Effects of Lactobacillus buchneri and Lactobacillus rhamnosus on Ryegrass Silage Fermentation and Aerobic Stability

Abstract

Italian ryegrass is a high-quality forage grass, and a full understanding of the changes in its microbiome and metabolome during aerobic exposure can prolong its aerobic stability and improve its utilization value. Italian ryegrass silage was prepared with deionized water (CK), Lactobacillus rhamnosus BDy3-10 (LR), Lactobacillus buchneri TSy1-3 (LB), and a mixture of these two lactic acid bacteria (M). The silage was maintained at ambient temperature for 60 days followed by aerobic exposure. The results show that the Italian ryegrass silage in the LB and M groups exhibited aerobic stability for up to 19 days. A total of 1881 chemicals were identified in Italian ryegrass silage. These metabolites are associated with bacterial communities, especially Lactobacillus. The addition of lactic acid bacteria resulted in a common differential metabolic pathway compared to CK: “phenylpropanoid biosynthesis”. “Flavone and flavonol biosynthesis” was the significant differential metabolic pathway between LB and LR. Inoculation with LB significantly increased the concentrations of lactic acid, acetic acid, vitexin, and luteolin. In conclusion, lactic acid bacteria (LAB) additives affect the microbial community and metabolites of silage. The application of LB inoculants is a feasible way to obtain well-fermented Italian ryegrass silage and improve aerobic stability, even at higher moisture content levels.

1. Introduction

Italian ryegrass (Lolium multiflorum Lamk.) is a major grass cultivated worldwide, with a fast growth rate, high yield, high nutritional value, and good palatability. It is one of the common forage varieties used in livestock production [1]. Silage is one of the most common methods of preserving roughage in ruminant feeding systems, extending the shelf life and retaining the nutritional value of the feed [2]. Silage is a common way to use it. The ensiled crop is subjected to four phases [3]. Much research effort has been placed on understanding the second and third phases, which are the fermentation phase and the stable storage phase, especially by adding fermentation inhibitors or enhancers at the time of ensiling to improve the efficiency of the second stage [4,5,6,7]. It is now understood that, from the perspective of retaining nutrients and upholding high sanitary quality until animals consume the silage, the changes that may take place during the fourth or feed-out period are just as significant as those in the first two stages [8]. Aerobic bacteria, yeasts, and molds oxidize fermentation acids and other substrates when silage is exposed to air when the silo is opened or after it has been removed [9]. The aerobic stability of silage is a key factor in ensuring that silage provides well-preserved nutrients to animals with minimal amounts of mold spores and toxins. Mold contamination in silage is associated with reduced palatability, reduced nutritional value and feed intake, animal health problems, reduced productivity and fertility, and increased disease susceptibility [10,11,12]. Mycotoxins are a unique group of secondary metabolites produced by certain fungi, including Aspergillus, Penicillium, Fusarium, Stachyobotris, and Cephalosprium [11,13].

As the freshwater content of ryegrass is up to 80% or more when mowed, it is susceptible to butyric acid fermentation if its silage is untreated [4,14]. Generally, if the raw material is silaged directly, the harmful phyllosphere microorganisms of the plant will hydrolyze the proteins during the fermentation process, resulting in a large loss of dry matter and thus affecting the quality of silage fermentation [15]. Therefore, it has become a conventional silage method to select appropriate additives to promote the silage fermentation process or improve the aerobic stability of silage during processing and preparation. In addition to promoting silage fermentation to improve feed quality, heterotypic LAB produce antimicrobial agents, bacteriocins, while converting lactic acid to acetic acid and 1,2-propanediol, which can effectively inhibit mold and delay mold spoilage time after aerobic exposure [16,17]. While homofermentative LAB are usually able to accelerate fermentation, reduce protein degradation, increase the lactic acid content, and reduce the acetic acid, butyric acid, and ethanol contents, with theoretical energy and dry matter recoveries of 100% and 99%, respectively, their effectiveness in improving aerobic stability is unsatisfactory [18].

Since new fungal metabolites are constantly being found, it is still unknown how much mycotoxin is actually present in animal feed [19]. Therefore, it is particularly important to pay attention to the changing patterns of fungi and mycotoxins during the aerobic exposure phase [20]. The effects of LAB addition on mycotoxins in silage and the dynamics of mycotoxins during aerobic exposure are the focus of our attention. With the development of metabolomics technology, an increasing number of metabolites are being elucidated, and their importance in silage is being determined, in addition to a focus on the microbiome [21]. For many years, most studies on silage metabolites have focused on the evaluation of the fermentation quality of organic acids. The metabolites of the silage process include a wide range of species, and many more metabolites remain unexplored.

As mentioned above, the purpose of this study was to characterize the bacterial and fungal core microbiota and metabolome associated with Italian ryegrass silage supplemented with Lactobacillus buchneri (L. buchneri) and Lacticaseibacillus rhamnosus (L. rhamnosus) during ensiling to aerobic exposure. A multi-omics approach was used to investigate whether ryegrass silage under high moisture conditions inhibits the effect of LAB addition, and to study the effect of homo- and heterotypic mixed LAB on the aerobic stability of ryegrass silage. In addition, emphasis was placed on studying how the introduction of heterotypic LAB enhances the aerobic stability of silage through metabolic pathways.

2. Materials and Methods

2.1. Silage Preparation

Italian ryegrass was planted in the Dushan County Grass Seed Farm, Guizhou Province (subtropical humid monsoon climate, 107.37° E, 25.33° N). The average annual temperature is 15 °C, the highest temperature is 35.5 °C, the lowest temperature is minus 4.0 °C, the annual rainfall is 1429.9 mm, and there are 297 days without frost per year. Italian ryegrass was planted in October 2020 with a 17-17-17 (N-P-K) compound fertilizer (New Wolfeng Agricultural Technology Development Co., Liaocheng, China) and no artificial irrigation treatments were conducted. The second growth of the Italian ryegrass row was harvested with a sickle on 5 January 2021, at noon on a sunny day, at the jointing stage of regrowth. The plant material was chopped into 1–2 cm pieces by a precision chopper and immediately transported to the laboratory for further processing. The nutritional characteristics and microbial counts of ryegrass are shown in

Table 1. Nutritional and microbial characteristics of ryegrass prior to fermentation.

