Effect of Nigella sativa L. as Saponin Sources on In vitro Rumen Fermentation, Enzyme Activity and Nutrients Digestibility

| The effect of Nigella sativa L. as saponin sources on in vitro rumen fermentation, enzyme activity, and nutrients digestibility was investigated in this study. The diet consisted of 70% Napier Grass (Pennisetum purpureum) and 30% wheat pollard containing 0%, 0.2%, 0.4%, and 0.6% saponin, respectively. In vitro fermentation was conducted using Menke and Steingass gas production and two-stage Tilley and Terry. The acquired data were subjected to variance analysis and Duncan’s Multiple Range Test (DMRT). The addition of Nigella sativa L. meal significantly decreases (P<0.05) methane (digested organic matter), protozoa population, and increased microbial protein but does not affect (P>0.05) pH, ammonia concentration, and volatile fatty acids (VFA) production. CMCase enzyme activity reduced (P<0.05) as saponin levels increased, but amylase and protein of enzyme were unaffected (P>0.05). Saponin reduced protein digestibility in the rumen, but there was no difference in dry matter (DM) or organic matter (OM) when compared to the control (P>0.05). Post-rumen nutrients digestibility did not affect by the saponin diet (P>0.05). In conclusion, saponins reduced protozoa population, CH4 production, increased microbial protein, and improve rumen protein digestibility. Therefore, we recommend the use of 0.4% saponins Nigella sativa L. to reduce CH4 production without affecting rumen fermentation.


INTRODUCTION
T he livestock industry contributes to global climate change as a result of greenhouse gases (GHG). The main GHGs from the livestock sector are carbon dioxide (CO 2 ), methane (CH 4 ), ammonia (NH 3 -N), and nitrous oxide (N 2 O) (Dopelt et al., 2019;Grossi et al., 2019). Methane emissions are not only related to environmental problems but also represent a large part of the energy lost, consequently not utilized for livestock production processes. According to Johnson and Johnson (1995) and Thompson and Rowntree (2020) about 2 to 12% of the consumed feed energy is lost due to the formation of CH 4 . Therefore, a mitigation strategy is needed to reduce the rate of GHG accumulation and energy loss that impacts ruminant performance.
An effort to reduce CH 4 emissions and improve ruminant productivity is to utilize the plant's secondary metabolites, one of which is saponins. Saponins decrease the population of rumen protozoa that contribute to CH 4 formation by lysing protozoa cells (Patra and Saxena, 2009;Ramos-Morales et al., 2017;Li et al., 2018). The use of saponins can increase the efficiency of microbial protein synthesis and protein flow into the duodenum (Ramaiyulis et al., 2018;Unnawong et al., 2021). In addition, a decrease in Nigella sativa L. is one of the plant sources of saponins. Nigella sativa L. meal is by-product of extracting oil rich in nutrient such as crude protein (33.13%), fat (12.72%), all essential amino acids and also fatty acids. According to Abbas et al. (2013) and Michel et al. (2011) Nigella sativa L. contains 4.54% saponins, and other secondary metabolites such as flavonoids and alkaloids. In addition, Nigella sativa L. has antioxidant, antifungal and antibacterial properties (Niu et al., 2020). Unnawong et al. (2021) reported that the use of saponins from Sesbania grandiflora pod meal significantly reduced the digestibility of protein, NH 3 -N, protozoa, NH 4 , while the total VFA increased when used in Thai Purebred Beef Cattle feed. The use of other plants containing saponins such as Delonix regia seed meal decreased CH 4 production and increased DM digestibility, but did not affect total VFA (Supapong et al., 2017). Anantasook et al. (2016) reported that the use of Terminalia chebula Retz. increased total VFA and decreased CH4, protozoa and acetate concentrations. However, the effect of using Nigella sativa L. meal containing saponins on the rumen of cattle has not yet been evaluated.
The hypothesis was made: Nigella sativa L. decreased CH 4 production and protozoa population in the rumen, but increased propionate concentration, enzyme activity and feed digestibility. Based on the description, this study aimed to determine the effect of adding Nigella sativa L. as a source of saponins to reduce CH 4 gas emissions on rumen fermentation and nutrient digestibility in vitro.

