Feasibility of ultrasound-assisted optimized process of high purity rice bran protein extraction. Feasibility of ultrasound-assisted optimized process of high purity rice bran protein extraction

: Rice bran is a by-product of the rice milling process, found worldwide in abundance and highlighted due its protein content. This study optimized the conditions for ultrasonic-assisted extraction of protein from defatted rice bran (DRB) and characterized the rice bran protein concentrate (RBPC). A sequential strategy of experimental design was employed; the effect of pH, temperature, ultrasound exposure time, and amplitude were evaluated regarding the percentage of protein extraction by a Full factorial design (FFD) with a fixed frequencies (FFD-A: 37 kHz; FFD-B: 80 kHz). Subsequently, the percentage of protein extracted was optimized employing a Central composite rotatable design (variables: pH and ultrasound exposure time) and RBPC obtained was characterized regarding chemical and functional properties. The pH and ultrasound exposure time had positive effect (P ≤ 0.05) on percentage of protein extraction; moreover, 37 kHz frequency was more effective in protein extraction. The optimized condition (frequency: 37 kHz; temperature: 30 °C; pH: 10; ultrasound exposure time: 30 min; and amplitude: 100%) allowed 15.07% of protein recovery and the RBPC presented 84.76 g 100 g -1 of protein. Magnesium and copper were the main mineral in RBPC (34.4 and 25.5 µg g -1 , respectively), while leucine was the limiting amino acid (0.42) and threonine presented the highest chemical score (1.0). The RBPC solubility was minimal at pH 4 and higher at pHs 6-10; the water and oil absorption capacity were higher than bovine serum albumin (BSA) and the emulsifying capacity was comparable to BSA, with a suitable stability. It was possible to obtain a higher purity RBPC than described in the literature, due to the optimization in the extraction process steps, with functional properties suitable for application in food products, especially emulsified ones.

supplying important nutritional but also functional properties, such as foaming, emulsifying and gelation capacity (PHONGTHAI et al., 2016a;HOU et al., 2017). In this context, it is evident the improvement on studies of rice bran protein functional properties, such as, the application of this protein on different food formulations.
The low application of rice bran protein as food ingredient is associated with their difficult obtainment. Rice bran proteins have different degrees of hydrogen, disulfide bonds that keep the polypeptides strongly bonded and different degrees of hydrophobicity may difficult the protein extraction process; thus, the conventional method in isolating rice bran proteins is alkali extraction followed by precipitation at the isoelectric pH (ROMERO, 2015). In this context, physical, enzymatic, and mechanical treatments may be associated in order to the release rice protein from cells. The ultrasonic waves is a physical technique highlighted as an alternative to reduce the extraction time and efficiency of rice bran protein extraction (CHITTAPALO et al., 2009;PHONGTHAI et al., 2016a).
Currently, the use of ultrasound for solid-liquid extraction have been highlighted due to high reproducibility, shorter time consumption, convenience, reduced consumption of solvent, and lower energy input (PHONGTHAI et al., 2016a). Mechanical effects of ultrasound provide a greater penetration of solvent into cellular materials and improves mass transfer in organic compounds extraction from a tissue (CHEMAT et al., 2011), showing as an alternative in the improvement of rice bran protein extraction.
Thus, the aim of this study was to evaluate the rice bran extraction process through ultrasonic waves in an alkaline medium for further protein concentrate precipitation. A sequential strategy of experimental design was employed in order to optimize the conditions for protein obtainment. Moreover, the protein concentrate composition and functional properties was evaluated.

Materials
Defatted rice bran (DRB) in pellets have been supplied by Indústria Riograndense de Óleos Vegetais (Irgovel, Pelotas-RS, Brazil) -2013/2014 harvest. Prior to protein extraction study, the DRB had been ground using a cutting mill (Solab, SL 31 Willye Type, Brazil) with 70 mesh average particle size, packed in polyethylene bags and stored at -20 ºC. All chemical reagents used have been commercially purchased from different sources.

