Food Chemistry 155 (2014) 64–73

Contents lists available at ScienceDirect

Food Chemistry

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

Chemical modification of New Zealand hoki (Macruronus novaezelandiae)skin gelatin and its properties

0308-8146/$ - see front matter � 2014 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.foodchem.2014.01.043

⇑ Corresponding author. Tel.: +64 9 9233156; fax: +64 9 373 7422.E-mail address: conradperera@gmail.com (C.O. Perera).

1 Current address: Department of Fisheries Science, Faculty of Fisheries and Aqua-Industry, University Malaysia Terengganu, 21030 Kuala Terengganu, Malaysia.

Nor Fazliyana Mohtar 1, Conrad O. Perera ⇑, Yacine HemarSchool of Chemical Sciences, Food Science Programme, The University of Auckland, Private Bag 92019, Auckland, New Zealand

a r t i c l e i n f o

Article history:Received 22 August 2013Received in revised form 16 December 2013Accepted 15 January 2014Available online 23 January 2014

Keywords:GelatinHoki (Macruronus novaezelandiae)Chemical cross-linking agentsGel strengthMelting pointRheological properties

a b s t r a c t

Chemical modifications of gelatin from New Zealand hoki (Macruronus novaezelandiae) skins were carriedout using three different cross-linking agents, namely, genipin, glutaraldehyde and caffeic acid, at differ-ent concentrations. The chemically modified gelatins exhibited better physical properties, such as highergel strength, melting point, and rheological properties than did the uncross-linked gelatin. Gelatin cross-linked with glutaraldehyde had higher gel strength and melting point (231 g, 21.9 �C) than those cross-linked with caffeic acid (229 g, 21.6 �C) and genipin (211 g, 20.5 �C) at concentrations of 0.133, 0.111, and0.044 M, respectively. The elastic modulus (G0) and the loss modulus (G00) of chemically cross-linked gel-atins were higher than those of the uncross-linked ones. These improved physicochemical properties ofgelatin could lead to the development of products in the food industry that meet consumer demands.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The skin gelatin derived from hoki (Macruronus novaezelandiae)exhibits lower gel strength and melting point than do bovine andporcine gelatins (Mohtar, Perera, & Quek, 2010). Hoki gelatin gelstend to melt and behave as a viscous liquid at room temperature,which limits their potential use in food applications. One possibil-ity for improving its physiochemical properties is through chemi-cal cross-linking. Cross-linked pollock and salmon gelatins werereported to have higher gel strength at room temperature thanhad uncross-linked gelatins (Chiou et al., 2006). Thus, this workdeals with the modification of hoki gelatin using three chemicalcross-linking agents, namely, genipin, glutaraldehyde and caffeicacid, at various concentrations.

Genipin is an aglycone derived from the glycoside, geniposide. Itis isolated from the fruits of Genipa americana and Gardenia jasmi-noides Ellis (Akao, Kobayashi, & Aburada, 1994; Djerassi et al.,1961). Almost 4–6% of the dry mass of the above dried fruits con-sists of genipin, having a molecular weight of 226 Da. It is an excel-lent natural cross-linking agent for protein, and it has a lowcytotoxicity. Its cytotoxicity is known to be lower than that of glu-taraldehyde (Nimni, Cheung, Strates, Kodama, & Sheikh, 1988;Sung, Chang, Chiu, Chen, & Liang, 1999; Sung, Huang, Huang, &

Tsai, 1999), formaldehyde (Nimni et al., 1988), dialdehyde starches(Rosenberg, 1978), and epoxy compounds (Imamura, Sawatani,Koyanagi, Noishiki, & Miyata, 1989). The cross-linking reaction ofgenipin and proteins involves two possible mechanisms, as shownin Supplementary Fig. S1A. The first mechanism is when the geni-pin molecule undergoes a nucleophilic substitution by the primaryamines of proteins, resulting in the heterocyclic linking of genipinto the amine in the protein (Butler, Ng, & Pudney, 2003). The sec-ond reaction takes place when the ester group on genipin under-goes a nucleophilic substitution by the secondary amide linkage(Butler et al., 2003). The occurrence of covalent cross-links be-tween the primary amine residues leads to minimal residual toxic-ity of the materials (Liang, Chang, Liang, Lee, & Sung, 2004; Tsai,Huang, Sung, & Liang, 2000).

Glutaraldehyde, a linear 5-carbon dialdehyde, is known to beone of the most widely used chemical cross-linking agents. It re-acts with amine groups in protein and is relatively inexpensive.It is commonly used to cross-link proteins through the reactionof the aldehyde functional groups (Farris, Song, & Huang, 2010).The mechanism of cross-linking is illustrated in SupplementaryFig. S1B, which shows a nucleophilic addition-type reaction ofaldehyde functional groups with free nonprotonated e-aminogroups (ANH2) of lysine. The initial reaction involves the nucleo-philic addition of the e-amino groups to the carbonyl groups ofthe aldehyde. This is followed by the formation of an intermediatecalled carbinoalamine. Formation of Schiff base occurs after theprotonation of the hydroxyl group, followed by the loss of a watermolecule, as shown in Supplementary Fig. S1B (Farris et al., 2010).

N.F. Mohtar et al. / Food Chemistry 155 (2014) 64–73 65

Caffeic acid (3,4-dihydroxycinnamic acid) is a plant-derivedphenolic compound like genipin, which is commonly found in cof-fee beans, tea leaves, potatoes, cell walls of fruits, and hulls of cer-eal grains. It is a good substrate for polyphenol oxidase and itundergoes an oxidation process in plant tissues. Caffeic acid is alsoused to cross-link amine-containing polymers (Butler et al., 2003;Nickerson, Patel, Heyd, Rousseau, & Paulson, 2006). The cross-link-ing reaction of caffeic acid is initiated by its conversion to an inter-mediate quinone (Strauss & Gibson, 2004). Subsequently, lysinegroups from proteins react reversibly with the quinone group itselfto form quinonimines (Cheftel, 1979). Further reactions lead to theformation of cross-linkages and complexes between lysine andpolymerised quinones, as illustrated in Supplementary Fig. S1C.Several studies on cross-linking of proteins, using caffeic acid, havebeen previously conducted (Isenburg, Karamchandani, Simionescu,& Vyavahare, 2006; Strauss & Gibson, 2004). The results have dem-onstrated that protein-caffeic acid cross-linking leads to the forma-tion of covalent bonds, thus improving the mechanical propertiesof the protein-caffeic acid complexes.

