An Insight into Saponins from Quinoa (Chenopodium quinoa Willd): A Review

Saponins are an important group found in Chenopodium quinoa. They represent an obstacle for the use of quinoa as food for humans and animal feeds because of their bitter taste and toxic effects, which necessitates their elimination. Several saponins elimination methods have been examined to leach the saponins from the quinoa seeds; the wet technique remains the most used at both laboratory and industrial levels. Dry methods (heat treatment, extrusion, roasting, or mechanical abrasion) and genetic methods have also been evaluated. The extraction of quinoa saponins can be carried out by several methods; conventional technologies such as maceration and Soxhlet are the most utilized methods. However, recent research has focused on technologies to improve the efficiency of extraction. At least 40 saponin structures from quinoa have been isolated in the past 30 years, the derived molecular entities essentially being phytolaccagenic, oleanolic and serjanic acids, hederagenin, 3β,23,30 trihydroxy olean-12-en-28-oic acid, 3β-hydroxy-27-oxo-olean-12en-28-oic acid, and 3β,23,30 trihydroxy olean-12-en-28-oic acid. These metabolites exhibit a wide range of biological activities, such as molluscicidal, antifungal, anti-inflammatory, hemolytic, and cytotoxic properties.


Introduction
The mother grain, quinoa (Chenopodium quinoa Willd.), is an annual plant native to the Andes. For over 5000 years, its seeds were the main crop of Andean cultures in South America; however, their cultivation and consumption were virtually eliminated and only remained in farming traditions after the arrival of the Spaniards [1,2].
The genus Chenopodium (Chenopodiaceae) comprises 250 species, among them C. quinoa [3]. The latter is considered a pseudocereal because its seeds are used as cereal grains [4]. The plants are tall, branched, and have a long vegetative period and broad leaves. They produce large to small flat grains, oval, usually pale yellow, but color can vary from pink to black [5].
Quinoa can be grown in the fertile soils and warm climate of the Andean region of South America, where the crop grows between sea level and around 4000 m above sea level [2,5]. Some quinoa varieties are able to adapt even to marginal environments [6].
The main quinoa-producing countries are Bolivia, Peru, and Ecuador, which produced 146,735 tons in 2017, compared with 59,115 tons in 2007 [7]. In 2017, quinoa production was recorded as 78,657 tons

Saponins Elimination Methods
Saponins are considered an antinutritional factor, which must be eliminated before the consumption of seeds. To this end, several methods have been reported to wash off seeds' saponins, including wet methods, dry methods, and combinations of both. Recently, genetic methods have also been developed [5,18]. Table 1 represents a summary of several wet washing methods. As early as 1978, Rios and his team initially washed quinoa seeds using water at 50 • C [19]. Twelve years later, Ridout et al. (1990) described a method for the separation and analysis of saponins in raw quinoa. They were also the first to classify quinoa according to its saponin content and degree of bitterness. The terms used to describe quinoa were bean, farinaceous, bitter, and astringent, and the rating scale of bitterness was 0: absent, 2: very weak, 4: weak, 6: moderate, 8: strong, and 10: very strong, recorded at times between 0 and 10 min after tasting. Additionally, the authors founded that three sequential washings decreased the total saponin content from 1.03% to 0.18%, a level at which the product lost its unpalatable astringent sensory characteristics. In the same year, Meyer et al. (1990) showed that quinoa seed bitterness is essentially due to the presence of quinoside A [20]. In another experiment,  detected two major saponins (1 and 2) in quinoa seeds. After polishing and washing the seeds, 56% of saponin 1 was removed, while saponin 2 was not detectable in the treated samples. Since quinoa did not taste bitter after that washing round, it was assumed that the bitter taste is most likely due to the contribution of saponin 2.

