Food Research International 41 (2008) 606615

Contents lists available at ScienceDirect

Food Research International

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

Antioxidant properties of soy proteinfructooligosaccharide glycation systems and its hydrolyzates

Mara D. Mesa a, Jose M. Silvn b, Josune Olza a, ngel Gil a, Mara D. del Castillo b,*
a b

Instituto de Nutricin y Tecnologa de Alimentos, Universidad Granada, Ramn y Cajal 4, 18071 Granada, Spain Instituto de Fermentaciones Industriales, Consejo Superior de Investigaciones Cientcas, Juan de la Cierva 3, 28006 Madrid, Spain

a r t i c l e

i n f o

a b s t r a c t
The present research aimed to assess how heating at 95 C for 1 h followed by proteolysis affects the antioxidant properties of Maillard reaction mixtures constituted by soy protein isolate (SPI) and fructooligosaccharides (FOS) in 0.5 M phosphate buffer pH 7.4. Solutions of protein and sugars were also heated alone as controls. Glycation and cross-linking of protein was estimated by o-phaldialdehyde assay. Samples were hydrolyzed in vitro mimicking gastro-intestinal conditions and subsequently submitted to analysis. In vitro antioxidant properties were determined by LDL oxidation and oxygen radical absorbance capacity assays. Simultaneously, undigested reaction mixtures constituted by SPI and sugars were heated, fractioned by ultraltration and the fractions were characterized. Although Maillard reaction might give rise antioxidants, present data seem to indicate that neoantioxidants able to prevent LDL oxidation and to scavenge peroxyl-alkyl radicals were majority formed by thermal degradation of FOS (caramelization). Peptides derived from soy protein scavenged peroxyl radicals and did not protect LDL against cupper oxidation. 2008 Elsevier Ltd. All rights reserved.

Article history: Received 4 December 2007 Accepted 16 March 2008

Keywords: Antioxidants Fructooligosaccharides LDL oxidation Maillard reaction Soy proteins

1. Introduction A reappraisal of the benecial effect of legume seed dietary intake is currently taking place. Soy proteins have been widely used as a functional ingredient in many processed foods because of their ability to form gels with high nutritional, sensory, and physiological qualities (Junfeng et al., 2006). Soy intake is of the current interest to human health because of its potential to decrease the risk of chronic diseases such as heart diseases and cancer. Protein and peptides are responsible for most of the biological/functional activities of the legume seeds (Scarafoni, Chiara, & Duranti, 2007). Several studies have been devoted to asses the antioxidant potential of the soy protein fractions as well as the structural characterization of the most active peptides. The amino acid sequences determine the antioxidant power. Processing conditions to obtain protein isolates as well as the type of protease and the degree of hydrolysis affect the antioxidant activity of native soy protein and its hydrolyzates, respectively (Moure, Domnguez, & Paraj, 2006). Maillard reaction (MR) has been employed for improving functional properties, such as solubility, heat stability, emulsifying properties, and anti-allergenicity of proteins. This reaction has also resulted reliable for obtaining neoglycopeptides exhibiting enhanced functional (gel forming and emulsifying) properties and
* Corresponding author. Tel.: +34 915622900; fax: +34 915644853. E-mail address: delcastillo@i.csic.es (M.D. del Castillo). 0963-9969/$ - see front matter 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodres.2008.03.010

antioxidative activity from soy protein hydrolyzates and curdlan, which is a polysaccharide having attractive functional and biological properties and an approved food additive by US FDA (Junfeng et al., 2006). This complex food chemical event, considered the most common food reaction occurring during food processing and storage, is initiated by natural condensation of the e-amino groups of proteins with reducing-end carbonyl group of carbohydrates giving rise numerous compounds named as Maillard reaction products (MRP). We are constantly exposed to MRP since they are present in daily diet at considerably amounts. MRP may result benecial (Lindenmeier, Faist, & Hofmann, 2002; Somoza, 2005), as antioxidant, chemopreventive and antimutagenic agents, or harmful (Bengmark & Gil, 2007) for human health. FOS (fructooligosaccharides) is a mixture of oligosaccharides possessing unique nutritional and functional properties. FOS are very well-known and proved prebiotics indigestible food ingredient(s) that benecially affects host health by selectively stimulating the growth and/or activity of one limited number of bacteria in the colon (Gibson & Robertfroid, 1995). FOS preparation employed in the present study (Raftilose P95) contains mostly repeating fructose subunits (FFn) with the terminal fructose having reducing capacity and able to participate in MR (Huebner, Wehling, Parkhurst, & Hutkins, 2008; Tromova & de Jongh, 2004; Van de Lagemaat, Silvn, Moreno, Olano, & del Castillo, 2007). Soy protein and FOS may react under heating conditions forming new products with different biological properties than those showed by the

