International Journal of Food Science and Technology 2012, 47, 23972404 2397

Original article
Process optimisation and physicochemical characterisation of
enzymatic hydrolysates of proteins from co-products of Atlantic
Salmon (Salmo salar) and Yellowtail Kingfish (Seriola lalandi)

Shan He,1,2,3 Chris Franco1,2,3 & Wei Zhang1,2,3*

1 Department of Medical Biotechnology, Flinders Medical Science and Technology, School of Medicine, Flinders University, Sturt Road,
Bedford Park, SA, 5042, Australia
2 Flinders Center for Marine Bioprocessing and Bioproducts (FCMB2), Flinders University, Sturt Road, Bedford Park, SA, 5042, Australia
3 Australian Seafood Cooperative Research Center, Box 26, Mark Oliphant Building, Science Park, Laer Drive, Bedford Park, Adelaide, SA,
5042, Australia
(Received 23 January 2012; Accepted in revised form 12 May 2012)

Summary The aim of this study was to develop an enzymatic hydrolysis process of protein co-products for two
major commercial sh species in Australia: Atlantic salmon (AS) and Yellowtail kingsh (YTK). The out-
comes are to produce high protein recovery of sh protein hydrolysates within controlled molecular
weight ranges that display enhanced physicochemical properties of oil binding and emulsication. Three
enzymes (Flavourzyme, Neutrase and Alcalase) were applied to processing co-products. Protein recovery
and physicochemical properties were evaluated with increasing hydrolysis time from 30 min to 180 min
and ratio of enzyme to substrate (E/S) from 0.5% to 3.0%. In order to achieve a product with
optimum emulsifying capacity (50 0.6 m2 g
1), an E/S ratio of 0.61.3% Flavourzyme was applied for
30111 min with a protein recovery of 55%; in order to achieve a product with optimum oil-binding capac-
ity (8.3 0.3 g oil g hydrolysates
1), an E/S ratio of 2.33.0% Flavourzyme was applied for 2564 min
with a protein recovery of 70%. YTK protein hydrolysates were further membrane-fractionated into ve
fractions (>100 kDa, 50100 kDa, 3050 kDa, 1030 kDa and <10 kDa), and of these, the 1030 kDa
exhibited the best properties of oil binding (19 0.3 g oil g hydrolysates
1) and emulsication
(57 0.7 m2 g
1). These results demonstrate the importance of enzymatic hydrolysis of seafood co-prod-
ucts into high-value ingredients for food products and processing.
Keywords Enzymatic processing condition, sh processing co-products, sh protein hydrolysates, fractionation, optimisation.

This inecient business model has been identied to

Introduction
Fish processing co-products can be dened as sh
material left over from primary processing during the
sh manufacturing process. Bechtel (2003) determined
the yield of sh processing co-products from twelve
sh species and found that they were all over 50% of
the total weight of the starting material. In Australia,
seafood industries discard over 100 000 tonnes of these
co-products annually. Currently, it costs AUD $150/
tonne to take away, which means that the Australian
seafood industry spends over AUD $15 million pa
1
to dispose of the products as waste rather than
generate benets via value-adding, because of legal
restriction, rising treatment costs and environmentally
conscious consumers (Arvanitoyannis & Ladas, 2008).
*
Correspondent: Fax: +61 8 72218555;
e-mail: wei.zhang@inders.edu.au
be not only cost-ineective but also environment
unfriendly. In Europe, numerous regulations are in
place to minimise the environmental problems caused
by sh processing wastes (Arvanitoyannis & Kassaveti,
2008). In Australia, the utilisation of sh co-products
to produce value-added products has been highlighted
as one of the high-priority areas by the Australian
Seafood Industry Council (Ian et al., 2004).
Fish processing co-products have been used as raw
materials to produce protein-based products such as
sh meal and animal feeds (Arvanitoyannis, 2008).
However, these products are of low economic value.
To upgrade their commercial potential, proteases such
as Flavourzyme (Nilsang et al., 2005), Neutrase
(Benjakul & Morrissey, 1997) and Alcalase (Wasswa
et al., 2007) were applied to convert these proteins to
protein hydrolysates with improved physicochemical
properties such as a higher capacity for oil binding

doi:10.1111/j.1365-2621.2012.03115.x
2012 The Authors. International Journal of Food Science and Technology 2012 Institute of Food Science and Technology
2398 Optimize fish protein hydrolysates process S. He et al.
and emulsication of fats. This process has been
applied to produce sh protein hydrolysates from vari-
ous sh species such as Capelin (Shahidi et al., 1995)
and Pacic Whiting (Pacheco-Aguilar et al., 2008).
Previous studies also demonstrated that longer
processing times and higher E/S ratio produced sh
protein hydrolysates with higher protein recovery but
reduced oil-binding capacity and emulsifying capacity
(Slizyte et al., 2005). However, to our knowledge, the
optimum enzymatic processing conditions to obtain
protein hydrolysates from sh processing co-products
with both protein recovery and market standard values
of physicochemical properties have not been
determined for any sh species.
