effects of protein separation conditions on the functional and thermal properties of canola protein...

E:FoodEngineering&PhysicalProperties

Effects of Protein Separation Conditionson the Functional and Thermal Propertiesof Canola Protein IsolatesWajira A. R. Manamperi, Dennis P. Wiesenborn, Sam K.C. Chang, and Scott W. Pryor

Abstract: Canola meal protein isolates were prepared from defatted canola meal flour using alkaline solubilization andacid precipitation. A central composite design was used to model 2nd-order response surfaces for the protein yield andthe functional properties of protein isolates. The solubilization pH and precipitation pH were used as design factors. Themodels showed that the protein yield and functional properties of isolates, such as water absorption and fat absorption, weresensitive to both solubilization pH and precipitation pH, whereas the emulsification was sensitive to only solubilizationpH. Gel electrophoresis analysis of protein fractions gave evidence to the compositional changes between proteins isolatedunder different conditions. Differences in glass transition temperatures suggest that proteins tend to be more denaturedwhen solubilized at highly alkaline conditions. These conformational and compositional changes due to different proteinseparation conditions have contributed to the changes in functional properties of protein isolates.

Keywords: canola, functional properties, predictive modeling, protein isolation, proteins

Practical Application: Protein isolation conditions may be determined primarily through optimization of total proteinyield. Improvements in protein functional properties may be achieved with a relatively small sacrifice in yield by alteringisolation conditions.

IntroductionThe plant-based protein isolate industry is fuelled by the avail-

ability of low-cost, high-volume byproducts from a range ofprotein-rich crops (Aluko and Yada1995). These products havea broad portfolio of potential food and nonfood applications(Ahmedna and others 1999; Kumar and others 2002). The soyprotein industry is a leader in both food and nonfood applicationsof protein isolates, serving as an excellent model for uses of othercrops (Kumar and others 2002). The canola protein industry isrelatively young and has not had as much opportunity or volumeto develop. Therefore, canola has found fewer food and nonfoodapplications of protein isolates, despite having a relatively high seedand meal protein content.

Commercial use of flour, protein concentrates, and protein iso-lates from canola meal has been limited by their low value as foodadditives due to the presence of antinutritional components such asglucosinolates, phytates, and phenolics (Diosady and others 1990;McCurdy 1990); however, breeding programs have reduced thesecomponents in canola meal (Tzeng and others 1990; Cumby andothers 2008; Wu and Muir 2008), and protein isolation proceduresmake canola protein isolates potentially an important food ingre-

MS 20101018 Submitted 9/9/2010, Accepted 1/10/2011. Authors Manamperi,Wiesenborn, and Pryor are with Dept. of Agricultural and Biosystems Engineering,North Dakota State Univ., Dept. 7620, P.O. Box 6050, Fargo, ND 58108-6050,U.S.A. Author Chang is with Dept. of Cereal and Food Sciences, North DakotaState Univ., Dept. 7620, P.O. Box 6050, Fargo, ND 58108-6050, U.S.A. Directinquiries to author Pryor (E-mail: [email protected]).

dient. Also, the excellent profile of essential amino acids in canolaproteins would make it a strong candidate for food applications(Dua and others 1996).

Oilseed protein isolates are generally prepared by solubilizingthe oilseed meal/flour in alkaline solutions to extract proteins, re-moving all nonsoluble material from the extract and then precipi-tating the protein-rich supernatant by the addition of acid (Kinsella1979). Protein extraction from meal/flour is affected by variousfactors such as particle size, thermal history, solvent-to-meal ra-tio, extraction time and temperature, pH, and ionic strength ofthe solution (Kinsella 1979). The pH of the solvent is one of themost sensitive and highly influential parameters among these fac-tors (Aluko and Yada 1995; Ragab and others 2004; Selmane andothers 2008). In protein isolation procedures, both solubilization(extraction) and precipitation (purification) steps contribute to theprotein yield. Therefore, protein recovery should be maximized ateach step in order to maximize the overall protein yield. Generally,high solubilization pH values (around 12) and low precipitationpH values (around 4.5) are expected to increase the extraction andpurification of canola proteins (Diosady and others 1990; Tzengand others 1990; Ghodsvali and others 2005). The extraction andprecipitation pH values may also have an effect on the behavior ofthe protein isolates in food and nonfood applications, but reportson such impacts are rare in literature.

