Introduction

The genus Allium (Amaryllidaceae) is one of the largest monocot genera comprising more than 800 species (Li et al. 2010; APG 2009). It is widely distributed in nature and has adapted to diverse habitats across the Holarctic region, with the exception of A. dregeanum, which is native to South Africa (Li et al. 2010). Some Allium species, such as garlic, onion and leek, are widely cultivated as vegetable products, spices and for medical purposes. The most characteristic constituents in Allium plants are sulfur compounds, which are the most important substances both in terms of chemotaxonomic value and biological activity (Rose et al. 2005). However, various researchers tend to attribute the potential pharmacological benefits of Allium plants to constituents other than sulfur compounds, such as steroidal saponins. Also, polyphenolic compounds, especially flavonoids, as well as fructans, N-cynnamic amides, and antioxidative enzymes are considered to be equally important (Matsuura 2001; Lanzotti 2005; Štajner et al. 2006; Amagase 2006; Lanzotti 2012).

Apart from the Amaryllidaceae family, steroidal saponins are widely distributed in other monocot families: Asparagaceae (Agave, Asparagus, Convallaria, Hosta, Nolina, Ornithogalum, Polygonatum, Sansevieria, Yucca), Costaceae (Costus), Dioscoreaceae (Dioscorea), Liliaceae (Lilium), Melanthiaceae (Paris), Smilacaceae (Smilax). Interestingly, these compounds have been reported as well in some dicotyledonous angiosperms: Zygophyllaceae (Tribulus, Zygophyllum), Solanaceae (Solanum, Lycopersicon, Capsicum), Plantaginaceae (Digitalis) and Fabaceae (Trigonella).

There are numerous reports referring to pharmacological activities of steroidal saponins. Some of them showed promising antifungal, cytotoxic, anti-inflammatory, antithrombotic, and hypocholesterolemic effects (Sparg et al. 2004; Lanzotti 2005; Güçlü-Üstündağ and Mazza 2007). Most importantly, these compounds are used as substrates in the production of steroid hormones and drugs.

Steroidal sapogenins and saponins have been identified so far in over 40 different Allium species. The earliest reports on Allium saponins date back to the 1970s and dealt with identification of diosgenin in A. albidum (Kereselidze et al. 1970) and alliogenin in the bulbs of A. giganteum (Khristulas et al. 1970). Further studies performed worldwide in the following years led to the isolation of a large number of new compounds. The first chemical survey of saponins from the genus Allium was published by Kravets in 1990, and this was followed by an update by Lanzotti in 2005 (Kravets et al. 1990; Lanzotti 2005). Since then, a large number of new compounds has been discovered, and there were also some that have not been included in the previous surveys.

A recent review by Lanzotti et al. (2014) compiled data on various compounds identified in Allium species with a reported cytotoxic and antimicrobial activity.

The present review is predominantly focused on the chemistry of Allium steroidal saponins and their biological activities.

Chemistry of Allium saponins

Steroidal saponins from the genus Allium can be divided into three groups on the basis of the sapogenin structure: spirostanols, furostanols, and open-chain saponins. The latter group is often referred to in the literature as “cholestane saponins” (Challinor and De Voss 2013). Allium saponins are mostly mono- or bidesmosides, however a tridesmodic cholestane glycoside has been reported in the bulbs of A. macleanii (Inoue et al. 1995). The sugar residue in Allium saponins consists of linear or branched chains made up most often of glucose (Glc), rhamnose (Rha), galactose (Gal), xylose (Xyl), and arabinose (Ara) units.

Spirostane-type saponins

A vast structural diversity of Allium spirostanols is associated with the differences in the structure of aglycones, especially their oxygenation patterns and stereochemistry (Table 1). In spirostane-type sapogenins, the steroid A/B ring junction is found mostly in a trans (5α), or more rarely in a cis (5β) fusion (e.g. anzurogenin A [48] and C [58]). Δ5(6) unsaturation is considered to be a quite common feature (diosgenin [4], ruscogenin [17], yuccagenin [19], lilagenin [20], cepagenin [44], karatavigenin C [45]). However, a double bond located at C25(27) was reported in the aglycones of saponins present in A. macrostemon and in one of the sapogenins identified in A. ursinum bulbs (He et al. 2002; Sobolewska et al. 2006; Cheng et al. 2013). The C-25 methyl group is found with either S or R absolute configuration. In many cases the isolated sapogenins appear to be a mixture of diastereomers R and S.

