Ciliate protozoa

A number of studies have demonstrated that pure saponins and saponin-containing plants or extracts have inhibitory effects on protozoa (Table 1)⁽Reference Klita, Mathison and Fenton^17–Reference Hristov, McAllister and Van Herk⁴³⁾. One of the earliest observations was made by Valdez et al. ⁽Reference Valdez, Bush and Goetsch²⁹⁾ who showed that a commercial sarsaponin (steroidal spirostanol saponin) extracted from the Yucca schidigera plant decreased the protozoal population with a small increase in bacterial numbers in continuous culture systems. In many subsequent studies, the extracts of Y. schidigera have been shown to suppress the growth of rumen protozoa in vitro ⁽Reference Wallace, Arthaud and Newbold⁴⁴⁾ without affecting DM disappearance and other rumen microbial populations⁽Reference Wang, MaAllister and Newbold⁴⁵⁾. The saponin-containing butanol extract of Y. schidigera has been shown to be responsible for all the plant's anti-protozoal activity⁽Reference Wallace, Arthaud and Newbold^44, Reference Makkar, Sen and Blümmel⁴⁶⁾. Protozoal activity, as measured by the breakdown of [¹⁴C]leucine-labelled Selenomonas ruminantium in rumen fluid incubated in vitro, was stopped by the addition of 1 % Y. schidigera extract⁽Reference Wallace, Arthaud and Newbold⁴⁴⁾. A number of in vivo studies have also confirmed the anti-protozoal effects of Yucca saponins⁽Reference Hristov, McAllister and Van Herk^43, Reference Hristov, Ivan and Neill⁴⁷⁾. Probably all types of saponins inhibit protozoal activity depending upon the concentrations in the diet. Lucerne saponins (triterpene saponins), isolated by ethanol extraction and partial acid hydrolysis, caused a decrease in protozoal numbers in the rumen of sheep by 34 and 66 % at high concentrations of 2 and 4 % of dietary DM, respectively⁽Reference Lu and Jorgensen¹⁸⁾. A methanol extract of Sapindus rarak fruit also depressed the protozoal population in the rumen of sheep by 57 %, which was attributed to the presence of saponins in the extract⁽Reference Thalib, Widiawati and Hamid⁴⁸⁾. Saponin-like material in the organic phase of a butanol–water mixture of Sesbania sesban foliage (containing oleanane-type triterpene saponins), a leguminous tree, was found to be inhibitory to protozoal activity⁽Reference Newbold, Hassan and Wang⁴⁹⁾. Similarly, the saponins from Q. saponaria and Acacia auriculiformis ⁽Reference Makkar, Sen and Blümmel⁴⁶⁾, theasaponins⁽Reference Hu, Liu and Ye^50, Reference Guo, Liu and Lu⁵¹⁾ and Sapindus saponaria ⁽Reference Kamra, Singh and Agarwal⁵²⁾ have also shown an anti-protozoal activity in vitro. Monforte-Briceno et al. ⁽Reference Monforte-Briceno, Sandoval-Castro and Ramirez-Aviles⁵³⁾ also reported defaunating effects of saponins present in Enterolobium cyclocarpum and Acacia angustissima. However, there is some evidence that saponins do not influence protozoal numbers in vitro (44 to 176 mg Yucca saponins and 100 to 400 mg Quillaja saponins per litre medium⁽Reference Hristov, Ivan and Neill⁴⁷⁾; 1 to 100 mg/kg DM substrate⁽Reference Sliwinski, Kreuzer and Wettstein³²⁾), and in sheep (1180 to 1477 mg Yucca or Quillaja saponins/kg DM intake)⁽Reference Pen, Takura and Yamaguchia²⁴⁾, dairy cows (60 g Yucca extract (10 % saponins) per d with feed intake of 21·8 kg, i.e. 275 mg Yucca saponins/kg DM intake)⁽Reference Benchaar, McAllister and Chouinard⁴²⁾ and heifers (8 g Quillaja extract (10 % saponins) per kg DM intake, i.e. 800 mg Quillaja saponins/kg DM intake)⁽Reference Baah, Ivan and Hristov²³⁾. The inefficacy of the saponins to cause protozoal inhibition in these studies is not clearly known. The lack of anti-protozoal effect of the saponins might involve the use of low doses of saponins in some studies, and an adaptation of rumen microbes to saponins in other studies.

Table 1

Effects of saponins or saponin-containing plants on protozoal numbers and rumen fermentation in vivo and performance of animals

F:C, forage:concentrate ratio in diet; Prot, protozoal numbers; NH₃, ammonia concentration; TVFA, total volatile fatty acid concentration; A:P, acetate:propionate ratio; Dig, digestibility of fibre components; MPS, efficiency of microbial protein synthesis; Met, methane; FI, DM intake; ADG, average daily gain; MY, milk yield; BW, body weight; ↓ , decrease; × , not reported; = , no effect; ↑ , increase; QS, Quillaja saponaria; YS, Yucca schidigera.

* Values in parentheses are the saponin content (%) in extracts or plants.

† Dose is given as g/kg diet unless otherwise stated, and was calculated as g/kg from feed intake and inclusion of saponin products whenever possible.

‡ Effect is relative to control.

§ DM digestibility.

∥ Wool growth also.

¶ ADG was greater at 0 to 28 d of feeding, after which there was no effect.

