Valorization of Glycine max (Soybean) Seed Waste: Optimization of the Microwave-Assisted Extraction (MAE) and Characterization of Polyphenols from Soybean Meal Using Response Surface Methodology (RSM)

,e present study aimed at determining the optimal conditions for extraction of total phenolic compounds from soybean (Glycine max) meal, a by-product of the soybean seeds industry, using a green protocol with microwave-assisted extraction (MAE). A facecentered composite design (FCCD) was used for optimization. Based on a screening aimed to determine the factors that significantly influenced the responses, a 50% hydro-ethanolic solution was used with solvent/dry matter ratio (60/1–110/1), power (120–270W), and time (0–10min) as factors, while the responses studied were total phenolic and flavonoid contents. FTIR, TLC, DPPH, and FRAP anti-oxidants tests were used to characterize the extracts obtained with optimum conditions. ,e factors that significantly influenced both responses were the individual effect of all factors, the interaction between solvent/dry matter ratio and extraction time, the quadratic effect of solvent/dry matter ratio, and power for total phenolic content, while only the quadratic effect of power significantly influenced the flavonoid content. ,e highest contents of phenols (13.09mg GAE/g) and flavonoid (7.39mg CE/g) were obtained at 120W for 0.16min with a solvent/dry matter ratio of 60/1. ATR-FTIR spectra indicated the presence of polyphenolic compounds in the extract, namely flavonoids. TLC indicated the presence of at least nine compounds in the extract, among which catechin and quercetin were identified with respective Rf of 0.98 and 0.93. DPPH assay showed the antioxidant capacity for the extract with an IC50 of 194.98 μg/ml. RSM permitted us to develop a green protocol for maximum extraction of polyphenols from soybean seeds waste using less solvent, low power, and a reduced time in MAE.


Introduction
Soybean meal is a by-product of the soybean seed industry, which is commonly used for animal feeding [1,2] or rejected directly in nature, then causing serious problems to municipalities and to the environment [3,4]. In order to reduce pollution related to soybean seed industry's by-products and wastes, numerous usages have been proposed and tested: broiler breeding [1], growing substrate or source of specific compounds for microorganisms [3,5], biosurfactants [6], and biodiesel production among others. But soybean waste and meal particularly have not been exploited for the production of polyphenols. Soybean (Glycine max) seeds have been reported as one of the richest flavonoid legume sources known nowadays; with up to 3 mg/g dry weight [7], these compounds are not totally destroyed during the different treatments that seeds undergo and may find themselves in soybean meal discarded by industries. Phenolic compounds have multiple applications in nutraceutical and pharmaceutical domains [8]. Flavonoids, for example, and isoflavones precisely are particularly abundant in soybean seeds and have long been exploited for their anti-oxidants, anti-inflammatory, and anti-cancer activities, among others [9]. Exigencies of green chemistry nowadays appeal to researchers to use ''green'' processes in science [10]. Recently, the use of supercritical fluid, ultrasound-assisted extraction (UAE), and microwave-assisted extraction (MAE) methods for extraction of phenols has contributed enough to reduce the cost of production and pollution and improve extraction yield [11,12]. MAE is said to be the most effective in terms of polyphenol yield, accessibility, greenest, and respect to the environment [13,14]. Literature has shown that yields of MAE are conditioned by many factors including the nature of the extracting solvent, time of extraction, power of the equipment, dry matter/solvent ratio, and the nature of the matrix [15]. So the determination of experimental conditions for extraction of the highest polyphenolic content from soybean meal using a green protocol will be helpful, since these seeds are among the richest sources of the named compounds. is study aimed at determining the optimal conditions for extraction of phenolic compounds from soybean meal, a soybean seed industry waste product, using a face-centered composite design in the response surface methodology with MAE. e meal was brought to the laboratory in a plastic bag and dried to constant weight at 45°C using an air oven.

Screening of Factors Affecting the Phenol and Flavonoid
Contents. On the basis of the literature, variables retained for screening were: time of extraction, dry matter/solvent ratio, proportion of ethanol, and power [16,17]. Factors influencing the yields of phenols and flavonoids were determined from Table 1.

