Journal of Food Engineering xxx (2017) 1e10

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier.com/locate/jfoodeng

Modication of extruded chicken collagen lms by addition of

co-gelling protein and sodium chloride
Anja Maria Oechsle a, b, Tanita Julia Bugbee a, Monika Gibis a, Reinhard Kohlus c,
Jochen Weiss a, *
a
Department of Food Physics and Meat Science, Institute of Food Science and Biotechnology, University of Hohenheim, Garbenstrasse 21/25, 70599
Stuttgart, Germany
b
Department Institute of Food Technology and Food Chemistry, Department of Food Technology and Food Material Science, Technische Universitaet Berlin,
Koenigin-Luise-Strasse 22, 14195 Berlin, Germany
c
Department of Food Process Engineering and Food Powders, Institute of Food Science and Biotechnology, University of Hohenheim, Garbenstrasse 21/25,
70599 Stuttgart, Germany

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

Article history: The properties of extruded chicken collagen sausage casings can be tailored by the addition of sodium
Received 14 July 2016 chloride (NaCl) or proteins prior to extrusion in order to gain lms with modied mechanical properties.
Received in revised form In this study, 4% (w/w) chicken skin collagen and telopeptide-poor collagen from bovine hide were
16 March 2017
modied by the addition of 0.05 mol/kg NaCl and/or partial substitution of collagen by soy protein
Accepted 20 March 2017
Available online xxx
isolate, at 1.25% (w/w). The collagen formulations were extruded to lms and characterized in terms of
their microstructure (light microscopy and scanning electron microscopy) and rheological properties
(storage and loss modulus and phase angle). Moreover, the tensile strength and lm thickness were
Keywords:
Chicken
examined. The addition of NaCl to the gel allowed for telopeptide-poor and chicken collagen lms with
Collagen high tensile strengths and elasticities to be formed. In contrast, a substitution of collagens with soy
Co-extrusion proteins decreased gel and lm strengths. The soy protein induced weakening of collagen networks
Soy protein isolate could be compensated by again adding NaCl leading to more homogeneous gels yielding lms with
Extruded lms higher storage moduli upon extrusion. The compensating effect of NaCl was more pronounced for
chicken skin than for telopeptide-poor collagen in the lm state suggesting differences in molecular
interactions and network formation between the two different collagen types. Overall, the modulation of
chicken collagen interactions by NaCl and soy protein addition enables the production of functional
chicken collagen lms, in turn providing the food and pharmaceutical industry with a viable alternative
to the increasingly scarce beef collagen.
2017 Elsevier Ltd. All rights reserved.

1. Introduction industry, the utilization of collagen to manufacture sausage casings

Collagen is a structural protein and major component of con-
nective tissue in vertebrates (Oechsle et al., 2014). Facilitating the
formation of a brillar network of the extracellular matrix, it fea-
tures unique properties such as high tensile strength, exibility,
biocompatibility and biodegradability. This makes the biopolymer
very useful in many applications such as in the elds of tissue en-
gineering, cosmetics and the food industry (Caliari and Harley,
2011; Campbell et al., 2009; Oechsle et al., 2014). In the meat
* Corresponding author.
E-mail address: j.weiss@uni-hohenheim.de (J. Weiss).
has gained in particular interest since demand and cost of natural
casings have risen (Oechsle et al., 2014). Currently, bovine hides are
the predominant source for collagen extraction, and there are
increasingly shortages thereof. However, pork skin, poultry skin
and sh skin have also been reported to be potential sources (Bker
et al., 2010; Cliche et al., 2003; Eckmayer et al., 2002; Liu et al.,
2015). Since diseases like bovine spongiform encephalopathy and
concerns over carbon footprints of meat have increased demand for
poultry products, an increasing amount of byproducts, such as skin,
bones and cartilages are available. The use of these raw materials
manufacture for collagen production could therefore be interesting
both from a consumer acceptance and an economical point of view
(Cliche et al., 2003). However, to this date, collagen products