The molecular weight (MW) range of protein
hydrolysates signicantly aects functions of sh pro-
tein hydrolysates. Raghavan & Kristinsson (2008)
demonstrated protein hydrolysates from Tilapia, with
the MW range between 3.5 and 10 kDa, possessing the
strongest antioxidative activity. Je et al. (2004) found
that the antihypertension activity of protein
hydrolysates from Cod frame increased from 59.6% to
87.6% when their MW changed from 30 to 10 kDa
to < 1 kDa. Production of bioactive sh protein
hydrolysates on an industrial scale has also been
studied recently (Picot et al., 2010). The relationship
between MW ranges of sh protein hydrolysates and
their bioactivities has been well studied before;
however, studies on the inuence of MW ranges of sh
protein hydrolysates on their physicochemical proper-
ties are limited to a few sh species, such as Cod (Jeon
et al., 1999).
The objective of this study was to optimise
enzymatic hydrolysis conditions to produce protein
hydrolysates from sh processing co-products of AS
and YTK, two major commercial sh species in
Australia, to achieve both high protein recovery and
market standards of physicochemical properties and to
further determine the inuence of their MW ranges on
their physicochemical properties. This research is part
of an ongoing project to develop value-added biopro-
cesses and products for the full utilisation of sh pro-
cessing co-products from AS and YTK to minimise
the environmental impact and maximise the nancial
returns for sh processors.
Materials and methods
Materials
Atlantic Salmon (AS) and Yellowtail Kingsh (YTK)
were supplied by Simplot Australia (Mentone, Victoria
3194, Australia). Chemicals were purchased from
Sigma-Aldrich Australia Pty. Ltd (Castile Hill, NSW,
1765, Australia). Commercial Flavourzyme, Neutrase
and Alcalase were provided by Novozymes Australia
International Journal of Food Science and Technology 2012
International Journal of Food Science and Technology 2012 Institute of Food Science and Technology
Pty. Ltd (North Rocks, NSW 2151, Australia). Eggs
were purchased from a local supermarket. Cow gelatin
was purchased from Ajax Chemicals (Cheltenham,
Vic. 3192, Australia). Spectrophotometer was
purchased from Bio-Tek (100 Tigan Street, Winooski,
VT, USA). Fifty-millilitre centrifugal lters were
purchased from Merck Millipore (290 Concord Road,
Billerica, MA, USA). Dual colour stain protein molec-
ular weight marker was purchased from Bio-Rad
(1000 Alfred Nobel Drive, Hercules, CA, USA).
Sample collection and preparation
Eight whole AS (length: 70 2.0 cm, weight:
4.2 0.6 kg) and YTK (length: 68 2.0 cm, weight:
2.9 0.7 kg) harvested from Tasmania and South
Australia, respectively, were packed in ice boxes and
transported within 24 h of harvesting to Angelakis
Bros. sh processing factory (Adelaide, SA 5000,
Australia) for lleting. Dierent sections of the
co-products including head, skin, frame, belly ap and
viscera of each sh were separated and weighed. These
co-products were then packed in ice boxes and trans-
ported to the cold room (04 C) at the Department
of Medical Biotechnology, Flinders University
(Bedford Park, SA 5042, Australia). They were minced
separately to uniformity with an electronic heavy duty
meat mincer (MGT0012, Nemco Food Equipment
Ltd., OH, USA) immediately, then stored at
20 C
until use. The period between harvesting AS and YTK
and their freezing was within 24 h.
Experimental design and analysis
Frozen mince of AS head, AS frame, YTK head and
YTK frame was packed in sealed bags, then thawed
for about 30 min under running tap water. Head and
frame of AS and YTK were mixed in the ratio of their
processing yield according to Table S2 (AS head/AS
frame = 8.8:13, YTK head/YTK frame = 21:15),
respectively, with the total weight of 50 g, then mixed
evenly with 50 mL tap water. The enzymatic process
was performed using three commercial proteases:
Flavourzyme 500MG, Alcalase 2.4L FG and Neutrase
0.8L. Temperature and pH of mince-water slurries
were adjusted to the optimal working conditions of
each enzyme (Flavourzyme 500MG: 45 C, pH 6.5;
Alcalase2.4L FG: 55 C, pH 8.5; Neutrase 0.8L: 45
C, pH 6.5) before enzymatic processing. The natural
pH of mince-water slurries was 6.5, and therefore,
optimal working pH of Flavourzyme and Neutrase did
not need to be adjusted and optimal working pH of
Alcalase was adjusted using NaOH pellets.