Sequential strategy of experimental design used for deffated rice bran (DRB) protein extraction
The use of ultrasonic waves in the alkaline medium has been studied in order to maximize the conditions for DRB protein extraction and a summary of experiments conducted in the present study were detailed in figure 1.
Initially, two fractional factorial designs (FFDs) 2 4-1 (3 central points, totalling 11 runs) have been applied at different ultrasonic frequencies (FFD-A: 37 kHz; and FFD-B: 80 kHz) to evaluate the effects of four variables on the extraction process, such as: pH (8-10), temperature (25-45 ºC), ultrasonic exposure time (5-30 min), and ultrasonic treatment amplitude (40-100%). Independent variable, percentage of extracted protein (g of protein 100 g -1 of DRB), have been selected from preliminary tests based on scientific studies (GNANASAMBANDAM & HELTIARACHCHY, 1995;GUPTA et al., 2008;FAN et al., 2012;ZHANG et al., 2012). Real and coded values of studied FFDs variables are shown in table 1. FFD runs have been conducted in accordance with a patented proposed (HYUN-JUN et al., 2009), with some modifications. Extraction procedure started by suspending 30 g of DRB in 170 mL of ultrapure water and correcting suspension pH with 2 mol L -1 NaOH solution, in accordance with proposed design levels. Subsequently, the suspension was disposed in a glass becker and submitted to indirect sonicate in an ultrasonic bath (Elmasonic, P60H, Elma, Germany) at 37 and 80 kHz frequencies, using the temperature, exposure time, and amplitude proposed by the design in table 1. Temperature was controlled by ultrasound bath heating system itself and cooling by the addition of ice. Phase separation has been done by centrifugation at 11979 g at 25 ºC for 15 min (Cientec, CT5000R, Brazil).
FFDs performance allowed the selection of statistically significant variables (P ≤ 0.05) in relation to DRB protein extraction process. Following that, a central composite rotatable design (CCRD) with the variable pH (8-10) and temperature (15-45 ºC), three central point repetitions, and four axis points (2 2 , totaling 11 runs) was done. The CCRD 2 2 matrix with real and coded values of studied variables is shown in table 2. Further, the CCRD experiments were validated in the optimized condition for DRB protein extraction.
The supernatants protein contents collected from FFDs and CCRD have been evaluated by Micro-Kjeldahl method (AOAC, 2005) and a 5.95 nitrogen conversion factor has been used. Protein recovery percentage has been calculated taking into account the supernatants protein obtained from the extraction process and initial protein value on DRB (Equation 1). Proteins content used for calculating the recovery percentage has been described in wet-basis. (1)

Obtainment of the rice bran protein concentrate (RBPC)
The protein precipitation stage was carried out after establishing the best extraction condition from the sequential strategy of experimental design. The collected supernatants had their pH adjusted to 4.5 with 3 mol L -1 HC1 which is the rice proteins isoelectric point (GUPTA et al., 2008). The solution was left to precipitate overnight at 10 ºC, followed by centrifugation at 11979 g at 25 °C for 15 min (Cientec, CT5000R, Brazil). The precipitate was washed three times with purified water and neutralized (pH 7.0) with 2 mol L -1 NaOH. Next, the precipitate was frozen (-18 ºC) and freeze-dried (Labconco, Free Zone, USA). The drying process was conducted for 24 h (-40 ºC, 0.8 mbar).