To the best of our knowledge, there is no available informationon the chemical cross-linking of hoki gelatin modified by genipin,glutaraldehyde and caffeic acid. Therefore, the scope of the currentwork was to investigate the effect of concentration of these threechemical cross-linking agents on the gel strength and meltingpoint of gelatin gels. Further characterisation of the properties ofchemically cross-linked gelatins was also conducted by theirmolecular weight distribution and rheological measurements.

2. Materials and methods

2.1. Materials

Frozen hoki skins were thawed at room temperature (15 �C)overnight. The adherent flesh and scales from the skin were re-moved using a sharp knife. The skin was washed in tap water to re-move obvious impurities and was then cut into approximately40 � 40 mm pieces. The cleaned skin was minced without water,using a Waring blender (Waring Commercial�, New Hartford, CT,USA) at the lowest speed setting for 15 min. The minced skins werepacked into 150 � 90 mm snap-lock plastic bags (GLAD, CloroxNew Zealand Ltd., Auckland, New Zealand), sealed after the head-space air was removed manually, and stored at �20 �C until usedwithin the next 1–2 weeks. All chemicals and reagents used wereof analytical grade.

2.2. Extraction of fish gelatin

Gelatin from hoki skin was extracted according to Mohtar et al.(2010). Briefly, the frozen minced skins were thawed at room tem-perature (15 �C) and were rinsed in tap water. The rinsed andminced skins were pre-treated in 0.75 M NaCl solutions, afterwhich they were rinsed again in tap water. These steps were re-peated twice. The minced skins were gently stirred in Milli-Qwater (Millipore Corporation, Billerica, MA, USA) for 60 min at acontrolled temperature of 49.3 �C in a shaking water bath (RatekInstruments, Boronia, Victoria, Australia). The samples were centri-fuged using Sorvall� RC-28S Centrifuge (Sorvall�, Newton, CT, USA)at 10,000�g for 30 min at 15 �C. The clear extract obtained was fil-tered using Whatman filter paper (No. 5) (Whatman InternationalLtd., Kent, UK), and the filtrate obtained was dialysed against threechanges of Milli-Q water before freeze-drying using a FreeZonePlus (Labconco, MO, USA). The freeze-dried gelatin samples werekept in sealed containers, wrapped in aluminum foil to avoid directexposure to light and stored in a vacuum desiccator prior to furtheruse within 1–6 months.

2.3. Preparation of gelatin gels

2.3.1. GeneralIn this experiment, gels were prepared using 5% (w/w) hoki gel-

atin with concentrations of 0.022, 0.044, 0.066, 0.111, 0.133, and0.177 M of the different chemical cross-linking agents. A sampleof 5% (w/w) hoki gelatin was used in these experiments, as re-ported by different authors for pollock (Chiou et al., 2006), salmon(Chiou et al., 2006), and porcine gelatins (Chiou et al., 2006; Strauss& Gibson, 2004). Gelatin gel solutions, with and without the addedcross-linking agents, were prepared according to the proceduresdescribed below.

2.3.2. Preparation of gelatin gels without added chemical cross-linkingagent

5% (w/w) gelatin solution was prepared by mixing 5 g of drygelatin in 95 g of Milli-Q water. The solution was left at room tem-perature for 10 min to form a visibly homogeneous gelatin suspen-sion, and then heated at 45 �C for 30 min until the gelatin wascompletely dissolved. The gelatin solution was cooled to roomtemperature before maturing in a refrigerator at 10 �C for 18 h,prior to gel strength analysis. In addition, 5% (w/w) samples of bo-vine and porcine gelatins were prepared in the same way as forhoki gelatin gels and used as controls.

2.3.3. Preparation of gelatin gels with added genipin5% (w/w) of gelatin solution with added genipin was prepared

according to Chiou et al. (2006), as given in Section 2.3.2. Theappropriate amount of genipin (Prod. No. G4796, Sigma–Aldrich,Auckland, New Zealand) was added to the gelatin solution andwas stirred for an additional 20 min at 45 �C. The concentrationsof genipin in the gelatin samples were set at 0.022, 0.044, 0.066,0.111, 0.133, and 0.177 M. The gelatin solutions with added geni-pin were cooled to room temperature before maturing in a refrig-erator at 10 �C for 18 h, prior to gel strength analysis.

2.3.4. Preparation of gelatin gels with added glutaraldehydeThe same procedure as in Section 2.3.2 was followed, except

that the appropriate amount of glutaraldehyde (Prod. No. G5882,Sigma–Aldrich, Auckland, New Zealand) was added to the gelatinsolution and stirred for 5 min at 25 �C. The concentrations of glu-taraldehyde in the gelatin samples were set at 0.022, 0.044,0.066, 0.111, 0.133, and 0.177 M. The gelatin solutions with addedglutaraldehyde were cooled to room temperature before maturingin a refrigerator at 10 �C for 18 h, prior to gel strength analysis.

2.3.5. Preparation of gelatin gels with added caffeic acidA 5% (w/w) gelatin solution with added caffeic acid was pre-

pared according to Kosaraju, Puvanenthiran, and Lillford (2010),as given in Section 2.3.2. The appropriate amount of caffeic acid(Prod. No. C0625, Sigma–Aldrich, Auckland, New Zealand) wasadded to the gelatin solution and was stirred for 20 min at 60 �C.The concentrations of caffeic acid in the gelatin samples were setat 0.022, 0.044, 0.066, 0.111, 0.133, and 0.177 M. Oxygen gas(BOC Gases Ltd., Auckland, New Zealand) was bubbled throughthe solution for 20 min to initiate the oxidation reaction and cooledto room temperature before maturing in a refrigerator at 10 �C for18 h, prior to gel strength analysis.

