EFFECT OF LERAK (SAPINDUS RARAK) EXTRACT IN HIGH ROUGHAGE DIET ON RUMEN MICROBIAL PROTEIN SYNTHESIS AND PERFORMANCE OF SHEEP

Secondary compounds containing plant extract as feed additive may improve the performance of livestock consuming high roughage diet. An in vivo trial was conducted to investigate the effect of Sapindus rarak extract (SRE) on ruminal fermentation products, microbial protein synthesis, and growth performance of sheep. Sheep (male, 28 heads) fed high roughage diet were arranged in a completely randomized design with four treatments: addition of SRE to the diet at 0, 4, 8 and 12 g head-1 day-1. The experiment was conducted for 105 days with 2 weeks adaptation period. At the end of the experiment, total faeces was collected for 1 week and rumen liquor was taken. Variables measured were ruminal fermentation products, microbial protein synthesis, daily intake, digestibility, N retention, body weight, and average daily gain. Protozoal numbers were significantly decreased with increasing SRE dose. The ruminal NH3 concentration tended to reduce by SRE addition (P = 0.06). SRE significantly increased propionate and efficiency of microbial protein synthesis from 6.4 to 10.5 g N kg-1 DOMR. SRE significantly improved average daily gain of sheep during the first 70 days of experiment without affecting intake and digestibility. SRE did not alter carcass percentage but tended to lower pancreas and liver weights (P<0.1). In conclusion, SRE has a significant role in partially defaunated rumen microflora, hence, increased microbial protein synthesis and propionate production in the rumen. Addition of SRE is useful to increase daily gain of sheep fed high roughage diet in 70 days of feeding. [


INTRODUCTION
In tropical countries, most feed especially during dry season contains limited amount of nitrogen that prevent rumen microorganism to have optimum growth. Feeding high roughage diet to their livestock has been practised by small farmers, hence, livestock performance usually in poor conditions especially during long dry season.
Feed additive is one of many strategies that could improve livestock performance. Plant extracts which contain secondary compounds have been reported to enhance livestock performance. Saponin which is one of the secondary compounds has the ability to defaunate the rumen by binding the sterol in protozoal membranes and lysed the protozoa. Protozoa in the rumen predates on bacteria causing further losses of microbial protein in the rumen, hence, less synthesis of microbial protein in the rumen and less nitrogen flow to duodenum. Several reports on the effects of saponins on rumen fermentation and feed digestibility are available (Wina 2012); Jayanegara et al. 2014;Liu et al. 2019).
One of saponin sources available in Indonesia is found in the fruit of Sapindus rarak known as soapnut or lerak. Lerak has been used widely in Indonesia as traditional soap for batik clothes. Saponin is the major compound in lerak fruit that caused foaming properties when the fruits were mixed in the water. As the bioactive compound, saponin content in lerak fruit was very high (32.54-35.98%) as reported by Pasaribu et al. (2014).
Experiments have been conducted to observe the effect of Sapindus rarak extract (SRE) on in vitro fermentation, the effect of interval feeding of SRE on feed digestibility (Wina et al. 2006b), and the effect of short and long term feeding of SRE on fibrolytic rumen microorganism (Wina et al. 2006a). However, there were only several data available on the effect of Sapindus saponins on microbial protein synthesis in the rumen, organ weight, and growth performance of livestock fed high roughage diet (Wang et al. 2012;Liu et al. 2019).
The objective of this study was to investigate the effect of saponins-containing methanol extract of SRE on ruminal fermentation products, microbial protein synthesis, organ weight, and growth performance of sheep consuming high roughage diet.

Preparation of Methanol Extract from Sapindus rarak
Sapindus rarak fruits (from Central Java, Indonesia) were dried in the oven at 60 o C, the the dried fruit powder was soaked with methanol (1:4w/v) for 16 hours. The extracts were separated from the residue, then the residue was soaked again with methanol for another 16 hours. The combined extracts were evaporated by rotary evaporator (Wina et al. 2006a). Then, the concentrated liquid was freeze dried. The products were very hygroscopic and had to be kept in airtight bag until used.

