Impact of food matrix and processing on the invitro bioaccessibility of vitamin C, phenoliccompounds, and hydrophilic antioxidant activityfrom fruit juice-based beverages

Mara Janeth Rodrguez-Roque a, Begoa de Ancos b,Concepcin Snchez-Moreno b, M. Pilar Cano c, Pedro Elez-Martnez a,Olga Martn-Belloso a,*a Department of Food Technology, University of Lleida, Agrotecnio Center, Av. Alcalde Rovira Roure 191,Lleida 25198, Spainb Department of Characterization, Quality and Safety, Institute of Food Science, Technology and Nutrition(ICTAN-CSIC), Spanish National Research Council (CSIC), C/ Jos Antonio Novais 10, Madrid 28040, Spainc Department of Food Biotechnology and Microbiology, Institute of Food Science Research (CIAL, CSIC-UAM),C/Nicols Cabrera 9, Campus de la Universidad Autnoma de Madrid, Madrid 28049, Spain

A R T I C L E I N F O

Article history:

Received 20 August 2014

Received in revised form 9 January

2015

Accepted 19 January 2015

Available online

A B S T R A C T

The effect of food matrix (water-, milk-, or soymilk-fruit juice beverages) and processing

[high-intensity pulsed electric fields (HIPEF); high-pressure processing (HPP); and thermal

treatment (TT)] on the in vitro bioaccessibility of vitamin C and phenolic compounds, as well

as on the hydrophilic antioxidant activity (HAA) of fruit juice-based beverages was analysed.

HIPEF and HPP improved or did not change the bioaccessibility of vitamin C and certain

phenolic compounds in comparison with untreated beverages. In contrast, TT diminished

the bioaccessibility of most of these compounds.The greatest vitamin C bioaccessibility was

obtained in soymilk-fruit juice beverages (SB), whereas water-fruit juice beverages (WB) fa-

voured the bioaccessibility of phenolic compounds and HAA. Milk-fruit juice beverages (MB)

reduced the bioaccessibility of these hydrophilic constituents. Results showed that both food

matrix and processing modulated the bioaccessibility of vitamin C and phenolic com-

pounds of fruit juice-based beverages. Furthermore, HPP and HIPEF allow obtaining beverages

with improved nutritional and functional quality.

2015 Elsevier Ltd. All rights reserved.

Keywords:

Fruit juice-based beverages

Bioaccessibility

Food matrix

Food processing

Vitamin C

Phenolic compounds

1. Introduction

Currently, the food industry is attracting consumer attentionthrough functional foods and beverages that besides beinghighly nutritious and healthy, are easy to prepare and consume

(Wootton-Beard & Ryan, 2011). Innovation, convenience andquality are considered as important marketing tools in the foodindustry to increase sales (Marsells-Fontanet, Elez-Martnez,& Martn-Belloso, 2012). In this context, fruit juice-based bev-erages are becoming more popular since they represent an easyand convenient way of consuming fruits, which are important

* Corresponding author. Department of FoodTechnology, University of Lleida, Agrotecnio Center, Av. Alcalde Rovira Roure 191, 25198 Lleida,Spain. Tel.: +97 3702593; fax: +97 3702596.

E-mail address: omartin@tecal.udl.es (O. Martn-Belloso).http://dx.doi.org/10.1016/j.jff.2015.01.0201756-4646/ 2015 Elsevier Ltd. All rights reserved.

j o u rna l o f f un c t i ona l f o od s 1 4 ( 2 0 1 5 ) 3 3 4 3

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/ locate / j ff

sources of health-promoting compounds, such as vitamin Cand phenolic compounds.

Vitamin C is an essential nutrient for the biosynthesis ofcollagen and certain hormones. Its intake has been related toreduced risk of cancer and cardiovascular diseases (Li &Schellhorn, 2007). On the other hand, it has been reported thatdiets rich in phenolic compounds correlate with the de-crease of neurodegenerative disease and some cancer types(Aboul-Enein, Berczynski, & Kruk, 2013). In addition, both kindsof bioactive compounds are good contributors to the antioxi-dant activity of food (Barba, Corts, Esteve, & Frgola, 2012a).

Blended fruit juices are often combined with milk andsoymilk to improve the sensory and nutritional characteris-tics of the final product. Moreover, milk is a rich source ofproteins, unsaturated fatty acids, vitamins, carotenoids, andminerals, among others (Antone, Sterna, & Zagorska, 2012;Claeys et al., 2013), while soymilk contains high amounts ofphenolic compounds, isoflavones, proteins, iron and niacin(Jinapong, Suphantharika, & Jamnong, 2008; Rodrguez-Roque,Rojas-Gra, Elez-Martnez, & Martn-Belloso, 2013b).