Composition and color
The centesimal composition (moisture, protein, ash and lipids) of DRB and RBPC were performed according to Association of Official Analytical Chemists (AOAC, 2005). The nitrogen content was determined by micro-Kjeldahl method, using 0.5 g of DRB or 2 mL of RBPC supernatant after centrifugation and 10 mL H 2 SO 4 in sample digestion performed at 300 ºC, followed by neutralization with NaOH 50% (v v -1 ) and nitrogen distillation. The titration was performed with 0.01 mol L -1 H 2 SO 4 and mixed indicator (methyl red and bromocresol green), applying a factor of 5.95 to convert nitrogen to protein.
Carbohydrate were quantified by difference. For mineral analysis, the RBPC samples were digested by microwave-assisted wet digestion method using a microwave oven (Multiwave 3000, Anton Paar, Graz, Austria) equipped with eight high-pressure quartz vessels with internal volume of 80 mL. The maximum operational temperature and pressure were set at 280 °C and 80 bar, respectively. A sample mass of 300 mg of RBPC were directly weighed into the quartz vessels and added of 6 mL of 14.4 mol L -1 HNO 3 solution. The following heat steps were used: 15 min at 1400 W and holding for 10 min, and 20 min at 0 W (cooling step). Obtained digests were transferred to a volumetric flask and the final volume was completed to 25 mL with ultrapure water.
The trace-elements determination was performed using an inductively coupled plasma mass spectrometer (ICP-MS) (Perkin-Elmer-SCIEX, model Elan DRC II, Thornhill, Canada), equipped with a concentric nebulizer (Meinhard Associates, Golden, USA), a cyclonic spray chamber (Glass Expansion, Inc., West Melbourne, Australia), and a quartz torch and injector tube (2 mm i.d.). Instrumental performance optimization, including nebulizer gas, flow rate, ion lens voltage and torch alignment, was carried out following the instructions of the manufacturer. The following operational conditions were used: radiofrequency power of 1400 W, plasma gas flow rate of 15 L min -1 , auxiliary gas flow rate of 1.20 L min -1 and nebulizer gas flow rate of 1.08 L min -1 . Isotopes monitored were 75 As, 112 Cd, 59 Co, 52 Cr, 65 Cu, 55 Mn, 98 Mo, 58 Ni, 208 Pb, 121 Sb, 82 Se, 120 Sn, 88 Sr and 64 Zn.
The RBPC amino acids profile and the protein nitrogen were evaluated using ultra violet high performance liquid chromatography (HPLC-DAD/ (2) RBPC surface color was measured using a Minolta Chromameter CR-400 (Osaka, Japan) with a specular component, a D illuminant and a 45º observer. Color results were expressed in the CIELAB (Commission International for Illumination) dimensions of lightness (L * ), redness (a * ) and yellowness (b * ).

Functional properties
The functional properties were determined according to MENEGOTTO (2019), with modifications. For solubility, the RBPC (0.2 g) was mixed with 40 mL of water and magnetically stirred (Fisatom, 752A, São Paulo, Brazil) for 10 min at room temperature (30 ºC). The pH suspension was adjusted to 2, 3, 4, 5, 6, 7, 8, 9 e 10 by addition of 2.0 mol L -1 HCl or 2.0 mol L -1 NaOH. Subsequently, the samples were transferred in a 50 mL falcon tubes and centrifuged (Rotina 420 R, Hettich Zentrifugen, Germany) at 10000 g at 25 ºC for 15 min. The supernatant was used to quantify the soluble protein content by LOWRY et al. (1951) method; the samples absorbance were measured using a spectrophotometer (UV-Vis Lambda XLS HP9 2FX, PerkinElmer, USA) at 660 nm wavelength. According to the same conditions, without pH adjusts and centrifugation, a control sample was performed. The results were calculated according Equation 3 and expressed as a percentage of soluble protein. (3) For water absorption capacity (WAC) and oil absorption capacity (OAC) the RBPC (0.1 g) was added to distillated water or corn oil, respectively, up to 5.0 mL in a 15 mL falcon tube, and vortex stirred (IKA, Staufen, Germany) at a temperature of 30 ºC for 1 min. The mixture was submitted to rest for 30 min, and then centrifuged (Cientec, CT 5000 R, São Paul, Brazil) at 3000 g at 25 ºC for 15 min. The supernatant was carefully drained on a clock glass for 25 min, thus just the water or the oil absorbed by proteins were measured by the weight. The results were expressed as g of water or oil held per g of RBPC and calculated by Equation 4 and 5.
(4) (5) The foaming capacity and stability of RBPC experiments were conducted on 50 mL falcon tubes and the dilutions were prepared with 0.02 g of RBPC and 5.0 mL of sodium phosphate buffer (0.01 mol L -1 ) on pH 4, 6 e 8. The samples were submitted to agitation (Fisatom, 952, n.12171, São Paulo, Brazil) for 10 min, and stirred for 1 min on a mixer with spiral beater (Britânia, BMM2, China). Thus, the increases of foam volumes by the stirred were evaluated and foaming capacity could be calculated by Equation 6. The foaming stability was determined considering the foam produced in foaming capacity test, which was left to rest for 30 min at room temperature (30 ºC) and the final volume was measured, allowing the foaming stability calculation by Equation 7. (6) The emulsifying capacity was conducted on 50 mL beaker and the dilutions were prepared with 0.04 g of RBPC and 10 mL of sodium phosphate buffer (0.01 mol L -1 ) on pH 2, 4, 6 e 8. After, the suspension was stirred for 5 min (Fisatom,952,São Paulo,Brazil) and the emulsifying occurred by the slow addition of corn oil droplets (2 mL) and under continuous agitation. The emulsion point was visually verified by the formation of a homogeneous phase (cream). The obtained emulsion was transferred to a 50 mL falcon tube and centrifuged (Cientec, CT 5000 R, São Paulo, Brazil) at 2087 g for 5 min. The emulsifying capacity was expressed based on Equation 8. A similar procedure was followed to determine the emulsifying stability, but after the emulsion formation the samples were incubated at 80 ºC for 30 min, before the centrifugation. The emulsifying stability was calculated according to Equation 9.
With the purpose to compare the results obtained to RBPC, dried bovine serum albumin (BSA) was also evaluated according to the same conditions for WAC, OAC, foaming and emulsifying properties. (8) Bernardi et al.