Table 1 Spirostane-type sapogenins identified in the genus Allium

The most common spirostanol sapogenins identified in Allium plants are: diosgenin [4], tigogenin [1], gitogenin [9], agigenin [34], alliogenin [49], and β-chlorogenin [12]. It was claimed that β-chlorogenin, a genin present in common garlic A. sativum, could be considered as a chemical marker for its identification in various food products, as the characteristic garlic sulfur compounds are very unstable (Itakura et al. 2001).

Until now, over 130 spirostanol glycosides have been identified in various Allium species. It should be mentioned however that some of these compounds were obtained as a result of enzymatic hydrolysis of furostanol saponin fraction by β-glucosidase (Ikeda et al. 2000).

Allium spirostane-type saponins are typically monodesmodic with the sugar residue usually at C-3 position. In rare cases, the sugar moiety was reported to be linked at other positions, such as C-1 (e.g. alliospirosides A-D [169, 170, 178, 179]) (Kravets et al. 1986a, b, 1987), C-2 (compounds from A. giganteum and A. albopilosum) (Sashida et al. 1991), C-24 (chinenoside VI [116], karatavioside F [181], and anzuroside [190]), or C-27 (tuberoside L [104]) (Jiang et al. 1998; Vollerner et al. 1984; Vollerner et al. 1989; Sang et al. 2001a).

Table 3 of ESM summarizes chemical structures of spirostane-type saponins that were reported in Allium species.

Furostane-type saponins

Furostanol aglycones possess either a cis or a trans fusion between ring A and B, or a double bond between C-5 and C-6 leading to 5α, 5β or Δ5(6) series. In the case of furostane-type sapogenins a double bond may also be located at 20(22) (e.g. ascalonicoside B [220], ceparoside C [230], chinenoside II [234]) or 22(23) (four furostanols from A. tuberosum) (Fattorusso et al. 2002; Yuan et al. 2009; Peng et al. 1996b; Sang et al. 2001b). The 27-Me group may be in either R or S configuration. Furostane-type compounds isolated from Allium species usually possess an OH or OMe group at C-22. However, sapogenins with a C-22 methyl ether are considered to be artifacts resulting from the use of methanol in the extraction/isolation procedures.

From among 140 furostanol glycosides identified in the Allium genus, sixteen compounds were found to be such methoxy-derivatives.

Furostanol saponins in Allium plants are bidesmodic glycosides with sugar chains attached usually at C-3 and C-26 positions. A rare glycosylation at C-1 with a galactose unit was reported in ascalonicosides A1/A2 [217, 218] (Fattorusso et al. 2002). A vast majority of furostanol saponins possess an O-linked glucose residue attached at position C-26. In compounds such as ceposides, persicoside C [205, 206], ascalonicosides A1/A2 [217, 218] a disaccharide chain was reported at C-26 (Lanzotti 2012; Sadeghi et al. 2013; Fattorusso et al. 2002).

Cholestane-type (open-chain) saponins

A review of available literature data shows that as much as 18 cholestane-type compounds have been identified in ten different Allium species.

Allium open-chain aglycones possess Δ5(6) unsaturation with an exception of schubertoside A [329]Δ4(5), and one of the glycosides found in A. albopilosum with a saturated aglycone (Kawashima et al. 1991b; Mimaki et al. 1993). Glycosides based on alliosterol—(22S)-cholest-5(6)-ene-1β,3β,16β,22-tetrol (Fig. 1 [196]), or related sapogenins showing the same oxygenation pattern at C-1, C-3, C-16 and C-22 are most common (Challinor and De Voss 2013). Sugar units are attached at one, two or, more seldom, at three separate positions (in A. macleanii) (Inoue et al. 1995). Most of these compounds are glycosylated at C-16, whereas in contrast to spirostanol and furostanol saponins, the attachment of sugar chain at position C-3 is almost unique (tuberoside U [353]) (Sang et al. 2003).

Fig. 1
figure 1

Alliosterol—(22S)-cholest-5(6)-ene-1β,3β,16β,22-tetrol [196]

Table 2 lists steroidal saponins/sapogenins identified in Allium species. Plant names are cited exactly as they were referred to in the original report. It is almost certain that some of them are synonyms but as the authors of the present review are not specialists in plant taxonomy no amendments have been made.

