Comparative analysis of silage fermentation and in vitro digestibility of tropical grass prepared with Acremonium and Tricoderma species producing cellulases

Objective To find out ways of improving fermentation quality of silage, the comparative analysis of fermentation characteristics and in vitro digestibility of tropical grasses silage applied with cellulases produced from Acremonium or Tricoderma species were studied in Thailand. Methods Fresh and wilted Guinea grass and Napier grass silages were prepared with cellulases from Acremonium (AC) or Trichoderma (TC) at 0.0025%, 0.005%, and 0.01% on a fresh matter (FM), and their fermentation quality, chemical composition and in vitro digestibility were analyzed. Results All silages of fresh Napier grass were good quality with lower pH, butyric acid, and ammonia nitrogen, but higher lactic acid content than wilted Napier grass and Guinea grass silage. Silages treated with AC 0.01% had the best result in terms of fermentation quality. They also had higher in vitro dry matter digestibility and in vitro organic matter digestibility at 6 and 48 h after incubation than other silages. Silages treated with lower levels at 0.005% or 0.0025% of AC and all levels of TC did not improve silage fermentation. Conclusion The AC could improve silage fermentation and in vitro degradation of Guinea grass and Napier grass silages, and the suitable addition ration is 0.01% (73.5 U) of FM for tropical silage preparation.


INTRODUCTION
Purple Guinea grass (Panicum maximum cv. TD 58) and Napier grass (Pennisetum purpureum ×Pennisetum americanum cv. Pak Chong 1) are widespread throughout tropical and sub tropical areas. They adapt and grow well in a variety of soil types, are tolerant to different conditions and require quite low inputs for growth [13]. Both grasses are important rough age sources for ruminant feed because of high dry matter biomass production; ranged from 15 to 18 t/ha per year for Guinea grass [1] and 63 to 87 t/ha per year for Napier grass [3,4]. However, both grasses are seasonal with high dry matter yield over the short rainy season; thus, silage making is an increasingly important method to preserve fresh forages for feeding ruminants yearround [5]. It is well known to be difficult to make good quality silage with these species because they are usually low in water soluble carbohydrate (WSC) content, insufficient for rapid decrease of silage pH. They are also low in dry matter and high in fiber content which promotes the growth of clostridia and decreases nutrient utilization of ru minants [68].
Cellulase enzymes are popular biological products for improving the fermentation quality of silages. It is widely accepted that they can improve the fermentation process of both grasses and legumes and thus enhance cattle performance [68]. The two main reasons for cellulase enzyme addition in silage making are firstly, to enhance plant fiber degradation to increase the content of WSC as a lactic acid bacteria (LAB) substrate to produce lactic acid and en hance the fermentation process [911] and secondly, to break down the structural carbohydrate component thus improving silage digestibility [10,12,13]. Many microorganisms that pro duce various cellulolytic enzymes have been studied for several decades. The genus of Acremonium and Trichoderma has been especially famous for producing cellulolytic enzymes with rela tively high enzymatic activity [14]. There is increasing interest in applying cellulase enzymes at ensiling due to reports that it can improve the fermentation quality of silage [15,16].
However, there is limited information available for compar ative analysis of silage prepared with these cellulases produced from Acremonium and Trichoderma. The objectives of this study were to determine the fermentation quality and in vitro digestibility of tropical grass silage prepared with two cellu lases at different concentration.