Statistical analysis
FFDs and CCRD runs were conducted randomly, and the data were analysed by Experimental Design procedure of Statistica 8.0 software (Statsoft Inc., Tulsa, USA). The adequacy of the models was expressed by the determination coefficient (R2), adjusted R2 (R2-adj) and statistical significance was determined by analysis of variance (Anova) (P ≤ 0.05). Tukey's test has been used in order to evaluate statistically significant differences in the average (P ≤ 0.05). Results in triplicate were expressed as the mean ± standard error.

Selection of variables for DRB protein extraction improvement
In the present study, ultrasonic-assisted extraction in an alkaline medium was used in DRB proteins extraction (14.89 ± 0.29% of proteins in wet basis) and results were presented in table 1. The protein extraction percentages observed for FFD-A (37 kHz) were higher (from 1.80% to 2.52%) if compared with the respective run in FFD-B at 80 kHz frequency (from 1.50% to 2.16%) (P ≤ 0.05). The process allows protein recovery percentages (R) from 14.89% to 16.93% for 37 kHz frequency, and between 10.35% and 14.51% for 80 kHz frequency. Thus, the protein content extracted at 37 kHz frequency was significantly higher than 80 kHz (P ≤ 0.05, Tukey test) for all runs, evidenced by the higher soluble protein content and higher protein recovery. Such difference is related to the fact that cavitation is greater at lower frequencies and that bubbles generated during cavitation are relatively large, making their collapse to result in strong shock waves, useful in mechanical cut applications (CHEMAT et al., 2011). Cavitation bubbles formation is more difficult at higher frequencies; however, such bubbles formation can occur by virtue of shorter rarefaction cycles due to time, temperature, and collapsing pressure, being mutually dependent of these factors. However, ultrasonic intensity cannot be increased indefinitely, as bubbles could grow so much in rarefaction that the time available for collapse becomes inadequate. In this way, it was opted to study only the 37 kHz frequency interference for the CCRD development.
Ultrasonic exposure time and pH were the variables with the effects (P ≤ 0.05) on the extraction process ( Table 3). The pH positively influenced the protein extraction in both FFD-A and FFD-B, indicating that the highest percentage of proteins extraction occurs at higher pH levels within the studied range (8 to 10). Normally, protein extraction from vegetables involves alkaline extraction, in which proteins are more negatively charged and repulsion is greater among molecules, increasing proteins solubility, consequently facilitating its extraction.
The temperature had no effect (P > 0.05) on the response evaluated in both FFDs, indicating that the temperature studied range (25-45 ºC) presented similar effects on the DRB proteins extraction. KARKI et al. (2010) observed that soy protein extraction was not affected (P > 0.05) by temperature fluctuation during sonication (29-38 ºC) in an ultrasonic bath with pH 8.5, 20 kHz frequency, and 1280 W power; protein extraction did not increase by thermal effects but by cellular rupture and particles size reduction. High-temperature protein extraction processes could be unfavorable as they induce protein denaturation, resulting in low-quality proteins (DAMODARAN, 2010). Therefore, this variable was fixed at 30 ºC, Table 3 -Effects of FFDs 2 4-1 factors on percentage of proteins extracted from defatted rice bran (DRB) (g protein 100 g -1 DRB).  more economically viable due to its easy control, because is near the room temperature. Ultrasonic exposure time presented a significant effect only for the 80 kHz-frequency extraction. The response was positively influenced (P ≤ 0.05) within the studied range (5-30 min), thus and increase in ultrasound exposure time increased the protein extraction. Ultrasonic exposure time on puerarin extraction efficiency from Radix Puerariae Lobatae -used in China as an antipyretic and antiemetic agent -was evaluated at 40 kHz frequency, at periods of 15, 20, 30, 40, and 50 min, and an increase in extraction was observed in ultrasonic exposure time of 30 min (FAN et al., 2012). Other studies indicated that a greater time exposure to ultrasonic waves could contribute towards greater extraction percentages of desired compounds (ZHAO et al., 2013). Therefore, this variable level was rearranged for further design in order to achieve a higher protein extraction.
The ultrasonic treatment amplitude had no significant effect (P > 0.05) in the evaluated response for both FFDs within the range studied (40-100%). However, the observed effects were positive, confirming the statements of several authors, who have indicated that by increasing the amplitude intensity, the sonochemical effects are deeper (FILGUEIRAS et al., 2000), benefiting extraction of the desired compounds. So, as the maximum amplitude level for the ultrasonic bath used in this study was 100%, it was fixed at this level for the next design (CCRD).

