Optimized methodology for alkaline and enzyme assisted extraction of protein from sacha inchi copy

For the alkaline extraction method, the optimal SI protein extraction conditions corresponded to 54.2C, solvent/meal 42/1 v/w ratio, NaCl concentration of 1.65 M, pH 9.5 for 30 min and y

OPTIMIZED METHODOLOGY FOR ALKALINE AND ENZYME-ASSISTED EXTRACTION OF PROTEIN FROM SACHA INCHI

(PLUKENETIA VOLUBILIS) KERNEL CAKE

ROSANA CHIRINOS1, MARTIN AQUINO1, ROMINA PEDRESCHI2and DAVID CAMPOS1,3

1

Instituto de Biotecnologıa, Universidad Nacional Agraria La Molina-UNALM, Lima, Peru

2 School of Agronomy, Pontificia Universidad Cat olica de Valparaıso, Quillota, Chile

3 Corresponding author.

TEL: 1051 1 6147800 ext 436;

FAX: 1051 1 3495764;

EMAIL: dcampos@lamolina.edu.pe

Received for Publication February 9, 2016

Accepted for Publication April 21, 2016

doi:10.1111/jfpe.12412

ABSTRACT The residue after oil extraction from sacha inchi (SI) presents a high protein content of59% that can be further exploited to extract proteins In this study, the protein extraction parameters for defatted SI cake meal (DSICM) were optimized using alkaline and enzyme-assisted extractions A central composed design (CCD) was used to optimize the protein yield for both methods The obtained response surface models (RSM) produced a satisfactory fitting of the results for both extraction methods (R250.9609–0.9761) For the alkaline extraction method, the optimal SI protein extraction conditions corresponded to 54.2C, solvent/meal 42/1 (v/w) ratio, NaCl concentration of 1.65 M, pH 9.5 for 30 min and yielded 29.7% protein For the enzyme-assisted method, optimal extraction conditions corresponded to an enzyme concentration of 5.6%, 40.4 min extraction, solvent/meal 50/1 (v/w) ratio, pH 9.0 and 50C and yielded 44.7% protein and hydrolysis degree of 7.8%

PRACTICAL APPLICATIONS The cake obtained after oil extraction from SI seed is an important source of protein Thus, efforts should focus on the development of protein extraction processes from the cake to add value to this by-product Up to date, studies are very limited Results obtained in this study allowed the optimization of the protein extraction process from SI cake meal The enzyme-assisted protein extraction resulted in a higher quantity of protein recovery (1.5 fold more) than the alkaline protein extraction The optimized protein extraction process will allow the food industry to obtain isolates or protein concentrates from SI cake meal to be used as techno-functional, nutritional and/or functional agent

INTRODUCTION

Proteins are macronutrients necessary for human beings and

they constitute an important nutritional contribution not

only as energy source but as source of nitrogen and essential

aminoacids Proteins are also important because they confer

physicochemical, functional and organoleptic properties to

foods (Scopes 1986)

Nonconventional sources of protein (e.g., by products

from agroindustry) could render added value as functional

ingredients, for nutritional purposes to fortify foods and for

pharmaceutical and cosmetics applications Currently, the

food industry is in need of alternative protein sources that can compete with the actual protein sources that dominate the market (Pszczola 2004) Within this context, sacha inchi (SI) or Plukenetia volubilis is a highly oil-containing seed (54%) and with a relatively high protein content (27%) (Hamaker et al 1992) The cake remaining after oil extrac-tion from SI presents a protein content of 59% dry weight (DW) (Sathe et al 2012; Ruiz et al 2013) Hamaker et al (1992) have reported in SI protein a high content of cysteine, tyrosine, threonine and tryptophan and a low content of phenylalanine In addition, Sathe et al (2012) reported that

the water soluble albumin fraction constituted25% (%) of

the defatted SI seed flour

In the last few years, there is an increasing interest on

methods for extracting plant protein based on acid, alkaline

and enzyme-assisted extraction (Sari et al 2013), being the

alkaline and enzyme-assisted methods more amenable for

practical applications The number of studies dedicated to

protein extraction from SI is very limited up to date and

none of the previous studies have focused on the

optimiza-tion of the protein extracoptimiza-tion parameters using the response

surface methodology (RSM) Sathe et al (2012) evaluated

the defatted flour protein solubility of SI using a step by step

alkaline extraction with yields of50% protein

RSM is an excellent statistical technique for the

optimiza-tion of complex processes (Box and Draper 2007) RSM

explores the existing relationships between explicative

varia-bles and one or more response variavaria-bles (Cao et al 2012)

