Optimization The Effect of Roasting Conditions by Central Composite Design (CCD) Method on The Antioxidant Compounds of Opuntia Ficus Indica Seeds From Morocco

The effect of roasting conditions on antioxidant activity Opuntia Ficus Indica seeds from Morocco and their phenolic compounds were optimized by the Central Composite Design (CDD) method. The CCD was used to optimize the roasting conditions: temperature (X 1 : 60–200 °C) and roasting time (X 2 : 10–50 min). The best roasting conditions were used in order to optimize the optimal value of the response : TPC (Total Phenolic Content): X1: 200°C, X2: 50min with 104.86±1.94GAE/g extract predicted response, TFC (Total Flavonoids Content) : X1: 200 °C, X2: 50 min with 81.23±0.90 mg QE/g extract predicted responses, TTC (Total Condensed Tannins Content): X1: 128.9°C, X2: 34.92min with 6.12±0.95 mg QAE/g extract) predicted response. Moreover, the optimal potential antioxidant activity by ABTS assay and DPPH assay was found in extracts of Opuntia Ficus Indica seeds roasted: at a high temperature of the roasting. Furthermore, the positive signicant correlations were determined by the principal component analysis (PCA) on the one hand, between the antioxidant capacity of the extracts and their antioxidants compounds (TPC and TFC), and on the other hand, between the two assays ABTS and DPPH. Consequently, the results of this work indicated that roasted Opuntia Ficus Indica seeds from Morocco can be considered as an essential ingredient for human foods.


Introduction
The Opuntia cactus is specie originally from Mexico (Mulas, 1992) and considered an important food for the indigenous populations (Barbera, Inglese, & Pimienta- Barrios, 1995). It is a specie of the Cactaceae family, the it is distributed in all continents as well as cultivated especially in the arid and semi-arid regions such as Africa and the Mediterranean region, South and Central America (Benayad et al., 2014).
Moreover, the prickly pear fruit are rich in sugar, vitamin C, minerals, antioxidant compounds, and pigments, consequently, it was recommended in the human diet (Di Cagno et al., 2016). Furthermore, the seeds from Opuntia sp are considered rich in nutritional value, such as polyphenol, tannins, avonoids, and fatty acids, in addition to that, the doses of these compounds are higher than in the pulp of the fruit, these fruits are eaten fresh with their seeds (Al Juhaimi et al., 2018). Also, the nutritional value and the chemical compounds of the oils of the seeds can be modi ed by roasting, this method is considered as a pretreatment which is done before extraction (Gao et al., 2019), and it can render the color, texture, and acceptability of grilled products (Khan & Saini, 2016). In addition to that, various studies showed that roasting pretreatment modi es the phenolic pro le of the seeds; it can improve the health bene t effects  (Zhang et al., 2011). In the present work, we want to apply the Central Composite Design (CDD) approach, in order to investigate the effect of roasting treatment: roasting temperature, roasting time to maximize the contents of antioxidants compounds, and the antioxidant activity of opuntia cus idica seeds. The main objective of this paper, we applied the CDD approach, in order to assess the effect of roasting treatment : roasting temperature and roasting time for to maximize the TPC(Total Phenolic Content),TFC (Total Flavonoids Content ),TTC(Total Condensed Tannins Content), and the antioxidant properties by ABTS and DPPH assays for opuntia cus indica from Morocco, Signi cant correlations between the antioxidant compounds and the antioxidant activities of Opuntia cus-Indica seeds were evaluated by principal component analysis (PCA).

Plant materials
The plant material (Opuntia cus-indica seeds) studied, were collected in the period between June and July 2019, in the region of Taza located in the East-North of Morocco.

Preparations of extracts
After the harvested of the fruits, the seeds were isolated, and then dried in the dark at room temperature for 72 hours, and they have been placed in an aluminum Petri dish (7 cm diameter) and roasted in a forced hot-air convection oven at 60, 130 and 200 °C for 10, 30 and 50 min. After that, the seeds will be crushed using a grinder, this ne powder will then be used for the preparation of the various extracts.
Moreover, 40 grams of ne powder was macerated with the ethanol solvent for 48 hours. After that, the solvent was evaporated using a rotary evaporator. The extracts obtained are preserved at a temperature of about -4 ° C until the subsequent analyzes.