http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
0260-8774/ 2017 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
2 A.M. Oechsle et al. / Journal of Food Engineering xxx (2017) 1e10
produced from chicken have not had the necessary tech-
nofunctionalities to manufacture sausage casing due to insufcient
tensile strengths and elasticities. Generally, boiled sausage casings
need to be able to withstand a certain pressure exerted by the raw
sausage masses during lling, and they need to be able to expand
and shrink upon heating and cooling.
Collagen extracted from homogenized and heated chicken skin
contains about 75% of type I collagen (Cliche et al., 2003). Collagen
type I is a heterotrimer consisting of two identical a1 chains and
one a2 chain (Zeugolis and Raghunath, 2011). Based on the
repeating Gly-X-Y amino acid sequence the three helical chains
form triple helical monomers with non-helical C- and N-terminal
extensions, called telopeptides, are responsible for collagen-
collagen crosslinking (Caliari and Harley, 2011; Oechsle et al.,
2014). Further, the monomers spontaneously assemble into brils
and bers (Prockop, 2013). It has been reported that the orientation
of the brils within a sausage casing is crucial to the strength of the
lm (Hoogenkamp et al., 2015). A homogenous distribution with
random orientation of brils could overcome possible weak spots
along the orientation direction. This often occurs when brils are
arranged in their natural, parallel quarter staggered conformation
(Lieberman, 1964).
The aim of this study was to assess the effect of two different
approaches to modify extruded collagen lms, i.e. by addition of
low concentrations of NaCl and by partially substituting collagen
with soy protein isolate. Based on previous studies, we hypothe-
sized that low amounts of NaCl may promote the formation of
larger collagen assemblies that could, in effect, lead to strengthened
collagen lms having increased elasticities (Oechsle et al., 2015b).
Furthermore, it was postulated that a partial substitution of
collagen by soy protein isolate while possibly slightly weakening
collagen gel strengths might still result in formation of stable lms
due to protein-protein interactions. There, addition of NaCl may
allow for a strength compensation. In our studies, a telopeptide-
poor collagen e established in previous studies - served as a
model system for less crosslinked samples (Oechsle et al., 2016a;
Oechsle et al., 2015a; Oechsle et al., 2016b; Oechsle et al., 2015b;
Oechsle et al., 2014). In addition, a skin collagen extract was used
to investigate animal specic alterations in collagen functionality.
2. Materials and methods
2.1. Preparation of collagen gels
Telopeptide-poor collagen was kindly provided by Kalle GmbH
(Wiesbaden, Germany). It was produced by cleaving off telopep-
tides severing intermolecular cross-links from native bovine
collagen to obtain single triple helices (Oechsle et al., 2015b). A
chicken skin collagen extract was generously supplied by Protein
Consulting (Singhofen, Germany), with manufacture procedures
having recently been published in form of a patent (Bker et al.,
2010). Soy protein isolate PRO FAM 974 was obtained from ADM
(Chicago, IL, USA). Commercially available NaCl (Bad Reichenhaller,
Bad Reichenhall, Germany) with a purity of >99% was used. Gels
with a total protein content of 4% (w/w) containing either no salt or
0.05 mol/kg of NaCl were produced. For some samples, 1.25% (w/w)
of the collagen was replaced by soy protein isolate. In addition,
samples containing 4% collagen and 1.25% soy protein were pre-
pared for comparison purposes.
All formulations were prepared in an Artisan 5KSM150 stand
mixer equipped with a ex edge beater 5KFE5T (KitchenAid, St.
Joseph, Minnesota, USA). Mixing of all components was done for a
total of 20 min at room temperature at a low to medium speed
setting. NaCl was added as a concentrated solution with water
content in the mixture being adjusted accordingly. Soy protein
Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
isolate was added after the rst 5 min of mixing. The pH was
adjusted to 2.7 in 1 min intervals, using 20% phosphoric acid during
the last 10 min of mixing. Gels were stored airtight over night at
1
C before homogenizing for 5 min at a low speed and evacuating
in a chamber machine (C400, Mulitvac, Wolfertschwenden, Ger-
many) to remove entrapped air prior to use.
2.2. Preparation of collagen lms
Collagen gels were extruded onto stainless steel rods (24 mm
diameter, 1 m length) acting as a receiving template to facilitate the
production of closed casings. Non-thermoplastic extrusion was
carried out on a QX Fresh System (Marel Townsend Further Pro-
cessing B.V., Boxmeer, Netherlands) with access kindly provided by
Kalle GmbH. The settings of the counter rotating extrusion cones
were set to their lowest speed as preliminary studies had shown
that otherwise lm rupture due to shear thinning would occur
(data not shown). The ring nozzle measured gap width of 25 mm.
Steel rods supporting extruded collagen lms were placed into a
saturated NaCl solution (5
C) immediately after extrusion to
facilitate precipitation and xation. Films were then carefully cut,
removed from the rod, and stored airtight at 1
C with sponge
cloths having been soaked in saturated NaCl solution acting as
separators (Kalle GmbH, Wiesbaden, Germany).
2.3. Imaging and microscopy
Photographic images of gels and lms were taken with a digital
single-lens reex camera (Type K-5, Pentax Ricoh Imaging,
Hamburg, Germany). Contrast and brightness of all collagen gel
images were enhanced using ImageJ (1.48v, Wayne Rasband, Na-
tional Institutes of Health, USA).
Light microscopic images were taken using a Scope.A1 Axio
transmission microscope combined with an AxioCam ICc3 camera
(Zeiss, Oberkochen, Germany). Images were taken with a total
magnication of 100 (10 objective lens, 10 ocular lens).
Scanning electron microscopy was conducted with a DSM 940
(Zeiss, Hamburg, Germany) at a high vacuum and with an accel-
erating voltage of 5000 kV. Pieces of the collagen lms,
0.5 mm  0.5 mm, were placed on a microscope slide and freeze-
dried in liquid nitrogen. The slides were sputtered with a gold/
palladium alloy (20:80) prior to microscopy.
2.4. Rheological measurements
Experiments were performed in a modular compact rheometer,
Physica MCR 502 (Anton Paar, Ostldern, Germany), at 5
C using a
plate-eplate geometry PP25/S with 24.98 mm diameter and a
sandblasted surface to avoid wall-slip (Anton Paar, Ostldern,
Germany). The phase angle d, the storage G (Pa) and loss modulus
G (Pa) of collagen gels and lms were determined via amplitude
sweeps (1 Hz, 0.1e100% strain) and frequency sweeps in the linear
viscoelastic region (1% strain for gels and 0.2% strain for lms,
0.1e100 Hz). Gels were measured at a plate-plate measurement
system gap of 1 mm, while gaps for lm samples were adjusted
individually to accommodate for their respective thickness. There,
the upper plate was automatically stopped when a force of 0.1 N
was registered. Films were cut into 25 mm diameter circles prior to
measurements.
2.5. Film thickness and tensile strength
Film thickness was determined during rheological measure-
ments as described in a previous study (Oechsle et al., 2016b). The
plate-plate geometry was lowered until a force of 0.1 N was
 A.M. Oechsle et al. / Journal of Food Engineering xxx (2017) 1e10 3