This methodology has been previously used in the

optimiza-tion of protein extracoptimiza-tion either using alkaline or

enzymatic-assisted methods from different food sources such as flaxseed

(Oomah et al 1994), pine seed (Wang et al 2011), palm

ker-nel cake (Chee et al 2012), soybean (Rosenthal et al 2001;

Rosset et al 2014), lentil (Jarpa-Parra et al 2014), etc

Since there are no previous studies on the optimization of

protein extraction yields from SI under alkaline and

enzyme-assisted methods in combination with RSM, the objectives

of this study were (1) to evaluate the effect of alkaline

extrac-tion parameters such as NaCl concentraextrac-tion, temperature

and solvent:meal ratio at pH 9.5 on the response variable

protein yield (%) from defatted SI cake meal (DSICM) by

applying RSM; (2) to evaluate the effect of enzyme-assisted

extraction parameters with Alcalase 2.4L enzyme

concentra-tion and time at pH 9.0, on the response variables protein

yield and hydrolysis degree (%, HD) by applying RSM The

optimization of the protein extraction parameters in DSICM

offers an alternative process for obtaining protein from a

nonconventional source (SI cake) and simultaneously this

agro-industrial by-product could be re-valorized

MATERIALS AND METHODS

Defatted Sacha Inchi Cake

SI kernel cake was provided by Olivos del Sur enterprise

(Lima, Peru) SI cake was obtained after oil extraction from

SI seed using an expeller Proximate analysis was performed

in SI kernel cake according to the method of AOAC (1995)

for nuts and nut products Protein content was calculated

using a conversion factor of 5.70 (Sathe et al 2012) The

cake was ground in a hammer mill to obtain particles of

500 lm Ground cake meal was defatted for 12 h using

petroleum ether at a solvent/meal ratio of 10/1 (w/v) under

300 rpm stirring conditions The DSICM was air-dried at

40C for 2 h in an oven then was packed in polyethylene bags and stored at 4C until use

Enzyme and Chemicals Alcalase 2.4L was provided by Novozyme (Bagsvaerd, Den-mark) All chemicals used were of reagent grade and pur-chased from Sigma (St Louis, MO) and Merck (Darmstadt, Germany)

Protein Analyses The soluble proteins were determined according to Lowry

et al (1951) and total protein with the Kjeldhal method (AOAC 1995) Protein yield (Y, %) was calculated as g of soluble protein from extract/100 g of total protein of DSICM

Hydrolysis Degree

HD (%) was determined in each hydrolyzed sample using the method of Adler-Nissen (1979) by assaying free amino groups with 2,4,6-trinitrobenzenesulphonic acid (TNBS) and using the following equation:

HD %ð Þ 5 h=hð totÞ3100 5 100 3ðAN2– AN1Þ=Npb

; where h is the number of peptide bonds broken, htotis total number of bonds per unit weight, AN1is the amino nitrogen content of the protein substrate before hydrolysis (mg/g protein), AN2is the amino nitrogen content of the protein substrate after hydrolysis (mg/g protein) and Npb is the amino nitrogen content of the peptide bonds in the protein substrate (mg/g protein) as determined after total hydrolysis with 6 M HCl at 110C for 24 h The values of AN2and AN1

were obtained from a standard curve at 340 nm absorbance versus mg/L amino nitrogen generated with L-leucine Protein Extraction

Alkaline Extraction of Sacha Inchi Protein Protein from DSICM was extracted using selected combinations of independent variables: temperature (C), solvent/meal ratio (v/w) and NaCl concentration (M) according to the experi-mental design All protein extractions were performed at pH 9.5, 300 rpm agitation for 30 min These parameters were kept constant based on preliminary studies After extraction, solutions were immediately centrifuged at 4,000 3 g for 30 min at 4C The supernatant were filtered through Whatman filter paper No 1 and the soluble protein and protein yield (%) were quantified All the experiments were carried out in triplicate