Table 2 List of steroidal saponins/sapogenins reported in Allium species

Biological and pharmacological properties of Allium saponins

Saponins are considered responsible for numerous pharmacological properties of many plants, and they are recognized as active constituents of Allium species as well. It should be mentioned, however, that Allium plants are not rich sources of these compounds. Results from quantitative studies indicate that saponin content is usually very low, for example A. nigrum total saponin content in different parts of the plant was determined as: 19.38 mg/g dw in the roots, 15.65 mg/g dw—bulbs, and 10.48 mg/g dw—leaves (Mostafa et al. 2013). Quantitative densitometric determination of diosgenin—the main sapogenin of A. ursinum, revealed some differences in its accumulation with respect to the vegetation period, nevertheless its highest percentage observed in the bulbs collected in March did not exceed 0.0029 % of fresh weight (Sobolewska et al. 2009). A significant exception, in terms of saponin content, is A. nutans, where the concentration of these compounds in the underground parts was established to be about 4 % of dry matter (Akhov et al. 1999).

It should be emphasized however that the results from many pharmacological in vitro and in vivo studies revealed several interesting activities of Allium saponins, for example antifungal, cytotoxic, antispasmodic, hypocholesterolemic, and other.

Cytotoxic properties

Cytotoxic activity of saponins was discussed in a number of experimental papers on Allium species. In vitro studies were performed on several human and animal cell cancer lines, including IGR-1—human melanoma cell line; HL-60—promyelotic leukemia cells; HCT-116, HT-29, and SW480—human colorectal cancer cell lines; DLD-1—human colon adenocarcinoma, HA549—lung cancer cell line, NCI-H460—human large-cell lung carcinoma, SF-268—human glioblastoma; MCF-7—human breast adenocarcinoma, HepG2—human hepatocellular liver carcinoma cell line; WEHI 164—murine fibrosarcoma cell line; J-774—murine monocyte/macrophage cell line; P-388 and L-1210—murine leukemia cell lines (Table 4 of ESM). Amongst tested spirostane saponins dioscin [135], isolated from A. ampleloprasum, seemed to be most potent, with an IC50 = 0.092 μg/mL against P388 cell line (Sata et al. 1998). This compound, which is widely distributed in species of the family Dioscoreaceae and Asparagaceae, revealed significant in vitro activity in tests performed on many other cancer cell lines (Podolak et al. 2010). Some authors claim that apart from the type of the cell line, the structure of the oligosaccharide chain, especially the site of interglycosidic linkages, rather than the sapogenin, are the modulating factors of cytotoxic properties (Rezgui et al. 2014). Some evidence that may substantiate such claims comes from the results obtained for a mixture of diosgenin tetrasaccharide and (25R)-spirost-5(6),25(27)-diene-3β-ol tetrasaccharide [141, 156] (A. ursinum) (Sobolewska et al. 2006). The sugar chain of these compounds differs from that of dioscin [135] (3-O-α-l-Rha-(1 → 2)-[α-l-Rha-(1 → 4)]-O-β-d-Glc) by an additional terminal rhamnose moiety. Both exhibited 100 % effect already at the concentration of 2 μg/mL on melanoma B16 and sarcoma XC. Similarly, deltonin [134] (diosgenin 3-O-β-d-Glc-(1 → 4)-[α-l-Rha-(1 → 2)]-O-β-d-Glc) isolated from A. schoenoprasum showed significant activity against HCT 116 and HT-29 cell lines with an IC50 = 0.40 and 0.75 μM, respectively (Timité et al. 2013). These results corroborate with those obtained by Mimaki et al. (2001), who suggested that an α-l-Rha-(1 → 2)-O-β-d-Glc sugar sequence attached to diosgenin is crucial for activity (Mimaki et al. 2001).