reached. This method was veried by measurements taken with a
digital micrometer (Andoer, Las Vegas, NV, USA). The lm thickness
was determined twice using triplicates.
For tensile strength investigation, the lms were cut into 50 mm
long and 10 mm wide strips. Collagen lm strips were then clam-
ped perpendicular into a texture analyzer 3365 equipped with a
100 N load cell, advanced screw side-action grips (500 N) and
50 mm  25 mm serrated jaw faces (Instron, Darmstadt, Germany).
Pre-stretching was conducted at 10 mm/min until 0.01 N were
registered, before proceeding with tensile strength measurement
at 50 mm/min. Tensile strength at maximum force was determined
by the software using the previously determined mean lm thick-
nesses for each sample.
2.6. Statistical analysis
All measurements were at least done three times using triplicate
samples. Means and standard deviations were calculated from
these measurements in Excel (Microsoft, Redmond, VA, USA). Sta-
tistical analysis was performed using One Way ANOVA and Tukey
Test embedded in SigmaPlot12.5 (Systat Software GmbH, Erkrath,
Germany). For statistically signicant differences P < 0.05 was
chosen.
3. Results and discussion
3.1. Collagen appearance and microstructure
Pictures of collagen gels and extruded collagen lms are shown
in Fig. 1. The telopeptide-poor collagen gels were transparent to
slightly opaque while chicken skin collagen gels were complete
opaque giving them a whitish appearance. In general, one could not
discriminate visually between the different chicken skin collagen
gel and lm preparations. In contrast, upon close inspection of the
telopeptide-poor collagen samples, gels there appeared to be more
turbid if soy protein isolate had been added. Films containing NaCl
were nearly transparent but contained distinct brous aggregates.
The transparent matrix around those structures suggested that
there, collagen was molecularly dispersed.
The formation of larger collagen assemblies has also been pre-
viously reported in studies investigating properties of low
concentrated telopeptide-poor collagen suspensions (Oechsle et al.,
2015b). There, addition of comparable concentrations of kosmo-
tropic potassium salts induced aggregation suggesting that
salting-out phenomena takes place both in low concentrated
suspensions as well as in gelled systems. Interestingly, the addition
of potassium chloride led to the development of ne structures in

Telopeptide-poor collagen Chicken skin collagen

Gel Film Gel Film

4% collagen

4% collagen
+ NaCl

2.75% collagen
+ 1.25% SPI

2.75% collagen
+ 1.25% SPI
+ NaCl

Fig. 1. Photographic images of modied telopeptide-poor collagen or chicken skin collagen gels and lms (lm width 40 mm). For modication 0.05 mol/kg NaCl was added and/
or 1.25% (w/w) of collagen was substituted with soy protein isolate (SPI).

Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
4 A.M. Oechsle et al. / Journal of Food Engineering xxx (2017) 1e10

suspensions, while in this study the addition of NaCl induced ber
formation. This can be attributed to the more chaotropic nature of
potassium compared to sodium as described by the reversed Hof-
meister series (Parsons et al., 2009). Chaotropic ions have also been
referred to as water-structure breaking compounds that is they
disrupt the ordering of water in turn affecting the hydration of
proteins due to entropic effects (Zhang and Cremer, 2006). Thus,
addition of potassium salts may induce formation of ner collagen
structures; whereas sodium salts trigger collagen-collagen in-
teractions and therefore aggregation. Our observation therefore
provides rst evidence on the previously presented hypothesis
postulating that low concentrations of sodium chloride may pro-
mote the formation of larger three-dimensional structures.
Formation of brous structures was not evident in the chicken
skin collagen lms suggesting that chicken skin collagen might not
be as sensitive to ions as telopeptide-poor collagen. In contrast to
our chosen model system, chicken skin collagen still contains
telopeptides, albeit the collagens are less crosslinked than in beef
collagen. The degree of covalent crosslinking is low in chicken
collagen because the young age at slaughter has not given the an-
imal time to sufciently crosslink collagen in their connective tissue
(Bailey and Shimokomaki, 1971; Maurer, 2003). Evidence of this has
also been provided by results of a previous study where chicken
skin collagen had a higher complex viscosity than telopeptide-poor
collagen (Oechsle et al., 2016a). More crosslinks would lead to more
collagen molecules being entangled and linked, which might in
turn hinder development of collagen aggregates due to changes in
ionic strength or ion type (Oechsle et al., 2015b). This would explain
the similarity in appearance of the different chicken skin collagen
lm samples.
For a more detailed examination of the microstructure, light mi-
croscopy pictures of telopeptide-poor and chicken skin collagen gels
and lms were taken (Fig. 2). The lm structure was generally denser
than the gel structure. This is likely due to a dewatering effect caused
by the saturated sodium chloride solution treatment of the extruded
collagen lms. During the treatment, the ions are competing with
collagen molecules for binding-sites of water molecules, and even-
tually collagen becomes dehydrated causing the network to become
denser due to increases in local biopolymer concentration. Moreover,
images reveal that telopeptide-poor collagen lms and gels con-
tained long bers, while chicken skin collagen contained short ones.
The formation of larger bers in gels and lms may be attributed
again to a hindrance of chicken skin collagen rearrangements due to
existing crosslinks. In contrast, telopeptide-poor collagens exist
mostly in the form of dispersed monomers allowing them to more
easily assemble and rearrange into bers.
In both telopeptide-poor and chicken skin collagen samples
containing soy protein, formation of large spherical aggregates