Item
Dry matter (g/kg FM)	177.41
Crude protein (g/kg DM)	113.77
Water-soluble carbohydrates (g/kg DM)	176.63
Neutral detergent fiber (g/kg DM)	181.33
Acid detergent fiber (g/kg DM)	53.89
Lactic acid bacteria (log10 cfu/g FM)	6.69
Yeast (log10 cfu/g FM)	6.83
Coliform bacteriai (log10 cfu/g FM)	3.00
Mold (log10 cfu/g FM)	5.47
pH	6.73

FM, fresh matter; DM, dry matter; cfu, colony-forming units.

. Four treatment groups were prepared after homogenization: (a) the control group (CK), (b) the L. rhamnosus BDy3-10 treatment group (LR), (c) the L. buchneri TSy1-3 treatment group (LB), and (d) the mixed LAB treatment group (M). The application rate of each inoculant to the fresh forage was 1 × 10⁷ cfu g⁻¹ FM, and an equal volume of distilled water was sprayed onto the fresh Italian ryegrass for the CK group. L. rhamnosus BDy3-10 and L. buchneri TSy1-3 were selected as additives in this study, which were the strains screened by our research group in Guizhou and identified as excellent [22]. There was approximately 5 kg of grass forage in each silo (10 L), with six silos per treatment. Silages were ensiled at room temperature, and all silos were opened after 60 days of ensiling. Subsamples from each group were taken into −80 °C lab freezers (Thermo Fisher Scientific, Waltham, MA, USA) for DNA extraction and detection of relevant fermentation parameters on the day of plastic bucket opening, and on the 2nd, 6th, and 9th day of aerobic exposure. Six replicates of each treatment were sampled to measure the metabolome.

2.2. Aerobic Stability

To simulate a realistic feeding environment, 60-day open silos of silage were stored at room temperature for 20 days. The real-time temperatures of the core area of silage buckets were measured every 5 min by a multiplex data logger (TP9000; TOPRIE Co., Ltd., Shenzhen, China) during this 20-day period. To assess the quality of fermentation, silages from each treatment were carefully combined and analyzed after two, six, and nine days of aerobic exposure (10 g, 3 repetitions per treatment), microbes (5 g, 3 repetitions per treatment), and metabolites (5 g, 6 repetitions per treatment). The aerobic stability was calculated based on the time at which the temperature of the silage exposed to air exceeded the ambient temperature by 2 °C.

2.3. Fermentation Analyses

Silage samples (10 g) from fresh, 60-day, and aerobic exposure for 2, 6, and 9 days were added to 90 mL of distilled water, refrigerated at 4 °C overnight, and then filtered through two layers of cheesecloth to obtain the silage extract. The pH of the extract was determined using a pH meter (PHSJ-5; LEICI, Shanghai, China), and the remaining samples were analyzed to ascertain their organic acid concentrations. Subsequently, the organic acids (lactic acid, acetic acid, propionic acid, and butyric acid) were analyzed by high-performance liquid chromatography (HPLC) with a UV detector (210 nm) and a column (TC-C18 (2) 250 × 4.6 mm 5 μm, Agilent, Santa Clara, CA, USA) [22]. To ascertain the DM and DMloss content, samples weighing roughly 100 g from each bag were dried in a forced-air oven for 48 h at 65 °C. Following a reaction with the anthrone reagent, the concentration of water-soluble carbohydrate (WSC) was determined using colorimetry. The Van Soest et al. [23] method was used to evaluate the contents of acid detergent fiber (ADF) and neutral detergent fiber (NDF). The total nitrogen (TN) content was measured via a Kjeldahl nitrogen analyzer (Kjeltec 8400 Analyzer; Foss, Skane, Sweden) and we multiplied the total nitrogen content by 6.25 to calculate the crude protein (CP) content.

2.4. Microbiome Analysis

The 16S rDNA V3-V4 region of the ribosomal RNA gene was amplified by PCR (94 °C for 2 min, followed by 30 cycles at 98 °C for 10 s, 62 °C for 30 s, and 68 °C for 30 s, with a final extension at 68 °C for 5 min). The 16S rRNA genes of distinct regions (16S V4) were amplified using the specific primers 341 F (CCTACGGGNGGCWGCAG) and 806R (GGACTACHVGGGTATCTAAT) [24]. ITS genes of distinct regions (ITS1/ITS2) were amplified using the specific primers ITS3_KYO2 (GATGAAGAACGYAGYRAA) and ITS4 (TCCTCCGCTTATTGATATGC) with barcode [25]. The purified amplicons were sequenced on the Illumina platform according to the standard procedure for double-end sequencing (PE250).

The tag assembly was conducted in accordance with the following principles: the overlap between paired-end reads was required to exceed 10 bp and the percentage of mismatches was limited to 2%. The unique tags were obtained by removing redundant tags using the MOTHUR (version.1.33.0) software [26]. The effective tags were clustered into operational taxonomic units (OTUs) of ≥97% similarity using the UPARSE (version 9.2.64) pipeline. The alpha diversity metrics, including observed species, Chao1, Shannon, Simpson, ACE, and Good-coverage, as well as the beta diversity metrics, such as weighted UniFrac and unweighted UniFrac, were calculated with the QIIME software (version 1.7.0).

The variable importance projection (VIP) values of the applied variables were ranked from largest to smallest, and the top 20 VIP values for different metabolites were selected to draw bubble plots for KEGG enrichment analysis. In the analysis of the fold-change in the relative concentration of fungal toxins: log₂ (Fold-change) = log₂ (case_mean/control_mean), with the difference multiplier taken as log₂; control_mean: the average value of the integral quantitative value of the control group; and case_mean: mean value of the integrated quantitative value of the treatment group.