The methods of extraction of saponins from plant materials may show different anti-protozoal activity. For example, Patra et al. ⁽Reference Patra, Kamra and Agarwal¹⁰⁾ noted that a methanol extract of pods of Acacia concinna exhibited stronger activity against protozoa than water and ethanol extracts. Similarly, Agarwal et al. ⁽Reference Agarwal, Kamra and Chaudhary⁵⁴⁾ noted that methanol, ethanol and water extracts of berries of Sapindus mukorossi inhibited protozoal numbers in an in vitro gas production system and methanol extract was more active that water and ethanol extracts. The dose-dependent effect of the saponins on protozoa has clearly been observed in many studies⁽Reference Patra, Kamra and Agarwal^10, Reference Santoso, Kilmaskossub and Sambodo^19, Reference Agarwal, Kamra and Chaudhary^54, Reference Pen, Sar, Mwenya and Kuwaki⁵⁵⁾. The oral administration of saponins from the plant Biophytum petersianum decreased protozoal numbers in goats by 35 and 40 % at the dose of 13 and 19·5 mg/kg body weight, respectively, beyond which protozoal counts did not further reduce⁽Reference Santoso, Kilmaskossub and Sambodo¹⁹⁾. Similarly, Pen et al. ⁽Reference Pen, Sar, Mwenya and Kuwaki⁵⁵⁾ reported that commercial extracts of Y. schidigera (9 % saponins; monodesmosidal steroid) and Q. saponaria (6 % saponins; bidesmosidal triterpene) decreased protozoal numbers linearly with increasing doses of both the extracts (0, 2, 4 and 6 ml/l of in vitro medium). From the studies of Santoso et al. ⁽Reference Santoso, Kilmaskossub and Sambodo¹⁹⁾ and Pen et al. ⁽Reference Pen, Sar, Mwenya and Kuwaki⁵⁵⁾, it could be noted that the inhibition of protozoal counts plateaued at a certain dose specific for saponin type (for example, decreased protozoal numbers of 53 % for Yucca, 29 % for Quillaja and 40 % for Biophytum petersianum), beyond which there was no appreciable inhibition of protozoa. Therefore, it appears that all of the protozoal species are not equally susceptible to the toxic effect of saponins, indicating that the range of activity of saponins might be dependent upon the type of saponins and protozoal species. Because the concentrations of saponins in Yucca (9 % saponins) and Quillaja (6 % saponins) extracts were different in the study of Pen et al. ⁽Reference Pen, Sar, Mwenya and Kuwaki⁵⁵⁾, after adjusting the concentration of saponins (i.e. 4 ml Yucca equivalent to 6 ml Quillaja extract) in the medium, it was noted that Yucca extract (53 % reduction of total protozoa compared with control) strongly inhibited the protozoal growth compared with Quillaja extract (29 % reduction of total protozoa compared with control). It could be tested whether a mixture of the saponins of different types might show a synergistic anti-protozoal activity. It has been noted that the degree of toxicity of saponins to rumen protozoa greatly differs in different accessions of the same plant, for example, Sesbania sesban ⁽Reference Teferedegne, McIntosh and Osuji⁵⁶⁾. This would suggest the presence of different types and/or concentrations of saponins in different accessions of plants. Additionally, the effects of saponins on ruminal protozoa could be diet-dependent. For example, Hess et al. ⁽Reference Hess, Kreuzer and Diaz⁵⁷⁾ reported that protozoa numbers were decreased by saponin-containing Sapindus saponaria fruits only when added to a diet supplemented with a high-quality legume (Arachis pintoi), but not with Cratylia argentea, a medium-quality legume. Conversely, anti-protozoal effects of lucerne saponins did not show a diet × saponin interaction, although inhibition was greater for concentrate- (75 %) than roughage- (47 %) based diets at 4 % saponin inclusion⁽Reference Lu and Jorgensen¹⁸⁾. The composition of the diet has a strong influence on the protozoal community⁽Reference Williams, Coleman, Hobson and Stewart⁵⁸⁾. Therefore, a diet-dependent effect might arise from the selectivity of saponins to protozoal species. However, no information is available on the effect of saponins on the species of rumen protozoa. Ivan et al. ⁽Reference Ivan, Koenig and Teferedegne²²⁾ reported that saponins of Enterolobium cyclocarpum reduced the protozoal numbers without changing the protozoal generic composition in sheep. Benchaar et al. ⁽Reference Benchaar, McAllister and Chouinard⁴²⁾ observed no effect of saponins on total protozoal counts as well as generic composition.

Bacteria and fungi

Relatively few studies have addressed the effects of saponins on pure cultures of rumen micro-organisms. Wallace et al. ⁽Reference Wallace, Arthaud and Newbold⁴⁴⁾ showed that Y. schidigera extract (1 %, v/v) inhibited the growth of pure cultures of Butyrivibrio fibrisolvens and Streptococcus bovis, while the growth of Prevotella ruminicola was stimulated and the growth of Selenomonas ruminantium was unaffected by Yucca extract. In another experiment, Yucca saponins inhibited the growth of Streptococcus bovis, Prevotella bryantii and Ruminobacter amylophilus by causing the alteration of the cell wall of the bacteria, while enhancing the growth of Selenomonas ruminantium ⁽Reference Wang, McAllister and Yanke⁵⁹⁾. Among the primary cellulolytic bacteria, the activity of Fibrobacter succinogenes was unaffected, but the cellulose-digesting ability of Ruminococcus albus and Ruminococcus flavefaciens was markedly reduced by Yucca saponins⁽Reference Wang, McAllister and Yanke⁵⁹⁾. These studies clearly suggest species-dependent effects of saponins on rumen bacteria, which might offer a selective manipulation of rumen metabolism. For example, saponins included in high concentrate-based diets may reduce the incidence of acidosis by inhibiting the growth of Streptococcus bovis, the numbers of which normally rapidly increase with concentrate-based diets, fermenting the soluble starch to lactate⁽Reference Russell and Rychlik⁶⁰⁾. The reason for the growth-promoting effect of saponins on some pure cultures of bacteria is unclear. However, it is possible that low doses of saponins might increase the permeability of the cell membrane in a controlled manner, thus allowing the increased absorption of nutrients into the bacterial cells⁽Reference Sen, Makkar and Muetzel⁶¹⁾. In studies with mixed rumen microbial population, saponin-containing plants increased bacterial numbers, presumably as a consequence of inhibition of protozoal numbers⁽Reference Hess, Beuret and Lotscher^27–Reference Valdez, Bush and Goetsch^29, Reference Thalib, Widiawati and Hamid^48, Reference Newbold, Hassan and Wang⁴⁹⁾. The study of Hess et al. ⁽Reference Hess, Beuret and Lotscher²⁷⁾ showed that supplementation with fruits of Sapindus saponaria, which supplied crude saponins at 0·6 g/kg body weight^0·75, increased total bacterial numbers, and decreased total ciliate protozoa count in lambs fed basal diets of grass hay or grass and legume mixture (1:2 or 2:1) along with concentrate (1:3). Cellulolytic bacterial counts have been shown to increase due to the feeding of Enterolobium cyclocarpum containing saponins⁽Reference Diaz, Avendan and Escobar²⁸⁾ and an extract of Sapindus rarak ⁽Reference Thalib, Widiawati and Hamid⁴⁸⁾. In a species-level study of cellulolytic bacteria, Muetzel et al. ⁽Reference Muetzel, Hoffmann and Becker⁶²⁾ found that the addition of saponin-containing Sesbania pachycarpa leaves did not affect the growth of F. succinogenes and Ruminococcus flavefaciens, but inhibited the growth of Ruminococcus albus in vitro. Another type of saponin, theasaponin (an oleanene-type triterpene) markedly (79 %) decreased the population density of rumen fungi; however, the numbers of F. succinogenes increased (41 %) and Ruminococcus flavefaciens were not affected by theasaponins at a concentration of 400 mg/l culture medium as quantified by real-time PCR⁽Reference Guo, Liu and Lu⁵¹⁾.