M.D. Mesa et al. / Food Research International 41 (2008) 606615

607

native protein. Since both food additives, soy protein and FOS, are health promoting compounds, a successful soy proteinFOS conjugate might preserve some of their benecial properties. It is also possible that during thermal treatment, FOS could have full or partial degradation to fructose or have formation of other compounds like caramelization products, possessing new biological properties (Huebner et al., 2008). Very few are known related to the antioxidant products of the latter compounds (Tsai, Hsieh, & Huang, 2004). The present study was undertaken to investigate how Maillard type conjugation between soy protein and FOS followed by proteolysis may affect the antioxidant properties of this food protein. The contribution of each compound forming the glycation mixture to its overall antioxidative activity was also assessed. 2. Materials and methods 2.1. Materials and reagents 2,20 -Azobis (2-methylpropionamidine) dihydrochloride (AAPH), 6-hydroxy-2,5,7,8-tetramethylchroma-2-carboxylic acid (trolox), bovine serum albumin (BSA), CuSO4, EDTA, butylated hydroxytoluene (BHT), furfural, 5-hydroxymethyl-2-furaldehyde (HMF), 2,5-dimethyl-4-hydroxy-3(2H)-furanone (DMHF), Triton X-100, sodium azide, H2O2 30% (w/v), sodium dodecyl sulfate (SDS), thiobarbituric acid, pepsin and pancreatin-bile extract were purchased from Sigma Chemical Co. (Germany). NaCl, HCl, NaHCO3 NaH2PO4, Na2HPO4, (NH4)2MoO4, Cl3C2O2H, and KBr were obtained from Panreac (Spain), while KI was purchased from Merk (Germany). Phosphate buffer saline (PBS) was from Gibco, Invitrogen (UK). Commercial FOS (Raftilose P95) was kindly provided by Orafti Espaa S.L. SPI 90% purity was purchased at Manuel Riesgo, S.A. (Madrid, Spain). 2.2. Glycation of soy protein with FOS and fructose SPI (25 mg/ml) and commercial FOS (500 mg/ml) were dissolved in 0.5 M phosphate buffer pH 7.4 in a molar ratio of primary amino groups to FOS 1: 52 (SPI-FOS). The mixtures were transferred to well-capped vials and heated in a water-bath at 95 C for 1 h with constant stirring. Since commercial FOS contains a signicant amount (6% w/w) of fructose, which could be an important competitor of the reducing FOS in the reaction, a SPI-fructose control system containing SPI (25 mg/ml) and fructose (38 mg/ml) in a weight ratio equivalent to that present in the SPIFOS system, which corresponded to a molar ratio of primary amino groups to fructose 1: 16.8 (SPIF), was included in the study. Solutions of SPI, fructose and FOS alone were also prepared and employed for controlling the occurrence of other thermal degradation reactions besides MR. After sampling, fractions were cooled immediately in an ice water-bath followed by storage at 20 C. 2.3. Extent of glycation The degree of lysine loss due principally to glycation was estimated with the uorogenic o-phtaldialdehyde (OPA) assay described in principle by Goodno, Swaisgood, and Catignani (1981) as modied by Ramirez-Jimenez, Garcia-Villanova, and GuerraHernandez (2004). 2.4. Sample fractionation Heated and unheated samples (SPI, FOS, fructose, SPI-FOS, and SPI-F) were submitted to ultraltration. This step of sample preparation allowed a preliminary characterization of the molecular

weight of new compounds occurring in the samples. Briey, two milliliters of samples (1 mg/ml) were placed in the sample reservoir of a Centricon YM-10 Centrifugal Filter Devices (Millipore) and centrifuged at room temperature, 5000g for 1 h. Proteins and other compounds possessing molecular weight higher than 10 kDa were retained (HMW), while those exhibiting lower mass were ltered (LMW). Filters were washed with water and concentrated samples recovered and dissolved in a nal volume of 2 ml. The ltrates containing compounds having molecular weight lower than 10 kDa that were obtained by washing of the samples were mixed, freeze-dried, reconstituted with 2 ml of distilled water and stored at 20 C until analysis. Protein concentration of the fractions obtained by ultraltration was determined by the micro BCA Protein Assay Kit (Pierce Biotech, Rockford, USA) (Smith et al., 1985). 2.5. In vitro gastrointestinal digestion mimicking physiological conditions Samples of native SPI, heated SPI, heated SPI-F, and heated SPIFOS were hydrolyzed. Since after consumption as part of the diet, proteins derived compounds are digested, absorbed and metabolized, we decided to investigate the production of antioxidant peptides from native and glycated soy proteins and their potential effect on key plasma targets like LDL. In vitro digestion was performed as described by Rubio and Seiquer (2002). A gastric digestion followed by an intestinal phase was carried out at 37 C and pH 2.0 and 7.5, respectively. The degree of digestibility of the samples was estimated as total nitrogen remaining in solution after in vitro digestion, which was determined by Kjeldalh (Lynch & Barbano, 1999). 2.6. In vitro antioxidant assessments Changes in antioxidant activity induced by thermal treatment and proteolysis of the samples were evaluated by means of two different methods known as in vitro LDL oxidation and oxygen radical absorbance capacity employing uorescein (ORACFL). 2.7. In vitro LDL oxidation 2.7.1. LDL isolation Fasting blood samples were collected into pre-chilled EDTAcoated tubes from healthy volunteers. Plasma was immediately separated by centrifugation at 1750g for 15 min and 4 C. LDL was isolated by single discontinuous density-gradient ultracentrifugation in a vertical rotor VTI-50 employing a vacuum Beckman ultracentrifuge, at 242,000g during 2.5 h and 4 C (Chung, Wilkinson, Geer, & Segrest, 1981). Isolated LDL was exhaustively dialyzed overnight against 150 mM NaCl, pH 7.47.6, at 4 C. Fresh LDL isolates were used for the oxidation experiments in the following four days. LDL protein was measured employing BCA reagent kit. 2.7.2. LDL oxidation LDL isolate was oxidized in PBS at 37 C with 20 lM CuSO4 as described below. The oxidation carried out in the absence (oxidation control) or presence of the samples was assayed. LDL oxidation was followed up by measuring conjugated dienes, lipid hydroperoxides, and thiobarbituric acid substances (TBARS) which are formed at different LDL oxidation stages. Trolox concentrations of 0.13, 0.53, 1.3 and 2.6 lg/ml were also assayed as a reference. Kinetic of LDL oxidation in presence of in vitro digested samples was followed by analysis of those three markers, while the effect of both undigested samples (whole protein) and their fractions containing molecules with higher and lower molecular masses than 10 kDa was assayed by analysis of conjugated dienes. Prior