Enzyme-Assisted Extraction of Sacha Inchi

Protein Protein from DSICM was extracted using Alcalase

2.4L and the selected combinations of independent variables:

enzyme concentration (% enzyme in relation to the DSICM

protein content) and time (min) according to the

experi-mental design Protein extraction was carried at pH 9.0, 50C,

300 rpm stirring and at a solvent/meal ratio of 50/1 (v/w)

These parameters were kept constant as recommended for

Alcalase (pH and temperature) and the solvent/meal ratio

based on preliminary studies After extraction, solutions

were immediately centrifuged at 4,000 3 g for 30 min at 4C

The supernatants were filtered through Whatman filter

paper No 1 and the soluble protein, protein yield (%) and

HD (%) were determined All the experiments were carried

out in triplicate

Experimental Design and Statistical Analysis

Alkaline Extraction of Protein The RSM was used to

determine the influence of three independent variables and

the optimal conditions for protein extraction from DSICM

The effect of the variables temperature (X1), solvent/meal

ratio (X2) and NaCl concentration (X3) on the protein

extraction yield (dependent variable, Y %) was investigated

The selection ranges within which each factor varied was

based on preliminary experiments (data not shown) Each

variable was coded at five levels: 21.68, 21, 0, 1 and 1.68

(Table 1) The conversion of real values to coded values was

as follows:

xi5 Xð i– XoÞ=DXi; (1) where xiis the dimensionless value of an independent vari-able, Xiis the real value of an independent variable, Xois the real value of an independent value at the center point and

DXiis the step change

A central composite design (CCD) was used to allow fit-ting a second-order model (Nakai et al 2006) A total of 19 randomized runs that included five central points were per-formed (Table 1) The proposed model for the response vari-able (Y (%), protein yield) corresponded to:

y5b01X4 i51

bizi1X4 i51

biiz2i1X4 i6¼j51

bijzizj; (2)

where b0is the value of the adjusted response to the central point of the design, bi, bii and bijare the linear, quadratic coefficients and the intercept, respectively

The optimum protein extraction conditions consisted on determining the maximum protein extraction yield (maxima desirability) through a combination of different variables or factors Predicted values (Y) were transformed into a desir-ability value (d) The generated RSM to obtain maximum protein yield from DSICM was experimentally validated with three experimental replicates and the obtained values compared to the ones predicted by the RSM model The

TABLE 1 CENTRAL COMPOSITE DESIGN ARRANGEMENT AND EXPERIMENTAL AND PREDICTED PROTEIN YIELD VALUES FOR ALKALINE EXTRACTION

Run

Coded variables Uncoded variables Protein yield (Y) %

x 1 x 2 x 3 X 1 X 2 X 3 Experimental Predicted

9 21.68 0 0 29.8 35 1.25 22.61 22.27

10 1.68 0 0 80.2 35 1.25 23.67 22.71

11 0 21.68 0 55 9.8 1.25 13.85 14.49

12 0 1.68 0 55 60.2 1.25 27.5 25.56

14 0 0 1.68 55 35 2.51 26.87 25.72

X 1 , temperature; X 2 , solvent/meal ratio; X 3 , NaCl concentration.

surface plots were generated by varying two variables within

the experimental range and holding the other constant

(zero) at the central point All the statistical analysis were

carried out with Statgraphics Centurion XV software 15.2.06

(Stat Point Inc., VA)

Enzyme-Assisted Extraction of Protein The RSM

was used to determine the influence of two independent

var-iables on the optimal conditions for enzyme-assisted protein

extraction from DSICM In addition, the influence of the

same variables on the HD (%) of protein was examined The

effect of the variables: enzyme concentration (X1) and time

(X2) on the protein extraction yield (maximum) and protein

HD (minimum) were investigated The selection ranges

within which each factor was varied based on preliminary

experiments (data not shown) Each variable was coded at

five levels: 21.41, 21, 0, 1 and 1.41 (Table 2) The

conver-sion of real values to coded values was conducted as

described in Eq (1) for the two evaluated responses (protein

yield and HD)

