducing agent able to break the disulfide bridges that 1984) by DTT, would lead to unfolding of the native structure and
644 A. Djoullah et al. / Food Hydrocolloids 77 (2018) 636e645
to better accessibility of the MTGase to the glutamine and lysine
residues. The crosslinking efficiency is thus improved (~2-folds
relatively to the native counterpart) without substantially
affecting the electrophoretic profile of the majority polypeptide
PA2 (26 kDa) of the Alb fraction (Fig. 3). This revealed that a large
part of the bonds formed between the glutamine and lysine resi-
dues are of intramolecular type, i.e. belonging to the same mole-
cule. The feature of the PA2 polypeptide (26 kDa) to form
intramolecular bonds has already been described in a previous
work of our group (Djoullah et al., 2016). Furthermore, analysis of
the CLSM microstructure of the albumin fraction treated with the
reducing agent DTT (Fig. 7) displayed the formation of small ag-
gregates through the possible contribution of intermolecular
bonds, which would remain minor compared to intramolecular
bonds as no gel has been formed. The hypothesis of the formation
of intramolecular G-L bonds, probably ascribed to the vicinity of
reactive groups within the albumin monomers, was also confirmed
in the case of the pre-heated albumin fraction. The reaction yield
was enhanced, but the initial thermal gelation of the Alb samples at
low protein concentration (>3%) prevented any evaluation of
MTGase action.
In the case of pea globulins, only legumin molecules can be
affected by the DTT treatment, as vicilins are sulfur-free proteins
(Croy et al., 1980b). Denaturation of the legumins with DTT led to
the separation of the acid (La, 40 kDa) and basic (Lb, 20 kDa)
polypeptides from the constitutive protein monomers modifying
substantially the hexameric native structure of this molecule. In
this new conformation, the glutamine and lysine residues of acid La
subunit would be better exposed to MTGase as revealed by the
disappearance of the band of this polypeptide in the electropho-
retic profile since it was consumed by the enzymatic reaction,
conversely to the basic Lb subunit. Thus, the overall reaction yield
for the DTT-treated Glob fraction improved comparatively to that
for the native Glob fraction (Fig. 2). Nevertheless, native legumins
participated in the enzymatic reaction through 360 kDa hexamers,
whereas the chemically denatured legumins participated in the
reaction though 40 kDa polypeptides. The destabilization of the
oligomeric structure of legumin by DTT would lead to the formation
of smaller cross-linked aggregates forming a thinner structure
network than for the native protein (Fig. 8). At the gel state the
protein network would be less continuous and would have fewer
junction points. The non-participation of the Lb polypeptide in
crosslinking reaction for the Globdtt fraction delayed the gelation
kinetics and weakened the formed network (Figs. 1 and 5) by
comparison to the Globnat fraction. This would explain the con-
tradictory results between increased reaction yield on the one hand
and higher MGCEz value and reduced final gel strength on the other
hand for the Globdtt sample.
The effect of chemical denaturation by DTT on enzymatic
crosslinking by MTGase has already been studied on b-lactoglob-
ulin by Eissa and Khan (2006). These authors observed an
improvement in the rheological properties of pre-denatured pro-
tein compared to its native state, which was not the case with the
Globdtt fraction in the present work. This difference is related to the
structure of the two proteins. Indeed, b-lactoglobulin is a globular
protein whose structure is stabilized by two disulfide bridges
(Nicolai, Britten, & Schmitt, 2011). The destruction of these covalent
bonds by DTT unfolds the protein without modifying its monomeric
primary structure (18 kDa), allowing better accessibility of MTGase
to the glutamine and lysine residues, resulting in enhanced enzy-
matic crosslinking. On the other hand, the chemical dissociation of
pea globulins into isolated polypeptides despite the improvement
in crosslinking efficiency modified the structure of the resulting
gels and affected unfavorably their rheological properties as
already explained.
Thermal denaturation of the Glob fraction led to an aggregated
state due to the predominantly noncovalent interactions generated
by the exposition of the hydrophobic groups following the
unfolding of pea proteins by heat denaturation (Mession, Sok,
Assifaoui, & Saurel, 2013). At this stage, the heat treatment did
not cause gelation even at high protein concentrations (<12%). This
can be explained by the repulsive forces existing between thermal
aggregates at the working pH (pH 7) away from the isoelectric point
of globulins (pH 4e6) and at low salt concentration in solution
(Djoullah et al., 2015a).