2.5. Metabolite Analysis

For the purpose of silage extraction, 100 mg of silage fodder samples from silages that had been exposed to air for either 0 or 9 days were added to 360 µL of pre-chilled methanol and 40 µL of the internal standard, L-2-chlorophenylalanine, at a concentration of 0.3 mg/mL in methanol [27]. The mixture was subjected to homogenization, ultrasound treatment, and centrifugation at 15,000 rpm at 4 °C for 20 min. Thereafter, 400 µL of the resulting supernatant was added to 80 µL of methoxyamine hydrochloride (15 mg/mL in pyridine) for oximation. Derivatization was then performed by adding 80 µL of BSTFA and 20 µL of n-hexane to the oximation product, which was subsequently heated to 70 °C for 60 min and shaken for 2 min. Subsequently, the sample was incubated at room temperature for a period of 30 min. A gas chromatography-mass spectrometry (GC-MS) analysis was conducted utilizing an Agilent 7890A/5975C GC-MS instrument (Agilent, Santa Clara, CA, USA). The analytical compounds (1 µL) were injected in splitless mode with an inlet temperature of 260 °C and separated with an HP-5MS (30 m × 0.25 mm × 0.25 µm) capillary column, using helium as a carrier gas at a constant flow rate (1 mL/min). The temperature of the GC column was increased from an initial temperature of 60 °C to 310 °C at a rate of 8 °C per minute, where it was held for a period of 6 min. The temperature of the ion source was 230 °C, while that of the quadrupole was 270 °C. The data were acquired using full-scan mode with an m/z range of 100–1500. The GC-MS data were converted from Chem Station analysis files (version E.02.02.1431, Agilent, Santa Clara, CA, USA) to netCDF format files and processed with Chroma TOF software (version 4.34, LECO, St. Joseph, MI, USA). The Chroma TOF software was employed for the extraction of raw peaks, filtering and calibration of data baselines, and the identification and integration of peaks. Metabolites were identified utilizing the Tracerfinder 3.2 software (Thermo Fisher Scientific, Waltham, MA, USA), employing the NIST14.0 library.

2.6. Statistical Analysis

The influence of aerobic exposure periods and inoculants on the chemical components, fermentation parameters, and microbial populations were evaluated through two-way ANOVA in SPSS 26.0 software (IBM Crop., Armonk, New York, NY, USA). The assumptions (normality, homogeneity of variances) were met using ANOVA. A p-value of less than 0.05 was regarded as statistically significant, and Duncan’s multiple range test was further applied to compare the group means, with post hoc tests adjusted accordingly.

3. Results

3.1. Nutritional Characteristics and Microbial Population of Ryegrass

When the silage is finished and the silo is opened, the nutritional qualities of Italian ryegrass are as follows (

Table 2. Nutritional qualities of Italian ryegrass silage..

		DM	DMloss	CP	WSC	ADF	NDF
		g/kg	g/kg	g/kg DM	g/kg DM	g/kg DM	g/kg DM
Treatment	CK	189.08 A	11.67 A	190.55 AB	45.31 A	85.41 A	163.87 A
	LB	154.09 B	−23.32 B	213.27 A	27.12 A	119.64 A	185.65 A
	LR	197.08 A	19.67 A	156.42 C	46.05 A	105.34 A	188.23 A
	M	165.1 AB	−12.31 AB	171.74 C	31.29 A	80.3 A	162.76 A
SEM		6.37	6.37	7.24	6.52	12.54	25.38
p value		<0.01	<0.01	<0.01	0.119	0.093	0.557

Different letters (A–C) indicate significant differences between treatments. DM, dry matter; DMloss, dry matter loss; CP, crude protein; WSC, water-soluble carbohydrates; ADF, acid detergent fiber; NDF; neutral detergent fiber. CK, untreated; LR, Lactobacillus rhamnosus BDy3-10; LB, Lactobacillus buchneri TSy1-3; M, two-LAB mixture.

). The dry matter (DM) content and dry matter loss (DMloss) in the untreated (CK) and L. rhamnosus (LR) groups were significantly higher than those in the other treatments, while the crude protein in the L. buchneri (LB) group was significantly higher than those in the other groups. There were no significant differences between the treatments with water-soluble carbohydrates (WSCs), acid detergent fiber (ADF), and neutral detergent fiber (NDF).

3.2. Aerobic Stability and Fermentation Quality

During aerobic exposure, the aerobic stability time of LB and the two-LAB mixture (M) was longer than that of LR and CK (Figure 1A). After 9 days of aerobic exposure, the temperature in the CK and LR treatment silages exceeded the ambient temperature by 2 °C. The difference between the temperature of the LR treatment group and the ambient temperature reached more than 10 °C between 12 and 15 days (reaching a maximum of 10.96 °C on 13 days). The temperature of the LB treatment group only exceeded the ambient temperature by 2 °C on day 20, demonstrating a very long aerobic stabilization period. The pH of all treatments remained essentially unchanged until 9 days of aerobic exposure, and the LR group maintained the lowest pH (approximately 4.0). From 9 days to 20 days of aerobic exposure, the pH of CK and LR significantly increased to 7.9 and 8.5, respectively. The pH of the LB group did not increase significantly until 16 days of aerobic exposure and reached 5.2 at 20 days. No significant change in pH was observed in the M group within 20 days of aerobic exposure (Figure 1B).

When the silo was opened, the LB group had significantly higher lactic acid (LA) and acetic acid (AA) than the other treatments, and the propanoic acid (PA) in the LB and M groups was significantly higher than that in the CK and LR groups, and no butyric acid (BA) was detected in the four treatments. After 2 days of aerobic exposure, LA content in the LB and LR groups was significantly higher than that in the CK and M groups (p < 0.05); after 6 and 9 days of aerobic exposure, the LA content in the CK group was the lowest, and that of the LB group was significantly higher than that of the other groups. The LA content in the CK, LB, and LR groups decreased significantly (p < 0.05) with the increase in aerobic exposure time, while there was no significant change in the M group. After 9 days of aerobic exposure, the LA content in the CK and LR groups was only 0.23 and 3.96 g/kg of dry matter (DM), respectively. Compared to the last day of ensiling, the AA concentration decreased significantly (p < 0.05) in the CK and LR groups, while it increased significantly (p < 0.05) in the LB and M groups after 9 days of aerobic exposure. PA levels were significantly lower in the CK group than in the remaining groups during the aerobic exposure period (p < 0.05), and they were highest in the LB group. As the duration of aerobic exposure increased, there was no significant difference between CK, LB, and LR in different aerobic exposure phases. In addition, significant differences were found between treatments in BA content on days 2 and 9 of aerobic exposure, while BA content on day 6 of aerobic exposure was significantly higher in the M group than in the rest of the groups (p < 0.05). No significant differences were found between CK, LB, and LR at different stages of aerobic exposure. Aerobic exposure time and different treatment interactions had significant effects on pH, LA, AA, PA, and BA (p < 0.05) (

Table 3. Changes in organic acid content (g/kg dry matter) dynamics during aerobic exposure of Italian ryegrass silage.

