Reducing the damage of quinoa saponins on human gastric mucosal cells by a heating process

Abstract Different food processing methods will influence the structure and activity of compounds. In this work, molecular structure and different content crude saponins that were extracted from quinoa, treated with water soaking, water boiling, and water steaming were analyzed by HPLC. Flow cytometry was employed to investigate the effects of the main saponins on the GES‐1 cell line. HPLC/MS analysis revealed that water soaking induced an extensive conversion of polar saponin Qc (424.41 ± 21.11 mg/g) to the less polar compound Qf (247.04 ± 15.71 mg/g). After treatment with 100 μg of Qf instead of Qc for 24 hr, the percentage of dead cells increased from 20.1 ± 2.2% to 86.2 ± 4.8%. One major reason of this result is that less polar saponins could damage membrane integrity more easier than polar saponins. The results indicate that saponin toxicity is enhanced after degradation, so it is necessary to avoid degradation before use.

It is worth noting that phytolaccagenic acid (PA) in quinoa, the main active ingredient of traditional Chinese medicine (Phytolacca acinosa Roxb), is a very controversial compound. PA-type saponins from quinoa are not acutely toxic to advanced organisms (Castillo- Ruiz et al., 2018), but PA which from P. acinosa Roxb could cause intestinal mucosal damage (Qi, Ma, Wang, & Zhang, 2017). Heating is a regular food processing method, and it could be optimized to have less effects on nutritional properties. For active substances, the heat processing could change the expected activity by transforming its chemical structure (Xue, Yao, Yang, Feng, & Ren, 2017). Most researches have focused on the retention of nutrients and the content of active ingredients (polyphenol/saponin) in quinoa, but there are few studies on the changes in the structure and function of saponins (Ramos Diaz et al., 2017;Gómez-Caravaca, Segura-Carretero, & Fernández-Gutiérrez, 2011).
Quinoa produces a large amount of bran during pearled processing, and this by-product comprises from 8% to 12% of the grain weight and contains from 200 to 300 g/kg saponins (Stuardo & San Martín, 2008). Therefore, quinoa bran was selected as the model to discover the changes in structure and activity of saponins during processing (Ruiz et al., 2017). In this work, the content and conversion of quinoa saponins were studied under three different processing methods (boiling, steaming, and infiltrating), which correspond to the three common methods of preparing edible quinoa, such as soup, steamed bread, and salade, respectively (López, Capparelli, & Nielsen, 2011). At the same time, the main compounds before and after conversion were also evaluated for the proliferation and toxicity of normal human gastric mucosal epithelial cells (GES-1), which form a barrier between the body and ingested substances, such as potential nutrients and toxins. The goal of this study is to determine which method can minimize the damage of quinoa saponins on human gastric mucosal cells, improve the safety of by-products, and provide data in support for the further utilization of saponins.

| Extraction and separation of the saponins
The quinoa bran was ground into powder and passed through a 60-mesh sieve. A total of 50 g powder was soaked in 30°C water at a ratio of 1:15. After ultrasonic extraction for 30 min, the supernatant was obtained by centrifugation for 5 min at 400 g. The product was extracted three times with n-butanol and combined.
The merged solution was recovered by rotary evaporation at 60°C and then heated and evaporated at 60°C in an evaporating dish to obtain a cold soak saponin extract (CS). In addition, 50 g of powder was suspended in cold water at a ratio of 1:15 and boiled for 30 min at 100°C. The saponin was dissolved in the hot water, and the subsequent treatment method was consistent with the method described above. Finally, the boiled quinoa saponin (BS) was obtained. Equal amounts of quinoa bran flour were wrapped with gauze and heated by steaming for 15 min at 105°C. Subsequently, the quinoa bran flour was removed and dried using a freeze dryer.
The subsequent treatment method was consistent with the method described above, and steamed saponins (SS) were obtained. Then, 50 g quinoa bran flour was dispersed into 750 ml of 70% methanol and sonicated for 30 min. The merged solution was recovered by rotary evaporation at 60°C, and the subsequent treatment method was consistent with the method described to obtain saponins in the n-butanol layer. (1.04 mg; 1.12 mg; 1.07 mg; 1.32 mg; 1.14 mg) were dissolved in 10 ml 70% (v/v) ethanol.

