Optimization of green PLE method applied for the recovery of antioxidant compounds from buriti (Mauritia flexuosa L.) shell

Optimization of green PLE method applied for the recovery of antioxidant compounds from buriti (Mauritia flexuosa L.) shell

Food Chemistry 298 (2019) 125061 Contents lists available at ScienceDirect Food Chemistry journal homepage: www.elsevier.com/locate/foodchem Optimi...

2MB Sizes 0 Downloads 38 Views

Food Chemistry 298 (2019) 125061

Contents lists available at ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Optimization of green PLE method applied for the recovery of antioxidant compounds from buriti (Mauritia flexuosa L.) shell

T

Adenilson Renato Rudkea, Simone Mazzuttia, Kátia Suzana Andradea, Luciano Vitalib, ⁎ Sandra Regina Salvador Ferreiraa, a

Laboratory of Thermodynamic and Supercritical Technologies (LATESC), Chemical and Food Engineering Department, Federal University of Santa Catarina, EQA/UFSC, C.P. 476, CEP 88040-900 Florianópolis, SC, Brazil b Department of Chemistry, Federal University of Santa Catarina, Florianópolis, SC, Brazil

ARTICLE INFO

ABSTRACT

Keywords: Antioxidants recovery High pressure extraction Optimization Response surface methodology

Buriti fruit, with high content in carotenoids and antioxidant compounds, is well appreciated for its organoleptic characteristics. However, its shell, an agroindustrial residue, is mostly discarded. Therefore, to verify the technological potential of the buriti shells, the aim of this this study was to evaluated the antioxidant potential of the extracts from buriti shell obtained by pressurized liquid extraction (PLE) with ethanol/water mixtures. PLE optimization was performed by response surface methodology, with all results maximized at the conditions of 71.21 °C and with 91.58% of ethanol. The yields values varied from 16.82 to 25.16%, total carotenoids from 23.38 to 1056.59 μg β-carotene equivalent g−1, total phenolic content from 143.37 to 172.02 mg Gallic acid equivalent g−1, DPPH from 31.04 to 48.62 μg.mL−1, and ABTS from 1.87 to 2.70 mmol TEAC. g−1. Therefore, considering the lack of studies about buriti shell, the present work provides valuable results that confirm the PLE relevance to enhance the value of this neglected material.

1. Introduction Considered the most abundant palm tree of Brazil, the “Mauritia flexuosa”, popularly known as buriti, belongs to the Arecaceae family (Koolen, Silva, da Silva, da Paz, & Bataglion, 2018). Buriti tree is present in at least 7 Latin American countries and the Amazonia natives called it “Tree of Life” because all its parts have many purposes (Pezoti Junior et al., 2014). This specie grows in flooded areas known as “buritizais” and is food source for several species of animals (Koolen, Silva, Gozzo, De Souza, & De Souza, 2013). The fruit is considered a functional food because it is an important source of carotenoids, rich in vitamins, antioxidants, unsaturated oils and dietary fibers (Cândido, Silva, & Agostini-Costa, 2015). Several biological benefits for humans (Bataglion, da Silva, Eberlin, & Koolen, 2014), like potent vermifuge and burn healer (Koolen et al., 2018), effective in treatment and prevention of xerophthalmia (Mariath, Lima, & Santos, 1989), and antimicrobial activity (Koolen et al., 2013) are related to buriti. Its pulp, rich in β-carotene, tocopherols, and oleic acid, is resistant to oxidation and comparable to olives (Vásquez-Ocmín, Alvarado, Solís, Torres, & Mancini-Filho, 2010). The above cited characteristics make buriti an important matrix that can be applied in cosmetic and food industries, mainly due to the pulp ⁎

oil (Koolen et al., 2018). However, the pulp only corresponds to 18–30% of the total fruit weight (Barbosa, Lima, & Mourão Junior, 2010), providing 70–82% of poorly used fruit parts, or residue. The shell, for example, is used as animal feed or, not often, as raw material for oil extraction. However, the buriti shell is normally treated as waste, with no proper disposal (Barbosa et al., 2010). About 2500 tons/year of buriti shells are generated by the industrial extraction of buriti oil (Resende, Franca, & Oliveira, 2019). From this residue, the buriti shell, there is a lack of information related to its use and, to the best of our knowledge, only few studies are available in the literature. Pezoti Junior et al. (2014) evaluated the production of activated carbon while Resende et al. (2019) used buriti shells for flour production. França, Reber, Meireles, Machado, and Brunner (1999) described the supercritical CO2 extraction from buriti pulp and shells, Tauchen et al. (2016) evaluated the Soxhlet extraction of different medicinal plants from the Peruvian Amazon, including the buriti shell, and PereiraFreire et al. (2018) studied the chemopreventive action of buriti parts (pulp, shell and endocarp). The “green extraction” concept emerged from low energy consumption processes, which allow the use of alternative solvents and natural renewable products, providing high quality and safe products (Vazquez-Roig & Picó, 2015).

Corresponding author. E-mail address: [email protected] (S.R.S. Ferreira).

https://doi.org/10.1016/j.foodchem.2019.125061 Received 11 March 2019; Received in revised form 31 May 2019; Accepted 22 June 2019 Available online 22 June 2019 0308-8146/ © 2019 Elsevier Ltd. All rights reserved.

Food Chemistry 298 (2019) 125061

A.R. Rudke, et al.

Pressurized liquid extraction (PLE) is based on the use of a pressurized solvent, usually at high temperatures, and is considered a “greener process” compared to conventional methods due to the lower amounts of solvent used (Mustafa & Turner, 2011). The PLE pressure is high enough to keep the solvent in the liquid state and, above this pressure, this variable has no influence on extraction performance (Herrero, Castro-Puyana, Mendiola, & Ibañez, 2013). Several process variables, such as temperature, pressure, extraction time, solvent type and volume, solid and solute characteristics, among others, affect the PLE performance, where temperature and solvent type are the most relevant ones (Carabias-Martínez, Rodríguez-Gonzalo, Revilla-Ruiz, & Hernández-Méndez, 2005). The solvent choice is based on the affinity between the solvent and the solute to be recovered from the raw material (Mustafa & Turner, 2011), but in most cases, the use of “GRAS” (Generally Recognized as Safe) solvents, such as water, ethanol and its mixtures, enhances the “green” aspect of the process (Herrero et al., 2013). In addition, it has been shown that mixtures such as methanol-water and ethanol-water are environmentally friendly and increase the extraction efficiency if compared with pure alcohol or propanol–water mixtures (Mustafa & Turner, 2011). Therefore, the aim of this study was to obtain PLE extracts from buriti shell and evaluate its technological potential. In this context, the response surface methodology (RSM) was used to verify the influence of temperature (T) and solvent (ethanol-water mixtures – E), on extraction yield. The quality of the extracts was evaluated in terms of total carotenoid content (TCC), total phenolic compounds (TPC), antioxidant activity by DPPH and ABTS methods, and analysis of phenolic compounds by LC-ESI-MS/MS. In addition, PLE results were compared with data from Soxhlet, a well stablished method, for buriti shell extraction.