Item	Day (d)	Treatment				SEM	p Value
		CK	LB	LR	M		D	T	D × T
pH	60	5.05 Aa	5.07 Ba	4.53 Ab	4.93 Aa	0.050	<0.001	<0.001	<0.001
	A2	4.39 Bb	4.56 Ca	3.98 Bc	4.51 Ba
	A6	4.55 Ba	4.51 Ca	4.04 Bb	4.46 Ba
	A9	4.77 ABb	5.60 Aa	4.05 Bc	4.76 Ab
LA (g/kg DM)	60	6.30 Ac	21.35 Aa	18.53 Ab	17.32 Ab	0.343	<0.001	<0.001	<0.001
	A2	5.80 Ab	12.18 Ba	11.42 Ba	5.24 Bb
	A6	2.77 Bc	11.84 Ba	5.37 Cb	5.51 Bb
	A9	0.23 Cd	9.61 Ca	3.96 Cc	5.26 Bb
AA (g/kg DM)	60	1.77 Ab	2.62 Ba	1.79 Ab	2.09 Bb	0.081	0.001	<0.001	<0.001
	A2	2.03 Ab	2.75 Ba	1.63 Ac	1.69 Cc
	A6	0.90 Bc	3.11 Aa	1.41 ABb	1.47 Cb
	A9	0.88 Bc	3.21 Aa	1.17 Bc	2.54 Ab
PA (g/kg DM)	60	1.09 c	3.56 a	2.48 b	3.29 Aa	0.138	0.385	<0.001	0.036
	A2	1.43 b	3.32 a	2.79 a	3.25 Ba
	A6	1.00 b	2.97 a	3.21 a	2.88 BCa
	A9	1.19 b	3.35 a	2.82 a	2.55 Ca
BA (g/kg DM)	60	ND	ND	ND	ND	0.528	<0.001	0.707	<0.001
	A2	7.88 a	5.96 a	5.83 a	5.91 Ba
	A6	5.18 b	6.40 b	5.20 b	10.3 Aa
	A9	5.77 a	5.27 a	5.94 a	2.94 Ca

Different lowercase letters (a–d) indicate significant differences between treatments on the same day during aerobic exposure. Different capital letters (A–C) indicate significant differences between days after opening for the same treatments (p < 0.05). CK, untreated; LR, L. rhamnosus; LB L. buchneri; M, two-LAB mixture. SEM, standard error of means. D, aerobic exposure time; T, treatment (silages with different additives); D × T, interaction between the aerobic exposure time and the treatment; DM, dry matter; LA, lactic acid; AA, acetic acid; PA, propanoic acid; BA, butyric acid; 60, silage for 60 days; A2, A6, and A9, 2nd, 6th, and 9th day of aerobic exposure; ND, content below detectable levels.

).

3.3. Bacterial and Fungal Communities

The sampling depth effectively captured the majority of the bacterial communities, as indicated by the good coverage of all samples, which was approximately 0.99%. Aerobic exposure time had a significant effect on bacterial diversity (Shannon and Simpson: p < 0.05), while the LAB inoculation treatment had a significant effect on both bacterial diversity and abundance (p < 0.05). The Shannon and Simpson indices were significantly lower in the LAB treatment groups at the same time than in the CK group. The Shannon and Simpson indices increased with longer aerobic exposure time in the CK and LB treatments, with a significant increase in the CK (

Table 4. Alpha diversity of bacterial microbiota under different additives and aerobic exposure times.

Item	Day (d)	Treatment				SEM	p Value
		CK	LR	LB	M		D	T	D × T
Shannon	60	198.3	194.3	164.7	190.3	0.123	0.013	<0.001	0.208
	A2	224.7	243.3	191.0	168.0
	A6	238.3 a	192.3 b	179.0 b	187.3 b
	A9	226.7 ab	230.3 a	174.3 b	185.3 ab
Simpson	60	0.79 Ba	0.52 b	0.20 Bc	0.19 c	0.031	0.007	<0.001	0.217
	A2	0.76 Ba	0.57 b	0.26 ABc	0.19 d
	A6	0.76 Ba	0.50 b	0.25 Bc	0.18 c
	A9	0.87 Aa	0.56 b	0.46 Ab	0.21 c
Chao 1	60	233.6	237.3	205.2	250.1	16.404	0.255	0.024	0.760
	A2	260.9	297.2	237.0	226.7
	A6	283.0	263.1	234.4	257.5
	A9	272.1 ab	296.3 a	213.4 b	263.7 ab
ACE	60	229.7	246.0	213.6	262.9	17.064	0.220	0.048	0.636
	A2	259.2	307.3	249.4	237.9
	A6	303.3	270.3	241.5	262.7
	A9	277.7	305.4	227.6	268.5
Goods coverage	60	0.9996	0.9995	0.9996	0.9994	-	-	-	-
	A2	0.9996	0.9994	0.9995	0.9995
	A6	0.9994	0.9994	0.9995	0.9994
	A9	0.9994	0.9993	0.9995	0.9994

Different lowercase letters (a–c) indicate significant differences between treatments on the same day during aerobic exposure. Different capital letters (A–B) indicate significant differences between days after opening for the same treatments (p < 0.05). CK, untreated; LR, L. rhamnosus; LB L. buchneri; M, two-LAB mixture. SEM, standard error of means. D, aerobic exposure time; T, treatment (silages with different additives); D × T, interaction between the aerobic exposure time and the treatment; 60, silage for 60 days; A2, A6, and A9, 2nd, 6th, and 9th day of aerobic exposure.

).

Shannon and Simpson indices of fungi were significantly affected by aerobic exposure time (p < 0.05), and the LAB treatment had a significant effect on all fungal α-diversity indices (p < 0.001). Fungal richness and diversity were significantly increased in the LAB treatment compared to CK (p < 0.05). Except for the M treatment, fungal richness and diversity significantly decreased with increasing aerobic exposure time for all treatments (p < 0.05) (

Table 5. Alpha diversity of fungal microbiota under different additives and aerobic exposure times.

