Food Chemistry 134 (2012) 1297–1306

Contents lists available at SciVerse ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

The role of molecular size in antioxidant activity of peptide fractions from

Pacific hake (Merluccius productus) hydrolysates
Imelda W.Y. Cheung, Lennie K.Y. Cheung, Nina Y. Tan, Eunice C.Y. Li-Chan ⇑
The University of British Columbia, Faculty of Land and Food Systems, Food Nutrition and Health Program, 2205 East Mall, Vancouver, British Columbia, Canada V6T 1Z4

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

Article history: The antioxidative properties of Pacific hake hydrolysates and their peptidic fractions varying in molecular
Received 14 November 2011 size were assessed. Hydrolysates produced by different proteases (Alcalase, bromelain, Flavourzyme,
Received in revised form 16 January 2012 Protamex, Protease A‘‘Amano’’2, Protease N‘‘Amano’’K, Protin SD NY10, Umamizyme-K, Validase BNP-L,
Accepted 29 February 2012
Validase FPexo) generally possessed good metal ion chelating (33–73% at 3 mg/ml), DPPH radical scav-
Available online 8 March 2012
enging (18–30% at 1 mg/ml), ferric ion reducing power (abs700nm 0.36–0.86 at 3 mg/ml) and ABTS radical
scavenging (47–85% at 0.067 mg/ml) activity, as well as a good capability to suppress lipid peroxidation
Keywords:
in a linoleic acid model system. Peptide size (<1.4 kDa) was important for ABTS radical scavenging
Fish protein hydrolysates
Enzymatic hydrolysis
activity, whereas specific peptide composition (which depended on the particular protease used) was
Antioxidant activity the governing factor for effective lipid peroxidation. Validase BNP-L was the most promising enzyme
Antioxidative peptides for producing Pacific hake hydrolysates with good antioxidative activity in various assays and similar
Molecular weight distribution effectiveness as the synthetic antioxidant BHT to inhibit lipid peroxidation.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction hydrolysates to their bioactivity. These studies showed that low

molecular weight fractions in general contained more potent
Lipid oxidation is a prominent problem in food high in fat and antioxidative peptides (Alemán, Giménez, Pérez-Santín, Gómez-
during storage, oxidation products may be formed, leading to Guillén, & Montero, 2011; Je, Park, & Kim, 2005; Kim, Je, & Kim,
decreased nutritional quality, altered texture and production of 2007; Liu, Kong, Xiong, & Xia, 2010; Mendis, Rajapakse, Byun, &
off-flavours (Sarmadi & Ismail, 2010). Traditionally, synthetic anti- Kim, 2005; Nalinanon, Benjakul, Kishimura, & Shahidi, 2011) while
oxidants such as butylated hydroxytoluene (BHT) and butylated peptide composition has also been proposed to be the important
hydroxyanisole (BHA), or naturally occurring antioxidants such factor governing antioxidant activity, as evidently demonstrated
as vitamin C and a-tocopherol, have been widely used to prevent by the difference in antioxidative activity in hydrolysates produced
lipid oxidation in food products. As the general public becomes using different proteases (Alemán, Pérez-Santín, et al., 2011;
increasingly health conscious, favouring food products that are Bougatef et al., 2009, 2010; Dong et al., 2008; Raghavan &
natural, with potential health benefits, minimal processing and Kristinsson, 2008; Raghavan, Kristinsson, & Leeuwenburgh, 2008;
close to no artificial additives, investigators have begun to look Ren et al., 2008a, 2008b; Tang, Wang, & Yang, 2009). Moreover, a
for additional sources of natural antioxidants. The early research chemometric analysis performed by Udenigwe and Aluko (2011)
mainly focused on the extraction of polyphenols and flavonoids revealed a potential relationship between amino acid composition
from plant materials such as herbs and fruits (Moon & Shibamoto, and some antioxidative activities. Nevertheless, there is still a
2009); nevertheless, these polyphenolic compounds are often paucity of information on the possible effects of peptide size and
insoluble in an aqueous system and may impart additional colour composition, as influenced by the protease used in the hydrolysis,
and/or taste leading to poor consumer acceptability. As a result, on the different mechanisms of antioxidant action.
some researchers have started to evaluate the potential of other Therefore, in this study, we utilized Pacific hake (Merluccius
sources of antioxidants, including antioxidative peptides and productus), a low market value fish widely caught in the Pacific
amino acids that are liberated by the hydrolysis of proteins coast region, to produce protein hydrolysates using 10 different
(Samaranayaka & Li-Chan, 2011; Sarmadi & Ismail, 2010). proteases. The hydrolysates were evaluated for antioxidative
Several studies have looked at the contribution of molecular activity using various assays and in the linoleic acid peroxidation
size and structural characteristics of peptide mixtures in protein system, with the aim to select an enzyme that could produce the
most potent antioxidative hydrolysate for use in food systems.
⇑ Corresponding author. Tel.: +1 604 822 6182; fax: +1 604 822 5143. Furthermore, these hydrolysates were fractionated based on their
E-mail address: eunice.li-chan@ubc.ca (E.C.Y. Li-Chan). molecular size distribution and the fractions were assessed to

0308-8146/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodchem.2012.02.215
1298 I.W.Y. Cheung et al. / Food Chemistry 134 (2012) 1297–1306

investigate the possible influence of peptide size and composition slurry and an incubated control were prepared as two additional
on antioxidant properties. samples for assessment; the former was not incubated but imme-
diately boiled, while the latter was incubated without the addition
of exogenous proteases.
2. Materials and methods
A 5 ml aliquot was taken from the whole samples after
hydrolysis for the determination of pH and a-amino group content.
2.1. Materials
After 1-h incubation, the whole samples were immediately boiled
for 15 min, then centrifuged at 18,000g for 15 min at room
Pacific hake (M. productus) captured off the coast of Vancouver
temperature using a DuPont Sorvall Centrifuge RC 5B (Mandel
Island, Canada, was transported within 2–3 days on ice to the
Scientific Co. Ltd., Guelph, ON). The supernatants, collected after
laboratory for filleting and skinning, then kept at 25 °C until use.
passing through a double-layered cheese-cloth, constituted the
AlcalaseÒ 2.4 l (from Bacillus licheniformis, P2.4 AU/g), Flavour-
soluble fractions and are hereinafter referred collectively as soluble
zymeÒ 500 L (from Aspergillus oryzae, P500 AU/g), ProtamexÒ (from
samples, or as FPH (supernatant of fish protein hydrolysate), PFS
Bacillus sp., P1.5 AU/g), trinitrobenzenesulphonic acid (TNBS), 2,
(pasteurized fish slurry supernatant) and autolysate (supernatant
20 -azino-bis(-ethylbenzothiazoline-6-sulphonic acid) diammonium
of incubated control), while the pellets were collected as the insol-
salt (ABTS), (±)-6-hydroxy-2,5,7,8-tetramethylchromane-2-carbox-
uble fractions. These fractions were lyophilized for further analysis
ylic acid (trolox), ferric chloride, ferrous chloride tetrahydrate,
and the soluble yield was expressed as a percentage of the total
potassium ferricyanide(II), 3-(2-pyridyl)-5,6-diphenyl-1,2,4-
solid yield.
triazine-40 ,400 -disulphonic acid sodium salt (ferrozine), linoleic acid
(P99%), 3,5-di-tert-4-butylhydroxytoluene (BHT), butylated
2.4. Extent of hydrolysis
hydroxyanisole (BHA), blue dextran, vitamin B12 and glycine were
products from Sigma–Aldrich (St. Louis, MO) while bromelain (from
The content of a-amino groups in the whole samples after 1-h
pineapple stem, 2000 GDU/g) was from Ultra Bio-Logics Inc.
incubation, or after pasteurization without incubation in the case
(Montreal, QC). Potassium persulphate and trichloroacetic acid
of pasteurized fish slurry, was determined in triplicate by the TNBS
(TCA) were from Fisher Scientific (Fairlawn, NJ) and ethylenedi-
method according to Liceaga-Gesualdo and Li-Chan (1999), and
aminetetraacetic acid tetrasodium salt (EDTA) was a product of
was used to represent the extent of hydrolysis. A standard curve
BDH Inc. (Toronto, ON). Protease A‘‘Amano’’2, Protease N‘‘Amano’’K,
was constructed using 0.2–3.0 mM leucine, and the a-amino
Protin SD NY10 and Umamizyme-K were kindly donated by Amano
content of the whole samples was expressed as milliequivalents
Enzymes USA Co., Ltd. (Elgin, IL) while Validase BNP-L and Validase
of leucine per gram of dried mince (meq/g).
FPexo were gifts from Valley Research (South Bend, IN).