Optimization of protein extraction from DRB
In order to optimize the protein extraction from DRB, a CCRD 2 2 (3 central points, totaling 11 runs) was performed considering the variables with effects obtained in FFDs. The frequency of 37 kHz, temperature of 30 ºC, and amplitude of 100% were fixed and the pH and ultrasonic exposure time were varied from 8 to 10 and from 15 to 45 min, respectively. The matrix with the real and coded values of the studied variables and the response are shown in table 2. The extraction percentages varied from 1.91% to 2.37%, allowing protein recovery between 12.83% and 15.92%. The higher protein extraction was observed in runs with higher pH values, especially above 9.
CCRD results analysis enable to obtain the regression coefficients for the protein extraction considering the variables studied. The coefficient of linear pH was the variable with higher and significant effect (P ≤ 0.05) in protein extraction. Non-significant effects were incorporated into the residue of the model. A valid linear model was obtained (p-value = 0.0095) and the variation percentage explained by the model was adequate (R 2 = 68%; adjusted R 2 = 60%). The model was considered adjusted to the experimental data, allowing response surface generation (Figure 2). Moreover, the predicted values ( Table 2) presented relative deviations lower than 10%, confirming the model's adequate adjustment to the experimental data.
It was possible to observe, through response surface (Figure 2), that pH was the variable with the greatest influence on protein extraction process, and pH values close to 10 allowed a greater extraction percentage. For ultrasonic exposure time variable was noticed that periods of time between 30 and 45 min presented better results.
Thus, the optimized conditions for DRB protein extraction was validated. The following conditions were employed: frequency of 37 kHz; temperature of 30 °C; ultrasound amplitude of 100%; pH of 10; ultrasound exposure time of 30 min. The validation runs were similar than obtained in run 6 of CCRD (P > 0.05), allowing 2.24% ± 0.04 of protein extraction, reaching a 15.07% of final protein recovery in relation to the initial protein percentage of DRB. Results obtained in the present study were considered higher than reported by PHONGTHAI et al. (2016a), which reached 4.73% of extracted protein from DRB at 76% amplitude, 18 min ultrasonic exposure time, and a DRB: water ratio of 0.99 g: 10 mL. The same group of researchers employing microwave for protein extraction from DRB and achieved a 4.37% of extraction (PHONGTHAI et al., 2016b), showing that ultrasound are more efficient technique for DRB proteins extraction.
There are only a few results obtained from studies describing DRB protein extraction processes if compared to other sources of protein that are raw material for concentrates preparation and protein isolate (whey and egg white proteins). However, taking into account that DRB is a by-product from rice processing industry, often either discarded or used as animal feed, the study for DRB proteins extraction and recovery is imperative for the industrial byproduct use. Ultrasound technology contributes to improve the results of protein extraction from rice bran due to the phenomenon of cavitation, which causes the rupture of cells and the transfer of mass to the medium. This cellular breakdown is very desirable in the extraction of substances from a given medium (CHEMAT et al., 2011).
Alkaline solutions have been widely used as an efficient and useful method of proteins extraction from vegetable sources. Vegetable proteins are hardly Bernardi et al. water-soluble due to their hydrophobic nature and the presence of disulphide bond among protein molecules. With alkalinization of the medium by NaOH addition, disulphide bonds breakdown occurs causing structural changes and an increase in proteins molecules superficial charge, making proteins more soluble (SHEN et al., 2008;DAMODARAN, 2010). In addition, the ultrasound could improve compounds extraction from a tissue due the mechanical effects and greater penetration of solvent, improving the mass transfer (CHEMAT et al., 2011).

