Bioactive compounds of organic goji berry (Lycium barbarum L.) prevents oxidative deterioration of soybean oil

Industrial Crops & Products 112 (2018) 90–97

Contents lists available at ScienceDirect

Industrial Crops & Products journal homepage: www.elsevier.com/locate/indcrop

Research paper

Bioactive compounds of organic goji berry (Lycium barbarum L.) prevents oxidative deterioration of soybean oil

MARK

⁎

Alessandra Cristina Pedroa, , Juliana Bello Baron Maurerb, Selma Faria Zawadzki-Baggiob, Suelen Ávilaa, Giselle Maria Macielc, Charles Windson Isidoro Haminiukc a

Programa de Pós-Graduação em Engenharia de Alimentos, Universidade Federal do Paraná, R. Cel. Francisco Heráclito dos Santos 210, Campus Politécnico, Curitiba, PR, Brazil b Departamento de Bioquímica e Biologia Molecular, Universidade Federal do Paraná, R. Cel. Francisco Heráclito dos Santos 210, Campus Politécnico, Curitiba, PR, Brazil c Programa de Pós-Graduação em Ciência e Tecnologia Ambiental, Universidade Tecnológica Federal do Paraná, Curitiba, PR, Brazil

A R T I C L E I N F O

A B S T R A C T

Keywords: Response surface methodology Natural antioxidant Superfruit Atoxic solvent

The protecting ability of the organic goji berry extract as an alternative to synthetic antioxidants against oxidation of soybean oil was investigated. For this, the extraction of phenolic compounds with high antioxidant activity was optimized. The main bioactive compounds of the optimized extract were identified and quantified. Different solvents were used in the extraction process. The effects of temperature (25–45 °C), time (60–180 min), and solid:solvent ratio (1:10–1:30, w/v) in the extraction were evaluated by a Box–Behnken experimental design. Analyses of phenolic acids and flavonoids were performed by Ultra Performance Liquid Chromatography (UPLC), in optimal extraction conditions. The effect of the organic goji berry extract on improving the soybean oil oxidative stability was evaluated using the Rancimat test. A solution of ethanol/water (70/30, v/v) was the most efficient in the extraction of phenolics with a high antioxidant activity. The optimum conditions for the extraction is obtained by the experimental design were 45 °C, 162 min and a solid:solvent ratio of 1:30, resulting in an extract with 1338.80 mg/100 g of phenolics, and antioxidant activity of 0.73, 3.66, and 2.81 mmolTE/ 100 g for FRAP, ABTS, and DPPH, respectively. The phenolic acids identified were as follows: syringic, chlorogenic, gallic, caffeic, p-coumaric, 4-hydroxybenzoic, ferulic, and trans-cinammic. The flavonoids were as follows: rutin, naringin, quercetin, catechin, and kaempferol. The protection factor (PF) values of soybean oil with organic goji berry extract (500–3000 mg/kg) were significantly higher than that of oil with butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT) (100 mg/kg). These results suggest that the organic goji berry extract may be a substitute for synthetic antioxidants in stabilizing oil against oxidative deterioration.

1. Introduction The fruits of the Lycium barbarum L., known as goji berry, wolfberry, or “gou-qi-zi” are grown mainly in Asian countries (Potterat, 2010). In China, goji berry is cultivated under two different systems, organic and conventional. The demand for organic food has increased throughout the world, mainly due to the awareness of the population about possible risks regarding the presence of chemical residues in food (LlorentMartínez et al., 2010). Goji berries are known as “superfruit” and have been used for over 2500 years in Asian countries for medicinal purposes (Amagase and Farnsworth, 2011). In scientific literature, several beneficial effects, such as the following, have been reported: treatment for eye problems (Amagase and Farnsworth, 2011; Qian et al., 2004), body weight reduction (Amagase, 2010), immunomodulatory effects (Arroyo-Martinez

⁎

et al., 2011) as well as antiproliferative activity (Hogan et al., 2010). The antioxidant properties of the goji berry are used for the development of drugs, cosmetics and special purpose foods (Pan et al., 2011). Accordingly, goji berry is considered to be a raw material for the extraction of phenolic compounds. These biologically active compounds have been added to food products to provide protection against oxidation. Many studies show that natural antioxidant compounds are as potent as synthetic antioxidants in vegetable oils (Yang et al., 2016; Franco et al., 2016; Comunian et al., 2016). These results are extremely important considering that the prolonged consumption of synthetic antioxidants may cause carcinogenic and mutagenic effects on the human body (Cordeiro et al., 2013). Therefore, in order to obtain the vegetable extract for applying as a natural antioxidant, some factors must be considered, such as the choice of a non-toxic solvent in an adequate volume and polarity for

Corresponding author. E-mail address: [email protected] (A.C. Pedro).

https://doi.org/10.1016/j.indcrop.2017.10.052 Received 23 May 2017; Received in revised form 12 September 2017; Accepted 29 October 2017 0926-6690/ © 2017 Elsevier B.V. All rights reserved.

Industrial Crops & Products 112 (2018) 90–97

A.C. Pedro et al.

(1997). Two grams of sample were mixed with 20 mL of methanol/ water (50/50, v/v) and stirred in an incubator shaker at 25 °C, 210 rpm for 60 min. The mixture was centrifuged at 4300 rpm for 20 min and the supernatant recovered. The sediment was added of 20 mL of acetone/water (70/30, v/v), and the steps of stirring and centrifugation were repeated. The extracts were combined and the volume was completed to 50 mL with ultrapure water. The sequential extraction using acetone/water/acetic acid (70/28/ 2, v/v/v) was performed according to Kevers et al. (2007), and the same extraction conditions used by Liyana-Pathirana and Shahidi (2005) (cited above) were employed. All the extracts were stored in amber glass bottles at −20 °C for further analysis.

extraction, the influence of temperature and time, biological activity and yield upon extracting bioactive compounds (Pedro et al., 2016). Food industries apply many technological operations to obtain vegetable extracts with a high yield and low cost. The response surface methodology (RSM) is a suitable approach to obtain an extract with such optimized features (Bassani et al., 2014). Therefore, to obtain a natural extract with high antioxidant content, the mathematical model of RSM may be efficiently applied, besides investigating the influences of different factors on extraction (Pedro et al., 2016). Studies regarding the organic goji berry have not yet been explored in the literature. In addition, this is the first study to indicate the application of organic goji berry extract as a natural antioxidant in edible oils, as there is no optimization study about the extraction process of these compounds with the use of non-toxic solvents for the application in the food industry. Therefore, the main objectives of this study were: (1) to optimize the extraction of total phenolic compounds with antioxidant activity by response surface methodology, (2) to identify and quantify the bioactive compounds in the optimized extract by UPLC–Ultra Performance Liquid Chromatography, and (3) to evaluate the capability of organic goji berry extract in protecting soybean oil against oxidation by the Rancimat method and comparing it with the efficacy of synthetic antioxidants: butylated hydroxyanisole (BHA), butylated hydroxytoluene (BHT), and t-butyl-hydroxyquinone (TBHQ).

2.4. Experimental design A Box–Behnken design (Table 1) was used to evaluate the effect of three independent variables in the extraction of the bioactive compounds and in vitro antioxidant activity: temperature (x1, 25–45 °C), time (x2, 60–180 min) and solid:liquid ratio (x3, 1:10–1:30 mL). These values were established previously, data not shown. The response parameters were: total phenolics, flavonoids, anthocyanins, carotenoids and antioxidant activity (FRAP, ABTS and DPPH assays). The complete design consisted of 17 combinations, including 5 replicates of the central point in order to estimate pure error and to assess the lack of fit of the proposed models. The experiments were performed randomly and in triplicate.

