Optimization of polysaccharides extraction from quince peels: partial characterization, antioxidant and antiproliferative properties

Abstract In this study, Box-Behnken Design was used to optimize the ultrasonic extraction of polysaccharides from quince peels (QPPs) by ascorbic acid and the effect of extraction temperature, extraction time and pH was evaluated. Under optimized conditions of temperature 90 °C, 60 min sonication time and pH = 3.26, the extraction yield, the galacturonic acid yield and the concentration of sample required to scavenge 50% of 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulphonic) acid (ABTS) values of QPPs were respectively 10.25%, 3.86% and 1.35 mg/mL. The QPPs extracted under optimum conditions was characterized by Fourier transform infrared spectroscopy (FTIR), Nuclear magnetic resonance (1 H NMR) and Size exclusion chromatography (SEC/MALS/VD/DRI). The monosaccharide analysis revealed that arabinose was the most abundant, followed by galactose, glucose, mannose and xylose. Moreover, QPPs showed significant antioxidant activities (2,2-diphenyl-1-picrylhydrazyl (DPPH) and Ferric- reducing antioxidant power (FRAP)) and reduced viability of human Caco-2 and murine B-16 cell lines in a dose-dependent manner. Hence QPPs could be used as antitumor agent in functional foods andpharmaceutical industries. Graphical Abstract


Introduction
Nowadays, pollution, stress, unbalanced diet and the consumption of fast food increased the possibility of getting cancer. In 2012, the number of cancer mortality was 8.2 million (Ferlay et al. 2015). The majority of drugs used in the treatment of cancer are organic therapy and chemotherapy which have several undesirable effects. In order to limit the side effects of these pharmaceuticals, researchers are interested to replace these drugs by naturally occurring ones such as polysaccharides mostly known for their antiproliferative activity and antioxidative properties (Zhou et al. 2008;Yu et al. 2017;Pandya et al. 2018). The quince (Cydonia Oblonga, Rosaceae family) is abundant in Tunisia and was used in the folk medicine for the treatment of diarrhea, gastric reflux, cardiovascular illnesses, diabetes and obesity. Otherwise, it was shown that quince contained large amounts of polysaccharides (11 g/100 g dry fruit) (Thomas et al. 2000). Thus, this paper reported the optimization of the extraction and the preliminary characterization of polysaccharides from fresh quince peels using ascorbic acid and response surface methodology. Moreover, the antiproliferative properties of extracted polysaccharides against B-16 and Caco-2 cells in vitro, as well as, the antioxidant activities were performed.

Statistical analysis
The optimization of the extraction condition using the response surface methodology and Box-Behnken design (BBD) was performed. Table S1 presents the predicted and experimental polysaccharides yield, galacturonic acid yield and ABTS scavenging activity of all the experimental runs and Table S2 presents the corresponding p value of different independent variables as well as their regression coefficients. The data showed that for all responses, the independent variables including extraction temperature (X 1 ), extraction time (X 2 ) and extraction pH (X 3 ) have significant quadratic model (p < 0.05). Also, for the three responses (Y, YGA, Y IC50 ), all the linear coefficients were significant with p value less than 0.05. In summary, ANOVA for the lack of fit test for the three latter responses was not significant (p > 0.05), which confirmed the validity of the model. Moreover, all the models had satisfactory multiple determination coefficients (R 2 ) and the values obtained were 0.889, 0.920, and 0.975 for the response extraction yield of QPPs (Y), yield of galacturonic acid (YGA) and ABTS scavenging ability (Y IC50 ), respectively. The extraction yield of QPPs values ranged from 6.1% to 10.1%. The yield was increased by lowering the pH and by increasing both temperatures and times (Table S1). The obtained model is given by the polymonial equation: Y ¼ 8:170 þ 1:790X 1 þ 0:670X 2 À3:750X 3 À0:089X 2 1 þ 0:019X 2 2 À0:016X 2 3 À0:180X1X 3 þ 0:047X 2 X 3 As shown in Table S2, (F tabulated(9,4,0.05) ¼ 3.85 <F regression ¼ 11.02), (F tabulated(3,4,0.05) ¼ 3.75 > F lack-of-fit ¼ 2) and (p < 0.05).
The yield of galacturonic acid (YGA) values ranged from 0.52% to 3.80%. The fitted model for the content of the galacturonic acid is given by Equation.
The equation Y IC50 presented the model that describes the relationship between the response of ABTS scavenging activity and the significant independent variables.
According to the data in Table S2, (F tabulated(7,9,0.05) ¼ 3.88 <F regression ¼ 5.61) and the p value was 0.0166. These results indicated that the model is significant. Two dimensional contours and three-dimensional response surface were plotted according to the model as shown in Fig. S1 representing the effects of combined factors including extraction time and temperature when the pH was fixed at 3.
The optimum conditions for the three independent variables were marked in the area by a point and were found as follow: a temperature of 90 C, an extraction time of 60 min and a pH of 3.26.

