Physicochemical and Functional Properties of Vegetable and Cereal Proteins as Potential Sources of Novel Food Ingredients

Currently there is a rising interest in protein isolation for their subsequent use as food ingredient. Sixty percent of Americans take into consideration protein content in food or beverages when making a buying decision (1). Of the three macronutrients (carbohydrates, proteins and fats), proteins are the most appealing for consumers concerned about their health. Nearly half of adults perceive proteins as ingredients that increase energy levels, support overall good health and improve muscle tone. These macronutrients are also considered important in diets aimed to complete a weight management program. Despite the awareness of protein importance in a balanced diet, nearly 25 % of adults believe that they cannot consume as much proteins as they would like because of the cost (2). The protein industry is segmented into animal (gelatin, egg white, casein and whey) or vegetable, of which soya bean is the only source of worldwide relevance. The former has the advantage of being of high nutritional quality, but with higher cost than the vegetable counterparts and frequently the supply is irregular and unreliable. The latt er ISSN 1330-9862 original scientifi c paper


Introduction
Currently there is a rising interest in protein isolation for their subsequent use as food ingredient.Sixty percent of Americans take into consideration protein content in food or beverages when making a buying decision (1).Of the three macronutrients (carbohydrates, proteins and fats), proteins are the most appealing for consumers concerned about their health.Nearly half of adults perceive proteins as ingredients that increase energy levels, support overall good health and improve muscle tone.These macronutri-ents are also considered important in diets aimed to complete a weight management program.Despite the awareness of protein importance in a balanced diet, nearly 25 % of adults believe that they cannot consume as much proteins as they would like because of the cost (2).The protein industry is segmented into animal (gelatin, egg white, casein and whey) or vegetable, of which soya bean is the only source of worldwide relevance.The former has the advantage of being of high nutritional quality, but with higher cost than the vegetable counterparts and frequently the supply is irregular and unreliable.The latt er in turn are cheaper, abundant and with a good nutritional value, mainly when combined, therefore making them a good option as food ingredients.
Vegetable proteins, as food ingredients, should perform specifi c functions within formulations such as to provide or enhance texture, gelling, emulsifying or foaming characteristics, among others.The best way to test the role of high-protein ingredients in food is in a practical scenario, unfortunately this is not always possible and therefore laboratory procedures for protein characterization are of utmost importance (3).Functional tests are required to evaluate and predict how proteins may behave in specifi c systems, off ering a pre-evaluation of the best application (4).The physicochemical and functional characterization thus should be clear before the use of proteins as food ingredients (5).
Currently there is a lot of information regarding functional properties of proteins starting with the overwhelming general data about soya and oilseed-derived materials (4,6,7).Several authors have compared the physicochemical and functional properties of buckwheat protein, soya protein isolate and casein (5,6,8).The functional characteristics of pseudocereals as quinoa and amaranth have also been reported in literature (5,9,10).Regarding cereals and other oilseeds, some authors have made comparisons among the protein functional properties of rice cultivars, peanut fl our and peanut protein concentrate as indicators of their potential use in food industry (11,12).Other high-protein crops, such as pulses, have been explored and characterized: marama bean (13), cowpea (14), pea, lentil, navy bean and chickpea (15).Despite the high quantity of information about specifi c crops and high protein materials, the characterization of specifi c and novel proteins is required to determine their physicochemical and functional characteristics.These data are valuable for the development of future protein sources through innovation and research, especially of new, less expensive materials capable of giving a well-balanced food in terms of health and sensorial characteristics.The aim of this work is to characterize the physicochemical and functional properties of a set of vegetable and cereal proteins (proposal of commercial and novel mixed materials from soya bean, pea and maize germ concentrates and hydrolysates) and also to explore their correlations in order to understand bett er the characteristics of proteins aimed to be used as food ingredients.