Functional properties, such as water absorption, fat absorption,and emulsifying activity, are important indicators that can be usedto predict the behavior of protein isolates in the food systems(Kinsella 1979; Ohren 1981). In general, higher water absorption,fat absorption, and emulsifying activity are beneficial for the foodindustry (Kinsella 1979; Dua and others 1996). Therefore, efforts

C© 2011 Institute of Food Technologists R©E266 Journal of Food Science � Vol. 76, Nr. 3, 2011 doi: 10.1111/j.1750-3841.2011.02087.x

Further reproduction without permission is prohibited

E:Fo

odEn

ginee

ring&

Phys

icalP

ropert

ies

Effects of protein separation on properties . . .

have been taken to enhance these properties in protein productsby various means (Dua and others 1996; Li and others 2005). Thefunctional properties of proteins can show differences accordingto the various processes used in oil extraction from seeds (such aspressing, cooking, flaking, and desolventizing under high pressuresand temperatures) as well as conditions used in protein isolationprocedures (such as solubilization and precipitation pH values andsolution ionic strength) (Kinsella 1979; McCurdy 1990; Aluko andYada 1995). Several reports have stated that pH plays a significantrole in determining the functionality (water absorption, emulsifi-cation, and gelation) of different protein sources such as canola,cowpea, and soy (Kinsella 1979; Aluko and Yada 1995; Uruakpaand Arntfield 2004).

In this study, the effects of protein solubilization and precipita-tion pH values on the yield and functional properties of canola pro-tein isolate were determined. Response surface analysis was usedto examine these effects and a central composite design (CCD)with 2 independent variables (solubilization pH and precipitationpH) was selected for model development.

Materials and MethodsCanola seeds (cv. Invigor 2573) grown in Cavalier County in

northeastern North Dakota) were purchased locally and cleanedaccording to USDA-GIPSA recommendations using a Carter-Day dockage tester (Minneapolis, Minn., U.S.A.) and hand sieves.Commercial-grade canola oil was acquired locally.

Preparation of canola meal flourCanola meal was prepared from raw canola seeds by screw press-

ing. Moisture content of the seeds was adjusted to 7% (wet basis)by adding distilled water and equilibrating overnight in sealed plas-tic bags. Screw pressing of canola seeds was then carried out usinga model S 87G Komet screw press (Monchengladbach, Germany)preheated to 70 ◦C at a feed rate at steady flow of around 80 g/minand a screw rotation speed of 24 rpm. A compression screw R6and a restriction die with a 6-mm die opening were used. Theresulting canola meal pellets were pressed again using the sameconditions (except for moisture adjusting) to further remove oil.The residual oil content of the twice-pressed meal was about 8%.Canola meal flour was produced by grinding the twice-pressedcanola meal pellets using a Retsch ZM1 mill (Brinkmann Instru-ments Inc., Westbury, N.Y., U.S.A.) with a 25-mesh screen.

Preparation of protein isolatesTo extract protein isolates, 100 g of twice-pressed meal flour

(approximately 38% protein) was dispersed in 400 mL of distilledwater. The pH of the suspension was adjusted to required solu-bilization pH using 6 N sodium hydroxide (NaOH) and stirredusing a magnetic stirrer for 1 h. The fiber and other suspendedsolids were removed by centrifuging at 5000 × g for 30 min. Theprotein-rich supernatant was further filtered through cheeseclothand through Whatman 41 filter paper. Protein in the supernatantwas then precipitated by drop-wise addition of 6 N hydrochloricacid (HCl) to lower the pH to the required precipitation pH. Theprecipitated proteins were recovered by centrifugation at 5000 ×g for 30 min and freeze-dried at a freezing temperature of −25 ◦Cand a drying shelf temperature of 25 ◦C. Lyophilized protein iso-lates were used for functional property testing. The percentageyield was based on the mass (g) of lyophilized protein isolate pro-duced from 100 g of twice-pressed canola meal flour.

Compositional analysis of protein isolatesTotal nitrogen content of the isolates was found using macro-

Kjeldahl procedure (988.05) and a conversion factor of 6.25 wasused to determine the crude protein content. Dry matter wasdetermined by oven-drying the protein isolates at 105 ◦C (950.46),and ash (920.153) and crude fat (976.21) was determined usingFoss-let procedure (AOAC 1990, 1995).

Functional properties of protein isolatesThe functional properties (water absorption, fat absorption, and

emulsifying activity) of protein isolates were measured accordingto the methods reported by Ghodsvali and others (Ghodsvali andothers 2005) with modifications. Three replicates were used foreach test.

To determine water absorption (WA), the lyophilized proteinisolate samples (2 g) were suspended in 16 mL of distilled water in50-mL centrifuge tubes. The tubes were vortexed for 30 s every10 min for 1 h, then centrifuged at 2000 × g for 15 min andinverted for 30 min to remove free water. WA was calculated asthe percentage increase of sample weight.