Liquid foods have traditionally been preserved by thermaltreatment (TT) to prevent microorganism spoilage and con-tamination with pathogens. However, this treatment leads tothe loss of healthy compounds and sensory properties offood (Odriozola-Serrano, Aguil-Aguayo, Soliva-Fortuny, &Martn-Belloso, 2013). Non-thermal food preservation tech-nologies, such as high-intensity pulsed electric fields (HIPEF)and high-pressure processing (HPP), have been developed asalternative to heat treatments in order to satisfy consumerdemand for nutritious, healthy and safe products with afresh-like appearance (Barbosa-Cnovas, Tapia, & Cano, 2005).Previous studies have reported that both technologies (HIPEFand HPP) inactivate microorganisms and enzymes withoutcompromising the nutritional and sensory quality of food(Kadam, Jadhav, Salve, & Machewad, 2012; Odriozola-Serranoet al., 2013; Sanchez-Moreno, De Ancos, Plaza, Elez-Martinez,& Cano, 2009).

Processing is expected to modify the food matrix. Thesechanges could exert a significant influence on the release, trans-formation and absorption of some nutrients during digestion(Parada & Aguilera, 2007). The fraction of bioactive com-pounds released from the food matrix following digestion thatis solubilised into the gut for intestinal uptake is usually knownas the bioaccessible fraction (Carbonell-Capella, Buniowska,Barba, Esteve, & Frgola, 2014). From the points of view of nu-tritional and functional value of beverages, information aboutthe concentration of bioactive compounds reaching thebioaccessible fraction is much more important than the con-centration of these compounds in the corresponding beverage.Moreover, it is also important to determine the influence of foodmatrix on the bioaccessibility of bioactive compounds, espe-cially in beverages because they are complex media that allowinteractions between bioactive compounds, nutrients and/orother constituents of food (Cilla et al., 2012; Kilara, 2006;Rodrguez-Roque, Rojas-Gra, Elez-Martnez, & Martn-Belloso,2014b). Thus, analysing the extent to which food matrix andprocessing may modify the interactions, the stability, and thebioaccessibility of bioactive compounds is an essential first stepfor better understanding the biological activity of foodconstituents.

Although in vivo studies, like human intervention studies,provide more specific information about the bioavailability ofbioactive compounds, in vitro digestion models are consid-ered valuable and useful methodologies for estimating pre-absorptive events as stability and bioaccessibility of nutrientsand bioactive compounds from food (Alminger et al., 2014).

While the influence of HIPEF and HPP on the concentra-tion of bioactive compounds of beverages has previously beenevaluated (Barbosa-Cnovas et al., 2005; Kadam et al., 2012;Odriozola-Serrano et al., 2013; Sanchez-Moreno et al., 2009),information about the impact of these technologies on thebioaccessibility of nutrients is really scarce (Cilla et al., 2012).These are the reasons why this research aimed to assess theeffect of the food matrix (water-, milk- and soymilk-fruit juicebeverages) and processing (HIPEF, HPP and TT) on the in vitrobioaccessibility of vitamin C and phenolic compounds, as wellas on the hydrophilic antioxidant activity of beverages basedon a blend of fruit juices (orange, pineapple, kiwi and mango).

2. Material and methods

2.1. Materials and reagents

Pepsin from porcine stomach, pancreatin from porcine pan-creas, bovine bile, phenol standards (caffeic, chlorogenic, ferulic,sinapic, p-coumaric and p-hydroxybenzoic acids; hesperidin,naringenin, rutin, quercetin and [+]-catechin), 1,1-diphenyl-2-picrylhydrazyl radical (DPPH) and cellulose dialysis membrane(molecular weight cutoff of 12,000 Da) were purchased fromSigma-Aldrich (St. Louis, MO, USA). Ascorbic acid and FolinCiocalteu (F-C) reagent were acquired from Scharlau ChemieS.A. (Barcelona, Spain).

2.2. Fruit juice-based beverages

Fruits (orange, kiwi, pineapple and mango) were purchased atcommercial maturity in a local supermarket (Lleida, Spain).Fruits were washed, peeled and juice extracted. Each fresh-squeezed juice was filtered with a cheesecloth using a vacuumpump. A blended fruit juice was obtained by mixing 40% (w/v)of orange, 33% (w/v) of kiwi, 13.5% (w/v) of pineapple and13.5% (w/v) of mango juices.