RBPC composition and color
The DRB centesimal composition evaluated in this study are present in table 4 and the results are in the range described by Romero (ROMERO, 2015). However, results obtained of RBPC composition presented higher content of protein and lipid and lower content of carbohydrate in comparison with YADAV et al. (2011); the differences in RBPC composition may be justified by the different strategy used on the extraction step: ultrasound waves at pH 10 and 30 °C for 30 min (current study) versus pH 11 at 60 ºC for 60 min (YADAV; YADAV; CHAUDHARY, 2011).
Moreover, other studies used sonication on DRB protein extraction and obtained a RBPC with a lower protein content (72.06 and 76.09%) (CHITTAPALO & NOOMHORM, 2009;LY et al., 2018). The greater protein content obtained in the current study can be attributed to the sequential strategy of the experimental design applied as a tool to optimize the protein extraction process.
Regarding the mineral composition, Mn was the predominant mineral on RBPC, followed by Cu, Zn, As, Ni, Mo, Sr, Co, Pb and Cr; Sb, Se and Sn were not detected (Table 5). In general, the mineral composition of plants may be easily influenced and controlled by their genetics, soil fertility and the environment (MILLER, 2010). However, even submitted through ultrasound and alkali treatment, the RBPC has a considerable amount of ash, which corresponded to minerals with effects on the human body. The RBPC is characterized as a light brown fine powder, with a velvety texture. In contrast, RBPC extracted in alkaline medium (pH 11) with ultrasound waves (100 W, 5 min) generates a darker product (L * : 46.17) (Table 5)  x 1 : pH; x 2 : ultrasound exposure time (min); y: percentage of protein extracted (g of protein 100 g -1 of DRB).
The amino acid profile of RBPC was shown in table 6 and compared with other RBPC, and BSA profiles available in the literature. Subsequently the RBPC amino acid profile was compared to the FAO protein requirement (preschool children 2-5 years), obtaining the essential amino acid chemical score. All essential amino acids content of RBPC are below the standard protein composition (WHO/FAO/ UNU, 1981, 2007; thus, a chemical scores of amino acids ≤ 1.0 are obtained. It is common in proteins obtained from a low biological value sources, as cereals including rice. Leucine is the RBPC limiting amino acid (chemical score 0.38), followed by lisine (0.48); in contrast, threonine had the highest chemical score (1.0). Other studies evaluated the amino acids profile of RBPC showing some variations, probably, due to the protein extraction procedure employed (TANG et al., 2003;HAN et al., 2015). Comparing the RBPC amino acids profile with BSA, lower amino acids content was observed for RBPC, an expected condition, since the BSA is a protein of high biological value (PRATA & SGARBIERI, 2008).
Although, non-essential amino acids were detected in the sample, this finding is not relevant, as it is known that these can be produced by the human body from essential amino acids. Protein quality is related to true nitrogen and amino acid retention, which is mainly related to the capacity of the protein to essential amino acids (LEE et al., 2016).