Forage sample and silage preparation
Purple Guinea grass (Panicum maximum cv. TD 58) and Napier (Pennisetum purpureum×Pennisetum americanum cv. Pak Chong 1) grass were grown in a Korat soil series (Oxic Paleustults) at the experimental farm, Faculty of Agriculture, Khon Kaen University, Thailand in May 2014. Before planting, the plot for Guinea grass was ploughed twice and harrowed once. Root stock of purple Guinea grass was planted by hand with a spacing of 75×75 cm in row and between rows in a 10× 30 m of the rectangular plot. The plot was fertilized with cattle manure at a rate of 24,000 kg/ha which 4 separate equal por tions of 6,000 kg/ha split applied for 4 cuts [16,17]. For Napier grass, one week before planting, the plot was ploughed and fertilized with basal fertilizers of NPK (15-15-15) and cattle manure at 300 and 12,500 kg/ha, respectively. The 11,111 stem cuttings of Napier grass were planted by hand with a spacing of 120×75 cm in row and between rows in a total area of 1,600 m 2 . Two weeks after planting, the plot was weeded out and nitrogen fertilizer (urea) applied at a rate of 60 kg/ha [4]. On 20 April 2015, guinea grass was cut to adjust the height to 10 cm above ground level and Napier grass was cut close to the soil surface. After 60 d of regrowth, both grasses were harvested in early morning and immediately chopped using a forage chopper (Supachai, Kanchanaburi, Thailand) to 1 cm particle size. Half of each chopped grass sample was wilted for 6 h in the shade to study the effect of moisture adjustment. Both fresh or wilted forages were then treated as follows: control (untreated), Acremonium cellulase (AC, Meiji Seika Pharma Co., Ltd, Tokyo, Japan): AC 0.0025%, AC 0.005%, AC 0.01%, Trichoderma cellulase (TC, Meiji Seika Pharma Co., Ltd, Japan): TC 0.0025%, TC 0.005%, and TC 0.01%. The pro duction strain, main composition and carboxymethylcellulase (CMCase) of AC were Acremonium cellulolyticus, glucanase and pectinase, and 7,350 U/g; for TC they were Tricoderma viride, xylanase and glucanase, and 2,720 U/g, respectively. Treatments were ensiled using a small scale fermentation sys tem [18]. Both AC and TC cellulases were added as % of fresh matter (FM). The design of this experiment was a 2 (two grasses)×2 (fresh vs wilted)×2 (two enzymes)×3 (three appli cation levels) factorial arrangement in a completely randomized design with triplicates for each treatment. The silage was pre pared with a laboratoryscale fermentation method [16]. The mixed grass with additive (1,000 g) was packed into a synthetic silo laminated from nylon and polyethylene (Hiryu KN type, Asahikasei, Tokyo, Japan), and sealed by using a vacuum sealer (SQ-303, Asahi Kasei Pax Corp., Japan). All silos were stored at room temperature (25°C to 37°C). On 30 d of ensiling, silos were opened for evaluation of fermentation products, chemical composition, and in vitro digestibility.

Microorganism analysis
Microorganism numbers were analyzed using the plate count method as described by [19], and reported as colonyforming unit (cfu)/g of FM. Fresh or wilted forage samples (1,000 g) or silage samples (200 g) were subsampled for 10 g samples. Then, each sample (10 g) was homogenized with 90 mL steril ized distilled water, and serial dilutions in 0.85% sodium chloride solution at 10 -1 to 10 -5 . Each dilution (20 μL) was dropped and spread on prepared agar plates. MRS agar (Difco Lab¬oratories, Detroit, MI, USA) was used for the LAB count, which the agar plates were put into an anaerobic box (TE-HER Hard Anaerobox, ANX-1; Hirosawa Ltd., Tokyo, Japan), and incubation at 30°C for 48 h in an incubator. Blue light broth agar (Nissuiseiyaku Ltd., Tokyo, Japan) was used for coliform bacteria counts after incubating at 30°C for 48 h. Nutrient agar (Difco, USA) and potato dextrose agar (Nissuiseiyaku, Japan) were used for aerobic bacteria, yeast, and mold. Mold was counted after 2 d of incubation. Yeasts were distinguished from molds or bacteria by colony appearance and cell mor phology observation.

Chemical composition
Cold water extraction of silage samples was used for fermen tation endproduct analysis following the method of Cai [20]. Ten grams of wet silage was soaked with 90 mL sterilized distilled water, and then, incubated at 4°C in a refrigerator overnight. Silage pH was measured immediately after in cubating using a glass electrode pH meter (FiveGo; Mettler Toledo, Greifensee, Switzerland), and 10 mL of cold water extract was sampled and stored at -20°C for further analysis of organic acids and ammonia nitrogen concentrations. Orga nic acid content of silages was measured by HPLC methods [20]. Ammonia nitrogen concentration was analyzed using www.ajas.info 1915 Khota et al (2018) Asian-Australas J Anim Sci 31:1913Sci 31: -1922 a spectrophotometer (UV/VIS Spectrometer, PG Instruments Ltd., London, UK) [21]. The lactate buffering capacity (LBC) of forages was analyzed by the titration method viz. 10 g of sample homogenized with distilled water, titrated with 0.1 M hydrochloric acid (HCl) to decrease pH from initial pH to 3, then titrated to pH 6 with 0.1 M sodium hydroxide (NaOH) [13]. The WSC content of forage was extracted and measured following the methods of [16].
Forage and silage samples on 30 d of ensiling were dried at 60°C for 48 h in a hot air oven, and ground through a 1 mm mesh screen. Dry matter (DM), organic matter (OM), crude protein (CP), and ether extract (EE) were analyzed following methods 934.01, 942.05, 976.05, and 920.39, respectively [22]. Neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) were analyzed using a fiber ana lyzer (ANKOM 200, ANKOM Technology, New York, USA), based on the method of [23].