Feeding Experiment
Twenty-eight male sheep (thin tailed local Indonesian breed) with an average body weight of 16.8+1.9 kg, about 1-1.5 year old were kept in individual wooden slatted cage, about 0.75 m above the ground. The sheep were adapted to a mixture diet containing sugarcane tops and wheat pollard in the ratio of 65:35 w/w for 2 weeks. The diet was arranged as low quality diet and did not follow the nutrient requirement of the sheep. The experiment had four treatments of feeding different levels of SRE, i.e. 0 (control without addition of SRE), 4, 8 and 12 g SRE head -1 day -1 . SRE was added to wheat pollard and was offered before the sugarcane tops to ensure complete uptake of the saponin extract. The proximate composition of sugarcane tops and wheat pollard is presented in Table 1. Crude protein, neutral detergent fibre, and gross energy content of the diet were 92.3 g kg -1 , 646 g kg -1 and 15.3 MJ kg -1 DM respectively. Calcium in the form of limestone and salt were added at the level of 12 g kg -1 and 3.3 g kg -1 , respectively The diets were offered to the animals for 105 days at the level of 4% of body weight. Animals had free access to drinking water except during the digestibility trials where 2 liters of water in a bucket were offered daily. The body weight of sheep was recorded every second week before morning feeding.

Sample Collection
Faeces and urine collections were done according to the method described in IAEA protocol (International Atomic Energy Agency 1997). Faeces were collected every day for two periods of 7 days each; those periods were from day 28 to day 35 and from day 70 to day 77 of the feeding experiment. A bulk of faeces was weighed every morning before morning meal and a representative subsample (about 10%) was taken, oven dried at 60 o C and the weight was recorded. Daily samples were then combined for further analysis.
Urine was collected every morning in the same period of faeces collection. It was collected in a bucket containing 100 ml of 10% sulphuric acid solution. Volume was made up to 2 liters with water and an aliquot was taken and stored at -20°C. Daily samples from each animal were then pooled at the end of the trial and were analysed for purine derivative and nitrogen concentrations. Rumen liquor (50 ml) was taken from each sheep using a plastic stomach tube and filtered through two layers of cheese cloth. Rumen liquor was taken twice at day 103 and day 105 (the last day of the experiment).
For short chain fatty acid (SCFA) analysis, an aliquot of 1.5 ml was centrifuged at 14,000 g for 10 minutes. The supernatant (0.9 ml) was mixed with 0.1 ml of formic acid containing internal standard (10 ml l -1 of methyl valeric acid) and kept at 4 o C overnight. The next day, the mixture was centrifuged at 14,000 g for 10 minutes and the clear supernatant was transferred into a glass vial and kept at 4 o C for further analysis.
An aliquot (0.3 ml) of the filtered rumen liquor was collected into a vial containing the fixative reagent for protozoal counts and stored at 4°C. Another aliquot (1 ml) of the filtered rumen liquor was collected for ammonia determination, which was done immediately using the Conway method (Conway and Byrne 1933)

Chemical Analysis
Feed, feed residues and faeces were subjected to analysis of DM (24 h at 105 o C) and total ash (Association of Official Analytical Chemists 1984), nitrogen, NDF, and ADF lignin. Nitrogen content was determined by Kjeldahl and nitrogen released after digestion was then analysed by autoanalyser. The NDF content of sugar cane tops and faeces were determined without amylase and a NDF content of wheat pollard was analysed with heat stable α-amylase using Van Soest et al. (1991) method. The NDF values were not corrected for ash content. Gross energy was measured using bomb calorimeter.
Protozoa were fixed in a solution containing 4% formaldehyde 13.5 mM NaCl and 0.6 mg ml -1 methyl green and counted in a Neubauer Chamber. The data of protozoal count were presented in log 10 . Log reduction (L) = log 10 (control) -log 10 (treatment).
Percent protozoa reduction (P)= (1-10 -L )*100 Concentrations of allantoin, uric acid, xanthine and hypoxanthine in the urine were determined by the method described in (Chen and Gomes 1992).