Processing time and E/S ratio were designed using
response surface methodology (RSM), and this was
facilitated by a Minitab software version 15 package
2012 The Authors

(Minitab Pty Ltd, Sydney, NSW 2000, Australia). The
independent variables of E/S ratio: x1 (0.53%) and
processing time: x2 (30180 min) were varied at three
levels of lowest, middle and highest: 0.5%, 1.8%, 3%
and 30 min, 105 min, 180 min, respectively. The
ranges of variables were chosen based on the highest
and lowest values from related studies (Benjakul &
Morrissey, 1997; Wasswa et al., 2007). Nine processing
conditions were generated by Minitab software
package (version 15) in a random order for the two
variables (Table S1). The response functions (y) were
protein recovery, oil-binding capacity and emulsifying
capacity of the resultant sh protein hydrolysates.
These values were related to the independent vari-
ables (x1, x2) by a second-degree polynomial:
y = b0 + b1x1 + b2x2 + b12x1x2 + b11x12 + b22x22.
The coecients of the polynomial were b0 (constant
term), b1 and b2 (linear eects), b12 (interaction eect),
b11 and b22 (quadratic eects). The eects of these
coecients were determined by ANOVA tables. Finally,
statistical calculations were used to generate overlaid
contour maps from the regression models.
The enzymes were inactivated at the end of each
processing time by increasing the temperature to 90 C
for 30 min. The heated material was then centrifuged
at 4000 g for 30 min, resulting in three layers: lipid
layer on the top, protein hydrolysate solution in the
middle and a semisolid layer at the bottom. The lipid
layer was removed by aspiration; then, the protein
hydrolysate solution was decanted, collected and
freeze-dried. The freeze-dried protein hydrolysates
were weighed to record yields, then placed in 50-mL
sealed tubes and stored in desiccators at room
temperature until use.
Physicochemical properties tests
Oil-binding capacity was measured by a published
method (Wasswa et al., 2007) with a slight
modication. Half a gram of each powder sample
was added to 9 g of canola oil in a 50-mL centri-
fuge tube, mixed for 1 min with a vortex mixer,
then centrifuged at 2000 9 g for 30 min at room
temperature. The weight of oil separated from the
hydrolysate was measured, and oil-binding capacity
was calculated as the amount of oil absorbed by 1 g
sample. Egg white powder (self-produced by freeze-
drying egg white), a commonly used oil binder, was
used as reference.
Emulsifying capacity was measured by a published
method (Klompong et al., 2007) with a slight
modication. Three grams of canola oil and 10 mL
1% solution of each powder sample were homogenised
in a bio-homogeniser (IKA Labortechnik, Staufen,
Germany) at a speed of 25 000 rpm for 1 min. 50 lL
of emulsion was pipetted from the bottom, then mixed
2012 The Authors
International Journal of Food Science and Technology 2012 Institute of Food Science and Technology
Optimize fish protein hydrolysates process S. He et al. 2399
with 5 mL of 0.1% sodium dodecyl sulphate (SDS)
solution. The absorbance (A500) of the SDS-diluted
solution was measured immediately at 500 nm by a
Bio-Tek spectrophotometer. Emulsifying activity was
calculated as follows:
2  2:303  A500
Emulsifying activity m2 =g
0:75  protein weight (g)
Egg white and cow gelatin powders, two commercial
emulsifying agents, were used as references.
Protein hydrolysate fractionation
One gram YTK processing co-products protein
hydrolysates hydrolysed by Flavourzyme from each
processing condition shown in Table S1 was dissolved
in double-distilled water at a concentration of
20 mg mL
1. The solution was vacuum-ltered using
Whatman No. 1 lter paper, then fractionated
sequentially using Merck Millipore 50-mL centrifugal
lters with dierent molecular weight cut-o of
100 kDa, to 50 kDa, to 30 kDa and 10 kDa. The
samples were centrifuged at 5000 9 g for 1 h. Five
groups of fractionated solutions of sh protein
hydrolysates with molecular weight ranges of
>100 kDa, 10050 kDa, 5030 kDa, 3010 kDa and
<10 kDa were freeze-dried, and the dried sh protein
hydrolysates were sealed in 50-mL centrifuge tubes
and stored in desiccators at room temperature
until use.
Molecular weight distribution of protein hydrolysates
The molecular weight distribution of the sh protein
hydrolysates was analysed by SDS gel electrophoresis
according to a published method (He et al., 2010)
with a slight modication. Concentrations of loading
sample solutions were adjusted to 3.3 mg mL
1 mea-
sured by the BCA method. Then, 15 lL of each
protein solution was mixed with 5 lL 49
SDS-PAGE sample-loading buer (40% glycerol, 4%
lithium dodecyl sulphate, 4% Ficoll-400, 0.8 M trieth-
anolamine-C1 pH 7.6, 0.025% phenol red, 0.025%
Coomassie G250, 2 mM EDTA disodium) followed
by heating in a hot water bath at 90 C for 3 min
before loading into wells on the gel. A Bio-Rad dual
colour stain protein molecular weight marker ranging
from 10 to 250 kDa was used. A gradient gel with
concentration from 4% to 20% was placed into a
gel tank lled with 4 mM MOPS gel buer (200 mM
MOPS, pH 7.0; 80 mM sodium acetate; and 10 mM
EDTA, pH 8.0), and then samples were loaded into
20-lL wells. The gel tank was connected to a power
pack supplying a voltage that increased to 200 V
with a current between 80 and 90 mA for 90 min.