SampleS ColleCtion and preparation
Pennisetum purpureum grass was oven-dried at 55°C for three days subsequently ground to pass through a 1 mm sieve. Samples of grass, pollard, and Nigella sativa L. meal were analyzed proximately according to the AOAC (2005) method. Analysis of the saponin content in Nigella sativa L. using the spectrophotometric method according to Uematsu et al. (2000).

animalS and preparation oF rumen inoCulum
One Bali Cattle with body weight (BW) 400kg was fed 70% forage and 30% concentrate separately for feed adaptation. Feed was given twice a day with a portion of 60% in the morning and 40% in the afternoon for ten days before rumen fluid was used. Rumen liquor was obtained from Bali cattle before the morning feeding. The rumen fluid was filtered through cheesecloth into pre-warmed thermo flasks with water at 39ºC and then transported to the laboratory.

In vItro Fermentation and Sample analySiS
In vitro fermentation was carried out using two techniques, gas production according to Menke and Steingass (1988), to measure fermentation parameters and enzyme activity in the rumen and two steps Tilley and Terry (1963) to determine feed digestibility. For each treatment, three replications was prepared.

menke and SteingaSS
Menke and steingass gas production was carried out using 300 mg of substrate consisting of 70% Pennisetum purpureum grass, 30% pollard, and Nigella sativa L. meal as a source of saponins weight into 100 mL syringe. Nigella sativa L. was added until the feed saponin content was 0%, 0.2%, 0.4%, and 0.6%. Each treatment was replicated three times, and each replication was duplicated. The chemical composition of feed ingredient and proportion of dietary treatment was shown in Tables 1 and 2. Ruminal fluid was mixed with the artificial saliva solution of Menke and Steingass (1988) in a proportion 2:1 (v/v) at 39°C under continuous flushing with CO 2 and 30 mL of rumen inocula mixture were added into each syringe under CO 2 flushing. The syringe was sealed with rubber stoppers and aluminum caps and incubated at 39°C (48 h) for the in vitro gas test. A syringe without feed with 2 replications was used as a blank containing rumen fluid and a standard syringe filled with 300 mg of Pangola grass. Gas production was measured at 1, 2, 4, 6, 8, 12, 24, 36, and 48 hours. When the volume reached its maximum, the gas was removed and calibrated syringe.

Advances in Animal and Veterinary Sciences
December 2021 | Volume 9 | Issue 12 | Page 2249 At the end of 48 hours of incubation, the fermented gas sample was used to analyze the levels of methane and carbon dioxide by gas chromatography method (Filípek and Dvořák, 2009). The fermented substrate was filtered, the residue was used to measure digestible dry matter and digested organic matter. The filtrate was used for pH measurement using pH meter (Hanna Ckecker 1 pH Taster, Hanna Instruments, Ann Arbor, Michigan, USA) and protozoa assays (Diaz et al., 1993). The other filtrate was centrifuged (3000 rpm, 10 min) to determine the ammonia content using the spectrophotometric method (Chaney and Marbach, 1962), and the remainder was centrifuged (10,000 rpm, 15 min). The precipitate was used for microbial protein assay by the Lowry method (Plummer, 1987), while the supernatant was used to determine VFA using gas chromatography method (Filípek and Dvořák, 2009) and the activity of CMCase and amylase enzymes (Halliwell, 1961).

tilley and terry
Two steps Tilley and Terry's in vitro fermentation was incubated for 48 hours and 96 hours, respectively. The first step was digestibility in the rumen, a total 56 test tubes with a volume of 50 mL were filled with 250 feed material substrate and Nigella sativa L. with levels of 0, 0.2%, 0.4%, and 0.6%, rumen fluid and medium solution. The medium solution was mixed with rumen fluid in a 4:1 ratio. The medium solution used for testing the digestibility of DM and OM was 50 mL and 100 mL for the CP digestibility. Tubes without feed samples were used as blanks. The mixed solution is then flushing with CO 2 and sealed with a rubber stopper that has a valve to release the gas from the fermentation. The tubes were incubated for 48 hours at 39°C. The 28 tubes containing the fermented substrate were filtered using a glass wool crucible. Tilley and Terry 48 h in vitro fermentation residue was employed to determine the digestibility of crude protein (CP, OM, and DM in the rumen. The second stage is post-rumen digestion. A total of 28 tubes containing substrate after being incubated for 48 hours were added with 3 ml of 20% HCl and 1 ml of 5% pepsin (3:1). The tubes that had been added with HCl and pepsin were incubated for another 48 hours at 39°C. After incubation, the fermented substrate was filtered, and the residue was analyzed for CP, OM, and DM to determine post-rumen nutrient digestibility.