pressure of 10 MPa, based on literature (Herrero et al., 2013). The Central Composite Design (CCD) was used for the optimization of the extraction conditions of temperature and solvent type (ethanol-water mixtures), as presented in Section 2.8. Based on literature data and on equipment limitations, temperature (T) and ethanol-water mixture (E) were set from lower to upper limits, with T of 50° C and E of 50% as central points. The CCD analysis defined variable limits of: (T) ranged from 28.79 °C to 71.21 °C, and (E) from 7.57 to 92.43 of ethanol concentration in water mixture. For practical purpose, the upper and the lower limits were set at 28 and 71 °C for temperature and at 7 and 92% for ethanol concentration, respectively. The recovered extracts were stored at −18 °C and absence of light for further solvent removal by rotary evaporator (Fisatom, model 801, São Paulo, Brazil). Results are expressed in global extraction yield (X0), calculated as the ratio between the mass dried extract and the mass of raw material. 2.2.2. Soxhlet extraction The Soxhlet extraction was performed according to 930.39C method of AOAC (AOAC, 2012). Briefly, 150 mL of ethanol recycling over 5 g of dried powder raw material, in a Soxhlet apparatus for 6 h extraction conducted at the solvent boiling temperature. The recovered extracts were stored at −18 °C and absence of light for further solvent removal by rotary evaporator (Fisatom, model 801, São Paulo, Brazil). Results, expressed in extraction yield (X0), represent the mean values ± standard deviation from triplicate experiments. 2.3. Total carotenoids content (TCC) The method is based on Kuhnen et al. (2009) with modifications. An amount of extract was diluted in 2:1:1 (v/v/v) mixture of methanol, hexane and acetone. The solution was placed in ultrasonic bath (Ecosonics-ultrasonique) for 2 min and centrifuged for 10 min (QuimisQ222T2) (3400 rpm). The supernatant was collected and the absorbance read at 450 nm in a spectrophotometer (FEMTO, 800XI). The total carotenoid content (TCC) of the extracts was based on a standard curve of β-carotene (≥97% UV, Sigma Aldrich, USA), prepared with 10 mg of β-carotene diluted in 50 mL 2:1:1 (v/v/v) mixture of methanol, hexane, and acetone to form concentrations from 0.3125 to 6 μg.mL−1. Then, 1 mL was collected and the absorbance was read at 450 nm. Results expressed as μg of β-carotene equivalent (βCE) per g of extract from triplicate measurements.

2. Materials and methods 2.1. Source material and sample preparation The fruits of buriti “Mauritia flexuosa” were collected in January 2018 in Jaru-Rondônia, North region, under Amazon biome (10° 22’ 14.0” S 62° 34’ 18.0” W). The collected fruits were transported to LATESC/UFSC (Laboratory of Thermodynamic and Supercritical Technologies from Federal University of Santa Catarina). Upon arrival, the fruits were selected, discarding the imperfect ones, and then washed, sanitized, manually pulped, with the parts (shell, pulp, seed, and endocarp) separated as shown in Supplementary material Fig. S1. The buriti shells (Fig. S1-B) were dried in an air circulation oven at 30 °C for 48 h, reduced to powder in a knife mill and stored in polyethylene packages at −18 °C until experiments were carried out.

2.4. Total phenolic content (TPC) The TPC of extracts from buriti shell were determined by FolinCiocalteu method, according to Koşar, Dorman, and Hiltunen (2005) with modifications. Firstly, the extract was diluted in ethanol to obtain a solution concentration of 5 mg.mL−1. Then, 10 μL of extract solution and 600 μL of distilled water were mixed to 50 μL Folin–Ciocalteu reagent (Sigma-Aldrich, USA). After 1 min, 150 μL of Na2CO3 20% (w/v) were added and the volume completed up to 1 mL with distilled water. The samples were incubated for 2 h at room temperature in the dark. The absorbance was measured at 760 nm. A standard curve was prepared with gallic acid solutions at different concentrations (from 0.03125 to 2 mg.mL−1). TPC results were expressed as mg of Gallic acid equivalent (GAE) per g of extract from triplicate measurements.

2.2. Extraction procedures 2.2.1. Pressurized liquid extraction A PLE kinetics was performed to determine the extraction time for the yield assays. The PLE apparatus and the procedure were described by Andrade, Trivellin, and Ferreira (2017). Briefly, the extraction bed was formed by 3 g of dried powder buriti shells, 33 g of glass beads and cotton. The solvent flow was set at 3 mL.min−1 by an HPLC pump (Waters, model 515, USA, volume flow rate from 0.001 to 10 mL.min−1), reaching pressure of 10 MPa (100 bar). The operation temperature was controlled with the aid of a coil inside a heating bath (Microquímica, model MQBTZ 99–20, Palhoça, SC, Brazil, 0.1 °C accuracy and 0.01 °C stability). The extraction kinetics was performed at 50 °C and ethanol as solvent at 3 mL.min−1, based on Andrade et al. (2017). At the extraction conditions, the samples were collected in test tubes every 3 min, until 30 min of extraction, and after that, every 5 min until 90 min of extraction. Based on the kinetics curve, the process time for the PLE yield assays was defined. The yield values were obtained by PLE in dynamic mode, which consists of the solvent flow in the extraction cell (Vazquez-Roig & Picó, 2015), performed at constants

2.5. DPPH free radical scavenging assay The buriti shell extracts obtained from PLE were evaluated using 1,1-diphenyl-2-picrylhydrazyl (DPPH) method, according to Mensor et al. (2001) with modifications. Sample stock solutions (1.25 mg.mL−1) were diluted to final concentrations (75, 50, 37.5, 25, 18.75 and 12.5 μg.mL−1) in ethanol-water mixture (70:30 v/v). Then, 290 μL of a 0.3 mM DPPH ethanolic solution was added to 710 μL of each sample solution at different concentrations to complete the final 2

Food Chemistry 298 (2019) 125061

A.R. Rudke, et al.

reaction medium (1 mL). After 30 min in the absence of light and at room temperature, the absorbance values were measured at 517 nm. The tests were performed in triplicate and accompanied by a control (710 μL de ethanol and 290 μL of DPPH) and the blank for each sample (710 μL of sample and 290 μL of ethanol 70% (v/v). Results were expressed in effective concentration at 50% (EC50). The EC50 values were calculated by linear regression of the AA curves obtained for all extract concentrations. The AA was calculated according to Eq. (1).

AA (%) = 100

(Abssample

chromatographic separation conditions and the mass spectrometer parameters for each phenolic compound were the same as describe by Schulz et al. (2015). System control and data analysis were performed using software Analyst (1.5.1). 2.8. Experimental design and statistical analysis A statistical experimental design was based on Central Composite Design (CCD) with the five dependent variables: X0, TCC, TPC, DPPH and ABTS, respectively coded as Y1, Y2, Y3, Y4, and Y5. These responses were measured according to the following variables: Temperature (T) (°C) and ethanol content (E) (%), coded as X1 and X2, respectively. These process variables were investigated at five levels (−1.41(−α), −1, 0, 1, +1.41(+α)), providing 11 combinations of process conditions, with three replicas at the central point of the experiment. The influence of (T) and (E) on the responses was evaluated using the following quadratic linear model (Eq. (2)).