Item	Day (d)	Treatment				SEM	p Value
		CK	LR	LB	M		D	T	D × T
Shannon	60	4.48 Ab	5.47 Aa	5.14 Aa	5.23 a	0.256	<0.001	<0.001	<0.001
	A2	4.09 Ab	3.46 Cb	4.56 Aab	5.90 a
	A6	0.25 Bb	4.86 Ba	4.86 Aa	5.42 a
	A9	1.44 Bb	0.16 Db	0.30 Bb	5.06 a
Simpson	60	0.92 Ab	0.96 Aa	0.94 Aa	0.95 a	0.04	<0.001	<0.001	<0.001
	A2	0.83 Aab	0.73 Bb	0.90 Aab	0.97 a
	A6	0.05 Bb	0.93 Aa	0.92 Aa	0.96 a
	A9	0.38 Bb	0.03 Cc	0.07 Bbc	0.92 a
Chao 1	60	87.60 ABc	206.70 Aab	154.92 b	222.60 a	16.94	0.191	<0.001	0.011
	A2	102.91 Ab	139.53 Bab	183.00 ab	196.39 a
	A6	26.14 Cc	133.83 Bb	156.75 b	281.23 a
	A9	78.61 Bb	84.55 Bb	105.09 b	254.48 a
ACE	60	97.14 Ac	207.93 Aa	150.32 ABb	224.56 a	15.71	0.141	<0.001	0.004
	A2	100.31 Ab	140.37 Bab	188.52 Aa	194.25 a
	A6	28.30 Bc	131.11 BCb	155.27 ABb	279.72 a
	A9	84.59 Ab	83.45 Cb	108.69 Bb	254.00 a
Goods coverage	60	0.9999	0.9999	0.9999	0.9999	-	-	-	-
	A2	0.9999	0.9999	0.9999	0.9999
	A6	0.9999	0.9999	0.9999	0.9999
	A9	0.9999	0.9999	0.9998	0.9999

Different lowercase letters (a–c) indicate significant differences between treatments on the same day during aerobic exposure. Different capital letters (A–D) indicate significant differences between days after opening for the same treatments (p < 0.05). CK, untreated; LR, L. rhamnosus; LB L. buchneri; M, two-LAB mixture. SEM, standard error of means. D, aerobic exposure time; T, treatment (silages with different additives); D × T, interaction between the aerobic exposure time and the treatment; 60, silage for 60 days; A2, A6, and A9, 2nd, 6th, and 9th day of aerobic exposure.

).

Firmicutes and Proteobacteria were the dominant bacterial phyla at the phylum level, with Firmicutes increasing drastically and Proteobacteria decreasing sharply after the end of ensiling (Figure 2A). At the genus level, Pseudomonas, Pantoea, and Lactobacillus were the most dominant genera in the fresh samples. Lactobacillus was the most dominant genus of ryegrass at the end of ensiling, in addition to Enterococcus and Weissella, which were also the main genera in the CK group. After 60 days of ensiling, the relative abundance of Lactobacillus increased in all groups, unlike after 9 days of aerobic exposure, where the relative abundance of Lactobacillus decreased in all treatment groups except for LR. The relative abundance of Lactobacillus in the additive-treated group was higher than that in the CK group both at the end of ensiling and after 9 days of aerobic exposure, indicating that the addition of LAB resulted in a more favorable microbial environment for fermentation in ryegrass (Figure 2B).

At the fungal phylum level, Ascomycota and Basidiomycota were the dominant phyla. The relative abundance of Ascomycota decreased at the end of fermentation, while the relative abundance of Basidiomycota increased. After aerobic exposure, the relative abundance of Ascomycota increased in all treatment groups except for the M group (Figure 3A). At the genus level, the fresh samples and silages had similar fungal structures, with Articulospora and Cladosporium being the dominant genera. The relative abundance of Hannaella increased, and it became the dominant genera after silage compared to before silage. Fungal communities changed considerably after aerobic exposure, with Penicillium becoming the dominant fungal genus in the CK group. No significant changes were observed in the M group compared to the previous group (Figure 3B).

3.4. Metabolomic Profiles of Italian Ryegrass Silage

As shown in Table S1 (attached as a Supplementary Materials), 143 lipids and lipid-like molecules, 81 benzenoids, 27 flavonoids, 11 isoflavonoids, 95 organic acids and derivatives, 15 alkaloids and derivatives, 25 nucleosides, nucleotides, and analogs, and 3 lignans, neolignans, and related compounds were identified in silage Italian ryegrass.

The partial least squares discriminant analysis of the metabolites is shown in Figure 4. Little variation was observed between the biological replicates and the six replicates from each group were clustered, demonstrating sufficient reproducibility and reliability of the experiment. FM was separated from the metabolites of each treatment group at the end of ensiling and after aerobic exposure for 9 days. A difference was also observed between the CK and supplement groups, suggesting that the addition of LAB also has an effect on silage metabolites. LB was not clearly separated from the M group, indicating little difference in metabolites between the two treatment groups (Figure 4).

By mapping the differential metabolites between CK and LR Italian ryegrass silage, the most affected metabolic pathways were “phenylpropanoid biosynthesis” (five differential metabolites were annotated to this pathway); “phenylalanine, tyrosine and tryptophan biosynthesis”; “cyanoamino acid metabolism”; and “glucosinolate biosynthesis” (p < 0.05) (Figure 5A). Compared with CK, the main metabolic pathways of differential metabolites involved in LB were “phenylpropanoid biosynthesis” (five differential metabolites were annotated to this pathway), “glycerophospholipid metabolism”, “tropane, piperidine and pyridine alkaloid biosynthesis”, “tyrosine metabolism”, “cyanoamino acid metabolism”, and “glucosinolate biosynthesis” (p < 0.05) (Figure 5B). “Phenylpropanoid biosynthesis” (five differential metabolites were annotated to this pathway), “phenylalanine, tyrosine and tryptophan biosynthesis”, “cyanoamino acid metabolism”, and “glucosinolate biosynthesis” were the most affected metabolic pathways in the CK compared with M (p < 0.05) (Figure 5C). Only a few differentially accumulated metabolites related to “Flavone and flavonol biosynthesis” and “tryptophan metabolism” were detected in the comparison between LR and LB (Figure 5D).