The most potent spirostanol glycosides include also eruboside B [79], leucospiroside A [97], yayoisaponin C [95] and aginoside [93] isolated from A. leucanthum, which showed in vitro cytotoxic activity, with relatively similar IC50 values against A549 WS1, and DLD-1 cells (Mskhiladze et al. 2008b). The two latter compounds, that were isolated from A. ampeloprasum, showed in vitro cytotoxicity against P388 cells at 2.1 μg/mL (Sata et al. 1998). Tigogenin pentasaccharide [67] (A. macleanii) and diosgenin 3-O-α-l-Rha-(1 → 2)-[β-d-Glc-(1 → 3)]-O-β-d-Glc [140] (A. senescens) were cytotoxic towards HeLa cells at the concentration of 50 μg/mL, whereas already at 5 μg/mL they exhibited 64.7 and 11.5 % inhibition, respectively (Inoue et al. 1995). Several spirostanol glycosides, that were isolated from different Allium species, revealed fairly high cytotoxic activity in tests on promyelotic leukemia cells HL-60. Yuccagenin tetrasaccharide (karatavioside A [151]) from the bulbs of A. karataviense exhibited considerable cytostatic activity with an IC50 value of 2.4 μg/mL as compared with etoposide (IC50 0.3 μg/mL) (Mimaki et al. 1999c). Tuberoside M [163] from the seeds of A. tuberosum inhibited the cells growth with IC50 = 6.8 μg/mL, while F-gitonin [72] isolated from the fresh bulbs of A. jesdianum—with an IC50 value of 1.5 μg/mL (Sang et al. 2002; Mimaki et al. 1999a). Other compounds isolated from this latter species were considered to be inactive. The authors concluded that the presence of an additional OH group at C-6 in gitogenin skeleton is detrimental to activity, while cholestane glycosides showed no effect. It is probable that the presence of a carbonyl at C-6 in a laxogenin glycoside [158] isolated by Timité et al. (2013) from the whole plant of A. schoenoprasum could be responsible for the loss of activity against two cancer cell lines HCT 116 and HT-29, an effect similar to that seen by Mimaki et al. when an additional OH group was introduced at C-6 of gitogenin (Timité et al. 2013; Mimaki et al. 1999c). In accordance with the studies of Mimaki et al. (1999a, b, c) were also the results obtained for cholestane glycosides, nigrosides C [303] and D [304] isolated from the bulbs of A. nigrum, which showed no effect (IC50 > 100 μM) on the HT-29 and HCT-116 cancer cell lines in the MTT assay (Jabrane et al. 2011). Opposite results were obtained however with two cholestane glycosides isolated from A. porrum—alliosterol 1-O-α-l-Rha 16-O-β-d-Glc [267] and alliosterol 1-O-β-d-Glc-(1 → 4)-O-α-l-Rha 16-O-β-d-Gal [308], which exhibited in vitro cytotoxic properties (IC50 4.0–5.8 μg/mL) against two murine cell lines: WEHI 164 and J-774 (Fattorusso et al. 2000).

Results of cytotoxicity assays of several spirostanol sapogenins indicated their weak activity or lack of it. Agigenin [34], porrigenin A [38] and porrigenin B [23] identified in A. porrum tested in vitro for their growth-inhibitory activity on four different cell lines (IGR-1, WEHI 164, J-774, and P-388) exhibited much weaker activity when compared with 6-MP and were virtually inactive (>100 μg/mL) (Carotenuto et al. 1997a). However, some of the steroidal glycosides isolated from the same plant exhibited quite a good activity towards J-744 and WEHI-164 cells, the most active being gitogenin and porrigenin C derivatives (IC50 ranging from 1.9 to 5.8 μg/mL) (Fattorusso et al. 2000).

From among tested furostanoles the majority of compounds showed weak activity or lack of it, for example two glycosides isolated from A. tuberosum showed no activity at concentrations below 5 μM against PC-12 and HCT-116 (Ikeda et al. 2004). Among numerous furostanoles obtained from A. macrostemon which were tested against NCI-H460, SF-268, MCF-7, and HepG2 cell lines, exclusively 26-O-β-d-Glc 5α-furost-25(27)-ene-3β,12β,22,26-tetrol 3-O-β-d-Glc-(1 → 2)-[β-d-Glc-(1 → 3)]-O-β-d-Glc-(1 → 4)-O-β-d-Gal [292] was found cytotoxic towards SF-268 cell line, while 26-O-β-d-Glc 5β-furost-20(22),25(27)-diene-3β,12β,26-triol 3-O-β-d-Glc-(1 → 2)-O-β-d-Gal [293] showed cytotoxicity towards SF-268 and NCI-H460 cell lines (Chen et al. 2009).

The differences in activity between compounds having the same aglycone but differing in sugar chain was observed by Zolfaghari et al. (2013). The equilibrated mixture of furostanols: vavilosides A1/A2–B1/B2 [355–358] and ascalonicosides A1/A2 [217, 218] isolated from A. vavilovii were tested against cell lines: J-774 and WEHI-164. The activity of all saponins was dose-dependent and varied in the following order: vavilosides B1/B2 > ascalonicosides A1/A2 > vavilosides A1/A2 (Zolfaghari et al. 2013). The substitution of a galactose residue (vavilosides A1/A2) with a xylose unit (vavilosides B1/B2) caused an increase in cytotoxic activity.

