Abstract

Introduction

Saponins are naturally occurring structurally and functionally diverse phytochemicals that are widely distributed in plants. They are a complex and chemically varied group of compounds consisting of triterpenoid or steroidal aglycones linked to oligosaccharide moieties. The combination of a hydrophobic aglycone backbone and hydrophilic sugar molecules makes the saponins highly amphipathic and confers foaming and emulsifying properties. These surface-active molecules have important roles in plant ecology and are also exploited for a wide range of commercial applications in the food, cosmetic and pharmaceutical sectors. Saponins in Quillaja saponaria extracts are used as foaming agents in carbonated beverages and cosmetics, as emulsifiers in preparations containing lipophilic colors or flavors, as preservatives, and for removal of dietary cholesterol (Güçlü-Ustündağ & Mazza, 2007; San Martín & Briones, 1999). Likewise, licorice saponin extracts are used as flavor modifiers in baked foods, chewing gum, beverages, candies, herbs, seasonings and dietary supplements (Güçlü-Ustündağ & Mazza, 2007). Saponins are also the major constituents of traditional folk medicines such as the ginsenosides produced by Panax species (Qi et al., 2011). A number of therapeutic properties have been ascribed to saponins (Francis et al., 2002). These molecules are potent membrane permeabilizing agents. They are also immunostimulatory, hypocholesterolemic, anti-carcinogenic, anti-inflammatory, anti-microbial, anti-protozoan, molluscicidal and have anti-oxidant properties. Furthermore, they can impair gut protein digestion, and uptake of vitamins and minerals. In general, the saponins and their biosynthetic intermediates display an array of properties that can have positive or negative effects in different animal hosts (Francis et al., 2002; Sparg et al., 2004).

Although the saponins are one of the largest classes of plant natural products, their biological functions are not completely understood. They are generally considered to have important roles in defense of plants against pathogens, pests and herbivores due to their antimicrobial, antifungal, antiparasitic, insecticidal and anti-feedant properties (Augustin et al., 2011; Morrissey & Osbourn, 1999; Osbourn et al., 2011; Sparg et al., 2004). Many plants synthesize and accumulate saponins during normal growth and development. The distribution of these natural products varies greatly among plant species, individual plants, organs and tissues, during development and maturation, and shows seasonal fluctuations. Some studies have suggested that variations in saponin distribution, composition and amounts in plants may be a reflection of varying needs for plant protection. In Phytolacca dodecandra and Dioscorea pseudojaponica, maximal saponin accumulation has been noted during fruit and tuber development, and has been suggested to prevent fruit loss, ensure seed maturation and protect reproductive organs (Lin et al., 2009; Ndamba et al., 1994). However, in several plant species, the production of saponins is induced in response to biotic stress including herbivory and pathogen attack. Abiotic stress factors such as humidity, nutrient starvation, light and temperature can influence both the quality and quantity of saponin content (De Costa et al., 2013; Szakiel et al., 2011). Increase in saponin levels in response to stress is often mediated by the transcriptional activation of biosynthetic genes through a complex signaling cascade involving the jasmonate and salicylate hormones. Hence, the biosynthesis of these molecules can be induced using elicitors and this feature has been exploited in several plant species to improve saponin yields (Shabani et al., 2009; Yendo et al., 2010).

Here, we provide a brief introduction to the vast diversity of saponin structures that are synthesized and accumulated by different plant species. Emphasis is given to differences in the composition and quantity of these natural products over different phases of plant growth, development and maturation. We provide a constructive framework to suggest that these compounds, which are traditionally regarded as secondary metabolites, have important roles in plants at various stages of development, including seed germination, vegetative growth and differentiation, fruiting and nodulation. A growing number of studies have reported effects on plant growth and development due to perturbation of saponin biosynthesis, suggesting roles for these molecules and their intermediates in primary plant processes. With these data, we highlight that maintaining a delicate balance of saponin biosynthesis is essential for overall plant integrity and well-being. The industrial applications of saponins, their intermediates and derivatives are also reviewed, together with structure–activity relationship data, to provide the reader with a perspective on the importance of this class of metabolites for pharmaceutical applications.

Biosynthesis of saponins

Most known saponins are found in angiosperms (Magnoliophyta), though some marine invertebrates such as sea cucumbers (Holothuroidea) and starfish (Asteroidea) also produce these molecules (Bordbar et al., 2011; Liu et al., 2008; Osbourn et al., 2011; Williams & Gong, 2007). While the synthesis of saponins is widespread in plants, the majority of the producing plant species is dicotyledonous and accumulates mainly triterpenoid-type saponins. The monocotyledonous angiosperms on the other hand mostly, but not exclusively, synthesize steroidal-type saponins. This broad classification of saponin types is based on the nature of the aglycone backbone from which the saponin molecule is derived. Both triterpenoid and steroidal aglycone backbones are derived from the 30-carbon linear precursor 2,3-oxidosqualene. During the synthesis of the committed precursor, the steroidal aglycone loses three methyl groups to result in a 27-carbon backbone, whereas the triterpenoid aglycone retains all 30 carbons in its backbone. The steroidal glycoalkaloids, which are also sometimes referred to as saponins, share their biosynthetic origin with the steroidal saponins and contain a characteristic nitrogen atom incorporated into the aglycone backbone (Augustin et al., 2011; Friedman, 2006; Ginzberg et al., 2009; Itkin et al., 2013).