Ruminal fungi, Neocallimastix frontalis and Piromyces rhizinflata, were also sensitive to Y. schidigera saponins⁽Reference Wang, McAllister and Yanke⁵⁹⁾. However, fungal numbers increased when sheep were fed with diets containing Sapindus saponaria ⁽Reference Diaz, Avendan and Escobar²⁸⁾ or Enterolobium cyclocarpum ⁽Reference Navas-Camacho, Laredo and Cuesta²¹⁾. Wina et al. ⁽Reference Wina, Muetzel and Becker⁶³⁾ noted that at low concentrations of Sapindus rarak saponins had a positive effect on Chytridiomycetes, a rumen fungus, during the long-term feeding, while no effect was observed with higher concentrations in sheep. These studies illustrate that a positive effect of saponins on rumen fungi might be obtained perhaps at low doses. Wina et al. ⁽Reference Wina, Muetzel and Hoffmann⁶⁴⁾ investigated the effect of saponins from a methanol extract of Sapindus rarak at concentrations of 0, 0·25, 0·5, 1·0, 2·0 and 4·0 mg/ml to an elephant grass and wheat bran (7:3, w/w) diet in vitro. The microbial biomass gradually increased with increasing doses of extract from 102·8 mg/g substrate incubated (control) to 147·3 mg/g substrate incubated at 4 mg/ml and the total protozoal counts decreased. Studies with membrane hybridisation (a molecular method to quantify rRNA of micro-organisms) also showed no eukaryotic RNA in the presence of the extract at concentrations greater than 1 mg/ml. The bacterial RNA concentration increased only at 1 mg/ml compared with the control. There was no effect on Fibrobacter spp.; however, there was a negative effect on Ruminococcus albus and Ruminococcus flavefaciens at 2 and 4 mg/ml of extract. The growth of Chytridiomycetes was reduced at the highest concentration. The in vivo trials with sheep and goats fed a saponin extract of Sapindus rarak up to 0·72 and 0·60 g/kg body weight, respectively, also confirmed that Fibrobacter spp. were not affected by the Sapindus rarak extract, while Ruminococcus albus, Ruminococcus flavefaciens and Chytridiomycetes showed a short-term negative effect to the application of saponins, but were not affected when this extract was fed for a longer period (105 d)⁽Reference Wina, Muetzel and Becker⁶³⁾. From the studies of Wang et al. ⁽Reference Wang, McAllister and Yanke⁵⁹⁾, Muetzel et al. ⁽Reference Muetzel, Hoffmann and Becker⁶²⁾, Wina et al. ⁽Reference Wina, Muetzel and Becker^63, Reference Wina, Muetzel and Hoffmann⁶⁴⁾ and Guo et al. ⁽Reference Guo, Liu and Lu⁵¹⁾, it is clear that F. succinogenes is relatively resistant to saponins compared with other cellulolytic bacterial species. The presence of 2-aminoethylphosphonic acid in its cell wall may enhance the membrane stability of Fibrobacter spp.⁽Reference Vinogradov, Egbosimba and Perry⁶⁵⁾, which might explain the insensitivity of saponins to this bacterium. 2-Aminoethylphosphonic acid covalently linked to membrane polymers renders the membrane resistant to enzymic hydrolysis, which may promote the longevity of the organisms⁽Reference Vinogradov, Egbosimba and Perry^65, Reference Sarkar, Hamilton and Fairlamb⁶⁶⁾.

Rumen archaea

There has been little experimentation demonstrating the direct effect of saponins on rumen archaea, but the direct effect cannot be ruled out. The studies on pure cultures of rumen methanogens are scant. The reduction of methane production by Sapindus saponaria was observed in a defaunated state without any change in methanogen counts, indicating that saponins may directly influence the activity of methanogens⁽Reference Hess, Monsalve and Lascano⁶⁷⁾. Guo et al. ⁽Reference Guo, Liu and Lu⁵¹⁾ noted that tea saponins (400 mg/l) had no effect on the growth and expression of the methyl coenzyme M reductase subunit A gene (mcrA) of Methanobrevibacter ruminantium. In mixed ruminal cultures, tea saponins also did not affect total archaeal numbers, but the activity of the mcrA gene decreased markedly (76 %) and methane production lowered by 8 %⁽Reference Guo, Liu and Lu⁵¹⁾. Saponins from Sapindus rarak decreased the methanogen RNA concentration at a dose of 4 mg/ml of in vitro medium, while lower concentrations ( < 4 mg/ml) had no effect on methanogens⁽Reference Wina, Muetzel and Hoffmann⁶⁴⁾. The study of Wina et al. ⁽Reference Wina, Muetzel and Hoffmann⁶⁴⁾ did not confirm that a decrease in methanogen RNA concentration is due to direct inhibition by saponins, but did suggest it may also result from the inhibition of protozoa activity. However, it has also been reported that saponins of Sapindus saponaria fruits increased the methanogen population in sheep although methane production reduced⁽Reference Hess, Beuret and Lotscher²⁷⁾. The reason for the increase in the methanogen population was not clear in this study.

Explaining the effects of saponins on micro-organisms

A schematic illustration of effects of saponins on rumen microbes and fermentation is presented in Fig. 2. The modifications of rumen fermentation by saponins are mediated via a direct influence on the composition and the activity of rumen protozoa, bacteria, fungi and archaea. The general mode of action of saponins on micro-organisms is their interaction with the sterol moiety⁽Reference Hostettmann and Marston¹⁵⁾, which is present in the membrane of protozoa⁽Reference Williams, Coleman, Hobson and Stewart⁵⁸⁾. Thus, an anti-protozoal activity of saponins may be due to the destruction of protozoal cell membranes, causing leaking of cell contents. It has been reported that the main sterols present in entodiniomorphs are stigmastanol, campestanol and cholestanol, whereas the major sterol in holotrichs is cholestanol⁽Reference Williams⁶⁸⁾. The different composition of sterols in rumen protozoa may relate to the different sensitivity of protozoa to saponins. Several factors such as saponin structure, dose level, adaptation process, diet composition and microbial community structure influence the effect of saponins on the rumen microbial population (Fig. 3). The structures of saponins and their varying proportions from different sources of the same plant material may elicit different responses on rumen metabolism. Also the method of saponin estimation in the extract or plant may differ⁽Reference Cheeke¹²⁾, which may explain the differences in response of saponins despite similar dose levels⁽Reference Holtshausen, Chaves and Beauchemin²⁵⁾. For example, Hostettmann & Marston⁽Reference Hostettmann and Marston¹⁵⁾ discussed that spectrophotometric methods are very sensitive, but not suitable for estimation of saponins in crude plant extracts, since colour reactions are not specific, and coloured products may form with phytosterols and flavonoids extracted with saponins. Klita et al. ⁽Reference Klita, Mathison and Fenton¹⁷⁾ mentioned that concentrations of saponins determined by HPLC procedures are commonly lower than those determined by biological procedures (for example, haemolytic micro-method). Additionally, saponins may not be completely extracted from vegetable materials⁽Reference Hostettmann and Marston¹⁵⁾.