Subsequently, the gel was stained with Coomassie
International Journal of Food Science and Technology 2012
2400 Optimize fish protein hydrolysates process S. He et al.
Blue for 1 h and destained with 100 mL destaining
solution (50 mL/100 mL double-distilled water,
45 mL/100 mL methanol and 5 g/100 mL acetic
acid) for 2 h.
Statistical analysis
Experiments were repeated at least three times. Data
are presented as the average with standard deviation
and subjected to one-way analysis of variance (ANOVA)
using SPSS software version 17.0. The signicance was
judged statistically by the F value at probability (P)
below 0.05.
Results and discussion
Processing yields of co-products from AS and YTK
The total yield of processing co-products from AS and
YTK was about 50% of the whole sh based on the
wet weight (Table S2). Utilisation of sh processing
co-products has been studied previously. It has been
found that sh viscera was a good resource to enhance
the growth of lactic acid bacteria (Dong et al., 1993),
and sh skin can be used to extract gelatin (Karim &
Bhat, 2009), which can be used in cosmetic products
(Arvanitoyannis & Kassaveti, 2008). And belly ap is
a good option to produce sh oil for biodiesel produc-
tion (Arvanitoyannis & Kassaveti, 2008) and omega-3
fatty acids owing to its high fat content (He et al.,
2011). Fish muscle proteins including the proteins
from llet, head and frame can be used to produce sh
protein hydrolysates with good physicochemical prop-
erties (Jeon et al., 1999) and bioactives (Je et al.,
2004).
This study focused on utilising sh processing co-
products, and therefore, head and frame of AS and
YTK were used as raw materials to produce sh
protein hydrolysates in this study.
Enzymatic processing yield and protein recovery of fish
protein hydrolysates
Enzymatic processing yields and protein recoveries of
sh protein hydrolysates produced by the three
enzymes under dierent conditions are presented in
Table S3. He et al. (2011) reported that YTK head
and frame contain 14.8% and 13.9% proteins, respec-
tively; AS head and frame contain 23% and 13% pro-
teins, respectively. In combination with the data in
Table S2, it can be calculated that protein contents of
50 g YTK and AS raw materials were 7.2 g and 8.6 g,
respectively. Protein recoveries were calculated based
on enzymatic processing yields and protein contents of
50-g raw materials. Higher yields and protein recover-
ies were associated with longer processing time and
International Journal of Food Science and Technology 2012
International Journal of Food Science and Technology 2012 Institute of Food Science and Technology
higher E/S ratio. The smaller peptides have more
chance to form hydrogen bonds between water
molecules and hydrophilic polar amino acid groups,
thus increasing their water solubility (Kristinsson &
Rasco, 2000) The highest enzymatic processing yields
were produced using 180-min processing time, 3% E/S
ratio, from AS processing co-products by Alcalase
(6.2 0.3 g), YTK processing co-products by
Flavourzyme (6.2 0.1 g), Neutrase (6.1 0.2 g) and
Alcalase (6.4 0.3 g), respectively, without signicant
dierences between these treatments.
Physicochemical properties of fish protein hydrolysates
Oil-binding capacity
Oil-binding capacity of sh protein hydrolysates is
reported in Table S4. Oil-binding capacity is an
important function used in meat and confectionery
industries (Sathivel et al., 2004). The oil-binding
capacity of protein hydrolysates from YTK was
generally higher than that from AS under the same
processing conditions. With the longer processing time
and higher E/S ratio, oil-binding capacity of protein
hydrolysates was reduced. Longer processing times
and higher E/S ratios destroy protein structure, result-
ing in the degradation of the network formed by the
protein to entrap oil, thus diminishing the oil-binding
capacity (Don et al., 1991). The highest oil-binding
capacity was produced from YTK processing co-prod-
ucts, by Flavourzyme with a processing time of 30 min
and an E/S ratio of 0.5% (14 1.2 g oil g hydroly-
sates
1) and 1.8% (13 0.1 g oil g hydrolysates
1)
and by Neutrase with a processing time of 30 min with
an E/S ratio of 0.5% (13 0.2 g oil g hydrolysates
1),
respectively, without a signicant dierence between
these treatments.
Oil-binding capacity of egg white protein powder
(8.3 0.3 g oil g hydrolysates
1) was used as reference
(Table S4). Protein hydrolysates from YTK produced
with short processing times and a low E/S ratio can
generally reach this value. However, similar values
could not be achieved using AS as a raw material.
