Effects of inoculant application on fermentation quality and rumen digestibility of high moisture sorghum-sudangrass silage

ABSTRACT This study estimated the effects of a new inoculant producing antifungal and esterase activity on quality of high moisture sorghum-sudangrass (SS) silage with two different hybrids (SX-17 and Speed-up). The SS hybrids were chopped, treated without an inoculant (CON) and with an inoculant containing Lactobacillus plantarum R48-27 and Lactobacillus buchineri R4-26 at ratio 1:1 (INO), then ensiled into 20-L mini silo in quadruplicate for 60 days. After ensiling, silage was placed under aerobic condition for 8 days to estimate yeast and mold. The INO silages had higher (P < 0.05) dry matter, crude protein, neutral detergent fibre, and acid detergent fibre than those of CON silages. The INO silages also had higher (P < 0.05) pH and acetate, but lower (P < 0.05) ammonia-N, butyrate, and lactate to acetate ratio than those of CON silages. Applied INO in both hybrids had lower (P < 0.05) yeast after 4–8 days of aerobic exposure than CON. In rumen, INO silages had higher (P < 0.05) in vitro dry matter digestibility, pH, ammonia-N, and acetate than those of CON silages. In conclusion, the new inoculant application improved not only fermentation quality, but also rumen digestibility of high moisture SS silage.


Introduction
In the warm season, annual grass such as sorghum-sudangrass (SS; Sorghum bicolor x S. bicolor var. sudanense) is one of the potential forages that could be applied in year-round crop systems. The SS has greater biomass production with high tolerance to environmental stress, especially in high temperature during the summer season (Venuto and Kindinger 2008). It could be an alternative forage for corn silage, which provides a suitable nutrient for dairy cows (Dann et al. 2008;Gurbuz et al. 2008). However, SS forage usually wilts long before reaching the ideal moisture content due to their stalk, which could reduce the silage quality (McDonald et al. 1991;Lim et al. 2009). Our preliminary study (Data unpublished) also had shown that SS forages still contained high moisture (>75%) after 3 days of the wilting process in the field. Generally, ensiling forage with high moisture content usually produces high butyrate concentration, which is associated with silage quality and aerobic deterioration (McDonald et al. 1991;Danner et al. 2003;Vissers et al. 2007). Microbial inoculant application on forages will stimulate organic acid production, followed by decrease of pH that can inhibit undesirable microbes and reduce nutrient loss during ensiling (McDonald et al. 1991). Moreover, Weinberg et al. (2007) reported that the digestibility of wheat and corn silages improved by applications of microbial inoculants. And, it had been reported that the applications of combo inoculants containing homo and heterofermentative lactic acid bacteria (LAB) could improve the quality of silage as well as aerobic stability (Filya 2003;Huisden et al. 2009;Queiroz et al. 2012;Joo et al. 2018). Several studies have shown that inoculants had various effects on silages by forage hybrids (Andrae et al. 2001;Beck et al. 2007;Kim et al. 2018). The SX-17 and Speed-up hybrids reported that it had high growth rate and dry matter (DM) yield (Zahid et al. 2002).
In our previous study, Lactobacillus plantarum R48-27 and Lactobacillus buchneri R4-26 were isolated from rye silage and confirmed by plate assay to produce the fibrinolytic enzymes and antifungal substances, respectively ). However, the effects of these new inoculants on aerobic deterioration and nutrient digestibility of SS silage did not confirm yet. The application of these inoculants on high moisture SS silage possibly inhibits the growth of undesirable microbes by antifungal substances as well as the improvement of digestibility by fibrinolytic enzymes. Therefore, this study estimated the effects of new inoculants on fermentation quality, aerobic deterioration, and rumen digestibility of high moisture SS silages.