Absblank ) × 100 Abscontrol

(1)

where: % AA (%): antioxidant activity; Abssample: sample absorbance; Absblank: blank absorbance and; Abscontrol: control absorbance. 2.6. ABTS assay The antioxidant capacity of buriti shell extracts by ABTS method was conducted according to Re et al. (1999) with some modifications. ABTS+ 2,2′-Azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt radical cation was produced by reacting 7 mM ABTS solution and 2.45 mM potassium persulfate solution in the dark at room temperature for 16 h. The aqueous ABTS+ solution was diluted with 5 mM sodium phosphate buffer at pH 7.4 till an absorbance of 0.7 ( ± 0.05) at 734 nm. Briefly, stock solutions (1.25 mg.mL−1) were diluted to final concentrations (750, 500, 375, 250, 187.5 and 125 μg.mL−1) in ethanol-water mixture (70:30 v/v). Then, 30 μL of each was mixed with 970 μL of ABTS%+, incubated in darkness for 45 min for absorbance reading at 734 nm. Trolox was used as a standard reference and the values were calculated from the standard curve (from 0.25 to 2 mM). Results were expressed as mmol of equivalent of trolox per g of dry extract (TEAC) from triplicate measurements.

Y=

0

+

1 X1

+

2 X2

++

12 X1 X2

+

2 11 X1

+

22 X2

2

(2)

where Y is dependent variable; β0 is the intercept; β1 and β2 are the linear coefficients; β11 and β22 are the quadratic coefficients; β12 is the interaction coefficient, X1 and X2 are the independent variables. The model quality and statistical significance of the coefficients were evaluated, and response surfaces and Pareto diagrams were drawn to evaluate the effect of temperature and ethanol content on the analyzed responses (Y1, Y2, Y3, Y4 and Y5). For this, the Statistica (StatSoft, Inc.; www.statsoft.com) v.13.0 program (USA) was used. The individual and global optimum conditions (parameter values that provide the individual optimum, for one response, or the global optimum, for combined responses) were calculated using the method proposed by Derringer and Suich (1980) through the function Desirability implemented in Statistica.

2.7. Determination of phenolic compounds using LC-ESI-MS/MS 2.7.1. Sample preparation of buriti shell extract The samples from buriti shell extracts were prepared according to procedure describe by Schulz et al. (2015), with modifications. Defatted buriti shell extracts (1.0 mL) were mixed with HCl 6 mol L−1 (5 mL), methanol (5 mL) and the system was maintained at 85 °C for 30 min. After hydrolysis reaction the pH was adjusted to 2 with NaOH 6 mol L−1. Then, samples were partitioned with diethyl ether (10 mL), shaken and centrifuged at 4000 rpm for 10 min, performed three times for each sample. The supernatants were rota-evaporated at 40 °C for solvent removal. The dried sample was resuspended with methanol (1 mL) and diluted ten times using methanol:water (30:70, v/v) before injection in the LC-ESI-MS/MS.

3. Results and discussion 3.1. Kinetic extraction assay of buriti shell in PLE method The PLE method was performed considering the extraction time defined according to the kinetic assay (Section 2.2.1), which provides the extraction curve as represented in Fig. 1 for an assay conducted at 50 °C, 10 MPa and 3 mL.min−1 of ethanol as solvent. In this figure it is shown the mass of extracted material with time and the colour of the recovered extracts for each time interval. Different curve periods are shown in Fig. 1, following description by Ferreira, Nikolov, Doraiswamy, Meireles, and Petenate (1999). The curve starts with the constant extraction rate period (CER), followed by the Falling Extraction Rate (FER) and the Diffusional period (DIF). According to Brunner (1994), CER and FER periods normally provide more than 70% recovery of extract. At the beginning of the extraction, controlled by convection, the solute from the particle surface is easily extracted up to 12 min, representing the CER period (rapid desorption), with samples of intense yellow color (Fig. 1-part B). Then, at FER, starts the recovery of the compounds within the particles, combining convection and diffusion. From this stage the extract color changes significantly from the intense yellow to a lighter yellow (from 12 to 40 min). The DIF period, after 40 min, states the diffusion as the dominant mass transfer mechanism. It is possible to observe from Fig. 1 a reduction in the mass of extract with progressing the time (after 40 min), which is accompanied by the discoloration of extract samples (Fig. 1-part B). During the DIF period, the extraction time has no significant influence on process yield. Also, at 40 min of extraction, more than 70% of the solute was extracted, as shown in Fig. 1. Therefore, considering these aspects, the PLE assays for the yield determination (following the CCD plan) were performed with 40 min of extraction time.

2.7.2. LC-ESI-MS/MS The identification and quantification of the phenolic compounds from the samples of buriti shell extract were performed using a highperformance liquid chromatography system (1200 Series, Agilent Technologies, Waldbronn-BW, Germany), according to methodology described by Schulz et al. (2015) and performed by Lima, Ferreira, Vitali, & Block, 2019). A Synergi column (4.0 μm, 2.0 × 150 mm d.i.; Phenomenex, Torrance-CA, USA) was employed for the separations under gradient elution with mobile phase composed of methanol:water (95:5, v/v) and aqueous solution of formic acid 0.1% (v/v). The LC system coupled to a mass spectrometry system composed by a hybrid triple quadrupole/linear ion trap mass spectrometer (Q Trap 3200 Applied Biosystems/MDS Sciex, Concord-ON, Canada) was used in the analysis. The mass spectrometer was operated in negative electrospray ionization mode (TurboIonSpray Applied Biosystems/MDS Sciex, Concord-ON, Canada) and MS/MS parameters were capillary needle maintained at − 4500 V; curtain gas at 10 psi; the temperature at 400 °C; gas 1 and gas 2 at 45 psi; CAD gas, medium. The 3

Food Chemistry 298 (2019) 125061

A.R. Rudke, et al.

Fig. 1. A) Kinetic assay in PLE of buriti shell at 10 MPa, 50 °C and 3 mL.min−1. B) The decay of the extracted color. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.2. Influence of process variables on PLE yield

and −α, enabling the evaluation of the quadratic effect of the process parameters. The results from the adjustment are listed in Table 2, through R2 and R2-adjusted values, which indicate the better performance of the second-order model. The best fit for yield (Y1) as a function of process variables (X1 and X2) is described by Eq. (3), which provided R2 and R2-adjusted values of 0.911 and 0.821, respectively.