This indicated that processing co-products of YTK are
a better resource than those of AS to produce protein
hydrolysates as commercial oil binders.
Emulsifying capacity
Emulsifying capacity of sh protein hydrolysates is
reported in Table S5. Protein hydrolysates adsorb to
the surface of formed oil droplets during homogenisa-
tion. This action results in the formation of a
protective water layer that inhibits coalescence of the
oil droplets (Gbogouri et al., 2004). Protein
hydrolysates are more surface active than intact
proteins, because of the exposure of hydrophilic and
2012 The Authors

hydrophobic amino acid groups. They also move
faster at the interface of water and oil than intact pro-
teins. Therefore in comparison with intact proteins,
protein hydrolysates absorb oil more eectively, so
making it easier to form the oil-in-water emulsion
(Klompong et al., 2007). However, the results in
Table S5 demonstrate the fact that emulsifying capac-
ity decreased with longer processing time and a higher
E/S ratio. This is attributable to low molecular weights
of protein hydrolysates produced by increasing the
processing time or the E/S ratio that are less ecient
in decreasing the wateroil interface tension because
they cannot unfold and reorient on the wateroil
surface to stabilise the emulsion system. Sathivel et al.
(2004) found that protein hydrolysates have to be
composed of a minimum length of 20 amino acids in
order to possess a good emulsifying capacity. In the
same manner as oil-binding capacity, protein hydroly-
sates from YTK generally possessed higher emulsifying
capacity than those from AS under the same process-
ing conditions. The highest emulsifying capacity was
produced from AS by Flavourzyme with a 30-min pro-
cessing time with a 0.5% E/S ratio (51 1.8 m2 g
1)
and by Neutrase under a 30-min processing time with
a 0.5% and 1.75% E/S ratio (53 0.7 m2 g
1); from
YTK by Flavourzyme under a 30-min processing time
with a 0.5% E/S ratio (53 1.7 m2 g
1), respectively,
without signicant dierence between these treatments.
The emulsifying capacity values of egg white pow-
der and cow gelatin powder, two food-grade emulsi-
ers in the market, were used as references, in order
to assess the possibility of developing sh protein
hydrolysates as food-grade emulsiers. Protein hydro-
lysates of AS and YTK that displayed emulsifying
capacity values within the range of 57 1.5 m2 g
1
(egg white powder) and 50 0.6 m2 g
1 (cow gela-
tin) could be obtained using a short processing time
and low E/S ratio, indicating their commercial
potential.
Optimisation of processing conditions by RSM
The aforementioned data showed that AS and YTK
protein hydrolysates could achieve or exceed market
values of oil-binding capacity, emulsifying capacity
and high protein recovery but under dierent
processing conditions. Because of this, a statistical
experimental design using RSM was conducted to
optimise processing conditions for producing sh
protein hydrolysates with optimised physicochemical
properties and protein recovery. YTK processing
co-products using Flavourzyme produced sh protein
hydrolysates with the highest values of oil-binding
capacity, emulsifying capacity and protein recovery in
this study. Therefore, experimental data related to this
combination were selected for optimisation purposes.
2012 The Authors
International Journal of Food Science and Technology 2012 Institute of Food Science and Technology
Optimize fish protein hydrolysates process S. He et al. 2401
Selected experimental data of protein recovery,
oil-binding capacity and emulsifying capacity of
protein hydrolysates under nine dierent processing
conditions were subjected to regression analysis using
second-order polynomial equations, and the results for
the linear, quadratic and interaction terms are
presented in Table S6. The R2 outcomes for enzymatic
processing yield, oil-binding capacity, emulsifying
capacity were all higher than 80%, which were judged
to be safe factorial ts that can be applied in RSM.
The variables of protein recovery, oil-binding
capacity and emulsifying capacity were considered in a
two-dimensional matrix to determine conditions of
optimum treatment. The lowest values of oil-binding
capacity and emulsifying capacity for optimisation
were chosen as 8.3 0.3 g oil g hydrolysates
1, the
oil-binding capacity of egg white powder (Table S4;
bold green line in Fig. S2), and 50 0.6 m2 g
1, emul-
sifying capacity of cow gelatin powder (Table S5; bold
green line in Fig. S1), as these are the values displayed
by the commercial/market standards. The highest
value of oil-binding capacity and emulsifying capacity
for optimisation was 14 1.2 g oil g hydrolysates
1
(Table S4; dash green line in Fig. S2) and
53 1.7 m2 g
1 (Table S5; dash green line in Fig. S1).
The data of oil-binding capacity and emulsifying
capacity were used to produce overlaid contour plots
against protein recovery, respectively, in order to
determine the processing conditions to optimum
functionalities with the highest protein recovery. The
analysis starts with the highest protein recovery of
about 85% and stepping down by 5% points until an
overlap is achieved. The white areas represent the
overlap of optimised processing condition of emulsify-
ing capacity against protein recovery in Fig. S1 and
oil-binding capacity against protein recovery in Fig. S2.