Materials and methods
2.1. Silage production and sampling harvested at the late heading stage from four plots, respectively, and then wilted for 24 h. The mean DM concentration of both hybrids was 23.1% at the end of wilting. The wilted SS forage chopped to 3-5 cm length separately and ensiled into 20 L bucket silo (4 kg) in quadruplicate for 60 days, following: silages applied 1% distilled water in fresh forage (CON); and silages applied combo inoculant containing Lactobacillus plantarum R48-27 and Lactobacillus buchneri R4-26 at 2 × 10 5 of fresh forage at a ratio of 1:1 (INO). The SS forage (500 g) and silage (1 kg) were sub-sampled for chemical compositions and in vitro rumen digestibility. Additionally, silages were also subsampled for fermentation indices (20 g) and aerobic deterioration (1 kg). The sub-sampled silage for aerobic deterioration was located in a polystyrene box (25 cm diameter × 45 cm depth) and covered with two layers of cheesecloth for 8 days at room temperature (10°C) as described by Kang et al. (2009). On 1, 2, 3, 4, 6, and 8 days of aerobic exposure, 50 g of silages were sub-sampled to measure the changes of yeast and mold.

Chemical composition
The sub-sampled (500 g) forage and silage were dried at 65°C for 48 h and ground to pass 1-mm screen using a cutting mill (Shinmyung Electric Co., Ltd, Gimpo, South Korea). The DM concentration was analyzed using a forced-air dry oven at 105°C for 24 h. Crude ash (CA) was determined with a muffle furnace at 550°C for 5 h. Crude protein (CP) and ether extract (EE) were measured by the producers of Kjeldahl (method number 984.13; AOAC 1995) and the Soxhlet (method number 920.39; AOAC 1995), respectively. Neutral detergent fibre (NDF) and acid detergent fibre (ADF) were determined using an Ankom 200 fibre analyzer (Ankom Technology, Macedon, NY, USA) following the method of Van Soest et al. (1991).

Fermentation indices
Twenty grams of silage were blended with 200 mL of sterile ultrapure water for 30 sec and filtered through two layers of cheesecloth for silage extraction. The analyses of pH, ammonia-N, lactate, volatile fatty acid (VFA), and microbial counts used silage extraction. The pH was measured by a pH metre (SevenEasy, Mettler Toledo, Greifensee, Switzerland). Ammonia-N was determined using a colorimetric method described by Chaney and Marbach (1962). The silage extraction was centrifuged at 5,645 × g for 15 min and collected the supernatant for lactate and VFA analyses. The concentrations of lactate and VFA were determined using HPLC (L-2200, Hitachi, Tokyo, Japan) fitted with a UV detector (L-2400, Hitachi, Tokyo, Japan) and a column (Metacarb 87H; Varian, Palo Alto, CA, USA) described by Muck and Dickerson (1988).

Microbial counts
Microbial counts were determined using silage extractions (first dilution), which were continued into several dilutions (10 −5 -10 −7 ). The injection of silage extraction was in triplicate in selective agar medium as described by Kim et al. (2017) and Joo et al. (2018). The LAB count used lactobacilli MRS agar media (MRS; Difco, Detroit, MI, USA) and yeast and mold counts used potato dextrose agar (PDA; Difco, Detroit, MI, USA). The MRS agar plates were put on a CO 2 incubator (Thermo Scientific, Waltham, MA, USA) at 30°C for 24 h, while PDA plates were put on an aerobic incubator at 30°C for 72 h in an aerobic incubator (Johnsam Corp., Boocheon, South Korea). Visible colonies from the plates were calculated and the number of colonies forming units (cfu) was per gram of silage. The microbiological data was transformed to log 10 .

In vitro rumen digestibility
Animal Care and Ethics Committee of Gyeongsang National University, Jinju city, Gyeongsangnam-do, South Korea, approved the animal procedure for the cannulated cow in the present study. Rumen fluid collected before morning feeding from two non-pregnant cannulated Hanwoo heifers, which had been fed rice straw hay and grains mixed at 8:2 ratio. The collected rumen fluid was composited and then filtered via two layers of cheesecloth. In vitro medium was prepared by mixing rumen fluid with Van soest medium at 1:2 ratio. A ground silage sample (0.5 g) and in vitro medium (40 mL) were placed into the incubation bottle at triplicate with two blanks. Then, the incubation bottle was filled with CO 2 to reach anaerobic condition (Tilley and Terry 1963). All incubation bottles were placed into an anaerobic incubator (Thermo Scientific, Waltham, MA, USA) at 39°C for 72 h. After incubation, the bottle was opened and transferred to 50 mL conical tube to separate the residue and supernatant (in vitro medium) through centrifugation at 2,568 × g for 15 min (Supra 21k, Hanil Electric Corporation, Gimpo, South Korea, with rotor A50S-6C No.6). The residue was used to measure in vitro digestibility of DM (IVDMD) and NDF (IVNDFD), while the supernatant was used to measure the rumen fermentation characteristics such as pH, ammonia-N, and VFA using the same protocol as described before.