Fig. 2

A schematic presentation of the proposed effects of saponins on rumen microbes and fermentation. Primary effects modify the composition of rumen microbes and secondary effects modify the rumen fermentation. +, Increase; − , decrease; OM, organic matter; VFA, volatile fatty acids; P, propionate.

Fig. 3

A schematic presentation of factors affecting the effects of saponins on rumen microbes.

The responses of rumen bacteria and fungi to saponins may be mediated through defaunation and/or direct effects. An increase in the number of bacteria and fungi due to the addition of saponins has been observed in some studies, which is believed to be as a consequence of reduced engulfment of these micro-organisms by protozoa⁽Reference Williams, Coleman, Hobson and Stewart⁵⁸⁾. The microbial growth in the rumen is also affected by the presence of saponins, which may decrease the bacterial and fungal numbers. Therefore, the observed effect of saponins on rumen microbes is the balance between anti-protozoal effects (hence increased microbial growth) and direct antibacterial effects of saponins. The antibacterial effects of saponins may be mediated by disrupting the cell membranes, rather than simply altering the surface tension of the extracellular medium⁽Reference Wina, Muetzel and Becker^5, Reference Sen, Makkar and Muetzel⁶¹⁾. Both Gram-positive and Gram-negative bacteria are sensitive⁽Reference Sung and Lee⁶⁹⁾, although their sensitivity depends upon the structure of the aglycone moiety of saponins⁽Reference Avato, Bucci and Tava⁷⁰⁾ and the fatty acid composition of cell membranes of target bacteria. For instance, Wallace et al. ⁽Reference Wallace, Arthaud and Newbold⁴⁴⁾ concluded from their study that Gram-positive bacteria were more susceptible compared with Gram-negative bacteria in the presence of Yucca saponins. Later, Wang et al. ⁽Reference Wang, McAllister and Yanke⁵⁹⁾ reported that Yucca saponins inhibited pure cultures of rumen cellulolytic bacteria and fungi, but their effects on amylolytic bacteria varied among species. Similarly, Gram-negative bacteria were negatively affected by saponins extracted from the leaves of Gymnema sylvestre and Eclipta prostrata ⁽Reference Khanna and Kannabiran⁷¹⁾, whereas the effect of medicagenic acid, a lucerne saponin, was more pronounced against Gram-positive bacteria⁽Reference Avato, Bucci and Tava⁷⁰⁾. Although the addition of sugar units decreases the antimicrobial properties⁽Reference Avato, Bucci and Tava⁷⁰⁾, the sugar moiety may not be important for activity in the rumen, as the side chain of sugars might be rapidly degraded by the mixed rumen microbial cultures⁽Reference Makkar and Becker⁷²⁾. From this perspective, the sapogenin structure–activity relationship of saponins with greater anti-protozoal properties and weak antibacterial (for example, rumen cellulolytic bacterial populations) and fungal activity could be targeted for rumen micro-organisms for improvement of fermentation. Further, Li et al. ⁽Reference Li, Du and Zou⁷³⁾ demonstrated that tea saponins (triterpenoid type) showed better antimicrobial activity at low pH in vitro. This indicates that the effect of saponins may be modulated by the pH in the rumen depending upon the concentrate- and roughage-based diets. The method of preparation and purification might also give rise to different responses to saponins. For example, Sen et al. ⁽Reference Sen, Makkar and Muetzel⁶¹⁾ noted that Quillaja saponins from different commercial sources showed maximum effects on growth of Escherichia coli at different concentrations of saponins in the medium.

The major mechanism suggested for the antifungal activity of saponins is their interaction with sterols⁽Reference Francis, Kerem and Makkar¹¹⁾, followed by pore formation and loss of membrane integrity. Some saponins may show a specific inhibitory activity to fungi but not to bacteria. For example, saponins isolated from Allium minutiflorum exhibited an antifungal activity depending upon the concentration and type of saponins, but did not show an appreciable antibacterial activity in the non-rumen bacteria⁽Reference Barile, Bonanomi and Antignani⁷⁴⁾. From the perspective of structure–activity relationships, saponins with the triterpene or spirostanol moiety generally demonstrate stronger antifungal activities, whereas furostanol saponins with bidesmosidic nature show little to no bacteriostatic and fungicidal effects⁽Reference Oleszek and Naidu⁷⁵⁾. The antifungal activity of these steroidal saponins is also associated with their individual steroidal sapogenins as well as the number and structure of monosaccharide units in their sugar chains. It has been demonstrated that diosgenin saponins with triglycosides were active against Candida albicans and Candida glabrata, while the monoglycoside and diglycoside did not show any activity⁽Reference Yang, Zhang and Jacob⁷⁶⁾.

Nitrogen metabolism

A number of studies have reported that saponins or saponin-containing plants decreased rumen ammonia N concentrations in vitro ⁽Reference Holtshausen, Chaves and Beauchemin^25, Reference Wallace, Arthaud and Newbold^44, Reference Hu, Liu and Ye^50, Reference Wang, McAllister and Yanke^86–Reference Busquet, Calsamiglia and Ferret⁸⁸⁾ and in vivo ⁽Reference Pen, Takura and Yamaguchia^24, Reference Santoso, Mwenya and Sar^39, Reference Hussain and Cheeke^41, Reference Hristov, McAllister and Van Herk^43, Reference Makkar, Sen and Blümmel⁴⁶⁾. The effect of saponins in reducing rumen ammonia concentrations was more pronounced for Yucca extract compared with Quillaja extract⁽Reference Holtshausen, Chaves and Beauchemin^25, Reference Pen, Sar, Mwenya and Kuwaki⁵⁵⁾, probably due to strong anti-protozoal properties of Yucca compared with Quillaja saponins. However, saponins do not always reduce rumen ammonia concentrations. Sapindus saponaria ⁽Reference Hess, Beuret and Lotscher²⁷⁾ and Yucca extract⁽Reference Benchaar, McAllister and Chouinard⁴²⁾ included in concentrate diets did not significantly reduce rumen ammonia concentrations, although it should be noted that the Yucca extract did not affect rumen protozoal numbers either. In studies with dairy cows, inclusion of sarsaponins up to 77 parts per million (ppm) of diet⁽Reference Valdez, Bush and Goetsch^29, Reference Goetsch and Owens³³⁾, 9 g/d⁽Reference Wilson, Overton and Clark³⁶⁾, 8 g/d⁽Reference Wu, Sadik and Sleiman⁴⁰⁾ and 10 g of Yucca whole plant (equivalent to 0·6 g saponins/kg DM) or 10 g Quillaja whole plant (equivalent to 0·3 g saponins/kg DM)⁽Reference Holtshausen, Chaves and Beauchemin²⁵⁾ did not influence ammonia concentrations in the rumen, probably due to low concentrations of saponins inclusion in the diets. Valdez et al. ⁽Reference Valdez, Bush and Goetsch²⁹⁾ and Hess et al. ⁽Reference Hess, Beuret and Lotscher²⁷⁾ have reported that by decreasing protozoal numbers slightly, saponins increased bacterial numbers, which may explain why saponin treatments are not always associated with reduced rumen ammonia concentrations.