Wet Methods
According to Kozioł (1992), the application of the afrosimetric method made it possible to classify the quinoa varieties of the Latinreco experimental plots as "sweet" or "bitter" via taste tests. Koziol suggested that the standard afrosimetric method was developed to allow enough time to extract saponins from quinoa seeds, but proved too long to be used in the field to identify low-saponin varieties or to test the effectiveness of the abrasive peeling. However, with this rapid afrosimetric method, the total time required to perform the analysis was reduced to about 7 min; quinoa is classified as sweet if it gives a foam height of 1.3 cm or less. The application of this method must take into consideration the height of the rapidly changing foam in both directions and the correlation between the height of the foam and the extraction time of the saponins [9].
Unlike the aqueous extraction used in the laboratory, one commercial washing procedure eliminated about 72% of the saponins from the grain, and the saponin content varied from 0% to 2.0% according to the variety-sweet or bitter quinoa [21]. Similarly, Pappier et al. (2008) showed that saponin content decreased to less than 0.06% in seeds treated by industrial process. On the other hand, the saponin content in seeds washed only with water at 50 • C was about 20% of the original content [22]. This suggests that the commercial process was considerably more intense than the laboratory procedure [21].
The initial saponin content in quinoa, 6.34%, reached a level as low as 0.25%-0.01% during the first half hour of washing in a study by Vega-Galvez et al. (2010). Based on these results, the minimum washing time required to extract the largest amount of saponin (represented by a 96% decrease in the saponin content) could be estimated as 60 min [23].
In a different approach, Quispe et al. (2012) described the kinetics of saponin leaching during the washing of quinoa seeds under different temperature conditions using empirical models based on Fick's diffusional equation for mass transfer to fit the experimental data. At the beginning of the washing process, the leaching of saponins was very rapid with an asymptotic value of saponin concentration in the seeds, and the value was negatively correlated with the washing temperature. The authors were also able to calculate an effective diffusion coefficient by applying the Arrhenius equation, which was found to increase with increasing washing temperature [10]. By the same means, Irigoyen and Giner (2018) used the diffusion equation to analyze the extraction kinetics of saponins, with the aim of perfecting the saponin leaching mechanism. In this study, treatment carried out at 40 • C for 6 min was considered optimum to reach a safe level of saponins for human consumption without visible damage to the seed [24].
From a different perspective,  reported the effect of washing on the nutritional quality of the protein content in quinoa seeds, where the digestibility of the quinoa proteins was studied based on pH variation and amino acid composition after digestion with enzymes. The study did not find any change in the amino acid composition after washing, which showed that the process of saponin removal has no effect on the protein nutritional quality of quinoa seeds [25]. In the same regard, Telleria Rios and Sgarbieri (1978) did not find any change in the amino acid composition [19]. However, Nickel et al. (2016) studied the effects of five types of treatment on the saponin content, total phenol content, and antioxidant capacity of quinoa grains grown in Brazil. It was found that washing the seeds with water was not highly efficient in decreasing the saponin content and, consequently, reducing bitterness. In addition, this method resulted in the increase of total phenols [26]. On the other hand, the total phenol content and the antioxidant capacity were found to be higher in water-cooked seeds compared to other treatments. The authors also ran a microscopic examination after washing the grain samples and after cooking to evaluate the damage to the pericarp and seed coat layers. Washed and abraded samples exhibited significant differences when examined under microscope before and after cooking; the abraded samples revealed more cellular content [27]. A study by Pappier et al. (2008) evaluated the effects of saponin removal on fungal contamination. The results of this work showed that all samples containing saponins had 100% fungal contamination. Penicillium and Aspergillus were the most prevalent contaminants [22].