Extraction of Phenolic
Compounds. For each trial, 1 g of soybean meal was mixed in a beaker with the appropriate amount of solvent according to the experimental conditions as given by the chosen design. e mixture was stirred using a magnetic agitator; afterward, it was allowed to rest for 10 minutes at room temperature and put in a microwave oven (SAMSUNG M735) for extraction, under specified conditions. Samples were centrifuged (4,000 rpm/5 min), and the supernatant was collected after filtration through Watman paper no. 1. e solvent was then evaporated in an air oven at 45°C until obtention of the dry extract. Dry extracts were immediately used for the determination of total phenolic and flavonoid contents.

Determination of Total Phenolic Content.
e total phenolic content was assessed according to the method proposed by Dohou et al. [1]. Briefly, 0.2 ml of Folin-Ciocalteu reagent (tenfold diluted) was added to a tube containing 0.01 ml of plant extract (5 mg/ml) and 1.39 ml of distilled water. e mixture was allowed to stand for 3 minutes before adding 0.4 ml of sodium carbonate (20% w/v) and then mixed using a vortex. e tube was then incubated at 40°C for 20 min in a water bath, and absorbance was read at 760 nm against a blank using a BIOMATE spectrophotometer. Gallic acid (0.2 g/l) was used to draw a calibration curve. All experiments were carried out in triplicate, and results were expressed as mg of gallic acid equivalent (GAE) per gram of dry extract (mg GAE/g dry weight).

Determination of Flavonoid Content.
Flavonoid content was obtained using the method described by Padmadja et al. [18]. A volume (0.03 ml) of sodium nitrite (5%) was added to a tube containing 1.49 ml of water and 0.1 ml of extract solution (5 g/ml). After 5 min, a volume (0.003 ml) of aluminum chloride (10%) was added to the tube, and the mixture was allowed to rest for 6 min. Afterward, 0.3 ml of NaOH (1 M) and 0.24 ml of water were introduced, respectively, in the tube and mixed with a vortex before absorbance was read using a BIOMATE spectrophotometer at 510 nm against a blank. Calibration curve was made using catechin. All experiments were made in triplicate, and the results were expressed as mg of catechin equivalent per g of dry extract (mg CE/g of dry weight). Figure 1 depicts the flowchart of the whole work.

Optimization of the Responses
Using the Central Composite Design. Factors such as time of extraction, dry matter/solvent ratio, and power were observed to influence the responses. A face-centered composite design was used, and the studied responses were total phenolic content (Y 1 ) and flavonoid content (Y 2 ). Ranges of different factors were taken according to the results of preliminary experiments. Experiments were randomized, and responses were evaluated in triplicate. e proposed model is as follows: where Y is the response (total phenolicor flavonoid content); X 1 , X 2 , and X 3 are the studied factors; a 0 is the offset term, while a 1 , a 2 , and a 3 are linear effects; a 11 , a 22 , and a 33 are the quadratic effects; and a 12 , a 13 , a 14 , a 23 , and a 34 are interaction effects. All experiments carried out are summarized in Table 2. Ranges are as follows: solvent/seed ratio

Fourier-Transform Infrared Spectroscopy (FTIR).
Soybean meal extract was analyzed by a Fourier-transform infrared (FTIR) apparatus coupled with an attenuated total reflectance (ATR) accessory. Spectra were collected at frequency regions of 4,000-400 cm −1 using an FTIR spectrometer (Alpha, Bruker, Germany) on a diamond plate at a resolution of 4 cm −1 . Two replicates spectra of 50 scans were recorded. Raw spectra were corrected.

2.4.2.
in-Layer Chromatography (TLC). TLC was performed using a precoated plate covered with 60F250 silica gel (MERCK). Plates of 5 × 8 cm were used and activated at 105°C for 30 min. Two milligram of the extract were dissolved in 10 ml water and centrifuged for 10 min at 3,500 rpm. Ten microliter of supernatant was deposited on the plates using a capillary tube. Two standards were used, namely catechin and quercetin (1 mg dissolved in 50 ml of ethanol, centrifuged, and supernatant used). Development was done for 20 min in a presaturated (30 min) rectangular development chamber. e mobile phase was made of ethyl acetate/formic acid/glacial acetic acid/water (96/11/11/30). After development (6.5 cm), the plate was removed and dried at 45°C in an air oven for 1 min, before visualization under UV light (254 nm). Bands were circled, and Rf calculated.