The PLE yields for buriti shell ranged from 16.82 to 25.16%, with results presented in Table 1, and compared to Soxhlet with ethanol. The data from Table 1 also show the results from the extract quality in terms of TCC, TPC and antioxidant potential, and compared with the BHT performance. The extract quality aspects are discussed in Sections 3.3, 3.4, 3.5. The influence of process variables temperature (T) and ethanol content (E) on PLE yield were evaluated by applying Eq. (2) and considering two approaches: (a) a first order model, applied to assays 1–7 and neglecting the quadratic parameters β11 and β22; and (b) a second order model, considering coefficients β0, β1, β2, β11, β22 and β12, performed for assays 1–11. The second order model included the levels +α

Y1 = 16.370 + 0.105X1 + 0.070X2 + 7.7. 10 4X1 X2

2.5. 10 4X12

0.002 (3)

X2 2

The generated response surfaces for all data are presented in Fig. 2. The surface related to extraction yield is presented Fig. 2a, while the other results (Fig. 2b–f) are discussed in Sections 3.3, 3.4 and 3.5. Fig. 2a was generated by the second order model and shows the higher yield values for the highest temperature level and the ethanol (in water

Table 1 Central composite design (CCD) with real and codified variables and influence the temperature and solvent (ethanol-water mixtures) in the yield of extraction, Total Carotenoids Content (TCC), Total Phenolic Content (TPC), antioxidant activity (DPPH and ABTS), in the extracts obtained by PLE. Assay

Codified variables T (°C)

E (%)

Real variable T (°C)

E (%)

Responses Y1 Yield (%)

1 2 3 4 5 6 7 8 9 10 11 Sox-EtOH BHT

−1 1 −1 1 0 0 0 −α α 0 0 – –

−1 −1 1 1 0 0 0 0 0 −α α – –

35 65 35 65 50 50 50 28 71 50 50 – –

20 20 80 80 50 50 50 50 50 7.57 92.43 – –

20.61 23.24 16.82 20.83 23.76 20.65 22.39 19.85 25.16 22.27 17.00 21.20 (1.58) –

Y2 TCC μg βCE.g

Y3 −1

29.52 23.38 678.77 615.90 93.60 73.56 69.27 60.12 128.21 104.56 1056.59 1132.15 (23.93) –

The values between parentheses refers to standard deviation. # Effective concentration at 50%. 4

Y4 −1

TPC mg GAE.g

163.88 169.26 164.73 170.05 151.68 147.88 145.05 162.32 172.02 143.37 154.34 162.36 (9.55) –

#

Y5 −1

DPPH EC50 (μg.mL 44.60 31.09 48.62 38.14 44.93 41.42 45.05 44.05 47.01 42.29 43.12 42.36 (0.22) 305.2

)

ABTS mmol TE.g−1 2.51 2.66 2.69 2.70 2.01 2.33 2.03 2.44 2.11 1.87 2.25 3.51 (0.30) 7.70

Food Chemistry 298 (2019) 125061

A.R. Rudke, et al.

a)

b)

d)

c)

e)

Fig. 2. Response surface plots for the studied responses. a) Yield of extraction; b) Carotenoids; c) TPC; d) DPPH expressed in EC50; e) ABTS.

solution) content between 30 and 40%. The Pareto diagrams are useful to evaluate the influence of process variable on extraction yield (as presented in Supplementary material: Fig. S2). The Pareto diagram related to extraction yield (Fig. S2a) indicate that temperature has a positive effect, i.e., the increase in temperature (T) implies in higher PLE yields, within the temperature range applied. Otherwise, the ethanol content (E) has a negative effect, i.e., the increase in ethanol content (from ethanol:water mixtures) reduces PLE yield (other Pareto results are discussed in Sections 3.3, 3.4 and 3.5 – Fig. S2b–e). As observed from Table 1, at constant T, the yield decreases with increasing ethanol concentration in the solvent mixture. These results are in agreement with those obtained by Syahariza, Torkamani, Norziah, Mahmood, and Juliano (2017) and by Tripodo, Ibáñez, Cifuentes, Gilbert-López, and Fanali (2018), for studies related to Momordica charantia and Lycium barbarum L., respectively. The use of solvent mixtures increases the number of solubilized compounds, decreasing selectivity, but also enhances the interaction of the target substance with the solvent mixture, since one solvent improves the analyte solubility, the other enhances its desorption. In this way, solvent mixture promotes higher yields (Mustafa & Turner, 2011). Otherwise, at constant ethanol concentration, the yield increases with enhancing temperature, within the variables ranges studied. At higher temperatures the diffusion rate and the analyte (solute) solubility enhances, favoring the yield (Syahariza et al., 2017). In addition, the strong solute-matrix interactions due to van der Waals forces, hydrogen bonding, and dipole attractions are more easily disrupted with increasing temperature, which also affect the reduction of solvent viscosity, enabling higher penetration of solvent in the solid matrix (Tripodo et al., 2018).

3.3. Influence of variables on TCC The results of total carotenoids content (TCC), expressed as μg of βcarotene equivalent (βCE) per g of extract, is also shown in (Table 1) for all extractions performed. The data, correlated with a standard β-carotene curve (R2 = 0.997), provided TCC values from 23.38 to 1056.59 μg βCE.g−1. The second order model (R2 = 0.996) was used to describe the relation between TCC and the independent variables (T and E). The model coefficients are described in Table 2. The generated response surface for TCC values as a function of process variables is shown in Fig. 2b, which clearly demonstrates the high influence of ethanol concentration in TCC behavior. Otherwise, the temperature effect was negligible, with a slight negative slope of TCC with temperature increase. Besides, the Pareto chart for TCC (Fig. S2b) detected no significant influence of T (p-value < 0.05), while the ethanol concentration and its quadratic influence affected positively the TCC results. The effect of ethanol concentration on TCC is associated to the solubility of carotenoid compounds in the solvent mixture. These molecules are mostly nonpolar and poorly soluble in water (Butnariu, 2016). Therefore, the increase in ethanol concentration in solvent mixtures enhances the carotenoids solubility. At 20 bar and ethanol concentration ranging from 0 to 10% (v/v), the solubility of β-carotene, the main buriti compound (Cândido et al., 2015), increased from 93.254 to 200.32 ppm, at 70 °C, and from 10.96 to 107.25 ppm at 100 °C (Mottahedin, Asl, & Lotfollahi, 2017). Mustafa, Trevino, and Turner (2012) performed PLE for carrot with a temperature range from 60 to 180 °C, different extraction cycles (1–5) and process time (2–10 min). The authors verified that the recovered amount of β-carotene and α-carotene was higher at 60 °C, indicating the carotenoids degradation with enhancing temperature. Also, Calvo, 5

Food Chemistry 298 (2019) 125061

A.R. Rudke, et al.

Table 2 Regression coefficients from first and second order model expressed in real variable terms. Regression coefficients β0

β1

β2

Yield (%) (Y1) TCC (μg βCE.g−1) (Y2) TPC (mg GAE.g−1) (Y3) DPPH (μg.mL−1) (Y4) ABTS (mmol TE.g−1) (Y5)

20.155 −312.406 149.230 47.119* 2.000

0.072 0.426 0.180 0.100 0.007

−0.090 11.924 0.016 −0.284 0.006

Yield (%) (Y1) TCC (μg βCE.g−1) (Y2) TPC (mg GAE.g−1) (Y3) DPPH (μg.mL−1) (Y4) ABTS (mmol TE.g−1) (Y5)

16.370* 230.055 271.777* 35.038 4.295

0.105 −1.561 −5.006* 0.542 −0.085

0.070 −15.500* −0.212 −0.123 −0.002

β11 First order model – – – – – Second order model −2.5 × 10−4 0.034 0.052* −0.006 8.7 × 10−4

β22

β12

R2

R2adj

Lack of fit

– – – – –

7.7 × 10−4 −0.032 −0.000 0.002 −8.0 × 10 −5

0.658 0.772 0.045 0.839 0.043

0.315 0.544 0.000 0.677 0.000

0.252 0.001 0.018 0.209 0.063

−0.002* 0.279* 0.003 −6.7 × 10−4 9.6 × 10−5

7.7 × 10−4 −0.032 −0.000 0.002 −8.0 × 10 −5

0.911 0.996 0.811 0.429 0.350

0.821 0.992 0.622 0.000 0.000

0.906 0.097 0.155 0.143 0.163

TCC – Total Carotenoids Content. TPC – Total Phenolic Content. * Significative.