The triterpenoid and steroidal aglycone backbones are isoprenoids that are synthesized from isopentenyl pyrophosphate (IPP) units generated by the mevalonate (MVA) pathway (Figure 1). The multistep MVA/3-hydroxy-3-methylglutaryl-CoA reductase (HMGR) pathway catalyzes the conversion of acetyl-CoA to the five-carbon terpene precursor IPP, which is then isomerized to its allylic isomer dimethylallyl pyrophosphate (DMAPP) by the enzyme isopentenyl diphosphate isomerase (IDI). The subsequent condensation of two IPP units with one DMAPP unit results in the formation of 15-carbon farnesyl pyrophosphate (FPP), the immediate prenylated precursor of the saponins. Finally, the condensation of two FPP units by squalene synthase (SQS) generates the linear 30-carbon precursor squalene, which is further epoxidized to 2,3-oxidosqualene by the enzyme squalene epoxidase (SQE). The 2,3-oxidosqualene is typically cyclized by a variety of oxidosqualene cyclases (OSCs) to polycyclic structures, which in higher plants marks the branch point between primary and specialized triterpene metabolism. The tetracyclic, primary triterpene precursor cycloartenol is generated by the cyclization of 2,3-oxidosqualene by cycloartenol synthase (CAS). In angiosperms, a mixture of phytosterols are generated from cycloartenol, including the 27-carbon cholesterol, the 28-carbon campesterol and the 29-carbon sitosterol. The steroidal saponins are synthesized by a series of oxygenations and glycosylations of the cholesterol backbone to spirostanol or furostanol derivatives with a fused O-heterocycle in their core aglycone structure (Thakur et al., 2011). The steroidal glycoalkaloids also utilize cholesterol as the committed precursor, but incorporate an amine group through side chain modifications to generate aglycones such as solasodine, solanidine, demissidine and tomatidine (Ginzberg et al., 2009; Itkin et al., 2013). The aglycones are then tailored with oxidoreductases before being glycosylated with multiple sugar moieties (Friedman, 2006).

All the remaining cyclization products of 2,3-oxidosqualene are committed precursors for the synthesis of specialized triterpenes. This cyclization reaction, involving multiple carbocation rearrangement steps, generates the first level of structural diversity inherent to the triterpenoid saponin aglycones, as the single substrate 2,3-oxidosqualene can be cyclized to an array of triterpene scaffolds. As such, nine main classes of triterpene backbones have been documented to be synthesized by plants (shown in orange in Figure 1; Vincken et al., 2007). The OSCs catalyzing these reactions are either specific or multifunctional in nature, and cyclize 2,3-oxidosqualene to a single product or to multiple products derived from a particular cyclization pathway in a single reaction. The triterpene aglycones are then modified, mainly oxidized, by a series of cytochrome P450-dependent monooxygenases (P450s). These modifications add a second level of complexity and expand the structural diversity of the backbone. Consecutive oxidations at different positions of the triterpene backbone increase the polarity of the scaffold and introduce reactive functional groups which are subsequently modified by an array of transferases, including UDP-dependent glycosyltransferases (UGTs), acyltransferases, malonyltransferases and methyltransferases. A comprehensive overview of the biosynthesis of triterpenoid saponins can be found in Thimmappa et al. (2014).