Whole milk (Hacendado, Cordoba, Spain) and soymilk (Yosoy,Girona, Spain) were also purchased at a local supermarket.Milk composition consisted of 3.6% (w/v) of fat, 3.0% (w/v) ofprotein and 4.5% (w/v) of carbohydrates, while 1.8% (w/v) offat, 3.6% (w/v) of protein, 0.7% (w/v) of carbohydrates and1% (w/v) of fibre were contained in soymilk (data provided bymanufacturers).

Afterwards, three different fruit juice-based beverages wereprepared by mixing 75% (w/v) of the blended fruit juice (orange,kiwi, pineapple and mango juices); 7.5% (w/v) of sugar; and17.5% (w/v) of milk (milk-fruit juice beverage, MB), soymilk(soymilk-fruit juice beverage, SB), or distilled water (water-fruit juice beverage,WB).The pH of the beverages was adjustedto 3.30 0.20 (Crison Instruments S.A., Alella, Barcelona, Spain)with citric acid. The soluble solid content was determined ina refractometer Comecta S.A., Abrera (Barcelona, Spain),

34 j o u rna l o f f un c t i ona l f o od s 1 4 ( 2 0 1 5 ) 3 3 4 3

resulting in 18.0 0.2, 18.5 0.2, 19.3 0.3 brix forWB, SB andMB, respectively. Beverage formulations were selected accord-ing to previous studies, in which a high bioaccessibility ofbioactive compounds was reached (Rodrguez-Roque,Rojas-Gra, Elez-Martnez, & Martn-Belloso, 2014a;Rodrguez-Roque et al., 2014b).

2.3. Processing technologies

2.3.1. High-intensity pulsed electric fields (HIPEF)HIPEF treatment was performed in a continuous-flow benchscale system (OSU-4F,The Ohio State University, Colombus, OH,USA), using square-wave pulses. Eight collinear chambers se-rially connected were used as treatment system. Each chamberconsisted of two stainless steel electrodes separated by a gapof 0.29 cm.The flow rate was adjusted to 60 mL/min and con-trolled by a variable speed pump (model 752210-25, Cole PalmerInstrument Company, Vermon Hills, IL, USA). HIPEF process-ing conditions for beverages were 35 kV/cm electric fieldstrength in bipolar mode, 4-s pulse width, 200 Hz pulse fre-quency and 1800 s total treatment time.Temperature was keptbelow 35 C using a cooling coil connected before and after eachpair of chambers and submerged in an ice-water shaking bath.HIPEF conditions were selected based on previous studiescarried out in our laboratory, where the nutritional and mi-crobiological stability of beverages based on fruit juice andmilk/soymilk was accomplished (Morales-de la Pea,Salvia-Trujillo, Rojas-Gra, &Martn-Belloso, 2011; Salvia-Trujillo,Morales-de la Pea, Rojas-Gra, & Martn-Belloso, 2011).

2.3.2. High-pressure processing (HPP)Beverages were treated in a high hydrostatic pressure unit witha vessel of 2925 mL capacity, a maximum pressure of 900 MPa,and a maximum temperature of 100 C (High Pressure Iso-Lab System, Model FPG7100:9/2C, Stansted Fluid Power LTD.,Essex, UK). Previously to HPP, all three beverages were vacuumpacked in flexible Doypack bags (Polyskin XL, Flexibles His-pania, S.L.) (300 mL). Afterwards, they were introduced in thepressure unit filled with pressure medium (water). Beverageswere processed at 400 MPa with a holding time of 5 min. Therates of compression and decompression were both 3 MPa/s.Because of adiabatic compression, the maximum tempera-ture in the vessel was 40 C at 400 MPa. Pressure, time andtemperature were controlled by a computer program, being con-stantly monitored and recorded during the process. Theseconditions were selected based on previous studies where thenutritional and microbiological stability of HPP fruit juice-based beverages were achieved (Muoz, De Ancos,Snchez-Moreno, & Cano, 2007; Snchez-Moreno et al., 2005).

2.3.3. Thermal treatment (TT)A tubular stainless-steel heat exchanger coil immersed in ahot water shaking bath was used to treat beverages by heat(University of Lleida, Lleida, Spain). The flow rate of bever-ages was maintained through a gear pump. Beverages werethermally treated at 90 C for 60 s. After heating, the bever-ages were immediately cooled down to 5 1 C in an ice-water bath.