In order to better understand the changes in the primary differential metabolites, the main metabolites based on the VIP were selected and sorted. A total of 1881 metabolites were identified after annotation, and these metabolites had individual metabolic characteristics at different time points after aerobic exposure and treatments. The top 20 significantly different metabolites during aerobic exposure were selected, and the VIP values of these 20 metabolites based on the PLS-DA model are shown (Figure 6). In terms of aerobic exposure, 4-acetamidobutanoic acid, L-pyroglutamic acid, choline, tyramine, and acetophenone all showed an increase, and L-palmitoylcarnitine, vanillin, and 5-methoxyindole-3-acetic acid all showed decreases, indicating that the effect of aerobic conditions on these several differential metabolites exceeded their modulation by the additive treatment, causing them to make unified changes under aerobic conditions. Except for the nine metabolites mentioned above, the changes in the remaining metabolites in the LR group showed a trend opposite to that in the LB and M groups. From different treatments, all differential metabolites showed the same trend in both the LB and M groups, which may be due to the absolute dominance of LB over LR in both treatment groups. After 9 days of aerobic exposure, six significantly different metabolites (betaine, tyramine, acetophenone, L-phenylalanine, ferulic acid, and luteolin) were enriched in the CK group, and eight were enriched in the LR group (L-pyroglutamic acid, choline, L-(-)-methionine, 5-methoxyindole-3-acetic acid, 5-OxoETE, luteolin, L-glutamic acid, and L-histidine). LB was enriched with only L-pyroglutamic acid as the main significantly different metabolite, while six significantly different metabolites (4-acetamidobutanoic acid, xanthine, betaine, L-tyrosine, L-pyroglutamic acid, and choline) were found in the M group.

Fold-changes in the relative concentrations of mycotoxins in Italian ryegrass silage with different treatments are shown in

Table 6. Fold-changes in relative concentrations of mycotoxins in Italian ryegrass silage with different treatments.

Compounds	Log2 (Fold-Change)	p-Value
CK vs. LR
Aflatoxin B1	0.734	0.452
Aflatoxin M1	−0.609	0.101
Aflatoxin G2	0.501	0.025
Aflatoxin G1	0.229	0.546
Sterigmatocystin	0.179	0.139
T-2 Triol	0.428	0.039
CK vs. LB
Aflatoxin B1	−1.761	0.236
Aflatoxin M1	−1.288	0.012
Aflatoxin G2	−0.378	0.244
Aflatoxin G1	−1.113	0.004
Sterigmatocystin	0.666	0.079
T-2 Triol	−1.370	0.000
CK vs. M
Aflatoxin B1	−0.380	0.710
Aflatoxin M1	−1.395	0.004
Aflatoxin G2	1.382	0.045
Aflatoxin G1	−0.545	0.130
Sterigmatocystin	0.078	0.558
T-2 Triol	0.406	0.055
LR vs. LB
Aflatoxin B1	−2.495	0.051
Aflatoxin M1	−0.679	0.063
Aflatoxin G2	−0.879	0.008
Aflatoxin G1	−1.342	0.006
Sterigmatocystin	0.487	0.175
T-2 Triol	−1.798	<0.001

. Most of the mycotoxins appeared to be downregulated in both the LB and M groups compared to the CK, indicating that both treatments inhibit the production of mycotoxins. In contrast, LR showed an upregulation of mycotoxins compared to CK, indicating that the addition of LR increased mycotoxin production. Except for sterigmatocystin, the remaining mycotoxins in the LB group showed a downregulation compared to LR.

Aflatoxin and sterigmatocystin in the LR group exposed aerobically for 9 days were not significant, although they were downregulated from the data at the end of silage. Moreover, T-2 Triol was significantly upregulated. The surprising result was that other mycotoxins in the LB group showed upregulation after aerobic exposure, except for aflatoxin M1. Upregulation of aflatoxin G1 and G2 in Italian ryegrass silage in the M group after aerobic exposure (

Table 7. Fold-changes in relative concentrations of mycotoxins in Italian ryegrass silage during aerobic exposure.

Compounds	Log2 (Fold-Change)	p-Value
CK60 vs. A9CK
Aflatoxin B1	−0.840	0.641
Aflatoxin M1	−1.004	0.073
Aflatoxin G2	0.258	0.367
Aflatoxin G1	−0.240	0.633
Sterigmatocystin	0.260	0.024
T-2 Triol	0.738	0.046
LR60 vs. A9LR
Aflatoxin B1	−0.050	0.968
Aflatoxin M1	−0.219	0.502
Aflatoxin G2	0.348	0.257
Aflatoxin G1	−0.577	0.333
Sterigmatocystin	−0.062	0.762
T-2 Triol	0.412	0.023
LB60 vs. A9LB
Aflatoxin B1	0.377	0.736
Aflatoxin M1	−1.737	0.027
Aflatoxin G2	1.620	0.011
Aflatoxin G1	0.220	0.495
Sterigmatocystin	1.155	0.050
T-2 Triol	1.133	0.006
M60 vs. A9M
Aflatoxin B1	−0.050	0.958
Aflatoxin M1	−0.201	0.520
Aflatoxin G2	2.931	0.002
Aflatoxin G1	1.028	0.036
Sterigmatocystin	0.329	0.159
T-2 Triol	−0.129	0.561

).

To explore the effects of organic acids, metabolites, microbial community composition, and aerobic stability (temperature over ambient) of Italian ryegrass silage in more detail, a structural equation model (SEM) was constructed to assess the direct and indirect effects between indicators and latent constructs (Figure 7A). It was found that bacterial and fungal community composition had positive or negative reciprocal direct effects on metabolites, as well as negative direct effects on pH. The aerobic stability of Italian ryegrass silage was only negative and directly influenced by the fungal composition. The direct effect of pH and organic acids on aerobic stability was not significant (p > 0.05).

From a network perspective, Figure 7B shows that Lactobacillus was positively correlated with the genera Hannaella and Pyrenochaetopsis, the metabolites L-tyrosine, 5-oxoETE, and 2-hydroxycinnamic acid, and the organic acids LA, AA, and PA and negatively correlated with the genera Enterococcus, Stenotrophomonas, Sphingomonas, Penicillium, and Issatchenkia, and the metabolite acetophenone. Penicillium was positively correlated with the genera Bacillus and Psychrobacillus and the metabolites L-phenylalanine and choline and negatively correlated with the genera Articulospora, Cladosporium, Erythrobasidium, Pyrenochaetopsis, and Hannaella, as well as pH and AA content.