Antifungal activity

Numerous steroidal saponins isolated from different plant sources have been reported to have antifungal/antiyeast activity, particularly against agricultural pathogens. Antifungal saponins require particular attention as there is a constant need for new agents that would be effective against opportunistic fungal infections and could provide an alternative to chemical fungicides used in the fight against plant pathogens. Unfortunately, only a few studies have been performed so far on Allium steroidal glycosides.

Antifungal activity of Allium saponins was modulated by both the sapogenin type and the number and structure of the sugar residue. Generally saponins with spirostanol skeleton exhibited higher antifungal activity than furostanols. Yu et al. (2013) observed several biochemical changes which could be involved in the possible mechanism of antimicrobial activity of saponins, such as reduced glucose utilization rate, decrease of catalase activity and protein content in microorganisms.

The results from in vitro assays against different plant and human pathogen strains are provided in Table 5 of ESM.

Studies by Barile et al. (2007), Lanzotti et al. (2012a, b), and Sadeghi et al. (2013) provide evidence for significant differences in the potency of saponins belonging to furostane or spirostane groups. Minutosides A-C [295, 119, 296] (A. minutiflorum) showed concentration-dependent antifungal activity against a number of pathogens: Alternaria alternata, A. porri, B. cinerea, Fusarium oxysporum, F. solani, Pythium ultimum, R. solani, Trichoderma harzianum P1, T. harzianum T39 (Barile et al. 2007). The most pronounced effect was seen with a spirostanol minutoside B [119], as compared to both furostanols (minutosides A [295] and C [296]). Persicosides A [120] and B [121]—compounds isolated from A. ampeloprasum ssp. persicum, showed a statistically significant activity against P. italicum, A. niger and T. harzianum, higher than furostanol and cholestane compounds (Sadeghi et al. 2013). The antifungal activity of isolated compounds against B. cinerea was not significant. Interestingly, all saponins inhibited the growth of P. italicum. Antifungal properties of persicosides A [120] and B [121], ceposides A1/A2 [209, 210], tropeosides A1/A2 [213, 214] and B1/B2 [215, 216] were dose dependent. Ceposides A-C [209, 232, 211] (isolated from A. cepa) showed antifungal activity, dependent on their concentration and the fungal species used: soil-borne pathogens (Fusarium oxysporum sp. lycopersici, Rhizoctonia solani and Sclerotium cepivorum), air-borne pathogens (A. alternata, A. niger, B. cinerea, Mucor sp., and Phomopsis sp.), antagonistic fungi (Trichoderma atroviride and T. harzianum), and a pathogen specific to the Allium genus—S. cepivorum (Lanzotti et al. 2012b). Their activity varied in the following order: ceposide B > ceposide A ~ ceposide C. The authors observed a significant synergism of action between those three saponins against B. cinerea and T. atroviride. Ceposide B [232] showed significant activity against all fungi with the exception of F. oxysporum sp. lycopersici, S. cepivorum and R. solani. Ceposides A [209] and C [211] were active against all fungi with the exception of A. niger, S. cepivorum and F. oxysporum sp. lycopersici. Agigenin 3-O-trisaccharide [90] and gitogenin 3-O-tetrasaccharide [73], isolated from the bulbs of A. sativum var. Voghiera, were more active against B. cinerea and T. harzianum than furostanol voghierosides isolated from that plant (Lanzotti et al. 2012a). All the compounds were effective towards T. harzianum in a dose dependent manner, but only spirostanol saponins and voghieroside C [323, 324]—against B. cinerea.

Mskhiladze et al. (2008a) in their studies on anti-yeast effects of saponins from A. leucanthum observed that β-chlorogenin as aglycone and the branched oligosaccharide chain substituted by xylose rather than glucose are beneficial for the activity. Yayoisaponin C [95], eruboside B [79], aginoside [93], agigenin 3-O-β-d-Glc-(1 → 2)-O-β-d-Glc-(1 → 4)-O-β-d-Gal [91] and β-chlorogenin 3-O-β-d-Glc-(1 → 2)-[β-d-Xyl-(1 → 3)]-O-β-d-Glc-(1 → 4)-O-β-d-Gal [80] exhibited antifungal activity on several Candida strains, including C. albicans, C. tropicalis, C. parapsilosis, C. glabrata, C. kefyr, C. krusei, C. lusitaniae, and also on Cryptococcus neoformans, however β-chlorogenin glycoside was the most active compound with MFC from ≤6.25 to 25 μg/mL (as compared to amphotericin B 0.78–12.5 μg/mL). In another study the same compound isolated from A. porrum showed antifungal activity towards Fusarium culmorum (ED50 = 30 μg/mL) (Carotenuto et al. 1999). Eruboside B [79] (A. sativum), β-chlorogenin glycoside as well, inhibited in vitro the growth of C. albicans (MIC 25 μg/mL) (Matsuura et al. 1988).