Calculation of Microbial Protein Synthesis
Microbial protein synthesis and efficiency of microbial protein synthesis were calculated by formula as described in (Chen and Gomes 1992) based on the total purine derivatives excretion in the urine.
Notes: The digestibility of microbial purines is assumed to be 0.83.
The N content of purines is 70 mg N mmol -1 . The ratio of purine N : total N in mixed rumen microbes is taken as 11.6 : 100.
Efficiency of microbial protein synthesis (g N kg -1 DOMR) = microbial N synthesis DOMR -1 Where DOMR (digestible organic matter apparently fermented in the rumen) = 0.65 of the DOMI (digestible organic matter intake)

Weight Measurement of Carcass, Skin, Organs, and Fat Deposit Around the Organs
All sheep were slaughtered at the end of experiment (day 105). The sheep were slaughtered 3 hours after morning feeding. The carcass and skin weights were recorded. The fat covering the organs was collected and weighed separately. Empty body mass (EBM) was calculated by subtracting the weight of gut contents from the body weight. The carcass, organs and organ fat weights were expressed as g kg -1 EBM.

Statistical Analysis
Values of digestibility are the average from two periods of faeces and urine collection. The values were then analysed using GLM procedure in SAS package version 8. Means were compared by Duncan's multiple range test.

Rumen Parameters
Total short chain fatty acid (SCFA) concentration in the rumen was not affected by SRE ( Addition of SRE linearly decreased ruminal ammonia concentration. The concentration at the highest inclusion of SRE tended to reduce ammonia by 18% from that of control treatment (P = 0.06, Table 2). SRE significantly decreased protozoal counts in a dose dependent manner with high correlation between protozoal count to dose of SRE addition (r 2 = 0.9865). Compared to control, the protozoal count reduction were 33.9%, 60.2% and 80.0% at the addition of 4, 8 and 12 g SRE head -1 day -1 , respectively.

Intake, Digestibility, and Microbial N Synthesis
Addition of SRE increased DM roughage intake about 11-13.9% compared to control treatment, but the total DM intake and other nutrient intake in gram per day were not significantly different (P>0.05) among treatments (Table 3). Compared to control, the highest inclusion level of SRE (12 g head -1 day -1 ) slightly reduced the digestibilities of organic matter, neutral detergent fibre and nitrogen of the diet by 4.5%, 4.3% and 8.0%, respectively, however, these differences were not significant (P>0.05) among treatments.
Based on purine derivatives excretion in urine, the microbial N synthesis and efficiency of microbial N synthesis were calculated. The result in Figure 1 shows an increase of both parameters in a dose dependent manner with high correlation (r 2 = 0.9038 and 0.9376, respectively).

Average Daily Gain and Feed Conversion Ratio
(FCR) Figure 2 shows the average body weight of sheep fed different levels of SRE for 105 days. From the first week of experiment, there was a difference in average body weight between treatments of 8 and12 g SRE head -1 day -1 and treatments of 0 and 4 g SRE head -1 day -1 . The difference became bigger up 70 days of experiment. The curve starts to plateau after 70 days of feeding experiment. Table 4 shows that the lowest level of SRE had no influence on growth performance. However, the addition of higher levels of SRE (8 and 12 g head -1 day -1 ) significantly increased average daily gain by 37% and 44%, respectively up to 70 days, but not up to 105 days of experiment. Feed conversion ratio (FCR) at 0-70 days decreased by 30% and 28% for sheep fed 8 and 12 g SRE head -1 day -1 of SRE, respectively, but was not significantly different from that of control. The FCR at 0-105 days was similar among treatments.

The Weights of Carcass, Skin, Fat Deposits Around Organs and Organs
No significant difference was found in the weights of carcass and skin for all treatments (Table 5). The total fat deposits around the organs were also not significantly different among treatments. SRE did not cause any significant difference in all organs weights. However, pancreas and liver weights tended to reduce in a dose dependent manner (P <0.1).

Effect of SRE on Rumen Fermentation
SRE exerted a persistent effect on suppressing protozoa numbers in a dose dependent manner up to 105 days with high correlation (r 2 = 0.9865) between level of SRE and    Day of experiment protozoal count in the rumen. The ability to reduce protozoal count is due to the ability of saponin in SRE to interact with cholesterol in the outer membrane of the protozoa, causing lysis of the protozoal membrane (Wina 2012). SRE seemed to be one of the saponin sources that partially defaunate the rumen consistently both on the in vitro experiment (Wina et al. 2006b;Suharti et al. 2011) and also sheep for short term and over 3 months (Wina et al. 2006b). In the in vivo trials, antiprotozoal effect of other saponins tend to disappear after 7-14 days (Patra and Saxena 2009).