| Determination of total saponin content by spectrophotometry
The total saponins content was evaluated by using a previously reported spectrophotometry method with some modifications (Chen, Xie, & Gong, 2007). Briefly, 0, 0.2, 0.4, 0.6, 0.8, or 1.0 ml esculentoside A was added to a 10 ml glass tube and dried in a water bath at 70°C. Then, 0.2 ml vanillin-glacial acetic acid solution and 0.8 ml perchloric acid solution were added to the 10 ml glass tube.
After shaking, the tubes were heated in a constant temperature water bath at 60°C for 15 min and then cooled under flowing cold water for 5 min. Then, 5 ml glacial acetic acid was added, and the mixture was allowed to stay for 30 min. The absorbance at 560 nm was then measured by visible spectrophotometry. The absorbance value was set to a standard curve on the ordinate. Meanwhile, the absorbance value of the 0.5 ml was added to the solution in the above manner, a solution without the sample was used as a blank, and the saponins content was determined using the standard curve.

| HPLC analyses
Chromatographic analysis was performed using a SHIMADZU Prominence LC-20A HPLC instrument (Shimadzu Corporation) equipped with a YMC-ODS Pack column (4.6 mm × 250 mm, YMC Co., Ltd.). The detection wavelength was set at 202 nm and the temperature of the column oven was 25°C. The mobile phase was consisted of water (A) and acetonitrile (B). A gradient elution was used as follows: 5 min 10% B; 10 min 15% B; 15 min 20% B; 35 min 28% B; 50 min 40% B; 60 min 60% B; 70 min 70% B; and 75 min 10% B. The flow rate was kept at 1 ml/min, and the injected volume was 10 μl.

| HPLC-ESI-MS conditions
The column effluent of HPLC was introduced into an Agilent LC-1100 mass spectrometer (Agilent) equipped with an electronic spray ionization source 6460 (ESI, Agilent). The parameters of the ESI were set according to a previous report with slight modifications (Gómez-Caravaca et al., 2011). Briefly, the collision gas (N 2 ) rate was maintained at 10 ml/min, and the column oven was maintained at 25°C. ESI-MS were acquired in negative mode to generate [M−H] − and ginsenoside ions by full scanning m/z over 50-2,000. The spray voltage was 4.5 kV, the capillary voltage was 10 V, and the capillary temperature was 250°C.

| Preparation of quinoa saponins curves
The injected volume of five quinoa saponins (Qb-Qf) was 1 μl, 2 μl, 4 μl, 8 μl, 10 μl, and 12 μl, respectively. The working standard solutions were analyzed by the established method in triplicate. Calibration curves were plotted as the peak area (y) versus the amount of each ginsenoside standard (x). The content of ginsenoside in each sample was evaluated by the standard curve for each analyte.
A recovery test was performed by a previously reported method . The spiked samples were analyzed in triplicate by the described HPLC method.
The cells were maintained in 5% CO 2 in an incubator at 37°C.

| Cell cytotoxicity assay
The quinoa saponins cytotoxicity activity was assayed as described by Yang and Liu with slight modifications (Yang & Liu, 2009).
Briefly, 200 μl of GES-1 cells were added to a 96-well plate (5 × 10 4 cells/well) and incubated at 37°C for 24 hr. Then, different concentrations (1, 50, 100, 200 μg/ml) of samples were mixed with the cells and incubated for 24 hr. For cytotoxicity testing, the growth medium was removed, and the cells were suspended in PBS. Next, 10 μl of WST-8 [2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium, monosodium salt was added to each well and incubated for approximately 1 hr. After removal of WST-8, the 96-well plate was completely washed by immersion in Pall water and 100 μl of elution solution (with PBS plus 50% of ethanol and 1% of acetic acid) was added. In the end, the cells were incubated for 20 min with gentle rotation at room temperature. The results were expressed in the terms of optical density at 570 nm using a Thermo Microplate Reader (Dynex Technologies).

| Cell proliferation assay
The proliferation activity was evaluated by the method reported by You, Zhao, Liu, and Regenstein (2011). Cells were seeded (2.5 × 10 4 cells/well) in a 96-well flat-bottom plate. The effect of quinoa saponins on cell viability was determined using the WST-8 to formazan assay. After incubation at 37°C for 6 hr, the growth medium was replaced with media containing ginsenosides at variable concentrations. The absorbance of each well was read at 570 nm using a microplate reader, and the cell proliferation (percent) was determined at 96 hr from the absorbance of the sample compared to the control.