2.5. Antioxidant activities

2.2. Enzyme activity using casein as a substrate
Assays for the DPPH radical scavenging activity, ABTS radical
Ten commercial proteases were selected based on their scavenging activity, metal ion chelating activity and ferric ion
suitability to perform hydrolysis at the chosen experimental reducing antioxidant power were performed on FPH, autolysate
conditions of substrate pH of 6.58–6.94 and incubation temperature and PFS according to Samaranayaka and Li-Chan (2008) with the
of 50 °C. The enzymatic activity of the protease preparations as following modifications. The final assay concentrations for samples
declared by each supplier was based on different assay methodolo- used in the DPPH radical scavenging assay was 1 mg/ml. In the
gies and could not easily be compared. Moreover, these protease metal ion chelating activity assay, sample solution (1 ml) was
preparations had been acquired at different times for ongoing added to 3.7 ml distilled deionized water and 0.1 ml 2 mM FeCl2
research in our laboratory, and the duration as well as conditions before incubation in the dark at room temperature for 30 min.
of storage could adversely affect their enzymatic activity. Therefore, Ferrozine solution (0.2 ml, 5 mM) was added and incubated in
the activity of the 10 protease preparations was determined by a the dark for 10 min, giving a 3 mg/ml final assay concentration of
standard enzyme assay using casein as a substrate; in this assay, FPH, before taking the absorbance reading at 562 nm.
one unit of activity is represented by a production of colour
equivalent to 1.0 lmol of tyrosine per min at pH 7.5 at 37 °C (Sigma 2.6. Effect on lipid peroxidation in a linoleic acid model system
Quality Control Test Procedure, for Enzymatic Assay of Protease,
Casein as a Substrate. SSCASE0.1.001, revised 04/02/99). Lipid peroxidation in a linoleic acid model system was moni-
tored according to the method of Samaranayaka and Li-Chan
2.3. Production of fish protein hydrolysates (2008) with the following modifications. BHT, BHA, FPH, autolysate
and PFS were each prepared to a final assay concentration of 40 lg/
Frozen Pacific hake fillets were thawed at 4 °C overnight and ml. FPH, autolysate and PFS were dissolved in 0.1 M sodium
minced with Beem-Gigant Grinder Model Type F5–10 using a phosphate buffer at pH 7.0, while BHT and BHA were prepared in
4-mm screen (BEEM Blitz-Elektro-Erzeugnisse Manufaktur 50 mM linoleic acid in absolute ethanol. FPH, autolysate or PFS
Handels-GmbH, Rosbach v.d.H., Germany) at 4 °C. Fish fillet mince (1 ml) was mixed with 0.5 ml ddH2O and 1 ml 50 mM linoleic acid,
(100 g) was prepared into a slurry by addition of 200 ml distilled after which the mixture was incubated at 60 °C in the dark with
deionized water, and incubated for 1 h at 50 °C with 216 units of shaking. For the standard, BHT or BHA solution (1 ml) was added
enzyme based on the activity determined in Section 2.2. The se- to 1 ml phosphate buffer and 0.5 ml ddH2O, whereas the control
lected level of enzyme activity represents an enzyme:substrate ra- was comprised of 1 ml phosphate buffer, 1 ml 50 mM linoleic acid
tio equivalent to 3% w/w Validase BNP-L:protein; this enzyme was and 0.5 ml ddH2O. FPH, autolysate, PFS, BHT, BHA and control were
reported previously for producing antioxidative hydrolysates from prepared at least in triplicates. The extent of lipid peroxidation was
Pacific hake (Samaranayaka & Li-Chan, 2008). The hydrolysis peri- monitored using the ferric thiocyanate method described by
od was selected based on previous work with Pacific hake hydrol- Mitsuda, Yasumato, and Iwami (1966). Briefly, a 50 ll aliquot
ysates, which indicated that 1 h incubation yielded hydrolysates was taken periodically during the 8 day incubation period and
with comparable antioxidative activities to those incubated for a mixed with 2.35 ml of 75% ethanol, 50 ll of 30% ammonium
longer time (Samaranayaka & Li-Chan, 2008). Pasteurized fish thiocyanate and 50 ll of 20 mM FeCl2 in 3.5% HCl. The mixture
 I.W.Y. Cheung et al. / Food Chemistry 134 (2012) 1297–1306 1299