608

M.D. Mesa et al. / Food Research International 41 (2008) 606615

to analysis, tested digested samples were diluted to obtain nal protein concentrations equivalent to those corresponding to Trolox (0.13, 0.53, 1.3 and 2.6 lg/ml). FOS concentrations were 2.6, 10.6, 26, and 52 lg/ml. Fructose concentrations were 0.2, 0.8, 2.0, and 4.0 lg/ml. After ltration, whole sample (undigested reaction mixture), retentates and ltrates were diluted in 2 ml of water and storage at 20 C until analysis. 2.7.3. Conjugated dienes LDL (20 lg/ml) was oxidized with 20 lM CuSO4 at 37 C, in a temperature-controlled multidetection microplate reader (Synergy HT, BIO-TEK Instruments, Vermont, USA). The formation of conjugated dienes was monitored at 234 nm for 300 min (Esterbauer, Striegl, Puhl, & Rotheneder, 1989). Lag phase results are expressed in minutes. The lag phase was calculated from the course time graph (OD234nm and time) as the time point, where the linear extrapolation of the propagation phase intercepts y = y(t0). Measurements were carried out in triplicate. 2.7.4. Lipid hydroperoxides LDL (100 lg/ml) was oxidized with 20 lM CuSO4 at 37 C. After 4 h oxidation was stopped and induced lipid hydroperoxide measured at 365 nm after 1 h incubation of 25 lg of oxidized LDL with 1 ml of tri-iodide reagent, as previously described (el-Saadani et al., 1989). Quantitative analysis of hydroperoxides formation was performed in triplicate, employing a calibration curve of H2O2. Data were expressed as nmol of hydroperoxyde/mg of LDL protein. 2.7.5. TBARS formation LDL (200 lg/ml) oxidation was induced by CuSO4 addition at 37 C for 16 h. TBARS were determined at 535 nm as described by Buege and Aust (1978). Quantitative analysis of the digested samples was performed employing a 1,1,3,3-tetraetoxypropane calibration curve. Data were expressed as nmol of malondialdehide/mg of LDL protein. Analyses were performed in duplicate. 2.8. ORAC assay The peroxyl radical scavenging activity of soy products was measured by ORACFL assay as described by del Castillo, Gordon, and Ames (2005). Fluorescein stock solution (100 lM) in phosphate buffer (75 mM, pH 7.5) was prepared and stored at 4 C in dark conditions. Trolox stock solutions in phosphate buffer (75 mM, pH 7.5) were aliquoted (2 ml) and stored at 20 C until analysis. A fresh working solution of uorescein (48 nM, 2.225 ml) was premixed with sample, standard or phosphate buffer (375 ll) and incubated for 30 s. The uorescence (FL) (kexc = 493 nm and kem = 515 nm) was measured in a spectrouorimeter (RF-1501, Shimadzu Corporation, Kyoto, Japan). This reading was FL at time zero (f0). The assay was initiated by adding AAPH (143 mM, 375 ll). Mixtures were kept in a water bath at 37 C. FL readings were taken every 5 min after AAPH addition. The FL decay curve was plotted and the area under the curve (AUC) calculated. Blanks were run by replacing sample with phosphate buffer. Sample FL values were corrected for the blank value. Trolox (20 lM) was analyzed with every batch of sample as a quality control measure. ORAC values were calculated as previously reported (del Castillo et al., 2005). Data were expressed as lg trolox equivalents (TE)/mg protein. All determinations were carried out at least in duplicate. 2.9. CE analysis of caramelization products Formation of carbohydrate breakdown products (caramelization products) was checked by employing a micellar electrokinetic CE method for analysis of HMF proposed by Morales and Jimnez-

Perez (2001). Separation was performed on an uncoated fused silica capillary and employing as running buffer 50 mM phosphate buffer pH 7.5 containing 100 mM SDS. A G1600A (Agilent, Madrid, Spain) capillary electrophoresis instrument equipped with a ChemStation software was employed for CE analysis. Electropherograms (e-grams) were monitored at 280 nm, and the spectra collected from 190 to 600 nm. Identity of the caramelization products (furfural, HMF, and DHMF) was conrmed by migration times, standard addition and spectral analysis. 2.10. Statistical analysis All data are presented as mean SEM, unless stated otherwise. Before any statistical analysis, all variables were checked for normality and homogeneous variance by using the KolmogorovSmirnov and the Levene tests, respectively. As all variables tested did not adjust to a normal distribution, to compare the effects of treatments and concentrations we preformed non-parametric statistical tests. The MannWhitney U-test was used for evaluation of differences between control and each tested groups, while KruskalWallis tests were done to evaluate mean differences in antioxidant capacity between different concentrations of tested samples. A value of P 6 0.05 was considered signicant. Data were analyzed by using a statistical software package (SPSS for Windows, 13.0, 2005, SPSS Inc. Chicago, IL, USA). 3. Results 3.1. Blockage of free amino groups during heating Glycation of SPI with FOS and fructose at 95 C during 1 h was followed by the OPA assay. A reduction in SPI primary free amino groups of 76% and 88% was observed in the glycation mixtures of SPIF and SPIFOS, respectively. Heating of native SPI caused a reduction of 15% of its free amino groups. 3.2. Degree of proteolysis Soluble total nitrogen values of 64%, 62%, and 60% were found for native SPI, heated SPIF, and heated SPIFOS samples, respectively. Data suggested that digestibility of SPI was practically unaltered by glycation under the conditions used in the present research. However, heating of the native protein dropped in 14% the degree of SPI proteolysis (47.9%), which may be ascribed to steric hindrance effects caused by protein cross-linking. 3.3. Effect of heating and proteolysis on the formation of in vitro LDL-oxidation inhibitors The oxidation of LDL gave rise to typical conjugated dienes in a time-dependent manner, both in the presence or absence of the tested compounds. The lag phase of the negative control, consisting only of LDL and CuSO4, was 55.1 0.1 min (mean SEM). Trolox was employed as positive reference at four different concentrations (0.13, 0.53, 1.3, and 2.6 lg/ml). Lag times of 160 min and 300 min were obtained by addition of Trolox solutions of 0.13 and 0.53 lg/ml that correspond to 0.5 and 2.0 lM, respectively. Fig. 1 shows how heating of SPI with reducing sugars based on fructose followed by digestion affect LDL oxidation in vitro. In Fig. 1A, the ability of the digested glycation mixtures affecting the formation of conjugated dienes can be observed. The presence of antioxidants in the samples able to prevent the formation of

