at SciVerse ScienceDirect

Food Hydrocolloids 32 (2013) 271e277

Contents lists available

Food Hydrocolloids

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

Interfacial cross-linking of b-casein changes the structure of theadsorbed layer

Riitta Partanen a,*, Pirkko Forssell a, Alan Mackie b, Eva Blomberg c,d

aVTT Technical Research Centre of Finland, P.O. Box 1000, FI-02044 VTT, Finlandb Institute of Food Research, Norwich Research Park, Colney, Norwich NR4 7UA, UKcKTH, School of Chemical Science & Engineering, Drottning Kristinas väg 51, SE-100 44 Stockholm, Swedend YKI, Institute for Surface Chemistry, Drottning Kristinas väg 51, SE-100 44 Stockholm, Sweden

a r t i c l e i n f o

Article history:Received 12 October 2012Accepted 16 January 2013

Keywords:b-CaseinInterfaceCross-linkingRheologyThickness

* Corresponding author. Tel.: þ358 40 825 5816; faE-mail address: riitta.partanen@vtt.fi (R. Partanen

0268-005X/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.foodhyd.2013.01.009

a b s t r a c t

The mechanism of transglutaminase-induced cross-linking of interfacial b-casein layer was investigatedin tetradecane/buffer system. Monolayer studies were carried out in a Langmuir trough, where incu-bation with the enzyme mostly affected the compression of the film through adsorption of trans-glutaminase to the interface. Interfacial shear rheology was used to follow the kinetics of formation ofa visco-elastic film upon cross-linking. Substrate concentration affected the rate of the interfacial cross-linking, when enzyme was dosed per protein concentration. This was most likely due to the saturatedsubstrate layer at the interface in all cases. SDS-PAGE revealed that most of the b-casein at the interfacewas not cross-linked by intermolecular links, but rather, intramolecular links were formed. Finally,studies of adsorbed b-casein layers on polystyrene beads revealed that cross-linking reduced thethickness of the adsorption layer from 11e12 nm to 8e9 nm. These results suggest that it may be mainlyintra-molecular cross-linking which modifies the physical interactions of b-caseins at the interfaceresulting in a higher layer density and thus, formation of a visco-elastic network.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Proteins contribute to food structure as three-dimensionalnetworks in gels and solids but also as two-dimensional net-works at interfaces of foams and emulsions. Interfacial properties ofproteins have been a subject of considerable interest as very dif-ferent films can be formed depending on the interactions and as-sembly of the protein molecules at the interface. Proteins withspecific function on surfaces in nature provide unique properties,such as the extreme elasticity of hydrophobin films at airewaterinterface (Cox, Aldred, & Russell, 2009). In an attempt to improvethe stability of food foams and emulsions, enzymatic means oftailoring interfacial properties of conventional food proteins havebeen sought for over a decade. While controlled hydrolysis can beused to improve foaming and emulsifying properties (Flanagan &FitzGerald, 2002), increased interfacial elasticity and stabilityhave been the aim using of cross-linking enzymes (Dickinson,1997). The impact of enzymatic cross-linking on the properties offoams and emulsions suggest application potential (Dickinson,Ritzoulis, Yamamoto, & Logan, 1999; Ma et al., 2012; Macierzanka

x: þ358 20 722 7071.).

All rights reserved.

et al., 2011; Partanen et al., 2009), even if the effect dependsstrongly on conditions of preparation, either prior to or after cross-linking as well as on the extent of reaction (Dickinson et al., 1999;Flanagan & FitzGerald, 2003). Improved foam stability has beenfound in case of active enzyme present at the time of foaming(Partanen et al., 2009). For emulsions, improved physical stabilityagainst Ostwald ripening (Dickinson et al., 1999) and improvedstability against oxidation (Ma et al., 2012) have been reported. Inaddition, digestibility of protein has been shown to be altered bycross-linking (Macierzanka et al., 2012).

To gain more fundamental understanding of changes in theinterfacial layer resulting from crosslinking, studies with modelinterfaces have been carried out. In particular, transglutaminase(TG, EC 2.3.2.13, g-glutamyl-peptide, amine-g-glutamyl trans-ferase) has been used in these studies (Faergemand &Murray,1998;Faergemand, Murray, & Dickinson, 1997, 1999). A few other studieswith tyrosinase (Ercili-Cura et al., 2012) as well as those withchemical cross-linkers have been reported (Romoscanu &Mezzenga, 2005; Tikekar, Johnson, & Nitin, 2011). The results byFaergemand et al. (1999) show a 100-fold increase in surface shearviscosity of the interfacial films and in dilatational surface elasticity.Dependence of the build-up of surface visco-elasticity on enzymedosage was also demonstrated, as application of high dose led toa steep increase, followed by a decrease in surface shear viscosity

R. Partanen et al. / Food Hydrocolloids 32 (2013) 271e277272

while application of lower doses were followed by a steady increaseof the viscosity. While interfacial cross-linking of non-globular ca-seins has systematically been associated with increased visco-elasticity of the layer, results on interfacial cross-linking of glob-ular proteins are more controversial (Ercili-Cura et al., 2012;Faergemand et al., 1997; Sharma, Zakora, & Qvist, 2002), indicatingthe importance of motional freedom and availability of reactivesites in the substrate protein.