data analySiS
Data were analyzed by one-way analysis of variance (ANOVA) using the Statistical Package for the Social Sciences (SPSS) program version 16.0. Differences between treatment means were further analyzed using Duncan's Multiple Range Test (DMRT) test (Gomez and Gomez, 1984).

eFFeCt oF nIgella satIva l. on rumen Fermentation and gaS produCtion parameterS
The effect of using Nigella sativa L. meal as a source of saponins on rumen fermentation parameters and gas production is presented in Table 3. The addition of Nigella sativa L. meal with 0.6% saponin content did not affect (P>0.05) the rumen pH value. Rumen pH values in this study ranged from 6.72 to 6.81 and were still in the normal pH range of 6.3 to 7, according to Reis et al. (2014). A previous study showed that the use of 5% Ageratum conyzoides leaf extract and 5% Zingiber officinale in the diet did not affect pH, with a pH range of 6.83 to 6.95 (Hapsari et al., 2018). Another study showed that supplementation of 0.4 and 0.6% of Sesbania grandiflora pod meal did not affect rumen pH, ranging from 6.76-6.80 (Unnawong et al., 2021). Variations in rumen pH value are influenced by the feed consumed. Feeds containing grains caused a decrease in the pH to less than 5.0, while fibrous feeds increased the pH by more than 7.0 (Sung et al., 2007;Kim et al., 2018;Ogata et al., 2019). The fiber content of feed can affect rumen pH. According to Dijkstra et al. (2012), the content of VFA as a result of fiber degradation by cellulolytic microbes is closely related to the rumen pH value.
The addition of Nigella sativa L. meal with 0.6% saponin content did not affect (P>0.05) the rumen NH 3 -N concentration. In previous studies, the use of 0.52% tea saponins did not affect the rumen NH 3 -N content of lactating dairy cows (Guyader et al., 2017). In another study conducted by Aazami et al. (2013), the use of 0.02% saponins in Baluchi sheep feed and 0.054% in Saanen kids did not affect rumen NH 3 -N concentrations. The concentration of NH 3 illustrates the degradation of protein by microbes. Microbes will degrade more than 60% of the protein in the rumen into amino acids, peptides, and NH 3 -N (Kamalak et al., 2005;Liu et al., 2019a). The higher the protein content, the higher the concentration of NH 3 -N produced (Griffin and Bradshaw, 2019). According to , Rumen NH 3 -N is utilized for microbial protein synthesis. Rumen NH 3 -N concentrations ranged from 8.5 to 30 mg/100 mL (McDonald et al., 2012), excess NH 3 -N more than 50 mg NH 3 -N/L did not affect microbial protein synthesis and was excreted (Satter and Slyter, 1974;Neto et al., 2019).
When compared to the control, the usage of Nigella sativa L. meal containing saponins 0.4 and 0.6% enhanced (P<0.05) 8.8 and 24.44% microbial protein, respectively, while the 0.2% level had no effect. Previous studies showed