These two areas signify that protein hydrolysates from
YTK produced with Flavourzyme using 30- to 111-min
processing time with 0.61.3% E/S ratio can meet the
market standard of emulsifying capacity with a 55%
protein recovery, and using 25- to 64-min processing
time with 2.33.0% E/S ratio can meet the market
standard of oil-binding capacity with a 70% protein
recovery.
It can be seen that although protein recovery is com-
promised to 55% and 70% to achieve market standard
values of oil-binding capacity and emulsifying capacity,
respectively, it is still higher or comparable to protein
recoveries of other sh species. For example, Pacheco-
Aguilar et al. (2008) produced sh protein hydrolysates
from Pacic whiting, using Alcalase with an E/S ratio
of 1.0%, 1.5% and 3.0% with a processing time of 2 h
and found the protein recoveries were 48.6%, 58.6%
and 67.8%, respectively. Shahidi et al. (1995) produced
sh protein hydrolysates from Capelin using Neutrase
and Alcalase with an E/S ratio of 2% and a processing
International Journal of Food Science and Technology 2012
2402 Optimize fish protein hydrolysates process S. He et al.
time of 2 h and found protein recoveries were 51.6%
and 70.6%, respectively.
Finally, the results produced by modelling were
assessed by three validation experiments on processing
time and E/S ratio obtained from the optimised region
of Fig. S1 (Table S7) and Fig. S2 (Table S8). It was
found that both protein recovery and physicochemical
properties of protein hydrolysates produced under
these three validation experiments meet or exceed the
market standard values that were chosen for Fig. S1
and Fig. S2. This conrmed the aforementioned
determination.
Physicochemical property dierences in fish protein
hydrolysates with dierent MW ranges
Fish protein hydrolysates from YTK produced by
Flavourzyme were also selected for fractionation
because of the same reasons mentioned previously.
The electrophoretic patterns of sh protein
hydrolysates with dierent MW ranges are presented
in Fig. S3. It showed that these protein hydrolysates
generally t the designed MW ranges: >100 kDa,
10050 kDa, 5030 kDa and 3010 kDa and
<10 kDa. This assured the quality of raw materials for
the next step of determining the inuence of MW
ranges on physicochemical properties of YTK sh pro-
tein hydrolysates. While this study used membrane
fractionation on a small scale, we envisage this appli-
cation at the industrial scale, thereby facilitating the
industrial implementation. Membrane fractionation
has been used to separate milk protein (Brans et al.,
2004) and pea proteins (Mession et al., 2012) for
desired MW ranges.
Oil-binding capacity and emulsifying capacity of
protein hydrolysates with dierent MW ranges are
presented in Table S9. Protein hydrolysates with the
MW range of 3010 kDa presented the highest
oil-binding capacity and emulsifying capacity. Jeon
et al. (1999) also found that 30-K (permeate from
30 kDa) and 10-K (permeate from 10 kDa) sh pro-
tein hydrolysates of Cod showed excellent emulsifying
properties, comparing with 5-K (permeate from
5 kDa) and 3-K (permeate from 3 kDa). By compar-
ing the results of crude protein hydrolysates without
fractionation, it can be seen that the following two
values of 19 0.3 g oil g hydrolysates
1 for oil-bind-
ing capacity and 57 0.7 m2 g
1 for emulsifying
capacity were both higher than the highest oil-binding
capacity (14 1.2 g oil g hydrolysates
1) in Table S4
and emulsifying capacity (54 1.7 m2 g
1) in
Table S5. It is important to note that the emulsifying
capacity of this fraction was also higher than that of
egg white powder (57 1.5 m2 g
1), whereas none of
the crude sh protein hydrolysates in Table S5 could
reach this value without fractionation. These results
International Journal of Food Science and Technology 2012
International Journal of Food Science and Technology 2012 Institute of Food Science and Technology
signify that fractionation based on dierent MW
ranges is worthy of consideration to obtain sh pro-
tein hydrolysates with improved physicochemical
properties.
Conclusion
The results from this study demonstrated that the
protein hydrolysates from AS and YTK processing
co-products can possess oil-binding capacity and
emulsifying capacity that meet or exceed the market
standard values by an optimised enzymatic hydroly-
sis process. Protein hydrolysates from YTK process-
ing co-products produced with Flavourzyme were
identied possessing the highest values for oil-bind-
ing capacity, emulsifying capacity and protein recov-
ery. Response surface methodology identied the
optimal processing conditions were 30111 min with
the E/S ratio of 0.61.3% for emulsifying capacity
and 2564 min with the E/S ratio of 2.33.0% for
oil-binding capacity. These functionalities were
achieved with protein recoveries of 55% and 70%,
respectively. Furthermore, protein hydrolysates from
YTK processing co-products produced by Flavour-
zyme were fractionated into ve groups with dier-
ent molecular weight ranges of >100 kDa, 100
50 kDa, 5030 kDa, 3010 kDa and <10 kDa. The
10- to 30-kDa fraction demonstrated the highest
capacities of both oil-binding and emulsifying, which
cannot be achieved by any crude sh protein hydro-
lysates without fractionation. These results showed
the strong potential for utilising protein hydrolysates
from YTK processing co-products by Flavourzyme
as a commercial oil binder and emulsier. Based on
this, further studies should focus on applying these
protein hydrolysates into the formulation of commer-
cial food products.