To determine fat absorption (FA), the protein isolate samples(2 g) were suspended in 12 mL of canola oil in 50-mL centrifugetubes. The tubes were vortexed for 30 s every 5 min for 30 min.Tubes were then centrifuged at 1600 × g for 25 min and the freeoil was decanted by inverting the tubes for 1 h. FA was expressedas the percentage increase of sample weight.

To measure the emulsifying activity (EA), 1.4-g protein isolatesamples were mixed with 20-mL distilled water in a 50-mL testtube and vortexed at speed 10 for 30 s using a VWR analogvortex mixer (VWR Intl., LLC, West Chester, Pa., U.S.A.). Themixture was combined with 10 mL of canola oil and vortexedfor 30 s. An additional 10 mL of canola oil was added to themixture, vortexed for 90 s, and centrifuged at 1100 × g for 5 minat 25 ◦C. The volume of emulsified layer was measured before andafter centrifugation. The EA was calculated as a percentage of thevolume of emulsified layer after centrifugation to the volume ofemulsion before centrifugation.

Gel electrophoresis of proteinsA modified Laemmli procedure was followed to prepare an 8%

to 16% acrylamide gradient gel to run samples (20 μL of 2 mg/mLsolutions) from each protein fraction and the molecular weightmarker (Laemmli 1970). The gel was run at constant voltage of100 V. After 7 to 8 h, the gel was stained with Coomassie BrilliantBlue solution for 8 h with constant shaking. Destaining was carriedout in a solution of 10% (v/v) acetic acid and 10% (v/v) methanolfor 24 h. The protein band intensities in the gels were scannedat 570 nm using a model GS-670 Bio-Rad imaging densitometer(Hercules, Calif., U.S.A.).

Surface hydrophobicitySurface hydrophobicity of protein isolates was measured using a

modified method of Kato and Nakai (1980) using the fluorescenceprobe, 1-anilino-8-naphthalene sulfonic acid (ANS, Tokyo KaseiKogyo Co., Ltd., Tokyo, Japan). Freeze-dried canola protein wassolubilized in 0.01 M phosphate buffer (pH 7.2) at room tem-perature while stirring continuously for 2 h to prepare 1 mg/mLsolutions. Protein solutions were serially diluted with buffer toprepare a series of protein concentrations in the range of 0.05 to100 μg/mL. In 1-mL aliquots of each diluted sample, 15 μL ofANS (8.0 mM in 0.01 M phosphate buffer at pH 7.2) was addedand incubated at room temperature for 15 min. The fluorescence

Vol. 76, Nr. 3, 2011 � Journal of Food Science E267



intensity (FI) of samples was measured using a Spectra Max-M5spectrophotometer (Molecular Devices, Sunnyvale, Calif., U.S.A.)at wavelengths of 360 nm (excitation) and 520 nm (emission). FIwas plotted against protein concentration and the slope of thestraight line obtained from linear regression analysis was reportedas the surface hydrophobicity (Kato and Nakai 1980). Duplicatesamples were used for the analysis.

Thermal properties of protein isolatesThe glass transition temperature (Tg) was measured using a

Q1000 differential scanning calorimeter (DSC) (TA Instruments,New Castle, Del., U.S.A.). A hermetic DSC pan was used toencapsulate a 5-mg freeze-dried protein isolate sample. A tem-perature ramp rate of 10 ◦C/min was used in a heat-cool-heatcycle in the range of 0 to 175 ◦C. The Universal Analysis 2000software (TA Instruments) was used to analyze the thermogramsto determine Tgs of the protein fractions. The Tg was determinedas the midpoint of the step transition using the plot of heat flowcompared with temperature (Brake and Fennema 1999).

Experimental design and statistical analysisCCD was used to analyze the impacts of solubilization pH

and precipitation pH on the various properties of protein isolates(Myers and Montgomery 1995). The design consisted of 4 factorialruns, 4 axial runs, and 5 replicates at the design center point. Thecenter point (solubilization pH 12.0 and precipitation pH 4.5) wasthe point of maximum protein yield found in previous work. Thehighest and lowest levels for solubilization and precipitation pHvalues were set as 1 pH unit above and below the center point,respectively. Table 1 shows design factors in their coded form andexperimental values (uncoded form), in an experimental designmatrix with a randomized treatment order.

A 2nd-order response surface model (Eq. 1) was developed foreach isolate property. The response variable R represents the mod-eled properties: yield, water absorption, FA, or EA. The equa-tion representing the expected responses (R) can be written asfollows:

R = β0 +∑2

i=1βιxi +

∑∑2

j<i=2βi j xi x j +

∑2

i=1βi i x2

i+ ε

(1)

The regression coefficients β0, β i, β ij, and β ii are calculatedby least squares method. The coded levels of the 2 design factors

Table 1–Coded levels and corresponding actual values of the 2factor CCD.