2. Materials and methods 2.1. Materials Folin-Ciocalteau reagent, TPTZ, DPPH, ABTS, chemical UPLC-grade standards (purity ≥ 95%), butylated hydroxyanisole (BHA), butylated hydroxytoluene (BHT) and t-butyl-hydroxyquinone (TBHQ) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Soybean oil without added antioxidants were provided by Granol Industry, Trade and Export S/A. Methanol, acetic acid, ethanol and acetone were purchased from Vetec® (Rio de Janeiro, RJ, Brazil). Ultrapure water (Milli-Q) was used in all experiments.

2.5. Total phenolic compounds (TPC) The TPC was determined using the Folin-Ciocalteau method (Singleton and Rossi, 1965). In test tubes of 1 mL were added 600 μL of ultrapure water, 200 μL of diluted sample (1/20, v/v) and 50 μL of Folin-Ciocalteau reactive. Afterwards, 150 μL of Na2CO3 (15%, w/v) was added to each tube followed by agitation. After 60 min of reaction, the absorbance was monitored at 760 nm. A calibration curve was prepared using gallic acid as standard and the results of TPC were expressed as mg of gallic acid equivalents per 100 g dry weight sample.

2.2. Preparation of samples The dehydrated samples of organic goji berry (Lycium barbarum L.) were acquired at Municipal Market of Curitiba, PR, Brazil, from the 2015 harvest and certified organic: IMO Control, Manufactured: Qingdao Ri Tai Food Co., Ltd. The fruits were ground in a mill (Marconi MA 630/1, São Paulo, Brazil) to 10 mesh and stored in dark in a vacuum packaging at 4 °C until analysis.

2.6. Total flavonoids content (TFC) The TFC was quantified using the aluminum chloride colorimetric assay (Jia et al., 1999). In test tubes of 1 mL were added 100 μL of sample, 400 μL of ultrapure water and 30 μL of NaNO2 (5%, w/v). After 5 min, 60 μL of AlCl3·6H2O (10%, w/v) was added and the solution was allowed to react for 6 min. Then, 200 μL of NaOH 1 mol/L solution and 210 μL of ultrapure water were added and mixed. After 5 min of reaction, the absorbance was measured at 510 nm. Catechin was used as a standard and the results were expressed as mg of catechin equivalents per 100 g of dried fruit.

2.3. Selection of solvents Various solvents, such as ultrapure water, acetone, ethanol and methanol, in different concentrations and mixtures were evaluated for the extraction of bioactive compounds from the goji berry. The selection of the best solvent was achieved in response to the presence of the highest quantity of phytochemicals in the extracts with high antioxidant activity (by FRAP, ABTS and DPPH assays). Solutions of ethanol/water and methanol/water were tested for the extraction of phytochemicals at the proportions of 20/80, 30/70, 50/ 50, and 70/30 (v/v). A solution of acetone/water at the proportion of 70/30 (v/v) was also used in the experiments of extraction (LiyanaPathirana and Shahidi, 2005). Two grams of sample were mixed with 20 mL of each extractive solution and stirred in an incubator shaker (Marconi MA 420, São Paulo, Brazil) at 25 °C, 210 rpm for 60 min. The supernatant was obtained by centrifugation (MPW-350R, Warsaw, Poland) at 4300 rpm for 20 min. For each extract, the volume was completed to 50 mL with the solution used in the extraction (ethanol/water, methanol/water or acetone/water). Sequential extraction was carried out according to Larrauri et al.

2.7. Total anthocyanins (TA) The TA was performed by the pH-differential method (Giusti and Wrolstad, 2001). The extract was diluted in a pH 1.0 solution (0.1 mol/ L HCl, 25 mmol/L KCl) and in a pH 4.5 solution (0.4 mol/L CH3COONa). The absorbance of the mixtures was then measured at 515 and 700 nm. The value of (Abs515–Abs700)pH1.0–(Abs515–Abs700)pH4.5 corresponds to the absorbance of the anthocyanins. Calculation of the TA was based on a cyanidin-3-glucoside, molar extinction coefficient of 26,900 and a molecular mass of 449.20 g/mol. The results were expressed as mg of cyanidin-3-glucoside equivalents per 100 g of dried fruit. 91

Industrial Crops & Products 112 (2018) 90–97

A.C. Pedro et al.

absorbance was determined at 517 nm. 2.10. UPLC analysis of phenolic compounds in optimum conditions The separation and quantification of phenolic compounds of the extract obtained in the optimized conditions was performed by Acquity UPLC H-Class (Waters, Miliford, MA, United States), column Acquity BEH C18 (Waters, Miliford, MA, United States) (2,1 × 50 mm, 1,7 μm). The extract was filtered through a 0.22 μm nylon syringe filter (Millipore, São Paulo, Brazil) and 2 μL were injected at 20 °C. Mobile phase was constituted of solvent A (ultrapure water with 0.1% formic acid, v/v) and solvent B (methanol with 0.1% formic acid, v/v). The flow rate was: 0.5 mL/min and gradient conditions: 0 min − 0% B; 8 min − 20% B; 15 min − 100% B; 18 min − 0% B and isocratic elution with 0% B to 20 min. The runs were monitored at 260, 270, 280 and 370 nm. The quantification was performed using calibration curves of standards of gallic acid, 3,4-dihydroxymethyl acid, 4-hydroxymethyl acid, caffeic acid, chlorogenic acid, syringe acid, p-coumaric acid, ferulic acid, trans-cinnamic acid, catechin, rutin, naringin, quercetin and kaempferol.

Fig. 1. Protection factor (PF) of different concentrations of organic goji berry extract and synthetic antioxidants (BHT, BHA and TBHQ) added to the soybean oil.

2.8. Total carotenoids (TC) 2.11. Rancimat analysis

The TC value was quantified using the colorimetric assay of Rodriguez-Amaya and Kimura (2004), with modifications. In a 5 mL volumetric flask, 2 mL of extracts were added and the volume was completed to ethanol 70% and the absorbance was measured at 445 nm. Calculation of the TC was based on zeaxanthin, molar extinction coefficient of 2592 (in ethanol 70%) and results were expressed as mg of zeaxanthin per 100 g of dry sample.

The phenolic extract of the organic goji berry obtained through optimized conditions was dried at around 37 °C under reduced pressure in a rotary evaporator (Fisatom–model 802, São Paulo, Brazil). The extract was mixed with refined soybean oil (without addition of antioxidants) provided by the company Granol Ind. Com. and Exp. S/A, in different concentrations (500, 1000, 1500, 2000, 2500 and 3000 mg/ kg). The result was compared with soybean oil added with synthetic antioxidants BHT, BHA and TBHQ at the concentration of 100 mg/kg. A control sample (without addition of antioxidants) was also oxidized under the same conditions. The samples were analyzed with the objective of evaluating them for oxidative stability, according to the method proposed by Abbeddou et al. (2013). The test was performed on the Rancimat (Metrohm Model 873, Switzerland) equipment, at Lacaut (Federal University of Paraná), under the following conditions: 3.0 g of oil, air flow of 20 L/h and temperature of 100 °C, whose measurement is based on the determination of the electrical conductivity of the volatile products of degradation.

2.9. Antioxidant assay Antioxidant activity was determined by three different in vitro assays [i.e., Ferric reducing antioxidant power (FRAP), 2,2-azinobis-3ethylbenzothiazoline-6-sulphonic acid (ABTS), 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical assay]. Standard curves for all assays were prepared with Trolox and results expressed in milimole of trolox equivalent per 100 g dry weight of sample. 2.9.1. FRAP antioxidant assay The ferric reducing antioxidant power (FRAP) was determined according to the method of Benzie and Strain (1996). The FRAP reagent was prepared by a mixture of acetate buffer (300 mmol/L, pH 3.6), a solution of 10 mmol/L of TPTZ in 40 mmol/L of HCl, and 20 mmol/L of FeCl3 at 10/1/1 (v/v/v). In test tubes of 1 mL were added 700 μL of acetate buffer pH 3.6, 200 μL of FRAP reagent and 100 μL diluted samples (1/20, v/v). The tubes were shaken and incubated for 40 min at 37 ± 1 °C and then the absorbance was monitored at 596 nm.