The genes involved in the biosynthesis of all three types of saponins (triterpenoid, steroid and steroidal glycoalkaloid) can therefore be grouped into four main categories based on their characteristic reactions, and include the OSCs, P450s, UGTs and other tailoring enzymes (mostly encoding transferases). Recently, numerous examples of gene clusters for the biosynthesis of different classes of specialized metabolites have been discovered in a variety of plant species, including clusters for saponin biosynthesis (Boycheva et al., 2014; Nützmann & Osbourn, 2014). Biosynthetic gene clusters have been identified for triterpenoid saponins and steroidal glycoalkaloids. The first triterpene gene cluster to be identified was the avenacin cluster from Avena strigosa (oat), which contains five biosynthetic genes encoding for an OSC, a P450, a UGT and two other tailoring enzymes (a methyltransferase and a serine carboxypeptidase-like acyltransferase) (Haralampidis et al., 2001a; Mugford et al., 2013; Papadopoulou et al., 1999; Qi et al., 2004, 2006). Other closely linked genes that form part of this cluster have been defined by genetics but not cloned (Mylona et al., 2008; Papadopoulou et al., 1999; Qi et al., 2004). In Arabidopsis thaliana, two triterpenoid gene clusters have been identified for the biosynthesis of thalianol- and marneral-derived triterpenes, containing four and three genes, respectively. The thalianol cluster is comprised of genes encoding an OSC, two P450s and an acyltransferase (Field & Osbourn, 2008). Likewise, the marneral cluster encodes an OSC and two P450s (Field et al., 2011). A triterpenoid gene cluster encoding an OSC and two P450s has also recently been identified in Lotus japonicus (Krokida et al., 2013). Gene clusters for the synthesis of the steroidal glycoalkaloids α-tomatine and α-chaconine/α-solanine have been discovered in Solanum lycopersicum (tomato) and Solanum tuberosum (potato), respectively (Itkin et al., 2013). The tomato gene cluster consists of eight characterized biosynthetic genes encoding for two P450s, four UGTs and two other tailoring enzymes (a dioxygenase and a transaminase). Similarly, the potato cluster contains six characterized genes encoding for two P450s, two UGTs, a dioxygenase and a transaminase. Although once thought to be an exception, the phenomenon of physical clustering of biosynthetic pathway genes for different types of secondary metabolites in plants is becoming common (Nützmann & Osbourn, 2014). The physical clustering of genes in the genome together with co-ordinate expression of the clustered genes is proving to be a useful tool to identify biosynthetic genes involved in new pathways for the synthesis of specialized metabolites (Castillo et al., 2013; Field & Osbourn, 2008; Field et al., 2011; Nützmann & Osbourn, 2014).

Enzymes encoded by multigene families for OSCs, P450s, UGTs and other specialized transferases are the key actors in the biosynthesis of plant triterpenoid, steroidal and steroidal glycoalkaloid saponins. Although the number of characterized enzymes belonging to these gene families has increased over the last decade, there is still a wide gap in our understanding of saponin biosynthesis, in how the synthesis of these molecules in different plant tissues and organs is regulated during development and in response to environmental stimuli, and in how these molecules are transported. Unraveling these molecular mechanisms will provide a better understanding of the role of saponins in plant processes, and will guide their exploitation in industrial applications.

Occurrence and distribution in plants

Saponins are synthesized and accumulated by many different plant families (Dinda et al., 2010; Sparg et al., 2004; Vincken et al., 2007). They occur in both wild and domesticated plants, with the triterpenoid-type saponins being predominant in cultivated crops. Some well-known triterpenoid saponin-accumulating plant families include the Leguminosae, Amaranthaceae, Apiaceae, Caryophyllaceae, Aquifoliaceae, Araliaceae, Cucurbitaceae, Berberidaceae, Chenopodiaceae, Myrsinaceae and Zygophyllaceae, among many others (Parente & Da Silva, 2009; Shi et al., 2004; Sparg et al., 2004). Legumes such as soybeans, beans and peas are a rich source of triterpenoid saponins. In contrast, cereals and grasses are generally deficient in saponins, with some notable exceptions such as Avena species (oats) which accumulate both triterpenoid and steroidal saponins (Osbourn, 2003). The steroidal saponins are typically found in members of the Agavaceae, Alliaceae, Asparagaceae, Dioscoreaceae, Liliaceae, Amaryllidaceae, Bromeliaceae, Palmae and Scrophulariaceae families and accumulate in abundance in crop plants such as yam, alliums, asparagus, fenugreek, yucca and ginseng (Hoffmann, 2003; Hostettmann & Marston, 2005). Diosgenin, the steroidal aglycone obtained by the hydrolysis of dioscin, a saponin abundant in the tubers of Dioscorea villosa (wild yam), is the precursor for the commercial synthesis of steroids such as cortisone, progesterone and pregnenolone (Patel et al., 2012; Peng et al., 2011). The steroidal glycoalkaloids are commonly found in members of the Solanaceae family including tomato, potato, aubergines and capsicum (Roddick & Melchers, 1985).

Several plants release saponins into the rhizosphere where these compounds might serve as allelopathic agents that suppress the growth of neighboring plants and prevent competition for natural resources. For instance, the germination, growth and yield of cotton was found to be lower when plants are grown on soil previously used for cultivation of Medicago sativa (alfalfa), compared with growth on soil after wheat cultivation or on uncultivated bare soil (Leshem & Levin, 1978). Similar effects were also observed for wheat. Alcoholic extracts of alfalfa roots suppressed germination of wheat seeds, and application of powdered alfalfa roots or compounds purified from alfalfa roots inhibited wheat seedling growth (Oleszek, 1993; Oleszek & Jurzysta, 1987). The antifungal triterpenoid saponins avenacins, which are synthesized in and released from oat roots, have also been suggested to have allelopathic potential, particularly because of the invasive nature of oat growth when compared to other cereals (Carter et al., 1999; Field et al., 2006; Hostettmann & Marston, 1995). A meta-transcriptomics-based analysis of bulk soil and rhizosphere of 4-week old pea, wheat and oat plants revealed major differences between their microbiomes. The relative abundance of nematodes and/or fungi was fivefold more in pea and oat rhizospheres than in the wheat rhizosphere and bulk soil. The pea rhizosphere was enriched for fungi, unlike the oat rhizosphere, which contained antifungal avenacins. Comparison of the rhizosphere microbiomes of oat wild-type and sad1 mutants (lacking avenacins) showed little differences in the prokaryotic community but large variation in the nematode and fungal community, suggesting a broader role for avenacins than just protecting from fungal pathogens (Turner et al., 2013).