Partial characterization of optimum polysaccharides extracted from quince peels
In this study, the percentage of total sugar content of QPPs extracted is 82.15% (Table S3).
The value of the uronic acid content in QPPs was 23.25%. This value is low compared to the value found by Thomas and Thibault (2002) which was superior to 38%. This result could be due to the nature of the harvesting environment and the conditions of extraction. Moreover, QPPs is a polysaccharide with appreciable purity since it contains no polyphenols and a quantity of protein less than 0.1%.
The Fourier transform-infrared spectroscopy (FTIR) spectrum of the extracted QPPs was depicted in Fig. S2, which is comparable to the spectrum of commercial apple pectin CAP.
A large absorption band was detected at around 3383 cm À1 , which is characteristic of hydroxyl group (O-H) (Grassino et al. 2016). Also, the C¼O stretching vibration of Oacetyl groups was assigned to the band at 1737 cm À1 , and the band at 1609 cm À1 was referred to the carboxylate stretching vibration (COO-). The band detected at 1440 cm À1 confirmed the presence of galacturonic acid (Grassino et al. 2016). Furthermore, the band at 1235 cm À1 is due to the stretching vibration of non-symmetrical C-O-C (Grassino 2016). In summary, the data from FTIR spectrum corroborated that the extracted QPPs is pectin-like structure.
The structural feature of QPPs was analyzed by NMR spectra (Fig. S3) and showed the signal of anomeric protons at d H 5.166-5.546 ppm, which indicated that the type of the bond was a-glycosidic (Chen et al. 2016). Compared to the literature, five signals at d H 5.16 ppm, 5.26 ppm, 5.40 ppm and 5.54 ppm may be attributed to the existence of a-Glcp, a-Galp, a-Araf, and a-Manp, respectively (Chen et al. 2016). Finally, the signal at 2.41 ppm could refer to the -CH 3 of the O-acetyl groups. These results could indicate that the QPPs is pectin-like structure.
The results determined by SEC give an average Mw of 134 K Da with a dispersity of 1.6 and a high intrinsic viscosity ([g] ¼ 330 mL.g À1 ) (Table S3). These results confirmed that QPPs is in the form of an expanded random coil.
Thus, the monosaccharide composition indicated that QPPs was composed of arabinose, galactose, glucose, xylose and mannose with a percentage of 40.28%, 22.58%, 19.90%, 8.68% and 8.56%, respectively (Table S3 and Fig. S4). These results are consistent with those found by Thomas and Thibault (2002). Besides, due to the arabinose and galactose abundance we notice that the extracted QPPs is probably a galactoarabinan-I-rich pectin. Figure S5. A depicted, in comparison with AA (Ascorbic Acid), the ability of QPPs to scavenge DPPH radicals is proven at different concentrations. Indeed, DPPH scavenging by QPPs polysaccharides increased from 25.93% to 51.79% with increasing concentrations from 0.6 mg/mL to 5 mg/mL. The chelating activity of QPPs as showed in Fig. S5.B rise with increasing concentrations and ranged from 623 mmol/L at 0.6 mg/mL to 1791.9 mmol/L at 10 mg/mL. In comparison with the literature, the assays suggested that the QPPs had a good reducing activity in FRAP (Chen et al. 2016;Liang et al. 2018).

Antiproliferative activity
In this study, the obtained results showed that QPPs revealed a significant activity against against B-16 and Caco-2 cancer cell lines. Fig. S5.C indicated that QPPs enhanced the antiproliferative activity which increased with increasing concentrations. Indeed, the proliferation inhibition ratios of QPPs at concentration of 4 lg/mL on B-16 and Caco-2 cells were up to 30.66 ± 1.61%, 28.59 ± 0.88%, respectively, however, at concentration of 400 lg/mL on B-16 and Caco-2 cells were up to 59.17 ± 1.43%, 87.54 ± 0.75%, respectively. It has been deducted that QPPs is more active in Caco-2 cell line than B-16 cancer cell line.

Conclusion
The optimization of the extraction condition of polysaccharides from quince peels using the response surface methodology and Box-Behnken design (BBD) was performed. The optimum conditions of extraction of QPPs were: extraction temperature, 90 C; extraction time, 60 min; and pH, 3.26. Under these, the experimental values of extracted polysaccharides yield, galacturonic acid yield and ABTS scavenging activity (IC 50 ) were 10.25%, 3.86% and 1.35 mg/mL, respectively. Moreover, the Preliminary characterization suggested that QPPs is probably galactoarabinan-I-rich pectin like. Finally, QPPs can be used as a product in functional medicines due to their good antioxidant ability and their high antiproliferative activity against B-16 and Caco-2 cells in vitro.