TG-induced reactions of milk proteins in bulk solution havebeen extensively characterized as a function of accessibility to thesubstrate as affected by heat, pH and reducing agents as reviewedby Jaros, Partschefeld, Henle, and Rohm (2006) and for side re-actions such as deamidation under conditions of limited availabilityof the glutamine groups (Sharma, Lorenzen, & Qvist, 2001). At highTG doses changes in electrophoretic mobility of monomers havebeen found, a property which has been linked with deamidationand intra-molecular cross-linking. The interfacial studies introducefurther complications, as in these, the concentration of substrate isusually kept low in order to detect properties of the interfacial layerrather than effects from the bulk. Therefore, the incubation timesare usually very long and/or the amount of the enzyme is abovecatalytic dosage meaning that it is not only affecting the system asa catalyst but potentially as a surface active protein.

In order to develop applications where stirred systemswith lowerenzyme dosages aremore relevant, it is of importance to demonstratethe role of catalytic activity of the enzyme for the induced changes inthe interfacial layer rather than contribution of the enzyme asabuildingblock inthe layer.Moreover, inorder todevelopmulti-phasesystems with interfaces promoting their stability, the structuralchanges taking place need to be assessed. Therefore the present studyinvestigated the mechanism behind the increased visco-elasticity ofthe b-casein film at oilewater interface upon transglutaminase-induced cross-linking. The adsorption of the enzyme to the interfacewas studied ina Langmuir trough tounderstandmore about the effectof the enzyme and its activity on b-casein monolayers. The effects ofenzyme activity onprotein oligomerization at the interface and on thedynamics of the reaction as affected by bulk protein concentrationwere studiedbySDS-PAGEand interfacial shear rheology, respectively.Finally, the effect of transglutaminase on the thickness of b-caseinadsorption layer on hydrophobic polystyrene beads was determined.

2. Materials and methods

2.1. Materials

b-Casein was purchased from Sigma (C6905, SigmaeAldrich Fin-landOy,Helsinki, Finland).Due tosignificantcontributionof impuritiesto the interfacial tension during compression, tetradecane (Aldrich87140,>99%purity)was extracted10 timeswithMilli Q-waterprior touse. Hence, in the Langmuir trough, the contribution of purified tet-radecane on interfacial tension at maximal compression was<2 mN/m, which was considered to be acceptable. The commercial Ca2þ-in-dependent TG product ActivaWM (Ajinomoto Inc., Tokyo, Japan), waspurchased from Vesantti Oy (Helsinki, Finland) and maltodextrin car-rierwas removedasdescribedbyLantto,Puolanne,Kalkkinen,Buchert,and Autio (2005). All other reagents used were of analytical grade.Throughout the study, beta-casein was studied in 10 mM phosphatebuffer (pH 7.0), which was freshly filtered prior to the layer thicknessmeasurements. Carboxyl latex beads (40 nm) were from Invitrogen(C37262, Life Technologies Europe BV, Bleiswijk, Netherlands).

2.2. Langmuir compression isotherms

An interfacial trough (KSV 5000, 362 mm � 54 mm, KSV Nima,Finland) equipped with perforated barriers to allow flow of the

upper bulk phase in compression was used to measure com-pression isotherms at 23.5 � 1.0 �C. The cleanliness of the troughwas first confirmed with water (p < 0.2 mN/m), thereafter waterwas replaced by buffer solution and 38 ml of refrigerated b-caseinsolution (0.1 mg/ml) was evenly applied on the surface drop bydrop with a Hamilton microsyringe. After equilibrating the surfacefor 20 min, tetradecane was carefully spread on the top outside thebarrier with a glass pipette in contact with the troughwall until theWilhelmy plate was submerged. The balance was zeroed, and thelayer was equilibrated for 30 min before starting the compressionat 1 mm/min. Transglutaminase was added either prior to com-pression (at an area per molecule of 8200�A2/molecule for b-casein)or after partial compression (area per molecule of 5000 or 4000�A2/molecule for b-casein). All calculations for molecular area wereperformed assuming negligible loss into the sub-phase duringspreading the protein. In preliminary trials, spreading of cold pro-tein solution led to a higher surface pressure in compression, whichwas taken as an indication of a higher retention of the protein onthe surface. Enzymewas injected by two injections with a Hamiltonsyringe to the buffer sub-phase and allowed to self-diffuse to thetetradecane/buffer interface. The amount of enzyme per interfacialarea was equal to that used in rheological experiments with a pro-tein concentration of 0.001%. The interface was incubated for 8 hwith the enzyme prior to compression. The monolayer was firstcompressed to 1800 �A2/molecule for casein, expanded to 9400 �A2/molecule, where it was relaxed for 2 h, after which a secondcompression cycle was performed. In the case of o/w measure-ments, evaporation was not a problem, but in case of a/w studies,a control experiment with buffer alonewas performed and found toaccount for an error of �0.2 mN/m per compression cycle due toevaporation of water.