Yucca, Quillaja and Acacia saponins have been shown to increase the efficiency of microbial protein synthesis in an in vitro fermentation system containing hay as a substrate⁽Reference Makkar, Sen and Blümmel⁴⁶⁾. Saponins in higher concentrations generally have adverse effects on microbial protein synthesis subject to saponin types. Yucca saponins increased microbial protein synthesis at 15 mg/l medium, but decreased it at higher concentrations⁽Reference Wang, McAllister and Yanke⁸⁶⁾, whereas tea saponins increased microbial protein synthesis up to a concentration of 267 mg/l⁽Reference Hu, Liu and Ye⁵⁰⁾. The greater bacterial protein synthesis, which could be the result of the prevention of predation of bacteria by protozoa, may overwhelm the decreased protozoal growth, and thus overall microbial protein synthesis could increase due to the administration of saponins. Therefore, saponin-specific doses needed for anti-protozoal effects, but with minimum inhibitory effects on bacteria and fungi, need to be established. In an animal feeding trial, Santoso et al. ⁽Reference Santoso, Mwenya and Sar³⁵⁾ reported that daily supplementation of a basal diet consisting of timothy silage and a concentrate (85:15 on a DM basis) with 240 ppm of Yucca saponins reduced rumen ammonia N concentrations in sheep. The microbial N supply and efficiency of microbial N synthesis were greater in sheep fed with Yucca diets v. those fed with control diets. In another experiment, Santoso et al. ⁽Reference Santoso, Kilmaskossub and Sambodo¹⁹⁾ reported that oral administration of saponins (0, 13, 19·5 and 26 mg/kg body weight) from Biophytum petersianum linearly decreased ruminal ammonia N concentrations, urinary N and total N excretion, and increased retained N as a proportion of N digested, efficiency of microbial N synthesis and microbial N supply in goats fed a diet consisting of elephant grass silage and concentrate (70:30). Klita et al. ⁽Reference Klita, Mathison and Fenton¹⁷⁾ reported that total N flow to the duodenum increased in sheep fed a saponin extract of lucerne root up to 4 % of DM intake. However, in studies with cattle receiving Y. schidigera saponins, a decrease in the efficiency of microbial protein synthesis was observed, probably due to the suppression of bacterial and protozoal growth⁽Reference Goetsch and Owens³³⁾. A diet-dependent response to saponins on bacterial protein synthesis has been clearly observed by Lu & Jorgensen⁽Reference Lu and Jorgensen¹⁸⁾, who noted that efficiency of bacterial protein synthesis and bacterial protein flow to the duodenum were decreased by lucerne saponins in a roughage-based diet, but had no effect in a concentrate-based diet.

Digestion of feeds and volatile fatty acid production

Mixed effects of saponins on feed digestion are reported in the literature. Saponins from Sapindus saponaria ⁽Reference Hess, Kreuzer and Diaz⁵⁷⁾, Yucca and Quillaja ⁽Reference Holtshausen, Chaves and Beauchemin²⁵⁾ plants reduced the digestibility of fibre components in vitro. Similarly, lucerne saponins (up to 4 % of DM intake) linearly decreased both ruminal and total tract digestibilities of organic matter, fibre components and digestibility in sheep fed a grass hay diet⁽Reference Klita, Mathison and Fenton¹⁷⁾. The xylanase activity and neutral-detergent fibre digestibility decreased at higher levels of Sapindus rarak extract⁽Reference Wina, Muetzel and Hoffmann⁶⁴⁾. The digestibilities of organic matter and neutral-detergent fibre also reduced in the presence of Sapindus saponaria fruits⁽Reference Hess, Beuret and Lotscher²⁷⁾ and extracts of Sapindus mukorossi ⁽Reference Agarwal, Kamra and Chaudhary⁵⁴⁾. In many other studies⁽Reference Navas-Camacho, Laredo and Cuesta^21, Reference Pen, Takura and Yamaguchia^24, Reference Hess, Beuret and Lotscher^27, Reference Lovett, Stack and Lovell^37, Reference Wang, Wang and Zhou³⁸⁾, saponins have been reported not to influence nutrient digestibilities. A number of studies have reported increases in digestibilities of organic matter and fibre components. Goetsch & Owens⁽Reference Goetsch and Owens³³⁾ reported a significant increase in the digestibility of organic matter in cattle fed 44 ppm of Yucca saponins. Valdez et al. ⁽Reference Valdez, Bush and Goetsch²⁹⁾ recorded an increased digestibility of organic matter, together with a tendency toward an increased digestibility of acid-detergent fibre with increasing application levels of Yucca extract (up to 77 ppm). Pen et al. ⁽Reference Pen, Takura and Yamaguchia²⁴⁾ reported an increase in neutral-detergent fibre digestibility by sheep supplemented with Quillaja saponins. Similarly, the total tract digestibilities of organic matter and fibre were increased by lucerne saponins in a concentrate-based diet for sheep⁽Reference Lu and Jorgensen¹⁸⁾. Saponins might change the site of digestion in the digestive tract. For example, in the latter experiment, the degradability of cellulose in the rumen was depressed by added saponins, but the degradability of fibre fractions in the hindgut was greatly increased.