Parameter
Other saponin such as tea seed saponin has various effects. It had antiprotozoal effect on the in vitro rumen fermentation on sheep (Wang et al. 2012), but not on dairy cow (Guyader et al. 2017). The long term effect to reduce protozoal count may be caused by the presence of 20 different structures of saponins that have been isolated and identified in S. rarak pericarp's extract (Asao et al. 2009). These saponins have the same sapogenin and hederagenin, with different amounts and types of sugars attached to it and they may have different antiprotozoal activities and time of degradation in the rumen. 2017) suggested that a persistent effect to depress protozoa can be done by chemical modification of the group that attached to the sapogenin. The type of diet fed to the animal, besides the saponin structure may also affect the antiprotozoal activity of these saponins (Patra and Saxena 2009;Dai and Faciola 2019). In this experiment, types of protozoa that survived in the presence of SRE were not identified, but (Zhou et al. 2011) showed that tea saponin could selectively inhibit specific type of protozoa causing reduction of protozoal diversity in the rumen of sheep.
In this experiment, the major effects of SRE on ruminal fermentation were a shift of SCFA pattern towards an increased proportion of propionate and a decreased butyrate and valerate. The higher propionate proportion may be the result of lower acetate and butyrate, which are the major fermentation end products of protozoa. Therefore, when the number of prozotoa was depressed by saponin, a significant increase in the proportion of propionate would be expected. The increased propionate may also due to the increased population of Prevotella ruminicola which produce propionate and succinate in the rumen as reported by (Suharti et al. 2011) in their study on the effect of SRE in the in vitro fermentation using cattle rumen fluid. With higher propionate production in the rumen, there will be less hidrogen available for methane production, hence, less methane produced by sheep fed  Organ fat is the sum of the fat around heart, kidney, testis, mesenteric tissue and omentum. ‡ approaching significant (P<0.1). sem = standard error for means. SRE (Thalib et al. 2010 ;Yulistiani et al. 2017). With higher propionate and lower acetate obtained in this experiment, it may lead to more energy available for nutrient utilization and animal growth as propionate is the most important precursor of gluconeogenesis (Martínez-Aispuro et al. 2018).
As a consequence of partly defaunation, ruminal ammonia concentrations tended to decrease by SRE addition (Table 2). With less protozoa in the rumen, lower turn over protozoa occurred, hence, less ammonia from protozoa lysis was produced in the rumen since protozoa contributed some 10-40% of the total rumen nitrogen. The decreased ammonia production may be also due to the lower bacterial protein breakdown by protozoa and lower feed protein degradability in the rumen (Newbold et al. 2015). Although ammonia concentration decreased, its concentration in the rumen was not critical since it was still close to the value needed for the optimum microbial growth (0.15-0.25 mg N ml -1 ).
Microbial protein synthesis and efficiency of microbial protein synthesis that were estimated from purine derivative analysis increased linearly with an increasing level of SRE (Figure 1). The higher the level of SRE added into the feed, the higher the microbial protein synthesis and efficiency of microbial protein synthesis. The same result was reported with tea saponin when fed sheep with 3 g head -1 day -1 of tea saponin (Mao et al. 2010) and when fed goat with 0.4% DM intake day -1 of tea saponin (Kumar et al. 2017). As saponins reduced protozoa, a lower predation of bacteria by protozoa occurred and resulted in a higher number of bacterial population or higher microbial protein synthesis and a slower microbial protein turnover in the rumen leading to increase bacterial N flow to the duodenum (Belanche et al. 2011). This bacterial protein that flow to the duodenum will be hydrolysed to amino acid for building block for protein and muscular development body during sheep growth (Lopes et al. 2019). The low quality of forage will only contribute low nitrogen for microbial protein synthesis in the rumen and the addition of SRE would be beneficial for low quality forage diet.