2.12. Statistical analysis Data were presented as mean ± standard deviation (SD) and the experimental design was evaluated using the Anderson-Darling normality test. The homogeneity of variances was tested by the BrownForsythe test (Action Software version 2.5, Estatcamp, Campinas, Brazil). One-factor analysis of variance (ANOVA) was performed to determine significant differences among treatments. Fisher’s LSD test was used to compare the means among treatments with significant differences (p < 0.05). When the probability value (p-value) was below 0.05, the response variable was subjected to multiple linear regression analysis using the RSM. The response function (Yn) was partitioned into linear, quadratic and interactive components. Data were modeled using Equation 1:

2.9.2. ABTS radical cation inhibition antioxidant assay The ABTS radical cation decolonization assay was determined according to the method of Re et al. (1999). The working solution was prepared by mixing: 5 mL of 7 mmol/L ABTS and 88 μL of 140 mmol/L potassium persulfate, and allowing them to react for 16 h at room temperature (25 ± 1 °C) in the dark. The solution was diluted by mixing 1 mL ABTS radical solution with 50 mL ethanol to obtain an absorbance of 1.10 at 734 nm. A total of 990 μL of an ABTS solution and 10 μL of diluted extract (1/8, v/v) were added to test tubes of 1 mL. The mixture was stored in the dark for 2 h, and then the absorbance was measured at 734 nm.

3

Yn (x ) =

∑ i=1

3

bix i +

3

∑ ∑ i≤j

bijx ix j + bijk x ix jxk

j

Where Yn is the predicted response, bi, bij, and bijk are the regression coefficients for linear, quadratic and cubic terms, respectively, xi, xj and xk are the independent variables, and 3 is the number of factors. The lack of fit test, percentage of total explained variance (r2 and adjusted r2 (r2adj)) and the model statistical significance (p-value) were used to verify the adequacy of the model. In order to maximize the extraction of total phenolics, flavonoids, anthocyanins, carotenoids and antioxidant

2.9.3. DPPH radical scavenging ability assay The ability of extracts to scavenge the DPPH• was determined according to Brand-Williams et al. (2011). In test tubes of 1 mL were added 550 μL of ethanol and 350 μL of a 223 μmol/L ethanol DPPH solution. The mixture was stored in the dark for 30 min, and then the 92

Industrial Crops & Products 112 (2018) 90–97

A.C. Pedro et al.

content ranged from 895 to 1036 mg/100 g. However, the different phenolic compound contents found in literature can be attributed to the drying process. Conventional drying (sun exposure) changes the chemical composition of the plant, releasing insoluble phenolic compounds (flavonoids) and increasing its total phenolic content (Nunes et al., 2016). In addition, phenolic compound content in goji berry depends on factors such as maturity stage of the fruit and the cultivation systems used to grow it (organic or conventional) (Pertuzatti et al., 2015). The TFC ranged statistically (p < 0.001) between 20.85 (MA) and 437.58 mg CE/100 g (AAA). Flavonoids are compounds that may be complexed with other compounds, such as sugars. The high flavonoid content extracted with the acidified acetone solution may be explained by the fact that acidified solvents provide membrane disruption of cellular material, decomplexation and protonation of the molecules, releasing soluble components (Naczk and Shahidi, 2004). The extracts obtained by the use of protic solvents, Met70 and Et70, present high flavonoids content, 100.10 and 95.90 mg/100 g, respectively. Flavonoid contents in dried goji berry determined by Qian et al. (2004) were 115.40, 120.70 and 149.70 mg/100 g in aqueous extract, 50 and 70% ethanol extract, respectively. In a study by Wang et al. (2010), the total flavonoid content in ethanol extract of goji fruit was 119.30 mg/100 g. The TA statistically varied (p < 0.001) of 0.02 (W) to 18.29 mg/ 100 g (Et70). Anthocyanins are polar molecules, due to the substituent groups (hydroxyl, carboxyl and methoxyl) and the residual glicosilas linked to aromatic rings. Because anthocyanins are glycosylated, protic polar solvents (Et70) tend to have a higher affinity with these molecules. In the literature, few studies of anthocyanins present in goji berry have been reported. Abdennacer et al. (2015) reported a low total anthocyanin content in the Lycium intractum boiss fruit (7.90 mg/100 g) and leaves (2.50 mg/100 g). The anthocyanin content in goji berry is considered low, compared with blackberry and blueberry, 5861.05 and 2972.28 mg/100 g, respectively (Souza et al., 2014). The goji berry’s color is given by carotenoids, whereas in blackberry and blueberry anthocyanins are the predominant pigment. The TC varied statistically (p < 0.001) between 31.20 (Et20) and 62.40 mg/100 g (Et80). Carotenoids are lipid-soluble compounds with the color ranging from yellow to red, and are present in large amounts in the L. barbarum L fruit (Wang et al., 2010). Zeaxanthin dipalmitate is a predominant carotenoid (31–56%) in Lycium fruits (Peng et al., 2005). Inbaraj et al. (2008) determined the zeaxanthin dipalmitate in the largest amount (114.37 mg/100 g) in L. barbarum L. fresh fruit. When compared with the literature, the dried goji fruit from the present study showed a lower carotenoid content. The results of Weng-Ping et al. (2008) showed that the content of zeaxanthin dipalmitate may decrease from 2 to 22 folds during the conventional drying process and storage. In addition, for reasons of affinity (solvent-solute), apolar solvents, such as hexane and petroleum ether, would be the most suitable for extracting a higher carotenoid content (apolar) from goji berry (Rezaiea et al., 2015). The in vitro antioxidant activity was determined by the FRAP, ABTS and DPPH assays and it varied significantly (p < 0.001) from 0.58 (MA) to 2.33 mmol TE/100 g (Met70), 4.71 (Et20) to 7.98 mmol TE/ 100 g (AAA) and 1.08 (Et20) to 2.33 mmol TE/100 g (AAA), respectively. The highest antioxidant activity by FRAP assay (Table 2) was observed in the Met70 extract, which also had a higher content of phenolic compounds. This demonstrates that the methanolic solvent is more selective for extracting phytochemicals present in organic goji berry. Alcoholic solvents can extract a wide range of glycosylated (flavonoids) and non-glycosylated phenolic compounds, which contribute to increased antioxidant activity (Philip and Christina, 2000). The AAA extract showed a higher antioxidant activity regarding ABTS and DPPH assays. The structure of flavonoids presents a high amount of hydroxyl groups, mainly the o-dihydroxy structure in the B ring, in favorable positions for electronic delocalization, contributing to a pronounced effect on antioxidant activity. Therefore, the polarity of the solvent affects the extraction of compounds with a different number

Table 1 The Box-Behnken design applied for compounds extraction of organic goji berry. True values and coded variables Assays.