In both monocotyledonous and dicotyledonous plants, the oleananes, one of the main classes of triterpenes derived from 2,3-oxidosqualene (Figure 1), represent the most common aglycone skeleton for the synthesis of triterpenoid saponins. The oleananes are pentacyclic triterpenoids that are found in almost all plant orders except the Solanales, Juglandales, Rhamnales and Zingiberales (Vincken et al., 2007). Although less abundant, the ursanes also occur in many plant families and are composed of a pentacyclic triterpene backbone that is almost identical to that of the oleanane backbone but differs in the location of one methyl group. The lupane and dammarane structures differ considerably from the oleananes and ursanes, and serve as aglycones for less common pentacyclic and tetracyclic saponins, respectively. The remaining classes of triterpenoid saponins, including the tirucallanes, hopanes, taraxasteranes, lanostanes and cucurbitanes are restricted to only a few plant families. The cycloartanes, also cyclized from 2,3-oxidosqualene, are an exception as they form the backbone for the synthesis of not only specialized metabolites (steroidal saponins and glycoalkaloids) but also phytosterols, and are present in all plants (Vincken et al., 2007). Furthermore, the phytosterols cholesterol, campesterol and sitosterol are the precursors of the brassinosteroid hormones that regulate plant growth and development (Fujioka & Yokota, 2003).

OSC genes encoding β-amyrin synthases for the cyclization of 2,3-oxidosqualene to the oleanane backbone β-amyrin have been identified in a number of different plant species (http://www.brenda-enzymes.org/php/result_flat.php4?ecno=5.4.99.39), including the model plant A. thaliana, which does not accumulate oleanane-derived saponins (Shibuya et al., 2009). In addition to β-amyrin, other simple triterpenes and triterpenoids have been detected as major components of the intracuticular wax layer in many plants (Buschhaus & Jetter, 2011). The widespread presence of simple triterpenes and triterpenoids in plants when compared with the taxon-specific accumulation of specialized glycosylated saponins in particular plant species suggests a possible fundamental role for these saponin biosynthetic precursors. These molecules may form constituents of plant membranes and may further have as yet unidentified signaling roles, given their structural similarities with sterols.

Saponins accumulate in an organ- and/or tissue-specific manner and the saponin content of plants is influenced by developmental factors, indicative of differential synthesis and regulation and suggesting distinct biological functions. Different types of saponins have been isolated from roots, stem, bark, leaves, seeds, fruits and flowers, and from whole plants (Vincken et al., 2007). In many plants, saponins are primarily synthesized and stored in underground organs. Glycyrrhiza glabra (licorice), Panax ginseng (ginseng) and the Chinese traditional medicine plant Polygala tenuifolia accumulate glycyrrhizin (G. glabra), ginsenosides (P. ginseng) and senegenin-derived saponins (P. tenuifolia) in the stolon, roots and/or rhizomes (Park et al., 2005; Seki et al., 2008; Teng et al., 2009). The roots of Bupleurum species and the model legume Medicago truncatula accumulate higher levels of saponins than the aerial organs (Huhman et al., 2005; Kapusta et al., 2005; Zhao et al., 2013). The total saponin content of M. truncatula roots is estimated at 5.92 µg/mg of dry weight, which is 5-, 10- and ∼200-times more than the saponin content of leaves and seeds, stems and seed pods, respectively (Huhman et al., 2005). Saponins can also accumulate in a tissue-specific manner in the roots of some plants. In P. ginseng, the ginsenosides are present in higher amounts in the phloem than the xylem, and are also found in parenchymal cells of the pith in the root (Fukuda et al., 2006; Yokota et al., 2011). In young Bupleurum roots, the saikosaponins are mainly in the pericycle and primary phloem, and in mature roots, they are distributed in the vascular cambium and secondary phloem (Tan et al., 2008). In oat roots, the UV-fluorescent saponin avenacin A-1 accumulates specifically in the epidermal layer of the root tips (Osbourn et al., 1994; Turner, 1960). However, saponins can also preferentially accumulate in aerial organs. Examples include the avenacoside steroidal saponins which are found in oat leaves, the asiaticosides in Centella asiatica leaves (Kim et al., 2005), the saponin tomatoside A in seeds of tomato (Moco et al., 2007), and saponins in flowers, fruits and seeds of Chenopodium quinoa (Kuljanabhagavad et al., 2008).