M.D. Mesa et al. / Food Research International 41 (2008) 606615

609

250

A. Lag Phase from Digested Samples

200

*d

min

150

100

*c b * * * * * * * * *a

50

C
500

cSPI

SPI

SPI+FOS

SPI+F

B. Hydroperoxides from Digested Samples

400

nmoles/mg prot

300

200

100

C
50

cSPI

SPI

SPI+FOS

SPI+F

C. TBARS from Digested Samples

40

nmoles/mg prot

30

*
20

10

cSPI

SPI

SPI+FOS

SPI+F

Control

0.13 g/ml

0.53 g/ml

1.3 g/ml

2.6 g/ml

Fig. 1. Role of the tested samples on the in vitro LDL oxidation. (A) Conjugated dienes, data are mean SEM. (B) Hydroperoxides, data are expressed in nmol of hydroperoxides/mg of LDL protein as mean SEM. (C) TBARS, data are expressed in nmol of malondialdehide/mg of LDL protein as mean SEM. Digested unheated SPI (cSPI) and heated SPI (SPI), SPI plus FOS (SPIFOS) and SPI plus fructose (SPIF) were added to LDL oxidation medium. *Denotes statistically signicant difference compared with control LDL (P 6 0.05). Different superscripts denote statistical differences between different amounts of each sample (P 6 0.05).

these compounds was estimated by the ability of tested samples to prolong the lag phase compared to a negative control. Peptides released by proteolysis of soy protein, both native and heated at

95 C for 1 h, did not inhibit the formation of conjugated dienes in vitro. Similarly, mixed SPI-F did not prevent the formation of these oxidation products. However, the mixture corresponding to

610

M.D. Mesa et al. / Food Research International 41 (2008) 606615

Table 1 Peroxyl scavenging activity (ORACFL values) of hydrolyzed samples lg of TE/mg protein Control SPI Digested samples SPI Heated SPI Heated SPIF Heated SPIFOS 1.3 0.065 9.5 0.475 6.2 0.310 9.6 0.480 17.2 0.860

FOS (Fig. 3B) result stronger scavenger of peroxyl radicals than the retentates constituted by compounds possessing HMW (Fig. 3A). Results agreed with those on LDL oxidation in vitro found by analysis of fractionated samples (Fig. 2). 3.5.3. Chemical characterization of the new antioxidants Datas on protein and browning obtained by analysis of the fractions exhibiting antioxidant properties are shown in Table 2. For comparison also data on fructose containing glycation mixture and solution of fructose alone is also provided in Table 2. Results indicated that $30% of protein is ltrated. Data on browning (absorbance value at 420 nm) suggested higher concentrations of colored compounds were formed in samples containing FOS compared with the rest of the samples. Thermal treatment of FOS and fructose alone induced a greater browning than MR did. Slight color development was observed in fructose samples compared to FOS samples which may be explained on the base of differences in sugar concentration. Therefore, it can be assumed that ltrates were constituted by compound possessing molecular weight lower than 10 kDa, including small proteins, like soy proteases inhibitors, MRP, depending on the presence or not of protein in the sample, free sugars and sugars derived products, while retentates contained molecules having molecular weight higher than 10 kDa, presumably soy proteins and MRP covalently attached to soy proteins. OPA data have supported the occurrence of the MR. Fig. 4 shows the e-grams obtained by analysis of fractions containing compounds with molecular mass lower than 10 kDa. In samples containing FOS (C and D) heating induced the formation of new compounds. Although the composition of these samples was different (C did not contain protein, while D contains 0.55 mg/ml of soy protein with molecular mass <10 kDa), very similar CE proles were observed suggesting the generation of the same products at the same range of concentration in both samples. Peaks corresponding to HMF and DMHF have been identied in thermal treated samples containing FOS. Migration time and spectral data of the peaks agreed with those of the pure standards (A). The identity of the two major peaks (tm 3.7 and 3.8 min, respectively) derived of FOS thermal degradation has been not established (C and D). Identity of the caramelization products (furfural, HMF, and DHMF) was conrmed by migration times, standard addition and spectral analysis. Spectral data of the two major peaks detected in the ltrates containing FOS are shown in Fig. 5. Spectra were similar to neither HMF nor furfural. The unidentied peaks shown similar spectra to that observed for DMHF (Fig. 5A) but different migration time. The neocompounds were in lM concentration. No degradation of fructose under the conditions here employed was observed. Data on FOS degradation are supported by color results. Results indicated that a partial thermal degradation of FOS and scarce formation of caramelization products have been taken. 4. Discussion In this study, it was shown how heating of Maillard models based on soy protein and FOS and its proteolysis affect soy protein antioxidant activity. The hydrolyzates obtained by enzymatic proteolysis of SPI and its MR mixtures were added to human LDL. In vitro LDL oxidation was used as a biological system to evaluate a potential protective activity of the tested samples against cardiovascular diseases. Oxidative hypothesis is widely accepted as a mechanism of atherogenesis (Navab et al., 2004; Zaman, Helft, Worthley, & Badimon, 2000). Dietary antioxidants may protect against oxidative events related to the pathogenesis of atherosclerosis in the body, in cases where insufcient levels of endogenous antioxidants cannot counteract reactive species (Lapointe, Couil-