| Apoptosis assay by flow cytometry
The GES-1 apoptosis assay by flow cytometry followed a reported method (Cui et al., 2010). Cells were cultured as described above and treated with different concentrations of crude saponins, or saponin Qc or Qe (10, 50 and 200 μg/ml) for 24 hr, and then, the cells were rinsed with PBS and harvested in buffer (PBS-0.05% trypsin). The apoptosis of the GES-1 cells was determined using the Annexin V-FITC/PI apoptosis detection kit according to the manufacturer's protocol. A flow cytometer was used to analyze the fluorescence of the GES-1 cells.

| Data analysis
The data are presented as the mean of three replicates ± standard deviation. One-way ANOVA with Duncan's multiple range test was used to analyze the results with SPSS 13.0 and Sigma Plot 10.0, respectively. A p value of <.05 was defined as statistically significant.

| Total saponins in the extracts by spectrophotometry
The regression equation of the esculentoside A standard curve was obtained as Y = 31.316X − 0.0051 (R 2 = .998) and is shown in Table 1.
It showed that the linear relationship was good in 0-0.017 mg/ml.
The content of the n-butanol layer and total saponins from quinoa bran, CS, BS, and SS were also listed in Table 1. There were significant differences for the n-butanol layer and total saponins in the methanol extract and the three different processing methods (p < .05).
The yield of the n-butanol layer extract in quinoa bran was 8%, and the saponin purity was over 60%, as reported by others (Zhao et al., 2019). The values reported in this study for the saponins in quinoa seeds were within the range of the results previously published (Stuardo & San Martín, 2008). The saponin content can be determined initially by spectroscopy and further explored by HPLC.

| Characterization of saponins
The linearity of the calibration curves of saponins Qb-Qc is shown in Table 1. The compounds showed good linearity (R 2 > .99). The average recovery of the saponins ranged from 98.6% to 102.9%, with a relative standard deviation of <3.0%.
Five reference saponins were simultaneously identified according to the standard RT (retention times) and spectra, HPLC-ESI-MS ion fragments, and nuclear magnetic resonance (results not shown).
The structures of those saponins are summarized in Figure 1a Table 1. As shown in Figure 2, 5 peaks were successfully separated under the gradient elution. Consistent with the spectrophotometer results, the total saponins in the quinoa methanol extract were significantly different from the treated sample according to their origin (p < .05). The highest amount of total saponins in SS was 554.10 ± 24.58 mg/g, followed by those in the methanol    process, and the degree of the reaction has been directly influenced by the temperature (Brady, Ho, & Rosen, 2007).

| Structural changes in the quinoa saponins degradation
To explore the effects of different processing on the detailed saponins profile of quinoa, these samples were also analyzed by mass spectrometry. Because the same elution conditions were used in HPLC/MS, every sample in the ion chromatograms corresponds to the UV spectra. Thus, the relative retention time and content trends for each saponin monomer are highly consistent with the UV spectra. HPLC/MS has a higher sensitivity and lower detection limit than HPLC, so more saponins are detected in the same sample. In the previous report, quinoa saponins can be divided into four categories according to the different aglycones, such as oleanolic acid (OA), phytolaccagenic acid (PA), hederagenin (Hed), and serjanic acid (SA), which are the common triterpenoid saponins (Kuljanabhagavad, Thongphasuk, Chamulitrat, & Wink, 2008). In addition to identifying the saponin species by standard compounds, the relative molecular mass and fragment ion peaks can also be utilized (Mad et al., 2006).
As Figure 1g shows, Qc loses one molecular glucose at the position of C-3 and is converted to Qd, and then finally transformed to Qf. Unlike the thermal degradation of saponins in ginseng, polar ginsenosides could transform into less polar ginsenosides upon heating (Xue et al., 2017). Conversion of the quinoa saponin may be related to the enzymes present in bran, which is similar to the Oenning's study (Oenning, Juillerat, Fay, & Asp, 1994).