Materials
The analyzed samples were identifi ed as: pea fl our (TECSA, Monterrey, Mexico), soya bean fl our national (SBFN; Food Proteins Corporation, Mexico City, Mexico), soya bean fl our 120 (SBF120; Productos Industriales Gaf, Mexico City, Mexico), soya bean fl our 200/20 (SBF200/20; Food Proteins Corporation), soya bean fl our Nutrisoy (SBFNutri; ADM, Chicago, IL, USA), soya bean fl our Ragasa (SBFRagasa; Ragasa, Monterrey, Mexico), concentrates of soya and maize (01 to 05) and hydrolysates of soya and maize (01 to 09).The number at the end of each code represents the sequence in which each protein was generated.All used materials were defatt ed.The mixtures of soya and maize proteins were obtained using a standard procedure of alkali extraction followed by acid precipitation (16).Briefl y, the pH of a fi nely ground mixture of defatt ed soya fl our and defatt ed maize germ (proportion 5:1 using 10 parts of water) was adjusted to pH=10 with 50 % NaOH.Contents were mixed for 30 min at 50 °C before separation of bagasse using an industrial centrifuge (Model SA14, GEA Westfalia, Oelde, Germany) operated at 15 L/min and 5500×g.The supernatant was then collected and the pH adjusted to 4.5 with 3 M HCl.The curd was separated using the centrifuge operated at the previously described conditions.The resulting product was washed with an equal volume of water, separated by centrifugation and then the pH was adjusted to 7.0 (with 50 % NaOH).The resulting material was dried using an industrial spray dryer designed by Nutrigrains (Monterrey, Mexico) with air inlet and outlet temperatures of 195 and 80 °C, respectively, and atomization pressure of 1726 N/cm 2 .For hydrolyzed proteins, enzymatic hydrolysis was performed before spray drying (Neutrase ® , 0.25 % of total solids in the curd, 30 min at 40 °C).The spray-dried samples were stored at room temperature in a dry and ventilated place.

Functional properties
The water absorption (WAI) and water solubility (WSI) indices were determined using 1 g of sample placed in 15 mL of distilled water according to Cheft el et al. (21).The nitrogen solubility index (NSI) was assayed using 0.5 g of sample dispersed in 50 mL of 0.1 M sodium chloride (pH=7.0)(21).Nitrogen was determined with micro-Kjeldahl method in total and soluble fractions (AOAC Offi cial Method 984.13-1994; 18).Fat absorption index (FAI) was determined based on a previously reported method (22).The turbidimetric procedure (23) was used for determining emulsifying activity index (EAI) in all samples, whereas emulsion stability (ES) was calculated according to Haque and Kito (24).Regarding functional properties related to protein and air interaction, foaming characteristics were evaluated: foaming activity (FA), foam stability (FS) and foam density (FD) in 3 % (by mass) protein dispersions in water (24).Urease activity (UA) was determined as a change in pH according to AOCS Method Ba 9-58 (25) and heat coagulation capacity (HCC) with the technique proposed by Regenstein and Regenstein (26).

Statistical analysis
All determinations were performed in triplicate and data were analyzed with ANOVA (Minitab Statistical Soft ware v. 16, Minitab Inc., State College, PA, USA).Mean values were compared with Tukey's test (α=0.05).
Pearson's correlations, linear regression and principal component analysis (PCA) were determined with the use of the same statistical soft ware.