Advances in Animal and Veterinary Sciences
December 2021 | Volume 9 | Issue 12 | Page 2250 that the use of 4% papaya (Carica papaya L.) leaf extract and powder increased 9.70% and 8.59% rumen microbial protein (Sairullah et al., 2016). Another study showed that the use of 0.6% Flemingia macrophylla silage as a source of saponins increased 43.64% of microbial protein synthesis (Viennasay and Wanapat, 2020). The main effect of saponins in the rumen appears to inhibit protozoa (defaunation), which might increase the efficiency of microbial protein synthesis and protein flow to the duodenum (Patra and Saxena, 2009). Protein fermentation in the rumen produces the final product of NH 3 -N, which is essential for synthesizing microbial protein in the rumen. Saponins can lyse protozoa by forming complex bonds with sterols on the surface of the protozoa membrane (Hanim et al., 2009;Ramos-Morales et al., 2019). The use of Nigella sativa L. as a source of saponins aims to defaunate protozoa so that bacteria and rumen microbes. Defaunation reduces NH 3 -N concentrations due to low levels of feed protein degradation (Harun and Sali, 2019). In this study, NH 3 -N decreased but not significantly while microbial protein increased. The efficiency of microbial protein synthesis is influenced by several factors such as NH 3 -N, availability of energy and carbon skeleton, mineral supply, consumption rate, and flow rate of feed particles in the rumen (Hackmann and Firkins, 2015;Harun and Sali, 2019).
The addition of Nigella sativa L. meal as a source of saponins significantly (P<0.05) increased 21.23% of the protozoa population at 0.2% levels but decreased the protozoa population to 19.86% at 0.4% levels compared to control. The 0.6% treatment had no significant effect when compared to the 0.4% level. Previous studies showed that the use of 0.50 g/L tea saponin powder reduced 9.64% of the rumen protozoa population (Guyader et al., 2017). Research conducted by Unnawong et al. (2021) showed that the use of 0.4% and 0.6% Sesbania grandiflora pod meal reduced the protozoa population from 9.22×10 5 cells/mL to 6.73×10 5 cells/mL. The decrease in protozoa with the addition of saponins was caused by the ability of saponins to lyse protozoa. Saponins are toxic to protozoa due to forming complexes with lipid membranes, which increase permeability, cause imbalance, and consequently promote cell lysis (Makkar et al., 1995;Fleck et al., 2019). In addition, saponins bind to sterols in the cell walls of protozoa, thereby changing the permeability of cell membranes (Patra et al., 2006;Belanche et al., 2016;Anggraeny et al., 2021). The decrease in the protozoa population led to increased total bacteria in the rumen but decreased bacterial methanogenesis (Ozutsumi et al., 2005;Li et al., 2018).
The addition of Nigella sativa L. saponins tended to decrease CH 4 production compared to control, but increased at 0.6% level. Expressed in mL per mg of digested DM, CH 4 production was similar but increased at the 0.6% level. In contrast, CH 4 production was reduced to a level of 0.4% but increased by 0.6% in mL per mg of ingested OM. The production of CO 2 gas increases with the addition of saponin levels. Supplementation of Nigella sativa L. reduced the protozoa population but was not accompanied by a decrease in methane gas production. Previous research showed that the use of Sesbania sesban leaves and Fenugreek seeds (Trigonella foenum-graecum L.) reduced the number of protozoa but did not reduce methane production (Goel et al., 2008a). Patra et al. (2006) showed that methanogenesis

Advances in Animal and Veterinary Sciences
December 2021 | Volume 9 | Issue 12 | Page 2251 is essentially unrelated to the density of protozoa. No inhibition of methane production with a decrease in methanogens could be due to (1) slow association between protozoa and methanogens due to the higher generation time of protozoa compared to methanogens, (2) increased metabolism of species-independent methanogenic microbes persisting after addition of saponins and/ or (3) by changes in the composition of the methanogenic community and increased efficiency of methane production . In addition, it can be said that at the inhibition of protozoa, species belonging to Methanobacteriaceae (living in association with protozoa) decreased with an increase in the number of free-living Methanobacteriales. Reduced levels of association of protozoa and methanogens may result in higher interspecies hydrogen transfer between an increased population of hydrogen-producing bacteria (R. flavefaciens and F. succinogenes) and free-living Methanobacteriales showing no effect on methane production (Goel and Makkar, 2012). Guyader et al. (2017) also reported that the use of tea saponins increased methane production in g/kg of digested DM and g/kg of digested OM lactating cows. In the study of Goel et al. (2008b), the saponin-rich fraction did not affect digestibility and a slightly higher tendency to produce gas, which may be due to the increase in saponin-mediated fiber-degrading bacteria. Therefore, it is possible to increase the production of CH 4 and CO 2 . According to Ridla et al. (2021), the use of saponins in low levels (less than 0.5% DM) beneficially reduces enteric methane emissions and stimulates nutrient digestibility. The addition of high amounts of saponins appears to cause adverse effects on nutrient digestibility without further reduction in methane emissions.
The addition of Nigella sativa L. meal as a source of saponins to a level of 0.6% did not affect (P > 0.05) the production of acetate (C2), propionate (C3), butyrate (C4), and total VFA production. The same result was shown by Anggraeny et al. (2021) that the use of 4% P. falcataria and S. saman leaf meal did not affect the total VFA production and the proportion of C2, C3, and C4 in the rumen. Another study showed that 50% Garcinia mangostana L. as a source of saponins did not affect the total and proportion of goat rumen VFA (Shokryazdan et al., 2016). According to Gunun et al. (2018), the use of 20 mg rambutan peel powder did not affect the production and proportion of VFA and the ratio of C2:C3. The decrease in the number of protozoa led to an increase in amylolytic bacteria that digest starch and produce propionate as part of the VFA (Li et al., 2018;Park et al., 2019). Hess et al. (2003), Qin et al. (2012), and Nguyen et al. (2020) stated that decreased levels of protozoa were associated with an increase in propionate and a decrease in the acetate to propionate ratio. The formation of C2 will produce H 2, which is the precursor for the formation of CH 4 , while the formation of C3 requires H 2 . Increased C3 production will reduce CH 4 formation in rumen fermentation (Moss et al., 2000;Santos et al., 2016;. The decrease in the number of protozoa in this study did not affect the proportion of rumen VFA.

