BACKGROUND
Alfalfa is considered difficult to ensile because of its relatively low water-soluble carbohydrate (WSC) content and high buffering capacity, factors that can increase nutrient loss during the aerobic phase of silage production. The authors note that dry matter (DM) loss from aerobic respiration can reach 4–20%, and may be 2% higher in alfalfa silage than in grass silage. Bacillus species may help consume oxygen and create a more anaerobic environment, potentially shortening the aerobic phase and improving fermentation. This study evaluated whether Bacillus coagulans (BC) could function as an alfalfa silage inoculant, alone or in combination with Lactobacillus plantarum (LP), and examined effects on fermentation characteristics, chemical composition, bacterial community structure, and predicted microbial functions.
METHODS
Fresh alfalfa (cultivar “WL 440”) was harvested at the early bloom stage, wilted to a DM content of 329.60 g/kg fresh weight within 8 h, and chopped into 1–2 cm pieces. Before ensiling, the wilted alfalfa had pH 6.49, WSC 19.61 g/kg DM, crude protein (CP) 201.29 g/kg DM, neutral detergent fiber (NDF) 476.73 g/kg DM, acid detergent fiber (ADF) 311.48 g/kg DM, lactic acid bacteria (LAB) 6.41 log10 CFU/g FW, yeasts 3.28 log10 CFU/g FW, and molds 2.26 log10 CFU/g FW. The inoculation groups were control (CON), LP, BC, and LP+BC. Inoculants were applied at 1 × 10^6 CFU/g FW. Silages were vacuum packaged, stored at 25 ± 2 °C, and sampled at 3, 7, 14, 30, and 60 d, with three replicates per treatment per time point. Fermentation end points included pH, lactic acid (LA), and acetic acid (AA), along with chemical composition and microbial counts. Bacterial communities were analyzed by 16S rRNA sequencing of the V3-V4 region, and functional prediction used PICRUSt2.
KEY RESULTS
Fermentation improved over time in all groups, with pH decreasing and LA increasing. At 3 d, pH values were 6.13 in CON, 5.42 in LP, 5.79 in BC, and 5.36 in LP+BC. At 60 d, pH values were 5.38 in CON, 4.89 in LP, 5.14 in BC, and 4.85 in LP+BC, with reported significance for CON vs. LP of <0.001, CON vs. BC of <0.001, and LP vs. LP+BC of 0.57; the overall inoculant effect was <0.001. LA concentrations also favored inoculated treatments. At 60 d, LA was 74.88 g/kg DM in CON, 91.55 g/kg DM in LP, 82.12 g/kg DM in BC, and 93.83 g/kg DM in LP+BC; reported p-values were 0.02 for CON vs. LP, 0.05 for CON vs. BC, 0.04 for LP vs. LP+BC, and <0.001 for the inoculant effect. AA at 60 d was 47.19 g/kg DM in CON, 42.19 g/kg DM in LP, 43.59 g/kg DM in BC, and 23.87 g/kg DM in LP+BC; reported p-values were 0.34 for CON vs. LP, 0.49 for CON vs. BC, <0.001 for LP vs. LP+BC, and <0.001 for the inoculant effect.
After 60 d, DM was 432.47 g/kg in CON, 388.31 g/kg in LP, 367.11 g/kg in BC, and 394.80 g/kg in LP+BC. DM loss was 10.61%, 8.25%, 9.65%, and 7.98%, respectively, but the authors reported no effects of LP or BC on DM and DM loss. WSC was 12.88 g/kg in CON, 7.93 g/kg in LP, 14.15 g/kg in BC, and 12.10 g/kg in LP+BC, with reported p-values of 0.01 for CON vs. LP, 0.33 for CON vs. BC, 0.01 for LP vs. LP+BC, and 0.001 for the inoculant effect. CP was similar across groups: 208.34, 211.09, 210.45, and 216.65 g/kg DM. In contrast, NH3-N was reduced in treated silages: 142.00 g/kg total N in CON, 126.51 in LP, 130.91 in BC, and 110.90 in LP+BC, with p-values of 0.003 for CON vs. LP, 0.03 for CON vs. BC, 0.004 for LP vs. LP+BC, and <0.001 overall. Fiber fractions were also lower in treated silages. NDF was 391.29 g/kg DM in CON, 363.00 in LP, 374.00 in BC, and 361.78 in LP+BC, with p-values of 0.01 for CON vs. LP, 0.10 for CON vs. BC, 0.11 for LP vs. LP+BC, and <0.001 overall. ADF was 269.53 g/kg DM in CON, 247.00 in LP, 254.00 in BC, and 214.83 in LP+BC, with p-values of 0.004 for CON vs. LP, 0.02 for CON vs. BC, <0.001 for LP vs. LP+BC, and <0.001 overall.
Microbiologically, LAB counts after 60 d were 7.41 log10 CFU/g in CON, 7.71 in LP, 7.61 in BC, and 7.91 in LP+BC. Yeasts were 2.79, 2.62, 2.71, and 2.53 log10 CFU/g, respectively, and molds were <2.0 in all groups. On beta diversity analysis, PC1 and PC2 explained 42.05% and 25.89% of variance. Before ensiling, Proteobacteria dominated at 75.97%, followed by Actinobacteriota 15.53%, Firmicutes 4.98%, and Bacteroidota 2.44%. After 60 d, Firmicutes became dominant, reaching 80.34% in CON, 87.41% in LP, 92.87% in BC, and 88.69% in LP+BC. At the genus level, fresh material was dominated by Pseudomonas 30.50%, Methylobacterium 11.48%, Sphingomonas 7.91%, and Enterobacter 7.00%. After ensiling, CON was characterized by Weissella 41.59% and Enterococcus 27.99%, while Lactobacillus dominated LP and LP+BC at 80.73% and 74.98%, respectively; BC had Lactobacillus 15.28% and Weissella 42.88%. LEfSe showed Lactobacillus and Lactobacillaceae enriched in LP-treated silage with LDA > 5.5, while Enterococcaceae and Enterococcus were enriched in BC-treated silage with LDA > 5.0. Spearman analysis showed LA was positively correlated with Lactobacillus, while pH was negatively correlated with Lactobacillus and positively correlated with Lactococcus, Enterococcus, and Aerococcus.
CLINICAL IMPLICATIONS
This was not a clinical study, but its applied significance is clear for animal feed preservation and agricultural microbiology. The main practical finding is that BC improved fermentation quality and that the LP+BC combination produced the best overall profile, including lower pH, higher LA, lower NH3-N, and lower ADF than LP alone. The authors also report that LP, BC, and LP+BC increased cofactor and vitamin metabolism abundance while decreasing drug resistance: antimicrobial pathway abundance, supporting the view that BC may be a useful bioresource for improving alfalfa silage fermentation quality.