Temperature (°C)

Time (min)

Solid:liquid ratio (mL)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

25(−1) 25(−1) 25(−1) 25(−1) 35(0) 35(0) 35(0) 35(0) 45(+1) 45(+1) 45(+1) 45(+1) 35(0) 35(0) 35(0) 35(0) 35(0)

120(0) 60(−1) 180(+1) 120(0) 180(+1) 60(−1) 180(+1) 60(−1) 120(0) 120(0) 180(+1) 60(−1) 120(0) 120(0) 120(0) 120(0) 120(0)

30(+1) 20(0) 20(0) 10(−1) 10(−1) 10(−1) 30(+1) 30(+1) 10(−1) 30(+1) 20(0) 20(0) 20(0) 20(0) 20(0) 20(0) 20(0)

(C) (C) (C) (C) (C)

(C) = Central point.

activity (FRAP, ABTS and DPPH) of the goji berry extract, the regression models were submitted to a multiresponse optimization approach (Derringer and Suich, 1980). Statistical analyses were performed using Statistica 7.0 software (StatSoft Inc. South America, Tulsa, United States). 3. Results and discussion 3.1. Preliminary experiments (solvent selection) The type of solvent strongly influences the efficiency of the extraction of antioxidant compounds from a plant matrix (Rezaiea et al., 2015). In the present study, a wide range of solvents with different physicochemical properties were used to extract phytochemical compounds from the organic goji berry: methanol at 20% (Met20), 30% (Met30), 50% (Met50) and 70% (Met70); ethanol at 20% (Et20), 30% (Et30), 50% (Et50) and 70% (Et70); acetone at 70% (A70); metanhol/ acetone (MA); acetone/water/acetic acid (AAA) and water (W). The solvents used showed a statistical difference (p < 0.05) in the total content of phenolics (TPC), flavonoids (TFC), anthocyanins (TA), carotenoids (TC), and antioxidant activity (Table 2). The TPC varied statistically (p < 0.001) from 822.45 (W) to 1736.36 mg GAE/100 g (Met70). As shown in Table 2, the TPC of the solvent extracts were in the order of Met70 > Et70 > Et50 > A70 > AAA > Met30 > Et20 > MA > Met20 > Et30 > Met50 > W. Solvents diluted in water in a ratio above 1:1 were more effective in the extraction of goji berry phenolic compounds. The mixture of solvents and water is more efficient than the mono-solvent system, since glycosylated phenolic compounds are more soluble in water. Besides presenting a selective behavior in the extraction of glycosylated compounds, the methanol and ethanol are also able to extract nonglycosylated compounds due to the high availability of the free electron pair (Philip and Christina, 2000). Among the analyzed solvents, Met70 and Et70 produced extracts with higher phenolic content, 1736.36 and 1351.45 mg/100 g, respectively. The high content of phenolic compounds can be explained by the high selectivity of alcoholic solvents. Phenolic compounds consist of a hydroxyl group linked with an aromatic hydrocarbon group. Different from other solvents, alcohols have a high affinity for these chemical compounds because they participate in carbon–oxygen and oxygen–hydrogen chemical interactions (Philip and Christina, 2000). Medina (2011) published similar results in the extraction of phenolic compounds from dried goji berry with 70% ethanol, where the 93

Industrial Crops & Products 112 (2018) 90–97

A.C. Pedro et al.

Table 2 Content of bioactive compounds extracted from organic goji berry with different solvents. Solvents

Phenolics (TPC) (mgGAE/100 g)

Flavonoids (TFC) (mgCE/100 g)

Anthocyanins (TA) (mg/100 g)

Carotenoids (TC) (mg/100 g)

Antioxidant activity FRAPa (mmolTE/ 100 g)

Methanol 20% (Met20) 30% (Met30) 50% (Met50) 70% (Met70) Ethanol 20% (Et20) 30% (Et30) 50% (Et50) 70% (Et70) Acetone 70% (A70) Methanol/Acetone (MA) Acetone/Acetic acid/Water (AAA) Water (W) p (Normality)d p (Levene)e p (ANOVA)f

1052.53 1084.42 1002.37 1736.36

± ± ± ±

42.89fgh 3.40ef 20.07h 3.99a

64.25 ± 1.25cd 38.18 ± 4.90gh 54.27 ± 0.00ef 100.10 ± 6.91b

1078.73 1025.15 1312.33 1351.45 1219.96 1067.11

± ± ± ± ± ±

30.46f 17.54gh 12.26c 25.06b 27.24d 19.45fg

32.54 45.34 64.70 95.90 74.19 20.85

± ± ± ± ± ±

4.34hi 6.31fg 6.12cd 15.70b 9.75d 5.01i

DPPHc (mmolTE/ 100 g)

59.97 37.28 33.47 35.00

± ± ± ±

2.47a 4.75fg 1.63hi 2.02gh

0.65 0.66 0.28 2.33

± ± ± ±

0.10efg 0.05efg 0.07gh 0.38a

6.26 5.51 6.31 5.44

± ± ± ±

0.05e 0.03f 0.09e 0.19f

1.51 0.94 1.62 1.51

± ± ± ±

0.07c 0.01e 0.05b 0.04c

1.32 ± 0.12d 0.66 ± 0.00ef 3.32 ± 0.25c 18.29 ± 0.28a 0.06 ± 0.01e 3.99 ± 0.39c

31.20 49.17 43.00 49.10 54.70 33.50

± ± ± ± ± ±

0.90i 2.31c 2.60de 2.00c 1.85b 0.79hi

1.04 1.03 0.14 1.25 0.74 0.58

± ± ± ± ± ±

0.09ef 0.00de 0.09h 0.07cd 0.16ef 0.13fg

4.71 6.19 6.35 6.88 6.99 7.40

± ± ± ± ± ±

0.47g 0.42e 0.50e 0.29cd 0.45c 0.39b

1.08 1.28 1.15 1.64 1.64 1.61

± ± ± ± ± ±

0.10g 0.08d 0.02f 0.03b 0.03b 0.05b

0.83 1.62 1.07 1.64

± ± ± ±

0.02ef 0.44d 0.12ef 0.68d

ABTSb (mmolTE/ 100 g)

1135.40 ± 40.17e

437.58 ± 11.96a

6.99 ± 0.55b

43.77 ± 1.70d

1.47 ± 0.26bc

7.98 ± 0.28a

2.33 ± 0.09a

822.45 ± 60.45i 0.20 0.85 < 0.001

57.27 ± 8.45ef 0.89 0.56 < 0.001

0.02 ± 0.01e 0.12 0.13 < 0.001

39.57 ± 0.64ef 0.42 0.98 < 0.001

0.66 ± 0.13efg 0.51 0.60 < 0.001

6.35 ± 0.24e 0.63 0.31 < 0.001

1.26 ± 0.04d 0.14 0.36 < 0.001

a

Ferric reducing antioxidant power. ABTS radical scavenging capacity. c DPPH radical scavenging capacity. d Value obtained by acording Anderson-Darling test to normality. e Value obtained acording to the Levene test for homogeneity of variances. f Value obtained by one-factor analysis of variance (ANOVA). Different letters in the same column represent results with statistical difference, according to the Fisher’s LSD test (p ≤ 0.05). b

(Souza et al., 2014). The organic goji berry can be considered as an excellent source of phenolic compounds (> 500 mg/100 g) (Vasco et al., 2008). The TFC varied from 102.82 (assay 6) to 183.32 (assay 5) mg/100 g. The model of flavonoid extraction did not present a lack of fit (p = 0.96), and it could explain 97.20% of variance in data (r2adj = 0.94). The temperature in the linear (x1) and quadratic model (x12), as well as the interaction between this variable and time (x1 × 2), tended to increase in the TFC (positive effects). The predicted model can be described by Eq. (2).

and type of functional groups in the aromatic ring, influencing the transfer of electrons and hydrogen atoms (Rezaiea et al., 2015). Based on the results obtained (Table 2), the Et70 was considered efficient for the extraction of bioactive compounds of goji berry with a high antioxidant activity. Extraction procedures using ethanol are desirable from the public health and industrial perspectives and are considered as safe for food processing, because it is not as toxic as methanol (Pedro et al., 2016). For this reason, the Et70 was chosen for the extraction of bioactive compounds of goji fruits in experimental design.