was incubated in the dark for 3 min before taking the absorbance In most studies reported in the literature, authors have usually
reading at 500 nm. compared two or more proteases based on the percent enzyme by
weight added to the substrates (Alemán, Pérez-Santín, et al., 2011;
Dong et al., 2008; Je, Lee, Lee, & Ahn, 2009; Jia et al., 2010; Mendis
2.7. Size exclusion chromatography
et al., 2005; Raghavan & Kristinsson, 2008; Raghavan et al., 2008;
Tang et al., 2009), while very few have studied the hydrolytic
Each FPH, autolysate or PFS was reconstituted in boiled and
ability of the proteases based on the same activity levels (Bougatef
filtered ddH2O (50 mg/ml), after which 1 ml of the solution was
et al., 2009, 2010; Ren et al., 2008a, 2008b). In the current study, it
filtered through an Acrodisc syringe filter with 0.45 lm HT Tuffryn
is evident from the results in Table 1 that equivalent weights of
membrane (VWR International, Mississauga, ON) and loaded onto
proteases could be associated with drastically different activity
a 1.75 cm diameter column packed with Sephadex G-25 media
levels. For example, the specific activity of Protease N‘‘Amano’’K
(Amersham Biosciences, Uppsala, Sweden) to a height of 100 cm
was 5.38 unit/mg, which was 7- and 40-fold higher than that
and a bed volume of 175 ml. The mobile phase was boiled and
observed for Validase BNP-L (0.72 unit/mg) and Flavourzyme
filtered ddH2O controlled at a flow rate of 1.0 ml/min. Fractions
(0.13 unit/mg), respectively (Table 1). The difference in activity
collected every 2 min, were monitored at 214 and 280 nm using
could be attributed to two causes: one being the inherent
a Shimadzu UV-160 UV–Visible Recording Spectrophotometer
difference in activity per mg of solids provided by the manufac-
(Mandel Scientific Company Inc., Guelph, ON). The pooled fractions
turer, while the other being the effect of storage time which might
were lyophilized for further analysis. The molecular weight
have reduced the proteolytic activity of the enzyme preparations.
markers used to calibrate the column included blue dextran
The enzymes investigated in this study were of bacterial, plant
(2,000,000 Da), antifreeze protein I (3240 Da), vitamin B12
or fungal origin, and each had an optimal pH and temperature
(1355.4 Da) and glycine (75.07 Da).
range falling within the experimental conditions of pH 7 and
50 °C pertinent for the current application (Table 1). The soluble
2.8. Antioxidant activity of size exclusion chromatographic fractions yields and extent of hydrolysis of the whole samples are shown
in Table 1. It is apparent that incubation at 50 °C for 1 h signifi-
The ABTS radical scavenging activity assay was performed cantly increased the product yield from 25% for pasteurized fish
according to procedures described in Section 2.5 with the slurry to 45% for the incubated control and over 58% for fish pro-
exception that FPH, autolysate and PFS were assayed at tein hydrolysates. Although Pacific hake contains some level of
0.033 mg/ml. Lipid peroxidation in a linoleic acid model system endogenous proteolytic activity, as demonstrated by the increase
was carried out as described in Section 2.6. in a-amino groups from 0.11 meq/g for pasteurized fish slurry to
0.88 meq/g for incubated control, the addition of exogenous
2.9. Statistical analysis enzymes further increased the extent of hydrolysis, yielding
a-amino group contents ranging from 1.54 to 3.46 meq/g (Table 1).
Analysis of Variance General Linear Model from version 16 of Nevertheless, there were only small differences in soluble product
Minitab statistical software (Minitab Inc., State College, PA) was yields among fish protein hydrolysates prepared using different
used to determine whether soluble samples and/or their molecular proteases in spite of their varying extent of hydrolysis as reflected
weight fractions had a significant effect on the antioxidant activi- by a-amino contents. The extent of hydrolysis usually increases as
ties at P < 0.05. The differences between FPH, autolysate and PFS, the proteolytic reaction progresses, since proteins or polypeptides
and their molecular weight fractions were analyzed using Tukey’s are being cleaved into shorter peptides leading to an increase in
test at P < 0.05. a-amino groups (Bamdad, Wu, & Chen, 2011). Some of the protease
preparations, such as Flavourzyme, Protease A‘‘Amano’’2, Umami-
zyme-K and Validase FP exo, contain both exopeptidases and
3. Results and discussion endoproteases, thus producing hydrolysates with peptides as well
as many free amino acids, each of which would lead to an increase
3.1. Enzymatic hydrolysis of fish protein hydrolysates in the a-amino content. Both soluble peptides and free amino acids
would contribute to the soluble yield. Although the content of
In this study, 10 proteases were screened for their potential to a-amino groups (and the extent of hydrolysis) would be expected
produce antioxidative fish protein hydrolysates from Pacific hake to be higher in hydrolysates with high contents of free amino acids,
within 1 h of proteolysis at 50 °C and the unadjusted substrate the yield would not necessarily differ from hydrolysates composed
pH (6.58–6.94). The characteristics of these proteases are listed primarily of soluble peptides.
in Table 1. The potential for endogenous proteolytic action and
1-h incubation to yield products with antioxidant activity was 3.2. Antioxidant activity of FPH
investigated by examining the properties of the incubated control
and pasteurized fish slurry, respectively. In the current study, FPH, autolysate and PFS were examined for
In order to be able to objectively compare the ability of the their ability to protect against oxidation by several mechanisms,
different protease products to generate hydrolysates with antioxi- including DPPH radical scavenging, ABTS radical scavenging, metal
dative activity, equivalent units of enzyme activity should be used ion chelating and ferric ion reducing properties. In general, FPH
for hydrolysate production. The hydrolytic activities of the 10 had antioxidant activities comparable to, if not better than, the
protease products supplied by different manufacturers were synthetic standards trolox and EDTA (Table 2).
reported in several different units (e.g. Anson unit for Alcalase, The DPPH radical scavenging activity of FPH was comparable to
Flavourzyme and Protamex, gelatin digesting unit for bromelain, those reported in the literature (Bamdad et al., 2011; Bougatef et al.,
etc.), making comparison of their proteolytic activity difficult. 2009; Jia et al., 2010; Liu et al., 2010; Nalinanon et al., 2011). FPH-U
Moreover, storage of the protease preparations prior to use in this showed 30% DPPH radical scavenging activity, close to 32% exhib-
study could have resulted in their actual enzymatic activity being ited by 15 lM trolox, while FPH-P and FPH-Protin only showed
lower than the declared activity. The activity of the 10 proteases 19% and 18%, respectively. Unexpectedly, PFS and autolysate had
was therefore determined by a standard enzyme assay using casein relatively high activities (26%), suggesting that extensive hydrolysis
as a substrate, as shown in Table 1. might not be necessary to release peptides with DPPH radical
1300 I.W.Y. Cheung et al. / Food Chemistry 134 (2012) 1297–1306

Table 1
Characteristics and activity of proteases used for hydrolysis and the soluble yield and extent of hydrolysis of the whole samples.a

Code Protease Whole samples

Name Origin Optimal pH Optimal Activity Soluble Extent of
temperature (°C) (units/mg)b yield (%)c hydrolysis (meq/g)d
A Alcalase Bacillus licheniformis 6–8 50–60 0.59 64.86 1.54
B Bromelain Pineapple stem 5–8 45–60 0.35 69.10 1.63
F Flavourzyme Aspergillus oryzae 5.5–7.5 50–55 0.13 62.96 1.90
P Protamex Bacillus 5.5–7.5 35–60 0.62 65.22 1.73
AA2 Protease A‘‘AMANO’’2 Aspergillus oryzae 7–8 50–55 0.94 63.10 2.04
NAK Protease N‘‘AMANO’’K Bacillus subtilis 7–8 50–60 5.38 66.63 2.02
Protin Protin SD NY10 Bacillus subtilis 7 50–55 0.36 64.30 3.09
U Umamizyme-K Aspergillus oryzae 7–8 45–50 1.19 67.46 1.76
VBNP Validase BNP-L Bacillus subtilis var 6.5–7.5 50–55 0.72 66.53 1.56
VFPexo Validase FP Exo Aspergillus oryzae 6–7 50–55 0.38 58.64 3.46
Incubated control Endogenous – – – – 45.06 0.88
Pasteurized fish – – – – – 24.90 0.11
slurry
a
Whole samples refer to fish slurry immediately pasteurized (pasteurized fish slurry) or fish slurry pasteurized after incubation for 1 h (fish protein hydrolysates and
incubated control).
b
Determined by Sigma’s enzymatic assay using casein as a substrate. One unit represents the liberation of 1.0 lmol tyrosine per min from the hydrolysis of casein at pH 7.5
and 37 °C.
c
Expressed as 100% * gram of soluble solids/gram of total solids.
d
Expressed as a-amino groups in milliequivalent leucine per gram dried mince, measured using TNBS method.