2.3. Interfacial shear rheology

Interfacial shear measurements were carried out in an AR-G2rheometer equipped with a Du Noüy ring (TA instruments, UK)with slight modification from Ercili-Cura et al. (2012). Visco-elasticproperties of the layer at 25 �C and b-casein concentrations of10�1%,10�3% and 10�4% were used in the studies. Exactly 45 g of thesolutions was weighed in the measuring cup (Ø70 mm) which wasplaced on the bottom plate of the instrument. The ring was wettedin the solution and its position was adjusted to the interface. It isdifficult to position the ring in a repeatable manner just by the eye.Consistent positioning in the measurements was ensured byweighing the liquid phase, and always using the same distancefrom the bottom plate. The proper position was determined in thepreliminary experiments as the position of the ring which gave thehighest value for elastic modulus. After adjusting the ring position,10 ml of purified tetradecane was gently added on the top withglass pipette via the cup wall after which the time sweep wasimmediately started. Transglutaminase was added after 1 h in thesub-phase with a single injection by Hamilton syringe in order tocause minimal disturbance to the formed interface. Trans-glutaminase was added at 50,000 nkat/g of b-casein in the case ofconstant enzyme/substrate ratio experiments. Experiments withconstant enzyme/interfacial area ratio were also performed, inwhich the dose was the same for 10�3% b-casein concentration, but10-fold greater for 10�4%, and 10-fold less for 10�2% b-casein so-lutions. The linear visco-elastic region was confirmed for the finalfilm by a frequency sweep and finally a strain sweep was used toevaluate the strain at which the interfacial layer ruptured. For timesweeps, a frequency of 0.005 Hz and 2% strain were used. Lowfrequency was used to improve sensitivity of the measurement forweak b-casein layers. Inactivated transglutaminase (o/n at 105 �C)was used as a control.

R. Partanen et al. / Food Hydrocolloids 32 (2013) 271e277 273

2.4. Oligomerization of interfacial protein

Cross-linked b-casein interfaces were formed under the sameconditions as the interfacial measurements. After 24 h incubation at23e25 �C, the reaction was stopped by 15 min incubation with0.45 ml of 1 M NH4Cl. Then, 40 ml of the sub-phase was gentlyremoved by a glass pipette with a narrow tip. The upper phase wasgently moved into a concentration tube (Vivaspin 2O, 3000 MWCOPES). Centrifugation was performed at 6800 rpm for 4 h, afterwhich excess tetradecane was removed from the top, and a 120 mlsample was taken from the aqueous phase above the membraneand mixed directly with 40 ml of sample buffer for SDS-PAGE. TriseGlycine gel (12%, Lonza, Basel, Switzerland) was used with pre-stained broad range standard (Bio-Rad, Hercules, CA, USA) and b-casein and transglutaminase were injected on separate lanes tofacilitate interpretation. Also, in addition to tetradecane buffersystem, the reactions were performed on airebuffer system.

2.5. Layer thickness estimation

The thickness of the interfacial layer as affected bytransglutaminase-induced reactions was studied by adsorptiononto polystyrene beads with the method by Mackie, Mingins, andNorth (1991). In order to maximize the resolution and preventany competitive adsorption by bead dispersing agent, covalentlystabilized beads (sulphonated) of 40 nm diameter were used. Thesize distribution of the beads was first measured in buffer solutionby photon correlation spectroscopy (PCS) with Zetasizer ZS3600(Malvern Instruments, UK), after which b-casein was added atdifferent concentrations per calculated surface area of the beads.After overnight incubation without mixing at 23e25 �C, the sizedistribution was measured again after which the enzyme wasapplied again followed by o/n incubation. The particle sizes wereagain measured and from the z-averaged cumulants, layer thick-ness was estimated by subtracting the radius of the original beadsfrom the radius of the bead with adsorbed protein. Duplicatesamples with triplicate measurements were made, of which aver-age value and standard deviation was calculated.

3. Results and discussion

3.1. peA isotherms on Langmuir interfacial trough

The mean molecular area of b-casein was determined for airebuffer and for tetradecaneebuffer interfaces in order to evaluate

-2

0

2

4

6

8

10

12

0 5000 10000

)m/

Nm(

erusserpecafruS

2)

A

Fig. 1. Surface pressure of b-casein in compression at 1 mm/min for A) airebuffer, and B) tetand 3rd compression (dotted line). Average curves of the replicate measurements were plotte27.5 � 0.5 mN/m for tetradecaneebuffer interface.

the monolayer properties of b-casein film. At the same mean mo-lecular area (1800 �A2), a much higher surface pressure wasobserved for tetradecaneebuffer interface (27.5 mN/m) comparedwith airebuffer interface (10.7 mN/m). Less difference betweenconsecutive compressions was found for tetradecaneebuffer sys-tem (Fig. 1), where some intermolecular interactions or orientationseem to remain after the first compression despite the relaxationstep between the compressions. Compression of b-casein at airebuffer interface leads to a systematic irreversible decrease of sur-face pressure. The small difference in the starting levels of thesecond and third compressions of airebuffer interface is most likelydue to evaporation of water from the trough during experiment.The differences between the airebuffer and tetradecaneebufferinterfaces suggest that the protein molecules are more firmlyattached to the oilewater interface.