YTFC = 119.20 + 21.58x1 + 25.08x12 + 13.62x1x2 3.2. Response surface methodology (RSM)

The TA varied from 4.72 (assay 3) to 21.02 (assay 5) mg/100 g. The RSM application of the TA values showed that the model could explain 98.10% of all variance in data (r2adj = 0.95). Moreover, it did not present a lack of fit (p = 0.97). The temperature in the linear (x1) and in the quadratic model (x12), solvent ratio (x3) and the interaction of time and solvent ratio (x2 × 3), showed negative effects in TA. The interactions of temperature and time (x1 × 22) in the quadratic model, temperature and solvent ratio in the linear (x1 × 3) and in the quadratic model (x12 × 3) showed positive effects in TA, as shown in Eq. (3).

In Table 3, for all response variables (content of phenolics, flavonoids, anthocyanins, carotenoids, and antioxidant activity), homogeneity of variances (p ≥ 0.05) and significant differences (p < 0.001) were observed. The TPC ranged between 1136.63 (assay 10) to 1401.43 (assay 5) mg/100 g. The multiple regression analysis showed that the factors, temperature and time significantly affect the TPC. The model did not present a lack of fit (p = 0.92) and it could explain 78.50% of all variance in data (r2adj = 0.71). The temperature in the linear model increased phenolic extraction (x1) and time in the quadratic model (x22) had a negative effect (Eq. (1)). YTPC = 1268.73 + 69.98x1–76.07x22

(2)

YTA = 13.25 − 3.22x1 − 2.69x12 − 3.98x3 + 1.92x1x3 − 3.23x2x3 + 6.66xc1x22 + 3.90x12x3

(3)

Temperature increases the diffusion coefficient, whereas solvent viscosity decreases it, and so the solubility of the compounds increases. However, high temperatures (above 25 °C) may degrade some flavonoids, such as anthocyanins, leading to breakage of covalent bonds and the hydrolysis of flavylium cation (Pedro et al., 2016). The TC varied from 4.86 (assay 10) to 6.79 (assay 9) mg/100 g. The model did not present lack of fit (p = 0.80), and it could explain 93% of variance in data (r2adj = 0.88). Time (x2) and the interaction between this factor and temperature (x1 × 2) and solvent ratio (x1 × 3) in the linear model, decreased the extraction of carotenoids (Eq. (4)).

(1)

Prolonged extraction times may lead to the decomposition of phenolic compounds and solvent vaporization, affecting mass transfer between solvent-solute (Pedro et al., 2016). On the other hand, raising the temperature increases the solubility of phenolic compounds and reduces the extraction time. The TPC in organic goji berry (1401.43 mg/100 g), compared with other varieties of berry fruit, was higher than in blueberry, strawberry, and blackberry, 305.38, 624.92 and 850.52 mg/100 g, respectively 94

Industrial Crops & Products 112 (2018) 90–97

A.C. Pedro et al.

Table 3 The Box-Behnken design with experimental data for the extraction of bioactive compounds from organic goji berry. Response variablesa Assays

Phenolics (TPC) (mgGAE/100 g)

1 2 3 4 5 6 7 8 9 10 11 12 13 (C) 14 (C) 15 (C) 16 (C) 17 (C) p (Normality)e p (Levene)f p (ANOVA)g

1263.72 ± 1147.72 ± 1168.06 ± 1207.80 ± 1136.63 ± 1250.95 ± 1194.56 ± 1218.83 ± 1376.74 ± 1401.43 ± 1340.17 ± 1228.78 ± 1271.99 ± 1334.21 ± 1176.16 ± 1265.47 ± 1265.48 ± 0.56 0.59 < 0.001

3.66de 41.56kl 3.56jkl 22.95ghi 7.30l 24.47def 13.21hij 24.62fgh 3.62ab 22.65a 14.12bc 9.73efg 25.29d 24.90c 21.03ijk 17.01de 17.61de

Flavonoids (TFC) (mgCE/100 g) 108.94 ± 136.66 ± 112.49 ± 132.69 ± 112.99 ± 102.82 ± 122.28 ± 137.62 ± 164.47 ± 162.66 ± 183.32 ± 153.00 ± 123.66 ± 127.02 ± 111.04 ± 113.80 ± 121.59 ± 0.07 0.34 < 0.001

27.19fgh 15.65cde 19.09fgh 13.06cdef 3.02efgh 2.62h 0.00defgh 20.57cd 3.02ab 15.77ab 11.78a 13.06bc 7.36defgh 12.83defg 14.81gh 4.64fgh 2.31defgh

Anthocyanins (TA) (mg/100 g)

Carotenoids (TC) (mg/100 g)

11.90 ± 2.16de 6.55 ± 2.05h 4.72 ± 0.88h 15.92 ± 1.14b 21.02 ± 0.99a 10.97 ± 1.97efg 6.58 ± 1.55h 9.47 ± 1.71fg 5.62 ± 2.46h 9.30 ± 1.63g 11.69 ± 0.22def 13.31 ± 1.40cd 13.98 ± 0.61bcd 14.52 ± 0.23bc 13.04 ± 0.70cde 11.87 ± 0.82de 12.35 ± 0.22cde 0.51 0.89 < 0.001

65.20 ± 61.70 ± 64.10 ± 49.20 ± 53.60 ± 58.10 ± 53.40 ± 62.90 ± 67.90 ± 48.60 ± 48.60 ± 64.30 ± 64.20 ± 62.40 ± 58.80 ± 58.60 ± 59.50 ± 0.11 0.89 < 0.001

2.80ab 2.10bcd 0.90abc 5.10fg 4.90ef 3.50de 5.10efg 3.00bcd 1.40a 0.90g 2.80g 1.60abc 2.20abc 0.50bcd 0.30d 0.60d 3.90cd

FRAPb (mmolTE/ 100 g)

ABTSc (mmolTE/ 100 g)

DPPHd (mmolTE/ 100 g)

1.10 ± 0.15b 0.37 ± 0.05fg 0.65 ± 0.07cd 0.55 ± 0.09de 0.24 ± 0.14hijk 0.32 ± 0.02ghi 0.26 ± 0.04hij 1.21 ± 0.01a 0.57 ± 0.09de 0.57 ± 0.10de 0.69 ± 0.16c 0.47 ± 0.16ef 0.20 ± 0.07jk 0.32 ± 0.02gh 0.15 ± 0.02k 0.26 ± 0.04hij 0.22 ± 0.04ijk 0.64 0.42 < 0.001

2.89 ± 0.08c 2.34 ± 0.16de 0.42 ± 0.04j 1.45 ± 0.16hi 2.23 ± 0.08e 2.23 ± 0.11e 1.31 ± 0.12i 2.31 ± 0.12de 3.00 ± 0.55bc 4.64 ± 0.35a 3.25 ± 0.27b 3.27 ± 0.49b 1.87 ± 0.12fg 2.59 ± 0.27d 1.62 ± 0.16gh 1.26 ± 0.04i 2.12 ± 0.31 ef 0.73 0.15 < 0.001

2.56 ± 0.35b 2.60 ± 0.12b 2.33 ± 0.02cd 2.09 ± 0.04ef 2.39 ± 0.29bcd 2.46 ± 0.19bc 2.23 ± 0.06cde 2.62 ± 0.07b 2.85 ± 0.18a 2.66 ± 0.22b 2.46 ± 0.39cde 2.23 ± 0.04cde 2.29 ± 0.17cd 2.35 ± 0.06bcd 2.24 ± 0.09de 2.13 ± 0.26efg 2.08 ± 0.13g 0.76 0.10 < 0.001

(C) = central points. a Values are expressed as mean (n = 3). b Ferric reducing antioxidant power. c ABTS radical scavenging capacity. d DPPH radical scavenging capacity. e Value obtained by Anderson-Darling test for normality. f Value obtained by Levene test for homogeneity of variances. g Value obtained by one-factor analysis of variance (ANOVA). Different letters in the same column represent results with statistical difference, according to the Fisher’s LSD test (p ≤ 0.05).