Coded level Actual value

Treatment Solubili- Precipitation Solubili- Precipitationorder zation pH pH zation pH pH

1 0 −1.414 12.0 3.52 0 0 12.0 4.53 1 1 12.7 5.24 0 0 12.0 4.55 1 −1 12.7 3.86 −1.414 0 11.0 4.57 1.414 0 13.0 4.58 0 0 12.0 4.59 −1 1 11.3 5.210 0 0 12.0 4.511 1 −1 11.3 3.812 0 1.414 12.0 5.513 0 0 12.0 4.5

are represented by xi and xj, and the statistical error is representedby ε.

Analysis of variance was performed for each response to deter-mine the model adequacy and coefficients. The adequacy of themodels was checked using parameters R2, R2(adj), and the lack-of-fit test (rejected for P > 0.05). Statistically insignificant terms(P > 0.1) were identified, and different forms of the 2nd-orderregression equations were tested by sequentially removing theseterms and determining the impact on the model adequacy. Cor-relations among properties and other statistical analyses were doneusing Minitab software (Minitab Inc., State College, Pa., U.S.A.).

Results and Discussion

Compositional analysis of protein isolatesThe compositional data of protein isolates are shown in Ta-

ble 2. The protein isolates showed relatively low concentrationof proteins within the range of 66% to 76%. The protein extrac-tion was carried out using fine flour with continuous stirring for1 h. These conditions provided sufficient contact between flourand the solvent (water) to facilitate protein extraction in to thesolution; however, the protein concentration was largely affectedby the high oil and ash contents in the isolates. High level of oilin the canola meal flour (approximately 8%) resulted in high oilcontents in the isolates due to formation of lipid-protein com-plex. The usage of NaOH in the solubilization and HCl in theprecipitation steps of protein isolation procedure resulted in highash (salt) content. The salt content increased with the increasingsolubilization pH used in protein isolate preparation. Increasingsalt concentrations tend to increase EA while decreasing waterabsorption (Ragab and others 2004; Ogungbenle 2008).

Protein yieldGenerally, proteins have higher solubility in strong alkaline so-

lutions and high solubility is a basic requirement for high overallyields after subsequent precipitation (Tzeng and others 1990). Athigh pH values (>12), however, the large negative charge cancause some proteins to unfold (Cao 2002). At very high pH values(>13) (especially with elevated temperatures), alkaline hydrolysisof peptide bonds can take place (Warner 1942) causing uncontrol-lable cleavage of the protein molecules (Whitaker 1978).

Figure 1 shows the contour plot of the model for yield com-pared with precipitation pH and solubilization pH. The resultsshow that the yield is almost equally sensitive to solubilization pHand precipitation pH. The highest yields (>28%) are predictedin a fairly large region near the center point of the experimentalspace. These predicted values are more than 56% higher than theyields reported in other studies for canola proteins (Tzeng and

Table 2–Compositional analysis (dry basis) of protein isolates.

Sol pH/Prec pH Ash (%) Protein (%) Fat (%)

11.0/4.5 2.3 (0.1) 71.0 (0.0) 15.4 (0.0)11.3/3.8 2.5 (0.0) 72.9 (0.2) 12.2 (0.1)11.3/5.2 2.2 (0.1) 73.0 (0.1) 13.6 (0.1)12.0/3.5 2.8 (0.0) 70.2 (0.0) 13.3 (0.3)12.0/4.5 2.8 (0.1) 70.3 (0.2) 14.3 (0.0)12.0/5.5 3.2 (0.0) 75.4 (0.1) 7.7 (0.1)12.7/3.8 5.4 (0.0) 68.5 (0.1) 15.1 (0.0)12.7/5.2 6.5 (0.0) 68.6 (0.1) 14.1 (0.0)13.0/4.5 7.7 (0.1) 66.0 (0.1) 18.7 (0.0)

E268 Journal of Food Science � Vol. 76, Nr. 3, 2011

E:Fo

odEn

ginee

ring&

Phys

icalP

ropert

ies


others 1990; Berot and others 2005). Low precipitation pH andlow solubilization pH values (within the experimental space) resultin the lowest protein yields but yields are also predicted to declineat combinations of high solubilization and precipitation pH; how-ever, the low solubilization efficiencies at lower pH values canbe compensated to some degree by carrying out the precipitationat a higher pH. There is also more flexibility in terms of yieldwith precipitation pH when solubilization is done at pH valuesnear 12.5. Such information could be important depending onthe impact of precipitation pH on other protein properties.