2.4. Determination of gel strength

The gel strengths of 5% (w/w) hoki gelatin gels (with and with-out added chemical cross-linking agents) were determined accord-ing to the British Standard Institution method (British StandardInstitution, 1975). The samples (prepared as given before) wereused immediately after they were removed from the refrigerator

66 N.F. Mohtar et al. / Food Chemistry 155 (2014) 64–73

at 10 �C, using a TA.XT2 Texture Analyser (Stable Micro Systems,Surrey, UK), fitted with a 12.7 mm diameter cylindrical probe ata speed of 1 mm per s with a force of 0.1 N. Gel strength was de-fined as the maximum force in grammes (g) recorded, that was re-quired for the probe to penetrate 4 mm into the gelatin gel (Mohtaret al., 2010).

2.5. Determination of melting point

The melting points of 5% (w/w) hoki gelatin gels (with andwithout added chemical cross-linking agents) were measured,using the slip-point method according to Mohtar et al. (2010). A10 mm length of gelatin solution was placed in the middle of a cap-illary tube 100 mm in length and with a diameter of 0.2 mm, andwas held at 10 �C in the refrigerator for 18 h. The capillary tubewas attached close to the bulb of an alcohol thermometer(305 mm in length), using transparent tape. The thermometerand tube were immersed in a beaker of water at 5 �C for 5 min toequilibrate the gelatin gel at that temperature. The beaker contain-ing the thermometer and capillary tube was placed on the hot platemagnetic stirrer (Stuart, Staffordshire, UK). The temperature of thewater bath was increased at a rate of 0.5 �C per minute. The melt-ing point was defined as the temperature at which the gelatin col-umn in the capillary tube started to move up. The experiment wasconducted in triplicate.

2.6. Determination of molecular weight distribution

Protein patterns of hoki gelatin cross-linked with genipin, glu-taraldehyde, and caffeic acid were determined using sodium dode-cyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE)according to the method of Laemmli (1970). Dry gelatin sampleswere dissolved in Nu PAGE� LDS sample buffer (Invitrogen™,Carlsbad, CA, USA), Milli-Q water and dithiothreitol (DTT) to finalconcentrations of 1.13, 0.75, and 0.38 mg of protein per ml. Gelatinsamples were incubated for 10 min at 70 �C and analysed usingpre-cast gels of thickness 1.0 mm � 10 wells (Nu PAGE� 12% Bis-tris gel, Invitrogen™, Carlsbad, CA, USA). The loading volume was15 ll in each well. The outer chamber of the gel cassette was thenfilled with Nu PAGE� MOPS SDS running buffer (Invitrogen™,Carlsbad, CA, USA). The electrophoresis apparatus was connectedto a power supply instrument (EPS 601 Power Supply, GE Health-care, USA) and operated at a constant voltage of 200 v and a cur-rent of 111 mA for 1 h. Protein bands were stained withCoomassie Brilliant Blue G-250 (Invitrogen™, Carlsbad, CA, USA)for 24 h and destained with Milli-Q water. For staining anddestaining of the gels, a KS 125 basic IKA-rotary shaker (IKA Labor-technik, Staufen, Germany) was used. The photographs of proteinsubunit patterns were taken using a photo scanner (Scanjet 4370digital Flatbed Scanner, HP, CA, USA). The apparent molecularweights of the protein bands used in the SeeBlue� Plus 2 Pre-stained Standard (Invitrogen™, Carlsbad, CA, USA) were: myosin(191 kDa), phosphorylase (97 kDa), BSA (64 kDa), glutamic dehy-drogenase (51 kDa), alcohol dehydrogenase (39 kDa), carbonicanhydrase (28 kDa), myoglobin red (19 kDa), and lysozyme(14 kDa).

2.7. Determination of rheological properties

2.7.1. Preparation of gelatin solutions for rheological measurementsA 5% (w/w) gelatin solution was prepared according to Sec-

tion 2.3.2. To investigate the effect of chemical cross-linking agentson the gelation properties of gelatin, samples were preparedaccordingly, as described in Sections 2.3.1, 2.3.2, 2.3.3, and 2.3.4.

2.7.2. Rheological measurementsRheological measurements of 5% (w/w) gelatin solutions were

carried out using a Paar Physica MCR 301 stress-controlled rheom-eter (Anton-Paar, Austria) fitted with a cup and bob (Couettegeometry). The inner diameter of the cup was 27.5 mm and thediameter of the bob was 26.5 mm, giving a gap between the cupand the bob of 1.0 mm. All the measurements (G’ and G00) were per-formed in triplicate.

Approximately 17 ml of 5% (w/w) gelatin solution were trans-ferred to the rheometer. The sample was sheared at a constantshear rate of 10 revolutions per second (10 s�1). The surface ofthe sample in the Couette geometry was covered with a thin layerof soya oil (one drop of oil) to prevent evaporation.

For gelation, the sample was initially maintained at a tempera-ture of 45 �C to allow for equilibration. The sample was then slowlycooled from 45 to 10 �C at a rate of 0.5 �C per min. The sample wasleft for 10 h at 10 �C to ensure that the gel formed was fully devel-oped. During this time, time sweep measurements were performedat a constant frequency of 0.1 Hz and a constant applied strain of0.5%, and G’ (Pa) was measured as a function of time (minute).

At the end of the time sweep measurement, frequency sweepmeasurements were performed at a constant strain of 0.5% and aconstant temperature of 10 �C as a function of frequencies rangingfrom 0.01 to 10 Hz.

Finally, at the end of the frequency sweep measurement, astrain sweep measurement was applied. The strain sweep mea-surement was applied at a constant frequency of 0.1 Hz by varyingthe strain from 0.1% to 10.000%. Strain sweep measurement allowsdetermination of the linear viscoelastic region and the behaviour ofthe sample at large deformation, which involves the determinationof the strain and stress at which the sample flows.