Dry Methods
Brady et al. (2007) evaluated by HPLC analysis the effects of three types of heat treatment, namely steam pre-conditioning, extrusion, and roasting of quinoa flour on their chemical profile. They showed that the heating temperature caused saponin degradation and therefore could have direct effects on sensory perception and pharmacological properties, and thus the application of extrusion and roasting can reduce the bitter taste conferred by saponins [29]. Gómez-Caravaca et al. (2014) studied the effect of two different degrees of beading (20% and 30%) on the elimination of saponins, and their impact on the phenolic composition of quinoa seeds. Their study proved that, according to the GC-MS analysis results, beading with a degree of abrasion of 30% allowed sweet quinoa to be obtained. The beading technique caused a 21.5% and 35.2% reduction in the free and bound phenolic compounds, and was an efficient method for obtaining sweet quinoa without affecting the total content of phenolic compounds in an excessive way [30]. In another study, Repo-Carrasco-Valencia et al. (2010) explored the effect of roasting on the nutritional value of the seeds of three cereals of Andean origin, among them quinoa (C. quinoa), and studied the effects of roasting and boiling on the nutritional value of these seeds, specifically the availability of iron, calcium, and zinc. According to their study, the analysis of zinc dialyzability in different raw samples showed that roasting treatment had no significant effect on the availability of these minerals. In contrast, boiling treatment decrease the dialyzability of iron and zinc in all samples and improved that of calcium [31]. Moreover, the roasting process, according to Nickel et al. (2016), resulted in a reduction of phenolic compounds [26]. In the same regard, the dry method was shown to significantly decrease the vitamin and mineral contents, especially potassium, iron, and manganese [25].
Scarification is the most used dry method at the industry level, and it consists of separating the episperm from the quinoa grain, where the highest content of saponins is concentrated. Studies have shown that the use of a scarificator for 6 min reduces the saponin content in quinoa seeds from 0.324% to 0.001%; however, it damages the seed structure, as small numbers of broken grains were observed, with a minor problem of dust remaining in the scarified grain [32]. Armada (2012) recommended the addition of a polishing operation after scarification, in order to improve the efficiency of the scarification equipment in the quinoa grain disaponification operation [32].

Genetic Methods
Other techniques of removal of quinoa seed saponins have been tested, including genetic methods [5]. A study by Ward (2000) performed selection of quinoa plants to estimate the heritability of saponin content from one generation to the next. This study showed a slow decrease in saponin content; however, the percentage of plants containing less than 1 mg/g of saponin in the grains increased from 3.57% at S1 to 11.1% at S4. This slow selection was explained by the dominance of the allele responsible for saponin expression, which required more time for the recessive inhibitory alleles of saponin synthesis to accumulate at the relevant loci [18]. On the other hand, Dick Mastebroek et al. (2001) selected sweet quinoa genotypes with low concentrations of saponins at different stages of development, based on the concentration of saponins in the leaves. The results showed that the seeds contain higher levels of sapogenins than the leaves, and that the difference in sapogenin content between leaves and seeds is higher in bitter quinoa. Nonetheless, the sapogenin content of the leaves generally increases during the development of the plant, with a maximum content of saponins reached 82 days after sowing. The authors demonstrated that following the sapogenin content in the leaves during the development of the plant did not allow the early selection of sweets genotypes before anthesis [33].
Concentrations of saponins vary between genotypes, from sweeter to bitter varieties. Water and nutrient availability have been found to be responsible for quinoa saponin biosynthesis in seeds, thus strengthening the hypothesis of the combined influence of genetic and environmental factors [33].
In 1989, Risi and Galway evaluated 294 accessions for their ability to adapt to the new environment of England. In this study, the differences between accessions were highly significative for saponin content. This opened the door for other studies to unlock the mystery of saponins in quinoa seeds [34]. In 1996, Jacobsen et al. studied the stability of various descriptive characteristics over a 5 year period in 14 lines of quinoa [35], where saponin content was assessed either by bitterness in a taste test of mature seeds, on a scale from 0-10, or by an afrosimetric method that involved the estimation of foam development [9]. Genotype-environment interactions were significant, with variance components accounting for 16% of the total variation in saponin content. This finding qualifies the saponin content as a quantitative trait. The complexity of understanding saponin expression is mainly due to the cytogenetics of quinoa as an allotetraploid species (2n = 4x = 36, with basic chromosome number of x = 9), with a diploid type of chromosomal segregation [36], although some tetrasomic inheritance occurs in the linkage groups [18]. Although quinoa displays disomic inheritance for most qualitative traits, this influences trait segregation and may cause segregation distortion [37].
Later, a microarray analysis by Reynold et al. (2009) allowed the identification of a set of candidate genes transcriptionally related to saponin biosynthesis, including genes with shared homology to cytochrome P450s, cytochrome P450 monooxygenases, and glycosyltransferases [38].
The development of the quinoa market in non-Andean countries and the development of molecular tools for breeding has inspired researchers to look for a potential new approach to quinoa grain improvement related to saponin content as an economically important trait. Maughan et al. (2004) and Jarvis et al. (2008) published linkage maps based on different molecular resources developed using quinoa RILs (recombinant inbred lines), including SSR (Simple Sequence Repeat), AFLP (Amplified Fragment Length Polymorphism), and RAPD (Random Amplified Polymorphic DNA) markers; 11S seed storage protein loci; the NOR (Nucleolar Organizer Regions); and the morphological betalain color locus [37,39]. The main objective was to provide a key starting point for the genetic dissection of important traits, including saponin content. To increase the precision of the linkage map and to narrow down the exact position of the candidate gene, Maughan et al. (2012) developed an SNP (Single Nucleotide Polymorphisms)-based map using 511 marker loci, which represented an important genomic tool needed in plant breeding programs. The complexity of the polyploidy of quinoa genome and thus the saponin content trait genetics make the inheritance patterns difficult to understand and to apply in agriculture [40].
A new era of research started in 2017 when the genome was sequenced, accelerating the identification of genes responsible for saponin synthesis [41]. The genome sequence facilitated the identification of the transcription factors likely to control the production of saponins, including a mutation that appears to cause alternative splicing, which is a source of a premature stop codon in sweet quinoa. The genes regulating the absence of saponins in sweet quinoa accessions are unknown. To identify these genes, Jarvis et al. (2017) made a linkage mapping and bulk segregant analysis (BSA) using two populations, segregating for the presence of saponins in the seeds. The results showed quantitative and qualitative differences between the saponins identified in bitter and sweet lines. Additionally, the presence and absence of saponins was correlated with the differences in the thickness of the seed coat, with a bitter line having a significantly thicker seed coat than sweet lines [41].