Anti-Oxidant
Assay. DPPH radical scavenging activity of the extract was determined using the method described by Mensor et al. [19], while the ferric reducing assay power (FRAP) was performed as described by Oyaizu [20].

Statistical Analysis
Designing and analysis of the results were done using Minitab 18. Experiments were carried out in triplicate. Statistical significance of the model variables was determined at a 5% probability level. Main effects and contour plots were plotted using Sigma Plot v11.0 (c) systat. Data on phenol and flavonoid contents were expressed as mean ± SD and compared using the Bonferroni test with the software SPSS version 22.

Screening Factors.
Results of the screening factors are indicated in Table 3: usage of high power leads to a reduction in both the total phenolic and flavonoid contents of the extracts. Long cooking time almost induced a linear reduction of the two studied responses. Short times were seen to be best for the highest responses. Variation of ethanol percentage in the extracting solvent showed the highest TPC and TFC at 50%. at observation means water/ethanol (50/50) may be more effective in extracting polyphenol from soybean meal. We noticed the effect of the dry matter/solvent ratio that any increase in the solvent proportion leads to an increase in total phenolic and flavonoid contents.

Optimization of the Responses Using the Central Composite Design.
Results of the screening permitted the optimization of the process using three main factors for extraction of phenolic compounds from soybean meal, namely the microwave power, the cooking time, and the solvent/dry matter ratio. Ethanol proportion was decided to be 50%, since we noticed a maximal production with 50% ethanol proportion (in the solvent) of TPC and TFC during the screening. Also, the literature indicates that dielectric properties of the solvent should be highly taken into consideration when planning to extract phenolic compounds using MAE. Compared to water, ethanol or its mixtures with water have a lower dielectric constant and are more transparent to microwave, thus not well converting them into heat. But its high capacity to dissolve and extract phenolic compounds [21][22][23] and its greenness oriented the choice of a hydro-ethanolic solution as extracting solvent.    Figure 2 shows that an increase in the solvent ratio induces an almost linear reduction in the total phenolic content of extracts when going from 60/1 to 80/1 (ml of solvent/g DW). But the responses measured started increasing as the solvent ratio passed from 90:1 to greater values. Such observations may be explained by the type of phenols extracted at each condition, since literature indicates that free or bound phenolic compounds are found in soybean and are not forcibly extracted in the same experimental conditions [16]. Nevertheless, it is well accepted nowadays that high solvent content increases mobility of compounds (mass transfer) from plants matrix, thus explaining the observed increase in the total phenolic content of extracts at a certain solvent ratio, since previous research had already reported that [27]. An increase in the solvent ratio only led to a progressive diminution of the flavonoid content.   Figures 2 and 3 depict the influence of power on the total phenolic and flavonoid contents of extracts. We can see from the figures that an increase in the power induces a reduction in the phenol and flavonoid contents of the meal extracts. is could be the consequence of degradation of these compounds exposed to high temperature, since high power in microwave induce a quick elevation of the solvent temperature even when exposure is for a short duration. Ðurović et al. [16] made a similar observation. Figures 2 and 3, we can see that any increase in time of exposure also led to a diminution of the TPC and TFC of extracts because of progressive destruction of these thermo-sensitive compounds under long exposure to heat. Previous authors also noticed the same effect [16,22,28]. Table 5 shows the ANOVA and the influence of each independent factor. We can see from the table that all independent factors significantly (p < 0.05) influenced both total phenolic and flavonoid contents. Quadratic effects of solvent ratio (X 1 X 1 ) and the power (X 3 X 3 ) significantly affected the total phenolic content of the extract obtained, while only the quadratic effect of the heating power (X 3 X 3 ) significantly influenced the flavonoid content of the extracts. Interaction between solvent ratio and the heating time (X 1 X 2 ) significantly impacted both the total phenolic and the flavonoid content, while only interaction between the time and the boiling microwave power (X 2 X 3 ) significantly affected the flavonoid content of the extracts. We can also see that interaction between solvent ratio and the heating time (X 1 X 3 ) contributed the most in the observed phenolic response (24.74%), followed by the quadratic effect of the solvent ratio (X 1 X 1 ), which contributed up to 22.17% to the final response. Talking about the phenolic content, time (X 3 ) and interaction between power and time (X 2 X 3 ) contributed the most to the observed response (30.06% and 18.33%, respectively). e mathematical model predicting the influence of the solvent ratio, boiling time, and working power on the phenol and flavonoid contents of the extracts is given by the following equation:

Assessment of Model Quality and Optimal Conditions
Experimental values show us that these mathematical models can well explain the observed results. According to Joglekar and May [29], a good mathematical model should predict at least 75% of the responses; R 2 should then range from 0.75 to 1. Our results give the determination coefficient for phenols and flavonoid, respectively, to be 0.95 and 0.94,  Journal of Chemistry falling in the good range, which means our second-order polynomial equations, really represented the experimental data. Also, obtaining values of AADM (analysis of the absolute average deviation) and Bf (Bial factor), respectively, equal to 0 and 0.99 for both total phenolic and flavonoid contents, thus confirming the suitability of the models since values were in the normal range (0 for AADM and 0.75 < Bf < 1.25 for Bf ).

Optimization of the Process
After validation of the model, the optimal extraction conditions for total phenolic and flavonoid contents were determined using iso-responses and responses surfaces curves.

Confirmation Experiments
In order to confirm the quality of our model to predict the optimal conditions for our responses, experiments were made, replicating optimal conditions and results compared with the predicted maximal values in Table 6. No significant differences were noticed between optimal predicted values and experimental values obtained for TPC and TFC, thus confirming the validity of the predicted optimal values given by the software.

IR Spectral Analysis.
e IR spectra of Glycine max seed extract is depicted in Figure 5. Table 7 shows the different absorbance peaks and their assignment. Polyphenols were identified among the compounds present in the extract.

in-Layer Chromatography.
in-layer chromatography of Glycine max seed extract showed a number of nine bands with Rf going from 0.15 to 0.98. Two spots were identified as catechin and quercetin, respectively, those with Rf of 0.98 and 0.93. Figure 6 shows the plates with respective spots under UV light and the schematic representation of the plate after development.
TLC revealed the presence of at least nine different compounds or groups of compounds in the obtained extract, with a high relative abundance of quercetin (8). A similar observation was made by Hanan et al. [39] who stated that quercetin is the most abundant flavonoid found in soybean seeds. Rf values of all spots observed under UV light are given in Table 8    In plane -C-OH (bending) Phenols [33] by Djordjevic et al. [40] who found IC 50 > 200 µg/ml. Table 9 shows DPPH and FRAP inhibitory results of the soybean meal extract.

Conclusion
e study aimed at determining the optimal conditions for extraction of total phenolic and flavonoid compounds from soybean meal using a green protocol. RSM was used to determine the conditions, and we found out that all factors, namely, solvent ratio, time, and power significantly, influenced both responses. Results suggest that low solvent, power, and time of exposure should be used when attempting to obtain high TPC and TFC extracts from soybean meal with MAE. RSM used in this research permitted us to define the conditions for green extraction of TPC and TFC from soybean waste as: 60/1 solvent/dry matter ratio, 120 W power, and 0.16 min time. FTIR confirmed the presence of polyphenolic compounds in the extract obtained, and TLC permitted to identify catechin and quercetin, while DPPH and FRAP tests showed that the obtained extract possesses moderate anti-oxidant capacities. RSM permitted to obtain a polyphenolic rich extract, containing catechin and quercetin, with anti-oxidant capacities, usable in nutraceutical, food, cosmetic and pharmaceutical industries from seeds waste; thus, soybean meal can be well valorized as good source of polyphenols.
Data Availability e data are available on request (Woumbo Cerile Ypolyte at woumbocerile@yahoo.fr).

Conflicts of Interest
e authors declare that there are no conflicts of interest.

Authors' Contributions
Woumbo Cerile Ypolyte and Kuate Dieudonné conceived the work, collected seeds, carried out experimentations, analyzed and interpreted data, and wrote the paper. Klang Mathilde Julie followed up the work, verified data analysis, and read the paper for correction. Womeni Hilaire Macaire supervised the work and read the paper.

Acknowledgments
is research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.