3. Results and discussion

3.1. Determination of gel strength, melting point, and molecularweight distribution

Fig. 1 shows the gel strength and melting point of gelatin gels asa function of concentration of added chemical cross-linking agents.The gel strength of hoki gelatin without the added chemicals was171.7 ± 0.39 g, and the gel strengths of bovine and porcine gelatinswere 246 ± 1.39 and 270 ± 0.36 g, respectively (Fig. 1A). In general,the gel strength of gelatin increased with increase in concentrationof the chemical cross-linking agents and remained relatively con-stant. The increase in gel strength could be explained as due tothe effects of cross-linking of peptide chains of gelatin that bringthem closer together, thus inducing the formation of collagen-liketriple helices (Jobstl, O’Connell, Fairclough, & Williamson, 2004). Itis presumed that the cross-linking agents form intermolecular andintramolecular bridges with the free amino groups of gelatinsubunits. Strauss and Gibson (2004) and Kosaraju et al. (2010) alsoreported similar increases in gel strength and melting point ofcross-linked porcine and bovine gelatins, respectively.

The highest gel strength of 231 ± 0.85 g, was found in gelatinwith added glutaraldehyde at 0.133 M, while that with added caf-feic acid was 229 ± 1.09 g at 0.111 M and that with genipin was211 ± 0.52 g at 0.044 M. These values were significantly differentfrom each other at P < 0.05. It could be seen that the highest gelstrength was observed at a lower concentration of genipin,whereas, those for gelatins with added glutaraldehyde and caffeicacid were obtained at higher concentrations. Chiou et al. (2006),working on pollock gelatin with added glutaraldehyde, also re-ported high gel strength. Their highest gel strength, of 92 g, wasobtained at 0.066 M, which was the highest concentration of

0

50

100

150

200

250

300

Control 0.022 0.044 0.066 0.111 0.133 0.177

Gel

stre

ngth

(g)

Concentration of chemical cross-linking agents (M)

a

b

c

aaa

a a a aaa a

b ba

b ca

b c

0

5

10

15

20

25

30

Control 0.022 0.044 0.066 0.111 0.133 0.177

Mel

ting

poi

nt (

C)

a

bc

a aaa b b a bb

a

cb

ac

b

acb

Concentration of chemical cross-linking agents (M)

A

B

Fig. 1. (A) Gel strength (g) as a function of concentration (M) of chemical cross-linking agents for 5% (w/w) hoki gelatin. (B) Melting point (�C) as a function of concentration(M) of chemical cross-linking agents for 5% (w/w) hoki gelatin. Types of chemical cross-linking agents used were: genipin ( ); glutaraldehyde ( ); and caffeic acid ( ). Bovine( ) and porcine ( ) gelatins (5%, w/w) were used as controls. The different letters on the bars within the same concentration of chemical cross-linking agents indicatesignificant differences (P < 0.05). Error bars represent standard deviation.

N.F. Mohtar et al. / Food Chemistry 155 (2014) 64–73 67

glutaraldehyde that they used; however, much higher concentra-tions, up to 0.177 M, were used in the present study.

Similar trends were observed in the melting points obtained, asshown in Fig. 1B. Hoki gelatin with added glutaraldehyde at0.133 M exhibited the highest melting point (21.9 ± 0.14 �C), fol-lowed by caffeic acid at 0.111 M (21.6 ± 0.10 �C), and genipin at0.044 M (20.5 ± 0.05 �C).

Supplementary Fig. S2 demonstrates the electrophoretic pro-files of gelatin with added genipin, glutaraldehyde, and caffeic acidat various concentrations. In general, the chemically cross-linkedgelatins demonstrated higher molecular weight subunits(>190 kDa) than did uncross-linked gelatins, reflecting the poly-merisation reaction of the gelatin subunits. These aggregates wereabsent in the gels without added chemical cross-linking agents.The electrophoretic profiles corresponding to 0.044 M genipin(Fig. S2, well ‘d’), 0.133 M glutaraldehyde (Fig. S2,well ‘g’), and0.111 M caffeic acid (Fig. S2, well ‘f’) showed the occurrence of sub-units much greater than 190 kDa, indicating a higher degree ofpolymerisation. These findings support the results reported onthe highest gel strength (Fig. 1A) obtained from gelatin with addedgenipin (211 ± 0.52 g), glutaraldehyde (231 ± 0.85 g), and caffeic

acid (229 ± 1.09 g). The difference in gel strengths, melting points,and molecular weight distribution of gelatin obtained in this studycould be due to the different reaction mechanisms of the chemicalcross-linking agents.

Glutaraldehyde has been used to cross-link protein moleculesthrough reaction of the aldehyde functional groups. It is knownto exhibit superior efficiency in stabilisation of collagenous mate-rials (Khor, 1997). Glutaraldehyde may also form covalent bondsbetween gelatin molecules at either intermolecular or intramolec-ular levels. Intermolecular bonds could form long distance bridgesthrough the condensation reaction of glutaraldehyde or polymer-ised glutaraldeyhde with the gelatin molecule in aqueous solution,thus leading to the formation of large molecular weight com-pounds, as illustrated in Fig. 2A. In this study, the reaction was con-ducted at an alkaline pH of 8, where the reaction between e-aminogroup of lysine from gelatin and aldehyde groups of glutaraldehydeis known to form stable cross-links (Walt & Agayn, 1994). This phe-nomenon could explain the higher gel strength obtained from gel-atin with added glutaraldehyde than from those with addedgenipin or caffeic acid. The result observed from this study is sup-ported by the findings from Hardy, Hughes, and Rydon (1979) who

Fig. 2. Schematic illustrations of the presumable intramolecular and intermolecular cross-linking structures of gelatin with added (A) glutaraldehyde, (B) genipin, and (C)caffeic acid. GTA, GP, and CA represent glutaraldehyde, genipin, and caffeic acid respectively.

68 N.F. Mohtar et al. / Food Chemistry 155 (2014) 64–73

reported the formation of stable cross-links through the reaction ofglutaraldehyde with amine groups of proteins, via the formation ofa Schiff base.