Extraction Methods of Quinoa Saponins
The extraction of quinoa saponin can be carried out by several methods ( Table 2). The most utilized methods are the conventional technologies of maceration, extraction by reflux, and Soxhlet [42]. These methods are based on the solubility of solute in the solvent. Ethanol and methanol are still considered the best solvents for the extraction of saponins because of their high solubilization properties. In a study by Gee et al. (1993), methanol was found to be the most efficient solvent with which to extract saponin. Unlike grain, infant wheat cereals' aqueous extracts exhibited similar saponin contents to their methanol extracts. In particular, aqueous cereal extracts showed higher oleanolic to hederagenine ratio than quinoa seed aqueous extracts [21]. Additionally,  showed that the aqueous methanolic extract, which is usually removed in the conventional analytical methods, contained some triterpenoid saponins that were not present in the butanol phase. Therefore, the conventional analytical procedure can lead to the loss of components if the water-soluble fraction is neglected [43]. Research is currently focused on microwave extraction, a technology that improves the efficiency of extraction by reducing extraction time and solvent consumption in addition to enhancing the extract quality [44]. In a study by Gianna et al. (2012), the authors evaluated the reliability of extracting saponins from quinoa seeds with solvents by the application of microwaves, and the effect of different parameters such as temperature, seed/solvent ratio, treatment time, and % solvent on saponin removal percentage were evaluated. The numerical analysis of the results showed that the optimal extraction conditions were using ethanol-water and isopropanol-water 20%, with a treatment time of 20 min at a temperature of 90 • C. The extraction of saponins by microwave method makes it possible to reduce the extraction time 20-fold compared to extraction by Soxhlet, and the use of the alcohol as a solvent makes it possible to extract more saponin than the use of water alone [45].

Chemical Characterization of Quinoa Saponins
Saponins constitute a large group of glycosides found in various plants. They are polar molecules consisting of a steroid or triterpene aglycone with one or more sugar chains and are characterized by their surfactant properties, which differentiate them from other glycosides [56]. According to their structure, saponins are classified into two types. Steroidal saponins dominate in angiosperm monocotyledonous plants (Figure 1a), and triterpene saponins are generally present in angiosperm dicotyledonous plants, the group to which quinoa belongs (Figure 1b) [57].