Upon finishing each compression step, an immediate and steepdecrease of the surface pressure was observed for both interfaceseven if the barriers were only stopped and not expanded. This in-dicatesmost likely lossofproteindue todesorption fromthe interfaceor major re-organization of the protein at the interface, such asexpulsion of tails and loops of the protein as suggested by Mellema,Clark, Husband, and Mackie (1998). At a 5 mm/min compressionrate, a higher surface pressure was obtained in the end, indicatingthat the ability of the interface to organize during rapid compressionwas less, and the systemwas therefore further from the equilibrium(result not shown). Thus, despite the slow compression rate, equi-libriumwas not maintained during compression. Moderate hystere-sis has been reported previously for peA isotherms of b-casein(Dauphas et al., 2005). The equilibrium and non-equilibrium aspectsof b-casein adsorption on tetradecaneewater interface have beendiscussed by Maldonado-Valderrama et al. (2005). The authors havedemonstrated the different interfacial conformations at oilewaterinterface and related the structural re-organization with interfacialcoverage in equilibrium state. Given long enough time, protein in-terfaces re-organize not only in conformation, but depending on thebulk concentration theymayalso formmultilayers, aggregates or gelson interfaces.

Transglutaminase was added in the sub-phase of equilibratedinterfaces of b-casein in expanded and compressed states followedby 8 h incubation prior to the 1st compression (Fig. 2A). It washypothesized that some motional freedom in the interfacial layer isneeded for cross-linking to take place. Incubation of the substrateprotein layer with the enzyme at different degrees of compressionresulted in difference in surface pressure levels. The treatment ofthe layers at 8200�A2/molecule and 5000�A2/molecule was followed

0

5

10

15

20

25

30

0 5000 10000

Surf

ace

pres

sure

(mN

/m)

2)

B

radecaneebuffer interface: 1st compression (solid line), 2nd compression (dashed line)d, where maximum surface pressure was 10.7 � 0.2 mN/m for airebuffer interface, and

0

5

10

15

20

25

30

35

0 5000 10000

Surf

ace

pres

sure

(mN

/m)

2)

A

0

5

10

15

20

25

30

35

40

0 5000 10000

Surf

ace

pres

sure

(mN

/m)

2)

B

Fig. 2. Addition of transglutaminase in the sub-phase of equilibrated tetradecaneebuffer interfaces of b-casein in expanded (8200 �A2/molecule: dashed line), and compressed(5000 �A2/molecule: dotted line; 4000 �A2/molecule: dotted dashed line) states followed by 8 h incubation and subsequent compression. A) 1st compression, the gray vertical lineshows the starting point of compression after incubation. The curve below the discontinuity point is the compression prior to enzyme addition. Mean molecular area is calculatedfor b-casein. B) 2nd compression, where the solid black line is for b-casein and solid gray line is for transglutaminase adsorbed from the sub-phase on equilibrated tetradecaneebuffer interface with 8 h equilibration time prior to compression. The mean molecular area is calculated assuming that all the enzyme is adsorbing. The arrow shows the x-axis shiftin the cross-linked curves in case of adsorption of transglutaminase.

R. Partanen et al. / Food Hydrocolloids 32 (2013) 271e277274

by an increase in surface pressure, but the treatment at 4000 �A2/molecule had minor effect on the surface pressure of the layer. Thesubsequent compression of the layer treated at 4000�A2/molecule isvery similar to the 1st compression of the pure b-casein layershown in Fig. 1B. After relaxation of the cross-linked layers inexpanded state for 2 h the differences had evened up (Fig. 2B).Thereafter, surface pressure response to the 2nd compression of allthe layers was nearly the same and considerably higher than thenon-cross-linked b-casein layer.

Upon cross-linking, increasing intermolecular interactions couldcontribute to compressibility of the film in different ways: desorp-tion of the protein from the interface could be hindered and also thefilm becoming more rigid, the proteins may be less able to re-organize at the interface upon compression. Both of these wouldbe expected to lead to decreased compressibility, and thus, to highersurface pressure. In the experiments with TG, however, surfacepressure already increased during the static incubation with TG, atsurface pressures<15mN/m. Yet another possibility is the formationof an elastic film, which may not be able to fully relax in theexpanded state. However, build-up of elasticity indicating change inmolecular interactions could not be observed in the shear rheologymeasurements during 8 h incubation with TG at correspondingsurface concentrations. These observations suggest the possibilitythat the enzyme can adsorb to the interface given enough time andspace, up to a surface pressure of 15 mN/m. In our previous workwith airewater interfaces, we have not observed any contribution ofTG on the surface pressure of equilibrated b-lactoglobulin interfaces(Ercili-Cura et al., 2012). In those experiments surface pressure of theequilibrated substrate protein was 20 mN/m and thus above thecritical limit indicated by the present study. In the present study, theadsorption of the enzyme to the interface would decrease the valuesof mean molecular area at the mixed interface, and thus shift thecurves to the left (indicated by an arrow in Fig. 2B). For comparison,the compression of TG after 8 h adsorption is plotted in Fig. 2B. Theposition of the isotherm is theoretical, assuming 100% efficiency inadsorption of TG. Still, this data reveals the presence of TG at theinterface, given high enough concentration.