A central composite design (CCD) allowed fitting of a

second-order model A total of 13 runs that included five

central points were performed (Table 2) The proposed

model for the response variables (second order polynomial),

desirability values, validation of RSM and generated surface

plots were calculated as described previously A multiple

response optimization was performed to determine the

com-bination of the experimental parameters (independent

varia-bles) that simultaneously maximize the protein yield and

minimize the protein HD The obtained result was

experi-mentally validated with three experimental replicates The

optimization of two variables was displayed as an overlaid

contour plot All the statistical analysis were carried out

with Statgraphics Centurion XV software 15.2.06 (Stat Point Inc., VA)

RESULTS AND DISCUSSION Proximal Composition and Protein Analysis

SI kernel cake presented 7.9% humidity and the contents of protein, fat, fiber, ash and carbohydrates in dry weight (DW) corresponded to 58.4, 8.9, 4.1, 5.7 and 22.7%, respec-tively, these values are close to the ones reported by Ruiz

et al (2013) DSICM reached values of 61.9% of protein (DW), this value was superior to the protein content reported for other defatted meals obtained from soybean, palm, and sesame (50, 16.8 and 42%, respectively) (Onsaard

et al 2010; Chee et al 2012; Rosset et al 2014) Moure et al (2006) reported that protein content of defatted meals from dehulled oilseeds depend on the seed type and ranges between 35 and 60% (DW)

Optimization of Alkaline Extraction The experimental design of five-levels, three-variable CCD and the experimental results of protein extraction are shown

in Table 1 Protein yield varied from 11.5 to 28.5% (or from 7.1 to 17.6 g protein/100 g of DSICM) Using alkaline extrac-tion and the RSM, protein recoveries from different defatted cakes from oilseeds ranged between 10.9 and 32.6; 3.3 and 5.7; 12.3 and 16.5; and 40.8 and 58.7 g of protein/100 g for flaxseed, pigeon pea, soybean and lentil (Oomah et al 1994; Jarpa-Parra et al 2014; Tan et al 2014)

The application of RSM yielded the following regression equation, which is an empirical relationship between protein yield (Y) and the evaluated variables (Eq (3)):

TABLE 2 CENTRAL COMPOSITE DESIGN ARRANGEMENT AND EXPERIMENTAL AND PREDICTED PROTEIN YIELD AND DEGREE HYDROLYSIS VALUES FOR ENZYME-ASSISTED EXTRACTION

Run

Coded variables Uncoded variables Protein yield (Y) % Hydrolysis degree (HD, %)

x 1 x 2 X 1 X 2 Experimental Predicted Experimental Predicted

1 21 21 2.00 15.00 28.97 28.60 0.96 0.92

2 1 21 5.00 15.00 34.45 34.80 4.03 5.01

3 21 1 2.00 45.00 29.18 28.51 3.75 3.91

4 1 1 5.00 45.00 40.33 40.37 5.38 6.55

5 21.41 0 1.38 30.00 28.79 29.46 1.56 1.71

6 1.41 0 5.62 30.00 42.58 42.23 7.72 6.42

7 0 21.41 3.50 8.79 28.41 28.36 2.97 2.54

8 0 1.41 3.50 51.21 31.85 32.23 6.44 5.74

9 0 0 3.50 30.00 34.80 34.99 4.07 4.12

10 0 0 3.50 30.00 35.00 34.99 4.25 4.12

11 0 0 3.50 30.00 34.91 34.99 4.49 4.12

12 0 0 3.50 30.00 35.07 34.99 4.03 4.12

13 0 0 3.50 30.00 35.18 34.99 3.72 4.12

X 1 , enzyme concentration (%); X 2 , time (min).