Telopeptide-poor collagen Chicken skin collagen

Gel Film Gel Film

4% collagen

4% collagen
+ NaCl

2.75% collagen
+ 1.25% SPI

2.75% collagen
+ 1.25% SPI
+ NaCl

Fig. 2. Light microscope images of modied telopeptide-poor collagen or chicken skin collagen gels and lms (scale bar 300 mm). For modication 0.05 mol/kg NaCl was added
and/or 1.25% (w/w) of collagen was substituted with soy protein isolate (SPI).

Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
 A.M. Oechsle et al. / Journal of Food Engineering xxx (2017) 1e10 5

could be observed. Soy protein isolate has an isoelectric point of
4.9e5.1 and becomes positively charged at the collagen gel condi-
tions (pH 2.7) (Roudsari et al., 2006; Wang et al., 2009). Due to the
electrostatic repulsion at pH 2.7, the soy proteins were actually
expected to distribute homogeneously. The reason for the aggre-
gation of soy protein isolate might be due to soy protein being quite
hydrophobic causing attractive interactions to occur (Hua et al.,
2005; Oechsle et al., 2015a). Soy protein isolate aggregates
appeared smaller and more evenly distributed when NaCl was
present. This might be explained by the fact that the addition of
NaCl increased their solubility by promoting hydrogen bonding
leading to a ner distribution of soy protein isolate. It should be
noted that at the same salt concentration and type, salting in can
occur for one protein while another one is simultaneously salted
out. Such phenomena in protein mixtures are currently the subject
of research due to the development of products containing mix-
tures of both animal and plant proteins.
To obtain insights into structural organization at the colloidal
and molecular level, scanning electron microscopy was conducted.
Fig. 3 depicts the images at a 100 fold magnication, since collagen

Telopeptide-poor collagen Chicken skin collagen

Film Film

4% collagen

4% collagen
+ NaCl

2.75% collagen
+ 1.25% SPI

2.75% collagen
+ 1.25% SPI
+ NaCl

Fig. 3. Scanning electron microscope (SEM) images (scale bar 200 mm) of modied telopeptide-poor collagen or chicken skin collagen lms. For modication 0.05 mol/kg NaCl
was added and/or 1.25% (w/w) of collagen was substituted with soy protein isolate (SPI).

Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
6 A.M. Oechsle et al. / Journal of Food Engineering xxx (2017) 1e10
structures did not differ considerably at higher magnications and
only the tetrahedral structure caused by sodium chloride crystals
was visible. Both the high NaCl concentration and the low tem-
perature made lms shrink causing crystalline deposits to be
formed due to the NaCl accumulating at the surface (Oechsle et al.,
2016b).
Interestingly, the microstructure of the telopeptide-poor
collagen lms did not noticeably differ for the various formula-
tions, even though light microscopic images showed structure
formation phenomena at larger length scales. It should be noted
though, that light microscopy is a transmission method revealing
intrinsic structure through their interaction with light, while
scanning electron microscopy reveals details about the surface of a
sputter-coated sample.
Conversely, changes in microstructure of chicken skin collagen
lms as a function of formulation were visible. The NaCl and soy
protein free samples featured a ne-grained, smooth surface. Upon
addition of NaCl, the surface became rougher containing larger
grains. This became even more pronounced when soy protein
isolate was added. Whether this is due to soy protein covering the
collagen bers or soy protein isolate altering the collagen assembly
and therefore the surface appearance, is currently unclear. How-
ever, the chicken skin collagen sample containing both soy protein
isolate and NaCl had again smother surfaces, supporting the hy-
pothesis of a ner soy protein isolate distribution when NaCl is
present.
3.2. Collagen gel and lm elasticity
The rheological properties of collagen gels and lms were
examined. Phase angles, as listed in Table 1, were determined from
the linear viscoelastic regions of the amplitude sweeps (graphs not
shown) of collagen gels. All samples behaved like viscoelastic gels
(45
> d > 0
) with chicken skin collagen featuring a slightly lower
d than telopeptide-poor collagen indicating that the elastic pro-
portion was higher in chicken skin collagen samples. Moreover,
chicken skin collagen also had extended linear viscoelastic regions
compared to telopeptide-poor collagen samples, and was stable up
to at least 10% strain (data not shown). In a previous study, a higher
degree of covalent crosslinks was associated with a reduced
viscoelastic region. Here, highly crosslinked gels were assumed as
less exible, since covalent bonds are more prone to be disrupted
compared to less crosslinked gels, which might rearrange based on
the self-assembly due to physical interactions. This supports the
assumption that chicken collagen features more covalent crosslinks
than telopeptide-poor collagen (Oechsle et al., 2016b; Oechsle et al.,
2014).
All collagen samples substituted with 1.25% (w/w) soy protein
isolate had decreased linear viscoelastic regions and a higher d than
Table 1
Phase angle d of modied telopeptide-poor collagen or chicken skin collagen gels in
the linear viscoelastic region of amplitude sweeps. For modication 0.05 mol/kg
NaCl was added and/or 1.25% (w/w) of collagen was substituted with soy protein
isolate (SPI).
Samples
Telopeptide-poor collagen
4% collagen
4% collagen NaCl
2.75% collagen 1.25% SPI
2.75% collagen 1.25% SPI NaCl
Chicken skin collagen
4% collagen
4% collagen NaCl
2.75% collagen 1.25% SPI
2.75% collagen 1.25% SPI NaCl
Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
samples without soy protein isolate, indicating that presence of the
soy protein isolate hindered the rearrangement and therefore the
elastic response regardless of the collagens origin. The higher d for
2.75% (w/w) compared to 4% (w/w) collagen samples might be
explained by the quantity of collagen molecules leading to more
entanglement and therefore to more elasticity. The addition of NaCl
increased the linear viscoelastic region of chicken skin collagen,
while d decreased.
In telopeptide-poor collagen, the addition of NaCl did not have
any effect on d. A higher mineral (NaCl) concentration in the
telopeptide-poor collagen (4.14 g/100 g) compared to the chicken
collagen (3.86 g/100 g), as found in a previous study, may be
responsible for this effect (Oechsle et al., 2016a). Salts are able to
increase the solubility of proteins by electrostatic screening,
altering hydrogen bonding and affecting hydrophobic interactions.
Once a critical ion concentration is reached, phase separation is
induced and the protein begins to precipitate (Arakawa and
Timasheff, 1984; Kuehner et al., 1996). Hence, the ionic strength
and therefore the effect on protein solubility in the telopeptide-
poor collagen could have been at an optimum below the critical
point and therefore the addition of 0.05 mol/kg did not further
increase collagen solubility or alter the viscoelastic range. In all
lms, d ranged between 16.7
and 29.7
with no systematic dif-
ferences observed between samples. The precipitated gels
remained viscoelastic after precipitation, although they were
showing a behavior much closer to an ideal elastic solid.
After identifying the linear viscoelastic region through ampli-
tude tests, frequency sweeps were performed to examine the time-
dependent behavior of the collagen gels and lms (Fig. 4). Gels had
much lower storage and loss moduli than the extruded collagen
lms, which may be attributed to the previously mentioned
dehydration induced by the submersion in the saturated sodium
chloride solution causing an increase in density and interactions of
biopolymers (Vasilev et al., 1973). The somewhat higher storage
and loss moduli of chicken skin collagen compared to telopeptide-
poor collagen in the gel state are not unexpected, and are as pre-
viously mentioned likely due to a higher degree of covalent cross-
links. This would limit intermolecular movements during
deformation resulting in higher G0 and G00 values in the low fre-
quency range (Metzger, 2010).
Furthermore, both collagen gels had their highest G0 and G00
without NaCl or soy protein. The addition of 0.05 mol/kg NaCl to
chicken collagen weakened gels due to changes in collagen solu-
bility. The substitution of 1.25% (w/w) of collagen by soy protein
isolate similarly led to a decrease in G0 and G00 values in collagens of
both origins. The known hydrophobicity of soy proteins contrib-
uted to the formation of soy protein aggregates, rather than soy
protein being nely distributed throughout the collagen matrix, as
would be the case for a co-dispersible, fully miscible system
(Nishinari et al., 2014). This was also evident from light microscopy
images where long collagen bers appeared to be wrapped around
large soy protein aggregates.
Although telopeptide-poor collagen and chicken skin collagen
gels had different rheological properties, mechanical properties of
lms formed from the two gels were surprisingly similar. This can
d (
) be ascribed to the fact that in gels, rheological properties are largely
dominated by molecules being loosely entangled due to polymers
21.8 being fully hydrated, while in precipitated lms polymer concen-
21.8 tration and density increases to the point where substantial me-
29.7
chanical forces can be transmitted across the innite network
28.8
(Oechsle et al., 2016b). Standard deviations were slightly smaller for
16.7 chicken skin collagen lms, suggesting that samples were more
19.8 homogeneous, which was also evident in photographic images.
20.8 When lms were handled, chicken collagen samples were notice-
22.7
ably dryer and rmer than their telopeptide-poor counterparts.
 A.M. Oechsle et al. / Journal of Food Engineering xxx (2017) 1e10 7

Telopeptide-poor collagen Chicken skin collagen

1000 1000
G' and G'' (Pa)

G' and G'' (Pa)

Gel

100 100

0.1 1 10 0.1 1 10
Frequency (Hz) Frequency (Hz)
G' and G'' (Pa)