scavenging activity. In a study on catfish hydrolysate, it was also study for FPH-F, FPH-AA2, and FPH-NAK, but in contradiction to the
found that hydrolysate with 5% degree of hydrolysis (DH) had a findings of Dong et al. (2008), who reported higher metal ion chelat-
higher DPPH radical scavenging activity than that with 30% DH ing activity in silver carp hydrolysate made with Alcalase than with
(Theodore, Raghavan & Kristinsson, 2008). Similarly, Bamdad Flavourzyme. Some studies have reported that higher DH, and thus
et al. (2011) found that a critical size was necessary for peptidic lower molecular weights, resulted in increased metal ion chelating
fractions to exhibit potent DPPH activity in barley hordein ability (Alemán, Pérez-Santín, et al., 2011; Dong et al., 2008;
hydrolysate. On the contrary, many studies have suggested that Nalinanon et al., 2011), while others proposed high molecular
high DH and low molecular weights were found to correlate well weights for potent metal ion chelators, which could form a caged
with DPPH radical scavenging activity (Bougatef et al., 2009; Liu structure for metal ion entrapment (Bamdad et al., 2011).
et al., 2010; Phanturat, Benjakul, Visessanguan, & Roytrakul, 2010; Ferric ion reducing antioxidant power measures the ability of the
Raghavan & Kristinsson, 2008; Udenigwe & Aluko, 2011). The potential antioxidant to reduce pro-oxidant metal ions. In this
discrepancy of these findings could be explained by the differences study, all soluble samples showed a strong ability to reduce ferric
in substrate, protease type as well as the enzyme-to-substrate ratio, ions to ferrous ions, with PFS showing the highest reducing power,
all of which will directly affect the extent of hydrolysis. In addition, while FPH-P showed the lowest reducing potential (Table 2). Strong
‘‘high DH’’ has been defined differently with some studies consider- ferric ion reducing power of protein hydrolysates was also reported
ing 8% as being ‘‘high’’ while others reported 30% as ‘‘high’’. by both Je et al. (2009) and Jia et al. (2010). Contradictory results
FPH generally demonstrated an excellent ability to scavenge have been reported with regard to the influence of molecular size
ABTS radicals, while PFS showed weaker activity (Table 2). Even on ferric ion reducing power. An increase in DH and hence lower
though the autolysate already displayed stronger ABTS radical molecular weight products has been suggested to decrease reducing
scavenging activity than 16.7 lM trolox, hydrolysis with the power of protein hydrolysates of catfish (Theodore et al., 2008) and
addition of exogenous proteases further increased this antioxidative barley hordein (Bamdad et al., 2011), which is in agreement with the
activity for most of the FPH. The potent ABTS radical scavenging findings of the present study that PFS and autolysate, the samples
activity of FPH, independent of protease type, suggests that with the lowest contents of a-amino groups, had the highest
molecular size of the peptides might be the key contributor to this ferric ion reducing power. Nevertheless, other studies have found
antioxidant activity. Increase in DH has been reported to increase higher reducing power with increasing DH (Alemán, Giménez,
ABTS radical scavenging activity in gelatin hydrolysate (Phanturat et al., 2011; Bougatef et al., 2010; Liu et al., 2010; Phanturat et al.,
et al., 2010) and tilapia hydrolysates (Raghavan et al., 2008). 2010; Raghavan et al., 2008). The inconsistency found in these
Nevertheless, Alemán, Giménez, et al. (2011) found lower activity reports suggests that size might not be the most important
in the lower molecular weight fractions than the unfractionated contributor to ferric ion reducing power. A positive relationship
<10 kDa fraction. The contradictory results suggest that other between sulphur-containing, acidic and hydrophobic amino acids
factors, such as substrate and peptide composition might also influ- with ferric ion reducing ability was postulated by Udenigwe and
ence the ABTS radical scavenging activity and should not be Aluko (2011).
overlooked. Overall, all FPH in this study displayed antioxidative properties
A range of metal ion chelating activity was observed for with a potential to decrease or inhibit oxidation reactions. While
hydrolysates produced by the different protease treatments. PFS demonstrated good DPPH radical scavenging activity, metal
Moreover, all FPH as well as PFS in the present study displayed as ion chelating ability and strong ferric ion reducing antioxidant
high or higher metal ion chelating ability at lower sample concen- power, it was poor at scavenging ABTS radicals. On the other hand,
tration than those reported for squid gelatin hydrolysates (Alemán, FPH-P was weak in DPPH radical scavenging activity, metal ion
Pérez-Santín, et al., 2011) and porcine plasma protein hydrolysate chelating ability and ferric ion reducing antioxidant power, but it
(Liu et al., 2010). Raghavan and Kristinsson (2008) found that tilapia had strong ABTS radical scavenging activity. These results together
protein hydrolysates made using Protease A‘‘Amano’’2 and illustrated that each mechanism possibly relies on different
Flavourzyme had better metal ion chelating ability than that using molecular size and/or peptide composition and sequence to give
Protease N‘‘Amano’’K, similar to the trends observed in the current the specific antioxidant activity.
 I.W.Y. Cheung et al. / Food Chemistry 134 (2012) 1297–1306 1301

Table 2
DPPH radical scavenging activity, ABTS radical scavenging activity, metal ion chelating activity and ferric ion reducing antioxidant power of the soluble fractions of fish protein
hydrolysates (FPH), incubated control (autolysate) and pasteurized fish slurry (PFS).

Samplea DPPH radical ABTS radical Metal ion Ferric ion reducing
scavenging activity scavenging activity chelating activity antioxidant power
(%)b,f (%)c,f (%)d,f (Abs at 700 nm)e,f
FPH-A 27.5 ± 0.3 79.6 ± 0.5 51.5 ± 0.2 0.477 ± 0.016
FPH-B 28.2 ± 0.3 83.2 ± 0.3 42.5 ± 1.1 0.453 ± 0.005
FPH-F 23.7 ± 0.1 81.2 ± 0.4 65.9 ± 1.9 0.447 ± 0.021
FPH-P 19.3 ± 0.5 82.7 ± 0.4 33.2 ± 1.3 0.362 ± 0.011
FPH-AA2 27.2 ± 0.6 80.9 ± 0.1 65.9 ± 1.1 0.537 ± 0.008
FPH-NAK 26.3 ± 0.3 83.5 ± 0.6 54.0 ± 0.1 0.536 ± 0.004
FPH-Protin 18.0 ± 0.5 81.6 ± 0.4 64.5 ± 2.6 0.470 ± 0.011
FPH-U 30.2 ± 0.3 84.1 ± 0.4 56.8 ± 0.9 0.476 ± 0.031
FPH-VBNP 28.2 ± 0.4 82.4 ± 0.6 58.0 ± 0.5 0.518 ± 0.016
FPH-VFPexo 22.1 ± 0.2 78.0 ± 0.9 67.5 ± 1.1 0.648 ± 0.025
Autolysate 26.1 ± 1.0 74.3 ± 1.0 50.1 ± 2.2 0.655 ± 0.018
PFS 26.2 ± 0.4 47.4 ± 0.0 73.2 ± 1.0 0.856 ± 0.018
Standard 32.4 ± 0.2 69.7 ± 0.0 80.2 ± 0.4 0.077 ± 0.004
a
Sample code was assigned according to information listed in Table 1. For example, FPH-A represents the soluble fraction of fish protein hydrolysate produced by Alcalase.
b
Samples were assayed at 1 mg/ml and standard used was 15 lM trolox.
c
Samples were assayed at 0.067 mg/ml and standard used was 16.7 lM trolox.
d
Samples were assayed at 3 mg/ml and standard used was 25 lM EDTA.
e
Samples were assayed at 3 mg/ml and standard used was 15 lM trolox.
f
Values were reported as mean ± standard deviation where n = 3.