The protein yield, Y (weight of protein isolates prepared as apercentage of starting meal flour), was expressed as a function ofsolubilization pH (S) and precipitation pH (P) (uncoded) in Eq. 2.The significances of coefficients in the model are shown in Table 3,and the lack-of-fit test for the model and correlation coefficientstatistics are shown in Table 4. The model shows the nonlinear

Figure 1–Contour plot showing the relationship of isolate yield with pre-cipitation pH and solubilization pH.

Table 3–Significance of coefficients in the models for yield andfunctional properties of protein isolates.

Terms (P-value)

Models S P S2 P2 SP

Yield (Y) 0.024 0.025 0.085 0.045 0.077Water absorption

(WA)0.027 0.082 0.048 0.196 0.278

Fat absorption(FA)

0.001 0.200 0.125 0.008 0.218

Emulsifyingactivity (EA)

0.002 0.563 0.140 0.209 0.213

S = solubilization pH; P = precipitation pH.

relationship of protein yield to solubilization and precipitation pHthat agrees with a physical understanding of the system as describedabove.

Y = −393.8+51.0(S)+46.3(P)−1.6(S)2 −2.0(P)2 − 2.2(S)(P)(2)

Water absorption of isolatesHigh WA was observed with high solubilization pH and high

precipitation pH values (Figure 2). The conditions that give highyields generally result in higher water absorption of the protein iso-lates. The solubilization and precipitation pH values that resultedin highest yield of protein isolates also showed highest WA. Thisphenomenon could be an added advantage if the protein isolatesare intended to be used in food-related applications where highWA is beneficial; however, it could be detrimental with respect toindustrial applications of the isolates where low WA (high waterresistance) is desired.

WA indicates the ability of protein isolates to physically holdwater against gravity (Kinsella 1979). The ability of proteins toretain water is very important in food systems since it affects theflavor and texture of foods (Yu and others 2007). The aminoacid composition, protein conformation, and hydrophobicityare among the important factors that affect the WA of proteins.Upstream processing steps involved in protein isolation cancontribute to alter these deterministic factors.

The effects of solubilization pH (S) and precipitation pH (P) onwater absorption of protein isolates (WA) can be expressed usingthe Eq. 3.

WA = −3549 + 591.8(S) + 14.3(P) − 23.8(S)2 (3)

Although the linear term of precipitation pH was significantin determining the water absorption of isolates, it has a muchsmaller impact than solubilization pH (Table 3). The lack-of-fittest indicates the model is significant (Table 4). The relationshipbetween solubilization pH and WA of isolates is nonlinear, whereasthe relationship between precipitation pH and WA of isolates islinear within the design space.

Fat absorption of isolatesUnlike WA, the FA showed a linear relationship with solubiliza-

tion pH while showing a nonlinear relationship with precipitationpH (Table 3). FA increased with solubilization pH (Figure 3);however, the precipitation pH had a relatively smaller impact onFA of protein isolates with the highest FA occurring at high andlow precipitation pH values. WA and FA in proteins take placewhen water or oil is physically entrapped by an unfolded proteinmatrix (Kinsella 1979; Ahmedna and others 1999). A possible rea-son for increased FA is the increased unfolding of proteins dueto denaturation at higher solubilization pH values. Other reports

Table 4–Lack-of-fit and correlation coefficient statistics for the models developed for yield and functional properties of proteinisolates.

P-value for lack-of-fit R2 (adj)

All factors Insignificant factors All factors Insignificant factorsincluded excluded included excluded Final R2

Yield (Y) 0.654 0.654 0.67 0.67 0.81Water absorption (WA) 0.911 0.741 0.56 0.49 0.62Fat absorption (FA) 0.107 0.082 0.79 0.73 0.79Emulsifying activity (EA) 0.630 0.266 0.65 0.58 0.61




have shown similar correlation with increasing FA with the in-creasing WA of protein isolates (Dua and others 1996; Yu andothers 2007).

The relationship of FA of protein isolates with solubilizationpH (S) and precipitation pH (P) (in uncoded form) is expressed byEq. 4. The lack-of-fit test for the model of FA has a relatively lowP-value (P = 0.082) compared to the models of other properties;however, the model is significant at 95% confidence level (Table 4).

FA = 215.0 + 12.5(S) − 124.3(P) + 14.2(P)2 (4)

FA is an important parameter in food applications of proteinisolates, where higher FA is generally desired (Kinsella 1979; Duaand others 1996). Extraction of proteins at higher pH values andprecipitation at either high or low pH values could therefore bebeneficial for these applications. Models show that shifting isola-tion parameters from maximum yield conditions to maximum FAconditions results in an increase of FA by 13% while sacrificing aloss of 10% in yield. It is also possible to produce protein isolateswith a range of FA properties within the area of near-maximumyield. This would allow a processer to modify isolation procedures

Figure 2–Contour plot showing the relationship of isolate water absorptionwith precipitation pH and solubilization pH.