G' and G'' (Pa)

10000 10000

Film

1000 1000
0.1 1 10 0.1 1 10
Frequency (Hz) Frequency (Hz)

4% collagen G'
4% collagen G''
4% collagen + NaCl G'
4% collagen + NaCl G''
2.75% collagen + 1.25% SPI G'
2.75% collagen + 1.25% SPI G''
2.75% collagen + 1.25% SPI+ NaCl G'
2.75% collagen + 1.25% SPI+ NaCl G''
Fig. 4. Frequency sweeps with storage (G0 ) and loss moduli (G00 ) of modied telopeptide-poor collagen or chicken skin collagen gels and lms. For modication 0.05 mol/kg NaCl
was added and/or 1.25% (w/w) of collagen was substituted with soy protein isolate (SPI).

The addition of NaCl to chicken collagen gels led to higher loss
and storage moduli of lms, although the addition of NaCl was
found to weaken gels. One explanation of this could be that the
lower viscosity of gels facilitates a better orientation of collagen
bers upon treatment in the salt bath yielding stronger lms after
extrusion. If that is generally the case, then lower viscosities of
collagen gels might even be benecial to ensure an enhanced
elasticity of extruded lms. In contrast though, telopeptide-poor
collagen did not yield lms that had increased storage moduli
when salt was added. It should be noted though that telopeptide-
poor collagen already contained a higher initial mineral content,
reducing the effect of any additionally added salt. This signies the
importance of the collagen gel composition prior to extrusion.
The addition of soy protein isolate reduced the storage and loss
moduli of all lms. Telopeptide-poor collagen was weakened to a
greater extent by the substitution of soy protein isolate than
chicken skin collagen, based again on the higher degree of covalent
crosslinks. This effect was also reported in a previous study where
highly crosslinked native collagen was found to incorporate soy
protein isolate more effectively in extruded lms than telopeptide-
poor collagen (Oechsle et al., 2016b). The weakening effect of soy
protein isolate could be partially compensated for in lms though

Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
8 A.M. Oechsle et al. / Journal of Food Engineering xxx (2017) 1e10

by addition of salt. This supports the assumption that some Table 2

collagen can be replaced by alternative proteins as long as those
proteins form aggregates that are small enough to be evenly
distributed throughout the collagen matrix.
Mean thicknesses of modied telopeptide-poor collagen or chicken skin collagen
lms. For modication 0.05 mol/kg NaCl was added and/or 1.25% (w/w) of collagen
was substituted with soy protein isolate (SPI). Values are means of six
replicates standard deviation. A-B means of samples indicating an identical letter
are not signicantly different (P > 0.05).

Collagen origin and modulation Mean lm thickness (mm)

3.3. Collagen lm thickness and tensile strength Telopeptide-poor collagen
4% collagen 0.58 0.17A
Finally, we measured tensile strength (Fig. 5) to gain further 4% collagen NaCl 0.38 0.12A
insights into the stability of collagen lms. Since the tensile 2.75% collagen 1.25% SPI 0.61 0.03AB
2.75% collagen 1.25% SPI NaCl 0.85 0.20 B
strength is inuenced by the dimensions of the lm, the lm
Chicken skin collagen
thickness was also measured (Table 2). Compared to the width of 4% collagen 0.45 0.10A
the extrusion die (25 mm), the lm thickness increased by a factor of 4% collagen NaCl 0.54 0.08A
up to 34 due to extrudate swelling. Among the chicken collagen 2.75% collagen 1.25% SPI 0.48 0.09A
lm samples, no signicant differences in lm thickness (P > 0.05) 2.75% collagen 1.25% SPI NaCl 0.50 0.05A

irrespective of formulations were observed. Telopeptide-poor

collagen lms containing both NaCl and soy protein isolate were
signicantly thicker than any chicken skin collagen lms (P < 0.05)
3.4. Key insights
and with NaCl (P < 0.001).
Although there were only few signicant differences in the
In summary, we can highlight a number of key insights obtained
lms thicknesses, a reverse correlation between tensile strength
from this study, as illustrated in Fig. 6:
and lm thickness could be established, e.g. the thickest lm
(telopeptide-poor collagen with NaCl and soy protein isolate)
- The substitution of collagen with soy protein isolate decreased
featured the lowest tensile strength (Table 1). This reverse corre-
the gels strength, as well as the lms elasticity and tensile
lation was also reported in a previous study, where increased
strength. A partially compensation of this structure weakening
entanglement was linked to thinner lms with higher tensile
effect could be achieved by addition of 0.05 mol/kg NaCl leading
strengths (Oechsle et al., 2016b). This also agrees well with
to a ner distribution of soy proteins in the collagen network. As
rheology results, where samples having the highest lm thickness
such, lms with lower collagen content can be manufactured by
also had the lowest storage moduli. Soy proteins might have made
use of alternative proteins using appropriate amounts of salt.
the lms more sensitive to rupture since they hinder collagen
- Chicken skin collagen generally yielded gels with higher
interaction, which provides the backbone of the network. Hence,
strengths. Microscopy and rheological measurements in com-
the gel strength is also inversely correlated to the lm thickness.
bination with data available in literature allow us to postulate
The so-called extrudate swell is responsible for the increase of the
that this is due to a higher degree of covalent crosslinks in
lm thickness, due to the collagen chains tend to recoil in the ow
chicken skin collagen compared to telopeptide-poor collagen.
direction and grow in the normal direction after exerting the die
Such gels also allow for a more effective incorporation of alter-
(Muksing et al., 2008). Extrudate swell depends on the elasticity,
native proteins such as those from soy.
the presence of bers and their entanglement (Wang, 2010). The
- The addition of 0.05 mol/kg NaCl induced ber formation and
entanglements will, to some extent, prevent the molecules from
increased tensile strength of telopeptide-poor collagen, as well
slipping past one another, thus preventing total relaxation of the
as elasticity in chicken skin collagen lms.
molecules (Dangtungee et al., 2005).