4. Discussion

4.1. Fermentation Characteristics of Ryegrass Silage

LAB inoculants are used to improve and enhance the quality of silage fermentation and may affect aerobic stability during opening and utilization. Upon the opening of the silage, the environment underwent a rapid transition from anaerobic to aerobic conditions. This resulted in the activation of yeast within the silage, which rapidly consumed LA and caused a notable increase in pH [28]. Similarly, feed exposure to air triggers higher aerobic microbial activity, and lactic acid assimilation yeast initiates the process of aerobic spoilage. This phenomenon eventually leads to the generation of excessive heat [29]. Following a nine-day aerobic exposure period, the pH of the CK and LR silages exhibited a marked increase, whereas the silages treated with LB and M demonstrated a more stable pH profile. This may be explained by the addition of L. buchneri, a heterofermentative LAB, to both treatment groups, which inhibits aerobic spoilage [18]. The pH of M and LB did not increase significantly until 19 days after aerobic exposure. The aerobic stability of the various treatments was also corroborated by these findings. The aerobic stability of the CK and LR silages was observed to persist for a period of nine days, whereas the LB and M silages demonstrated a longer duration of stability, exceeding 19 days. It has been demonstrated that the L. buchneri strain has the capacity to enhance the aerobic stability of silage [30], a number of experiments [8,31,32] have examined this bacterium as a typical heterofermentative lactic acid bacterium (LAB) in silage, with the aim of increasing aerobic stability by producing acetic acid (AA) as a means of inhibiting yeast growth, even in silages with high water content. L. buchneri can improve their aerobic stability [33]. In addition, acetobacterium has also been shown to oxidize ethanol to acetic acid, thereby increasing aerobic stability [34].

4.2. Effects of LAB on the Microbial Community of Italian Ryegrass Silage

Generally, epiphytic bacterial communities were highly correlated with feed type [35], and a very low natural abundance of Lactobacillus was found in grasses [36] and grains [37]. It is well-documented that Lactobacillus species play a pivotal role in silage production, given their rapid growth and ability to reduce pH by producing lactic acid. Strains belonging to the Lactobacillus genus are typically employed as silage additives with the objective of enhancing the quality of the fermentation process [38,39]. After prolonged ensiling, exogenous LAB decreased unwanted bacteria like Paenibacillus, Rosenbergiella, and Providencia while increasing the growth of Lactobacillus and ensuring their continued dominance of the bacterial community. The aforementioned findings demonstrated how exogenous LAB helps to stabilize the microbiota over extended storage [40]. As expected, in our study, LR and LB increased the abundance of Lactobacillus (Figure 2B). Generally, Clostridium is undesirable in silage because it competes with LAB, leading to an increase in pH and a slowdown in nonprotein-N accumulation [41]. Clostridium produces BA and leads to poor fermentation quality. Some species may also produce pathogenic toxins during silage that are harmful to animals [42]. In the present study, fewer Clostridium were observed in the LAB-treated group than in the CK group, which is consistent with the better fermentation quality of the LAB-treated silage. The bacterial diversity was consistently higher for the CK than for the other treatment groups, probably because the LAB in the CK group is less dominant than in the treatment group, hence leading to a higher diversity. Lactobacillus was the dominant bacterial genus in all treatment groups, both at the end of fermentation and after 9 days of aerobic exposure. This may be due to the lower ambient temperature (15 °C) at the time of aerobic exposure that inhibited the activity of all microorganisms, making their structural changes very insignificant. The increase in fungal diversity at the end of fermentation may be because the particular silage environment of high-moisture Italian ryegrasses reduced the inhibition of fungi even in the microbial environment with a higher relative abundance of Lactobacillus. In contrast, the rapid loss of water in Italian ryegrass after aerobic exposure reduced the suppression of Lactobacillus and made the emergence of fungi more repressive in a microbial environment dominated by Lactobacillus [43].

The relative abundance of Issatchenkia in the CK increased substantially at 9 days of aerobic exposure and became the dominant fungal genus, while the abundance in the Lactobacillus-supplemented treatments increased only slightly, which is consistent with previous findings [44] and may be due to the inhibition of Issatchenkia growth by Lactobacillus during aerobic exposure. Penicillium is a vital fungus in the aerobic spoilage process of forages, producing not only various toxins but also penicillic acid and mycophenolic acid. At 9 days of aerobic exposure, Penicillium was the dominant genus in the CK and LR groups, while its relative abundance was minimal in the LB group and the M group, probably because of the metabolic production of AA and phenyllactic acid by L. buchneri, and these two organic acids are very effective in inhibiting the growth and Ochratoxin A production of Penicillium [45].

4.3. Metabonomic Analysis of Italian Ryegrass Silage

The fermentation process of silage aerobic spoilage is complex and involves interactions between metabolites and various microorganisms. The addition of LAB not only affected the composition of silage microorganisms but also changed the metabolism of silage microorganisms. Studies have revealed that LAB also produce some metabolites with antifungal properties, such as 4-hydroxybenzoic acid, 3-hydroxydecanoic acid, acetaldehyde, 3-(R)-hydroxytetradecanoic acid, vanillic acid, 2,3-butanedione, and bacteriocins [46,47]. Different metabolites affected various metabolic pathways, but the same differential metabolic pathway (phenylpropanoid biosynthesis) was observed in all of the LAB-treated Italian ryegrass silage compared to the CK (Figure 5A–C). The differential metabolites in this metabolic pathway were trans-2-hydroxy-cinnamate and ferulic acid with antioxidant properties [48] and phenylalanine, tyrosine, and cinnamaldehyde for flavor addition [49,50,51]. This indicates that the use of LAB additives through the “phenylpropanoid biosynthesis” pathway increased the aerobic stability of Italian ryegrass silage. Flavonoids, a major class of plant secondary metabolites, are known to be a phenolic family with beneficial health effects due to their antioxidant, anti-inflammatory, antimicrobial, antimutagenic, and anticancer properties [52]. Xu et al. [53] reported that many phenolics, such as flavonoids and flavones, were identified as differential metabolites between control and inoculant sainfoin silages. The differential metabolic pathways in the LB and LR groups were “flavone and flavonol biosynthesis”, “tryptophan metabolism”, and “porphyrin and chlorophyll metabolism” (Figure 5D). Among the differential metabolites in the “flavone and flavonol biosynthesis” metabolic pathway, including vitexin with antioxidant properties and luteolin with antimicrobial properties, we inferred that the inoculant L. buchneri can regulate flavone and flavonol polyphenol during silage. This may be the cause of the difference between L. buchneri and L. rhamnosus as the main metabolite responsible for the difference in the ability of the two additives to inhibit aerobic spoilage of Italian ryegrass silage.