Determination of Total Phenolic Content (TPC)
The TPC of these extracts seeds was quanti ed by the method of singleton (Singleton, Orthofer, & Lamuela-Raventós, 1999). 200 μl of OFI seed extracts, add a volume of 1.5 ml of Folin Ciocalteu reagent (diluted 10 times). After 4 minutes, a volume of 1.5 ml of 5% sodium carbonate (Na 2 CO 3 ) was added on to the solution. The tubes were placed in darkness. After two hours, Gallic acid was used as a standard for the calibration curve. The results were read on a spectrophotometer at 750 nm. The concentration of total polyphenols is deduced according to a calibration interval established with Gallic acid (0-100 μg / ml), as well as are presented in milligrams equivalent of a Gallic acid gram of extract (mg EGA / g extract).
Determination of Total Flavonoids Content (TFC) The TFC was determined according to the method discussed by yeddes et al (Yeddes, Chérif, Guyot, Sotin, & Ayadi, 2013). 1 ml of extract is added to 1 ml of a solution of AlCl 3 (2% in methanol). After ten minutes of reaction, the absorbance is read at 430 nm. The avonoid content is determined using a linear regression equation deduced from the calibration curve and expressed in milligrams equivalent of Quercetin per gram of extract (mg EQ / g extract). The avonoid concentration is deduced from a calibration range established with Quercetin (0-100 μg / ml).

Determination of Total Condensed Tannins Content (TTC)
The TTC was done by the method described by the method described by sun et al (Sun, Ricardo-da-Silva, Spranger, & chemistry, 1998).50 mL of extracts seeds (50-600 mg / mL) was added to 3 mL of 4% methanolic vanillin solution and 1.5 mL of H 2 SO 4 . The absorbance was read at 430 nm after 15 minutes.
The Catechin calibration curve was in the range of 50-600 mg / mL. TTC was presented as mg Catechin equivalent (CE) per gram of extract Radical scavenging activity of DPPH1(1-diphenyl picrylhydrazyl) The free radical removal capacity of the extracts was evaluated by the method of Grzegorczyk et al (Grzegorczyk, Matkowski, & Wysokińska, 2007). One ml for each extract at different concentrations (50 to 1000µg / ml) was mixed with one ml of a methanolic solution of DPPH at 0.1 mM and let sit for 30 min at 27 ° C. Methanol and DPPH were used as controls. After incubation at 37 ° C in the dark for 20 min, the absorbance was read at 517 nm. The antiradical capacity was quanti ed according to the following equation: % Radical scavenging activity DPPH = 1-[A sample / A control ] × 100, where A sample and A control were the absorbance of the sample and control. The results obtained for each extract tested are compared with those obtained for ascorbic acid taken as standard antioxidant.

Radical cation inhibition activity (ABTS)
The radical cation capacity of OFI seed extracts was determined by the method described by Yim el al (Yim et al., 2013). 88µL of 140mM of potassium persulfate (K 2 S 2 O 8 ) was added to 5ml of 7mM ABTS .+ solution. After that, the whole was stored in the dark at room temperature for 16h.Then,the absorbance of the ABTS .+ mixture was adjusted by ethanol to 0.70 ±0.05 at 734nm.10µ of OFI seed extracts at different concentration was mixed with 1 ml of ABTS reagent (100 to 1000µg / ml).The absorbance was read against the blank reagent at 734nm.The inhibition capacity was quanti ed uation according to the following equation:: % Radical inhibition activity ABTS= 1-[A extract / A control ] × 100, where A extract and A control were the absorbance of the extract and control

Experimental design
In this study we used the CDD method, this method consists of 11 experimental assays (Table2) was employed for the optimization of roasting variables. The independent variables include roasting temperature and roasting time. These variable had a three levels (-1.0.+1) which are lower, medium, and higher (Table 1). TPC, TFC and TTC and antioxidant activity by ABTS assay and DPPH assay were selected such as the response of model design(Y) in this study; they are presented in Table 2. The regression coe cients (β) were obtained by the adjustment the experimental results to a second order polynomial model; this model was used to response analysis surface as below.
Where, Y is the response variable, X i and X j are the independent variables. β 0, β i,, β ii , and β ij are the   Data Analysis Table 2 present the results of the responses TPC, TFC, and TTC, DPPH, and ABTS of extracts from opuntia cus indica seeds .This optimization of the roasting conditions was achieved in eleven randomized trials in order to evaluate the effects of different roasting factures on the studied responses. Central Composite Design (CCD) was done to the optimization of roasting conditions for Opuntia Ficus Indica seeds using, it was analyzed by JMP 11(SW) software. Moreover, the XLSTAT 2014 software (Zielinski et al., 2014) was used to determine the Pearson correlation between all data responses and for a graphical representation (PCA). Software IBM SPSS Statistics 21 was used to express the data in means ± standard error of the mean. The signi cation of data was done by the Tukey test at alpha = 0.05.