SPIFOS was able to prolong signicantly the lag phase in a dosedependent manner when it was added in a concentration higher than 0.53 lg/ml of protein. Addition of 1.3 lg/ml of digested SPIF decreased hydroperoxides formation by LDL oxidation compared to the negative control (Fig. 1B), while different concentrations of sugar glycated SPI reduced TBARS formation compared to the tested negative control (Fig. 1C). 3.4. Effect of heating and proteolysis on the formation peroxyl radical scavengers Table 1 shows data on peroxyl radical scavenging activity of native SPI, glycation mixtures and its hydrolyzates. The antiradical activity of SPI was improved by proteolysis. However, the thermal treatment of SPI prior to its proteolysis negatively affected the antioxidant power of soy protein hydrolyzates. Fructose did not have any contribution to the scavenger activity of soy hydrolyzates, while FOS did. The overall antioxidant activity of SPIFOS MR system hydrolyzates was dramatically increased compared to native SPI hydrolyzates (from 9.5 to 17.2 lg of TE/ mg protein). Both processes, heating and proteolysis, contributed to the formation of new scavenger of peroxyl radicals. 3.5. Fractionation of Maillard reaction system The antioxidant properties of the fractions obtained by ultraltration of the MR systems and the corresponding control samples were analyzed and the fraction containing the new antioxidants submitted to further chemical characterization. 3.5.1. Inhibition of LDL oxidation by the fractions Fig. 2 shows data on inhibition of LDL oxidation measured as the ability of the fractions to prolong the lag phase of conjugated dienes formation. As can be observed, native SPI, fructose, FOS and the corresponding mixtures, in the range of concentration studied in the present paper, did not prevent LDL oxidation induced by copper (Fig. 2A). However, heating of FOS induced the formation of neocompounds able to delay in vitro LDL oxidation (Fig. 2B). Filtrates of heated FOS and SPI-FOS samples elongated the lag phase of dienes formation in a dose-dependent manner, while the rest of the samples did not affect LDL oxidation (Fig. 2C). In addition, the retentates did not caused any inhibition of LDL oxidation (Fig. 2D) suggesting that the HMW products do not greatly contribute to the overall antioxidant activity of the tested samples. Therefore, according to the ultraltration process, antioxidants had molecular weights lower than 10 kDa. 3.5.2. ORAC assay results obtained by analysis of the fractions Fig. 3 shows the corresponding FL decay curves induced by AAPH in presence of different samples. Addition of AAPH causes a rapid decay of the FL signal due to uorescein, which may be avoided by the addition of antioxidants. In the presence of antioxidants, the AUC was increased compared to that obtained for a negative control of uorescein and AAPH. Data presented in this gure indicate that the ultraltrates containing LMW compounds from

M.D. Mesa et al. / Food Research International 41 (2008) 606615

611

200

A. Untreated Samples.

150

min

100

50

ab b

b *a

*b ab a ab

ab ab b a

C
200

SPI

FOS

SPI+FOS

SPI+F

B. Treated Samples
*d *c *c a ab b ab a *b * * *d

150

min

100

*b a

50

C
200

SPI

FOS

SPI+FOS

SPI+F

C. Ultrafiltrates

*c *c *b *b *a a *a *a

150

min

100

50

C
200 150

SPI

FOS

SPI+FOS

SPI+F

D. Retentates

min

100

50 0

* *

SPI

FOS

SPI+FOS

SPI+F

Control

0.13 g/ml

0.53 g/ml

1.3 g/ml

2.6 g/ml

Fig. 2. Effect of high (retentates) and low molecular (ulfraltrates) weights compounds fractionated from tested samples by ultraltration on the formation of conjugated dienes. Data are mean SEM. Unheated (A) and treated total (B), ultraltrated (C) and retained (D) fraction of SPI (SPI), FOS (FOS), fructose (F), SPI plus FOS (SPIFOS) and SPI plus fructose (SPIF) samples were analyzed. *Denotes statistically signicant difference compared with control LDL (P 6 0.05). Different superscripts denote statistical differences between different amounts of each sample (P 6 0.05).

lard, & Lemieux, 2006; Osawa & Kato, 2005; Rietveld & Wiseman, 2003). Data presented in this paper demonstrated that heating of SPIFOS mixtures, under specic conditions of temperature, pH and time, results in the generation of strong antioxidants active against LDL oxidation and this benecial activity seems to be associated to FOS.