YTC = 6.02 − 0.34x2 − 0.42x1x2 − 0.88x1x3

positively influenced the antioxidant activity. The interaction between both factors (x1 × 3) in the linear model, showed negative effects, as shown in Eq. (7).

(4)

Carotenoids are extremely sensitive to high temperatures. Therefore, high temperatures combined with long extraction times accelerate the process of carotenoid degradation (Weng-Ping et al., 2008). The ferric reducing antioxidant power (FRAP) varied from 0.22 (assay 17) to 1.21 (assay 8) mmolTE/100 g. The RSM application of the FRAP values showed that the model could explain 98.60% of variance in data (r2adj = 0.97) and the model did not present a lack of fit (p = 0.83). The temperature (x12) and the interaction between the temperature and time (x12 × 2) in the quadratic model, the solvent ratio in the linear (x3) and in the quadratic model (x32), showed a tendency to increase the antioxidant activity. The interaction between the solvent ratio and the temperature (x1 × 3) with time (x2 × 3) in the linear model, showed a tendency to decrease the antioxidant activity. The predicted model can be described by Eq. (5).

YDPPH = 2.23 + 0.15x12 + 0.17x32 − 0.16x1x3

For all the tested antioxidant assays, the temperature was the factor that had the greatest effect on antioxidant activity. This may be because temperature increases the solubility of compounds with different molecular structures, related to a greater mass transfer, with consequently greater penetration of the solvent into the vegetable matrix (Cacace and Mazza, 2003). Similar effects were found in other studies, i.e., apple extract (Alberti et al., 2014), Berberis asiatica extract (Belwal et al., 2016) and tomato extract (Li et al., 2012). A simultaneous optimization based on the desirability function was performed, aiming at maximizing the content of phenolic compounds and antioxidant capacity of organic goji berry extracts. The conditions of 162 min at 45 °C and solid:liquid ratio of 1:30 (w/v) could be considered as suitable to obtain the optimized solution for combining these variables. The overall desirability function obtained for this solution was 0.75. The optimized conditions were applied to the experimental extraction of phenolic compounds and determination of antioxidant activity. The observed and predicted mean values, along with the computed absolute errors (AE), were as follows: total phenolic compounds (mg/100 g) (observed (1338.80) and predicted (1335.54); AE = −0.44%); and antioxidant activity (mmol TE/100 g) by FRAP (observed (0.73) and predicted (0.74), AE = 1.42%), ABTS (observed (3.66) and predicted (3.63), AE = −0.51%) and DPPH assays (observed (2.81) and predicted (2.87), AE = −1.13%). A model can be used to predict a response based on the values of specific factors (independent variables) when the regression equations generated by RSM are statistically significant (p-value is below the stipulated α-value) and the equations can explain 70% more of the variability in the data (r2 > 0.70) and the p-lackoffit > 0.05 (Granato

YFRAP = 0.23 + 0.23x3 − 0.14x1x3 − 0.22x2x3 + 0.25x12 + 0.21x32 + 0.12x12x2 (5) The antioxidant activity for the ABTS assay varies between 0.42 (assay 3) and 4.64 (assay 10) mmolTE/100 g. The RSM application of ABTS values showed that the model could explain 84.50% of variance in data (r2adj = 0.75) and the model did not present a lack of fit (p = 0.57). The factor that positively affects the antioxidant activity was temperature in the linear (x1) and the quadratic model (x12) (Eq. (6)). YABTS = 1.95 + 0.88x1 + 0.71x12

(7)

(6)

The free-radical scavenging activity (DPPH) varied from 2.08 (assay 17) to 2.85 (assay 9) mmolTE/100 g. The RSM application of the DPPH values showed that the model could explain 84% of variance in data (r2adj = 0.74) and the model did not present a lack of fit (p = 0.54). The temperature (x12) and solid:solvent ratio (x32) in the quadratic model 95

Industrial Crops & Products 112 (2018) 90–97

A.C. Pedro et al.

The protection factor (PF) for soybean oils containing optimized organic goji berry extract, rich in bioactive compounds (500 at 3000 mg/ kg), and the synthetic antioxidants BHT, BHA and TBHQ (100 mg/kg) are presented in Fig. 1. High PF values indicate high antioxidant activity. The phenolic extracts (G500, G1000, G1500, G2000, G2500, and G3000) exhibited a concentration dependent increase in activity, presenting a significantly better efficiency (p < 0.05) in stabilizing the oil against oxidative deterioration than BHT and BHA (Fig. 1). However, the highest protection factor was observed for TBHQ. The high soybean oil protection factor observed by TBHQ can be explained by the high stability at elevated temperature and less volatility than BHT and BHA. In addition, the presence of two hydroxyls in their structure allows greater ease in the transfer of hydrogen atoms to stabilize free radicals formed during the oxidative process (Araújo, 2006). The high protective factor values presented by the organic goji berry extract can be explained by the presence of different antioxidant compounds identified by UPLC. The syringic and chlorogenic acids, rutin and naringin, may have been the main compounds responsible for the protective effect. Benzoic acids, such as syringic acid, present a high antioxidant action, due to the presence of two methoxyl groups in the molecular structure. The antioxidant activity of chlorogenic acid is mainly related to the presence of the ortho-dihydroxy phenyl group (catechol moiety). In addition, it presents important structural features in the stabilization of reactive species: alterations in the number and position of hydroxy groups and insertion of electron donating or withdrawing moieties as well as the modifications of the carboxylic functions that include the esterification and amidation process (Razzaghi-Asl et al., 2013). Flavonoids, like rutin, act as antioxidants primaries by attaching phenolic hydroxyl groups to ring structures, and can act as reducing agents, hydrogen donators, singlet oxygen quenchers, superoxide radical scavengers, and even as metal chelators (Fennema, 1996). Results like the present study were found in the literature. Comunian et al. (2016) showed that the addition of the rutin extract was effective in the process of oxidation protection accelerated by Rancimat of echium oil microencapsulated. Potato peel extracts, rich in chlorogenic acid, showed a high oxidative stability of soybean oil (Franco et al., 2016). Yang et al. (2016) found that rosemary extract, when acting as a natural antioxidant, had a better antioxidant capacity than BHT and BHA. The antioxidant capacity of this extract was attributed to the presence of phenolic diterpenes that scavenge singlet oxygen, hydroxyl radicals, and lipid peroxyl radicals, thereby preventing lipid oxidation. The FP values in oils with organic goji berry extract were significantly higher than that of oil with added synthetic antioxidants (BHT and BHA), indicating that the goji berry extract may be a substitute for synthetic antioxidants in stabilizing oil against oxidative deterioration.