Acknowledgments
The authors are grateful for the Australian Seafood
Cooperative Research Centre and Simplot Australia
for funding support and for awarding a Postgraduate
Scholarship to Shan He, and their advice and supply
of AS and YTK.
References
Arvanitoyannis, I.S. (2008). Waste Management for the Food Indus-
tries. London, UK: Academic Press (an imprint of Elsevier).
Arvanitoyannis, I.S. & Kassaveti, A. (2008). Fish industry wastes:
treatments, environmental impacts, current and potential uses.
International Journal of Food Science and Technology, 43, 726
745.
Arvanitoyannis, I.S. & Ladas, D. (2008). Meat waste treatment
methods and potential uses. International Journal of Food Science
and Technology, 43, 543559.
2012 The Authors

Bechtel, P.J. (2003). Advances in seafood byproducts: 2002
conference proceeding. Published by Alaska Sea Grant College
Program, University of Alaska Fairbanks. Pp. 114115.
Benjakul, S. & Morrissey, M.T. (1997). Protein hydrolysates from
Pacic whiting solid wastes. Journal of Agricultural and Food
Chemistry, 45, 34233430.
Brans, G., Schroen, C.G.P.H., Van der Sman, R.G.M. & Boom, R.
M. (2004). Membrane fractionation of milk: state of the art and
challenges. Journal of Membrane Science, 243, 263272.
Don, L.S.B., Pilosof, A.M.R. & Bartholomai, G.B. (1991).
Enzymatic modication of soy protein concentrates by fungal and
bacterial protease. Journal of the American Oil Chemists Society,
68, 102105.
Dong, F.M., Fairgrieve, W.T., Skonberg, D.I. & Rasco, B.A. (1993).
Preparation and nutrient analyses of lactic acid bacterial ensiled
salmon viscera. Aquaculture, 109, 351366.
Gbogouri, G.A., Linder, M., Fanni, J. & Parmentier, M. (2004).
Inuence of hydrolysis degree on the functional properties of
salmon byproducts hydrolysates. Journal of Food Science, 69, C615
C622.
He, S., Elisabeth, G., Stefan, K. & Andreas, L. (2010). Optimization
of hydrogen-peroxide washing of common carp Kamaboko using
Response Surface Methodology. LWT-Food Science and
Technology, 43, 765770.
He, S., Franco, C. & Zhang, W. (2011). Characterization of process-
ing wastes of Atlantic salmon (Salmo salar) and Yellowtail kingsh
(Seriola lalandi) harvested in Australia. International Journal of
Food Science and Technology, 46, 18981904.
Ian, K., Clive, S., Aravind, S. & William, A. (2004). Utilization of
seafood processing waste-challenges and opportunities. 3rd
Australian New Zealand Soils Conference, Published on CDROM.
Je, J.Y., Park, P.J., Kwon, J.Y. & Kim, S.K. (2004). A novel
angiotensin I converting enzyme inhibitory peptide from Alaska
pollock (Theragra chalcogramma) frame protein hydrolysates.
Journal of Agricultural and Food Chemistry, 52, 78427845.
Jeon, Y., Byun, H. & Kim, S. (1999). Improvement of functional
properties of cod frame protein hydrolysates using ultraltration
membranes. Process Biochemistry, 35, 471478.
Karim, A.A. & Bhat, R. (2009). Fish gelatin: proerties, challenges
and prospects as an alternative to mammalian gelatines. Food
Hydrocolloids, 23, 563576.
Klompong, V., Benjakul, S., Kantachote, D. & Shahidi, F. (2007).
Antioxidative activity and functional properties of protein hydroly-
sates of Yellow stripe trevally (Selaroides leptolepis) as inuenced
by the degree of hydrolysis and enzyme type. Food Chemistry, 102,
13171327.
Kristinsson, H.G. & Rasco, B.A. (2000). Fish protein hydrolysates:
producttion, biochemical and functional properties. Critical
Reviews in Food Science and Nutrition, 40, 4381.
Mession, J.L., Assifaoui, A., Cayot, P. & Saurel, R. (2012). Eect of
pea proteins extraction and vicilin/legummin fractionation on the
phase behavior in admixture with alginate. Food Hydrocolloides,
29, 335346.
Nilsang, S.N., Lertsiri, S., Suphantharika, M. & Assavanig, A.
(2005). Optimization of enzymatic hydrolysis of sh soluble con-
centrate by commercial protease. Journal of Food Engineering, 70,
571578.