3.2. Interfacial shear rheology

The effect of cross-linking on the interactions of proteins andthus, visco-elasticity of the layer was studied by interfacial shear

rheology. An equilibrated b-casein layer was too weak for elastic orviscous modulus to be measured, which is in accordance with itswell-reported surfactant-like character (Faergemand et al., 1999).Upon addition of active transglutaminase however, both modulistarted to build up simultaneously after a lag phase (Fig. 3A and B).In all cases, build-up of both elastic (G0) and viscous (G00) modulitook place simultaneously but the elastic response dominated thebehavior. A maximum in surface shear viscosity, which wasreported by Faergemand et al. (1999) for transglutaminase-inducedcross-linking of b-casein, was not observed for the viscous inter-facial shear modulus in the present study. However, the filmsformed from lower concentrations had longer lag phases and werestill developing at the end of the incubation time. The increase inviscous modulus had leveled off in the 0.0001% system. Interest-ingly, not only the 0.001% system with active enzyme but also thesystem with inactivated enzyme slowly developed a detectableviscous modulus.

Saturation of a b-casein layer at around 3 mg/m2 is indicated bythe work of Husband, Wilde, Mackie, and Garrood (1997). In thepresent study, the most dilute concentration led to a theoreticalsurface concentration of 11 mg/m2. Thus, the layer to be cross-linked could be saturated in all experiments, and therefore inde-pendent of the bulk concentration. As the enzymewas dosed basedon the amount of total substrate protein in the system, the amountof enzyme was 100-fold for 0.01% solution as compared with0.0001% solution. The higher rate of modulus increase was there-fore to be expected. However, another experiment was performed,where enzyme was dosed per interfacial area (results not shown),i.e. equal amount of enzyme for all the samples, the rise in elasticmodulus was fastest in the most dilute system. In order to act onthe interfacial layer, the enzyme has to diffuse through the bulkphase. Possibly, the collisions with the substrate in the bulk andsubsequent catalytic activity, which is not contributing to theinterfacial but bulk rheology, reduce the rate of diffusion of theenzyme through the bulk solution.

The low frequencyused in the time sweepwas tominimize inertiain themeasurement and thereby increase the sensitivity to change ofa fluid layer into a visco-elastic film. However, as the oscillation fre-quency determines the time-scale of themolecular processes that areprobed in the measurement, a frequency sweep (0.001e1 Hz) wasperformed after each time sweep (Fig. 3C). Up to 0.1 Hz, there wasslow, linear increase in G0. Toward higher frequencies, the increase in

0.000

0.002

0.004

0.006

0.008

0.010

0 200 400 600 800

G' (

N/m

)

Time (min)

+TG

0.000

0.001

0.002

0.003

0 200 400 600 800

G''

(N/m

)

Time (min)

+TG

A B

0.0001

0.001

0.01

0.1

0.001 0.01 0.1 1

G' (

N/m

)

Frequency (Hz)

C

0.0001

0.001

0.01

0.1

0.1 1 10 100

G' (

N/m

)

Strain (%)

D

Fig. 3. Time sweeps of transglutaminase-treatment of b-casein films at 25 �C. Single measurement is shown and the final modulus value with standard deviation is given for eachsample after description. The time of TG addition is shown by arrow. Shear interfacial A) elastic and B) viscous modulus of b-casein at 0.01% (dotted line, G0 ¼ 8.1 � 0.6 mN/m;G00 ¼ 2.0 � 0.2 mN/m), at 0.001% (dashed line, G0 ¼ 5.9 � 0.6 mN/m; G00 ¼ 1.1 � 0.3 mN/m), at 0.0001% (solid line, G0 ¼ 4.8 � 0.8 mN/m; G00 ¼ 0.4 � 0.1 mN/m), and at 0.001% withinactivated TG (dash dotted line, G0 ¼ 0.2 mN/m; G00 ¼ 0.2 mN/m, single measurement). C) Frequency and D) strain sweeps for the cross-linked films. The gray vertical bars show thefrequency and strain used in the other two sweeps.

Fig. 4. SDS-PAGE gel interfacial layers of b-casein cross-linked with transglutaminase(TG) at 25 �C overnight. Lanes: 1) molecular weight standard; 2) 0.0001% b-casein þ TG(a/w); 3) 0.001% b-casein þ TG (a/w); 4) 0.01% b-casein þ TG (a/w); 5) TG control; 6) b-casein control; 7) 0.001% b-casein þ TG (o/w) and 8) 0.01% b-casein þ TG (o/w).