Figure 3–Contour plot showing the relationship of isolate fat absorptionwith precipitation pH and solubilization pH.

according to the application while decreasing the need for moreexpensive downstream protein modification.

Emulsifying activity of isolatesProteins can act as good emulsifiers in food systems, since they

possess both hydrophilic and hydrophobic characteristics due tothe availability of polar and nonpolar amino acids, respectively (Yuand others 2007). The ability of proteins to interact with both oiland water determines the EA of protein isolates. The relationshipbetween EA and the solubilization pH can be illustrated in Figure4 and expressed in the form of Eq. 5.

EA = −37.3 + 10.1(S) (5)

EA showed a positive relationship with the solubilization pHof the protein isolates, but the impact of precipitation pH wasnot found to be significant (Table 3). Although the final R2 valuewas only 0.61, the result of the lack-of-fit test (P = 0.266) showsthat the model is significant (Table 4). Emulsification of proteinsdepends largely on the ionic charge and hydrophobicity. Excessiveionic charge can play a negative role in the EA of proteins, sincethe proteins can repel the negatively charged oil droplets (Beuchat1977). At high pH values, unfolding and hydrolysis can exposemore hydrophobic groups in the protein chains (Yu and others2007). The high surface activity of the proteins extracted at highpH values may also have contributed to the increased interaction ofproteins with oil and water, thus increasing the EA (Ahmedna andothers 1999); however, precipitation of proteins (at any acidic pH)results in minimal excess charge and insensitivity of emulsificationto the precipitation pH value.

Correlation of yield and protein isolate propertiesThe functional properties of isolates generally did not show

high correlation between each other (Table 5). The functional

Figure 4–Model of emulsifying activity of isolates compared with solubi-lization pH.

Table 5– Correlation matrix for protein yield and isolate func-tional properties.

Y WA FA EA

Y 1.000WA 0.875 (0.000) 1.000FA 0.323 (0.281) 0.479 (0.098) 1.000EA 0.271 (0.370) 0.357 (0.231) 0.487 (0.091) 1.000

P-values are denoted in parentheses. Y = yield; WA = water absorption; FA = fatabsorption; EA = emulsifying activity.


E:Fo

odEn

ginee

ring&

Phys

icalP

ropert

ies


properties (WA, FA, and EA) are governed by widely varyingphenomena such as protein conformation, hydrophobicity, andionic charge that led to low correlation between each other.WA of isolates, however, showed a high positive correlation withyield.

Solubilization pH was an influential parameter in all properties,while precipitation pH was less significant in most. This could beattributed to the fact that significant changes to the compositionand quality of proteins (especially canola proteins) take place in thehigh pH range. Closer to the upper end of the solubilization pH(pH 13), significant denaturation of proteins can take place thatsignificantly impacts the isolate properties (Yu and others 2007).However, when considering the precipitation pH range (pH 3.5to 5.5), yield was affected significantly while there were less sig-nificant changes to the properties of the resulting protein isolates.With an expanded precipitation pH range, however, propertiesof protein isolates may exhibit more significant differences (at theexpense of yield) with the changing precipitation pH.

The fact that all functional properties (WA, FA, and EA) ofprotein isolates were positively correlated with the solubilizationpH (P < 0.05) could be beneficially used in the protein-basedfood additives industry. By operating toward the higher end ofthe precipitation pH range, the functional properties could beimproved with smaller impacts on yield.

Gel electrophoresis of proteinsSodium dodecyl sulfate-polyacrylamide gel electrophoresis was

carried out in order to investigate the possible compositionalchanges among the different fractions of protein isolates that ledto the differences in isolate functional properties. The gel elec-trophoresis patterns showed slight differences between variousprotein isolates (Figure 5). The band intensities were comparedusing a densitometer and the globulin bands at lower solubiliza-tion pH values (11 to 11.3) were more than 30% darker than thebands corresponding to higher solubilization pH values (12.7 to13). The relative intensities of the globulin bands (approximately30 kDa) confirmed the decrease in globulins with the rising sol-ubilization pH. This indicates that the protein isolates extractedat lower pH values possess a higher fraction of more hydrophobicglobulins compared to those extracted at higher pH values. Atthe lower end of solubilization pH, the albumins were extractedindicating that they are still soluble near their isoelectric point (pH11) (Schmidt and others 2004). Therefore, the protein isolates ex-tracted at lower solubilization pH values were comprised of bothalbumins and globulins.