Y ð Þ 5 241:894 1 0:9801  X% 1 1 0:9972 X2

1 29:3638 X3 2 0:0086X1

1 0:0:0042X1 X2 20:1323 X1 X3

2 0:0125 X2 2 0:1074 X2 X3

2 5:35729 X3 :

(3)

Analysis of variance (Table 1 and Supporting Information)

revealed that DCC application resulted in a highly significant

model (P < 0.000) indicative of a good generated response

model for optimization with R250.9609 and an adjusted

R250.9218 These coefficients suggest good fitting of the

model given that at least the R2 should be higher than

0.8000 (Joglekar and May 1999) Within the experimental

evaluated range, the factor time did not significantly affected

(P > 0.05) protein extraction yield meanwhile the other

components solvent/meal ratio and NaCl concentration had

a high significant effect (P < 0.01) These results indicate

that solvent/meal ratio and NaCl concentration are the main

factors contributing to protein extraction from DSICM

Similar results have been previously obtained for extracted

protein from flaxseed and cowpea flour (Oomah et al 1994;

Mune et al 2008)

The surface responses are displayed in Fig 1 The effect of

temperature and solvent/meal ratio on protein yield is

dis-played in Fig 1a Solvent/meal ratio exerted a quadratic

effect on protein yield that can be evidenced in Fig 1a, where

its interaction with temperature is also displayed and with

solvent/meal ratios between 40/1 and 50/1 displaying the

highest protein yields In Fig 1a, a linear effect of

tempera-ture on protein yield can be observed, and within the 30–

70C range, no big variations were observed for the different

evaluated solvent/meal ratios Temperature exerted a slight

quadratic effect on protein yield at different NaCl

concentra-tions (Fig 1b) The interaction of solvent/meal ratio and

NaCl concentration is displayed in Fig 1c, where the

quad-ratic effect of both components is evidenced The quadquad-ratic

effect of NaCl concentration was also evidenced in Fig 1c,

reaching the maximum protein yield at NaCl concentrations

close to 1.5 M Results indicate that solvent/meal ratio and

NaCl concentration are the main contributors to the protein

extraction from DSICM

The desirability maxima function was used to obtain the

optimal extraction conditions The dependent variable was

set to the maximum possible (d 5 1), the optimal conditions

corresponded to 54C, a solvent/meal ratio of 42/1 (v/w),

NaCl 1.65 M at a pH of 9.5 and 30 min extraction time,

obtaining a protein yield of 29.7% (18.4 g protein/100 g

DSICM) Higher extraction yields (40.9 and 47 g solubilized

protein/100 g defatted SI flour) were reported by Sathe et al

(2012) for SI meal defatted with hexane, using a step by step

methodology, at conditions of 1M NaCl, 15–30 min

extrac-tion time, two extracextrac-tion steps, pH 9–12 and room tempera-ture The differences with that study can be attributed to the solvent/meal ratio and differences of the raw materials The

SI cake used in our study was exposed to a mechanical force and friction generated in the expeller during the process to obtain SI oil that could have affected the physicochemical characteristics of the SI protein disfavoring its extraction Finally, the suitability of the generated mathematical model

to predict maximum protein yield was experimentally vali-dated using the conditions determined in the optimization Thus, the experimental protein yield at the optimum condi-tions was 30.2 6 0.33% being this value close to the value generated by the mathematical model

Optimization of Enzyme-Assisted Extraction This study evaluated Alcalase 2.4L with the aim to increase the protein yield from DSICM with a low HD A low HD is key to obtain a not highly hydrolyzed protein that can be used as raw material in different food applications

FIG 1 RESPONSE SURFACE PLOTS AND CONTOURS FOR THE EFFECTS OF (a) TEMPERATURE VERSUS SOLVENT/MEAL RATIO, (b) TEMPERATURE VERSUS NaCl CONCENTRATION AND (c) SOLVENT/ MEAL RATIO VERSUS NaCl CONCENTRATION ON PROTEIN YIELD FOR ALKALINE METHOD OF PROTEIN EXTRACTION