Page 7/27
The data are presented in the form of the average of two individual repetitions (n = 2e ± SEM), the means followed by similar letters exposing in the same colum are not different (P < 0.05  its quadratics effects X 2* X 2 were shown to have a signi cant positive effect respectively (p-value < 0.05).Furthermore, the interaction effect between the two parameters studied X 1 * X 2 had signi cant (pvalue < 0.05) on TPC.

Response Surface Methodology (RSM) analysis
The effects of both the roasting temperature and the roasting time and their reciprocal interactions on TPC can be visualized on the generating 3D response surface plots shown in Fig. 1. According to Fig. 1, the TPC content increased when X 1 roasting temperature (°C) increased at a roasting time xed, also it increased rapidly when the X2 roasting time (min)exceeds 30 min. Thus, the maximal extraction of TPC was found at the strong levels of both the roasting temperature (X 1 ) and roasting time (X 2 ). However, to get optimization overall of all variables studied. Optimization of the response was used by desirability function (d) in order to obtain the maximum response in TPC of the opuntia cus indica seeds, thereby the maximum response precision is obtained when the desirability close to 1 ( Figure.6  Roasting time(X 2 ) had a positive signi cant linear effect (p-value = 0.0011 < 0.05) on TFC, and it doesn't have a signi cant quadratic effect (p-value = 0.2944).As well as, the Roasting temperature X1 had signi cant positive linear (p-value = 0.0084), but its quadratic effect is not signi cant (p-value = 0.2944). Moreover, the not signi cant interaction effect between the two parameters studied was observed (Table 4). Response Surface Methodology (RSM) analysis Figure 2 shows the response surface plot of roasting temperature and roasting time on total avonoid content. The TFC increased with the increased both the roasting time and the roasting temperature. Accordingly, the higher TFC yield was detected in regions of high roasting temperature and hard roasting time. Consequently, the optimum extraction of TFC was at: roasting temperature 200°Cand roasting time 50 min, and it was assigned for the predicted response is 80.22623 (mg QE/g extract) with the desirability is d=0.84( Figure.6.b.). Our results are similar with various studies as, Lin et al (Lin et al., 2016) reported that the TFC increased signi cantly after 5min of the roasting, as well as, the avonoid aglycones and acids are increased according to roasting temperature and time. Furthermore, Kumar et al (Kumar & Pandey, 2013) mentioned that, the fractions of sugar in avonoids glycosides have an important role in antioxidants capacity, as well as, the aglcycones had a high effect on the antioxidant capacity more than the glycosides. Interpretation of the response surface model of TTC Second-order polynomial model Table 5 shows the coe cients of regression and their signi cance for the TTC yield. The regression model was signi cant (p-value = 0.0231). Also, the determination coe cients (R 2 ) for the TTC response variable (0.881256) and the lack-of-t values (0.0537) were not signi cant (P > 0.05), which indicates that the model can explain all data. So the response variable was included in roasting optimization. Besides, the R 2 adjs was 0.762512,it indicated that 76.25% of the variability was estimated by the model. Therefore, the second-order polynomial model was applied (Eq. 4). TTC = 5.8052632-0.133333X 1 0,75X 2 -0,4X 1* X 2 -3.463158X 1* X 1 -1.513158X 2* X 2 (Eq. 4) According to p-value < 0.05, the X 1 * X 1 is the quadratic effect of roasting temperature was positive signi cant for TTC, on the contrary, its linear effect did not have signi cant because of the p-value = 0.791. As well as, the X2 and X2 * X 2 of roasting time were not had signi cant in TTC because, their pvalue was respectively: 0.1768, 0.0944. The X 1 * X 2 had also not signi cant according to its p-value was 0.5241.

Response Surface Methodology (RSM) analysis
The 3D of response surface of regression Eq. (3) were constructed using RSM to illustrate the effects of the roasting temperature and roasting time and their interaction on the TTC (Fig. 3). Accordingly, the TTC content increased before the roasting temperature increased at 130 °C, after that it decreased quickly. Also, the TTC increased with the roasting time in the range of 10 to 35 min and then decreased.. The optimum extraction of TTC was roasting temperature 128.9 °C, roasting time 34.92 min with 5.901 (mg QAE/g extract) predicted responses and the desirability is d = 0.82( Figure.6.c.).These results are similar to these reported by Lin et al (Lin et al., 2016), they showed that, during the roasting at 200 °C for 20 min the content of condensed tannins from ethanol extracts had a high levels in TTC. Interpretation of the response surface model of DPPH assay Second-order polynomial model The ANOVA results from DPPH assay content based on the RSM design are reported in Table 6.The pvalue of the model was (< 0.0001),which indicated that the model was signi cant. Moreover, the R 2 and R 2 adj were 0.997921 and 0.995841 respectively, that con rms the adequacy of the model because R 2 > 0.75 (LI et al., 2019). Additionally, the lack of t (p-value > 0.05) con rms also the adequacy the model for prediction of the antioxidant power for opuntia cus indica seeds roasted. Therefore, the second-order polynomial model was applied (Eq. 5).