Lower gel strengths were observed for gelatin with added gen-ipin than for those with added glutaraldehyde and caffeic acid, andthis could be explained by the reaction mechanism involved be-tween genipin and the gelatin molecule. It is well known that gen-ipin is a high molecular weight compound (226 Da) compared toglutaraldehyde (100 Da) and caffeic acid (180 Da). The reactivebinding sites on genipin are responsible for linking two amine sidegroups of the gelatin molecule, as illustrated in Fig. 2B. When gen-ipin cross-links with one strand of the gelatin molecule, it is diffi-cult for another strand to interact with the other reactive bindingsite of genipin, due to the effect of steric hindrance (Mi, Sung, &Shyu, 2000). Specifically, steric hindrance arises when the neigh-bouring groups of gelatin crowd the reactive binding sites of gen-ipin, thus leading to the slower reaction rate. Also, if the genipinand gelatin molecules are brought too close together, it is possibleto have an increase in energy, which may slow the reaction mech-anism of genipin.

In forming gel networks, two reaction mechanisms betweengenipin and gelatin molecules are important in the system. Thefirst reaction, involving the formation of heterocyclic amine, mayoccur faster than the substitution of the ester group on the genipinmolecule in the second reaction. A slower reaction rate is sug-gested due to the effect of steric hindrance between genipin andgelatin molecules as mentioned above. Such a hypothesis is sup-ported by Butler et al. (2003) who cross-linked bovine serum albu-min (BSA), porcine gelatin, and chitosan with genipin, and foundthat the second reaction mechanism was slower than the first.

Caffeic acid has the ability to bind to gelatin through its differ-ent phenolic groups at many sites within the same gelatin strand.This mechanism is schematically illustrated in Fig. 2C. The bindingaction of phenolic compounds on several sites of the same gelatinmolecule leads to a coiled structure. This may cause structuralchanges of the gelatin molecule to become more compact and so-lid. Aggregation starts to occur with increase in the concentrationof caffeic acid added, leading to the formation of gelatin-caffeicacid complexes, as shown in Fig. 2C. The presence of suchcomplexes also promotes cross-linking reactions between different

N.F. Mohtar et al. / Food Chemistry 155 (2014) 64–73 69

gelatin molecules, resulting in increase of large molecular weightcompounds, thus directing to greater formation of gel network sys-tems. This hypothesis is supported by the current findings ob-tained from gel strength (Fig. 1A) and also from molecularweight analyses (Supplementary Fig. S2). The highest gel strengthwas observed at 0.111 M concentration, and it was slightly de-creased when the concentration went beyond that point. The de-crease in gel strength at higher concentrations than 0.111 Mcould be due to the excessive formation of gelatin–caffeic acidcomplexes, resulting in the formation of more compact and solidmolecules, but leading to a decrease in the gel strength due tonon-homogeneous gel formation. Strauss and Gibson (2004), work-ing with porcine gelatin also reported an increase in gel strengthwith added caffeic acid; however, the maximum concentrationthey studied was only 0.05 M while the highest concentration usedin this study was 0.177 M. Further differences in the properties ofgelatin with added chemical cross-linking agents are noteworthy,based on the results obtained from the rheological measurements.

3.2. Determination of rheological properties of chemically cross-linkedgelatin

Time sweep measurements at a constant frequency of 1 Hzwere performed to determine the gelation kinetics and gel proper-ties of gelatin. The elastic modulus (G’) and loss modulus (G00) ofgelatin with added genipin (A), glutaraldehyde (B), and caffeic acid(C) as a function of time are reported in Supplementary Fig. S3. Theresults show that gelatins (with and without added chemical cross-linking agents) exhibited more or less liquid behaviour with verylow G’ values at the initial gelation time. As the proteins in the gel-atins start to aggregate, a gel begins to form and G’ increases mark-edly as gelation time is advanced. The transition of solution to gel(sol–gel) of gelatin samples corresponds to a state in which a cross-linked material undergoes phase transition from a liquid to a solid-like state (Ross-Murphy, 1992; Te Nijenhuis, 1997). This gelationprocess happened when the gelatin samples were cooled to lowtemperatures, which in this study was 10 �C. The occurrence of fas-ter gelation process with time could be explained by the formationof physical cross-links such as ionic and hydrogen-bondinginteractions. Kosaraju et al. (2010) worked on bovine gelatin with

Fig. 3. Elastic modulus, G’ as a function of the chemical cross-linking concentration. Chem( ). Error bars represent standard deviation.

added caffeic acid, and they made similar observations on the in-crease of G’ values due to the increase in gel rigidity during thegelation time.

In general, the G’ of gelatin samples increased with increase inconcentration of the chemical cross-linking agents. However, theadded chemical cross-linking agents at certain concentrations gavemarkedly higher increases in the G’ values than did those at otherconcentrations, indicating faster or greater cross-linking effects.Specifically, the gelatin with added genipin showed the highestG’ value at a concentration of 0.044 M (Supplementary Fig. S3A).This behaviour was expected because of the higher molecularweight subunits formed due to cross-linking reactions at that con-centration, as shown in Supplementary Fig. S2A (well ‘d’). In partic-ular, higher molecular weight subunits, corresponding toapproximately 210, 190 and 90 kDa, were observed for gelatin withadded genipin at a concentration of 0.044 M (well ‘d’). For the gel-atin samples with added glutaraldehyde, the highest G’ value wasobserved at a concentration of 0.133 M (Supplementary Fig. S3B).Again, according to the electrophoretic profile in SupplementaryFig. S2B, much higher molecular weight subunits (220 and200 kDa) than the control samples were identified at that concen-tration. Similar observations were made by Chiou et al. (2006) onthe rapid increase in storage modulus (G’) of pollock gelatin withadded glutaraldehyde compared to those with added genipin.

Supplementary Fig. S3C shows that gelatin with added caffeicacid had the highest G’ value at the concentration of 0.111 M andit corresponded with the electrophoretic pattern obtained inSupplementary Fig. S2C (well ‘f’). Several high molecular weightsubunits, corresponding to 200, 191, and 180 kDa, were observed,indicating the presence of gelatin-caffeic acid complexes. Ashypothesised earlier in Fig. 2C, the formation of such complexeswould result in the formation of more rigid gels.