At the higher end of solubilization pH, the higher molecularweight bands (around 50 kDa) appear to be darker than those ex-tracted at lower solubilization pH values. This could be due to theformation of unusual amino acids such as lysinoalanine in highly

Figure 5–SDS-PAGE of canola proteins isolatedusing various solubilization and precipitationpH values.

Figure 6–Surface hydrophobicity of canola protein isolates prepared using different conditions (common letters denote no significant difference at P =0.05).




Figure 7–DSC thermograms for protein isolates prepared under different conditions. Curve labels represent solubilization pH/precipitation pH used inisolate preparation.

alkaline solutions (Satterlee and Chang 1982). Lysinoalanines areformed by the combination of smaller subunits of albumins andglobulins via cross-linking of amino acid residues that appear ashigher molecular weight subunits in the gel. The formation oflysinoalanines results in a decrease in available lysine and also af-fects functional properties due to cross-linking. It is evident thatthe formation of lysinoalanine can be reduced by using lower pHvalues for the extraction of proteins. Also Figure 5 shows thatthe proteins extracted at solubilization pH 13 showed less distinctbands in the gel indicating significant uncontrollable cleavage anddegradation of protein chains due to highly alkaline conditions.

Surface hydrophobicity of proteinsSurface hydrophobicity of different protein isolates are shown in

Figure 6. Surface hydrophobicity generally did not show high re-sponse to varying extraction and precipitation pH values; however,water absorption of isolates showed a decreasing trend with de-creasing solubilization pH values. This agrees well with the highersurface hydrophobicity values seen at lower solubilization pH. TheEA and FA, however, did not show positive correlations with in-

Table 6– Glass transition temperatures of protein isolates pre-pared under different conditions.

Solubilization Precipitation Glass transitionpH pH temperature (◦C)

11.0 4.5 6111.3 3.8 6411.3 5.2 6112.0 3.5 5412.0 4.5 5612.0 5.5 5312.7 3.8 5612.7 5.2 5413.0 4.5 50

creasing surface hydrophobicity of the protein isolates at the lowersolubilization pH values.

Thermal properties of protein isolatesThe thermal properties of protein isolates were investigated to

examine the impact of protein isolation parameters on the pro-tein conformation. The representative thermograms of differentprotein isolates are shown in Figure 7. Tg was determined as themidpoint of the step transition that appeared in the range of 50 to65 ◦C for different protein isolates; however, major phase changesin proteins are not likely in this temperature range. Since theproteins are denatured in higly alkaline solutions, there were noprominent endotherms representing denaturation temperature.

The Tg values of the various protein isolates are shown in Ta-ble 6. It is evident that Tg gradually decreased when the solubi-lization pH was increased from 11.0 to 13.0. This effect can beattributed to the protein denaturation in highly alkaline solutionsthat contributes to the reduction of Tg of protein isolates (Bell andHageman 1996). These conformational changes of protein isolateshad a significant impact on protein isolate properties as describedabove.

ConclusionsThe functional properties of protein isolates showed high sen-

sitivity to solubilization pH and lesser, but still significant, impactfrom precipitation pH. The protein yield showed high positivecorrelation with WA but did not have significant correlations withother properties. The yield was robust with respect to both solubi-lization and precipitation pH values showing a fairly large regionfor near-maximum yield conditions. This allows processers to ad-just pH values according to the protein isolate property require-ments without affecting the yields greatly. Upstream processingconditions, such as solubilization and precipitation pH values, can


E:Fo

odEn

ginee

ring&

Phys

icalP

ropert

ies


be used as controllable parameters to modify the properties ofprotein isolates according to specific needs and applications.

AcknowledgmentsThis research was supported by the USDA/CSREES NRI pro-

gram through grant 2008–35504-18667. We gratefully acknowl-edge the technical support given by Dr. Zhisheng Liu, Dr. DarrinHaagenson, Dr. Stefan Vetter, Jaidev Sehrawat, Gloria Nygard,and Heidi Docktor. We also thank Mr. Tom Borgen (Langdon,N. Dak.) for providing canola seed.

ReferencesAhmedna M, Prinyawiwatkul W, Rao RM. 1999. Solubilized wheat protein isolate: functional

properties and potential food applications. J Agric Food Chem 47:1340–5.Aluko RE, Yada RY. 1995. Structure-function relationships of cowpea (Vigna unguiculata) glob-

ulin isolate: influence of pH and NaCl on physicochemical and functional properties. FoodChem 53(3):259–65.

AOAC. 1990. Official methods of analysis, 15th ed. Washington, D.C.: Assn. of Official AnalyticalChemists.

AOAC. 1995. Official methods of analysis, 16th ed. Washington, D.C.: Assn. of Official AnalyticalChemists.