Effect of SRE on Animal Performance
Addition of SRE 12 g head -1 day -1 to wheat pollard did not affect the palatability and consumption of concentrate or grasses. Saponin was reported to have bitter taste and bitter taste would reduce feed intake (Favreau et al. 2010). Favreau et al. (2010) used quinine solution as the bitter compound at the level of 2-3 g -1 kg sprayed on the hay, but at this experiment, the bitter taste of 4-12 g SE day -1 may be disappeared after mixing with wheat pollard and did not cause any reduction in daily intake of feed. There were also no reduction of DM, OM, NDF digestibilities or nitrogen retention by addition of 4-12 g (0.7-2 g SE 100 -1 g feed consumption). The results were similar to those reported by Yulistiani et al. (2017) that the addition of 1-2% complete rumen modifier (CRM) which also contained S. rarak pericarp powder did not affect DM consumption, DM, OM, NDF digestibilities, but with higher inclusion (3% CRM), DM, OM and NDF digestibility was significantly reduced compared to control (without CRM). Protozoa have several fibrolytic enzymes (Newbold et al. 2015), therefore NDF and ADF digestibilities reduced with elimination of protozoa but those perhaps were not affected with partly defaunation.
SRE increased the body weight gain of male sheep (Table 4). It was reported that saponins enhanced the performance of ruminants fed roughage-based diets (Patra and Saxena 2009). However, various results were reported on the effect of saponin on growth performance. Complete rumen modifier that contained S. rarak fruit powder had a significantly positive effect on weight gain of sheep fed rice straw diet (Thalib et al. 2010), but not significant when sheep fed corn cob silage compared to control (Yulistiani et al. 2017). S. rarak fruit extract when fed to Ongole cattle at the level of 100-200 mg kg -1 feed caused only a trend increase of body weight gain (Suharti et al. 2015). Other saponin source, tea saponin, significantly increased weight gain of goat fed at the level of 0.4% DM intake for 120 days (Kumar et al. 2017). The effect of SRE on body weight gain of sheep was observed immediately and significantly at the early weeks up to 70 days of feeding trial, but not significant up to 105 days of feeding trial. The variations among animals in this experiment were quite high and the weight gain of the animals to 105 days of experiment became smaller and was not signficantly different among treatments.
It was reported by Patra and Saxena (2009) that the different types and levels of saponins and also the different other secondary compounds in the saponin extract or saponin containing plant materials may affect the various results of the effect of saponin on growth performance. It may also be due to the different compositions, levels of protein and energy of diet .
There are very limited data on the effect of saponins on carcass weight. In the present experiment, the carcass weight was not affected by SRE feeding. The percentage of carcass to the empty body weight was in average of 54%. Sen et al. (2000) reported that total defaunation did not cause any difference in carcass weight and the proportion of carcass to the empty body mass was 52%. Eugène et al. (2004) summarised the data from 17 trials and reported that defaunation increased the ruminal volume, which might contribute to the increase in weight gain of defaunated animal. The fat deposits were not different for all treatments (Table 5) indicating that the increase of body mass gain was not distributed to the fat deposit around organ tissues.
Feed conversion ratio up to 70 days was reduced by 30% when sheep was fed 8 g SRE head -1 day -1 suggesting that saponin-fed sheep may be more efficient in the utilization of feed than control sheep. Newbold et al. (2015) reported that the efficiency of energy utilization for growth improved even though body energy retention did not increase.
In this experiment, the increasing level of SRE added to the diet seemed to cause smaller weights of pancreas and liver, but these weights were not significantly different among treatments. The occurrence of liver damage has been reported on the animal that suffered from photosensitization caused by some saponin containing plant materials (Low 2015). Saponin that ingested into digestive tract will be degraded in the rumen and the sapogenin may be modified or transformed to conjugated sapogenins that present in the bile and liver of animals and cause some cells damaged in liver and kidney (Low 2015). Sheep during feeding trial were inside the animal house, therefore there was no report on photosensitization case in this experiment. Smaller weight of pancreas and liver may not affect growth since higher weight gain was obtained with higher SRE level addition up to 70 days of trial.
Saponin was reported to inhibit lipase activity produced in pancreas (Marrelli et al. 2016), however, the relation between lipase activity inhibition and smaller size of pancreas has not yet been clear. Inhibition of lipase activity resulted in lower cholesterol in blood. In the present study, we did not find any visual changes in the outside appearance of the liver, kidney or pancreas. Further study on the effect of S. rarak saponin extract on the health of the animal is needed.

CONCLUSION
Sapindus rarak saponin extract has a persistent effect on suppressing the protozoal numbers and shift SCFA profile. Sapindus rarak saponin extract increased weight gain of the animals fed high roughage diet up to 70 days of experiment by enhancing microbial protein synthesis and molar proportion of propionate.