Animal care
The animal experiment was approved by the Animal Care and Use Committee of Khon Kaen University (KKU), Khon Kaen, Thailand and performed at KKU in August 2015. The experiment was performed according to recommendations proposed by the European Commission [24] and to minimize the suffering of animals.

In vitro digestibility analysis
In vitro digestibility of DM (IVDMD) and OM (IVOMD) were measured after incubating samples in buffered rumen fluid at 6 and 48 h. The buffered rumen inoculum was prepared following the method of [25]. Rumen fluid was collected from two dairy steers by a stomach tube sucker before morning feeding. The pH of rumen fluid was 7.2. Ground silage samples (0.5 g) was put into serum bottles of 50 mL capacity (3 repli cations per sample). The bottles were closed by a rubber stopper with an aluminum seal cap. The buffered rumen inoculum (40 mL) was injected into each sample bottle using a 60 mL syringe (Nipro Thailand corporation, Ltd., Phra Nakhon Si Ayutthaya, Thailand) with 18 gauge×3.5 cm needle (Nipro Corporation, Osaka, Japan), and flushed with carbon diox ide gas to produce an anaerobic condition. All samples were incubated in a water bath at 39°C, swirled by hand and the head space gas production released using a 21 gauge×3.5 cm needle (Nipro, Japan) at 2 h intervals. Two blanks of 40 mL of rumen inoculum per bottle were also incubated.

Microorganism population and chemical composition of forages
The microorganism counts of forages are shown in Table 1. Fresh Guinea grass and fresh Napier grass showed similar numbers of LAB (10 3 cfu/g of FM), aerobic bacteria (10 7 cfu/g of FM), and coliform bacteria (10 7 cfu/g of FM). The counts of yeasts and molds of both fresh grasses ranged from 10 4 to 10 5 and 10 2 to 10 3 cfu/g of FM, respectively. After wilting process, the counts of LAB, yeasts and molds of both grasses increased from 10 4 to 10 5 , 10 5 to 10 6 , and 10 4 cfu/g of FM, respectively, the counts of coliform bacteria and aerobic bacteria did not change by the wilting process. The chemical composition and LBC of Guinea grass and Napier grass are shown in Table 2. The OM content was higher (p<0.01) and WSC content was lower (p<0.01) in Napier grasses than those in Guinea grasses. During the wilting process, the DM of Guinea grass and Napier grass significantly (p<0.01) increased by 7% and 8%, respectively, but the contents of ADL and LBC contents did not differ. Compared with fresh Guinea grass, the contents of CP and ADF were higher (p<0.01), but the NDF content was lower in wilting Guinea grass. Also, wilted Napier grass showed higher (p<0.01) CP content and lower EE content than fresh Napier grass.

Fermentation quality of silages
The DM, pH, and fermentation products of Guinea grass and Napier grass silages at 30 d of ensiling are shown in Table 3. The interaction between A, B, C, and D influenced (p<0.01) the silage DM, pH, all organic acid contents, and ammonia nitrogen content. Guinea grass silage treated with AC 0.01% showed good quality with lower (p<0.01) pH and ammonia nitrogen, and higher (p<0.01) lactic acid content than other silages. Regarding the differences in production strain (A. cellulolyticus vs T. viride), main composition (glucanase and pectinase, vs xylanase and glucanase), and CMCase (7,350 vs 2,720 U/g) between AC and TC, the highest (p<0.01) pH and ammonia nitrogen content with lowest (p<0.01) lactic acid content were found in wilted Guinea grass silages treated with TC 0.05%. The AC treatments improved grass silage quality more than TC treatments. The addition level of AC at 0.01% improved fermentation quality more than 0.005% and 0.0025%.