The experimental design corresponded to five-levels and

two CCD variables The experimental results of protein yield

as well as the HD are shown in Table 2 Protein yield and

HD varied from 28.4 to 40.3% (or from 17.3 to 24.5 g

pro-tein/100 g of DSICM) and from 0.96 to 5.38%, respectively

Protein yield significantly increased (40%) when the

Alca-lase was employed in the protein extraction process in

com-parison to the alkaline extraction The efficiency of

proteolytic enzymes during protein extraction from different

sources has been extensively reported Sari et al (2013) by

using different proteases (1% of enzyme for 3 h) extracted

more protein from rapeseed, microalgae and soybean meals

(60, 80 and 90%, respectively) in comparison to alkaline

extraction (pH 9.5;15, 30 and 80%, respectively) A

signif-icantly (P < 0.05) higher trypsin extracted protein yield

(61.9 g/100 g) was obtained from palm kernel in comparison

to the alkaline (pH 9.5) method (10.2 g/100 g) (Chee et al

2012) Also Latif and Anwar (2011) found that proteases

Protex 7L and Alcalase 2.4L successfully extracted proteins

from sesame meal (87.1 and 79.6%, respectively) For the

HD, the maximum obtained corresponded to 5.38% Sari

et al (2013) reported that certain amount of hydrolysis is

needed and acceptable for protein extraction but a high

hydrolysis is detrimental because proteins would be

con-verted to peptides displaying an increased solubility and

thus altering the functional properties of the extracted

pro-teins (Rosenthal et al 2001; Taha and Ibrahim 2002) and the

bitterness associated with a high HD Taha and Ibrahim

(2002) reported that low protein HD (range between 8.8

and 9.5%) for soybean, sesame and rice bran meals

enzy-matically hydrolyzed with papain (0.06–0.21%) for 5 min

produced improvements in wettability, flow ability and

emulsifying capacity properties and a direct relation between

increasing HD, nitrogen solubility and dispersibility was

found

A quadratic (Eq (4)) and lineal (Eq (5)) relationship was

found between protein yield and HD with the different

extraction parameters evaluated by SRM The relationship

established between protein yield (Y) and HD (in real

val-ues) with the evaluated parameters is shown:

Y ð Þ 5 21:2596 2 0:2097  X% 11 0:4973

 X21 0:1901X1 1 0:063X1 X2– 0:0104 X2 ; (4)

HD %ð Þ 5 22:0681 1 1:1176  X11 0:0753 X2: (5)

Analysis of variance (Table 2 and 3 in Supporting

Informa-tion) revealed that CCD application resulted in a highly

sig-nificant model (P < 0.000) indicative of a good generated

response model for optimization with a good R250.9934

and the adjusted R250.9887 for protein yield and with a

moderate R250.8551 and adjusted R250.8261 for HD

Within the experimental evaluated range, the factors enzyme

concentration and time significantly affected (P < 0.05) pro-tein extraction yield as well as HD Thus enzyme concentra-tion and time contribute to protein extracconcentra-tion from DSICM and protein HD Similar results were reported for soybean meal, soybean, sesame and rice bran meals and palm kernel meal (Rosenthal et al 2001; Taha and Ibrahim 2002; Chee

et al 2012)

The surface response for protein yield is displayed in Fig 2a The effect of the enzyme concentration (%) in the evaluated range on protein yield presented an increasing trend (from20 to 28%) as the enzyme concentration was incremented (from 1.38 to 5.62%) Time exerts a quadratic effect on protein yield High protein yields were observed between 40 and 50 min of extraction The dependent variable was set to the maximum possible (d 5 1.00) and the obtained optimal conditions corresponded to a time of 54 min and enzyme concentration of 5.5% considering a 50/1 (v/w) sol-vent/meal ratio, pH 9.0 and 50C, respectively, obtaining a protein yield of 43.4% (26.8 g protein/100 g DSICM) Using the predicted optimum conditions, experiments carried out

in triplicates gave good results (43.78 6 0.28%) that coin-cided with the predicted value implying that the model was adequate The surface response for HD is displayed in Fig 2b The effect of the enzyme concentration (%) in the evaluated range of HD presented a lineal effect Increasing concentrations of Alcalasa 2.4L resulted in higher HD Also time exerted a lineal effect but its effect was less pronounced

on HD The optimization consisted on a minimization (d 5 0) because a low as possible HD was aimed The obtained optimal conditions corresponded to 8.78 min and

FIG 2 RESPONSE SURFACE PLOTS AND CONTOURS FOR THE EFFECTS OF ENZYME CONCENTRATION AND EXTRACTION TIME ON (a) PROTEIN YIELD AND (b) HYDROLYSIS DEGREE FOR THE ENZYME-ASSISTED METHOD OF PROTEIN EXTRACTION

1.37% of enzyme concentration obtaining a HD of 0.13%.