Response Surface Methodology (RSM) analysis
The response surface (3D) of regression Eq. (5) were constructed using RSM to illustrate the effects of X1 and X 2 and their interactionX 1 *X 2 on the IC 50 of DPPH assay (Fig. 4.). We know that the antioxidant power is inversely proportional with the value of the IC 50 . We observed that the increase in the antioxidant power was made thanks to the increasing of the X 2 roasting time and the increasing of the roasting temperature (X 1 ). The optimum of the antioxidant power by DPPH assay observed at roasting temperature 200 °C and roasting time 50 min with 96.60% of inhibition, which matches 86.3845 µg/ml predicted responses, and also the desirability is d = 0.93( Figure.6.d.).Our results con rm those founded by Lin et al (Lin et al., 2016), they reported that, the higher antioxidants capacity was observed at strong roasting temperature for ethanol extracts from almond(Prnus duclis)kernel, and during the roasting at 200 °C for 20 min the power for scavenging DPPH radical was strong than the raw sample. Moreover, Chandrasekra et al (Chandrasekara et al., 2011), reported that in their study, the scavenging capacity of DPPH radical increased signi cantly with the increase of the roasting temperature for soluble phenolic extract from testa, as well as, they justi ed that increase due to Maillard reaction products MRPs. Indeed, during at roasting, a reaction between the reducing sugars and amino acids can be done, this reaction can produce the new compounds, which are Maillard reaction products MRPs, these formed products can contribute to TPC, avour, antioxidant activity and color of food (Chandrasekara et al., 2011). Interpretation of the response surface model of ABTS assay Second-order polynomial model Experimental modeling results for antioxidant power by ABTS assay were shown in Table 7. From the model analysis, the R 2 and R 2 adj of the model were 0.99075, 0.9815 respectively, also did not present lack of t (p-value = 0. 0528). Moreover, the model was signi cant because its p-value was < 0.0001,which showed that the model equation was acceptable to predict the antioxidant power by ABTS assay. this model equation is shown in Eq. 6 as follows: ABTS (IC 50 ) =415.29474-135,2122X 1 -107.6177X 2+ 23.584X 1* X 2 -68.64034X 1* X 1 +9.1731579X 2* X 2 (Eq 6) Roasting time (X 1 ) and roasting temperature (X 2 ) had a negative signi cant linear effect (p < 0.05). Also, the quadratic effect of the roasting temperature X1 * X1 had signi cant effect (p-value = 0.0022), and the interaction effect between the two parameters roasting was not signi cant (Table 7).
Response Surface Methodology (RSM) analysis Figure 5 shows the IC 50 of ABTS assay, we observed that the antioxidant power increase signi cantly when the roasting temperature X 1 and roasting time X 2 increased, because the IC 50 decreased. According to, the more IC 50 decreases the more the antioxidant power increases. The optimum of the antioxidant power by ABTS assay was at: roasting temperature 200°C and roasting time 50 min with 96.75% of inhibition, which matches 130.581 µg/ml predicted responses, and also the desirability is d=0.89 ( Figure.6.e.). These results are similar to several works as Gao et al (Gao et al., 2019) mentioned that, the ABTS capacity was increased signi cantly during the roasting at 160°C for 10min compared to raw sample. Also Yin et al (Yin et al., 2019) reported that the ABTS scavenging increased during the heating between 130°C-140°C after 60min.Moreover, these results can depend on several conditions such as, the plants have a bound antioxidant phenol and bound polymeric compounds, during the thermal treatment these molecules can be degraded and released which leads to an increase in the antioxidants activity (Lee, Kim, Kim, Jang, & chemistry, 2002). Furthermore, after the antioxidants characteristics can be improved thanks to the degradation of the heat-labile antioxidants compounds or training the new compounds by Maillard reaction (Nicoli, Anese, Parpinel, & Technology, 1999). Also, the solubility of nonphenolic molecules was improved by roasting (Dewanto, Wu, Adom, Liu, & chemistry, 2002).