RBPC functional properties
The results of functional properties are shown in figure 3. The RBPC protein solubility presented a U-shaped trend, with a minimum solubility (22.2%) at pH 4, corresponding to the pI of rice bran proteins as previously demonstrated (GUPTA et al., 2008;ZHANG et al., 2012). In contrast, the RBPC showed great solubility at pH 6, 7, 8, 9 and 10, with no significant difference (P > 0.05); the maximum solubility was 99.7% (pH 7). It is recognized that in the pI the protein solubility is minimum, due to the lack of electrostatic repulsion, which promotes aggregation and precipitation through to the hydrophobic interactions. However at pH values above or below pI, proteins have positive and negative charges, which allows the electrostatic repulsion and the hydration of charged residues improving the protein solubility, an essential characteristic that may affect the other functional properties (DAMODARAN, 2010).
The RBPC presented higher results for WAC and OAC than BSA (P ≤ 0.05). The results obtained in the present study for WAC was lower and, for OAC was higher than values of RBPC extracted by agitation (1 h) at 25 ºC in pH 9.0 (WAC: 3.87 to 5.60 g g -1 ; and OAC: 3.74 to 9.18 g g -1 ) (CHANDI & SOGI, 2007). The differences may be associated with the protein extraction procedure, that can be able to improve some functional properties of protein (JAMBRAK et al., 2008). Ultrasound treatment may affect WAC and OAC properties since ultrasonic waves may form free radicals Table 4 -Centesimal composition of defatted rice bran (DRB) and rice bran protein concentrate (RBPC).

Parameters DRB RBPC
---------Centesimal composition (g 100 g -1 ; dry basis)---------  Parameters RBPC -------------Mineral composition (µg g -1 ; dry basis)-------------  that increase WAC (HU et al., 2013); furthermore, ultrasonic waves may increase the protein nonpolar groups exposure, favoring its connection with fat or oil polar fractions, improving OAC (LIN; HUMBERT;SOSULSKI, 1974). It is important to emphasize that greater values of WAC and OAC are considered positive, especially in processed foods, improving water and oil retention in the matrix and obtaining a food product with desired sensory characteristics. The obtained RBPC had lower values for foaming capacity and stability than BSA (P ≤ 0.05) and both proteins, RBPC and BSA, have their foaming properties influenced by pH as previously reported in the literature (ROMERO, 2015). The lower RBPC foaming capacity and stability is probably associated with the high lipid content, which influenced foaming stability in negative way because the lipids are quickly adsorbed in airwater interface, impeded the protein adsorption (DAMODARAN, 2010;CANO-MEDINA et al., 2011). The RBPC emulsifying capacity was comparable to BSA in the pHs evaluated (P > 0.05), and both proteins present lower emulsion capacity in the pH 2 (P ≤ 0.05). However, the RBPC present lower emulsifying stability than BSA (P ≤ 0.05), with the greater emulsifying stability observed at pH 4, 6, and 8. Although, the emulsifying properties are lower than BSA, they are still considered suitable and in the range described for concentrates obtained from DRB (PHONGTHAI et al., 2016a).
The RBPC functional propertied of solubility, OAC, emulsion capacity and stability are a desirable characteristics in emulsified foods, especially at a neutral or slight acid pH. Furthermore, the RBPC potential use must be considered due to it be a highly available byproduct and because it may be combined with other proteins to improve its functional properties and amino acid biological value. Table 6 -Rice bran protein concentrate (RBPC) amino acids profile (mg g protein -1 ) obtained in the current and other scientific study and bovine serum albumin (BSA), essential amino acid requirements of preschool children, and amino acid chemical scores of RBPC.

CONCLUSION
DRB protein extraction contributed to add value to rice bran and the use of a combination of ultrasound exposure in alkaline medium can improve this process. The sequential strategy of experimental design allowed to reach 15.07% of protein recovery and the RBPC presented 84.76 g 100g -1 of protein, higher values than all reported in the literature. The optimized and validations conditions were: frequency of 37 kHz; temperature of 30 °C; ultrasound amplitude of 100%; pH of 10; ultrasound exposure time of 30 min. The extraction using ultrasonic waves ensure a suitable RBPC functional property, especially the solubility in pH above 6, oil absorption and emulsifying properties. The RBPC presented excellent potential as alternative protein source and food supplement, evidenced from its amino acids and mineral profile. Thus, RBPC can be considered an important food ingredient that can improve the quality of food products contributing with the development of novel foods. Results are expressed by mean ± standard error (n = 3); different lowercase letters indicate statistically significant differences by Tukey test between samples in the same condition (P ≤ 0.05); different uppercase letters indicate statistically significant differences by Tukey test between pH for the same sample (P ≤ 0.05).