2.4. In vitro gastrointestinal digestion

Once processed, beverages were digested following the in vitromethodology described by Rodrguez-Roque, Rojas-Gra,Elez-Martnez, and Martn-Belloso (2013a). This method con-sisted of two sequential stages: gastric (pH 2, containing pepsin)and small intestinal digestions with dialysis (pH 7, contain-ing a pancreatin-bile mixture). Aliquots of digested beverageswere collected from the dialysed fraction at the end of the di-gestive process and immediately placed in a cold water bathduring 10 min. All samples were frozen (45 C) until analy-sis. The samples were analysed within a period time of twomonths.

2.5. Bioactive compounds analyses

2.5.1. Vitamin CVitamin C was extracted, separated, identified and quantifiedby HPLC according to Rodrguez-Roque et al. (2013a). VitaminC identification was carried out by comparing its retention timeand spectra with the standard (ascorbic acid) and using a cali-bration curve (R2 = 0.9989, at the concentration range of 10to 1000 mg/L). Results were expressed as mg of ascorbicacid/100 mL of sample.

2.5.2. Phenolic compounds analysed by HPLCExtraction, separation, identification and quantification of phe-nolic compounds by HPLC were performed following themethodology of Rodrguez-Roque et al. (2013b). Individual phe-nolic compounds were identified by comparison of theirretention time and spectra with those of the standards (caffeic,chlorogenic, ferulic, sinapic, p-coumaric and p-hydroxybenzoicacids; hesperidin, naringenin, quercetin, rutin and (+)-catechin).Quantification was carried out by integrating the peak areasand using calibration curves (R2 in the range of 0.9978 to 0.9999,concentration between 5 and 500 mg/L). Results were ex-pressed as mg of phenolic compound/100 mL of sample. Totalphenolic acids and total flavonoids were determined as the sumof individual compounds of each family of phenolic sub-stances. The concentration of total phenolic compounds (TPCby HPLC) was the sum of total phenolic acids and total flavo-noids determined individually by HPLC.

2.5.3. Total phenolic content analysed by FolinCiocalteu(TPC by F-C) methodologyTotal phenolic compounds were determined according toRodrguez-Roque et al. (2013b), based on the method previ-ously described by Singleton, Orthofer, and Lamuela-Ravents(1998). A calibration curve of gallic acid (R2 = 0.9990, concen-tration in the range of 50 to 2000 mg/L) was used to quantifythe concentration of total phenolic compounds in each sample.Results were expressed as mg of gallic acid equivalents/100 mLof sample.

2.5.4. Hydrophilic antioxidant activity (HAA)Extraction of hydrophilic fraction of non-digested or digestedbeverages was carried out based on the procedure reported byRodrguez-Roque et al. (2013b). The antioxidant activity wasevaluated through the colorimetric method (DPPH) reported

35j o u rna l o f f un c t i ona l f o od s 1 4 ( 2 0 1 5 ) 3 3 4 3

by Brand-Williams, Cuvelier, and Berset (1995). Results wereexpressed as percentage of DPPH inhibition.

2.6. Bioaccessibility calculations

Bioaccessibility was determined using Eq. 1 and was ex-pressed as percentage.

Bioaccessibility xBC

BCdigested

non digested

%( ) =

100 Eq. 1

where BCdigested corresponded to the bioactive compound con-centration in the digested beverage and BCnon-digested was thebioactive compound concentration in non-digested beverage.

2.7. Statistical analysis

Each bioactive compoundwas extracted and analysed two timesfrom two independent experiments. Results were reported asthe mean standard deviation. Analysis of variance (ANOVA)followed by the least significant difference test (LSD) wereapplied to the results obtained to verify whether there weresignificant differences (p < 0.05) in the concentration andbioaccessibility of bioactive compounds from beverages in re-lation to the factors studied in this research (food matrix andprocessing). Multifactorial analysis of variance (ANOVA) wasperformed to study separately the main effects (food matrixand treatment) and the interaction effect (food matrix treat-ment). As a significant interaction effect was observed in mostof the variables,ANOVA, comparing the means within the samefood matrix for different treatments and within the same treat-ment for different food matrix, was performed. All analyseswere carried out with the statistical program Statgraphics Plus5.1 (Statistical Graphics Corporation, Inc., Rockville, MD, USA).