Tryptophan and phenylalanine are both aromatic amino acids and essential amino acids [54]. In livestock feeding, low tryptophan intake will reduce appetite [55]. Tryptophan can be used as a feed additive to regulate appetite in pigs, sheep, and cattle, thereby improving farming efficiency [55]. Nathalie et al. [56] produced a more severe inflammatory response in pigs fed food with low tryptophan content. Phenylalanine is a common precursor to many common precursors of many phenolic compounds, including flavonoids, polyflavonoids, lignin, and phenylpropanoids [57]. In phycology, derivatives of phenylalanine have a wide range of physiological functions; since animals cannot synthesize phenylalanine, it has become an essential requirement in animal diets [58]. In all additive-treated ryegrass silages, the relative abundances of DL-tryptophan and L-phenylalanine were downregulated to varying degrees at 9 days of aerobic exposure, indicating that the decrease in feed palatability after aerobic spoilage is likely to be related to these two amino acids. In addition, the downregulation of tryptophan after aerobic exposure may be because it can be metabolized by Lactobacillus to indole-3-carboxaldehyde and by Bifidobacterium to indole-3-lactic acid [59,60], whereas at 9 days of aerobic exposure, the LB group and M group still had a higher relative abundance of Lactobacillus. This also explains the upregulation of indole-3-lactic acid in these two treatment groups after aerobic exposure. Penicillium was positively correlated with L-phenylalanine (Figure 7B), while the extremely high relative abundance of Penicillium in the CK and LR groups (>50%, Figure 3B) at 9 days of aerobic exposure may explain the upregulation of L-phenylalanine in these two treatment groups (Figure 6). In addition, Issatchenkia had a high relative abundance in the CK at 9 days of aerobic exposure (Figure 3B) and was negatively correlated with the contents of 2-hydroxycinnamic acid and indole with aroma-enhancing effects, the fatty acid 5-OxoETE, and LA and the abundance of Lactobacillus (Figure 7B). These findings suggest that the loss of nutrients and reduced palatability of silage under aerobic conditions without the addition of Lactobacillus are associated with Issatchenkia.

In the present research, the only mycotoxins screened were aflatoxin, sterigmatocystin, and T-2 triol, which may be the result of the average temperature being below 15 °C, while fungi proliferation and mycotoxin production are enhanced in hot and humid environments [61]. Aflatoxin is one of the most abundant and mutagenic mycotoxins present in forage [62]. Aflatoxin B1 is the most carcinogenic chemical known, being strongly poisonous to humans and animals, and its toxic effects mainly cause liver damage [63]. Several studies have shown that L. rhamnosus can effectively resist aflatoxin B1, M1, and G1 through the adsorption mechanism [64,65,66], which explains why there was a slight downregulation of aflatoxin B1, M1, and G1 in the LR group after aerobic exposure in our study. Although it is widely accepted that LA and AA produced by LAB are the main metabolites that inhibit fungal growth and mycotoxin production, it is clear that AA is the most effective antifungal metabolite of LAB [67]. The minimum inhibitory concentration value of AA was 33 mM to inhibit 50% Penicillium nordicum growth and 100% ochratoxin production [45]. This may explain the fact that the LB group with the highest concentration of AA had less Penicillium and mycotoxin than the other treatment groups (Figure 3B and Table 4). It has been shown that LA has no effect on the growth of several fungal species under the genus Penicillium [68], which explains why Italian ryegrass silage under L. rhamnosus additions had a very high relative abundance of Penicillium, although LA was detected at considerable levels after aerobic exposure. T-2 triol is a fungal toxin produced by the metabolism of T-2 toxin, which is less poisonous than T-2 toxin. T-2 toxin inhibits the growth of Lactobacillus in the growth medium [69]. In our results, the relative abundance of Lactobacillus decreased while T-2 triol appeared significantly upregulated in the CK, LB, and LR groups as aerobic exposure proceeded (Figure 2B and Table 5).

To explore the complex relationships among organic acids, metabolites, microbial community composition, and the aerobic stability (temperature over ambient) of Italian ryegrass silage, we conducted an SEM analysis. We found that aerobic stability was not directly affected by metabolites or the bacterial community. This result contradicted our hypothesis that metabolites and bacterial communities were the main and direct contributors to the aerobic stability of Italian ryegrass silages. In this regard, we speculate that high moisture conditions inhibit the direct effect of bacteria on aerobic stability and instead act indirectly on aerobic stability through interactions between metabolites. Consistent with previous studies, the fungal community was a direct factor influencing the aerobic stability of silage [70], and in our model, the fungal community was also the only direct negative influence. In addition to the indirect effects of bacteria and metabolites, a very stable fungal community structure was maintained before and after aerobic exposure, which may explain the extra-long stability of the M group after aerobic exposure. Metabolites were significantly affected directly by bacterial and fungal communities, while in Figure 3, we see that bacteria and fungi showed similar community compositions in all additive treatment groups for 60 days of silage, which may explain the small difference in metabolites between these three treatment groups (Figure 4). In our results, bacteria had a direct positive but not significant effect on organic acid levels. This may be because our data are in the aerobic contact phase, when the activity of the dominant bacterial genus, anaerobic Lactobacillus, is inhibited, reducing its ability to produce LA and AA, and thus the causal relationship between the two is not significant. This can also explain the insignificant correlation between bacteria and fungi, where LAB, which receive inhibition after aerobic exposure, have a diminished inhibitory or promotional effect on fungi and lead to the regulation of fungi through interactions with metabolites. Organic acids have a negative direct effect on fungi, which is well understood because AA, PA, and BA have a good inhibitory effect on the growth and reproduction of fungi [67].

5. Conclusions

The addition of LB and M improved the feed quality and enhanced the aerobic stability of Italian ryegrass. The fungal community was a direct factor influencing the aerobic stability of silage, while bacteria indirectly affected aerobic stability through interactions with metabolites and fungi. A total of 1881 chemical compounds were identified in Italian ryegrass silage. These metabolites were associated with bacterial communities, especially Lactobacillus. The use of Lactobacillus additives increased the palatability and resistance to aerobic spoilage of Italian ryegrass silage through the “phenylpropanoid biosynthesis” metabolic pathway. The loss of nutrients in Italian ryegrass silage under aerobic conditions was associated with Penicillium and Issatchenkia. The application of LB and M is a feasible way to obtain well-fermented Italian ryegrass silage and to improve aerobic stability, even at relatively high moisture content levels. Therefore, in order to retain biomass, enhance fermentation quality, and enrich biofunctional metabolites, it is recommended that homo-LAB be inoculated in high-moisture silage.