Physicochemical characterization of vegetable and cereal proteins
Table 1 shows the physicochemical properties of the array of analyzed vegetable and cereal proteins.The pH, an important parameter associated with protein solubility, ranged from 6.42 to 7.57, except for the soya and maize concentrates 01 and 02 with values of 5.63 and 5.89, respectively.At lower or higher values than the isoelectric point, the electrostatic repulsion increases and consequently the solubility of proteins improves.This parameter is also related to the water absorption capacity of the material: ionized amino acid groups bind more water than non-ionized ones.Lowering the pH below 4 changes the carboxyl groups into non-ionized forms, thus reducing water-binding properties of the protein.
Electrical conductivity (EC) of all samples was between 4.32 and 1.32 mS/cm (soya and maize hydrolysates 02 and 04).This parameter is a measure of the ability of material to conduct electrical current and is aff ected by the content of protein, fat and minerals, among others.Foods with electrolytes such as salts, acids, certain gums and thickeners contain charged groups that have a notable eff ect on the EC.Protein charge and amino acid composition also aff ect this property.Foods such as apples, strawberries and potatoes have EC of 0.7, 1.9 and 0.4 mS/ cm at 25 °C, respectively (27).
The EC of soya and maize concentrates ranged from 2.47 to 3.64 mS/cm, comparable to the results reported by Režek Jambrak et al. (28) of 3.28 mS/cm of a soya protein concentrate.The soya and maize protein hydrolysates 02 and 04 showed the lowest and the highest conductivity of 1.32 and 4.32 mS/cm.Therefore, neither the protein content nor the higher degree of hydrolysis infl uenced the high EC of the protein, which depended on the charge density of the protein and confi guration acquired aft er a particular process (thermal or enzymatic hydrolysis).EC is important for the development of foods or beverages because of its infl uence on solubility, emulsifying and foaming activities, and consequently on the interaction with other ingredients and the protein stability in a given food system.EC is also important because it determines the heating rate and eff ectiveness of novel food processes, such as ohmic heating and pulsed electric-based operations (dehydration, extraction, pasteurization, etc.) (29).
Table 1 shows that the protein content and the degree of hydrolysis aff ect the EC signifi cantly (p<0.05).The pea fl our and soya and maize protein hydrolysate 04 contained the lowest and the highest protein fractions of 20.78 and 94.24 % (on dry mass basis), respectively.The concentrates and some soya and maize hydrolysates contained approx.70 % protein, similar to soya bean concentrates available on the market.The free amino nitrogen (FAN) content, which determines free amino acids or small peptides and therefore the degree of protein hydrolysis, solubility and water absorption capacity, ranged between 0.54 and 2.87 mg/g.As expected, the hydrolyzed proteins (treated with protease) had a higher FAN content (>2.0 mg/g), except for soya and maize hydrolysate 03, which contained 1.52 mg/g, similar to SBFRagasa and soya and maize protein concentrate 03.FAN values between 12 and 27 % in bean pods, 5 and 12 % in spinach and 34 and 56 % in potato tubers have been reported (30).The percentage of FAN in total nitrogen ranged between 0.6 and 2.3 % (Table 1), and these values were below those reported by Eppendorfer and Bille (30) in vegetable protein products.The diff erences, besides the availability of proteins and amino acids, could be associated with the method used for FAN determination.
The reducing sugar (RS) assay is highly relevant because the amounts of sugars relate with the stability or retention of protein functionality during storage (3).Foods can deteriorate during storage due to both enzymatic and Maillard-type reactions of primary amino groups with RS (31).Determination of reducing sugars by dinitrosalicylic acid method is thus a good index for characterization of high-protein materials (Table 1).Pea fl our had the highest content of RS of 136.65 mg/g, followed by SBFRagasa with 85.09 mg/g.The lowest content of RS was measured in SBF120, SBF200/20 and soya and maize hydrolysate 02 with only 5 mg of glucose reducing equivalents per gram.
The urease activity (UA) indicates the intensity of heat treatment during processing of protein meals.A value of 0.3 or less suggests that the protein source retains slight urease activity but has received suffi cient heat treatment for the inactivation of antibiological factors.A product with a pH increase of 0.02 or less during urease activity test (25) was surely overheated, yielding thus a material with diminished functional properties.All UA results shown in Table 1 were between 0.08 and 2.20 (soya and maize protein concentrate 01 and SBFRagasa, respectively), indicating that these protein sources received high and low thermal treatments and thus contained low and high residual enzymatic activity, respectively.In the specifi c case of pea fl our, its UA was similar to that of SBF120, SBF200/20 and soya and maize protein concentrate 04.Values are also similar to the ones reported by Valencia et al. (32), who compared UA activity of pea protein vs. soya bean protein concentrates.