Telopeptide-poor collagen Chicken skin collagen

0.08 0.08 a
A a
Tensile strength (MPa)

Tensile strength (MPa)

b
0.06 0.06

0.04 B 0.04

C BC
0.02 0.02

0.00 0.00
4% 4% 2,75% 2,75% 4% 4% 2,75% 2,75%
+ NaCl + SPI + SPI + NaCl + SPI + SPI
+ NaCl + NaCl
Fig. 5. Tensile strength of modied telopeptide-poor collagen or chicken skin collagen lms. For modication 0.05 mol/kg NaCl was added and/or 1.25% (w/w) of collagen was
substituted with soy protein isolate (SPI). Equal letters (a-b or A-B) are used to label samples that are not signicantly different (P > 0.05).

Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
 A.M. Oechsle et al. / Journal of Food Engineering xxx (2017) 1e10 9

Fig. 6. Schematic illustration of extruded telopeptide-poor and chicken skin collagen lms modied by addition of 0.05 mol/kg NaCl and/or substitution of 1.25% (w/w) collagen
with soy protein isolate (SPI) with a very lower (YY), lower (Y), higher ([), or very higher ([[) tensile strength and storage modulus G0 .

4. Conclusion Acknowledgments

This study demonstrated the potential of chicken collagen to be
utilized as lm formers in co-extrusion processes. The partial
substitution of collagen is feasible if soy protein addition is com-
bined with low concentrations of NaCl supporting an even distri-
bution of foreign proteins in the collagen matrix. However, to gain
further insights into practical implementations of such modica-
tions approaches, the impact of a broader salt concentration range,
different salt types, and different protein types will need to be
examined. Addition of salt alone improved lm properties due to
modulations in collagen molecular interactions. For telopeptide-
poor collagen lms, NaCl increased the tensile strength, which is
a particularly important feature if lms are to be used as sausage
casings. There, mechanical properties inuence snap and bite of
sausages in turn inuencing consumer acceptance of products. In
chicken skin collagen lms, addition of salt increased the storage
modulus, which is important for lms maintaining the ability to
expand and shrink during heating and cooling of sausages. Taken
together, the results show that end product attributes of collagen-
based products can be rationally designed by understanding and
tailoring molecular interaction of collagen biopolymer.
The authors would like to thank Robert Wilfer, Katja Mader and
Dennis Schll, all with Kalle GmbH, for providing samples and
machinery as well as technical support. Also we would like to thank
Gert and Marion Bker (Protein Consulting) for sample provision
and Barbara Maier for her assistance with SEM images. This
research project was supported by the German Ministry of Eco-
nomics and Energy (via AiF) and the FEI (Forschungskreis der
hrungsindustrie e.V., Bonn). Project AiF 17478 N.
Erna
References
Arakawa, T., Timasheff, S.N., 1984. Mechanism of protein salting in and salting out
by divalent cation salts: balance between hydration and salt binding.
Biochemistry 23 (25), 5912e5923.
Bailey, A.J., Shimokomaki, M.S., 1971. Age related changes in the reducible cross-
links of collagen. FEBS Lett. 16 (2), 86e88.
Bker, G., Bker, M., Schmidt, M., Stalberg, S.D., 2010. Edible Food Casing Based on
Proteins, and Process for Production Thereof.
Caliari, S.R., Harley, B.A.C., 2011. CollageneGAG materials. In: Ducheyne, P. (Ed.),
Comprehensive Biomaterials. Elsevier, Oxford, pp. 279e302.
Campbell, N.A., Reece, J.B., Wasserman, S.A., 2009. Struktur und Funktion von
Bindegeweben. In: Biologie, 8 ed., pp. 1158e1162 (Pearson Studium).

Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
10 A.M. Oechsle et al. / Journal of Food Engineering xxx (2017) 1e10
Cliche, S., Amiot, J., Avezard, C., Gariepy, C., 2003. Extraction and characterization of
collagen with or without telopeptides from chicken skin. Poult. Sci. 82 (3),
503e509.
Dangtungee, R., Yun, J., Supaphol, P., 2005. Melt rheology and extrudate swell of
calcium carbonate nanoparticle-lled isotactic polypropylene. Polym. Test. 24
(1), 2e11.
Eckmayer, Z., Do rstewitz, R., Schlosser, L., Bohni, J.A., Geistlich, P., 2002. In: Method
of Making a Collagen Membrane from Porcine Skin (In).
Hoogenkamp, H.R., Bakker, G.-J., Wolf, L., Suurs, P., Dunnewind, B., Barbut, S.,
Friedl, P., van Kuppevelt, T.H., Daamen, W.F., 2015. Directing collagen bers
using counter-rotating cone extrusion. Acta Biomater. 12, 113e121.
Hua, Y., Cui, S.W., Wang, Q., Mine, Y., Poysa, V., 2005. Heat induced gelling prop-
erties of soy protein isolates prepared from different defatted soybean ours.
Food Res. Int. 38 (4), 377e385.
Kuehner, D.E., Blanch, H.W., Prausnitz, J.M., 1996. Salt-induced protein precipitation:
phase equilibria from an equation of state. Fluid Phase Equilib. 116 (1e2),
140e147.
Lieberman, E.R., 1964. In: Ofce, U.S.P. (Ed.), Apparatus and Method for Producing
Collagen Tubing by Continuous Extrusion. Johnson & Johnson.
Liu, D., Zhang, X., Li, T., Yang, H., Zhang, H., Regenstein, J.M., Zhou, P., 2015.
Extraction and characterization of acid- and pepsin-soluble collagens from the
scales, skins and swim-bladders of grass carp (Ctenopharyngodon idella). Food
Biosci. 9 (0), 68e74.
Maurer, A.J., 2003. Poultry | chicken. In: Caballero, B. (Ed.), Encyclopedia of Food
Sciences and Nutrition, 2 ed. Academic Press, Oxford, pp. 4680e4686.
Metzger, T.G., 2010. Das Rheologie Handbuch. Vincentz Network, Hanover,
Germany.
Muksing, N., Nithitanakul, M., Grady, B.P., Magaraphan, R., 2008. Melt rheology and
extrudate swell of organobentonite-lled polypropylene nanocomposites.
Polym. Test. 27 (4), 470e479.
Nishinari, K., Fang, Y., Guo, S., Phillips, G.O., 2014. Soy proteins: a review on
composition, aggregation and emulsication. Food Hydrocoll. 39, 301e318.
Oechsle, A.M., Akgn, D., Krause, F., Maier, C., Gibis, M., Kohlus, R., Weiss, J., 2016a.
Please cite this article in press as: Oechsle, A.M., et al., Modication of extruded chicken collagen lms by addition of co-gelling protein and
sodium chloride, Journal of Food Engineering (2017), http://dx.doi.org/10.1016/j.jfoodeng.2017.03.017
Microstructure and physicalechemical properties of chicken collagen. Food
Struct. 7, 29e37.
Oechsle, A.M., Ha upler, M., Gibis, M., Kohlus, R., Weiss, J., 2015a. Modulation of the
rheological properties and microstructure of collagen by addition of co-gelling
proteins. Food Hydrocoll. 49 (0), 118e126.
Oechsle, A.M., Ha upler, M., Weigel, F., Gibis, M., Kohlus, R., Weiss, J., 2016b. Mod-
ulation of extruded collagen lms by the addition of co-gelling proteins. J. Food
Eng. 171, 164e173.
Oechsle, A.M., Landenberger, M., Gibis, M., Irmscher, S.B., Kohlus, R., Weiss, J., 2015b.
Modulation of collagen by addition of Hofmeister salts. Int. J. Biol. Macromol. 79
(0), 518e526.
Oechsle, A.M., Wittmann, X., Gibis, M., Kohlus, R., Weiss, J., 2014. Collagen entan-
glement inuenced by the addition of acids. Eur. Polym. J. 58 (0), 144e156.
m, M., Maceina, T.J., Salis, A., Ninham, B.W., 2009. Why direct or
Parsons, D.F., Bostro
reversed hofmeister Series? Interplay of hydration, non-electrostatic potentials,
and ion size. Langmuir 26 (5), 3323e3328.
Prockop, D.J., 2013. Collagens. In: Lane, W.J.L.D. (Ed.), Encyclopedia of Biological
Chemistry. Academic Press, Waltham, pp. 545e549.
Roudsari, M., Nakamura, A., Smith, A., Corredig, M., 2006. Stabilizing behavior of soy
soluble polysaccharide or high methoxyl pectin in soy protein isolate emulsions
at low pH. J. Agric. Food Chem. 54 (4), 1434e1441.
Vasilev, M.P., Kotetskii, V.V., Pukhova, Z.I., Volf, L.A., Istranov, L.P., 1973. Collagen
precipitation from solutions. Fibre Chem. 4 (5), 486e488.
Wang, D., Sun, X.S., Yang, G., Wang, Y., 2009. In: Engineers, A.S. o. A. a. B. (Ed.),
Improved Water Resistance of Soy Protein Adhesive at Isoelectric Point.
Wang, K., 2010. Description of extrudate swell for polymer nanocomposites. Ma-
terials 3 (1), 386e400.
Zeugolis, D.I., Raghunath, M., 2011. Collagen: materials analysis and implant uses.
In: Ducheyne, P. (Ed.), Comprehensive Biomaterials. Elsevier, Oxford,
pp. 261e278.
Zhang, Y., Cremer, P.S., 2006. Interactions between macromolecules and ions: the
Hofmeister series. Curr. Opin. Chem. Biol. 10 (6), 658e663.