Table 4 Chromatographic parameters and quantification of phenolic acids and flavonoids from organic goji berry analyzed by UPLC. Chemical Standards

Retention time (min)

UV band (nm)

Concentration (mg/ 100 g)

Syringic acid Chlorogenic acid Gallic acid Caffeic acid p-coumaric acid 4-Hydroxybenzoic acid Ferulic acid Trans-cinnamic acid Rutin Naringin Quercetin Catechin Kaempferol

6.31 5.57 0.97 5.13 7.00 3.31 8.42 10.83 10.26 10.38 11.22 4.89 11.70

270 300 260 300 300 260 300 280 270 280 300 280 260

946.33 ± 7.89 130.00 ± 2.52 86.33 ± 3.02 86.33 ± 3.24 55.33 ± 1.33 37.33 ± 2.26 18.67 ± 0.65 2.67 ± 0.12 665.00 ± 4.55 213.00 ± 2.71 69.33 ± 3.78 37.00 ± 1.22 35.67 ± 2.08

et al., 2014). In the present study, all created regression models are statistically significant (p ≤ 0.05), explained more than 78% of the dataset (r2adj > 0.71) and adequately adjusted the experimental data (p-lackoffit > 0.54). The predicted and observed values are in agreement (relative standard error below 10%). Therefore, the RSM model can be regarded as significant and predictive under the studied range. 3.3. Phenolic compounds in optimum point (UPLC analysis) The contents and identification of phenolic acids and flavonoids in the optimized extract (162 min at 45 °C and solid:liquid ratio 1:30 (w/ v)) from organic goji berry were determined by UPLC analysis. In the optimized extract, eight phenolic acids (syringic, chlorogenic, gallic, caffeic, p-coumaric, 4-hydroxybenzoic, ferulic, and trans-cinammic) and five flavonoids (rutin, naringin, quercetin, catechin and kaempferol) were identified (Fig. S1, S2, S3 and S4). Table 4 shows the concentration of the phenolic acids and flavonoids from organic goji berry analyzed by UPLC. High syringic acid and chlorogenic acid contents were determined, 946.33 and 130.00 mg/100 g, respectively. The main flavonoids present in the optimized organic goji berry extract were rutin and naringin, 665.00 and 213.00 mg/100 g, respectively. Wang et al. (2010) also presented high contents of chlorogenic acid and rutin in the Lycium barbarum L fruit. However, in studies conducted by Donno et al. (2015), only chlorogenic acid was found in high amounts, and rutin was not detected in goji berry samples. The syringic acid has previously never been reported in the goji berry. Several authors mention rutin as the main flavonoid in goji berry. In addition, another flavonoid, quercetin is also found, as it is a derivative of rutin deglycosylation (Potterat, 2010). Inbaraj et al. (2010) found quercetin-rhamno-di-hexoside (43.86 mg/100 g) and chlorogenic acid (23.70 mg/100 g) in the goji berry powder. Qian et al. (2004) identified a higher content of rutin (3.40 mg/mL) and chlorogenic acid (0.23 mg/ mL), and a low content of protocatechuic acid, hyperoside, morin and quercetin, 0.17, 0.13, 0.05 and 0.16 mg/mL in L. chinense Mill, respectively. Based on the results obtained in this study and data from literature, it can be said that the goji berry used (organic production) presents higher contents of phenolic acids and flavonoids compared with goji fruits produced in the conventional method, in addition to presenting a high content of syringic acid not yet mentioned in the literature regarding this fruit.

4. Conclusion Methanol and ethanol were the most effective solvents to extract phenolic compounds with a higher antioxidant activity. Et70 was chosen for the extraction of bioactive compounds from the goji fruit in experimental design because it is a solvent desirable from an industrial perspective as safe for food processing. The RSM proved to be useful to evaluate the effect of temperature, time and solid-liquid ratio on the extraction of TPC from organic goji berry. The statistical analysis based on a Box–Behnken design showed that an extraction at 45 °C for 162 min using a solid:solvent ratio of 1:30 (w/v) maximizes the TPC and in vitro antioxidant activity. The main phenolic compounds of the organic goji berry identified by UPLC analysis were syringic and chlorogenic acid, rutin and naringin. Results showed that, in comparison with synthetic antioxidants (BHT and BHA) organic goji berry extract was more effective against oxidative deterioration of soybean oil. The use of organic farming techniques and the high content of

3.4. Rancimat test The Rancimat test evaluates the antioxidant activity of a sample. It is referred to as the increase in electrical conductivity due to the formation of secondary products of lipid oxidation (Rezaiea et al., 2015). 96

Industrial Crops & Products 112 (2018) 90–97

A.C. Pedro et al.

antioxidants extracted by a non-toxic solvent, suggested that the organic goji berry extract can be applied in the food industry as a safe and natural antioxidant.

Inbaraj, S.B., Lu, H., Hung, C.F., Wu, W.B., Lin, C.L., Chen, B.H., 2008. Determination of carotenoids and their esters in fruits of Lycium barbarum Linnaeus by HPLC–DAD–APCI–MS. J. Pharm. Biomed. Anal. 47, 812–818. Inbaraj, S.B., Lu, H., Kao, T.H., Chen, B.H., 2010. Simultaneous determination of phenolic acids and flavonoids in Lycium barbarum Linnaeus by HPLC-DAD-ESI–MS. J. Pharm. Biomed. Anal. 51 (3), 549–556. Jia, Z., Tang, M., Wu, J., 1999. The determination of flavonoid contents in mulberry and their scavenging effects on superoxide radicals. Food Chem. 64, 555–559. Kevers, C., Falkowski, M., Tabart, J., Defraigne, J.O., Dommes, J., Pincemail, J., 2007. Evolution of antioxidant capacity during storage of selected fruits and vegetables. J. Agric. Food Chem. 55, 8596–8603. Larrauri, J.A., Rupérez, P., Saura-Calixto, F., 1997. Effect of drying temperature on the stability of polyphenols and antioxidant activity of red grape pomace peels. J. Agric. Food Chem. 45, 1390–1393. Li, H., Deng, Z., Wu, T., Liu, R., Loewen, S., Tsao, R., 2012. Microwave-assisted extraction of phenolics with maximal antioxidant activities in tomatoes. Food Chem. 130, 928–936. Liyana-Pathirana, C., Shahidi, F., 2005. Optimization of extraction of phenolic compounds from wheat using response surface methodology. Food Chem. 93, 47–56. Llorent-Martínez, E.J., Fernández-de-Córdova, M.L., Ortega-Barrales, P., Ruiz-Medina, A., 2010. Characterization and comparison of the chemical composition of exotic superfoods. Microchem. J. 110, 444–451. Medina, M., 2011. Determination of the total phenolics in juices and superfruits by a novel chemical method. J. Funct. Foods 3, 79–87. Naczk, M., Shahidi, F., 2004. Extraction and analysis of phenolics in food. J. Chromatogr. A 1054, 95–111. Nunes, J., Lago, M., Castelo-Branco, V., Oliveira, F., Torres, A., Perrone, D., Monteiro, M., 2016. Effect of drying method on volatile compounds, phenolic profile and antioxidant capacity of guava powders. Food Chem. 197, 881–890. Pan, M.H., Chiou, Y.S., Tsai, M.L., Ho, C.T., 2011. Anti-inflammatory activity of traditional Chinese medicinal herbs. J. Tradit. Complement. Med. 1, 8–24. Pedro, A.C., Granato, D., Rosso, N.D., 2016. Extraction of anthocyanins and polyphenols from black rice (Oryza sativa L.) by modeling and assessing their reversibility and stability. Food Chem. 191, 12–20. Peng, Y., Ma, C., Li, Y., Leung, K.S., Jiang, Z., Zhao, Z., 2005. Quantification of zeaxanthin dipalmitate and total carotenoids in Lycium fruits (Fructus Lycii.). Plant Foods Hum. Nutr. 60, 161–164. Pertuzatti, P.B., Sganzerla, M., Jacques, A.C., Barciab, M.T., Zambiazi, R.C., 2015. Carotenoids, tocopherols and ascorbic acid content in yellow passion fruit (Passiflora edulis) grown under different cultivation systems. Food Sci. Technol. 64, 259–263. Philip, S., Christina, A., 2000. Organic Chemistry. Prentice Hall, California. Potterat, O., 2010. Goji (Lycium barbarum and L. chinense) Phytochemistry, pharmacology and safety in the perspective of traditional uses and recent popularity. Plant Med. 76, 7–19. Qian, J., Liu, D., Huang, A., 2004. The efficiency of flavonoids in polar extracts of Lycium chinense Mill fruits as free radical scavenger. Food Chem. 87, 283–288. Razzaghi-Asl, N., Garrido, J., Khazraei, H., Borges, F., Firuzi, O., 2013. Antioxidant properties of hydroxycinnamic acids: a review of structure-activity relationships. Curr. Med. Chem. 20, 4436–4450. Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., Iceevans, C., 1999. Antioxidant activity applying an improved ABTS radical cation decolorization assay. Free Radic. Biol. Med. 26, 1231–1237. Rezaiea, M., Farhoosh, R., Iranshahi, M., Sharif, A., Golmohamadzadeh, S., 2015. Ultrasonic-assisted extraction of antioxidative compounds from Bene (Pistacia atlantica subsp. mutica) hull using various solvents of different physicochemical properties. Food Chem. 173, 577–583. Rodriguez-Amaya, D.B., Kimura, M., 2004. General procedure for carotenoid analysis. Handbook for Carotenoid Analysis. HavervestPlus, Washington, NW, USA. Singleton, V.L., Rossi, J.A., 1965. Colorimetry of total phenolics with phosphomolybdicphosphotungstic acid reagents. Am. J. Enol. Vitic. 16, 144–158. Souza, V., Pereira, P.A.P., Lima, T.L.T., Pio, R., Queiroz, F., 2014. Determination of the bioactive compounds, antioxidant activity and chemical composition of Brazilian blackberry red raspberry, strawberry, blueberry and sweet cherry fruits. Food Chem. 156, 362–368. Vasco, C., Ruales, J., Kamal-Eldin, A., 2008. Total phenolic compounds and antioxidant capacities of major fruits from Ecuador. Food Chem. 111, 816–823. Wang, J., Hu, Y., Wang, D., Liu, J., Zhang, J., Abula, S., Zhao, B., Ruan, S., 2010. Sulfated modification can enhance the immune-enhancing activity of Lycium barbarum polysaccharides. Cell. Immunol. 263, 219–223. Weng-Ping, M., Zhi-Jing, N., He, L., Min, C., 2008. Changes of the main carotenoid pigment contents during the drying processes of the different harvest stage fruits of Lycium barbarum L. Agric. Sci. China 7 (3), 363–369. Yang, Y., Song, X., Sui, X., Qi, B., Wang, Z., Li, Y., Jiang, L., 2016. Rosemary extract can be used as a synthetic antioxidant to improve vegetable oil oxidative stability. Ind. Crops Prod. 80, 141–147.