Microorganism counts of silages
The microbiological analysis of Guinea grass and Napier grass silages at 30 d of ensiling are shown in Table 4. The A, B, C, D and their interaction (A×B×C×D) were not significantly dif ferent (p<0.05) in LAB count, aerobic bacteria and yeasts. Coliform bacteria and molds were below the detectable level (10 2 cfu/g of FM) in all silages. The LAB population of all silages was dominant ranging from 10 7 to 10 8 cfu/g of FM. Aerobic bacteria were 10 3 to 10 4 cfu/g of FM and yeast were 10 2 to10 4 cfu/g of FM.

Chemical composition of silages
The chemical compositions of Guinea grass and Napier grass silages at 30 d of ensiling are shown in Table 5. The significances of A×B×C×D (p<0.001) influenced all chemical compositions. The OM content of AC or TC treated Guinea grass silage were higher (p<0.01) than other silages. The wilted Napier grass silages treated with AC 0.01% showed higher (p<0.01) CP con tent and lower (p<0.01) NDF and ADF content than other silages. The highest (p<0.01) fiber contents were observed in TC treated fresh Guinea grass silages. When silages treated with AC, the CP, EE, and ADL contents were significantly (p< 0.01) higher and the NDF and ADF contents were significantly (p<0.01) lower than TC treatments; and the application levels of enzymes at 0.01% decreased fiber content more than 0.005% and 0.0025%.

In vitro digestibility of silages
The in vitro digestibility of Guinea grass and Napier grass silages at 30 d of ensiling are shown in Table 6. Fresh Napier grass silages treated with AC 0.01% had higher (p<0.05) IVDMD than other treatments. Napier grass silages had higher (p<0.01) IVDMD and IVOMD than Guinea grass silages. The wilted grasses silages had higher (p<0.05) IVOMD than fresh grasses silages. When silages treated with AC, the IVDMD and IVOMD at 6 h after incubation were significantly higher (p<0.01) than those of TC treatments. The IVDMD and IVOMD of 0.01% cellulase treated silage at 6 and 48 h after incubation were greater (p<0.05) than those of 0.0025%. The significances of A×B×C×D (p<0.05) influenced IVDMD at 48 h after incu bation in buffered rumen fluid, while IVDMD at 6 h after incubation and IVOMD at both 6 and 48 h after incubation did not.