eFFeCt oF nIgella satIva l on rumen enZymeS aCtivity
The effect of Nigella sativa L. meal as the source of saponin on enzyme activity in the rumen is shown in Table 4. The addition of Nigella Sativa L. meal as a source of saponins significantly (P<0.05) decreased the activity of the CMCase enzyme in rumen fluid. The use of 0.2%, 0.4% and 0.6% of saponin caused a decrease in CMCase enzyme activity by 25.92%, 41.8% and 27.5%, respectively. However, the increase in saponins from 0.2% to 0.4% and 0.6% did not cause a decrease in the activity of the CMCase enzyme. Another study conducted by Belanche et al. (2016) showed that the use of 15% saponins from Ivy fruit (Hedera helix) did not affect the enzyme activity of the CMCase in the Rusitec system. Hristov et al. (2003) reported that the use of saponins from the extract of Yucca schidigera decreased the activity of CMCase, xylanase, and amylase in vitro.
CMCase is a cellulase enzyme secreted by bacteria and fungi in the rumen (Kirn et al., 2018). Wang et al. (1998) noted a 30% reduction in cellulolytic bacteria when Yucca schidigera was added to the fermenter in a rumen simulation (Rusitec) device. According to Wina et al. (2005a) the use of Sapindus rarak saponins reduced the concentration of the bacteria Ruminococcus albus, Ruminococcus flavefaciens and anaerobic rumen fungi (Chytridiomycetes) causing a decrease in CMCase enzyme activity. However, some researchers revealed that decrease in xylanase or CMCase activity in the rumen seemed closely related to decreasing protozoa rather than decreasing cellulolytic microbes. There is a significant relationship between the number of protozoa and the activity of cellulolytic enzymes (Williams et al., 2020). Wina et al. (2005b) and Patel and Ambalam (2018) stated that protozoa also secrete cellulolytic enzymes and contribute 19-28% of the total rumen cellulolytic activity. In addition, Newbold et al. (2015) reported that based on meta-analysis, the absence of protozoa caused a decrease in the concentration of anaerobic fungi (−92%), Ruminococcus albus (−34%), and Ruminococcus flavefaciens (−22%), thereby affecting the activity of cellulolytic enzymes. Differences in substrates and bioactive components in plants are assumed to cause differences in research results.
The use of saponins as feed additives with levels of 0.2%, 0.4%, and 0.6% did not cause a significant difference (P>0.05) on the amylase enzyme activity in rumen fluid fermentation. Previous research showed that the use of 200 ppm Hibiscus tiliaceus leaf extract did not affect the amylase enzyme activity (Bata and Rahayu, 2016).