There are varying reports on the effects of saponins on ruminal volatile fatty acid (VFA) concentrations, as digestibility of feeds varied among the studies. No difference in total VFA concentrations with the supplementation of saponins⁽Reference Klita, Mathison and Fenton^17, Reference Pen, Takura and Yamaguchia^24, Reference Holtshausen, Chaves and Beauchemin^25, Reference Santoso, Mwenya and Sar^39, Reference Hristov, McAllister and Van Herk^43, Reference Guo, Liu and Lu^51, Reference Pen, Sar, Mwenya and Kuwaki⁵⁵⁾ have been reported. However, reductions in VFA concentration due to supplementation of lucerne saponins at 4 % of DM intake⁽Reference Lu and Jorgensen¹⁸⁾ and Yucca extract at levels of 25 and 50 g/d to Holstein cows⁽Reference Lovett, Stack and Lovell³⁷⁾ have also been reported. Saponins in higher concentrations might adversely affect rumen fermentation. Gas and VFA production were greater at a Yucca saponins concentration of 75 mg/l compared with the control, but decreased at higher concentration⁽Reference Wang, McAllister and Yanke⁸⁶⁾. The rate of gas production decreased in the presence of Yucca and Quillaja saponins, but increased in the presence of Acacia saponins when a mixture of hay and concentrate (70:30) was used as a substrate⁽Reference Makkar, Sen and Blümmel⁴⁶⁾. The authors⁽Reference Makkar, Sen and Blümmel⁴⁶⁾ also reported that Yucca, Quillaja and Acacia saponins either decreased the rate of gas production significantly or had a tendency to decrease it depending upon the dose levels when hay was used as a substrate. The potential extent of gas production was either not affected (Quillaja saponins) or decreased (Yucca and Acacia saponins) in this study⁽Reference Makkar, Sen and Blümmel⁴⁶⁾. A number of studies reported that saponins from different plants increased propionate and decreased acetate proportions⁽Reference Holtshausen, Chaves and Beauchemin^25, Reference Santoso, Mwenya and Sar^39, Reference Hussain and Cheeke^41, Reference Hristov, McAllister and Van Herk^43, Reference Guo, Liu and Lu^51, Reference Agarwal, Kamra and Chaudhary^54, Reference Pen, Sar, Mwenya and Kuwaki^55, Reference Lila, Mohammed and Kanda^87–Reference Devant, Anglada and Bach⁹⁰⁾. Patra et al. ⁽Reference Patra, Kamra and Agarwal¹⁰⁾ reported that extracts of Acacia concinna containing saponins, which inhibited protozoa number to the extent of 80 %, did not affect total VFA productions, but increased propionate production, and there was a decrease in the acetate:propionate ratio. The molar proportion of acetate was lower in sheep receiving the Yucca diets compared with the control diet, and molar proportions of butyrate and iso-acids were higher⁽Reference Santoso, Mwenya and Sar³⁵⁾. Similarly, an increased molar proportion of propionate concentration in the rumen by the Biophytum extract that contains saponins has been reported⁽Reference Santoso, Kilmaskossub and Sambodo¹⁹⁾. These characteristics of VFA pattern are often noted in studies with defaunated animals⁽Reference Williams, Coleman, Hobson and Stewart⁵⁸⁾. Conversely, Hussain & Cheeke⁽Reference Hussain and Cheeke⁴¹⁾ found a numerically lower propionate concentration in vivo with 75 ppm of Yucca saponins. However, Valdez et al. ⁽Reference Valdez, Bush and Goetsch²⁹⁾ and Wu et al. ⁽Reference Wu, Sadik and Sleiman⁴⁰⁾ did not observe a significant change in propionate concentration in vivo up to 77 and 125 ppm, respectively, of Yucca saponins. The responses of saponins on rumen VFA production are often diet dependent. The study of Wang et al. ⁽Reference Wang, McAllister and Yanke⁸⁶⁾ clearly demonstrated that gas and total VFA productions from barley grain were increased by Y. schidigera, whereas they were reduced from lucerne hay. Kil et al. ⁽Reference Kil, Cho and Kim⁹¹⁾ found a significant increase in propionate concentration in vitro when supplying 125 ppm of Yucca saponins with a mixed forage–concentrate diet, but reported a marginal effect with a concentrate diet. Additionally, Cardozo et al. ⁽Reference Cardozo, Calsamiglia and Ferret⁹²⁾ showed that Yucca saponins at dose levels up to 30 mg/l increased the molar proportion of propionate and reduced that of acetate at pH of 5·5, whereas no effect was noted at pH 7·0, which indicated that effects of saponins on VFA profiles might be modulated by rumen pH.

Methane production

Methane produced during anaerobic fermentation in the rumen represents a feed energy loss and contributes to the greenhouse effect. Therefore, decreasing methane emission has been an important objective for ensuring the sustainability of ruminant production. Methane is produced normally during fermentation of feeds in the rumen by methanogenic archaea. The removal of protozoa can decrease methane production, as some populations of methanogens remain associated with protozoa as ectosymbionts and endosymbionts⁽Reference Hess, Kreuzer and Diaz^57, Reference Finlay, Esteban and Clarke^93, Reference Tokura, Ushida and Miyazaki⁹⁴⁾. Theasaponin caused an increase in propionate production, while methane and isobutyrate production were decreased⁽Reference Ye, Liu and Shi⁸⁹⁾.

In a study by Hess et al. ⁽Reference Hess, Monsalve and Lascano⁶⁷⁾, Sapindus saponaria in Rusitec reduced methane production by 20 %, without affecting the methanogens relative to control and there was a 54 % reduction in protozoa counts. This was confirmed in vivo – daily methane release was decreased by Sapindus saponaria supplementation in lambs fed basal diets of grass hay or grass and legume mixture (1:2 or 2:1) along with concentrate⁽Reference Hess, Beuret and Lotscher²⁷⁾. Agarwal et al. ⁽Reference Agarwal, Kamra and Chaudhary⁵⁴⁾ reported that the depression in methane production was 96, 39·4 and 20 % with ethanol, water and methanol extracts of Sapindus mukorossi, respectively, compared with controls. Tea saponins deceased in vitro methane production by 13 % at a concentration of 67 mg/l to 22 % at a concentration of 133 mg/l, whereas greater concentrations (200 and 267 mg/l) had no further inhibitory effect on methane production⁽Reference Hu, Liu and Ye⁵⁰⁾. Methane production also decreased linearly with increasing concentrations of Yucca saponins in an in vitro medium containing different substrates such as potato starch, maize starch and hay-concentrate⁽Reference Lila, Mohammed and Kanda⁸⁷⁾. However, saponins from Acacia concinna extracts did not affect methane production or DM digestibility, although it reduced protozoa numbers⁽Reference Patra, Kamra and Agarwal¹⁰⁾. Similarly, Q. saponaria extract (0·12, 0·24 and 0·36 g saponins per litre) did not influence in vitro methane production⁽Reference Pen, Sar, Mwenya and Kuwaki⁵⁵⁾ whereas Y. schidigera extract (0, 0·18, 0·36 and 0·54 g saponins per litre)⁽Reference Pen, Sar, Mwenya and Kuwaki⁵⁵⁾ and whole Yucca plant (0, 15, 30 and 45 g/kg DM equivalent to 0, 0·023, 0·046 and 0·069 g saponins per litre)⁽Reference Holtshausen, Chaves and Beauchemin²⁵⁾ decreased in vitro methane production, with increasing concentrations of saponins⁽Reference Pen, Sar, Mwenya and Kuwaki⁵⁵⁾ accompanied by decreased protozoal numbers. Klita et al. ⁽Reference Klita, Mathison and Fenton¹⁷⁾ demonstrated that saponin extract (27·8 % saponins) from lucerne root did not affect methane production in sheep when administered intra-ruminally at the rate of 0, 1, 2 and 4 % of DM intake although protozoal numbers decreased linearly with increasing concentrations of saponins.