Statistical analysis
The experiment was conducted by a 2 (hybrid; SX vs. SU) × 2 (inoculant; CON vs. INO) factorial design with four replicates per treatment. All data was analyzed using general linear model (GLM) procedure of Statistical Analysis System (SAS), version 9.3 package programme to test the effects of hybrid, inoculant, and its interaction (hybrid × inoculant). The model was Y ijk = µ + α i + β j + (αβ) ij + e ijk , where Y ijk = response variable, µ = overall mean, α i = the effect of hybrid treatment, β j = the effect of inoculant treatment, (αβ) ij = the interaction effect of hybrid and inoculant, e ijk = error term. The significant differences were declared at P < 0.05.
Application of INO in SS silage resulted in higher yeast count (P = 0.015) than CON only on day 1 of aerobic exposure (Table 4). However on 4, 6, and 8 days of aerobic exposure, both hybrids inoculated with INO had lower (P < 0.05) yeast count than CON. Yeast count was not affected by hybrid during aerobic exposure, except on 8 days, which SX hybrid was higher (P = 0.039; 7.96 vs. 7.93 log 10 cfu/g) than SU hybrid. Also an interaction effect between hybrid and inoculant was detected (P = 0.010) on yeast count on day 8 of aerobic exposure, while yeast count only decreased in SU silage by inoculant application.