The enzymatic treatment of the thermally denaturated Glob
fraction led to an acceleration of the gelation kinetics with a final
gel strength 3-folds higher than that obtained for its native coun-
terpart. This improvement in gelling properties was consistent with
the concomitant increase in crosslinking efficiency and was
attributed mainly to vicilin polypeptides (30, 35 and 50 kDa) which,
after thermal denaturation, effectively participated in the formed
protein network (Fig. 1). The unfolding of globular proteins by
thermal effect allowed better exposition of glutamine and lysine
residues as reported elsewhere for whey protein (Eissa & Khan,
2006; Gauche, Barreto, & Bordignon-Luiz, 2010). Thus better con-
nectivity between aggregates via G-L bonds resulted in a compact
and homogeneous gel structure for pea globulins (Fig. 8). At the
same time, the La (40 kDa) polypeptides appeared to be less
affected by enzymatic treatment than vicilin polypeptides whereas
the Lb (20 kDa) polypeptides were still not involved in the enzy-
matic reaction (Fig. 1). The reassociation of the legumin acid sub-
units via disulfide bridge exchanges during the formation of
thermal aggregates of globulins (Mession et al., 2013) could restrict
the accessibility of the enzyme to the glutamine and lysine residues
of these polypeptides, despite the loss of legumin structure. How-
ever, the overall effect of thermal pre-denaturation on MTGase
action remained drived by vicilin behavior, which accounted for
about two-thirds of the protein content in the present isolated Glob
fraction.
4. Conclusion
The behavior of pea protein isolated fractions upon trans-
glutaminase treatment was investigated after chemical reduction
or heat treatment.
This study showed that albumin fraction is not a good candidate
for gelation by enzymatic treatment even in the denatured state
which was never demonstrated before this study. Although the
reaction yield increased due to improved accessibility to the lysine
and glutamine groups of the protein, formation of intramolecular
bonds appeared to dominate the behavior of the albumin fraction.
Even at high protein concentrations, this effect led to the genera-
tion of isolated protein aggregates rather than to gelation. The
enzymatic treatment of the native globulin fraction resulted in the
formation of a strong and elastic chemical gel which improved
considerably with thermal pre-denaturation. As new findings, this
was attributed to an enhanced participation of the majority vicilin
subunits in the enzymatic reaction. On the other hand, the loss of
the oligomeric structure of legumin after chemical denaturation,
despite the improvement of the reaction efficiency, substantially
weakened the gel and increased the minimum gelling concentra-
tion, which was a contradictory effect not commonly observed for
other proteins such as whey proteins.
The composition of the protein isolate, but also the structure
and conformation of the protein, play important roles in the
enzymatic reaction with transglutaminase. Regarding our results,
protein denaturation is not always an effective way to improve the
gelling pathway of plant proteins. The behavior of a protein mixture
containing the two pea protein fractions as substrates for MTGase
 A. Djoullah et al. / Food Hydrocolloids 77 (2018) 636e645
would be interesting to investigate in order to evaluate possible
synergistic effects.
Acknowledgement
This work was supported by the Regional Council of Bourgogne -
Franche Comte
(PARI ALIMþ program) and the European Union
through the PO FEDER-FSE Bourgogne 2014-2020 programs. We are
grateful to Ajinomoto Foods Europe SAS for providing the MTGase
powder.
References

rot, S., Le Goff, E., Foucault, A., & Quillien, L. (2007). Centrifugal partition chro-
Be
matography as a tool for preparative purification of pea albumin with enhanced
yields. Journal of Chromatography B, 845, 205e209.
Boye, J., Zare, F., & Pletch, A. (2010). Pulse proteins: Processing, characterization,
functional properties and applications in food and feed. Food Research Inter-
national, 43, 414e431.
Clark, A., & Ross-Murphy, S. (1987). Structural and mechanical properties of
biopolymer gels. Biopolymers, 83(1), 57e192.

vieu, I., Be
Cre
rot, S., & Gueguen, J. (1996). Large scale procedure for fractionation of
albumins and globulins from pea seeds. Food/nahrung, 40(5), 237e244.
Croy, R. D., Gatehouse, J. A., Evans, I. M., & Boulter, D. (1980a). Characterisation of
the storage protein subunits synthesised in vitro by polyribosomes and RNA
from developing pea (Pisum sativum L.). I. Legumin. Planta, 148, 49e56.
Croy, R. D., Gatehouse, J. A., Evans, I. M., & Boulter, D. (1980b). Characterisation of
the storage protein subunits synthesised in vitro by polyribosomes and RNA
from developing pea (Pisum sativum L.). II. Vicilin. Planta, 148, 57e63.
Croy, R. R., Gatehouse, J. A., Tyler, M., & Boulter, D. (1980). The purification and
characterization of a third storage protein (convicilin) from the seeds of pea
(Pisum sativum L.). Biochemical Journal, 191, 509e516.
Croy, R. R., Hoque, M. S., Gatehouse, J. A., & Boulter, D. (1984). The major albumin
proteins from pea (Pisum sativum L). Purification and some properties.
Biochemical Journal, 218, 795e803.
De Jong, G. A. H., & Koppelman, S. J. (2002). Transglutaminase catalyzed reactions:
Impact on food applications. Journal of Food Science, 67, 2798e2806.