Agigenin glycosides: aginoside [93] together with yayoisaponins A [96] and C [95] isolated from A. ampeloprasum showed antifungal activity against Mortierella ramanniana at 10 μg/disc (Sata et al. 1998). None of the saponins was active against Penicillium chrysogenum at concentrations up to 100 μg/disc. Ampeloside Bs1 [90], agigenin 3-O-β-d-Glc-(1 → 4)-O-β-d-Gal [87], and furostane-type ampeloside Bf1 [202] isolated from the same species did not inhibit the growth of Aspergillus niger; spirostanols showed weak activity against Candida albicans (Morita et al. 1988). Aginoside [93] at 400 ppm completely inhibited the growth of C. gloeosporioides, Fusarium verticillioides, and Botrytis squamosa and partially suppressed F. oxysporum f. sp. cepae and F. oxysporum f. sp. radicis-lycopersici (Mostafa et al. 2013). The influence of the structure of the sugar chain on the observed anti-fungal activity of compounds bearing the same aglycone was revealed in studies by Teshima et al. (2013).

Alliospirosides A [169] and B [170] (both (25S)-ruscogenin glycosides), which are present mainly in the basal plates and roots of A. cepa Aggregatum group, to a different extent inhibited in vitro a wide range of plant pathogenic fungi: Alternaria ssp., Botrytis ssp., Colletotrichum spp., Curvularia lunata, Epicoccum nigrum, Fusarium ssp., Magnaporthe oryzae, S. cepivorum, and Thanatephorus cucumeris (Teshima et al. 2013). Alliospiroside A [169] strongly inhibited (>80 % growth inhib.) the growth of Colletotrichum spp. isolates. It was also more effective against M. oryzae and S. cepivorum compared to alliospiroside B [170], however, its antifungal activity against B. cinerea, F. oxysporum and F. solani was relatively low.

Enzyme inhibitory properties

Saponin fraction isolated from the methanol extract of A. chinense inhibited cAMP PDE (43.5 %) and Na+/K+ATP-ase (59.3 %) at the concentration of 100 μg/mL (Kuroda et al. 1995). Both enzymes were also inhibited by (25R,S)-5α-spirostane-3β-ol tetrasaccharide [65, 110] (IC50 7.0 × 10−5 and 4.0 × 10−5 M respectively). Laxogenin glycosides exhibited significant activity only on cAMP phosphodiesterase, one of which, with an acetyl group in the saccharide moiety, was almost as potent as papaverine used as a positive control (IC50 3.3 × 10−5 and 3.0 × 10−5 M respectively).

Also, saponins isolated from A. giganteum bulbs inhibited cAMP phosphodiesterase (Mimaki et al. 1994) and in concordance with previously cited results, an acetyl derivative—3-O-acetyl-(24S,25S)-5α-spirostane-2α,3β,5α,6β,24-pentaol 2-O-β-d-Glc [193] exhibited inhibitory activity almost equal to that of papaverine (IC50 4.1 × 10−5 and 3.0 × 10−5 M respectively). In the same study, furostanol saponins were revealed to be much more potent than the corresponding spirostanol glycosides. The results were in contrast to the previous studies of these authors which showed that furostanol glycosides were less active, exhibiting only weak inhibitory activity or none. The authors concluded that the anti-enzyme activity could be dependent on the number of hydroxyls in the A and B rings as in the present study the tested furostanol saponins contained several OH groups.

Saponins isolated from the fruits of A. karataviense and A. cepa as well as the products of chemical modifications of karatavioside A, were studied on a highly purified porcine kidney Na+/K+ATP-ase, in the concentration range from 1 × 10−4 to 1 × 10−7 M (Mirsalikhova et al. 1993). All the compounds affected the enzyme activity being capable of its inhibition, and/or activation. As was showed, the presence of a hydroxyl group in the F-ring at C-24 led to a decrease in the percentage inhibition of Na+/K+ATP-ase. At the concentration of 1 × 10−4 M the inhibitory effect of karatavioside A [151] was 19.8 %, karatavioside B [152]32.4 %, karatavioside C [268]4.9 %, karatavioside E [180]1.7 %, karatavioside F [181]7.5 %, alliospiroside A [169]99.7 %, alliospiroside B [170]76.3 %, alliospiroside D [179]67.1 %; while alliospiroside C [178] activated the enzyme by 13.4 %. A keto group at C-6 of sapogenin slightly increased the inhibition level of Na+/K+ATP-ase.