Data on free amino groups suggested the successful formation of MRP in the tested samples, which was also supported by SDS PAGE (data not shown). According to our data, MRP formed under the conditions employed in the present work are not major contributors to the overall antioxidant activity of the assayed samples. These compounds in the concentrations generated in this study did

612

M.D. Mesa et al. / Food Research International 41 (2008) 606615

200

A. Retentates
Fluorescence Units
150

100

50

0 0 5 10 15 20 25 30 35

Time (min)
250

B. Ultrafiltrates

Fluorescence Units

200

150

100

50

0 0 10 20 30 40 50 60 70 80 90 100

Time (min)

Control

Trolox

FOS FO

SPI-FOS

SPI-F

Fig. 3. FL decay curves induced by AAPH in the presence of fractionated samples under study. (A) Shows samples possessing molecular mass higher than 10 kDa: retentates, while (B) fractions containing compounds having molecular mass lower than 10 kDa: ultraltrates. Data are expressed as arbitrary FL units.

Table 2 Chemical characterization of ltrates (<10 kDa) obtained by ultraltration Samples FOS FOS (95 C, 1 h) FOSSPI FOSSPI (95 C, 1 h) Fructose Fructose (95 C, 1 h) FSPI FSPI (95 C, 1 h) Protein content (mg/ml) Browning (A420 0.0380 0.0002 0.6640 0.0127 0.0380 0.0002 0.2520 0.0014 0.0370 0.0014 0.0760 0.0003 0.0370 0.0014 0.0545 0.007
a nm)

0.5510 0.0037 0.5106 0.006

0.5057 0.0021 0.5290 0.0013

a Samples were diluted at a nal concentration of 1 mg of protein/ml and the absorbance values of the fraction obtained by ultraltration read at 420 nm.

not prevent in vitro LDL oxidation and showed weak scavenge activity of peroxyl radicals. The capacity of some MRP to inhibit copper induced in vitro LDL oxidation has been previously reported by Dittrich et al. (2003). These authors concluded that the amino reductones formed by MR between glucose and equimolar amount of amino acids (glycine, lysine and arginine) during 1 h under reux have strong potential to inhibit the oxidation of LDL in vitro. In the present work, MRP have been obtained from soy proteins

and FOS in 0.5 M phosphate buffer pH 7.4 at 95 C for 1 h. As can be observed reactants and reaction conditions are very different than those employed by Dittrich et al. (2003). The MR is inuenced by many factors such as temperature, time, pH, water activity and reactants. Changes of any of these parameters will alter the reaction rate and reaction pathways. Therefore, the formation of MR compounds having different structure and function is expected. Our ndings agreed with those reported by Jing and Kitts (2004) who have found that the source of sugar composing the MRP determine their antioxidant properties. According to the data published by these authors, MPR of ribose and casein exhibited both hydroxyl and DPPH free radical scavenging capacity, while those of glucoseand fructose-casein MRP demonstrated only hydroxyl free radical scavenging activity in the Fenton assay. Under our conditions, a high concentration of reducing sugar was employed to increase the rate of the chemical reaction. However, other chemical reactions beside Maillard, like caramelization, inducing formation of new antioxidants are also favored (Tsai et al., 2004). The occurrence of this thermal degradation reaction was also supported by browning of the samples containing fructose of FOS (Table 2). Caramelization of ketose sugar fructose, which also forms FOS, has been suggested to contribute to the browning (Jing & Kitts, 2002).

M.D. Mesa et al. / Food Research International 41 (2008) 606615

mAU 4 3 2 1 0 -1 -2 0
mAU 4 3 2 1 0 -1 -2 0 1 2 3 4 5 6 7 min

613

HMF Furfural

DMHF

min

mAU 4 3 2 1 0 -1 -2 0 mAU 4 3 2 1 0 -1 -2 0 mAU 4 3 2 1 0 -1 -2 0

min

min

min

Fig. 4. Electropherograms of carbohydrates breakdown products. Mixture of HMF, DMHF, and furfural (A). Ultraltrate of unheated FOS (B). Low molecular weight fraction (LMW < 10 kDa) from FOS treated at 95 C for 1 h (C). LMW fraction from SPIFOS mixture heated at 95 C for 1 h (D). LMW fraction from SPI-fructose mixture heated at 95 C for 1 h (E).

614

M.D. Mesa et al. / Food Research International 41 (2008) 606615

mAU 15 12.5 10 7.5 5 2.5 0 -2.5 -5 200 mAU 12 10 8 6 4 2 0 -2 -4 -7.5 -6 200 mAU 250 300 350 400 450 500 550 nm mAU 200 250 300 350 400 450 500 550 nm 250 300 350 400 mAU 450 500 550 nm

B1

10 7.5 5 2.5 0 -2.5 -5

B2

8 6 4 2 0 -2 200 250 300 350 400 450 500

C1

10 8 6 4 2 0 -2 -4 -6

C2

550

nm

200

250

300

350

400

450

500

550 550

nm

Fig. 5. UVVis spectra of the major peaks detected by CE analysis of carbohydrates breakdown products. Panel A, pure DMHF; panel B, major peaks from the low molecular weight (LMW) fraction of heated FOS, B1, and B2 eluted at 3.72 and 3.84 min, respectively; panel C, spectra of the major peaks found in LMW from glycation mixture SPIFOS, C1 and C2 showed migration times of 3.72 and 3.84 min, respectively.