Functional analysis of vegetable and cereal proteins
Functional characterization of the array of analyzed proteins is summarized in Table 2.The water absorption index (WAI) is one of the most important parameters to take into consideration for product development, particularly for dairy products and foods exposed to thermal treatments such as baking and thermoplastic extrusion (33).WAI is defi ned as the water absorbed per gram of tested material and it is regularly used as synonymous with water holding, water binding or water retention capacity (34).WAI values were between 0.41 and 18.52 (Table 2).The protein concentrates exhibited higher WAI values compared to the other vegetable or cereal protein sources (average of 8), followed by the soya bean and pea fl our (4.31 to 5.38 and 4.97, respectively) and the hydrolysates (around 1.0).Despite the direct relationship between water holding capacity and protein concentration (7), the higher protein concentration of hydrolysates did not improve the WAI compared with the other samples.This is infl uenced by protein structure and composition.According to Barbut (34) water can be divided in two general types according to its relationship with the protein molecule: absorbed and retained.The fi rst is the water bound to the protein molecule and therefore no longer available for its use as solvent, whereas the second is trapped within the protein matrix.The fi rst kind depends mainly on the amino acids and pH of the system, and the second is more dependent on the same relationship among protein molecules.Because of the type of water absorption procedure used herein, the second type of water is the one that varied the most among samples (Table 2), which is mainly due to protein structure organization.On the other hand, the smaller protein molecules that form hydrolysates reduce the interaction among molecules, yielding structures that do not hold water.Nitrogen solubility index (NSI) is another parameter related to the hygroscopic properties of proteins: it is a measurement of the protein dispersability in a NaCl solution.The NSI values of all the analyzed proteins were between 10.14 (of the SBFNutri) and 74.89 (soya and maize protein hydrolysate 05).These values were similar to the amounts reported for commercial high--protein soya bean products which ranged from 10 to 90 % (33).NSI is generally related to the extent of heating or protein denaturation, and is also important because it affects the solubility of proteins at diff erent ionic strengths.It off ers a more realistic approach to the performance of the protein in foods (since these are complex ionic systems).As expected, the hydrolysates showed the highest NSI because these proteins were hydrolyzed with protease beforehand, which according to Kinsella and Melachouris (3) markedly improves nitrogen solubility.
The water solubility index (WSI) of the proteins is the most important functional property because it aff ects other functional characteristics such as EAI, FA and HCC.WSI depends on the protein ability to interact with water.The WSI of soya bean fl our and soya and maize concentrates was around 35 %, while that of the hydrolysates (02 and 09) ranged from 32.71 to 95.65 %.Arrese et al. (6) reported the WSI values of soya proteins of 36.3 to 83.6 % and according to Tomotake et al. (5), WSI of a buckwheat isolate at similar pH value as used in our study was around 50 %, followed by the soya protein isolate, with values below 20 % and that of peanut fl our of 30 % (12).Therefore, the values of WSI obtained in some analyzed samples (Table 2) are higher than those reported for similar products.
Fat absorption index is the ability of the vegetable and cereal proteins to physically bind fat by capillary attraction.This is a parameter of paramount importance in food development, because fats act as fl avour retainer and also increase the mouth feel of the foods.The FAI of pea, soya bean fl our and soya and maize concentrates and hydrolysates ranged from 2.59 to 4.72.Meng and Ma (35) reported a value of FAI of a commercial soya protein of 1.52.FAI variation may be due to the diff erent surface hydrophobicities of vegetable proteins, because the absorption of fat has been att ributed to physical entrapment within the protein and non-covalent bonds such as hydrophobic, electrostatic and hydrogen bonding, the forces involved in lipid-protein interactions (36).Another property related to the hydrophobicity is the emulsifying activity index (EAI), i.e. the ability of a protein to form and stabilize the emulsion by creating electrostatic repulsion on oil droplet surface.The emulsion stability index (ESI) refl ects the ability of the proteins to form and maintain a stable emulsion over a period of time by preventing fl occulation and coalescence of the oil globules (9).The soya and maize protein hydrolysate 05 had the lowest EAI of 3936.62 m 2 /g, while the pea fl our had the highest, 52 399 m 2 /g.These results may be due to the high content of nonprotein solids in pea fl our, favourable for emulsion.In fact, the good performance of pea protein as egg replacer in mayonnaise-like products has received wide coverage in the media (2).In general, the diff erences in protein emulsifying activity may be related to their solubility and conformational stability.This property is widely utilized in totally or partially emulsifi ed foods, such as mayonnaise, cream, sauces, desserts, comminuted meat products and some beverages.Moreover, the emulsions of pea, soya bean fl our and soya and maize concentrates and hydrolysates were all stable for 24 to 48 h, i.e. they all have similar capacity to stabilize an emulsion.The high stability of the emulsions of vegetable and cereal proteins is due to their conformation.They are globular protein structures that reduce surface tension and form more rigid interfacial fi lms.
Similar to emulsion characteristics, another related property with two-phase interaction is foaming capacity or activity (FA).Foam can be defi ned as a two-phase system where air cells are separated by a continuous liquid layer, and foam stability (FS) is the capacity of a protein to reduce the surface tension by forming strong interfacial membranes via protein-protein interactions at the air-water interface.FA values of pea fl our and soya and maize hydrolysate 04 were 75 and 475 %, respectively (Table 2).The high content of nonprotein solids of the pea fl our may have increased the surface tension of the dispersion, reducing signifi cantly the FA.The FS of pea fl our was one of the lowest, with a value of zero, and the highest of soya and maize concentrate 02 (93.20 %).The foam density (FD) was similar in all samples.Protein foams can provide unique textures (as in meringue and nougat) that are associated with many foods such as angel and pound cakes, ice cream and confectionary products.
Heat coagulation capacity (HCC) was also determined and results are shown in Table 2. Results ranged from 66.35 to 95.50 % for pea fl our and soya and maize concentrate 05, respectively.Coagulation is the capacity of the protein to form a clot or a semisolid mass aft er an initial denaturation (driven by diff erent factors, such as heat).It involves the rupture of hydrogen bonds within peptide chains, and when an advanced state is reached, denaturation becomes irreversible.According to Kinsella (7), in soya proteins an initial heating above 60 °C is necessary to induce dissociation of quaternary globulins.This thermal treatment causes unfolding of polypeptides of the protein subunits with an increase in viscosity.Upon cooling, the unfolded polypeptides reassociate via hydrophobic associations, hydrogen bonding, ionic interactions and possibly some disulphide linkages, forming a gel.HCC is thus an important property in food applications such as processed meat, sausages and cheese.