Bell LN, Hageman MJ. 1996. Glass transition explanation for the effect of polyhydroxy com-pounds on protein denaturation in dehydrated solids. J Food Sci 61(2):372–5.

Berot S, Compoint JP, Larre C, Malabat C, Gueguen J. 2005. Large scale purification of rapeseedproteins (Brassica napus L.). J Chromatogr B. 818(1):35–42.

Beuchat LR. 1977. Functional and electrophoretic characteristic of succinylated peanut flourproteins. J Agric Food Chem 25:258–60.

Brake NC, Fennema OR. 1999. Glass transition values of muscle tissue. J Food Sci 64(1):10–5.Cao Y. 2002. The functional properties of soybean film. Fargo, N. Dak.: North Dakota State

Univ.Cumby N, Zhong Y, Naczk M, Shahidi F. 2008. Antioxidant activity and water-holding capacity

of canola protein hydrolysates. Food Chem 109:144–148.Diosady LL, Rubin LJ, Chen BK. 1990. Membrane based processes for the production of

rapeseed protein isolates. Int Cong Eng Food 3:280–9.Dua S, Mahajan A, Mahajan A. 1996. Improvement of functional properties of rapeseed (Brassica

campestris Var. Toria) preparations by chemical modification. J Agric Food Chem 44:706–10.Ghodsvali A, Khodaparast MHH, Vosoughi M, Diosady LL. 2005. Preparation of canola protein

materials using membrane technology and evaluation of meals functional properties. FoodRes Int 38(2):223–31.

Kato A, Nakai S. 1980. Hydrophobicity determined by a fluorescence probe methodand its correlation with surface properties of proteins. Biochim Biophys Acta 624:13–20.

Kinsella JE. 1979. Functional properties of soy proteins. J Am Oil Chem Soc 56:242–58.

Kumar R, Choudhary V, Mishra S, Varma IK, Mattiason B. 2002. Adhesives and plastics basedon soy protein products. Ind Crop Prod 16(3):155–72.

Laemmli UK. 1970. Cleavage of structural proteins during the assembly of the head of bacterio-phage T4. Nature 227(259):680–5.

Li CP, Enomoto H, Ohki S, Ohtomo H, Aoki T. 2005. Improvement of functional propertiesof whey protein isolate through glycation and phosphorylation by dry heating. J Dairy Sci88:4137–45.

McCurdy SM. 1990. Effects of processing on the functional properties of canola/rapeseedprotein. J Am Oil Chem Soc 67(5):281–4.

Myers RH, Montgomery DC. 1995. Response surface methodology: process and product op-timization using designed experiments. 2nd ed. New York, N.Y.: A Wiley-Interscience pub-lication.

Ogungbenle HN. 2008. Effects of salt concentrations on the functional properties of somelegume flours. Pak J Nutr 7(3):453–8.

Ohren J. 1981. Process and product characteristics for soya concentrates and isolates. J Am OilChem Soc 58(3):333–5.

Ragab DM, Babiker EE, Eltinay AH. 2004. Fractionation, solubility and functional properties ofcowpea (Vigna unguiculata) proteins as affected by pH and/or salt concentration. Food Chem84:207–12.

Satterlee LD, Chang KC. 1982. Nutritional quality of deteriorated proteins. Food proteindeterioration. Washington, D.C.: American Chemical Society. p 409–31.

Schmidt I, Renard D, Rondeau D, Richomme P, Popineau Y, Axelos MAV. 2004. Detailedphysicochemical characterization of the 2S storage protein from rape (Brassica napus L.). JAgric Food Chem 52(19):5995–6001.

Selmane D, Christophe V, Gholamreza D. 2008. Extraction of proteins from slaughterhouseby-products: influence of operating conditions on functional properties. Meat Sci 79(4):640–7.

Tzeng YM, Diosady LL, Rubin LJ. 1990. Production of canola protein materials by alkalineextraction, precipitation, and membrane processing. J Food Sci 55(4):1147–56.

Uruakpa FO, Arntfield SD. 2004. Rheological characteristics of commercial canola proteinisolate–k-carrageenan systems. Food Hydrocolloid 18:419–27.

Warner RC. 1942. The alkaline hydrolysis of egg albumin. J Biol Chem 142:741–56.Whitaker JR. 1978. Biochemical changes occurring during the fermentation of high protein

foods. Food Tech 32:175–80.Wu J, Muir AD. 2008. Comparative structural, emulsifying, and biological properties of 2 major

canola proteins, cruciferin and napin. J Food Sci 73(1):C210–16.Yu J, Ahmedna M, Goktepe I. 2007. Peanut protein concentrate: production and functional

properties as affected by processing. Food Chem 103:121–9.


effects of protein separation conditions on the functional and thermal properties of canola protein...

Documents