Dado, and Santa-María (2007) have demonstrated that a temperature variation from 25 °C to 60 °C is sufficient to degrade carotenoids, such as β-carotene, phytofluene and phytoene, by up to 1/3. Therefore, the low influence of temperature on TCC could be related to the solubility of carotenoids and its degradability, in relation to solvent and temperature conditions. In general, the increase in temperature enhances the solubility of compounds, however, above certain point, the increase in temperature no longer affects the solubility, starting components degradation (Mottahedin et al., 2017). The highest TCC value from PLE method was provided by sample 11, obtained at 50 °C and 92.43% of ethanol concentration (Table 1: 1056.59 μg βCE g−1). This value was higher than reported by PereiraFreire et al. (2018) for buriti peel (or shell), which detected the value of 883 μg βCE g−1 for extract recovered by maceration with methanol as solvent. This difference was probably due to the source of the raw material, which also affects TCC. Carotenoids are secondary metabolites and depend on the plant interaction with the environment. These metabolites make the plants adapt to their habitats, protecting them against physical and climatic factors and against predators and pathogens (Valli, Russo, & Bolzani, 2018). Regions with high sunlight incidence such as the Amazon region (sample from the present work) provide fruits with high carotenoid content, as demonstrated by Cândido et al. (2015), which detected TCC of 311.3 and 528.6 μg.g−1 of the buriti pulp from Cerrado and Amazon biome, respectively.

by Tripodo et al. (2018) also showed that the maximum TPC was obtained at the highest temperature and ethanol content. The molecular dispersion reduces viscosity and surface tension, aiding the solute-matrix disruption and increasing the mass transfer rates (Ciulu et al., 2017; Mustafa & Turner, 2011). The TPC results from this work (from 143.37 to 172.02 mg GAE. g extract) were lower than obtained by Tauchen et al. (2016) for buriti shell extract recovered by Soxhlet with ethanol 70% (461.5 mg GAE. g extract). Besides, Cândido et al., 2015; Koolen et al., 2013; Schiassi, De Souza, Lago, Campos, and Queiroz (2018) found TPC values of 4,35 mg GAE. g−1, 3,78 mg GAE. g−1 and 1,1 mg GAE. g−1 for buriti pulp extracts, which were inferior to the data from the present work. This behavior is attributed to difference in extraction method and raw material. 3.5. Influence of variables on antioxidant activity In order to evaluate the functional properties of the buriti shell extracts, the antioxidant activity was detected by the common methods DPPH and ABTS. The DPPH results are expressed in EC50 (μg.mL−1), while ABTS data are expressed in mmol Trolox Equivalent per g extract. The first order model was used to describe DPPH (R2 = 0.839) and the second order model was used to represent ABTS (R2 = 0.350). The generated response surfaces (Fig. 2d–e) for the two methods show that the maximum values were obtained for low temperature and high ethanol content. The Pareto diagrams (Fig. S2d–e) show a significant positive influence (p > .05) of ethanol concentration on DPPH, although no influence was detected for ABTS analysis. Koolen et al. (2013) obtained EC50 of 19.58 mg.mL−1 (equivalent to 19580 μg.mL−1) for buriti pulp, which is lower than results from the present study, ranging from 31.09 to 47.01 μg.mL−1 (Table 1). This comparison shows the importance of buriti shell to provide antioxidant compounds detected by DPPH method, compared to the fruit pulp. The ABTS data from this work varied from 1.87 to 2.70 mmol TE per g of extract (or from 283.45 to 382.90 μmol TE per g of buriti shell), values higher than obtained from buriti pulp by Cândido et al. (2015) (46.63 μmol TE per g of buriti pulp) and by Schiassi et al. (2018) (6.03 μmol TE per g of buriti pulp). Again, these results confirm the importance of buriti shell compared to buriti pulp in terms of antioxidant potential detected by ABTS method. The differences compared to literature data may be due to the extraction technique and the raw materials source, affected by harvesting and climate conditions, interfering on plant characteristics (Cândido et al., 2015; Schiassi et al., 2018). And, because there are very few results in the literature about buriti shell, the data from the present work was compared with the buriti pulp. Moreover, this comparison

3.4. Influence of variables on TPC Total phenolic content is expressed in mg of Gallic acid equivalent (GAE) per g extract and ranged from 143.37 to 172.02 (Table 1). The results obtained are correlated with a standard Gallic acid curve (R2 = 0.980). The coefficients of the second order model are listed in Table 2, reaching R2 = 0.811, for the relation between TPC and the independent variables (T and E). The response surface generated for TPC as a function of process variables is shown in Fig. 2c, where the highest total phenol content was detected at the highest temperature and ethanol concentration. Besides, the Pareto chart for TPC (Fig. S2b) detected significant influence only for the quadratic temperature (p-value < 0.05), while other variables show no significant influence on TPC. The positive effect of temperature on TPC is probably due to the solubility increase with enhancing temperature, due to higher molecular agitation of the solvent. A positive influence of extraction temperature on TPC values was also detected by Ciulu et al. (2017), which conducted PLE from Stevia rebaudiana, and by and Tripodo et al. (2018), which performed PLE from Goji berry (Lycium Barbarum L.). The response surface generated 6

Food Chemistry 298 (2019) 125061

A.R. Rudke, et al.

shows that, in general, buriti shell presents higher antioxidant activity than buriti pulps, since they protect the fruits against external agents such as microorganisms and insects (Resende et al., 2019). Then, the results from this work demonstrated that the PLE is a viable alternative to recover phenolic compounds and carotenoids from buriti shell, showing higher antioxidant activity than buriti pulp.