3. Results and discussion

3.1. Vitamin C

The concentration of vitamin C from the three beveragesanalysed (WB, MB and SB) is reported in Table 1. In untreatedbeverages, the concentration of vitamin C was between 29.5and 30.8 mg of ascorbic acid/100 mL with no significant dif-ferences among beverages.These results were within the rangereported in similar products, which varied from 9.3 to 41.6 mg/100 mL (Andrs, Villanueva, Mateos-Aparicio, & Tenorio, 2014;Morales-de la Pea, Salvia-Trujillo, Rojas-Gra, &Martn-Belloso,2010).

According to the results obtained in this study, the vitaminC concentration was significantly influenced by treatments(HIPEF, HPP andTT), but not by the food matrix (p > 0.05). HIPEFprocessing reduced the concentration of this compound in therange of 8 to15% as compared with those untreated. Bever-ages treated by HPP did not change their content of vitaminC in comparison with untreated ones, with the exception ofSB, where a decrease of 10.5% was obtained.When both treat-ments (HIPEF and HPP) were compared, no significant changesin the concentration of vitamin C from WB and SB were

observed. On the other hand, the greatest losses of this bioactivecompound were reached in TT samples (up to 31% as com-pared with untreated beverages). In line with these results, Cillaet al. (2012) and Morales-de la Pea et al. (2010) reported thatfruit juice-based beverages processed by HIPEF and HPP, re-spectively, had their vitamin C concentration decreased between11 and 13%. Other authors also showed that the content ofvitamin C in HIPEF or HPP samples was very close to those ofuntreated beverages, whereas the losses of this bioactive werealways higher in TT beverages (Barba et al., 2012a; Morales-dela Pea et al., 2010; Snchez-Moreno et al., 2005; Zulueta, Barba,Esteve, & Frgola, 2013).

3.1.1. Bioaccessibility of vitamin CVitamin C bioaccessibility of untreated and treated (HIPEF, HPPand TT) beverages was in the range of 10.9 to 23.2% (Table 1).Previous studies have shown a vitamin C bioaccessibilitybetween 11.5 and 15% in beverages based on fruit juices plussoymilk or milk, and in blended fruit juices (Rodrguez-Roqueet al., 2013a; Rodrguez-Roque, Rojas-Gra, Elez-Martnez, &Martn-Belloso, 2014a; Rodrguez-Roque et al., 2014b). Simi-larly, this compound was between 7.2 and 12.58% bioaccessiblein HPP treated and untreated fruit juice-beverages contain-ing soymilk (Cilla et al., 2012).

The treatment by non-thermal processing (HIPEF and HPP)did not modify the bioaccessibility of vitamin C in compari-son with untreated beverages, except for MB treated by HPPwhich increased by 8%. In contrast, significant losses in thevitamin C bioaccessibility were observed inTT beverages, beingreduced 16.5 and 11.6% inWB and SB, respectively. It has beenreported that thermal treatment of food promotes the nutri-ent release through cell rupture (breakage of the cell wall) orcell separation (individual cells or clusters of cells become de-tached from each other) and can also enhance the bioavailability

Table 1 Concentration and bioaccessibility of vitamin Cin fruit juice-based beverages.a

Vitamin C

WB MB SB

Concentration (mg/100 mL)b

Untreated 30.4 2.0bA 30.8 2.1cA 29.5 1.7cAHIPEF 25.8 1.3aA 27.9 1.6bA 27.1 1.5bAHPP 28.0 1.8abAB 29.5 2.0bcB 26.4 1.5bATT 25.7 1.1aC 21.3 0.9aA 23.9 1.0aB

Bioaccessibility (%)c

Untreated 13.3 0.8bB 11.3 0.5aA 23.2 1.2bCHIPEF 14.2 0.9bB 11.9 0.5aA 21.3 1.0abCHPP 13.7 0.6bA 13.1 0.8bA 23.0 1.6bBTT 11.1 0.6aA 10.9 0.7aA 20.5 1.4aB

a Values are expressed as the mean standard deviation. Differ-ent lower case letters in the same column show significantdifferences (p < 0.05) within treatments. Different capital lettersin the same row indicate significant differences (p < 0.05) withinbeverages.

b Concentration of vitamin C in non-digested fruit juice-basedbeverage.

c The bioaccessibility of vitamin C was calculated using equation 1.WB, water-fruit juice beverage; MB, milk-fruit juice beverage; SB,soymilk-fruit juice beverage. HIPEF, high-intensity pulsed electricfields; HPP, high-pressure processing; TT, thermal treatment.