Simple triterpenes may also accumulate in a tissue- and organ-specific manner during plant growth. For example, in legumes, β-amyrin and lupeol have been detected in nodule tissues (Hartmann et al., 2002), and sprouting has been reported to increase both β-amyrin and saponin content of seedlings (Ayet et al., 1997; Baisted, 1971; Guajardo-Flores et al., 2012; Livingston et al., 1984). Likewise, in Solanum species, differential tissue- and organ-specific synthesis and accumulation of steroidal glycoalkaloids has been observed (Mweetwa et al., 2012). Synthesis commences during germination and peaks in the flowering period. The leaves are the first organs to attain maximum concentration, followed by unripe fruits and flowers which accumulate even higher amounts of the metabolites (Friedman, 2006). In potato tubers, the steroidal glycoalkaloids are concentrated in the peel with highest levels around the eyes constituting of the periderm, cortex and outer phloem (Friedman et al., 2003). The model plant A. thaliana has 13 OSCs which catalyze the cyclization of 2,3-oxidosqualene to different major triterpene products. These OSC genes differ in their organ-specific expression patterns, suggesting specific functions for the different triterpene products in different organs during different stages of plant growth and development (Thimmappa et al., 2014; Figure 2A).

The isoprenoids for the synthesis of both sterols and specialized triterpenes are derived from the shared mevalonate pathway, and there is evidence for inverse regulation of the biosynthetic pathways for the synthesis of phytosterols and triterpenes in plants. During the early stages of pea germination, sterol synthesis is very low and β-amyrin production is predominant. During development, however, sterol synthesis is prevalent. In the late fruit maturation period, sterol synthesis decreases considerably and there is significant synthesis of β-amyrin (Baisted, 1971). Similarly, sorghum grains accumulate β-amyrin early in root development, while sterol biosynthesis becomes active only later (Palmer & Bowden, 1977). The end products of sterol and triterpenoid biosynthesis also manipulate the flux through these biosynthetic pathways (Haralampidis et al., 2001b). For example, feeding radiolabeled acetate to Gypsophila paniculata and Saponaria officinalis cell suspension cultures pre-treated with gypsogenin 3,O-glucuronide (a saponin precursor in G. paniculata, but not S. officinalis) resulted in reduced incorporation of radioactivity in saponins and their precursors, but not sterols in G. paniculata. Compared to untreated controls the ratio of β-amyrin to cycloartenol reduced by ∼10-fold in G. paniculata but remained unchanged in S. officinalis (Henry et al., 1992). Also, in cell suspension cultures of Tabernaemontana divaricate, inhibition of the activity of cycloartenol synthase resulted in the induction of β- and α-amyrin synthases, which in turn increased flux through the triterpenoid pathway (van der Heijden et al., 1989). In many cases, there is a good correlation between the accumulation of metabolites and the expression pattern of the respective biosynthetic genes. In Glycyrrhiza glabra seedlings, the expression of the β-amyrin synthase gene is maximal two days after germination and thereafter decreases rapidly, whereas the expression of the cycloartenol synthase gene remains relatively constant throughout germination (Hayashi et al., 2004). In oat roots, a marked difference is observed in the spatial expression patterns of the cycloartenol synthase and β-amyrin synthase genes, which are required for synthesis of sterols and triterpenoid saponins, respectively. In young oat roots, the β-amyrin synthase gene Sad1 is expressed in the epidermal cells of the root tips, in accordance with the accumulation of the β-amyrin derived avenacins. In contrast, expression of the cycloartenol synthase gene CS1 is restricted to the cortex in the elongation zone above the root tip (Figure 2B; Wegel et al., 2009).

Genotype- and cultivar-dependent variation in the saponin content of plants has also been reported. For instance, the grain saponin content of different cultivars of C. quinoa is estimated to range from 0 to ∼10 mg/g dry weight, and based on the total saponin content these cultivars are classified as bitter (high saponin) or sweet (low saponin) (Koziol, 1992). Between the two Centella species, C. asiatica and C. glabrata, the triterpenoid saponin asiaticoside and related compounds are major constituents in the former species and absent in the latter (Long et al., 2012). Similarly, a systematic analysis of the triterpene glycosides asiaticoside and madecassoside in sixty C. asiatica accessions from South India grown under identical environmental conditions revealed huge variations in the content of these saponins, with three accessions having no detectable asiaticoside/madecassoside (Thomas et al., 2010). Analysis of the saponin composition of 800 cultivated (Glycine max) and 329 wild (Glycine soja) soybean seeds showed clear variety-specific differences between the accessions. These differences were attributed to variety-specific expression of biosynthetic genes that utilize soyasapogenol glycosides as substrates (Tsukamoto et al., 1993). In G. max, the levels of group B soyasaponins represent one of the most discriminatory metabolic indicators of closely related salt-sensitive and salt-tolerant soybean genotypes (Wu et al., 2008). Large variation in the seed saponin composition of 3025 G. soja accessions from nine regions of Korea has also been reported. Based on the saponin composition of the seed hypocotyls, the accessions were grouped into seven phenotypes (Panneerselvam et al., 2013). In addition, comparison of six Cajanus cajan (chickpea) and five Phaseolus vulgaris (common bean) genotypes showed genotypic variations in their sapogenol content (Vasishtha & Srivastava, 2011, 2012).