Correlation analysis between physicochemical and functional properties of vegetable and cereal proteins
Results of correlation analyses between functional properties and physicochemical parameters are summarized in Table 3.Protein content was positively correlated with the FAN (R=0.75),NSI (R=0.69),WSI (R=0.78),FA (R=0.59) and FS (R=0.62), because higher protein mass fraction was observed in the hydrolyzed samples (last nine rows in Tables 1 and 2).On the other hand, protein mass fraction correlated negatively with RS (R=-0.56),EAI (R= -0.77) and FD (R=-0.72),which means that a higher protein content lowered EAI and FD, despite the fact that high-protein materials are good emulsifi ers (37).This result is reinforced by the fact that EAI showed an inverse relationship with WSI.Protein content was also positively correlated with FS and FA, and as expected, inversely with FD.
WSI had a good correlation coeffi cient with FA, perhaps because of the reduced size of the protein molecules that favoured protein-water interaction, which also increased the water-air interface.NSI also correlated positively with protein and FAN content (R=0.69 and 0.87).The positive and highly signifi cant relationships among WSI and NSI with FAN clearly indicate that the degree of hydrolysis of a protein promotes solubility.Proteolysis enhances the protein-water interactions, because as the molecular mass decreases, it simplifi es the secondary structure, increases the number of ionizable groups and exposes the hydrophobic groups, changing the physicochemical interactions of the protein with the medium (38).WAI showed an inverse relationship with UA.Since the latt er is an indicator of thermal treatment or heating index, it would be expected that the denaturation of high-protein materials aff ected its water absorption capacity.The other important group of functional properties is the fat-associated indices (FAI and EAI).EAI was correlated with foam capacity and, as described previously, inversely with protein.The EAI had positive correlation with FD (R=0.55) and negative with FS (R=-0.63)and HCC (R=-0.57).These correlations make reference to the balance between hydrophilic and hydrophobic groups exposed on the surface of vegetable and cereal proteins.
FA and FS were, on the other hand, positively correlated with UA, with correlation values of R=0.73 and 0.56, respectively.These corr elations can be associated with the heating used during the extraction process and not directly with the residual enzyme activity.FA and FS could then be associated with the degree of denaturation of the protein structure.HCC, the only functionality evaluated for protein--protein interaction showed a good correlation coeffi cient with UA (R=0.67), which meant that high UA increased HCC values.Therefore, proteins with lower denaturation due to lower exposure to heat treatments were more prone to coagulation.