36 j o u rna l o f f un c t i ona l f o od s 1 4 ( 2 0 1 5 ) 3 3 4 3

of several nutrients by releasing them from the food matrix(Wollstonecroft, Ellis, Hillman, & Fuller, 2008). However, vitaminC is a thermo-labile compound, which is very susceptible tochemical and enzymatic oxidation during processing (Hotz &Gibson, 2007).The vitamin C and oxidative enzymes (i.e. ascor-bic acid oxidase and peroxidase) may bring into contact whenfood matrix is disrupted by the thermal treatment(Martnez-Hernndez, Arts-Hernndez, Gmez, & Arts, 2013).Therefore, oxidative reactions of vitamin C inTT beverages couldexplain why in this research the lowest bioaccessibility ofvitamin C was obtained in TT beverages. On the other hand,HIPEF and HPP are able to inactivate some of these oxidativeenzymes (Snchez-Moreno et al., 2005), avoiding the oxida-tion of vitamin C and thus maintaining its active form andbioaccessibility in beverages treated by these methods. Similarresults were reported in a study carried out with twelve healthysubjects, in which an orange juice treated by HIPEF or HPP pre-served the in vivo bioavailability and the antioxidantcharacteristics of vitamin C of the fresh product(Snchez-Moreno et al., 2003, 2004). Conversely, Cilla et al. (2012)reported that HPP diminished significantly the bioaccessibilityof vitamin C in both milk- and soymilk-based fruit bever-ages, while it increased in TT products.

Overall, the food matrix had a significant influence (P < 0.05)on the bioaccessibility of vitamin C.The highest bioaccessibilityof vitamin C was observed in SB products, followed byWB andMB. It is well known that the stability of vitamin C is influ-enced by several factors, such as oxygen availability,temperature, light, pH, metal catalyst, the presence of otherantioxidants and reducing agents, as well as possible pres-ence of ascorbic acid oxidase (Eitenmiller, Ye, & Landen, 2008).Soymilk is a rich source of phenols and isoflavones; thus, itcould be hypothesised that these antioxidant compounds couldprevent vitamin C oxidation in SB beverage. In contrast, an-tagonistic interactions between vitamin C and other foodconstituents (i.e. proteins, vitamins, metal ions) could occur,mainly in products containingmilk (MB).Milk contains vitamin-binding proteins that could lead to the formation of complexes(Claeys et al., 2013). In addition, other vitamins (B1, B2 and B12)and metal ions (Fe, Cu and Zn) contained in milk are able tointeract with vitamin C, thus increasing its degradation rate(Ball, 2006).

To the best of our knowledge, there are few literature reportsconcerning the food matrix effect on the bioaccessibility ofbioactive compounds, including vitamin C, from blended bev-erages. Cilla et al. (2012) reported that the highest bioaccessibilityof vitamin C was obtained in fruit beverages containing 16.5%(v/v) of whole milk, while the lowest was in that made with42.5% (v/v) of soymilk. Differences in the results obtained inthe present research and in that reported by Cilla et al. (2012)could be explained by the proportion of juices, milk or soymilkused to prepare the beverages.Additionally, these authors (Cillaet al., 2012) determined the vitamin C bioaccessibility in thesupernatants obtained after centrifugation of the intestinaldigesta (3300g/1h at 4 C) instead of the dialysed fraction (asin this research). In a previous study carried out in our labo-ratory, it was found that a beverage made with a similar blendof juices and 42.5% of soymilk (2.4 times higher than in thisresearch) led to 13% of vitamin C bioaccessibility(Rodrguez-Roque et al., 2014b). Taking into account the

behaviour of vitamin C in similar beverages with differentamounts of soymilk, it could be speculated that the propor-tion in which beverages are made is very important in termsof synergistic and antagonistic interactions, which also affectthe bioaccessibility of the biologically active compounds in foodsand beverages. On the other hand, Rodrguez-Roque et al. (2014a)reported that the addition of milk to a blend of fruit juicesreduced the bioaccessibility of vitamin C by 22.8%, likely dueto the interaction of this compound with other milk constitu-ents, such as proteins, vitamins and minerals.

3.2. Phenolic compounds

The concentration of individual phenolic compounds, total phe-nolic acids, total flavonoids, total phenolic compoundsdetermined by HPLC as the sum of individuals (TPC by HPLC)and total phenolic compounds determined by FolinCiocalteumethod (TPC by F-C) is presented in Tables 2 and 3.

In untreated beverages, the concentration of TPC by HPLCwas in the range of 25.4 to 34.4 mg/100 mL, and between 76and 125 mg of gallic acid equivalents/100 mL by F-C method.Flavonoids were the predominant phenolic compounds of allthe analysed beverages, while sinapic acid and (+)-catechin werenot detected.