In addition to genetic determinants of saponin content, environmental and agronomic factors may also influence the saponin composition of plants (Etebu, 2012; Schwarzbach et al., 2006). For example, potato tubers show pronounced drought- and light-induced accumulation of steroidal glycoalkaloids in their outer layers when stored in the light instead of the dark (Bejarano et al., 2000; Percival, 1999). In contrast, ginsenoside biosynthesis in P. ginseng is enhanced in the dark (Kim et al., 2014). Seasonal variation in the expression of the β-amyrin synthase gene and the content of glycyrrhizin has also been observed in G. glabra plants (Hayashi et al., 1998, 2004). The combination of environmental stress and agricultural factors such as water-logging or drought, fertilization and low or high temperatures influences the content of triterpenoids and saponins in plants. In the medicinal herb Hypericum brasiliense, betulinic acid content increases in response to drought and low or high temperatures (Nacif de Abreu & Mazzafera, 2005). Drought stress significantly increases the production of the triterpenoids rosmarinic, ursolic and oleanolic acid in spicas (reproductive parts) of Prunella vulgaris (Chen et al., 2011). Also in Bupleurum chinense, mild water stress increases the saikosaponin content of roots (Zhu et al., 2009).

The occurrence and distribution of saponins in plants in a tissue- and organ-specific manner at different growth and development stages suggests that these metabolites may have functions in physiological processes in plants, in addition to their traditionally accepted roles as phytoalexins and defense molecules.

Effects on plant growth and development

The ability of different plant species to synthesize distinct secondary metabolites is generally thought to be essential for their survival. Indeed, the saponin content and composition of a plant is variable and influenced by its surrounding environment and growth stage (Szakiel et al., 2011). Stress- or developmentally regulated changes in saponin composition contribute to the total plant secondary metabolite pool where they may ensure protection of plant organs and overall plant survival. However, as described above, in many plants, saponins and their biosynthetic intermediates accumulate during normal growth and development. Manipulating the biosynthesis of these molecules causes morphological and physiological effects, suggesting additional functions (Abe et al., 1988; Guhling et al., 2006). In this section, we present growing evidence linking the saponins and their biosynthetic intermediates to various physiological processes during plant growth, development and maturation (Figures 3 and ​and4).4). We also highlight the impact of modulating saponin synthesis and accumulation in plants and suggest their underexplored primary role in plant well-being.

Effects of exogenous application

The external application of saponins to plants impacts different biological processes (Geuns, 1978). Seeds of pea and corn absorb water more rapidly in the presence of saponins, leading to more rapid germination. Crude saponin extracts from Quillaja, Polygala, Saponaria and Sapindus species can double the growth rate of isolated wheat and pea embryos (Helmkamp & Bonner, 1953). Asiaticosides extracted from Centella asiatica stimulate growth and chlorophyll biosynthesis in radish, pea, Lupinus and Linum species (Boiteau & Ratsimamanga, 1956). Saponins from pea have been proposed to regulate gravitropism and cellulose synthesis in different plant species (Ohana et al., 1998; Rahman et al., 2001). Chromosaponin I, isolated from 7-day-old etiolated Pisum sativum seedlings, stimulates cortical cell division and elongation leading to enhanced growth of roots in A. thaliana, Lactuca sativa, Chrysanthemum coronarium, Medicago sativa, Trifolium repens, Astragalus sinicus, Brassica juncea, Brassica campestris, Cryptotaenia japonica, Phleum pretense and Lolium multiflorum (Rahman et al., 2000; Tsurumi & Wada, 1995). In addition to enhanced elongation of roots, a reduction in root diameter was also observed in A. thaliana (Figure 3A) and L. sativa, suggesting a potential growth regulatory role for chromosaponin I, which appears to be modulated by the auxin influx carrier protein AUX1 (Rahman & Tsurumi, 2002; Rahman et al., 2001). In contrast, betulin and its glycosides show growth inhibitory effects on alfalfa radicles, with the inhibitory activity tending to increase with an increase in the number of β-d-glucose residues. The hypocotyls were less affected than the radicles for all tested betulin derivatives, suggesting that these compounds primarily affect the roots (Figure 3B). In the absence of sugar molecules, the triterpenoid betulinic acid, with a carboxyl group at carbon position 28, showed stronger inhibitory activity than betulin, which has a hydroxyl group at the same position. Oleanolic acid, also with a carboxyl group on position 28, but having an oleanane instead of a lupane backbone (Figure 8), showed no growth effects on alfalfa radicles or hypocotyls (Ohara & Ohira, 2003).

Open in a separate window

Figure 8.

Overview of chemical structures. (A) Carbon and ring numbering of oleanane, ursane, lupane and dammarane aglycones. (B) Examples of triterpenoids and saponins discussed in structure-activity relationship studies. α-l-Ara, α-l-arabinose; α-l-Rha, α-l-rhamnose; Glc, glucose; Xyl, xylose.