Frequency sweep tests were conducted to observe the elasticmodulus, G’ (square symbols) and loss modulus, G00 (circle symbols)as a function of frequency for gelatin with added genipin, glutaral-dehyde, and caffeic acid at various concentrations (SupplementaryFig. S4). It could be clearly seen that, for gelatin samples with eachtype and level of added chemical cross-linking agents, and at allapplied frequencies, the elastic modulus (G’) values were greaterthan the loss modulus (G00) by a factor of 10. This finding indicates

ical cross-linking agents used were genipin ( ), glutaraldehyde ( ), and caffeic acid

B

A

10

100

1000

10000

0000100010010111.0

Strain (%)

G' (

Pa)

and

G"

(Pa)

10

100

1000

10000

0000100010010111.0

Strain (%)

G' (

Pa)

and

G"

(Pa)

A

B

0.1

1

10

100

1000

10000

0.1 1 10 100 1000 10000

Strain (%)

G'(

Pa) a

ndG"

(Pa)

C

Fig. 4. Elastic modulus, G’ (circle symbols) and loss modulus, G’’ (square symbols) as a function of strain (%) for (A) genipin, (B) glutaraldehyde, and (C) caffeic acid. Theconcentration of gelatin was 5% (w/w) and the concentrations of chemical cross-linking agents were: Control ( , ); 0.022 M ( , ); 0.044 M ( , ); 0.066 M ( , );

0.111 M ( , ); 0.133 M ( , ); and 0.177 M ( , ).

70 N.F. Mohtar et al. / Food Chemistry 155 (2014) 64–73

N.F. Mohtar et al. / Food Chemistry 155 (2014) 64–73 71

that the gelatin samples were gelled and a strong gel network wasformed (Lapasin & Pricl, 1995). The formation of gel network thattook place when the gelatin solution was cooled, probably emu-lates the triple helical structures of collagen (Babin & Dickinson,2001). The triple helix structure is the basic unit of collagen fromwhich gelatin was derived. Thus, it is suggested that, to some ex-tent, the gelatin molecules revert back to the collagen structureduring gelation process. This speculation is supported by Gómez-Guillén et al. (2002), Simon et al. (2003) and Haug, Draget, andSmidsrød (2004), who confirmed that gelatin samples with greaterG’ values had higher concentration of triple helical structures. TheG’ values obtained for all of the gelatin samples with added chem-ical cross-linking agents showed a constant value in the range offrequency from 0.01 to 1 Hz. This is an indication that stablecross-linked network systems were formed. This frequency-inde-pendent characteristic of G’ also demonstrates a solid-like behav-iour in the gel network system (Moura, Figueiredo, & Gil, 2007).

At higher frequency values (>1 Hz), the gelatin samples demon-strated a slight increase in G’, corresponding to the highest gelstrength obtained for each chemical cross-linking agent used (referto the gel strength data obtained in Fig. 1A), but they were notsignificantly different at P < 0.05. The findings from this study re-

10

15

20

25

30

35

40

45

50

55

0 0.02 0.04 0.06 0.08

Concentration of chemical

γγ linear

(%)

0

200

400

600

800

1000

1200

0 0.02 0.04 0.06 0.08

Concentration of chemic

σ linear

(Pa)

A

B

Fig. 5. (A) The critical strain (clinear, %) and (B) critical stress (rlinear, Pa) values of gelatChemical cross-linking agents used were genipin ( ), glutaraldehyde ( ), and caffeic acistandard deviation.

vealed that the different concentrations of the added chemicalcross-linking agents had very little influence on the G’ and G00

values.To compare the result of the frequency sweep between the dif-

ferent chemical cross-linking agents, the value of G’ at 1 Hz is re-ported as a function of the concentration of added chemicalcross-linking agents (Fig. 3). The G’ values increased for all the con-centrations of the chemical cross-linking agent used in this study.However, different optimum concentrations were obtained for thedifferent chemical cross-linking agents. For instance, the maxi-mum values of G’ were found to be 1905, 2385, and 2205 Pa at con-centrations of 0.044, 0.133, and 0.111 M for genipin,glutaraldehyde, and caffeic acid, respectively. This finding confirmsthe time sweep measurements above, where different addedchemical cross-linking agents resulted in different maximum val-ues of the final G’ (value of G’ after 10 h).

The results of the strain sweep measurements are shown inFig. 4 for the different chemical cross-linking agents used. Qualita-tively, all the gelatin samples showed similar behaviour of G’ andG’’ as a function of the applied strain. First, up to a critical strain(clinear), which corresponds to a critical stress (rlinear), both G’ andG’’ remained independent of the applied strain. This region of

0.1 0.12 0.14 0.16 0.18

cross-linking agents (M)

0.1 0.12 0.14 0.16 0.18

al cross-linking agents (M)

in samples with added chemical cross-linking agents at linear viscoelastic region.d ( ) at control, 0.022, 0.044, 0.066, 0.111, 0.133, and 0.177 M. Error bars represent

100

200

300

400

500

600

700

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18

Concentration of chemical cross-linking agents (M)

γγ bre

akin

g(%

)

200

600

1000

1400

1800

2200

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18

Concentration of chemical cross-linking agents (M)

σ bre

akin

g(P

a)

A

B

Fig. 6. (A) The critical strain (cbreaking, %) and (B) critical stress (rbreaking, Pa) values of gelatin samples with added chemical cross-linking agents at gel breaking region.Chemical cross-linking agents used were genipin ( ), glutaraldehyde ( ), and caffeic acid ( ) at control, 0.022, 0.044, 0.066, 0.111, 0.133, and 0.177 M. Error bars representstandard deviation.

72 N.F. Mohtar et al. / Food Chemistry 155 (2014) 64–73

strain is known as the linear viscoelastic region. When the appliedstrain is further increased, both G’ and G’’ increase to reach a max-imum value. This region of strain depicts a typical strain-hardeningbehaviour. Strain-hardening behaviour has attracted a lot of inter-est in the last few years; however, this area is still not yet wellunderstood (Hyun et al., 2011). Further analysis of the strain-hard-ening behaviour observed for the gelatin samples is outside thescope of this paper. However, at a very high applied strain, bothG’ and G’’ start to decrease, indicating that the gelatin sample isstarting to break. For simplicity, the strain (cbreaking) and the corre-sponding stress (rbreaking) of G’ = G’’ are considered as the strain andcorresponding stress at which the sample breaks. For strains higherthan the strain at which G’ = G’’, the sample starts to flow (G’’ be-comes higher than G’).