The concentration of most phenolic acids increased between10 and 44% after applying HIPEF or HPP, with the exception offerulic acid from both MB and SB, p-coumaric acid from WBand p-hydroxybenzoic acid fromWB, which diminished theirconcentration (in the range of 9 to 31%). In general, phenolicacids from TT beverages were reduced (up to 36%) or did notchange their concentration as compared with untreated bev-erages. Hesperidin, naringenin and quercetin from both MB andSB, as well as rutin from SB were the flavonoids that in-creased their concentration (varying from 9 to 47%) aftertreatment by the technologies employed herein. The incre-ment in the phenolic concentration suggests that phenols linkedto the food matrix or to other food constituents are releaseddue to the influence of treatment, improving their extractabil-ity and therefore their content. In fact, Wang, He, and Chen(2014) reported that high pressure and high temperature in-creased the content of phenolic substances due to thebreakdown of cell wall structure and hydrolysis of polysac-charides. Additionally, it is possible that these treatmentsinactivate enzymes related to the loss of phenolic substances(such as the polyphenol oxidase) or increase the activity ofenzymes that participate in the biosynthesis of phenols (i.e.the phenylalanine ammonia-lyse) (Morales-de la Pea et al.,2011). Some studies have reported a similar trend in fruit juice-soymilk beverages treated by HIPEF or TT (Morales-de la Peaet al., 2011); in HPP orange juice-milk beverages (Barba et al.,2012a) and in orange juice treated by HIPEF or HPP(Snchez-Moreno et al., 2005).

The food matrix had a significant influence on the concen-tration of phenolic compounds (p < 0.05). SB was the beveragewhich contained the highest concentration of total phenolicacids and flavonoids, as well as TPC (by HPLC and by F-C), fol-lowed by MB andWB. However,TPC by F-C were not statisticallydifferent in MB and WB products. These results suggest thatthe addition of soymilk or milk to blended fruit juices fa-voured the concentration of these constituents, in spite of the

37j o u rna l o f f un c t i ona l f o od s 1 4 ( 2 0 1 5 ) 3 3 4 3

interactions between phenolic compounds and proteins.Thereare very few reports evaluating the effect of the food matrixon the concentration of phenolic compounds from liquid food.However, in contrast to the results obtained in this study,Sharma, Vijay Kumar, and Jagan Mohan Rao (2008) observedthat the greatest content of phenols was observed in black teabrew, followed by black tea with sugar, black tea with milk andsugar, and finally in black tea with milk. On the other hand,Barba, Esteve, and Frgola (2012b) reviewed the influence of HPPon the nutritional properties of different fluid foods, such asjuices, purees/pastes based on fruits and vegetables.They con-cluded that although some general trends were observed, theeffect of HPP depended on both the treatment intensity andthe food matrix (Barba et al., 2012b). Therefore, the effect offood processing must be separately studied in each foodmatrix.

3.2.1. Bioaccessibility of phenolic compoundsThe bioaccessibilities of individual and total phenolic com-pounds are shown in Tables 4 and 5. Overall, the bioaccessibilityof phenolic compounds in untreated beverages was in the rangeof 10.1 to 28.3%, except for rutin from MB, which was not re-covered in the dialysed fraction. In line with these results, thebioaccessibility of these compounds was in the range of 0 to29% in untreated soymilk,milk, blended fruit juices, and in bev-erages containing a blend of fruit juices plus soymilk or milk(Rodrguez-Roque et al., 2013a, 2013b, 2014a, 2014b). In addi-tion, flavanones from orange juice were around 11 to 36%bioaccessible (Gil-Izquierdo, Gil, Ferreres, & Toms-Barbern,2001).

Processing had a variable influence on the bioaccessibilityof phenolic compounds from beverages. An improvement upto 38% in the bioaccessibility of several phenolic substances(caffeic and p-coumaric acids from both WB and MB; chloro-genic and ferulic acids from MB; hesperidin and rutin from allbeverages) was observed after treatments, mainly by non-thermal methods (HIPEF and HPP). Processing (HIPEF, HPP andTT) did not change the bioaccessibility of caffeic and chloro-genic acids from SB, as well as naringenin from bothWB andMB. On the contrary, all treatments diminished thebioaccessibility of ferulic acid fromWB. The bioaccessibilitiesof chlorogenic and p-hydroxybenzoic acids fromWB were alsosignificantly reduced by HIPEF (between 10 and 11%) and TT(between 11 and 24%). Among the treatments studied in thisresearch, the lowest bioaccessibility of phenolic compoundswas found inTT beverages, specifically in ferulic acid fromWB,which was reduced 31% as compared with that untreated. Itis also interesting to note that rutin from MB was notbioaccessible in untreated and TT products; however, it dis-played bioaccessibility of 7.2 and 8.4% in HIPEF and HPPbeverages, respectively.