obtained in this study. The desirability function was used for this purpose. The optimization (individual or global) considered the maximum response values, except for EC50 where minimum values are desired. The global optimum condition, obtained by the models, was 71.21 °C and 91.58% of ethanol (in solvent mixture) and the predicted response values are listed in Table 3. The level of desirability, obtained maximizing all responses, was 0.762. This value was already expected because it represents a combination of five responses (Y1 to Y5), where one individual optimal may differ from global ones. For instance, while TPC and yield require high temperature values to reach the best performances, the TCC requires lower temperatures for its best result. Similarly, Gilbert-López, Barranco, Herrero, Cifuentes, and Ibáñez (2017) studied extracts from Phaeodactylum tricornutum and obtained a desirability value of 0.699, also comparing responses such as extraction yield with TPC, ABTS and TCC values. Therefore, the global optimum condition was tested experimentally in triplicate (for practical purposes the operational conditions were considered as 71 °C and 91% of ethanol) to verify the model prediction in terms of the evaluated responses. The experimental results obtained at optimum conditions are compared to the values predicted by the second order model (Table 3). The relative standard deviation between the predicted and the experimental values were 2.94%, 2.28%, 6.65%, 5.53% and 18.51% for the responses Y1, Y2, Y3, Y4 and Y5, respectively. These data provide a good model representation for most results, except Y5 (ABTS response) which had the lowest R2 value (0.350 from Table 2). Also, the PLE method performed at the optimum conditions (Table 3) was compared to Soxhlet extraction from buriti shell using ethanol as solvent (Table 1). The results for all responses (Y1 to Y5) were very similar for both techniques, with no significant difference at 95% confidence between PLE and Soxhlet responses, except for Y1 and Y2 (yield and total carotenoid content, respectively). Although PLE and Soxhlet are different procedures, with advantages and drawbacks, the comparison is useful to aid the selection of the extraction for specific purposes. It is also worth mention that Soxhlet was conducted during 6 h extraction, with solvent recycling, compared to the 40 min for PLE procedure, which allows more solute-solvent interactions at Soxhlet method compared to PLE, leading to higher yield. Additionally, Soxhlet was more efficient than PLE for the recovery of carotenoid compounds, most probably due to the solvent differences (100% ethanol at Soxhlet and 91% ethanol:water at PLE). It is still worth mention that PLE is comparable to classical extraction methods in terms of product quality, but can also be conducted faster.

3.6. Phenolic profile Because the total phenolic content was highly influenced by the extraction temperature, compared to ethanol content, the phenolics profile was carried out for extract samples recovered at constant ethanol content (50%) and the temperature of 28; 50 and 71 °C. The assays selected for phenolics evaluation were 8, 5 and 9. Among the 46 phenolic compounds tested as standards, a total of 27 were detected from the buriti shell extracts, and the results are presented in Table 4. The parameters for the identification and quantification of phenolic compounds were retention time, parent ion, quantitative ion and limits of identification (LOD) and quantification (LOQ). The compounds were predominantly phenolic acids and flavonoids, although other phenolic classes were also detected. From the phenolic compounds quantified the ones with higher concentration in the extracts of buriti shell were: Isoquercetin > Ferulic acid > Vanilic acid > Protocatechuic acid > Caffeic acid, recovered at 28 °C; Ferulic acid > Protocatechuic acid > Isoquercetin > Vanilic acid > Quercetin and recovered at 50 °C; and Epicatechin > Isoquercetin > Ferulic acid > Catechin > Protocatechuic acid for the extract obtained at 71 °C. It should be noted that seven compounds were identified for the first time in association with buriti, in this case from buriti shell extracts (compounds from Table 4 marked with asterisk). To the best of our knowledge, only two studies from the literature quantified phenolic compounds from buriti samples, i.e., from the shell (Tauchen et al., 2016) and from the pulp (Bataglion et al., 2014). These authors found chlorogenic acid, protocatechuic acid and isoquercetin as the main compounds. From the present work, Table 4 shows that ferrulic acid and isoquercetin were the main compounds detected for all samples, with concentrations above the content of chlorogenic acid. Vanillic acid and protocatechuic acid were also detected in high content for all samples. Therefore, the good antioxidant activity detected by these extracts, obtained by PLE, may be due to the presence of such compounds, which have their antioxidant activity reported in the literature (Cruz-Zúñiga et al., 2016). Besides, the antioxidant activity of the standard Butylated Hydroxytoluene (BHT) a synthetic antioxidant, presented in Table 1, was lower than that obtained from buriti shell extracts recovered by PLE. This behavior is probably due to the main compounds from buriti extract (ferrulic acid, isoquercetin and vanillic acid) and to the presence of caffeic acid and protocatechuic acid. For instance, caffeic and protocatechuic acids have EC50 (by DPPH method) of 0.11 and 0.14 mol/L, respectively, compared to 0.24 mol/L from BHT (Brand-Williams, Cuvelier, & Berset, 1995). Which confirms the best DPPH performance (lower EC50 values) of the buriti extracts compared to BHT. On the other hand, the ABTS results provided better antioxidant performance for BHT compared to buriti extracts. This was probably due to the mixture of components present in the extracts, which increases the system complexity to be detected by the antioxidant reaction method (DPPH or ABTS). Also, Karadag, Ozcelik, and Saner (2009) suggest that, although the methods DPPH and ABTS are based on electron transfer, the reaction rates are different, which affect the antioxidant results. Therefore, because antioxidant components react differently, it is necessary to combine more than one method for the antioxidant analysis.

4. Conclusions The “shell residue” obtained from the buriti processing to produce beverages, jellies, ice cream, desserts, wine and sometimes oil can be a good source of antioxidant compounds such as β-carotene. The extracts recovered from this residue, with desirable bioactivities, can be useful for pharmaceutical, cosmetic and food industries. The processing of this agroindustrial residue reduces the environmental impact generated by the disposal of this waste and provides an additional income for the buriti processing. Also, PLE method combined with “green” solvents contribute to low environmental impact, within the conditions used. Except for yield, all responses were maximized with increasing ethanol content in the solvent mixture. On the other hand, the highest PLE temperature maximized yield and TPC but had a negative influence on antioxidant activity and on TCC. Therefore, a global optimum response was obtained at 71.21 C and ethanol content of 91.58%. A second order model was used and provided a quite promising method to predict the responses studied, with good approximation to experimental data, except for the ABTS values. Furthermore, the results from the present work demonstrated the importance of the use of a neglected raw material, the buriti shell (considered as a residue), to recover valuable bioactive compounds by a promising high pressure method.

3.7. The optimum conditions Table 3 shows the individual and global optimum conditions 7

Food Chemistry 298 (2019) 125061

A.R. Rudke, et al.

Table 3 Individual and global optimal responses (yield, TCC, TPC, DPPH and ABTS) for PLE.

Individual optimal Yield (%) (Y1) TCC (μg βCE.g−1) (Y2) TPC (mg GAE.g−1) (Y3) DPPH (μg.mL−1) (Y4) ABTS (mmol TE.g−1) (Y5) Global optimal Yield (%) (Y1) TCC (μg βCE.g−1) (Y2) TPC (mg GAE.g−1) (Y3) DPPH (μg.mL−1) (Y4) ABTS (mmol TE.g−1) (Y5)

Temperature (°C)

Ethanol (%)

Predicted Response

Experimental Response

71.21 37.70 71.21 28.79 35.15

38.12 92.43 92.43 92.43 92.43

24.91 1056.65 184.10 28.21 2.71

– – – – –

71.21

91.58

20.11 1010.50 183.84 39.21 2.69

19.29 ± 0.45 978.42 ± 52.45 167.33 ± 6.38 42.28 ± 0.42 3.50 ± 0.12

TCC – Total Carotenoids Content. TPC – Total Phenolic Content.

and CAPES, Brazilian founding agencies, for the financial support and fellowship. We also thank Baratela family from Jaru, RO, Brazil, for kindly provide the buriti fruits.