DISCUSSION
Generally, the DM content of forages influences fermentation quality of the silage; and optimal DM content ranges from 30% to 40% for good quality silage making [13,27]. If dry matter content is less than 20%, the fermentation process would be dominated by clostridium, resulting in low quality silage pro duction [28]. In this study, the DM content of fresh Guinea grass and Napier grass were lower than 18%, which was not ideal to preserve the forage. When both grasses were wilted   Khota et al (2018) Asian-Australas J Anim Sci 31:1913Sci 31: -1922 for 6 h in the field, DM contents of Guinea grass and Napier grass were higher than 21%, much more suitable for ensiling. The decreasing of moisture during the wilting process under this experiment is negligible and less than those of the optimal DM content ranges. In addition, the wilted both grasses as shown Table 2 had a higher CP and lower ADF compared with that of fresh forage. However, we cannot fully explain the mech anism of these effects, this may be influenced by the climatic conditions such as low temperature and high humidity during the wilting process. Future study is necessary to study the effect of climatic condition on the chemical composition of forage during wilting process. The CP, NDF, and ADF of Guinea grass and Napier grass used in this study were similar to our previous work which reported that Guinea grass and Napier grass harvested at 60 d of regrowth were lower in CP contents, but high in fiber [16]. The WSC in the ensiling material also plays an important role in silage fermentation and it is the main energy source for LAB growth to produce lactic acid [29,30]. Wilkinson [27] suggested that WSC needs to be greater than 10% of DM for good fermentation. However, the WSC contents of fresh for ages in this present study were less than 2.7% of DM (Table 2). Although a lower WSC content was found in Guinea grass compared to Napier grass, it would be difficult to make good quality silage from both grasses [5].
At 30 d of fermentation, Guinea grass silages were poor quality. As shown in Tables 1, 2, both fresh and wilted Guinea grasses had low LAB counts and WSC content which were not sufficient to produce more lactic acid for good fermen tation. Thus, the pH of Guinea grass silages did not decrease below 4.2, which cannot inhibit the growth of harmful bacteria, especially clostridia [13]. As a result, a harmful fermentation could occur with clostridia using WSC, amino acid, and other organic acids to produce butyric acid and ammonia nitrogen.
Some protein of forages breaks down to simpler or non protein nitrogen during wilting process, result in decreased true protein solubility [31]. Nevertheless, this is consistent with previous studies [32] that total protein content of intact as paragus spear and excised tip sections increased approximately 10% after 6 to 12 h of harvested. Also, our results indicated that wilted forages silages were lower (p<0.01) lactic acid content than fresh forage silages. Consistently, our previous study [16] evidenced a lower lactic acid concentration in silages prepared from wilted forages compared to fresh forages. The significantly higher DM in wilted forages silages may attribute to decrease LAB growth and activity during silage fermentation.
When grass silages treated with AC 0.01%, their fermenta tion quality were more improved compared to other treatments. Our findings are in agreement with [9,3335]. that the addition of cellulase enzyme resulted in a decrease of pH and increased lactic acid content in sorghum straw, oil palm frond and mixed hullessbarley straw with corn silages. This could be attributed to the cellulase enzyme degrading plant cellulose so increas ing the WSC, an essential substrate for LAB growth and more lactic acid production. As a result, the pH of silages decrease rapidly inhibiting the growth of clostridia and preventing pro teolysis in the silo [33,36]. Some studies [15,29,34,37] reported that addition of cellulase resulted in a decrease in fiber con tent of mixed silages of hullessbarley straw and corn, wheat straw, Leymus chinensis, and Napier grass silage. On the oth er hand, cellulase addition had no effect on NDF and ADF content of barley and vetch silage [38]. In the present study, the AC improved silage fermentation more than TC, and the AC 0.01% cellulase effectively increased CP content and re duced NDF and ADF contents of silages. Perhaps the type, composition and enzyme activity of cellulase may affect cel lulose degradation and silage fermentation. Generally, cellulases catalyze the hydrolysis of cellulose, which are mainly three types: endoglucanases, cellobiohydrolases and βglucosidases, and it is used for any occurring mixture or complex of various such enzymes, that act serially or synergistically to decompose cellulosic material. Acremonium cellulolyticusproducing cel lulases contain a strong glucanase and pectinase, and cellulase produced by Trichoderma viride contains mainly xylanase and glucanase. Therefore, Acremonium cellulolyticusproducing cellulase is said to be more potent than Trichoderma viride producing cellulase. This indicated that AC is more effective for improvement of silage fermentation than TC, and its ad dition concentration is necessary to be 0.01% on a FM basis. The digestibility of roughage also plays an important role in animal production [10,29]. The IVDMD and IVOMD de pend on physical characteristics of forage, especially the fiber content, low NDF and ADF contents resulted in a rapid in crease in digestibility of DM and OM [29,39]. In the present study, Napier grass silages had higher IVDMD and IVOMD when compared to Guinea grass. This could be attributed to the differences in chemical composition of Guinea grass and Napier grass. Guinea grass had higher fiber content than Napier grass as shown in Table 2. Furthermore, the cellulase enzyme addition, especially AC 0.01% treated silages had greater IVDMD and IVOMD values than other enzyme application levels in both 6 and 48 h of incubation (Table 6). This study is consistent with previous studies reported by [10,40] that the addition of an exogenous cellulase enzyme could increase IVDMD of forage crops, but inconsistent with [29] that cellu lase addition did not increase the IVDMD of Leymus chinensis after 48 h of incubation. The greater effect of AC 0.01% treat ment in the present study due to the composition of enzymes. The AC used in this study has both glucanase and pectinase; the synergistic effects of these enzymes probably increase silage digestibility. The CMCase activity is also higher in AC (7,350 U/g) than TC (2,720U/g).
These results confirmed that Napier grass has more suitable ensiling characteristics than Guinea grass, and the wilted silage can decrease the feed nutrient loss and improve the in vitro digestibility. The addition of AC 0.01% cellulase (73.5 U/g of FM) was effective in improving the silage fermentation and promoting the in vitro degradation of tropical grass.

CONCLUSION
Silage fermentation and in vitro digestibility of Guinea grass and Napier grass prepared with cellulase enzyme at different addition concentration were studied. The AC has the potential to improve the fermentation quality, chemical composition and in vitro degradation of Guinea grass and Napier grass, but the TC did not affect the tropical silage fermentation. The AC addition at 0.01% (73.5 U/g) of FM was the most promising for improving silage fermentation containing low WSC.

CONFLICT OF INTEREST
We certify that there is no conflict of interest with any financial organization regarding the material discussed in the manu script.