Pacheco-Aguilar, R., Mazorra-Manzano, M.A. & Ramirez-Suarez, J.
C. (2008). Functional properties of sh protein hydrolysates from
Pacic whiting (Merluccius productus) muscle produced by a
commercial protease. Food Chemistry, 109, 782789.
Picot, L., Ravallec, R., Fouchereau-Peron, M. et al. (2010). Impact
of ultraltration and nanoltration of an industrial sh protein
hydrolysate on its bioactive properties. Journal of the Science of
Food and Agriculture, 90, 18191826.
Raghavan, S. & Kristinsson, H.G. (2008). Antioxidative ecacy of
alkali-treated Tilapia protein hydrolysates: a comparative study of
2012 The Authors
International Journal of Food Science and Technology 2012 Institute of Food Science and Technology
Optimize fish protein hydrolysates process S. He et al. 2403
ve enzymes. Journal of Agricultural and Food Chemistry, 56, 1434
1441.
Sathivel, S., Bechtel, P.J., Babbitt, J., Prinyawiwatkul, W.,
Negulescu, I.I. & Reppond, K.D. (2004). Properties of protein
powders from Arrowtooth ounder (Atheresthes stomias) and
Herring (Clupea harengus) byproducts. Journal of Agricultural and
Food Chemistry, 52, 50405046.
Shahidi, F., Han, X.Q. & Synowiechi, J. (1995). Production and
cheracteristics of protein hydrolysates from Capelin (Mallotus
villosus). Food Chemistry, 53, 285293.
Slizyte, R., Dauksas, E., Falch, E., Storro, I. & Rustad, T. (2005).
Characteristics of protein fractions generated from hydrolysed Cod
(Gadus morhua) by-products. Process Biochemistry, 40, 20212033.
Wasswa, J., Tang, J., Gu, X.H. & Yuan, X.Q. (2007). Inuence of
the extent of enzymatic hydrolysis on the functional properties of
protein hydrolysates from Grass carp (Ctenopharyngodon idella)
skin. Food Chemistry, 104, 16981704.
Supporting Information
Additional Supporting Information may be found in
the online version of this article:
Figure S1. Overlaid contour plot (white area) of pro-
tein recovery (red line) and emulsifying capacity (green
line) for protein hydrolysates of YTK co-products pro-
duced by Flavourzyme with protein recovery of 55%
(a) and 60% (b).
Figure S2. Overlaid contour plot (white area) of pro-
tein recovery (red line) and oil-binding capacity (green
line) for protein hydrolysates of YTK co-products pro-
duced by Flavourzyme with protein recovery of 70%
(a) and 75% (b).
Figure S3. SDS-PAGE pattern of protein hydroly-
sate fractions with dierent MW ranges from YTK
processing co-products produced by Flavourzyme.
Lane 1: YTK protein hyrolysates with MW range
of <10 kDa; Lane 2: YTK protein hyrolysates with
MW range of 1030 kDa; Lane 3: YTK protein hyrol-
ysates with MW range of 3050 kDa; Lane 4: YTK
protein hyrolysates with MW range of 50100 kDa;
Lane 5: YTK protein hyrolysates with MW range of
>100 kDa; Lane 6: Marker.
Table S1. Experimental design for two variables:
enzyme/substrate ratio and enzymatic processing time.
Table S2. Industry processing yields of sh process-
ing co-products from AS and YTK.
Table S3. Enzymatic processing yield* (g) and pro-
tein recovery of AS and YTK processing co-products
obtained with dierent processing conditions and dif-
ferent enzymes.
Table S4. Oil-binding capacity* (g oil g hydroly-
sates
1) of sh protein hydrolysates from AS and
YTK co-products after hydrolysation with dierent
processing conditions by dierent enzymes.
Table S5. Emulsifying capacity* (m2 g
1) of sh
protein hydrolysates from AS and YTK co-products
after hydrolysation with dierent processing conditions
by dierent enzymes.
International Journal of Food Science and Technology 2012
2404 Optimize fish protein hydrolysates process S. He et al.
Table S6. Regression coecients and R2 for three
variables of sh protein hydrolysates from YTK pro-
cessing co-products produced by Flavourzyme.
Table S7. Protein recovery and emulsifying capacity
of three validation experiments for RSM model.
Table S8. Protein recovery and oil-binding capacity
of three validation experiments for RSM model.
International Journal of Food Science and Technology 2012
International Journal of Food Science and Technology 2012 Institute of Food Science and Technology
Table S9. Oil-binding capacity* and emulsifying
capacity* of protein hydrolysate fractions from YTK
processing co-products produced by Flavourzyme.
Please note: Wiley-Blackwell are not responsible for
the content or functionality of any supporting materi-
als supplied by the authors. Any queries (other than
missing material) should be directed to the correspond-
ing author for the article.
2012 The Authors