Chemical Characterization of Quinoa Saponins
Saponins constitute a large group of glycosides found in various plants. They are polar molecules consisting of a steroid or triterpene aglycone with one or more sugar chains and are characterized by their surfactant properties, which differentiate them from other glycosides [56]. According to their structure, saponins are classified into two types. Steroidal saponins dominate in angiosperm monocotyledonous plants (Figure 1a), and triterpene saponins are generally present in angiosperm dicotyledonous plants, the group to which quinoa belongs (Figure 1b) [57]. Saponins are the major secondary metabolites present in quinoa, found mainly in the outer layer of the seeds [25]. Chemically, quinoa saponins are triterpene glycosides composed of a hydrophilic oligosaccharide bound at C-3 and C-28 to a hydrophobic aglycone. Glucose, galactose, arabinose, glucuronic acid, and xylose are the common oligosaccharide sugars. The aglycone is usually derived from oleanolic acid, hederagenin, phytolaccagenic acid, serjanic acid, spergulagenic acid, gypsogenin, 3β-hydroxy-27-oxo-olean-12-en -28-oic acid, and 3β, 23 α, 30 β-trihydroxy-olean-12-en-28-oic acid [43,49,53,55,58,59].
The biosynthesis of quinoa saponins takes place through the mevalonate pathway via farnesyl diphosphate (FPP), a 15 C structure [60]. Pairs of FPP molecules are linked to give squalene (30 C) [61]. The latter is oxidized to oxydosqualene [62], followed by the formation of β-amyrin (the precursor for all quinoa saponins) under the action of β-amyrin synthase. The first aglycone is oleanolic acid, which is then subjected to several modifications (oxidation, glycosylation, and esterification) producing the other seven aglycones [5].
The biosynthesis of quinoa saponins takes place through the mevalonate pathway via farnesyl diphosphate (FPP), a 15 C structure [60]. Pairs of FPP molecules are linked to give squalene (30 C) [61]. The latter is oxidized to oxydosqualene [62], followed by the formation of β-amyrin (the precursor for all quinoa saponins) under the action of β-amyrin synthase. The first aglycone is oleanolic acid, which is then subjected to several modifications (oxidation, glycosylation, and esterification) producing the other seven aglycones [5].
In general, the majority of saponins are isolated from the seeds of quinoa; however, a few studies have reported some saponins isolated from other parts such as leaves, stem, and roots. Dick Mastebroek et al. (2000) reported the presence of saponins in quinoa leaves, but, so far, major saponins have not been isolated from plant parts other than seeds and flowers.