Table 4 Phenolic profile of buriti shell extract (μg.g−1 of extract). Phenolic compounds

ASSAY 8

5

9

T (°C)

28 °C

50.00 °C

71 °C

E (%)

50.00

50.00

50.00

158.88 52.77 < LOQ 1352.19 34.20 367.97 < LOQ 3.21 101.24 469.40 4.46

112.14 11.60 < LOQ 679.52 6.69 403.48 nd 3.71 120.56 376.22 < LOQ

79.74 12.42 < LOQ 403.81 < LOQ 302.02 nd 2.00 116.85 283.95 < LOQ

2.00 6.78 14.15 < LOQ < LOQ 2.74 1560.72 nd 11.53 146.15 < LOQ 5.53 1.99

1.56 4.87 137.89 < LOQ 14.01 2.63 382.62 < LOQ 10.46 139.99 nd 19.75 < LOQ

2.10 nd 355.78 < LOQ 559.20 3.49 462.52 21.63 11.27 125.78 nd 29.54 1.14

< LOQ

3.25

3.74

11.74

6.55

6.43

nd

< LOQ

nd

Phenolic acids 1 Caffeic acidI,II,III 2 Chlorogenic acidI,II,III 3 Ellagic acid* 4 Ferulic acidI,II 5 Gallic acidI 6 Protocatechuic acidII 7 Salicylic acidI 8 Sinapic acidI 9 Syringic acidI 10 Vanillic acidI 11 ρ-coumaric acidI,II Flavonoids 12 ApigeninI,II 13 Aromadendrin* 14 CatechinII,III 15 Chrysin* 16 EpicatechinI,II 17 Hispidulin* 18 IsoquercetinI 19 KaempferolI,II 20 MyricetinI,II,III 21 QuercetinI,II,III 22 RutinI,III 23 Taxifolin* 24 VitexinIII Phenolic Aldeyde 25 Vanillin* Coumarin 26 Umbeliferone* Stilbene 27 ResveratrolI

Appendix A. Supplementary data Supplementary data to this article can be found online at https:// doi.org/10.1016/j.foodchem.2019.125061. References Andrade, K. S., Trivellin, G., & Ferreira, S. R. S. (2017). Piperine-rich extracts obtained by high pressure methods. The Journal of Supercritical Fluids, 128, 370–377. https://doi. org/10.1016/j.supflu.2017.05.001. AOAC (2012). In G. W. Latimer (Ed.). Official methods of analysis of AOAC International. AOAC. Barbosa, R. I., Lima, A. D., & Mourão Junior, M. (2010). Biometria de frutos do buriti (mauritia flexuosa l.f. – arecacea‘e): Produção de polpa e óleo em uma área de savana em Roraima. Amazônia: Ciência & Desenvolvimento, 5(10), 71–86. Bataglion, G. A., da Silva, F. M. A., Eberlin, M. N., & Koolen, H. H. F. (2014). Simultaneous quantification of phenolic compounds in buriti fruit (Mauritia flexuosa L. f.) by ultra-high performance liquid chromatography coupled to tandem mass spectrometry. Food Research International, 66, 396–400. https://doi.org/10.1016/j. foodres.2014.09.035. Brand-Williams, W., Cuvelier, M. E., & Berset, C. (1995). Use of a free radical method to evaluate antioxidant activity. Lebensmittel-Wissenschaft and Technologie, 28, 25–30. Brunner, G. (1994). Gas Extraction: An Introduction to Fundamentals of Supercritical Fluids and the Application to Separation Processes. (1a). Steinkopff, Darmstadt. Butnariu, M. (2016). Methods of analysis (extraction, separation, identification and quantification) of carotenoids from Natural Products. Journal of Ecosystem and Ecography, 6(2), https://doi.org/10.4172/2157-7625.1000193. Calvo, M. M., Dado, D., & Santa-María, G. (2007). Influence of extraction with ethanol or ethyl acetate on the yield of lycopene, β-carotene, phytoene and phytofluene from tomato peel powder. European Food Research and Technology, 224(5), 567–571. https://doi.org/10.1007/s00217-006-0335-8. Cândido, T. L. N., Silva, M. R., & Agostini-Costa, T. S. (2015). Bioactive compounds and antioxidant capacity of buriti (Mauritia flexuosa L. f.) from the Cerrado and Amazon biomes. Food Chemistry, 177, 313–319. https://doi.org/10.1016/j.foodchem.2015. 01.041. Carabias-Martínez, R., Rodríguez-Gonzalo, E., Revilla-Ruiz, P., & Hernández-Méndez, J. (2005). Pressurized liquid extraction in the analysis of food and biological samples. Journal of Chromatography A, 1089(1–2), 1–17. https://doi.org/10.1016/j.chroma. 2005.06.072. Ciulu, M., Quirantes-piné, R., Spano, N., Sanna, G., Borrás-linares, I., & Segura-carretero, A. (2017). Evaluation of new extraction approaches to obtain phenolic compoundrich extracts from Stevia rebaudiana Bertoni leaves. Industrial Crops and Products, 108, 106–112. https://doi.org/10.1016/j.indcrop.2017.06.024. Cruz-Zúñiga, J. M., Soto-Valdez, H., Peralta, E., Mendoza-Wilson, A. M., RoblesBurgueño, M. R., Auras, R., & Gámez-Meza, N. (2016). Development of an antioxidant biomaterial by promoting the deglycosylation of rutin to isoquercetin and quercetin. Food Chemistry, 204, 420–426. https://doi.org/10.1016/j.foodchem.2016. 02.130. Derringer, G., & Suich, R. (1980). Simultaneous optimization of several response variables. Journal of Quality Technology, 12, 214–219. Ferreira, S. R. S., Nikolov, Z. L., Doraiswamy, L. K., Meireles, M. A. A., & Petenate, A. J. (1999). Supercritical fluid extraction of black pepper (Piper nigrun L.) essential oil. The Journal of Supercritical Fluids, 14, 235–245. França, L. F., Reber, G., Meireles, M. A. A., Machado, N. T., & Brunner, G. (1999). Supercritical extraction of carotenoids and lipids from buriti (Mauritia flexuosa), a fruit from the Amazon region. The Journal of Supercritical Fluids, 14, 247–256.

nd Not detected. < LOQ – not quantifiable. * Reported first time. I Reported by (Tauchen et al., 2016) in buriti pulp and shell. II Reported by (Bataglion et al., 2014) in buriti pulp. III Reported by (Koolen et al., 2013) in buriti pulp.