3.3. Inhibition of lipid peroxidation in a linoleic acid model system hydrolysate, produced using Alcalase, was also shown to be more
effective in the linoleic acid peroxidation system than hydrolysate
Assays measuring DPPH radical scavenging, ABTS radical made with Flavourzyme (Dong et al., 2008), in agreement with
scavenging, metal ion chelating and ferric ion reducing power have what was found in this study (that in comparison to FPH-B and
been widely used to assess the antioxidative potential of phenolics, FPH-F, FPH-A was a more effective inhibitor of lipid peroxidation
flavonoids and recently, peptide mixtures. However, these assays in the model system).
each measure an antioxidant property representing a different When the results from the linoleic acid model system were
mechanism, which may not reflect the multiple mechanisms by compared to those of the various individual chemical assays of
which samples may act as antioxidants to retard or inhibit lipid antioxidant activity discussed in Section 3.2, no consistent trends
oxidation in a food system. Therefore, in this section, the ability could be observed, suggesting that different specific structural
of the soluble samples to suppress lipid peroxidation in a linoleic requirements are associated with each mechanism of antioxidant
acid model system was investigated. action. PFS showed good DPPH radical scavenging, excellent metal
All the soluble samples demonstrated an ability to delay linoleic ion chelating and ferric ion reducing antioxidant power, but
acid peroxidation and many of them showed similar effectiveness relatively low ABTS radical scavenging activity, yet it was the least
to BHT (Fig. 1) as well as BHA (data not shown) when monitored effective antioxidant among all in the linoleic acid peroxidation
over 8 days. Even though PFS displayed the weakest protection system. On the other hand, FPH-P, which had weak DPPH radical
against oxidation of linoleic acid, in comparison to the control scavenging activity, metal ion chelating and reducing ability but
without the addition of potential antioxidant, it still delayed the strong ABTS radical scavenging activity, was effective against lipid
onset of lipid oxidation (Fig. 1). FPH and autolysate offered similar peroxidation in the model system. This lack of congruence between
or better protection against linoleic acid peroxidation compared to the results for antioxidative activity from the different assays and
what has been reported for sardinelle by-product hydrolysates the in vitro lipid peroxidation model system was also observed
(Bougatef et al., 2010), silver carp hydrolysates (Dong et al., for hydrolysates from sardinelle by-products hydrolysates, where
2008), Alaska pollack frame hydrolysate (Je et al., 2005), hoki frame samples showing good reducing power, but poor DPPH scavenging
protein hydrolysate (Kim et al., 2007), jumbo squid skin gelatin activity, only showed 15% inhibition of lipid peroxidation after
hydrolysates (Mendis et al., 2005) and pea seed hydrolysate 5 days at 40 °C (Bougatef et al., 2010). Similarly, pea seed hydroly-
(Pownall, Udenigwe, & Aluko, 2010), especially considering the sate, which exhibited strong metal ion chelating activity yet poor
lower hydrolysate concentration (0.040 mg/ml) and harsh to moderate DPPH, OH, O2 and H2O2 radical scavenging and
temperature of incubation (60 °C) used in the current study reducing ability, was excellent in inhibiting lipid peroxidation over
compared to the other reports. 7 days at 60 °C (Pownall et al., 2010).
Some studies reported that high DH and thus low molecular
weight peptides were more effective against linoleic acid peroxida- 3.4. Molecular weight distribution by size exclusion chromatography
tion (Je et al., 2005; Ren et al., 2008a). In the current study, FPH-F,
FPH-VFPexo, FPH-U, FPH-AA2, as well as the autolysate, were To date, there is still a lack of evidence to explain the relation-
slightly less effective than other hydrolysates in inhibiting ship between characteristic structural properties of peptides, such
oxidation, as monitored by the higher absorbance at 500 nm. These as molecular size, hydrophobicity, amino acid composition and
particular FPH samples have in common the fact that they were sequence, and their antioxidative property specific to each
prepared using proteases comprising endoproteases and exopep- mechanistic action.
tidases; it is possible that amino acid residues at the terminal In order to better understand what contributes to the observed
end of the peptides, which may be important for the antioxidant differences between the samples in their antioxidative activity
ability, were cleaved off by the exopeptidases. Ren et al. (2008b) measured by the different assays and in the linoleic acid peroxida-
found better protection against lipid peroxidation for grass carp tion system, each soluble sample was fractionated using size
hydrolysates made with Alcalase than bromelain, while silver carp exclusion chromatography. Size, especially in the range of
1302 I.W.Y. Cheung et al. / Food Chemistry 134 (2012) 1297–1306