To compare the strain sweep measurements of the gelatin sam-ples with different added chemical cross-linking agents, the valuesof clinear and rlinear are shown in Fig. 5. Similarly, the values ofcbreaking and rbreaking are shown in Fig. 6. From Fig. 5A, it can beclearly seen that the linear viscoelastic region occurs at strainshigher than 15% (see control sample). This confirms that the fre-quency sweep measurements that were performed at a constantstrain of 0.5% were all carried out within the linear viscoelastic

region. Fig. 5 also shows that both clinear and rlinear increased withthe increase in the concentration of chemical cross-linking agents.Furthermore, a similar trend to the elastic modulus G’ at 1 Hz (Sup-plementary Fig. S4) was observed. In fact, maximum clinear values of32%, 50%, and 25% were obtained at concentrations of 0.044, 0.133,and 0.111 M for genipin, glutaraldehyde, and caffeic acid, respec-tively (Fig. 5A). Similarly, maximum rlinear values of 578, 1205,and 369 Pa were obtained at concentrations of 0.044, 0.133, and0.111 M for genipin, glutaraldehyde, and caffeic acid, respectively(Fig. 5B). The behaviour of the gelatin gels at very high strains alsofollows a similar trend. A maximum cbreaking of 284%, 631%, and379% was obtained at concentrations of 0.044, 0.133, and0.111 M for genipin, glutaraldehyde, and caffeic acid, respectively(Fig. 6A). Similarly, a maximum rbreaking values of 1425, 2040,and 1500 Pa were obtained at concentrations of 0.044, 0.111, and0.133 M for genipin, glutaraldehyde, and caffeic acid, respectively(Fig. 6B). The results obtained for both linear viscoelatic and gelbreaking regions indicate that the added chemical cross-linkingagents enhanced the strength of the gelatin gels. The rheologicalproperties determined above confirmed the gel strengthmeasurements shown in Fig. 1A that glutaraldehyde was a more

N.F. Mohtar et al. / Food Chemistry 155 (2014) 64–73 73

effective chemical cross-linking agent than were genipin or caffeicacid.

4. Conclusion

Results showed that the use of chemical cross-linking agentswhich involved one synthetic chemical (glutaraldehyde) and twonaturally occurring chemicals (genipin and caffeic acid) could in-crease the functional properties of hoki gelatin. Gelatin with addedglutaraldehyde exhibited higher gel strength and melting pointthan did those with added genipin and caffeic acid. For gelatinsamples containing glutaraldeyhde, the highest gel strength valueof 231 ± 0.85 g was achieved at the concentration of 0.133 M. Incontrast, caffeic acid at a concentration of 0.111 M was needed toachieve comparable gel strength of and 229 ± 1.09 g. Genipin at0.044 M gave the highest gel strength of 211 ± 0.52 g, which wassignificantly lower than those obtained from glutaraldehyde andcaffeic acid. The gel strengths obtained were in good agreementwith the results achieved from the rheological measurements.The elastic modulus (G’) values for cross-linked gelatin gels werehigher than were those for uncross-linked ones.

This study was aimed at obtaining fundamental information onthe effect of cross-linking gelatin with glutaraldehyde, genipin, andcaffeic acid. Glutaraldehyde, one of the most commonly used pro-tein cross-linking agents was used for comparison with the othertwo naturally occurring cross-linking agents, genipin and caffeicacid. The addition of different cross-linking agents to gelatin at dif-ferent optimal concentrations has resulted in increased functionalproperties of hoki gelatin due to the differences in the mechanismof cross-linking. From these results, it is possible to conclude thatcaffeic acid is a better cross-linking agent than genipin, with lowtoxicity that could provide high gel strength, although glutaralde-hyde gave the best results.

Acknowledgments

The authors would like to acknowledge the University of Auck-land for the financial support throughout the research. Furthergratitude is extended to the Ministry of Higher Education Malaysiaand University Malaysia Terengganu for a scholarship to Nor Fazli-yana Mohtar. They also thank the Independent Fisheries Limitedfor the fish skin samples and the International Food Agencies Lim-ited Company for providing the commercial mammalian gelatins.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.foodchem.2014.01.043.

References

Akao, T., Kobayashi, K., & Aburada, M. (1994). Enzymic studies on the animal andintestinal bacterial metabolism of geniposide. Biological Pharmacology Bulletin,17, 1573–1576.

Babin, H., & Dickinson, E. (2001). Influence of transglutaminase treatment on thethermoreversible gelation of gelatin. Food Hydrocolloids, 15, 271–276.

British Standard Institution (1975). Methods for sampling and testing gelatin (physicaland chemical methods). London, UK: BSI.

Butler, M. F., Ng, Y.-F., & Pudney, P. D. A. (2003). Mechanism and kinetics of thecrosslinking reaction between biopolymers containing primary amine groupsand genipin. Journal of Polymer Science Part A: Polymer Chemistry, 41,3941–3953.

Cheftel, J. C. (1979). Proteins and amino acids. In S. R. Tannenbaum (Ed.), Nutritionaland safety aspects of food processing (pp. 153–215). New York: Marcel Dekker.

Chiou, B. S., Avena-Bustillos, R. J., Shey, J., Yee, E., Bechtel, P. J., Imam, S. H., et al.(2006). Rheological and mechanical properties of cross-linked fish gelatins.Polymer, 47, 6379–6386.

Djerassi, C., Nakano, T., James, A. N., Zalkow, L. H., Eisenbraun, E. J., & Shoolery, J. N.(1961). Terpenoids. XLVII. The structure of genipin. Journal of Organic Chemistry,26, 1192–1206.

Farris, S., Song, J., & Huang, Q. (2010). Alternative reaction mechanism for thecrosslinking of gelatin with glutaraldehyde. Journal of Agricultural and FoodChemistry, 58, 998–1003.