Results shown in Figs. 2 and 3 suggest that FOS caramels with molecular masses lower or equal to 10 kDa might prevent in vivo LDL oxidation by means of free radicals scavenging. These products may be used as natural antioxidants to be added to food as a multifunctional food ingredient. The antioxidant properties of caramelization products of fructose have been previously described by Pittoti, Elizalde, and Anese (1995). Our ndings (Figs. 4 and 5) suggested the formation of small antioxidants such heterocyclic compounds including furans (HMF), which may be major responsible of the overall antioxidant activity ascribed to the samples containing FOS. Heterocyclic compounds like furans have shown very strong antioxidant activity comparable to that of a-tocopherol (Yanagimoto, Ochi, Lee, & Shibamoto, 2004). Thermal processing generally enhances the digestibility of proteins and carbohydrates, although if Maillard browning occurs in baked foods, protein quality and digestibility can be reduced (Gibson, Perlas, & Hotz, 2006; Oste, 1991). In the present work, thermal treatment of soy proteins was carried out at 95 C during 1 h fol-

lowed by in vitro proteolysis mimicking gastrointestinal digestion. We have observed that digestibility of soy proteins was signicantly decreased by thermal processing (64% vs 49.7%) but not after glycation with both fructose and FOS. Although our data indicated a slight decrease of the proteolysis due to glycation with FOS (64% vs 60%), it can be concluded that under our specic conditions of time, temperature and pH, the degree of glycation achieved did not greatly affect the activity of the digestive enzymes. Antioxidant properties of soy have been associated to soy protein and its interaction with isoavones, specically the fractions 11S, or glycinin, has been reported as the precursor of antioxidant peptides (Gibbs, Zougman, Masse, & Mulligan, 2004). Antioxidants peptides have been obtained either from soy-fermented foods (Gibbs et al., 2004) and soy protein fractions (Moure et al., 2006). In agreement, the results presented in this study demonstrated that the antiperoxyl radical of soy protein is enhanced by proteolysis (from 1.3 to 9.5 lg of TE/mg of protein). Therefore, the scavenger properties of hydrolyzates of glycated SPI with FOS can be

M.D. Mesa et al. / Food Research International 41 (2008) 606615

615

ascribed to an additive or synergic effect of products formed by both processes thermal treatment and proteolysis. In the present study was assumed that the antioxidants formed under the conditions here described might be absorbed, reaching uids and tissues and protecting LDL against oxidation. Absorption of Maillard derived antioxidants from bread crust was demonstrated by Somoza et al. (2005), since they found that supplementation with this food enhanced total antioxidant capacity and plasma tocopherol and decreased plasma TBARS concentrations. Data demonstrated soy peptides are able to scavenge peroxyl radicals; however, they did not prevent LDL oxidation in vitro in the concentrations here assayed. Other radicals besides peroxyls may be linked to LDL oxidation, which cannot be scavenged by those peptides. Thus, further research should be conducted in order to fully elucidate bioavailability and the antioxidant mechanism of soy peptides under physiological conditions. Results also showed that the digestion process does not affect the formation of FOSderived antioxidants, suggesting that they might be available for later absorption. In conclusion, our results demonstrated that heating of FOS at 95 C for 1 h in a phosphate buffered solution pH 7.4 leads to the formation of antioxidants, and pinpoint FOS as important source of antioxidants able to scavenge peroxyl radicals and to prevent in vitro LDL oxidation. As far as we know, no study of the antiperoxyl radical activity of FOS and related products has been reported. Results are of interest for food industry because FOS is commonly used as functional food ingredient in thermally processed foods. Therefore, its thermal stability is relevant. Further investigations should be carried out in other to know how thermal processing may change FOS biological activity, as well as to fully identify the compounds formed during heating responsible for this activity, pathway of formation, stability under conventional processing conditions, bioavailability and mechanism of action. Moreover, it has been observed that heating of FOS and soy protein hydrolysis allow the generation of peroxyl radical scavengers. MRP formed under our conditions seem do not possess the antioxidant properties assessed in the present study. Acknowledgements The authors would like to thank Orafti Espaa S.L. for providing fructooligosaccharide samples. This study was funded by the Spanish Ministry of Education and Science (Fellowship SB2003-0045, Project AGL2004-05031/ALI). References
Bengmark, S., & Gil, A. (2007). Productos nales de la glicacin de la lipoxidacin como amplicadores de la inamacin: papel de los alimentos. Nutricin Hospitalaria, 22, 625640. Buege, J. A., & Aust, D. S. (1978). Microsomal lipid peroxidation. Methods in Enzymology, 52, 302310. Chung, B. H., Wilkinson, T., Geer, J. C., & Segrest, J. P. (1981). Preparative and quantitative isolation of plasma lipoprotein: rapid, single discontinuous density gradient ultracentrifugation in a vertical rotor. Journal of Lipid Research, 21, 284291. del Castillo, M. D., Gordon, M. H., & Ames, J. M. (2005). Peroxyl radical-scavenging activity of coffee brews. European Food Research and Technology, 221, 471477. Dittrich, R., el-Massry, F., Kunz, K., Rinaldi, F., Peich, C. C., Beckmann, M. W., et al. (2003). Maillard reaction products inhibit oxidation of human low-density lipoproteins in vitro. Journal of Agricultural and Food Chemistry, 51, 39003904. el-Saadani, M., Esterbauer, H., el-Sayed, M., Goher, M., Nassar, A. Y., & Jrgens, G. (1989). A spectrophotometric assay for lipid peroxides in serum lipoproteins using a commercially available reagent. Journal of Lipid Research, 30, 627630. Esterbauer, H., Striegl, G., Puhl, H., & Rotheneder, M. (1989). Continuous monitoring of in vitro oxidation of human low density lipoprotein. Free Radical Research Communications, 6, 6775. Gibbs, B. F., Zougman, A., Masse, R., & Mulligan, C. (2004). Production and characterization of bioactive peptides from soy hydrolysate and soyfermented food. Food Research International, 37, 123131.