R. Partanen et al. / Food Hydrocolloids 32 (2013) 271e277 275

G0 becamenon-linear. Even in the high concentration case, where theincrease of themoduli has leveled off in the time-sweep experiment,the filmwas still going through minor re-organizations. Similar me-chanical spectra are also typical for 3D gel networks, which, in con-trast to entangled networks show little dependency on frequency(Clark & Ross-Murphy, 1987). Also the strain sweep measurement(Fig. 3D) confirmed the linear visco-elastic region for the time sweep,where build-up of visco-elasticity was followed. Most of the filmscollapsed between 20 and 40% strain, but the lowest strain beforecollapse was 12% for the 0.01% system, and the highest stain beforecollapsewas 60% for the 0.0001% system. Thus the higher elasticity ofthe film was generally associated with earlier breakage, but withinthe replicates deviation was high.

3.3. Extent of protein cross-linking by SDS-PAGE

Themodulus could be increasing for two reasons. Firstly, if moreprotein becomes associated with the interfacial layer due to cross-linking, and secondly, if the interactions between the adsorbedproteins are changed. The covalent links forming between proteinsare known to increase the hardness of transglutaminase-treatedsodium caseinate gels (Myllärinen, Buchert, & Autio, 2007), andcould be expected to work in a similar way in an interfacial layer. Inorder to evaluate the extent of intermolecular cross-links, the re-actions were performed in the same vessel but without the oscil-lating probe, and the interfacial layer was separated from the bulk.The SDS-PAGE for tetradecaneebuffer (lanes 7 and 8) and airebuffer (lanes 2e4) interfaces is shown in Fig. 4. The intensities ofbands between different cross-linked interfaces are not comparable

due to the extraction and concentration treatments. The presenceof transglutaminase enzyme (main band, lane 5) is observed at allinterfaces. Only traces of native b-casein (lane 6) could be observedat the cross-linked interfaces. Instead, an intense band of proteinwith higher mobility than the native monomer was observed on allcross-linked samples (lanes 2e4, 7 and 8). This band is most likelydue to formation of non-native, intra-linked monomers formedupon interfacial cross-linking. Such monomers could be formed viacovalent linkswithin single b-casein molecules, causing incompletelinearization of the protein by SDS and b-mercapto-ethanol. Similarbehavior due to TG-induced cross-linking has been previously

Table 1Interfacial layer thickness for polystyrene beads incubated for two days with b-casein, or one day with b-casein and one with active or inactivated TG. Standarddeviation (�) is given after the average value.

Concentration (mg/m2) Layer thickness (nm)

b-Casein b-Casein withinactivated TG

b-Casein withactive TG

5 11.6 � 0.2 11.6 � 0.2 8.9 � 0.310 11.2 � 0.1 11.3 � 0.4 8.2 � 0.4

R. Partanen et al. / Food Hydrocolloids 32 (2013) 271e277276

reported for ovalbumin and b-lactoglobulin by Giosafatto et al.(2012) and Ercili-Cura et al. (2012), respectively. The presence ofnon-native monomers, intra-linked monomers was suggested alsoin these studies. In size-exclusion chromatography, altered elutiontime of enzymatically cross-linked monomers has been demon-strated (Hellman, Mattinen, Fu, Buchert, & Permi, 2011). In thepresent study, cross-linked oligomers are found in addition to theintra-linked monomers. Due to multiple cross-linking sites for TGin b-casein, the oligomers appear as smear rather than distinctdimers, trimers and tetramers. As compared with the intensity ofaltered monomer band, the intensity of oligomers seems to be less.The presence of insoluble, high-molecular weight polymersremaining in the sample loading wells was not observed. Therefore,it is possible that rather than forming a continuously cross-linkednetwork, transglutaminase is catalyzing formation of non-nativemonomers and small oligomers. Thus, the increased elasticity ofthe layer could be explained by physical interactions betweenthese.

3.4. Thickness of the b-casein adsorption layer

In order to understand whether the change in rheology of thelayer upon TG-induced crosslinking is due to increased amount ofprotein associated with the layer or due to the decreased repulsionswithin the existing layer, the effect of TG on the thickness ofadsorbed b-casein layer was studied by PCS. Real oil-droplets arenot suitable as an adsorption surface as they can become unstable,mainly through coalescence in the case of the present study. Thuscarboxylated polystyrene beads were used as a stable hydrophobicadsorption media. The absolute thickness of the layer most likelydiffers to some extent from the layer at tetradecaneebuffer inter-face due to the penetration of the protein into tetradecane. Smallsize of the beads was selected to keep them dispersed by Brownianmotion, but also, to improve the resolution of the layer thicknessmeasurement.