Acknowledgements The authors would like to thank the following fomenting agencies for financial support: CAPES/PROAP, UFPR, Fundação Araucária and CNPq. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.indcrop.2017.10.052. References Abbeddou, S., Petrakis, C., Pérez-Gálvez, A., Kefalas, P., Hornero-Méndez, D., 2013. Effect of simulated thermo-Degradation on the carotenoids, tocopherols and antioxidant properties of tomato and paprika oleoresins. J. Am. Oil Chem. Soc. 90, 1697–1703. Abdennacer, B., Karim, M., Yassine, M., Nesrine, R., Mouna, D., Mohamed, B., 2015. Determination of phytochemicals and antioxidant activity of methanol extracts obtained from the fruit and leaves of Tunisian Lycium intricatum Boiss. Food Chem. 174, 577–584. Alberti, A., Zielinski, A.A.F., Zardo, D.M., Demiate, I.M., Nogueira, A., Mafra, L.I., 2014. Optimisation of the extraction of phenolic compounds from apples using response surface methodology. Food Chem. 149, 151–158. Amagase, H., Farnsworth, N.R., 2011. A review of botanical characteristics, phytochemistry, clinical relevance in efficacy and safety of Lycium barbarum fruit (Goji). Food Res. Int. 44, 1702–1717. Amagase, H., 2010. Comparison of Lycium barbarum − containing liquid dietary supplements to caffeinated beverages on energy/caloric metabolism activity and salivary adrenocortical hormone levels in healthy human adults. FASEB J. 24–540. Araújo, J.M.A., 2006. Química de alimentos: teoria e prática. UFV, Viçosa. Arroyo-Martinez, Q., Sáenz, M.J., Arias, F.A., Acosta, M.S., 2011. Lycium barbarum: A new hepatotoxic “natural” agent? Digest. Liver Dis. 743–749. Bassani, D.C., Nunes, D.S., Granato, D., 2014. Optimization of phenolics and flavonoids extraction conditions and antioxidant activity of roasted yerba-mate leaves (Ilex paraguariensis A. St.-Hil., Aquifoliaceae) using Response Surface Methodology. An. Acad. Bras. Ciênc. 6, 923–933. Belwal, T., Dhyani, P., Bhatt, I.D., Rawal, R.S., Pande, V., 2016. Optimization extraction conditions for improving phenolic content and antioxidant activity in Berberis asiatica fruits using response surface methodology (RSM). Food Chem. 207, 115–124. Benzie, I.F., Strain, J.J., 1996. The ferric reducing ability of plasma as a measure of ‘antioxidant power’: The FRAP assay. Anal. Biochem. 239, 70–76. Brand-Williams, W., Cuvelier, M.E., Berset, C., 2011. Use of a free radical method to evaluate antioxidant activity. Food Sci. Technol. 28, 25–30. Cacace, J.E., Mazza, G., 2003. Mass transfer process during extraction of phenolic compounds from milled berries. J. Food Eng. 59, 379–389. Comunian, T.A., Boillon, M.R.G., Thomazini, M., Nogueira, M.S., Castro, I.A., FavaroTrindade, C.S., 2016. Protection of echium oil by microencapsulation with phenolic compounds. Food Res. Int. 88, 114–121. Cordeiro, A.M.T.M., Medeiros, M.L., Silva, M.A.A.D., Silva, I.A.A., Soledade, L.E.B., Souza, A.L., Queiroz, N., Souza, A.G., 2013. Rancimat and PDSC accelerated techniques for evaluation of oxidative stability of soybean oil with plant extracts. J. Therm. Anal. Calorim. 114, 827–832. Derringer, G., Suich, R., 1980. Simultaneous optimization of several response variables. J. Qual. Tec. 12, 214–219. Donno, D., Beccaro, G.L., Mellano, M.G., Cerutti, A.K., Bounous, G., 2015. Goji berry fruit (Lycium spp.): antioxidant compound fingerprint and bioactivity evaluation. J. Funct. Foods (Part B), 1070–1085. http://dx.doi.org/10.1016/j.jff.2014.05.020. Fennema, O.R., 1996. Food Chem. Marcel Dekker Inc., New York. Franco, D., Pateiro, M., Amado, I.R., Pedrouso, M.L., Zapata, C., Vazquez, J.A., Lorenzo, J.M., 2016. Antioxidant ability of potato (Solanum tuberosum) peel extracts to inhibit soybean oil oxidation. Eur. J. Lipid Sci. Technol. 118, 1891–1902. Giusti, M.M., Wrolstad, R.E., 2001. Characterization and measurement of anthocyanins by UV-visible spectroscopy. Current Protocols in Food Analytical Chemistry. John Wiley & Sons, Hoboken, NJ, USA. Granato, D., Calado, V.M.A., Jarvis, B., 2014. Observations on the use of statistical methods in Food Science and Technology. Food Res. Int. 55, 137–149. Hogan, S., Chung, H., Zhang, L., Li, J., Lee, Y., Dai, Y., 2010. Antiproliferative and antioxidant properties of anthocyanin-rich extract from açaí. Food Chem. 18, 208–214.

97