0.5–3 kDa, has been suggested to be a crucial factor affecting the linoleic acid model system were selected to examine the effect of
antioxidant activity of protein hydrolysates (Bamdad et al., 2011; size on these antioxidant activities.
Je et al., 2005; Kim et al., 2007; Nalinanon et al., 2011; Phanturat
et al., 2010; Samaranayaka & Li-Chan, 2011) and consequently, in 3.5.2. ABTS radical scavenging activity
this study, three fractions were collected for further determination The ABTS radical scavenging activity of unfractionated FPH,
of antioxidative activities. autolysate, PFS, and their molecular weight fractions (I, II and III)
The elution profiles of FPH, autolysate and PFS are presented in are compared in Table 3. In general, fraction II had similar or lower
Fig. 2, with three fractions I, II and III representing the approximate ABTS radical scavenging activity than fraction I, which in turn
molecular weight ranges of 2000–3600, 1400–2000, and <1400 Da, showed weaker activity than the unfractionated counterpart, while
respectively. In general, all soluble samples had a wide spectrum of fraction III displayed the most potent ABTS radical scavenging
materials with molecular weight <3600 Da, as monitored by activity. The only exception was observed in PFS, where all the
absorbance at 214 and 280 nm. PFS, which was the soluble fraction fractions had stronger activity than the unfractionated counter-
obtained from fish slurry heated immediately to inactivate endog- part, suggesting a potential antagonistic effect of the peptides
enous enzymatic activity, had a very distinct elution profile show- when simultaneously present in PFS, thus reducing the radical
ing two dominant peaks, one around the molecular size above scavenging activity.
3000 Da and the other below 1500 Da. These peaks suggest that Although fraction III had the lowest yield of all fractions, its
PFS consisted primarily of peptides with large molecular weight, ABTS radical scavenging ability was approximately double those
with a relatively smaller amount of materials under 1000 Da observed in fractions I and II, further confirming oligopeptides with
(Fig. 2). In comparison, the profile of the autolysate, resulting from sizes smaller than 1400 Da to be the important contributors to
incubation with endogenous enzymes in the fish mince, had a ABTS radical scavenging activity. Some variations in ABTS radical
decrease in the size of the peak containing peptides larger than scavenging activity were observed in fraction III from the different
3000 Da and an increase in materials with molecular weight below FPH, suggesting that peptide composition had some effect on the
2500 Da. Moreover, FPH, which were produced by addition of antioxidant activity; however, it is clear that the small molecular
exogenous proteases, all showed materials with size below size below 1400 Da is the dominant factor contributing to the
3000 Da as the major components, with an increase in peptides potent ABTS radical scavenging activity in Pacific hake protein
eluting in region III of the chromatogram (Fig. 2). Similar elution hydrolysates.
profiles were observed among FPH, with the exception of the
increase in peaks below 1000 Da for FPH-F, FPH-AA2, FPH-U and
3.5.3. Inhibition of lipid peroxidation in a linoleic acid model system
FPH-VFPexo, which all had relatively bigger peaks beyond 200 ml
Fig. 3 shows the progression of linoleic acid peroxidation for
elution volume, suggesting that more free amino acids were
FPH, autolysate and PFS, and their molecular weight fractions over
released by the exopeptidase activity in these enzyme
8 days of incubation at 60 °C with constant shaking. Unlike ABTS
preparations. Similar results were reported by Tang et al. (2009),
radical scavenging activity, each sample behaved differently in
in which hemp protein hydrolyzed with Flavourzyme had more
the linoleic acid peroxidation system, and the most effective
peptides with low molecular weight than Alcalase and Protamex.
fraction for suppressing lipid oxidation varied among the samples.
Numerous studies have investigated the effect of molecular size
PFS, which had the weakest ABTS radical scavenging activity
on bioactivity through fractionation by ultrafiltration (Je et al., 2005;
amongst all samples, also did not provide good protection against
Kim et al., 2007; Liu et al., 2010; Mendis et al., 2005) or size exclusion
lipid peroxidation. In fact, the unfractionated PFS promoted
chromatography (Alemán, Giménez, et al., 2011; Bougatef et al.,
oxidation in comparison to the control without the addition of
2009, 2010; Nalinanon et al., 2011; Phanturat et al., 2010) but very
potential antioxidants (Fig. 3). While fractions I and II of PFS were
few aimed to study how the use of different proteases might
not better than the control, fraction III was able to exert some
produce hydrolysates containing different antioxidative activity
inhibitory activity against linoleic acid peroxidation although it
for understanding structural requirements. In the next section, the
was not as good as the action of BHT (Fig. 3). On the other hand,
antioxidative potential of the three fractions (I, II and III shown in
fraction I of the autolysate had no antioxidant ability in the linoleic
Fig. 2) was studied to provide insight into whether molecular size
might play a role in determining bioactivity.

3.5. Antioxidative activity of fractions from size exclusion

chromatography

3.5.1. Fraction yield

The relative yields of fractions I, II and III were 68%, 19% and
13%, respectively, for PFS and 57%, 34% and 9%, respectively, for
the autolysate (Table 3), indicating the action of endogenous
proteases on the proteins and long chain polypeptide molecules
of the fish mince to give rise to a higher content of medium-sized
peptides in the autolysate. FPH-NAK and FPH-Protin, despite
having a more extensive hydrolysis than PFS and the autolysate,
did not give higher yields in fraction III, while the other FPH all
had a yield of above 15% in fraction III. Several studies have also
reported different yields of various molecular weight fractions
demonstrating the difference in size distribution of protein
hydrolysates produced using different proteases and substrates
(Bamdad et al., 2011; Dong et al., 2008; Jia et al., 2010).
Due to the limited amount of samples and considering the Fig. 1. Progression of lipid peroxidation in a linoleic acid model system, averaged
results shown in Sections 3.2 and 3.3, ABTS radical scavenging among three replicates, as a function of time for FPH, autolysate, PFS, BHT (assay
activity and inhibitory activity against lipid peroxidation in the concentration at 0.040 mg/ml) and control (without the addition of antioxidants).
 I.W.Y. Cheung et al. / Food Chemistry 134 (2012) 1297–1306 1303

Fig. 2. Molecular weight distribution profile of FPH, autolysate and PFS by Sephadex G-25 column chromatography. The absorbance was monitored at 214 nm ( for
undiluted and for 10-fold dilution) and 280 nm ( ). The estimated molecular weights of peptides in Fractions I, II and III were 2000 – 3600, 1400 – 2000 and <1400 Da,
respectively, based on the elution of the molecular markers blue dextran (2,000,000 Da), antifreeze protein I (3240 Da), vitamin B12 (1355.4 Da) and glycine (75.07 Da) at 86,
90, 176 and 192 ml, respectively, as depicted by the arrows.

Table 3
Fraction yield from size exclusion chromatography (shown in Fig. 2) and ABTS radical scavenging activity of each fraction and the unfractionated counterpart.

Samplea Unfractionated Fraction I Fraction II Fraction III

activity (%)b
Yield Activity Yield Activity Yield Activity
(%)c (%)b (%)c (%)b (%)c (%)b
FPH-A 51.4 ± 0.9 41 41.5 ± 0.8 42 41.0 ± 0.4 17 83.6 ± 0.4
FPH-B 54.9 ± 0.2 42 44.6 ± 1.0 40 44.7 ± 0.4 18 81.5 ± 0.4
FPH-F 51.4 ± 0.3 44 41.1 ± 0.6 37 36.8 ± 0.7 19 78.7 ± 0.2
FPH-P 54.4 ± 0.8 36 41.6 ± 1.0 39 37.5 ± 0.4 25 75.8 ± 0.7
FPH-AA2 56.2 ± 1.3 39 39.8 ± 1.0 43 38.4 ± 0.4 18 85.4 ± 0.2
FPH-NAK 55.2 ± 0.3 51 41.7 ± 0.8 39 36.6 ± 0.7 10 90.9 ± 0.2
FPH-Protin 54.0 ± 0.3 44 46.7 ± 0.8 45 40.4 ± 1.0 11 89.6 ± 0.1
FPH-U 57.3 ± 1.4 32 46.7 ± 1.7 41 38.1 ± 0.5 27 80.0 ± 0.6
FPH-VBNP 56.7 ± 0.8 40 50.6 ± 1.2 40 44.3 ± 0.7 20 82.2 ± 0.3
FPH-VFPexo 53.0 ± 0.7 43 44.7 ± 1.6 40 38.9 ± 0.4 17 83.6 ± 0.0
Autolysate 44.5 ± 0.5 57 45.5 ± 0.3 34 33.6 ± 0.5 9 72.6 ± 0.7
PFS 22.3 ± 0.2 68 31.2 ± 0.2 19 29.3 ± 1.2 13 42.1 ± 1.3
a
Sample code was assigned according to information listed in Table 1. For example, FPH-A represents the soluble fraction of fish protein hydrolysate produced by Alcalase.
b
Samples were assayed at 0.033 mg/ml. Values were reported as mean ± standard deviation where n = 3.
c
Expressed as 100% * gram of fraction/gram of total solids.
1304 I.W.Y. Cheung et al. / Food Chemistry 134 (2012) 1297–1306

Fig. 3. Progression of lipid peroxidation in a linoleic acid model system, averaged among a minimum of two replicates, as a function of time for FPH, autolysate and PFS, and
their size exclusion chromatography fractions at assay concentration of 0.040 mg/ml ( unfractionated (U), fraction I, fraction II, fraction III, s control (C), without the
addition of antioxidants and x BHT (B)).