Using the predicted optimum conditions, experiments

car-ried out in triplicate gave good results (1.51 6 0.11%) very

close to the values predicted by the generated SRM model

Finally, the obtained optimization results for both responses

did not offer concluding results when they were evaluated in

separate Thus, a multiple optimization response was

gener-ated and the factors time and enzyme concentration that

resulted in a high protein yield and a low HD (lower to 10%)

were included The optimization of these two responses is

displayed as an overlaid contour plot in Fig 3 After the

mul-tiple response optimization, values of 5.62% Alcalase 2.4L

enzyme and 40.4 min at pH 9.0, 50C and 50/1 solvent/meal

ratio resulted in a maximum protein yield of 44.7% (27.6 g

protein/100 g DSICM) with a HD of 7.86% Same conditions

were experimentally validated resulting in protein yield and

HD of 44.7 6 0.4 and 7.86 6 0.14%, respectively Our results

indicate that the enzyme-assisted protein extraction was able

to extract 1.46 fold more protein than the alkaline extraction

from DSICM

CONCLUSIONS

RSM allowed optimization of the alkaline and

enzyme-assisted protein extraction conditions from DSICM For the

protein alkaline method, the factors: solvent/meal ratio and

NaCl concentration significantly affected the extraction

con-ditions, but not extraction time For the enzyme-assisted

protein extraction method, the Alcalase 2.4L enzyme

con-centration and time of hydrolyses affected the protein yield

and the HD By means of a multiple response methodology

(MRM) with the responses: protein yield and HD which

were maximized and minimized, respectively, it was possible

to obtain the maximum protein extraction with a low HD

Results of the MRM for the enzyme-assisted protein

extrac-tion method indicated that maximum protein yield (optimal

conditions, enzyme concentration of 5.6%, 40.4 min extrac-tion, solvent/meal 50/1 (v/w) ratio, pH 9.0 and 50C) was 46% higher in comparison to the alkaline method (optimal conditions, temperature: 54.2C, solvent/meal 42/1 (v/w) ratio, NaCl concentration of 1.65 M, pH 9.5 for 30 min) The predicted values for protein yield from all generated models were consistent and experimentally validated These results indicate that the enzyme-assisted protein extraction from sacha inchi kernel cake is an alternative protein extrac-tion method with higher yields than the tradiextrac-tional alkaline method Additionally, the recovered protein from this by-product could be considered as potential source of proteins

to be used in multiple industrial applications

ACKNOWLEDGMENT This research was supported by the grant in Science and Technology (2013–2014) supported by the Universidad Nacional Agraria La Molina (Lima, Peru)

REFERENCES

ADLER-NISSEN, J 1979 Determination of the degree of hydroly-sis of food protein hydrolysates by trinitrobenzenesulfonic acid

J Agric Food Chem 27, 1256–1262

AOAC 1995 Officials Methods of Analysis, 15th Ed., Association

of the Official Analytical Chemists, Washington, D.C., Gaithers-burg, Maryland

BOX, G and DRAPER, M 2007 Response Surfaces, Mixtures, and Ridge Analyses 2nd Ed., John Wiley & Sons Inc., Hoboken, New Jersey

CAO, W., ZHANG, C., JI, H and HAO, J 2012 Optimization of peptic hydrolysis parameters for the production of angiotensin I-converting enzyme inhibitory hydrolysate from Aceteschinen-sis through Plackett–Burman and response surface methodo-logical approaches J Sci Food Agric 92, 42–48

CHEE, L., LING, H.K and AYOB, K 2012 Optimization of trypsin-assisted extraction, physochemical characterization, nutritional qualities and functionalities of palm kernel cake protein LWT - Food Sci Technol 46, 419–427

HAMAKER, B.R., VALLES, C., GILMAN, R., HARDMEIER, R.M., CLARK, D., GARCIA, H., GONZALES, A.E., KOHLSTAD, I., CASTRO, M., VALDIVIA, R., et al., 1992 Amino acid and fatty acid profiles of the inca peanut (Plukene-tia volubilis) Cereal Chem 69, 461–463