Gómez-Guillén, M. C., Turnay, J., Fernández-Díaz, M. D., Ulmo, N., Lizarbe, M. A., &Montero, P. (2002). Structural and physical properties of gelatin extracted fromdifferent marine species: A comparative study. Food Hydrocolloids, 16, 25–34.

Hardy, P. M., Hughes, G. J., & Rydon, H. N. (1979). The nature of the cross-linking ofproteins by glutaraldehyde. Part 2. The formation of quaternary pyridiniumcompounds by the action of glutaraldehyde on proteins and the identification ofa 3-(2-piperidyl) pyridinium derivative, anabilysine, as a cross-linking entity.Journal of the Chemical Society Perkin Transactions, 1, 2282–2288.

Haug, I. J., Draget, K. I., & Smidsrød, O. (2004). Physical and rheological properties offish gelatin compared to mammalian gelatin. Food Hydrocolloids, 18, 203–213.

Hyun, K., Wilhelm, M., Klein, C. O., Cho, K. S., Nam, J. G., Ahn, K. H., et al. (2011). Areview of nonlinear oscillatory shear test: Analysis and application of largeamplitude oscillatory shear (LAOS). Progress in Polymer Science, 36, 1697–1753.

Imamura, E., Sawatani, O., Koyanagi, H., Noishiki, Y., & Miyata, T. (1989). Epoxycompounds as a new cross-linking agent for porcine aortic leaflets:Subcutaneous implant studies in rats. Journal of Cardiac Surgery, 4, 50–57.

Isenburg, J. C., Karamchandani, N. V., Simionescu, D. T., & Vyavahare, N. R. (2006).Structural requirements for stabilization of vascular elastin by polyphenolictannins. Biomaterials, 27, 3645–3651.

Jobstl, E., O’Connell, J., Fairclough, P. A., & Williamson, M. P. (2004). Molecular modelfor astringency produced by polyphenol/protein interactions.Biomacromolecules, 5, 942–949.

Khor, E. (1997). Methods for the treatment of collagenous tissues for bioprostheses.Biomaterials, 18, 95–105.

Kosaraju, S. L., Puvanenthiran, A., & Lillford, P. (2010). Naturally crosslinked gelatingels with modified material properties. Food Research International, 43,2385–2389.

Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of thehead of bacteriophage T4. Nature, 227, 680–685.

Lapasin, R., & Pricl, S. (1995). Rheology of industrial polysaccharides: theory andapplications. New York: Aspen Publishers Inc.

Liang, H.-C., Chang, W.-H., Liang, H.-F., Lee, M.-H., & Sung, H.-W. (2004). Cross-linking structures of gelatin hydrogels crosslinked with genipin or a water-soluble carbodiimide. Journal of Applied Polymer Science, 91, 4017–4026.

Mi, F.-L., Sung, H.-W., & Shyu, S.-S. (2000). Synthesis and characterisation of a novel-based network prepared using naturally occurring crosslinkers. Journal ofPolymer Science Part A: Polymer Chemistry, 38, 2804–2814.

Mohtar, N. F., Perera, C., & Quek, S. Y. (2010). Optimisation of gelatin extraction fromhoki (Macruronus novaezelandiae) skins and measurement of gel strength andSDS–PAGE. Food Chemistry, 122, 307–313.

Moura, M. J., Figueiredo, M. M., & Gil, M. H. (2007). Rheological study of genipincross linked chitosan hydrogels. Biomacromolecules, 8, 3823–3829.

Nickerson, M. T., Patel, J., Heyd, D. V., Rousseau, D., & Paulson, A. T. (2006). Kineticand mechanistic considerations in the gelation of genipin-cross-linked gelatin.International Journal of Biological Macromolecules, 39, 298–302.

Nimni, M. E., Cheung, D., Strates, B., Kodama, M., & Sheikh, K. (1988). Bioprosthesisderived from cross-linked and chemically modified collagenous tissues. In M. E.Nimni (Ed.), Collagen (pp. 1–38). Florida: CRC Press.

Rosenberg, D. (1978). Dialdehyde starch tanned bovine heterografts development.In P. N. Sawyer & M. J. Kaplitt (Eds.), Vascular grafts (pp. 261–270). New York:Appleton-Century-Crofts.

Ross-Murphy, S. B. (1992). Physical gelation of biopolymers. In: W. G. Glasser, H.Hatakeyama (Eds.), Viscoelasticity of biomaterials (pp. 204–216). Washington:ACS Symposium Series 489.

Simon, A., Grohens, Y., Vandanjon, L., Bourseau, P., Balnois, E., & Levesque, G. (2003).A comparative study of the rheological and structural properties of gelatin gelsof mammalian and fish origins. Macromolecular Symposia, 203, 331–338.

Strauss, G., & Gibson, S. M. (2004). Plant phenolics as cross-linkers of gelatin gelsand gelatin-based coacervates for use as food ingredients. Food Hydrocolloids,18, 81–89.

Sung, H. W., Chang, Y., Chiu, C. T., Chen, C. N., & Liang, H. C. (1999). Crosslinkingcharacteristics and mechanical properties of a bovine pericardium fixed with anaturally occurring crosslinking agent. Journal of Biomedical Materials Research,47, 116–126.

Sung, H. W., Huang, R. N., Huang, L. L., & Tsai, C. C. (1999). In vitro evaluation ofcytotoxicity of a naturally occurring cross-linking reagent for biological tissuefixation. Journal of Biomaterials Science Polymer Edition, 10, 63–78.

Te Nijenhuis, K. (1997). Thermoreversible networks. Advances in Polymer Science,130, 1–252.

Tsai, C. C., Huang, R. N., Sung, H. W., & Liang, H. C. (2000). In vitro evaluation of thegenotoxicity of a naturally occurring crosslinking agent (genipin) for biologictissue fixation. Journal of Biomedical Materials Research, 52, 58–65.

Walt, D. R., & Agayn, V. (1994). The chem istry of enzyme and proteinimmobilization with glutaraldehyde. Trends Analytical Chemistry, 13, 425–430.