In contrast to the effects observed with triterpenoid saponins, which mainly affect the roots, the exogenous application of the steroidal glycoalkaloid saponin α-tomatine inhibits stem elongation and chlorophyll accumulation in seedlings of Sesbania exaltata, Senna obtusifolia, Vigna radiata, Triticum aestivum and Sorghum vulgare (Figure 3C). α-Tomatine also causes electrolyte leakage from leaf discs of Zea mays, Cucumis sativus, Nicotiana tabacum and Nicandra physaloides, but not from tomato and potato tissues (Hoagland, 2009). In comparison, the aglycone tomatidine causes greater electrolyte leakage in Z. mays, C. sativus, N. tabacum and N. physaloides. Application of pure tomatidine to A. thaliana seedlings caused root growth effects without affecting their germination rate. The roots of seedlings grown on medium supplemented with tomatidine were shorter than those of seedlings grown on control medium or medium supplemented with α-tomatine (Itkin et al., 2011).

Industrial potential as bioactive principles

With the broad diversity in structures and biological activities of saponins and their biosynthesis intermediates, these compounds have been sought for various applications in the pharmaceutical, cosmetic, food and agronomic industry (Augustin et al., 2011; Moses et al., 2013; Sparg et al., 2004). Saponins from Camellia oleifera and Sapindus mukorossi are useful in detergents and cosmetics as they have excellent foaming properties, moderate detergency and anti-dermatophytic activity (Chen et al., 2010; Tamura et al., 2012; Yang et al., 2010). Quil A, a mixture of over 25 different saponins extracted from Quillaja saponaria, is the most widely used saponin adjuvant in veterinary vaccines (Dalsgaard, 1978). Yucca schidigera and Q. saponaria extracts are used as livestock dietary additives for fecal odor control (Cheeke, 2009; Killeen et al., 1998). Several C-27 steroidal saponins show antifungal activity (Morrissey & Osbourn, 1999; Yang et al., 2006), and glycyrrhizin accumulating in licorice roots is an efficient natural sweetener (Kitagawa, 2002). The wide spectrum of biological activities of this specialized class of compounds has been attributed to the diverse array of structures synthesized naturally by plants. Although such structural diversity is beneficial, the vast spectrum of biological activities can prove to be antagonistic to their application, particularly in the pharmaceutical sector. For instance, many saponins are cytotoxic and therefore potential chemotherapeutic agents, but also have hemolytic activity hindering their application as anti-cancer drugs (Podolak et al., 2010).

Several studies have hence focused on generating semi-synthetic derivatives inspired by the naturally accumulating saponins and their biosynthetic intermediates to generate a new range of candidate drugs with enhanced activity and reduced adverse effects. For example, synthetic derivatives of the pentacyclic triterpenoids oleanolic, ursolic and betulinic acid, which naturally accumulate to high concentrations in plants, have been evaluated to be strikingly more active than the parent compounds in several clinical applications (Liby & Sporn, 2012; Liby et al., 2007a,b). The semi-synthetic derivatives of oleanolic acid (Figure 6), bardoxolone (CDDO) and bardoxolone methyl (CDDO-Me) were clinically evaluated as anti-cancer agents (Hong et al., 2012; Tsao et al., 2010). CDDO-Me was also clinically tested for anti-inflammatory activity in patients with chronic kidney disease and type 2 diabetes to improve kidney function (Pergola et al., 2011). However, the worldwide phase III trial of the drug was terminated following severe adverse effects and mortality in patients administered the drug (http://www.clinicaltrials.gov/show/{"type":"clinical-trial","attrs":{"text":"NCT01351675","term_id":"NCT01351675"}}NCT01351675). It has been speculated that CDDO and its derivatives interact with multiple targets or entire networks in the cell, and therefore might be most effective in the early stages of disease progression prior to irreversible tissue damage, rather than at advanced stages of disease (Liby & Sporn, 2012; Sporn et al., 2007). The patients in the phase III trial with CDDO-Me had moderate-to-severe chronic kidney disease with type 2 diabetes. A detailed analysis of data collected during this terminated trial can provide essential insight into the relation between adverse effect, mortality and dose of CCDO-Me given to patients, which can then be utilized for future tests with other CDDO analogues (Abboud, 2013). The imidazolide and ethylamide derivatives of bardoxolone, CDDO-Im and CDDO-EA, respectively, induce chondrogenic differentiation in organ cultures and could be used for the treatment of diseases where enhancing chondrogenesis is therapeutically useful, including the prevention or treatment of osteoarthritis (Suh et al., 2012). Multiple in vitro and in vivo studies with CDDO-Me have also demonstrated the effect of this compound in the prevention and treatment of solid tumors (Wang et al., 2014). The combination of CDDO-Me or CDDO-EA with vorinostat, a histone deacetylase inhibitor, effectively delayed tumorigenesis in mouse models, proving to be a promising combination for chemoprevention (Tran et al., 2013). Another semi-synthetic derivative evaluated clinically in phase I and II trials is bevirimat (Figure 6). This betulinic acid derivative is a potent inhibitor of human immunodeficiency virus (HIV) maturation, which showed promising clinical outcomes (Smith et al., 2007). Other derivatives of betulin, betulinic acid and dihydrobetulin also have demonstrated potent anti-HIV and anti-leishmanial activities (Sousa et al., 2014; Xiong et al., 2010).