De Jong, G. A. H., Wijngaards, G., Boumans, H., Koppelman, S. J., & Hessing, M.
(2001). Purification and substrate specificity of transglutaminases from blood
and Streptoverticillium mobaraense. Journal of Agricultural and Food Chemistry,
49(7), 3389e3393.
Djoullah, A., Djemaoune, Y., Husson, F., & Saurel, R. (2015a). Native-state pea al-
bumin and globulin behavior upon transglutaminase treatment. Process
Biochemistry, 50(8), 1284e1292.
Djoullah, A., Krechiche, G., Husson, F., & Saurel, R. (2016). Size measuring techniques
as tool to monitor pea proteins intramolecular crosslinking by transglutaminase
treatment. Food Chemistry, 190, 197e200.
Djoullah, A., Sok, N., Djemaoune, Y., Penouilh, M. J., Husson, F., & Saurel, R. (2015b).
Monitoring of transglutaminase crosslinking reaction by 1H NMR spectroscopy
645
on model substrates. Colloids and Surfaces a: Physicochemical and Engineering
Aspects, 475, 69e74.
€fer, C., Neidhart, S., & Carle, R. (2007). Texturisation and modification
Dube, M., Scha
of vegetable proteins for food applications using microbial transglutaminase.
European Food Research and Technology, 225, 287e299.
Eissa, A. S., Bisram, S., & Khan, S. A. (2004). Polymerization and gelation of whey
protein isolates at low pH using transglutaminase enzyme. Journal of Agricul-
tural and Food Chemistry, 52(14), 4456e4464.
Eissa, A. S., & Khan, S. A. (2006). Modulation of hydrophobic interactions in dena-
tured whey proteins by transglutaminase enzyme. Food Hydrocolloids, 20(4),
543e547.
Gan, C. Y., Cheng, L. H., & Easa, A. M. (2009). Assessment of crosslinking in combined
crosslinked soy protein isolate gels by microbial transglutaminase and maillard
reaction. Journal of Food Science, 74(2), 141e146.
Gaspar, A. L. C., & de Go
es-Favoni, S. P. (2015). Action of microbial transglutaminase
(MTGase) in the modification of food proteins: A review. Food Chemistry, 171,
315e322.
Gauche, C., Barreto, P. L. M., & Bordignon-Luiz, M. T. (2010). Effect of thermal
treatment on whey protein polymerization by transglutaminase: Implications
for functionality in processed dairy foods. LWT-food Science and Technology,
43(2), 214e219.
Jaros, D., Partschefeld, C., Henle, T., & Rohm, H. (2006). Transglutaminase in dairy
products: Chemistry, physics, applications. Journal of Texture Studies, 37(2),
113e155.
Kieliszek, M., & Misiewicz, A. (2014). Microbial transglutaminase and its application
in the food industry. A review. Folia Microbiology, 59, 241e250.
Larre, C., Chiarello, M., Dudek, S., Chenu, M., & Gueguen, J. (1993). Action of
transglutaminase on the constitutive polypeptides of pea legumin. Journal of
Agricultural and Food Chemistry, 41, 1816e1820.
Mession, J. L., Sok, N., Assifaoui, A., & Saurel, R. (2013). Thermal denaturation of pea
globulins (Pisum sativum L.) e molecular interactions leading to heat-induced
protein aggregation. Journal of Agricultural and Food Chemistry, 61(6),
1196e1204.
Motoki, M., & Kumazawa, Y. (2000). Recent research trends in transglutaminase
technology for food processing. Food Science and Technology Research, 6(3),
151e160.
Nicolai, T., Britten, M., & Schmitt, C. (2011). B-Lactoglobulin and WPI aggregates:
Formation, structure and applications. Food Hydrocolloids, 25(8), 1945e1962.
Schafer, C., Zacherl, C., Engel, K. H., Neidhart, S., & Carle, R. (2007). Comparative
study of gelation and cross-link formation during enzymatic texturisation of
leguminous proteins. Innovative Food Science & Emerging Technologies, 8,
269e278.
Shand, P. J., Ya, H., Pietrasik, Z., & Wanasundara, P. K. J. P. D. (2008). Trans-
glutaminase treatment of pea proteins: Effect on physicochemical and rheo-
logical properties of heat-induced protein gels. Food Chemistry, 107, 692e699.
Sun, X. D., & Arntfield, S. D. (2011). Gelation properties of salt-extracted pea protein
isolate catalyzed by microbial transglutaminase cross-linking. Food Hydrocol-
loids, 25, 25e31.
Walsh, D. J., Cleary, D., McCarthy, E., Murphy, S., & FitzGerald, R. J. (2003). Modifi-
cation of the nitrogen solubility properties of soy protein isolate following
proteolysis and transglutaminase cross-linking. Food Research International,
36(7), 677e683.