In Fig. 5, the layer thickness values for b-casein adsorption layersbefore and after TG treatment are presented. The adsorption layerof b-casein was ca. 7 nm for 1 mg/m2. At higher concentrations, 2e10 mg/m2, the layer was saturated with 10e12 nm thickness. Thesevalues are in accordance with the study by Husband et al. (1997). Inthe present study, as TG was added to the bead dispersion withequilibrated adsorption layers of b-casein, a clear reduction in layerthickness was observed after overnight incubation. The dosage ofthe enzyme (1000 nkat/g or 10,000 nkat/g) did not affect the

0

2

4

6

8

10

12

14

0 2 4 6 8 10

Lay

er th

ickn

ess

(nm

)

-casein/bead surface area (mg/m2)

Fig. 5. Thickness of b-casein adsorption layers on polystyrene beads. The symbols arefor b-casein control (circles), for b-casein treated with 1000 nkat/g TG (squares) and b-casein treated with 10 000 nkat/g TG (triangles). Standard deviation is given in errorbars.

thickness of the layer, indicating that rather than acting asa building block in the layer, the catalytic activity of TG was ofimportance. At high dosage, however, particle size distributionrevealed a minor fraction of aggregated material in addition to theindividual beads with increased radius. This could be due to partialloss of repulsion or formation of covalent links between the beadsdue to TG activity. Reduction in the layer thickness was small butrepeatable. Further workwould be needed to determine the preciseorigin of this densification.

In order to confirm that the reduction in layer thickness was notdue to desorption of b-casein with time from the surface or due tothe enzyme as a physically interacting macromolecule, controlmeasurements for extended incubation without enzyme and withinactivated enzyme were performed at 5 mg/m2 and 10 mg/m2

concentrations (Table 1). Neither the inactivated enzyme nor theprolonged incubation time had an effect on layer thickness. Theseobservations support the interpretation that the effect of TG on theadsorption layer is due to its catalytic action.

4. Conclusions

Given high enough concentration and “free” interface, TG wasshown to adsorb on the tetradecaneebuffer interface. However,with the low surface concentration of the b-caseins in the Langmuirtrough studies, no elastic film could be formed by the addition ofTG. At saturation concentrations of b-casein, TG-induced cross-linking was confirmed to lead to the development of visco-elasticity of the adsorbed layer. The build-up of visco-elasticitywas explained by an increase in the density of the layer causedby increased interaction between b-caseins as observed in theadsorption layer thickness measurements. It is suggested thatphysical interactions between b-caseins also play a role in thisprocess as little evidence for covalently linked three-dimensionalnetwork was found. Instead, b-casein oligomers and non-nativemonomers with altered mobility, most likely because of incom-plete linearization due to the formation of intra-molecular cross-links by TG, were observed. Increased density of b-casein adsorp-tion layer was indicated by the reduction in the thickness of thelayer 11e12 nm to 8e9 nm due to cross-linking. In addition to theimproved mechanical strength of the interfacial layer, the higherdensity may also protect the layer from desorption of b-caseins andreduce the contact between phases in an emulsion system. Thesemay be important in consideration of mass transfer propertiesacross the interface which affect not only physical but also oxida-tive stability of emulsions, and are crucial in development ofemulsion-based encapsulation systems.

Acknowledgments

The work was carried out with financial support from theAcademy of Finland in Project “Novel Protein-Based Emulsions byEngineering Interfacial Mass Transfer” (No. 133339). Heljä Heikki-nen (VTT) is thanked for excellent technical assistance.

R. Partanen et al. / Food Hydrocolloids 32 (2013) 271e277 277

References

Clark, A., & Ross-Murphy, S. (1987). Structural and mechanical properties of bio-polymer gels. Advances in Polymer Science, 83, 57e192.

Cox, A. R., Aldred, D. L., & Russell, A. B. (2009). Exceptional stability of food foamsusing class II hydrophobin HFBII. Food Hydrocolloids, 23, 366e376.

Dauphas, S., Mouhous-Riou, N., Metro, B., Mackie, A. R., Wilde, P. J., Anton, M., et al.(2005). The supramolecular organisation of b-casein: effect on interfacialproperties. Food Hydrocolloids, 19, 387e393.

Dickinson, E. (1997). Enzymic crosslinking as a tool for food colloid rheology controland interfacial stabilization, a review. Trends in Food Science & Technology, 8,334e339.

Dickinson, E., Ritzoulis, C., Yamamoto, Y., & Logan, H. (1999). Ostwald ripening ofprotein-stabilized emulsions: effect of transglutaminase crosslinking. Colloidsand Surfaces B: Biointerfaces, 12, 139e146.

Ercili-Cura, D., Partanen, R., Husband, F., Ridout, M., Macierzanka, A., Lille, M., et al.(2012). Enzymatic cross-linking of b-lactoglobulin in solution and at airewaterinterface: structural constraints. Food Hydrocolloids, 28, 1e9.

Faergemand, M., & Murray, B. S. (1998). Interfacial dilatational properties of milkproteins cross-linked by transglutaminase. Journal of Agricultural and FoodChemistry, 46, 884e890.

Faergemand, M., Murray, B. S., & Dickinson, E. (1997). Cross-linking of milk proteinswith transglutaminase at the oilewater interface. Journal of Agricultural andFood Chemistry, 45, 2514e2519.