Biological Activities of Quinoa Saponins
Saponins are characterized by a bitter taste and are considered toxic in high concentrations. They are present throughout the quinoa plant. To date, there have been no studies justifying the role of quinoa saponins and the limiting factors of their production at the level of quinoa, but their presence is generally considered a defense mechanism against the plant's natural enemies [21,28].
Saponins exhibit several physicochemical and biological properties (Table 4) [64]. These include antioxidant, analgesic, immunostimulant, antimicrobial, antiviral, and cytotoxic activities, and anti-inflammatory and hemolytic effects. They also affect the absorption of some minerals and vitamins and the growth rate of the consuming organisms [60,65]. In order to improve the yield and the biological activities of saponins and their applications in the food, cosmetics, agricultural, and pharmaceutical sectors, several techniques of extraction and identification have been developed [42,44]. Woldemichael and Wink (2001) reported the antifungal activity of a crude saponin extract from C. quinoa. The extract inhibited the growth of C. albicans at the concentration of 50 µg/mL. However, pure saponins (mono-, bi-, and tridesmosidic saponin) exhibited poor antifungal activity with minimal inhibitory concentrations (MICs) of 100 and 500 µg/mL for mono-and bidesmosidic saponin, respectively; the superior activity of the extract might be attributed to synergistic interactions between the multiple components of the extract, in which they may affect several molecular targets at the same time. However, it has also been reported that the carbohydrate chain in saponins has an effect on membrane permeability [52]. This can be explained in terms of the amphiphilic structure of monodesmosidic saponins [52,66,67]. The lipophilic triterpene aglycone allows the saponins to dive into the microbial membrane, while the carbohydrate chain unit at C3 helps the molecule to binds to glycolipids and extracellular glycoproteins [52]. Quinoa saponins have also shown a biocidal effect against Pomacea canaliculata, a snail that attacks rice seeds, with a very positive effect on the germination rate of rice seeds [68,69].
To date, saponins have received much attention due to their wide spectrum of pharmacological and biological proprieties [70][71][72]. Saponins have interesting hemolytic properties against erythrocytes and can interact with lipid and cholesterol membranes as surface active molecules and disrupt these membranes by forming pores that that destabilize the membrane [70][71][72][73][74][75]. According to Gee et al. (1993), quinoa saponins can induce membranolytic activity against the mucosal cells of the small intestine. However, the hemolytic activity of saponins depends on the structure of the saponins in question [21]. Vo et al. (2016) investigated the hemolytic activity of 41 triterpenoid saponins and sapogenins with three different types of structural skeletons, and showed a strong relationship between the chemical structure and the biological activity [75]. In fact, according to this study, almost all of the oleanane sapogenins exhibited stronger hemolytic activity, owing to the presence of a carboxyl group at position 28, an α-hydroxyl group at position 16, and a β-hydroxyl at position 2. However, a reduction of hemolytic effect was observed with the introduction of methyl hydroxyl at positions 23 or 24 and α-OH at position 2 [52,75,76].
Inflammation is the first response to infection or injury via the complex biological response of vascular tissues to harmful stimuli such as pathogens, damaged cells, or irritants [77].
The anti-inflammatory activity of saponins from quinoa seeds has been thoroughly investigated and reported in the literature. In a study by Yao et al. (2014), four quinoa saponin fractions, namely Q30, Q50, Q70, and Q90, extracted from C. quinoa were assessed for their anti-inflammatory activity. First, seeds were extracted with methanol by maceration, re-dissolved in water, and partitioned with ethyl acetate and butan-1-ol. After that, butan-1-ol saponins were fractionated in a resin column and eluted with 30% (Q30), 50% (Q50), 70% (Q70), and 90% (Q90) methanol. Results showed that the saponin fractions decreased the response of inflammatory mediators, inhibiting the release of inflammatory cytokines including tumor necrosis factor-α and interleukin-6 in lipopolysaccharide-induced RAW264.7 cells [77]. Verza et al. (2012) evaluated the adjuvant activity of C. quinoa saponins towards the humoral and cellular immune responses of mice immunized subcutaneously to ovalbumin (OVA). In the study, two fractions of quinoa saponin were tested (FQ70 and FQ90). Results revealed that the effect of FQ70 was significantly greater than that of FQ90. The quinoa saponins were shown to have a very positive effect on the production of cellular and humoral immune responses to OVA in mice [46]. Moreover, quinoa saponins showed a high capacity to potentiate antibody responses (IgG and IgA) by causing an increase in the permeability of the mucosa, allowing increased absorption of the antigen [78].   [49] Other biological properties, such as the effect of quinoa saponin on adipocyte differentiation, have also been demonstrated. Quinoa saponins were able to decrease cell viability and suppress 3T3-L1 adipocytes during the differentiation process [80]. In addition, bidesmosidic and aglycones saponins from quinoa showed cytotoxic effects through apoptosis induction in colon carcinoma Caco-2 cells [49].

Conclusions
Altogether, 40 saponin structures with eight different aglycones have been isolated and characterized from quinoa (C. quinoa). They exhibit a wide array of biological activities, among them antimicrobial, hemolytic, and anti-inflammatory activities, which make them very beneficial agents, especially in the medical sector. Saponins are also considered an antinutritional factor, necessitating their elimination before the consumption of seeds; to this end, several methods are employed, including wet, dry, and genetic. However, to set up a treatment process for the removal of quinoa saponin on an industrial scale, the best design parameters and the most profitable treatment conditions must be determined. Therefore, additional studies are still needed to reduce energy waste, optimize water flow in a continuous wash process, and improve saponin extraction.