Moreover, it was revealed that alliospirosides A [169] and B [170] were both uncompetitive enzyme inhibitors, while alliospiroside D [179]competitive. Interestingly, alliospiroside C [178], although bearing the same aglycone as alliospiroside Dcepagenin [44], did not inhibit Na+/K+ATP-ase at all.

Drugs acting via inhibition of the activity of this transport enzyme may be of potential use in the treatment of many diseases of the cardiovascular system, the kidneys, the immune system, which are connected with disturbances in the active transport of ions.

Cardioprotective activity

Three saponins from A. chinense and their aglycones were tested for the protective effects against oxidative stress-induced cardiac damage (Ren et al. 2010). Their activities were evaluated on H2O2-injured cardiac H9C2 cells. The cytotoxicity was measured using MTT assay while the oxidative damage by determination of MDA and NO contents. All tested compounds protected cultured H9C2 cells from death in the concentration range of 5–20 μΜ. It was shown that glycosides exhibited less protective efficacy than sapogenins. Among these, laxogenin [6] and tigogenin [1] displayed stronger effects than furostane-type aglycones. The authors concluded that the presence of F ring in spirostanols may enhance their protective activity whereas oxidation in the B ring might be detrimental as laxogenin was less active than tigogenin.

Nine furostane saponins isolated by Lai et al. from A. fistulosum were tested for antihypoxic activity against hypoxia/reoxygenation (H/R)-induced human umbilical vein endothelial cell (HUVEC) injury (Lai et al. 2010). Cell viability was determined by MTT assay. It was observed that the saponin treatment significantly improved the survival of H/R-treated HUVEC (P < 0.05) in a dose-dependent manner. Fistulosaponin A [250] was the most effective compound with a cell viability of 59.5 ± 3.0, 76.3 ± 3.3, 80.1 ± 3.6, 82.7 ± 4.1, 86.3 ± 4.6, and 78.2 ± 2.8 % for the six dose groups (0.5, 1, 5, 10, 50, and 100 μM), respectively.

In animal studies, alloside B [334], isolated from fruits of A. suvorovii and A. stipitatum, exhibited a statistically reliable hypotriglyceridemic activity in experimental hyperlipidemia caused by 1-day starvation, Triton WR-1339 and vitamin D2–cholesterol, when compared with lipanthyl (Aizikov et al. 1995).

The hypocholesterolemic activity of saponins was reported in many animal studies.

The cholesterol-lowering effect of garlic is probably partially due to the steroid saponin presence. In a rat model of experimental hyperlipidemia induced by feeding a 0.5 % cholesterol-enriched diet saponin-rich fraction from raw garlic administrated at 10 mg/kg/day led to a decrease of plasma total and LDL cholesterol concentration level without affecting HDL cholesterol levels after 16 weeks (Matsuura 2001). It was claimed that the reduction of concentration of plasma cholesterol concentration is the result of inhibition of cholesterol absorption by saponins in the intestine or a direct effect on cholesterol metabolism.

Antispasmodic effect

Furostanol saponins hirtifoliosides C1/C2 [264, 265] and a spirostanol glycoside agapanthagenin 3-O-Glc [85] isolated from A. hirtifolium, along with four saponins elburzensosides A1/A2 [238, 239] and C1/C2 [242, 243] and the sapogenin agapanthagenin [31], from A. elburzense, were subjected to biological assays on the guinea-pig isolated ileum in order to evaluate their possible antispasmodic activity (Barile et al. 2005). Apart from the agapanthagenin glycoside, all the tested compounds were able to reduce induced contractions, as measured by the reduction of histamine release, in a concentration-dependent manner. Elburzensosides C1/C2 [242, 243] and agapanthagenin [31] showed the highest activity with a maximum effect at 10−5M (approx. 50 % inhibition).

The authors concluded that the positive effect is associated with the presence of a hydroxyl group at position C-5 and of a glucose unit at position C-26. On the other hand, hydroxylation at C-6 and glucose attachment at C-3 seem to be structural features responsible for the loss of activity. Furostane-type saponins that were isolated from A. cepa var. tropea, namely tropeosides A1/A2 [213, 214] and B1/B2 [215, 216] were able to dose-dependently relieve acetylocholine- and histamine-induced contractions (50 % inhibition of contractions was seen at the concentration of 10−5 M) (Corea et al. 2005). Interestingly, other furostanols identified in this plant, such as ascalonicosides A1/A2 [217, 218], were inactive.