JARPA-PARRA, M., BAMDAD, F., WANG, Y., TIAN, Z., TEMELI, F., HAN, J and CHEN, L 2014 Optimization of lentil protein extraction and the influence of process pH on protein structure and functionality LWT - Food Sci Technol 57, 461–469 JOGLEKAR, M and MAY, T 1999 Product excellence through experimental design In Food Product and Development: From Concept to the Market Place (E Graf and I.S Saguy, eds.), Aspen Publishers Inc., Gaithersburg, Maryland

FIG 3 SUPERIMPOSED CONTOUR PLOT FOR PROTEIN YIELD AND

HYDROLYSIS DEGREE (HD) AS A FUNCTION OF ENZYME

CONCENTRATION (%) AND EXTRACTION TIME (min) AT 50C,

SOLVENT/MEAL RATIO OF 50/1 AND pH 9.0

LATIF, S and ANWAR, F 2011 Aqueous enzymatic sesame oil

and protein extraction Food Chem 125, 679–684

LOWRY, H., ROSEBROUGH, J., FARR, L and RANDALL, J

1951 Protein measurement with the Folin phenol reagent J

Biol Chem 193, 265–275

2006 Functionality of oilseed protein products: A review Food

Res Int 39, 945–963

MUNE, M.A., MINKA, S.R and MBOME, I.L 2008 Response

surface methodology for optimisation of protein concentrate

preparation from cowpea [Vigna unguiculata (L.)Walp] Food

Chem 110, 735–741

NAKAI, S., LI-CHEN, Y and DOU, J 2006 Experimental design

and response surface methodology In Handbook of Food and

Bioprocess Modeling Techniques (S Sablani, A Datta, M.S

Rah-man and A Mujumdar, eds.), CRC Press, Boca Raton, FL

ONSAARD, E., POMSAMUD, P and AUDTUM, O 2010

Func-tional properties of sesame protein concentrate from sesame

meal Asian J Food Agro-Ind 3, 420–431

OOMAH, B.D., MAZZA, G and CUI, W 1994 Optimization

of protein extraction from flaxseed meal Food Res Int 27,

355–361

PSZCZOLA, D 2004 Ingredients of food technology J Food Sci

58, 56–69

ROSENTHAL, A., PYLE, D.L., NIRANJAN, K., GILMOUR, S

and TRINCA, L 2001 Combined effect of operational

varia-bles and enzyme activity on aqueous enzymatic extraction of

oil and protein from soybean Enzyme Microb Technol 28,

499–509

extrac-tion from defatted soybean flour with Viscozyme L pretreat-ment J Food Process Preserv 38, 784–790

RUIZ, C., DIAZ, C., ANAYA, J and ROJAS, R 2013 Analisis proximal, antinutrientes, perfil de acidos grasos y de aminoaci-dos de semillas y tortas de 2 especies de sacha inchi (Plukenetia volubilis y Plukenetia huayllabambana) Rev Sociedad Quımica Peru 79, 29–36

protein extraction from rapeseed, soybean and microalgae meals Ind Crops Prod 43, 78–83

SATHE, S., KSHIRSAGAR, H and SHARMA, G 2012 Solubiliza-tion, fractionaSolubiliza-tion, and electrophoretic characterization of Inca Peanut (Plukenetia volubilis L.) proteins Plant Foods Hum Nutr 67, 247–255

SCOPES, R (1986) Protein Purification Principles and Practice, 3rd Ed., Springer, New York

TAHA, F.S and IBRAHIM, M.A 2002 Effect of degree of hydro-lysis on the functional properties of some oilseed proteins Grasas y Aceites 53, 273–281

TAN, E.S., NGOH, Y.Y and GAN, C H.Y 2014 A comparative study of physicochemical characteristics and functionalities of pinto bean protein isolate (PBPI) against the soybean protein isolate (SPI) after extraction optimization Food Chem 152, 447–455

WANG, S., JIANG, L., LI, Y., LI, D and SUI, X 2011 Optimi-zation on aqueous enzymatic extraction conditions of pine seed protein by response surface method Proc Eng 15, 4956–4996