Advances in Animal and Veterinary Sciences
December 2021 | Volume 9 | Issue 12 | Page 2252  Another study conducted by Belanche et al. (2016) showed that 15% of saponins from Ivy fruit (Hedera helix) did not affect the amylase enzyme activity in the Rusitec system. Several studies have shown that saponins inhibit the activity of the amylase enzyme activity (Ali et al., 2006;Ercan and El, 2016;Hanh et al., 2016). It is conceivable that the lack of variation in amylase enzyme activity was due to saponins' lack of effect on amylolytic bacteria. Yucca schidigera extract saponins were found to have a direct negative effect on cellulolytic bacteria while being harmless to amylolytic bacteria (Wang et al., 2000;Patra and Saxena, 2009). Abdel-Raheem et al. (2019) stated that the use of saponins decreased NH 3 -N levels, thereby reducing the population and activity of rumen fibrolytic bacteria (mostly Gram-positive) and increasing amylolytic bacteria (predominantly Gram-negative); therefore, amylase activity increased. Several studies have shown that saponins inhibit the activity of amylase enzyme activity (Hristov et al., 2003;Ercan and El, 2016;Hanh et al., 2016;Samtiya et al., 2020)  The effect of giving Nigella sativa L. meal as saponin sources on nutrient digestibility in the rumen and postrumen is presented in Table 5. The use of 0.6% Nigella sativa L. meal as a source of saponins significantly (P<0.05) decreased 12.18% rumen digestibility of CP but did not affect the digestibility of OM and DM. Previous research has shown that 80% Tithonia diversifolia flower extract can reduce the digestibility of CP ( Jamarun et al., 2016). Hanim et al. (2017) showed that the addition of 0.2 mg of saponin Hibiscus rosa-sinensis L. leaves/100 ml of medium did not affect the digestibility of OM and DM. Another study conducted by Khoiriyah et al. (2016) showed that the addition of 4% papaya leaf powder and extract (Carica papaya L) did not affect the digestibility of OM and DM. Saponins reduce protein digestibility probably by forming difficulty digestible saponins-protein complexes, thereby inhibiting the microbial fermentation of protein (Potter et al., 1993;Francis et al., 2002;Das et al., 2012;Lakram et al., 2019). In some studies, there was a substantial decrease in the protozoa population and no compensatory increase in the bacterial population, and there was a decrease in digestibility. Differences in CP digestibility could be related to many factors, including supplement sources, sources' form, study dose, and diet composition ( Jafari et al., 2019).
The effect of using 0.6% saponin Nigella sativa L. meal did not affect (P > 0.05) post-rumen digestibility of CP, OM, and DM. Although the activity of the CMCase enzyme decreased with 0.2% saponin administration (Table 4), this may not have caused a decrease in crude fiber digestibility, so the digestibility of OM and DM was not affected. The same result was shown by Aazami et al. (2013) that the use of 0.02% Quillaja saponaria powder did not affect the digestibility of CP, OM, and DM Baluchi sheep. Another study showed that the use of 0.6% S. graniflora pods meal

Advances in Animal and Veterinary Sciences
December 2021 | Volume 9 | Issue 12 | Page 2253 decreased the digestibility of CP but did not affect the digestibility of OM and DM Thai Purebred Beef Cattle (Unnawong et al., 2021). Kumar et al. (2017) showed that the use of 5.2% tea seed and 0.8% tea saponins did not affect the digestibility of CP, OM, and DM Gaddi kids. Makkar et al. (1993) stated that saponins might affect or have no effect on nutrient digestibility. The addition of saponins in the feed gave an inconsistent effect. These differences appear to be related to the chemical structure and dosage of saponins, diet composition, microbial community, and adaptation of the microbiota to saponins (Patra and Saxena, 2009;Singh and Kaur, 2020). The use of saponins has been reported to increase nutrient digestibility (CP, OM, DM, Acid detergent fiber (ADF), and Neutral detergent fiber (NDF)) (Wei et al., 2012;McMurphy et al., 2014;Liu et al., 2019b), but several studies have shown that saponins reduce nutrient digestibility (CP, OM, DM, and NDF) (Santoso et al., 2007;Jadhav et al., 2018;Dai and Faciola, 2019).

CONCLUSIONS AND RECOMMENDATION
The addition of Nigella sativa L. meal as a source of saponins decreased protozoa population and CH 4 production. As saponin levels increased, microbial protein increased and rumen protein digestibility was improved but CMCase enzyme activity was reduced. The authors recommend the addition of 0.4% Nigella sativa L. saponin to reduce CH 4 production without affecting rumen fermentation.

AUTHOR'S CONTRIBUTION
Satyaning Widyarini and Faradista Sekar Nagari performed the experiment, collected the sample, and wrote the manuscript. Chusnul Hanim designed and supervised the experiment. Zaenal Bachruddin supervised the study and revised the manuscript. Muhlisin analyzed the data and supervised the experiment. Lies Mira Yusiati designed and supervised the experiment and revised the manuscript. All authors read and approved the final manuscript for publication.