Other activities

Macrostemonoside A [65] inhibited ADP-induced rabbit ertythrocyte aggregation with IC50 = 0.065 mM (Peng et al. 1992). An in vitro inhibitory activity of ADP-induced platelet aggregation was also reported for macrostemonosides E [272], F [273] and G [274] (IC50 = 0.417; 0.020; 0.871 mM, respectively) (Peng et al. 1993, 1995). 26-O-β-d-Glc (25R)-5α-furostane-3β,12β,22,26-tetrol 3-O-β-d-Glc-(1 → 2)-[β-d-Glc-(1 → 3)]-O-β-d-Glc-(1 → 4)-O-β-d-Gal [289] and 26-O-β-d-Glc (25R)-5α-furostane-3β,12α,22,26-tetrol 3-O-β-d-Glc-(1 → 2)-[β-d-Glc-(1 → 3)]-O-β-d-Glc-(1 → 4)-O-β-d-Gal [290] exhibited significant inhibitory activity on CD40L expression on the membrane of ADP stimulated platelets (Chen et al. 2010).

β-chlorogenin 3-O-β-d-Glc-(1 → 2)-[β-d-Glc-(1 → 3)]-O-β-d-Gal 6-O-β-d-Glc [78], isolated from the bulbs of A. ampeloprasum var. porrum, demonstrated in vivo antiinflammatory and gastroprotective effects in a carrageenan-induced oedema assay and by measuring acute gastric lesions induced by acidified ethanol (Adão et al. 2011a). Saponin administrated orally (100 mg/kg) inhibited oedema formation similar to dexamethasone (25 mg/kg). Cytoprotective activity of β-chlorogenin glycoside resulted in a significant reduction in gastric hyperemia and also in the severity and number of lesions.

Macrostemonoside A [65] increased the synthesis and release of visfatin in 3T3-L1 adipocytes and elevated mRNA levels in this cytokine in a dose- and time-dependent mode (Zhou et al. 2007). In a study on C57BL/6 mice fed on a high-fat diet, this saponin when administered at the dose of 4 mg/kg/day for 30 days moderately inhibited glucose level, glycogen hepatic content, total plasma cholesterol level and abdominal adipose tissue (Xie et al. 2008).

In the molluscicidal bioassay with Biomphalaria pfeifferei diosgenin 3-O-β-d-Glc-(1 → 4)-[β-d-Glc-(1 → 6)]-O-β-d-Glc-(1 → 4)-O-α-l-Rha-(1 → 4)-[α-l-Rha-(1 → 2)]-O-β-d-Glc [146], isolated from A. vineale, exhibited 100 % effect at 25 ppm in <24 h (Chen and Snyder 1989). The authors observed that the molluscicidal activity of isolated compounds increased with an increasing number of monosaccharides in a sugar moiety.

Aginoside [93] was found to be toxic to leek-moth larvae Acrolepiopsis assectella (Harmatha et al. 1987). The compound caused mortality and ecdysial failures 56 ± 10 and 19 % respectively in larvae of A. assectella reared on semisynthetic diet at a concentration of 0.9 mg/g of diet.

Conclusions

In this paper steroidal saponins reported in various Allium species from early 1970 to March 2014 are reviewed, including their skeletal structures and sugar chains.

Until now, as many as 290 saponins have been identified, including a certain number of methoxyl derivatives originating from furostanol compounds, that should be considered as artifacts resulting from the use of methanol in the extraction/isolation procedures.

Allium genus is characterized by a great diversity of structures. Apart from spirostane- and furostane-type compounds, a rare group of open-chain saponins has been identified in several species. Allium genus is also a source of unique steroidal sapogenins, such as 25(S)-5β-spirostane-1β,3β-diol [8] and 2,3-seco-porrigenin [64]. Despite a relatively low content of steroidal glycosides in Allium species, they are considered to contribute, in addition to sulfur compounds, to the overall biological activity of these plants. Undoubtedly, stability of saponins is their advantage as compared to fairly unstable sulfur compounds, thus, they in fact may be predominant active constituents of Allium products. Bearing this aspect in mind it seems highly feasible to develop antifungal Allium preparations against animal and plant pathogens. Also, reports on high in vitro cytotoxic activity of steroidal saponins from Allium species makes them potential candidates for further development as anti-cancer agents.