| Proliferative effects of quinoa saponins on GES-1
The proliferative activities of quinoa saponins toward the human gastric epithelial cell line GES-1 at a concentration range of 1-200 μg/ ml were investigated. The proliferative activity of saponin Qc and Qd toward GES-1 cells in vitro was enhanced, as the concentration of saponin increased at 12 hr and 24 hr (Figure 4a,b). As a positive control, aspirin slightly improved GES-1 cell proliferation at doses from 10 μg to 50 μg, but it shows a tendency to reduce the growth rate above 100 μg. The behavior of crude saponins was consistent with the positive control drug. Interestingly, saponin Qf minimally promoted cell growth.
Saponins consist of a lipophilic aglycone core and a hydrophilic sugar chain, which can perturb biologic membranes and reduce the surface tension of aqueous solutions (Bottger & Melzig, 2013).
However, the activity of saponins acting on the cell membrane is mainly related to their structure (Bottger, Hofmann, & Melzig, 2012), which explains the different activities exhibited by the different saponins of quinoa.

F I G U R E 4
Aspirin, crude saponins, Qc, Qe, and Qf affects proliferation on GES-1 cells. Samples improve GES-1 cells in a concentration-dependent manner after treatment for 12 hr (a), and 24 hr (b); samples inhibit proliferation on GES-1 cells after treatment for 24 hr (c)

| Cytotoxicity effects of saponins Qc and Qf on GES-1
The data for the cytotoxicity effects of aspirin, crude saponins and saponins Qc, Qd, and Qf on GES-1 are shown in Figure 4c. It was observed that Qd and crude saponins did not exhibit any cytotoxicity toward GES-1 cells. After treatment with 10 or 50 μg saponin Qf for 24 hr, inhibition of cell growth was not observed. When the dose was increased to 100 μg, the percentage of dead cells increased to 86.2 ± 4.8%. After treatment with 10, 50, 100, or 200 μg saponin Qc and aspirin for 24 hr, the percentage of dead cells increased. However, the toxicity of these saponins is much lower than saponin Qf.
The apoptosis of GES-1 cells treated with Qc and Qf was detected by flow cytometry to further verify the toxic effects of quinoa saponins. As Figure 5 shows, Qc alone at 100 μg/ml did not induce apoptosis, with 91.87% of cells remaining viable (Q3) compared to the untreated cells, with a viable cells rate of 91.57%. As shown in Figure 5e-g, after treatment with Qc (10-100 μg/ml) for 24 hr, the percentage of Q2 and Q3 was increased significantly in a concentration-dependent manner; after treatment with 10, 50, and 100 μg/ ml Qf, the total apoptotic cells increased to 9.85, 10.84, and 14.75%, respectively. The results confirmed that the ability of Qf to induce the apoptosis of GES-1 cells was stronger than QC. Consistent with previous work, transforming ginseng by heat produced significantly higher anticancer effects than polar saponins (Quan et al., 2015).
In addition, saponin anemoside B4 with five sugars had a weak cytotoxic effect on the K562, B16, Hela, and HUVEC cell lines, while pulchinenoside A with two sugars could significantly inhibit cell proliferation . Quinoa saponin can effectively inhibit mycelial growth and spore germination with lye treatment by destroying fungal cell membranes, which may be related to the tighter connection between saponin and cell membranes (Stuardo & San Martín, 2008). Perhaps less polar quinoa saponins are more likely to bind to cholesterol in cells alone (Bottger et al., 2012;Bottger & Melzig, 2013).

| CON CLUS IONS
Most saponins in quinoa were hydrolyzed by cold soaking. One of the main reasons for this hydrolysis is that the presence of glycosidase in bran can hydrolyze glycosides from saponins because organic reagents and heating can destroy the enzyme, thereby preventing saponin hydrolysis. Quinoa saponin was more toxic to GES-1 after hydrolysis than the polar saponin because less polar quinoa saponin is more likely to bind to cholesterol in cells. Quinoa bran contains various nutrients and could prevent endogenous enzymes from degrading saponins during consumption or feeding.

ACK N OWLED G M ENTS
This work was supported by the Doctoral research start-up funds of Weifang Medical College and project no. ZR2019PH043 supported by Shandong Provincial Natural Science Foundation.

CO N FLI C T O F I NTE R E S T S
No conflict of interest was declared by the authors.

E TH I C A L A PPROVA L
This study does not involve any human or animal testing.