acid peroxidation system, and fraction III only slightly delayed the the most effective fraction. Both fractions II and III of FPH-Protin
onset of lipid oxidation, whereas the unfractionated autolysate in were effective, with fraction II being able to suppress lipid peroxi-
addition to fraction II were both able to reduce lipid peroxidation, dation similarly to BHT. Fractions II and III of FPH-P both displayed
with fraction II being similar to BHT. FPH-AA2 did not appear to similar effectiveness against lipid peroxidation in comparison to
show any fraction with a better inhibitory effect than the unfrac- BHT. Furthermore, fraction I of these four hydrolysates all
tionated counterpart, leading to the speculation that the different displayed significantly weaker inhibitory activities against lipid
molecular weight components acted synergistically in the unfrac- peroxidation.
tionated hydrolysate. Nevertheless, the inhibitory activity of On the other hand, FPH-F, FPH-NAK, FPH-U and FPH-VFPexo
unfractionated FPH-AA2 was not better than unfractionated and their different molecular weight fractions all showed good
autolysate, indicating that addition of Protease A‘‘Amano’’2 to inhibitory activities against lipid peroxidation (Fig. 3). Many of
hydrolyze Pacific hake did not improve the antioxidant ability these samples had fraction III as the least effective in the linoleic
compared to that achieved by the endogenous proteases. acid model system, in contrast to the strong activity displayed by
The different molecular weight fractions of FPH-A, FPH-B, FPH-P fraction III of FPH-B and FPH-A, further suggesting that peptide
and FPH-Protin showed different degrees of protection against composition is a major factor contributing to antioxidative activity
lipid peroxidation (Fig. 3). For FPH-A and FPH-B, fraction III was in the linoleic acid peroxidation system. Since FPH-F, FPH-U and
 I.W.Y. Cheung et al. / Food Chemistry 134 (2012) 1297–1306 1305

FPH-VFPexo were produced using proteases comprising both 4. Conclusions

endo- and exo-peptidase activity, fraction III in these particular
hydrolysates may have contained peptide sequences in which the The results of this study showed that Pacific hake fillet mince
exopeptidases had cleaved terminal amino acid residues important could be used to produce hydrolysates with potent antioxidative
for the antioxidant activity of small molecular weight peptides in peptides, especially after hydrolysis with exogenous proteases.
the lipid peroxidation model system. In addition, they could The fish protein hydrolysates not only had good DPPH and ABTS
contain a significant amount of free amino acids, which have radical scavenging activity, metal ion chelating activity and ferric
been reported to be less effective antioxidants than peptides ion reducing antioxidant power, but also showed at least 50%
(Samaranayaka & Li-Chan, 2011). inhibition of lipid peroxidation. Furthermore, oligopeptides with
Although small molecular size has often been associated with sizes <1400 Da were found to be major contributors to the ABTS
high antioxidant activity, there is a lack of evidence to suggest radical scavenging activity with little dependence on peptide
which molecular weight range encompasses the most potent composition, while the latter appears to be a dominant factor
peptides. Some researchers reported peptides below 1500 Da as dictating antioxidative properties in the linoleic acid peroxidation
containing potent antioxidative activity (Je et al., 2005), while model system. Among the proteases investigated in the current
some suggested 1000–3000 Da to be the effective size (Kim et al., study, Validase BNP-L was found to be the most effective in produc-
2007) and others reported <3000 Da to be the potent fraction ing fish protein hydrolysates with good reducing power, metal ion
(Mendis et al., 2005; Ren et al., 2008a). In addition, Wu, Chen, chelating ability and excellent DPPH and ABTS radical scavenging
and Shiau (2003) found that mackerel hydrolysate peptides of activity. In addition, fish protein hydrolysate produced by Validase
molecular weights around 1400 Da exhibited much stronger BNP-L, whether as a whole or fractionated by size, was as effective
inhibition against lipid oxidation than smaller molecular weight as BHT and BHA in suppressing lipid peroxidation during 8 days of
fractions at 900 and 200 Da. This discrepancy indicates that incubation at 60 °C. Therefore, it is speculated that Validase BNP-L
size might not be the most important factor leading to inhibitory would be a promising protease to be used to produce antioxidative
activity of peptides in the linoleic acid peroxidation system and hydrolysates to be applied in food products. Further experiments
in fact, the specific peptide composition and properties might play should be conducted to verify the potential of the fish protein
a more crucial role, as evidently shown by the results of the hydrolysates so produced to serve as an antioxidant to delay or
present study. suppress lipid oxidation in a food matrix such as fish, and to
Most of the reports that have documented more effective determine their effect, if any, on the sensory quality of the resulting
inhibition against linoleic acid peroxidation by lower molecular food product.
weight fractions have only studied either one or at most two
proteases concurrently, and thus did not consider the potential Acknowledgements
impact of variable peptide composition generated by different
protease treatments as an additional factor influencing the antiox- The funding for this research was provided by the Natural
idative activity. In the present study, three fractions obtained by Sciences and Engineering Research Council of Canada. We thank
size-exclusion chromatography of 12 samples (the FPH produced Nicholas Knott (Steveston Seafood Direct Ltd.) for providing the
by 10 proteases, autolysate and PFS) were compared; it is evident Pacific hake used in this research as well as Amano Enzyme Inc.,
that molecular size was not the only property governing the Neova Technologies Inc., and Valley Research for the donations of
effectiveness of peptides in suppressing linoleic acid peroxidation. proteases used in this study.
In some samples such as FPH-NAK, FPH-Protin and autolysate,
fraction II with peptides in the 1400–2000 Da range, had higher References
antioxidative abilities, while others like FPH-A, FPH-B and PFS
showed peptides with sizes <1400 Da being the best antioxidants. Alemán, A., Giménez, B., Pérez-Santín, E., Gómez-Guillén, M. C., & Montero, P.
(2011). Contribution of Leu and Hyp residues to antioxidant and ACE-inhibitory
Therefore, peptide composition is believed to also be of great activities of peptide sequences isolated from squid gelatin hydrolysate. Food
importance in determining antioxidative activity towards lipid Chemistry, 125, 334–341.
peroxidation in the linoleic acid model system. Peptide composi- Alemán, A., Pérez-Santín, E., Bordenave-Juchereau, S., Arnaudin, I., Gómez-Guillén,
M. C., & Montero, P. (2011). Squid gelatin hydrolysates with antihypertensive,
tion, sequence, configuration and hydrophobicity have been
anticancer and antioxidant activity. Food Research International, 44(4),
proposed to affect antioxidative activity (Sarmadi & Ismail, 2010), 1044–1051.
while others have shown that hydrophobic amino acid residues Bamdad, F., Wu, J., & Chen, L. (2011). Effects of enzymatic hydrolysis on molecular
including Val or Leu at the N-terminus as well as Pro, His, Tyr, structure and antioxidant activity of barley hordein. Journal of Cereal Science,
54(1), 20–28.
Trp, Met and Cys in the sequence were important in antioxidant Bougatef, A., Hajji, M., Balti, R., Lassoued, I., Triki-Ellouz, Y., & Nasri, M. (2009).
activity (Samaranayaka & Li-Chan, 2011). Antioxidant and free radical-scavenging activities of smooth hound (Mustelus
Although most of the samples investigated in this study had mustelus) muscle protein hydrolysates obtained by gastrointestinal proteases.
Food Chemistry, 114, 1198–1205.
one or more fractions displaying good inhibition of peroxidation Bougatef, A., Nedjar-Arroume, N., Manni, L., Ravallec, R., Barkia, A., Guillochon, D.,
in the linoleic acid model system, FPH-VBNP showed the best et al. (2010). Purification and identification of novel antioxidant peptides from
potential to be used as antioxidants in food application owing enzymatic hydrolysates of sardinelle (Sardinella aurita) by-products proteins.
Food Chemistry, 118, 559–565.
to the fact that the unfractionated FPH-VBNP and its different Dong, S., Zeng, M., Wang, D., Liu, Z., Zhao, Y., & Yang, H. (2008). Antioxidant and
fractions all displayed similar antioxidant activity to BHT. This biochemical properties of protein hydrolysates prepared from Silver carp
suggests that Validase BNP-L was able to hydrolyze Pacific hake (Hypophthalmichthys molitrix). Food Chemistry, 107, 1485–1493.
Je, J.-Y., Lee, K.-H., Lee, M. H., & Ahn, C.-B. (2009). Antioxidant and antihypertensive
fillet mince and release peptides with a range of molecular protein hydrolysates produced from tuna liver by enzymatic hydrolysis. Food
weights that were all effective in inhibiting lipid peroxidation in Research International, 42, 1266–1272.
the linoleic acid model system. It is postulated that Validase Je, J.-Y., Park, P.-J., & Kim, S.-K. (2005). Antioxidant activity of a peptide isolated from
Alaska pollack (Theragra chalcogramma) frame protein hydrolysate. Food
BNP-L may cleave at specific amino acid residues that release
Research International, 38, 45–50.
peptides having particular structural characteristics associated Jia, J., Zhou, Y., Lu, J., Chen, A., Li, Y., & Zheng, G. (2010). Enzymatic hydrolysis of
with potent antioxidant activity, regardless of peptide size. This Alaska pollack (Theragra chalcogramma) skin and antioxidant activity of the
finding is advantageous for large-scale production because no resulting hydrolysate. Journal of Science and Food Agriculture, 90, 635–640.
Kim, S.-Y., Je, J.-Y., & Kim, S.-K. (2007). Purification and characterization of
purification steps will be required to generate antioxidative antioxidant peptide from hoki (Johnius belengerii) frame protein by
peptides to apply to food products. gastrointestinal digestion. Journal of Nutritional Biochemistry, 18, 31–38.
1306 I.W.Y. Cheung et al. / Food Chemistry 134 (2012) 1297–1306