Comparisons of predict (models) and experimental results. The veri cation experiments for ve responses such as antioxidants activity by DPPH IC 50 (µg/ml), ABTS•+ inhibition activityIC 50 (µg/ml), Total phenolic Contents (mg GAE/g extract, Total avonoids (mg QE/g extract), and Total Tannins Content (mg QAE/g extract) were reported in Table 8. These experiments were done at the conditions of responses optimal and in the experimental range. these results showed that the values of responses experimental are close to those predicted.  Table 9 showed the correlations coe cients data between all responses studied. Moreover, Table 10 presents the p-value of these correlation coe cients. Additionally, the DPPH (1/DPPH IC 50 ) and ABTS (1/ABTS IC 50 ) represent the power to inhibit DPPH free radical and ABTS. + radical respectively.  The values is bold are different from 0 at a signi cance level alpha = 0.05. According to Table 9 and Table 10, we observed that, the TPC had high positive correlations signi cant (p-value < 0.05) between the antioxidants power. The correlation coe cients of TPC were 0.949 and 0.966 with free radical scavenging effect DPPH and ABTS .+ respectively. These positive correlations are justi ed that the antioxidants capacity depends on the presence of phenolic compounds in opuntia cus indica seeds extracts,these results are similar with those reported by several study (Amri et  correlation coe cients were 0.784 and 0.727 respectively. Therefore, these results are con rmed by the strong positive correlation signi cant (p-value < 0.05) between TFC and TFC (r 2 = 0.652). We observed that, the p-values of TTC with the antioxidants power were not signi cant (p-value > 0.05), which indicates that Tannin contribute slightly in this bioactivity. Furthermore, the strong positive correlation signi cant between tow antioxidant capacity (r 2 = 0.920),indicates that the same bioactive molecules in our extracts are responsible to the scavenging power of two free radicals DPPH and ABTS .+ ..
Principal Component Analysis (PCA). According to Fig. 8, the projections of the responses studied and the experiment assays (extracts) were done by the factorial plan reported in Fig. 8.The cumulative percentage was 95.14%, which indicates that it was is representative of the variables because it was more the 50%. Moreover, the two axe are suitable for explains the all information, with the frits (F1) and second (F2) main components have explained 70.63% and 24.51 the information respectively. The correlations between all variables studied were explained by a plan formed by F1 and F2 axes. Besides, the F1 axe was formed by the positive correlation between TPC,TFC,ABTS(1/IC 50 ),DPPH(1/IC 50 )),on the contrary the F2 axe was constructed by TTC ( Fig. 7).Ours 11 extracts studied from opuntia cus indica seeds, were distributed in three groups according the responses (Fig. 8). Group I:this group was formed by four extracts (2,3,5,6) ,these extracts had a strong values of the TPC and TFC, as well as, they had also a high power antioxidants by DPPH and ABTS assays. Group II: it contains four extracts (1, 4, 10, and 11), these extracts are characterized by a strong value of TTC, and by lower values of TPC and TTC. Therefore, their antioxidants activity is lower compared to group I. Group III is formed by a three extract (7, 8 and 9), these extracts are characterized by the low values of TPC and TFC, and its antioxidant activity is low compared to extracts of the other groups.
The extracts from Group I are characterized by a high roasting temperature varies between 130 °C and 200 °C, and a high time of roasting (50 min) for the extracts roasted at 130 °C, which shows that their a strong antioxidants capacity more than the extracts from the Group II and III obtained by low roasting temperature. Therefore, the roasting makes it possible to increases the extraction of bioactive compounds responsible for antioxidants power.

Conclusion
Dry thermal processing of seeds from the Opuntia cus indica increased the amounts of active compounds. As well as, antioxidant activity power was signi cantly improved in extracts roasted especially at stronger temperature. Furthermore, the results indicated that the temperature and time of roasting had signi cant effects.PCA showed that on one hand the positive correlation between the photochemical compounds (polyphenol, avonoids) and the antioxidant capacity (ABTS, DPPH),Consequently, this study showed thermal processing can be used as a pre-treatment to increase the antioxidants capacity of Opuntia cus Indica seeds.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
Not applicable

Competing interests
The authors declare that they have no competing interests.

Not applicable
Authors' contributions C-EG, H-H, A-Z, and M-T conceived and designed the experiments. C-EG and H-EM have carried out experiments. All authors discussed the results and co-wrote the manuscript. All authors read and approved the nal paper.