It has been observed that the effect of Sapindus saponaria on methane was more pronounced in the defaunated (29 %) than faunated (14 %) rumen fluid, indicating that reduced methane production was not entirely due to associated depression in protozoal counts⁽Reference Hess, Monsalve and Lascano⁶⁷⁾. Again, Goel et al. ⁽Reference Goel, Makkar and Becker⁹⁵⁾ noted that Sesbania sesban saponins decreased methanogen populations by 78 % and the methane inhibition effect of Sesbania sesban was more pronounced in concentrate-based diets compared with roughage-based diets.

Explaining the effects of saponins in modifying digestion and rumen fermentation

The presence of protozoa in the rumen has been shown to stabilise rumen pH and decrease the redox potential of rumen digesta, which should indirectly stimulate the cellulolytic bacterial activity⁽Reference Russell and Wilson⁹⁶⁾. Furthermore, approximately one-fifth of fibre degradation is of protozoal origin⁽Reference Dijkstra and Tamminga⁹⁷⁾. However, the engulfment and digestion of bacteria, fungal zoospores and archaea by protozoa causes a considerable proportion of microbial protein turnover in the rumen, decreasing the efficiency of protein utilisation in ruminants⁽Reference Wallace and McPherson⁹⁸⁾. It has been recognised that many of the negative effects of defaunation may disappear as the bacterial and fungal populations increase and occupy the niches previously filled by the protozoa⁽Reference Williams, Coleman, Hobson and Stewart⁵⁸⁾, which may increase the efficiency of microbial protein synthesis and protein flow to the duodenum and improvement of N retention in animals⁽Reference Santoso, Kilmaskossub and Sambodo¹⁹⁾. All of the studies that have reported anti-protozoal properties of saponins or plants containing saponins were shown to cause partial defaunation. Veira et al. ⁽Reference Veira, Ivan and Jui⁹⁹⁾ found that partially defaunated sheep (8·3 % of faunated sheep) were intermediate between observations in the completely defaunated and faunated sheep, and differences between partially and completely defaunated sheep were not significant. These findings suggest that responses of saponins to the efficiency of microbial protein synthesis and flow of amino acids to the intestine might be in between complete faunated and faunated animals depending upon the extent of the partial defaunation.

One of the pronounced effects of saponins reported in many of the studies is a decrease in ammonia concentration in the rumen, which may be primarily due to reduced bacterial lysis⁽Reference Wallace and McPherson⁹⁸⁾ as a consequence of anti-protozoal properties of saponins. The inhibition of protozoa may also directly depress the proteolytic and deaminative activities of protozoal origin. The reduced proteolysis and deamination may also be the result of inhibition of bacteria by saponins. Moreover, it is possible that direct inhibition of microbial urease by saponins may lead to reduced ammonia concentrations in the rumen⁽Reference Ellenberger, Rumpler and Johnson¹⁰⁰⁾. Additionally, the glycofractions of saponins are known to bind ammonia when concentrations of ammonia are low and when a high dose of saponins is added⁽Reference Wallace, Arthaud and Newbold⁴⁴⁾. This might improve N efficiency by maintaining adequate rumen ammonia (trapping of rumen ammonia when ammonia concentrations are high after feeding and then slowly releasing it with the degradation of the glycofraction and desaturation of ammonia binding) for use during microbial protein synthesis when ammonia concentrations decrease⁽Reference Wilson, Overton and Clark^36, Reference Wu, Sadik and Sleiman⁴⁰⁾. However, it has been shown that this ammonia binding appears to be negligible at ammonia concentrations typical of the rumen⁽Reference Wallace, Arthaud and Newbold⁴⁴⁾. Collectively, all of these factors may result in reduced ammonia concentrations in the rumen. Although the greater bacterial activities may increase proteolysis and deamination, the contribution to the total ammonia pool in the rumen from this process could be minor due to the inclusion of saponins in diets. Therefore, it is apparent that the variations in ammonia concentrations due to the addition of saponins are caused by the different contributions that protozoa and bacteria make in different studies.

The ligno-cellulosic feedstuffs are degraded in the rumen by the synergistic activities of the bacteria, protozoa and fungi, with bacteria and fungi contributing approximately 80 % of the degradative activity, and the protozoa only 20 %⁽Reference Dijkstra and Tamminga⁹⁷⁾. The most active fibrolytic bacteria F. succinogenes, Ruminococcus albus and Ruminococcus flavefaciens are generally considered as the primary organisms responsible for the degradation of plant cell walls in the rumen while Butyrivibrio fibrisolvens, Clostridium locheadii and Clostridium longisporum are some of the secondary fibrolytic bacteria⁽Reference Chesson, Forsberg, Hobson and Stewart¹⁰¹⁾. Ruminal fungi produce a broad array of enzymes and generally degrade a wider range of substrates than do rumen bacteria⁽Reference Chesson, Forsberg, Hobson and Stewart^101, Reference Trinci, Davies and Gull¹⁰²⁾. Furthermore, ruminal fungi are able to degrade the most resistant plant cell wall polymers and the cellulases and xylanases produced by them are among the most active fibrolytic enzymes⁽Reference Forsberg, Cheng, Bills and Kung¹⁰³⁾. The ruminal protozoa also contribute to the degradation of plant cell wall polymers, but their contribution in fibre degradation is considered not as important as that of the bacteria and fungi⁽Reference Chesson, Forsberg, Hobson and Stewart^101, Reference Lee, Ha and Cheng¹⁰⁴⁾. Some studies have demonstrated that saponins decrease the passage rate of digesta from the rumen⁽Reference Lu and Jorgensen¹⁸⁾, which may increase the ruminal degradation of feeds. However, physiological effects of saponins are usually overridden by microbiological effects in the rumen⁽Reference Lu and Jorgensen¹⁸⁾ because of comparatively greater effects of saponins on microbial populations. Consequently, the positive effects of saponins on the digestibility of feeds in some studies might be attributed to the increased bacterial populations, whereas negative effects reported in other studies are due to decreased hydrolytic enzyme activities from protozoa and/or bacteria and fungi when saponins affect these populations. The substrate-dependent effect of saponins on digestibility might arise due to the effects of saponins on specific bacterial populations⁽Reference Hussain and Cheeke⁴¹⁾. Wang et al. ⁽Reference Wang, McAllister and Yanke^59, Reference Wang, McAllister and Yanke⁸⁶⁾ observed that supplementation with Yucca extracts might be beneficial to ruminants fed a high-grain diet. Yucca saponins were found to have a direct negative effect on cellulolytic bacteria while being harmless to amylolytic bacteria⁽Reference Wang, McAllister and Yanke⁵⁹⁾. Additionally, Killeen et al. ⁽Reference Killeen, Madigan and Connolly¹⁰⁵⁾ proposed that a surfactant or flocculent action of saponins on the feed constituents that alters the rate of digestion may account for the substrate-dependent nature of the effect of saponins on rumen digestibilities of feeds.