Liceaga-Gesualdo, A. M., & Li-Chan, E. C. Y. (1999). Functional properties of fish Ren, J., Zhao, M., Shi, J., Wang, J., Jiang, Y., Cui, C., et al. (2008a). Purification and
protein hydrolysates from herring (Clupea harengus). Journal of Food Science, identification of antioxidant peptides from grass carp muscle hydrolysates by
64(6), 1000–1004. consecutive chromatography and electrospray ionization-mass spectrometry.
Liu, Q., Kong, B., Xiong, Y. L., & Xia, X. (2010). Antioxidant activity and functional Food Chemistry, 108, 727–736.
properties of porcine plasma protein hydrolysate as influenced by the degree of Ren, J., Zhao, M., Shi, J., Wang, J., Jiang, Y., Cui, C., et al. (2008b). Optimization of
hydrolysis. Food Chemistry, 118, 403–410. antioxidant peptide production from grass carp sarcoplasmic protein using
Mendis, E., Rajapakse, N., Byun, H.-G., & Kim, S.-K. (2005). Investigation of jumbo response surface methodology. LWT – Food Science and Technology, 41,
squid (Dosidicus gigas) skin gelatin peptides for their in vitro antioxidant effects. 1624–1632.
Life Sciences, 77, 2166–2178. Samaranayaka, A. G. P., & Li-Chan, E. C. Y. (2008). Autolysis-assisted production of
Mitsuda, H., Yasumato, K., & Iwami, K. (1966). Antioxidative action of indole fish protein hydrolysates with antioxidant properties from Pacific hake
compounds during the autoxidation of linoleic acid. Eiyoto Shokuryo, 19, (Merluccius productus). Food Chemistry, 107, 768–776.
210–214. Samaranayaka, A. G. P., & Li-Chan, E. C. Y. (2011). Food-derived peptidic
Moon, J.-K., & Shibamoto, T. (2009). Antioxidant assays for plant and food antioxidants: A review of their production, assessment, and potential
components. Journal of Agricultural and Food Chemistry, 57, 1655–1666. applications. Journal of Functional Foods, 3(4), 229–254.
Nalinanon, S., Benjakul, S., Kishimura, H., & Shahidi, F. (2011). Functionalities and Sarmadi, B. H., & Ismail, A. (2010). Antioxidative peptides from food proteins: A
antioxidant properties of protein hydrolysates from the muscle of ornate review. Peptides, 31, 1949–1956.
threadfin bream treated with pepsin from skipjack tuna. Food Chemistry, 124, Tang, C.-H., Wang, X.-S., & Yang, X.-Q. (2009). Enzymatic hydrolysis of
1354–1362. hemp (Cannabis sativa L.) protein isolate by various proteases and
Phanturat, P., Benjakul, S., Visessanguan, W., & Roytrakul, S. (2010). Use of pyloric antioxidant properties of the resulting hydrolysates. Food Chemistry, 114,
caeca extract from bigeye snapper (Priacanthus macracanthus) for the 1484–1490.
production of gelatin hydrolysate with antioxidative activity. LWT – Food Theodore, A. E., Raghavan, S., & Kristinsson, H. G. (2008). Antioxidative activity of
Science and Technology, 43, 86–97. protein hydrolysates prepared from alkaline-aided channel catfish protein
Pownall, T. L., Udenigwe, C. C., & Aluko, R. E. (2010). Amino acid composition and isolates. Journal of Agricultural and Food Chemistry, 56, 7459–7466.
antioxidant properties of pea seed (Pisum sativum L.) enzymatic protein Udenigwe, C. C., & Aluko, R. E. (2011). Chemometric analysis of the amino acid
hydrolysate fractions. Journal of Agricultural and Food Chemistry, 58, 4712–4718. requirements of antioxidant food protein hydrolysates. International Journal of
Raghavan, S., & Kristinsson, H. G. (2008). Antioxidative efficacy of alkali-treated Molecular Sciences, 12, 3148–3161.
tilapia protein hydrolysates: A comparative study of five enzymes. Journal of Wu, H.-C., Chen, H.-M., & Shiau, C.-Y. (2003). Free amino acids and peptides as
Agricultural and Food Chemistry, 56, 1434–1441. related to antioxidant properties in protein hydrolysates of mackerel (Scomber
Raghavan, S., Kristinsson, H. G., & Leeuwenburgh, C. (2008). Radical scavenging and austriasicus). Food Research International, 36, 949–957.
reducing ability of tilapia (Oreochromis niloticus) protein hydrolysates. Journal of
Agricultural and Food Chemistry, 56, 10359–10367.