Declaration of Competing Interest None. Acknowledgments The authors wish to thank CNPq (project number 404347/2016-9) 8

Food Chemistry 298 (2019) 125061

A.R. Rudke, et al. Gilbert-López, B., Barranco, A., Herrero, M., Cifuentes, A., & Ibáñez, E. (2017). Development of new green processes for the recovery of bioactives from Phaeodactylum tricornutum. Food Research International, 99, 1056–1065. https://doi. org/10.1016/j.foodres.2016.04.022. Herrero, M., Castro-Puyana, M., Mendiola, J. A., & Ibañez, E. (2013). Compressed fluids for the extraction of bioactive compounds. TrAC – Trends in Analytical Chemistry, 43(2), 67–83. https://doi.org/10.1016/j.trac.2012.12.008. Karadag, A., Ozcelik, B., & Saner, S. (2009). Review of methods to determine antioxidant capacities. September 2008 Food Analytical Methods, 2, 41–60. https://doi.org/10. 1007/s12161-008-9067-7. Koolen, H. H. F., Silva, F. M. A., da Silva, V. S. V., da Paz, W. H. P., & Bataglion, G. A. (2018). Buriti fruit - Mauritia flexuosa. In S. Rodrigues, E. de O. Silva, & E. S. de Brito (Eds.). Exotic fruits (pp. 61–67). Londres: Academic Press. Koolen, H. H. F., Silva, F. M. A., Gozzo, F. C., De Souza, A. Q. L., & De Souza, A. D. L. (2013). Antioxidant, antimicrobial activities and characterization of phenolic compounds from buriti (Mauritia flexuosa L. f.) by UPLC – ESI-MS/MS. Food Research International, 51(2), 467–473. https://doi.org/10.1016/j.foodres.2013.01.039. Koşar, M., Dorman, H. J. D., & Hiltunen, R. (2005). Effect of an acid treatment on the phytochemical and antioxidant characteristics of extracts from selected Lamiaceae species. Food Chemistry, 91(3), 525–533. https://doi.org/10.1016/j.foodchem.2004. 06.029. Kuhnen, S., Lemos, P. M. M., Campestrini, L. H., Ogliari, J. B., Dias, P. F., & Maraschin, M. (2009). Antiangiogenic properties of carotenoids: A potential role of maize as functional food. Journal of Functional Foods, 1(3), 284–290. https://doi.org/10.1016/j.jff. 2009.04.001. Lima, R. S., Ferreira, S. R. S., Vitali, L., & Block, J. M. (2019). May the superfruit red guava and its processing waste be a potential ingredient in functional foods? Food Research International, 115, 451–459. https://doi.org/10.1016/j.foodres.2018.10. 053. Mariath, J. G. R., Lima, M. C. C., & Santos, L. M. P. (1989). Vitamin A activity of buriti (Mauritia vinifera Mart) and its effectiveness in the treatment and prevention of xerophthalmia. American Journal of Clinical Nutrition, 49(5), 849–853. https://doi. org/10.1093/ajcn/49.5.849. Mensor, L. L., Menezes, F. S., Leitão, G. G., Reis, A. S., Santos, T. C., Coube, C. S., & Leitão, S. G. (2001). Screening of Brazilian plant extracts for antioxidant activity by the use of DPPH free radical method. Phytotherapy Research, 15, 127–130. https://doi.org/ 10.1002/ptr.687. Mottahedin, P., Asl, A. H., & Lotfollahi, M. N. (2017). Experimental and modeling investigation on the solubility of β-carotene in pure and ethanol-modified subcritical water. Journal of Molecular Liquids, 237, 257–265. https://doi.org/10.1016/j.molliq. 2017.04.036. Mustafa, A., Trevino, L. M., & Turner, C. (2012). Pressurized hot ethanol extraction of carotenoids from carrot by-products. Molecules, 17(2), 1809–1818. https://doi.org/ 10.3390/molecules17021809. Mustafa, A., & Turner, C. (2011). Pressurized liquid extraction as a green approach in food and herbal plants extraction: A review. Analytica Chimica Acta, 703(1), 8–18. https://doi.org/10.1016/j.aca.2011.07.018. Pereira-Freire, J. A., Oliveira, G. L. D. S., Lima, L. K. F., Ramos, C. L. S., Arcanjo-Medeiros,

S. R., De Lima, A. C. S., ... Ferreira, P. M. P. (2018). In vitro and ex vivo chemopreventive action of Mauritia flexuosa products. Hindawi, 2018. https://doi.org/10. 1155/2018/2051279. Pezoti Junior, O., Cazetta, A. L., Souza, I. P. A. F., Bedin, K. C., Martins, A. C., Silva, T. L., & Almeida, V. C. (2014). Adsorption studies of methylene blue onto ZnCl2-activated carbon produced from buriti shells (Mauritia flexuosa L.). Journal of Industrial and Engineering Chemistry, 20, 4401–4407. https://doi.org/10.1016/j.jiec.2014.02.007. Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., & Rice-Evans, C. (1999). Antioxidant activity appliying an improved ABTS radical cation decolorization assay. Free Radical Biology and Medicine, 26(98), 1231–1237. https://doi.org/10.1016/ S0891-5849(98)00315-3. Resende, L. M., Franca, A. S., & Oliveira, L. S. (2019). Buriti (Mauritia flexuosa L. f.) fruit by-products flours: Evaluation as source of dietary fibers and natural antioxidants. Food Chemistry, 270(July 2018), 53–60. https://doi.org/10.1016/j.foodchem.2018. 07.079. Schiassi, M. C. E. V., De Souza, V. R., Lago, A. M. T., Campos, L. G., & Queiroz, F. (2018). Fruits from the Brazilian Cerrado region: Physico-chemical characterization, bioactive compounds, antioxidant activities, and sensory evaluation. Food Chemistry, 245(October 2017), 305–311. https://doi.org/10.1016/j.foodchem.2017.10.104. Schulz, M., Campelo, S., Valdemiro, L., Katia, S., Seraglio, T., Simon, I., ... Fett, R. (2015). Chemical composition, bioactive compounds and antioxidant capacity of juçara fruit (Euterpe edulis Martius) during ripening. Food Research International, 77, 125–131. https://doi.org/10.1016/j.foodres.2015.08.006. Syahariza, Z. A., Torkamani, A. E., Norziah, H. M., Mahmood, W. A. K., & Juliano, P. (2017). Optimisation of pressurised liquid extraction for antioxidative polyphenolic compound from Momordica charantia using response surface methodology. International Journal of Food Science and Technology, 52(2), 480–493. https://doi.org/ 10.1111/ijfs.13304. Tauchen, J., Bortl, L., Huml, L., Miksatkova, P., Doskocil, I., Marsik, P., ... Kokoska, L. (2016). Phenolic composition, antioxidant and anti-proliferative activities of edible and medicinal plants from the Peruvian Amazon. Brazilian Journal of Pharmacognosy, 26(6), 728–737. https://doi.org/10.1016/j.bjp.2016.03.016. Tripodo, G., Ibáñez, E., Cifuentes, A., Gilbert-López, B., & Fanali, C. (2018). Optimization of pressurized liquid extraction by response surface methodology of Goji berry (Lycium barbarum L.) phenolic bioactive compounds. Electrophoresis, 1–10. https:// doi.org/10.1002/elps.201700448. Valli, M., Russo, H. M., & Bolzani, V. da S. (2018). The potential contribution of the natural products from Brazilian biodiversity to bioeconomy. Anais Da Academia Brasileira de Ciencias, 90(1), 763–778. https://doi.org/10.1590/00013765201820170653. Vásquez-Ocmín, P. G., Alvarado, L. F., Solís, V. S., Torres, R. P., & Mancini-Filho, J. (2010). Chemical characterization and oxidative stability of the oils from three morphotypes of Mauritia flexuosa L.f, from the Peruvian Amazon. Grasas Y Aceites, 61(4), 390–397. https://doi.org/10.3989/gya.010110. Vazquez-Roig, P., & Picó, Y. (2015). Pressurized liquid extraction of organic contaminants in environmental and food samples. TrAC – Trends in Analytical Chemistry, 71, 55–64. https://doi.org/10.1016/j.trac.2015.04.014.

9