The effect of processing on the concentration of phenoliccompounds depends on the type of food, the nature and lo-cation of phenolic compounds in food, as well as the intensityand duration of treatment (Chandrasekara, Naczk, & Shahidi,2012). Processing is known to change some physicochemicalfeatures of phenolic compounds and thus, it may also modify(increase or reduce) the bioaccessibility of these compounds.For instance, several changes in the phenol structure (hydrox-ylation, methylation, isoprenylation, dimerisation, andglycosylation, among others) and/or the formation of phenolic

Table2Con

centrationof

phen

oliccompou

ndsin

non

-digestedfruitjuice-ba

sedbe

verages.

a

Phen

olic

compou

ndsco

nce

ntration(m

g/10

0mL)

Bev

erag

esTrea

tmen

tsCaffeic

acid

Chloroge

nic

acid

Feru

licac

idp-co

umaric

acid

p-hyd

roxy

benzo

icac

idHesperidin

Naringe

nin

Querce

tin

Rutin

WB

Untrea

ted

0.57

00.22

aA1.95

0.04

cA0.61

0.03

aA1.58

0.07

cA4.8

0.3c

A9.2

0.3b

A3.13

0.20

bcA

1.90

0.08

dA

1.64

0.06

bAHIPEF

0.64

0.04

cA1.63

0.08

bA0.65

60.02

4bA

1.37

0.03

bA3.31

0.21

aA8.2

0.5a

A2.90

0.15

bA1.47

0.10

bA1.30

0.07

aAHPP

0.63

0.03

bcA

1.87

0.09

cA0.63

0.04

abA

1.44

0.04

bA3.96

0.25

bA8.9

0.4b

A3.23

0.17

cA1.61

0.03

cA1.27

0.04

aATT

0.58

60.01

8abA

1.31

0.08

aA0.60

20.18

aA1.23

0.04

aA3.09

0.12

bA7.7

0.3a

A2.36

0.10

aA1.27

0.06

aA1.26

0.04

aAMB

Untrea

ted

0.66

0.04

aB2.40

0.08

aB0.81

30.01

9cB

1.66

0.10

aA4.4

0.3a

A11

.00.3a

B4.55

0.25

aB1.71

0.04

aB1.85

0.08

cBHIPEF

0.76

70.02

2bB

3.25

0.16

cB0.72

80.01

2bB

1.84

0.09

bB5.4

0.3b

B13

.80.3c

B5.83

0.24

cB2.17

0.11

cB1.67

0.06

bBHPP

0.80

0.04

bB3.43

0.22

cB0.73

0.05

bB1.86

0.05

bB5.7

0.4b

B15

.00.4d

B6.4

0.3d

B2.30

0.16

cB1.47

0.04

aBTT

0.70

0.04

aB2.77

0.11

bB0.66

0.03

aB1.75

0.01

abB

4.3

0.3a

B13

.00.3b

B5.33

0.13

bB1.93

0.11

bB1.38

0.08

aBSB

Untrea

ted

0.58

50.02

3aA

2.65

0.10

aC1.16

0.07

aC1.94

0.04

aB4.8

0.3a

A13

.20.3a

C6.5

0.4a

C2.13

0.11

aC1.54

0.05

aAHIPEF

0.66

0.04

bA3.72

0.18

cC0.95

0.03

bC2.21

0.15

bC5.6

0.3b

B14

.40.4b

B7.4

0.3b

C2.67

0.12

cC1.77

0.10

bBHPP

0.65

0.03

bA3.81

0.12

cC0.94

0.03

cC2.43

0.11

cC5.78

0.16

bB15

.50.9c

B8.9

0.5c

C3.0

0.14

dC

2.27

0.09

dC

TT

0.60

0.03

aA3.28

0.23

bC0.76

70.02

3aC

1.96

0.01

3aC

5.04

0.25

aC14

.60.5b

C7.5

0.3b

C2.45

0.06

bC1.90

0.04

cC

aValues

areex

pressed

asthemea

nstan

darddev

iation

.Differentlower

case

lettersin

thesa

meco

lumnforea

chbe

verage

show

sign

ifica

ntdifference

s(p