Gibson, R. S., Perlas, L., & Hotz, C. (2006). Improving the bioavailability of nutrients in plant foods at the household level. Proceedings of the Nutrition Society, 65, 160168. Gibson, G. R., & Robertfroid, M. B. (1995). Dietary modulation of the human colonic microbiota: Introducing the concept of prebiotics. Journal of Nutrition, 125, 14011412. Goodno, C. C., Swaisgood, H. E., & Catignani, G. L. (1981). A uorimetric assay for available lysine in proteins. Analytical Biochemistry, 115, 203211. Huebner, J., Wehling, R. L., Parkhurst, A., & Hutkins, R. W. (2008). Effect of processing conditions on the prebiotic activity of commercial prebiotics. International Dairy Journal, 18, 287293. Jing, H., & Kitts, D. D. (2002). Chemical and biochemical properties of casein-sugar Maillard reaction products. Food and Chemical Toxicology, 40, 10071015. Jing, H., & Kitts, D. D. (2004). Chemical characterization of different sugar-casein Maillard reaction products and protective effects on chemicalinduced cytotoxicity on Caco-2 cells. Food and Chemical Toxicology, 42, 18331844. Junfeng, F., Yanyan, Z., Szesze, T., Fengjuan, L., Manyu, Z., Saito, M., et al. (2006). Improving functional properties of soy protein hydrolysate by conjugation with curdlan. Journal of Food Science, 71, C-285C-291. Lapointe, A., Couillard, C., & Lemieux, S. (2006). Effects of dietary factors on oxidation of low-density lipoprotein particles. Journal of Nutrition and Biochemistry, 17, 645658. Lindenmeier, M., Faist, V., & Hofmann, T. (2002). Structural and functional characterization of pronyl-lysine, a novel protein modication in bread crust melanoidins showing in vitro antioxidative and phase I/II enzyme modulating activity. Journal of Agricultural and Food Chemistry, 20, 69977006. Lynch, J. M., & Barbano, D. M. (1999). Kjeldahl nitrogen analysis as a reference method for protein determination in dairy products. Journal of AOAC International, 8, 13891398. Morales, F. J., & Jimnez-Perez, S. (2001). Effect of malindialdehyde in the determination of furosine in milk-based products. Journal of Agricultural & Food Chemistry, 48, 680684. Moure, A., Domnguez, H., & Paraj, J. C. (2006). Antioxidant properties of ultraltration-recovered soy protein fractions from industrial efuents and their hydrolysates. Process Biochemistry, 41, 447456. Navab, M., Ananthramaiah, G. M., Reddy, S. T., Van Lenten, B. J., Ansell, B. J., Fonarow, G. C., et al. (2004). The oxidation hypothesis of atherogenesis: the role of oxidized phospholipids and HDL. Journal of Lipid Research, 45, 993 1007. Osawa, T., & Kato, Y. (2005). Protective role of antioxidative food factors in oxidative stress caused by hyperglycemia. Annals of the New York Academic of Sciences, 1043, 440451. Oste, R. E. (1991). Digestibility of processed food protein. Advances in Experimental Medicine and Biology, 289, 371388. Pittoti, A., Elizalde, B. E., & Anese, M. (1995). Effect of caramelization and Maillard reaction products on peroxidase activity. Journal of Food Biochemistry, 18, 445457. Ramirez-Jimenez, A., Garcia-Villanova, B., & Guerra-Hernandez, E. (2004). Loss of ophthaldialdehyde reactivity during storage of infant cereals. International Journal of Food Science and Nutrition, 55, 143148. Rietveld, A., & Wiseman, S. (2003). Antioxidant effects of tea: evidence from human clinical trials. Journal of Nutrition, 133, 32853292. Rubio, L. A., & Seiquer, I. (2002). Transport of amino acids from in vitro digested legume proteins or casein in Caco-2 cell cultures. Journal of Agricultural and Food Chemistry, 28, 52025206. Scarafoni, A., Chiara, M., & Duranti, M. (2007). Molecular nutraceutics as a means to investigate the positive effects of legume seed proteins on human health. Trends in Food Science and Technology, 18, 454463. Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gartner, F. H., Provenzano, M. D., et al. (1985). Measurement of protein using bicinchoninic acid. Analytical Biochemistry, 150, 7685. Somoza, V. (2005). Five years of research on health risks and benets of Maillard reaction products: An update. Molecular Nutrition and Food Research, 49, 663672. Somoza, V., Wenzel, E., Lindenmeier, M., Grothe, D., Erbersdobler, H. F., & Hofmann, T. (2005). Inuence of malt, bread crust and a pronylated protein on the activity of chemopreventive enzymes and antioxidative defense parameters in vivo. Journal of Agricultural and Food Chemistry, 53, 81768182. Tromova, D., & de Jongh, H. H. J. (2004). Modication of beta-globulin and oligofructose: impact on protein adsorption at the airwater interface. Langmuir, 20, 55445552. Tsai, P. J., Hsieh, Y. Y., & Huang, T. C. (2004). Effect of sugar on anthocyanin degradation and water mobility in a roselle anthocyanin model system using 17O NMR. Journal of Agricultural and Food Chemistry, 52, 30973099. Van de Lagemaat, J., Silvn, J. M., Moreno, F. J., Olano, A., & del Castillo, M. D. (2007). In vitro glycation and antigenicity of soy proteins. Food Research International, 40, 153160. Yanagimoto, K., Ochi, H., Lee, K. G., & Shibamoto, T. (2004). Antioxidative activities of fractions obtained from brewed coffee. Journal of Agricultural and Food Chemistry, 52, 592596. Zaman, A. G., Helft, G., Worthley, S. G., & Badimon, J. J. (2000). The role of plaque rupture and thrombosis in coronary artery disease. Atherosclerosis, 149, 251266.