Naturally, accumulating plant compounds also have biological activities of interest for clinical applications and are routinely evaluated for potential novel therapeutics. YQ2, a 6:4 (v:v) blend of partially purified Y. schidigera and Q. saponaria extracts, has been clinically studied as an anti-obesity medicine. Administration of YQ2 to 51 hypercholesterolemic patients showed a decrease in total and low-density lipoprotein cholesterol levels in blood plasma, compared to 36 patients given a placebo treatment. YQ2 also improved gastrointestinal motility in the treated patient group, thereby enhancing gastrointestinal functions (Kim et al., 2003). The triterpenoid saponin escin (Figure 6), isolated from seeds of Aesculus hippocastanum (horse chestnut), has been clinically evaluated in gel formulations as combination therapy with diethylammonium salicylate and heparin. The gel combination of the three active principles showed positive effects on tenderness reaction, edema formation and joint function, compared to placebo. This randomized, double blind, controlled, group comparative study with 158 patients, topically treated or not with escin, concluded that the gel combination containing escin was well tolerated and led to rapid pain reduction, implicating its recommended use for the treatment of blunt impact injuries (Wetzel et al., 2002). Celastrol (Figure 6), extracted from roots of the traditional Chinese herb Tripterygium wilfordii, has potent anti-inflammatory and anti-cancer activities, and holds great potential for the treatment of multiple myelomas and other hematological malignancies (Chiang et al., 2014; Kannaiyan et al., 2011). QS-21 (Figure 6), is a soluble fraction of triterpenoid saponins from Q. saponaria with immunological adjuvant activity. QS-21 is currently being clinically evaluated in phase I and II studies for use in human vaccines against HIV (Leroux-Roels et al., 2014), Mycobacterium tuberculosis (Montoya et al., 2013), varicella-zoster virus (Lal et al., 2013), cancer (Garçon & Van Mechelen, 2011) and malaria (Ndungu et al., 2012).

Despite the various potential applications of saponins, their biosynthetic intermediates and derivatives, very few compounds are currently commercially exploited. A major concern hindering their widespread application as therapeutics is the lack of information about the mechanisms of adsorption, distribution, metabolism, excretion and toxicity of these compounds in human and animal cells. Well-designed, associated, systematic in vitro and in vivo studies will play a key role in deepening our understanding of the mechanisms involved in the action of saponins, before their widespread usage as drugs in the future.

Exploring structure–activity relationships

Saponins, their biosynthetic intermediates and derivatives have been ascribed a number of pharmacological activities (Figure 7), most notably permeabilization of cell membranes, cytotoxic effects on cancerous cells, adjuvant properties, immunomodulatory potential and lowering of serum cholesterol (Francis et al., 2002; Hostettmann & Marston, 2005; Man et al., 2010; Sjölander et al., 1998; Sun et al., 2009). Despite their structural diversity these compounds share distinct biological properties, and therefore several studies have focused on understanding the various biological activities associated with them in relation to their structural features. Identification of effective combinations of key functional entities and aglycones responsible for specific bioactivities allows targeted synthesis of novel active principles with enhanced efficacy and reduced adverse effects. Here, we summarize some of the most investigated structure–activity relationships of these compounds.

Concluding remarks

Simple and elaborated triterpenoids affect fundamental cellular processes both in plants and animals. The molecular targets that are subject to the action of these compounds have not yet been identified in plants. Nevertheless, there is a remarkable functional conservation of groups of proteins that regulate cell proliferation, cell fate decisions and signal transduction pathways in plants and mammalian systems (Jonak & Hirt, 2002). The non-exhaustive series of proven protein targets of simple triterpenes in mammalian cell lines could be used as a navigation tool in comparative trans-kingdom studies to illustrate the physiological role of saponins in processes pertaining to multicellular organization. On the other hand, knowledge acquired by deciphering the biosynthetic pathways of saponins in plants can provide vital clues to the role of saponins and their specialized intermediates in plant well-being, as well as provide pivotal elements for the rational production of new drugs in synthetic biology approaches. The identification of enzymes responsible for the functionalization of precursors at biologically relevant positions may be exploited for structure-guided compound synthesis in heterologous microbial, plant and animal hosts. Moreover, the identification of regulatory proteins such as transcription factors, protein kinases and phosphatases involved in saponin biosynthesis can be used to augment natural product synthesis in the native plant species. Ancillary to this, systematic studies of the environmental and physiological factors and regulatory proteins that modulate the synthesis of this family of compounds are likely to reveal further as yet unidentified roles for these developmentally and industrially relevant class of compounds.

Acknowledgements

Footnotes

References