Discussion
The chemical compositions of SX and SU forages were in the expected range (Beck et al. 2007;Dann et al. 2008). After 60 days of ensiling, SU silage presented higher concentrations of DM, EE, CA, and NDF than SX silage. Generally, these results could occur by chemical compositions of those forages (Johnson et al. 2002;Kim et al. 2018). The higher DM and CP concentrations by INO application might indicate a lower nutrient loss during ensiling. The previous study also reported that an applied combo inoculant consisting of Lactobacillus plantarum and Lactobacillus buchneri produced higher DM concentration than untreated silage Filya 2003). In addition, the application of inoculant could reduce proteolysis caused by rapid acidification during ensiling (Winters et al. 2000). These results could support higher CP concentration of INO silage than CON silage in the present study. Huisden et al. (2009) reported that inoculant application led to increases of NDF concentration due to the higher degradation of nonstructural carbohydrates. Similarly in the present study, NDF and ADF concentrations were slightly higher in INO silage than in CON silage. Additionally, this result also might be supported partially by higher microbial counts in INO silage than in CON silage (Table 3).
In general, the forages containing high CP concentration produce a high ammonia-N concentration in the silages by proteolysis during the ensiling period (McDonald et al. 1991;Kim et al. 2018). The SX forage had a slightly higher CP concentration (11.8 vs. 11.2%) than SU forage. This might have influenced higher ammonia-N concentration in SX silage than in SU silage, which also could lead to higher pH in SX silage.
Inoculant applications on silages lead to lower pH and ammonia-N concentration due to the stimulation of organic acid, which also could inhibit the growth of undesirable microbes (McDonald et al. 1991;Filya 2003). However, application of heterofermentative LAB had shown an increase of pH due to the conversion of lactate into acetate and propionate Danner et al. 2003;Joo et al. 2018). In addition, it could increase yeast count without a negative effect on aerobic stability of silage (Oliveira et al. 2017). The results of fermentation indices in the present study were in agreement with those previous studies, which were higher in pH, acetate concentration, and yeast count, and lower in ammonia-N, lactate and butyrate concentrations by new inoculant application. Additionally, the effects of new inoculants in the present study also supported our previous study that had demonstrated fibrinolytic and antifungal effects . As an interaction effect between hybrid and inoculant, application of INO increased acetate concentration greater in SU silage than in SX silage. The previous studies have also reported an interaction effect between hybrid and inoculant on acetate concentration (Kang et al. 2009;Kim et al. 2018). The presence of butyrate in the present study might be due to the high moisture of SS forage (McDonald et al. 1991). However, the new inoculant application in the present study proved a decrease of butyrate concentration, which is an indicator of silage fermentation quality and feed value for ruminant (McDonald et al. 1991;Danner et al. 2003;Vissers et al. 2007;Nkosi and Meeske 2010).
Microbial counts of silage during aerobic exposure are the indicators to evaluate aerobic stability (McDonald et al. 1991;Nkosi and Meeske 2010;Dolci et al. 2011). Yeast is especially an initiator of aerobic deterioration in silage, followed by mold growth (Danner et al. 2003;Dolci et al. 2011). Application of heterofermentative LAB had reported inhibiting the growth of yeast and mold in feedout phase due to the presence of antifungal substances such as acetate, propionate and so on (Danner et al. 2003;Kim et al. 2017). The previous studies also had reported that application of combo inoculant containing homofermentative LAB and Lactobacillus buchneri improved aerobic stability of silages (Filya 2003;Huisden et al. 2009;Queiroz et al. 2012;Joo et al. 2018). Although yeast count (7.93 vs. 7.54 log 10 cfu/g) of INO silage was higher than CON silage on 0 d of aerobic exposure in the present study, the opposite was observed (7.79 vs. 7.91 log 10 cfu/g) from 4 days of aerobic exposure. Moreover, the increase of yeast count from 1 to 8 days of aerobic exposure was higher in CON silage than in INO silage (0.46 vs. 0.18 log 10 cfu/g), which indicated CON silage spoiled rapidly compared to INO silage. This result occurred by higher acetate concentration of INO silage in the present study (Table 3). The result of aerobic deterioration in the present study also supported the previous study that reported improvement on aerobic stability by application of antifungal inoculant (Kleinschmit et al. 2005). The effects of Lactobacillus buchneri R4-26 on aerobic deterioration of SS silage in the present study also   confirmed our previous study that had demonstrated antifungal effects by plate assay. Moreover, its antifungal activity was reported to inhibit the growth of mycotoxins fungal such as Fusarium moniliforme and Aspergillus parasiticus in our previous study .
Application of fibrinolytic inoculant had been reported to improve apparent total tract digestibility of silage DM and CP that might increase concentrations of VFA and ammonia-N in the rumen, respectively (Kamalak et al. 2002;Kang et al. 2009;Arriola et al. 2011;Romero et al. 2015). In addition, Lactobacillus plantarum R48-27 as silage inoculant in the present study was reported to release cellulose, xynalase, chitinase, and esterase enzymes ). An increase of rumen acetate concentration indicates an improvement of structural carbohydrate degradation (Hobson and Stewart 1997;Sutton et al. 2003), which can be caused by an excreted fibrinolytic enzyme from microbial inoculants. The results of rumen digestibility and fermentation characteristics in the present study agreed with those previous studies, where INO silage had higher rumen pH, IVDMD, and concentrations of ammonia-N and acetate than CON silage. In addition, higher acetate with lower propionate proportions in the rumen of INO silage compared to CON silage reflects the fibrinolytic activity of new inoculant. The results of this study was in agreement with previous study that also reported improvement of corn silage digestibility by fibrinolytic inoculant (Kang et al. 2009) or by fibrinolytic enzymes (Arriola et al. 2011;Romero et al. 2015).

Conclusion
The present study concluded that applied SU silage presented higher rumen digestibility than SX silage. Applications of the new inoculant in high moisture sorghum-sudangrass might help to improve not only the silage quality by inhibitions of undesirable substances (butyrate and yeast), but also rumen digestibility.

Disclosure statement
No potential conflict of interest was reported by the authors.