Principal component analysis
Principal component analysis (PCA) was used to visualize the correlation among the physicochemical and functional properties of the twenty vegetable proteins.A 3D graphic presentation of the fi rst three components (PC1, PC2 and PC3) described 67.3 % of the variance.The PCA showed two correlated clusters as shown in Fig. 1.The fi rst group (A) is formed by pH, EC, UA, protein content (P), WSI, NSI, FAN, FA, FS, HCC and ES aft er 24 and 48 h, whereas the second group (B) includes FAI, EAI, RS, soluble solids (SS) and FD.
Group A is characterized by association with the charging properties of the (pH, EC, UA and FAN) that infl uenced the protein-water interactions.Group B included properties related to solid content.Aluko et al. (39) reported a signifi cant eff ect of the content of soluble solids on the emulsifying activity of coriander fl our and protein concentrate.This is similar to the results discussed previously regarding the high emulsifi cation capacity (52 399 m 2 /g) observed in pea fl our mainly due to its high RS mass fraction (136.65 mg/g).

Conclusions
This research characterized and compared chemical and functional properties of some vegetable and cereal proteins including commercial and new protein concentrates and hydrolysates obtained from a mixture of soya bean and maize germ.Correlations were obtained between physicochemical and functional properties in order to acquire a bett er comprehension of vegetable and cereal proteins as food ingredients.Water-related properties, such as WSI and NSI, in the soya and maize hydrolysates were higher, thus making them good options for use as ingredients in beverages.WAI was bett er in soya and maize concentrates, indicating their best suitability as extenders for sausages and related products.Fat-related properties (mainly FAI and EAI) were bett er in the pea fl our, making it a good emulsifi er option for dressings and other high-fat formulations.FA and FS were on average bett er in the soya and maize hydrolysates, which also had the best air trapping or foaming properties.The degree of protein hydrolysis was positively correlated with solubility-related parameters.Fat-associated characteristics were inversely correlated with water-associated characteristics.Foam and coagulation properties were bett er in low-heat-treated materials, which had high UA.The PCA of pea fl our and soya and maize concentrates and hydrolysates was linked within two groups, the fi rst mainly associated with foam and coagulation properties and the second related to emulsifi cation characteristics.This research characterized a set of vegetable and cereal proteins from a wide range of samples of raw materials and demonstrated relationships among their physicochemical and functional properties.

Table 1 .
Physical and chemical characterization of vegetable and cereal proteinsMean values are the average of at least three replicates±standard deviation.Mean values with diff erent lett er(s) in superscript within columns are statistically diff erent (p<0.05).EC=electrical conductivity, RS=reducing sugars, FAN=free amino nitrogen, UA=urease activity

Table 2 .
Functional properties of vegetable and cereal proteins bcdMean values are the average of at least three replicates±standard deviation.Mean values with diff erent lett er(s) in superscript within columns are statistically diff erent (p<0.05).WAI=water absorption index, NSI=nitrogen solubility index, WSI=water solubility index, FAI=fat absorption index, EAI=emulsifying activity index, ES

Table 3 .
Pearson's correlation coeffi cients between physicochemical parameters and functional properties of vegetable and cereal proteins