Faergemand, M., Murray, B. S., Dickinson, E., & Qvist, K. B. (1999). Cross-linking ofadsorbed casein films with transglutaminase. International Dairy Journal, 9,343e346.

Flanagan, J., & FitzGerald, R. J. (2002). Functionality of Bacillus proteinase hydro-lysates of sodium caseinate. International Dairy Journal, 12, 737e748.

Flanagan, J., & FitzGerald, R. J. (2003). Functional properties of Bacillus proteinasehydrolysates of sodium caseinate incubated with transglutaminase pre- andpost-hydrolysis. International Dairy Journal, 13, 135e143.

Giosafatto, C. V. L., Rigby, N. M., Wellner, N., Ridout, M., Husband, F., & Mackie, A. R.(2012). Microbial transglutaminase-mediated modification of ovalbumin. FoodHydrocolloids, 26, 261e267.

Hellman, M., Mattinen, M. L., Fu, B., Buchert, J., & Permi, P. (2011). Effect of proteinstructural integrity on cross-linking by tyrosinase evidenced by multidimen-sional heteronuclear magnetic resonance spectroscopy. Journal of Biotechnology,151, 143e150.

Husband, F., Wilde, P., Mackie, A., & Garrood, M. (1997). A comparison of thefunctional and interfacial properties of b-casein and dephosphorylated b-casein. Journal of Colloid and Interface Science, 195, 77e85.

Jaros, D., Partschefeld, C., Henle, T., & Rohm, H. (2006). Transglutaminase indairy products: chemistry, physics, applications. Journal of Texture Studies,37, 113e155.

Lantto, R., Puolanne, E., Kalkkinen, N., Buchert, J., & Autio, K. (2005). Enzyme-aidedmodification of chicken-breast myofibril proteins: effect of laccase and trans-glutaminase on gelation and thermal stability. Journal of Agricultural and FoodChemistry, 53, 9231e9237.

Ma, H., Forssell, P., Kylli, P., Lampi, A.-M., Buchert, J., Boer, H., et al. (2012). Trans-glutaminase catalyzed cross-linking of sodium caseinate improves oxidativestability of flaxseed oil emulsion. Journal of Agricultural and Food Chemistry, 60,6223e6229.

Macierzanka, A., Bordron, F., Rigby, N. M., Mills, E. N. C., Lille, M., Poutanen, K., et al.(2011). Transglutaminase cross-linking kinetics of sodium caseinate is changedafter emulsification. Food Hydrocolloids, 25, 843e850.

Macierzanka, A., Böttger, F., Rigby, N. M., Lille, M., Poutanen, K., Mills, E. N. C., et al.(2012). Enzymatically structured emulsions in simulated gastrointestinalenvironment: Impact on interfacial proteolysis and diffusion in intestinalmucus. Langmuir, 28, 17349e17362.

Mackie, A. R., Mingins, J., & North, A. (1991). Characterisation of adsorbed layers ofa disordered coil protein on polystyrene latex. Journal of the Chemical Society,Faraday Transactions, 87, 3043e3049.

Maldonado-Valderrama, J., Fainerman, V. B., Aksenenko, E., Galvez-Ruiz, M. J.,Miguel, A., Cabrerizo-Vilchez, M. A., et al. (2005). Dynamics of proteinadsorption at the oilewater interface: comparison with a theoretical model.Colloids and Surfaces A: Physicochemical and Engineering Aspects, 261, 85e92.

Mellema, M., Clark, D. C., Husband, F. A., & Mackie, A. R. (1998). Properties of b-casein at the air/water interface as supported by surface rheological measure-ments. Langmuir, 14, 1753e1758.

Myllärinen, P., Buchert, J., & Autio, K. (2007). Effect of transglutaminase on rheo-logical properties and microstructure of chemically acidified sodium caseinategels. International Dairy Journal, 17, 800e807.

Partanen, R., Paananen, A., Forssell, P., Linder, M., Lille, M., Buchert, J., et al. (2009).Effect of transglutaminase-induced cross-linking of sodium caseinate on theproperties of equilibrated interfaces and foams. Colloids and Surfaces A: Phys-icochemical and Engineering Aspects, 344, 79e85.

Romoscanu, A. I., & Mezzenga, R. (2005). Cross linking and rheological character-ization of adsorbed protein layers at the oilewater interface. Langmuir, 21,9689e9697.

Sharma, R., Lorenzen, P. C., & Qvist, K. B. (2001). Influence of transglutaminasetreatment of skim milk on the formation of ε-(g-glutamyl)lysine and the sus-ceptibility of individual proteins towards crosslinking. International DairyJournal, 11, 785e793.

Sharma, R., Zakora, M., & Qvist, K. B. (2002). Characteristics of oilewater emulsionsstabilised by an industrial a-lactalbumin concentrate, cross-linked before and afteremulsification, by a microbial transglutaminase. Food Chemistry, 79, 493e500.

Tikekar, R. V., Johnson, A., & Nitin, N. (2011). Real-time measurement of oxygentransport across an oilewater emulsion interface. Journal of Food Engineering,103, 14e20.