The VFA concentration may generally follow the trend of feed digestion in the rumen. However, a part of the digested substrate may be partitioned to microbial mass, decreasing VFA concentrations⁽Reference Makkar, Sen and Blümmel⁴⁶⁾. Besides, saponins may affect intestinal cell permeability by forming complexes with sterol moieties in mucosal cell membranes⁽Reference Johnson, Gee and Price¹⁰⁶⁾. It has also been reported that saponins may reduce active nutrient transport mechanisms and increase membrane permeability in the small intestine to facilitate uptake of materials to which the gut would normally be impermeable⁽Reference Johnson, Gee and Price^106, Reference Knudsen, Jutfelt and Sundh¹⁰⁷⁾. Moreover, saponins may reduce the transmural potential differences in the small intestine⁽Reference Oleszek, Nowacka and Gee¹⁰⁸⁾. Thus, although there appears to be no study demonstrating the effects of saponins on absorption of VFA from the rumen, a reduction of VFA absorption would be possible, which may increase VFA concentrations in the rumen. It is well established that VFA in their undissociated forms are predominantly absorbed by passive transport mechanisms from the reticulo-rumen⁽Reference Gabel and Sehested¹⁰⁹⁾. But transport of dissociated VFA is coupled to the operation of an active transport mediated via a non-selective, electroneutral anion exchange system⁽Reference Gabel and Sehested¹⁰⁹⁾. Therefore, considering the predominant role of passive transport on absorption of VFA, the effect of saponins on VFA transport is likely to be small.

An increase in the proportion of propionate production and a decrease in the acetate:propionate ratio have been observed. These characteristics of VFA profile are often noted in studies with defaunated animals⁽Reference Williams, Coleman, Hobson and Stewart⁵⁸⁾. Acetate and butyrate are the major endproducts of fermentation in ruminal protozoa⁽Reference Williams, Coleman, Hobson and Stewart⁵⁸⁾. Hence, a decrease in protozoal numbers by saponins may result in an increase in the proportions of propionate and the propionate:acetate ratio. Besides, saponins sometime stimulate the growth of Selenomonas ruminantium ⁽Reference Wang, McAllister and Yanke⁵⁹⁾, which is predominantly responsible for propionate production from succinate metabolism⁽Reference Wolin, Miller, Stewart, Hobson and Stewart¹¹⁰⁾.

Several factors may determine the effect of saponins on methanogen populations and methane production, which are discussed below. The direct effect of saponins on methanogen populations has not been proved so far, but cannot be ruled out. Nevertheless, some studies have demonstrated that saponins may decrease the activity of methane-producing genes or rate of methane production per methanogenic cell⁽Reference Guo, Liu and Lu^51, Reference Hess, Kreuzer and Diaz⁵⁷⁾. Many methanogens are found in both etco- and endosymbiotic associations with protozoa, which are responsible for up to 37 % of rumen methanogenesis⁽Reference Finlay, Esteban and Clarke⁹³⁾. It has been estimated that a single protozoan cell may contain 10³–10⁴ methanogens before feeding, decreasing to one to ten methanogens after feeding⁽Reference Tokura, Ushida and Miyazaki⁹⁴⁾. Sharp et al. ⁽Reference Sharp, Ziemer and Stern¹¹¹⁾ reported that Methanobacteriaceae are the most abundant archaeal population both in the rumen fluid (89·3 %) and in a protozoal fraction (99·2 %), but it declined from 84 to 54 % in fermenters with the loss of protozoa. Methanomicrobiales are the second most abundant archaeal family (12·1 %) in the rumen and they were found to be predominantly free-living. The Methanosarcinales can be free-living or associated with protozoa. Hence, defaunation, which disrupts the symbiotic relationship, has been suggested as a method to reduce methane production in ruminants. Contrary to this, protozoa engulf rumen methanogens⁽Reference Tokura, Ushida and Miyazaki⁹⁴⁾. As a result, there may be an increase in methanogen numbers not associated with protozoa when protozoa are washed out in the rumen. For instance, Sharp et al. ⁽Reference Sharp, Ziemer and Stern¹¹¹⁾ showed that the numbers of Methanomicrobiales increased from 9·6 to 26·3 % to replace the methanogens lost via protozoal displacement, whereas Methanosarcinales remained constant in the fermenters. Additionally, methane release may also be affected by saponins as a result of a reduced rate of methanogenesis via diminished activity of the methane-producing mcrA gene without changing the total methanogen population⁽Reference Guo, Liu and Lu⁵¹⁾. Since saponins favour the production of an increased proportion of propionate, the channelling of hydrogen from methanogenesis to propionate production may cause lower methane production. In contrast, methane production may be increased because of greater digestion of feeds and fibre components, as reported in some studies on saponins⁽Reference Pen, Takura and Yamaguchia^24, Reference Valdez, Bush and Goetsch^29, Reference Goetsch and Owens³³⁾, probably resulting from increased bacterial and fungal populations. Besides, as there is an inverse relationship between digesta passage rate and methane production, probably due to an increased fibre degradation in the rumen, and saponins decrease the rate of digesta passage⁽Reference Klita, Mathison and Fenton^17, Reference Lu and Jorgensen¹⁸⁾, methane production may be increased as a result of an influence of saponins on passage rate. However, physiological effects of saponins appear to be insignificant compared with the microbiological effects⁽Reference Lu and Jorgensen¹⁸⁾. The rate of digesta passage could be affected by saponins via changes in the motility of the digestive tract. Klita et al. ⁽Reference Klita, Mathison and Fenton¹⁷⁾ demonstrated that the introduction of lucerne saponins inhibited rumen motility in sheep, while increasing the intestinal motility. No mechanism of action has yet been determined for the inhibitory effects of saponins on rumen motility. Klita et al. ⁽Reference Klita, Mathison and Fenton¹⁷⁾ assumed that epithelial receptors in the luminal epithelium might be involved, which is more likely to be a result of interactions of saponins with the sterols in biological membranes. The mechanism of saponins on intestinal motility might be through the modulation of pace-making cells in the intestinal muscles that generate the rhythmic oscillations in the membrane potential and this modulation is mediated through non-selective cation channels and intracellular Ca²⁺ mobilisation in a protein kinase C-independent manner, as demonstrated with ginseng saponins in mice⁽Reference Kim, Parajuli and Yeum¹¹²⁾.