Bacterial diversity and composition of alfalfa silage as analyzed by Illumina MiSeq sequencing: Effects of Escherichia coli O157:H7 and silage additives

The first objective of this study was to examine effects of adding Escherichia coli O157:H7 with or without chemical or microbial additives on the bacterial diversity and composition of alfalfa silage. The second objective was to examine associations between the relative abundance of known and unknown bacterial species and indices of silage fermentation quality. Alfalfa forage was harvested at 54% dry matter, chopped to a theoretical length of cut of 19 mm, and ensiled in quadruplicate in laboratory silos for 100 d after the following treatments were applied: (1) distilled water (control); (2) 1 × 10 5 cfu/g of E. coli O157:H7 (EC); (3) EC and 1 × 10 6 cfu/g of Lactobacillus plantarum (EC+LP); (4) EC and 1 × 10 6 cfu/g of Lactobacillus buchneri (EC+LB); and (5) EC and 0.22% propionic acid (EC+PA). After 100 d of ensiling, the silage samples were analyzed for bacterial diversity and composition via the Illumina MiSeq platform (Illumina Inc., San Diego, CA) and chemically characterized. Overall, Firmicutes (74.1 ± 4.86%) was the most predominant phylum followed by Proteobacteria (20.4 ± 3.80%). Relative to the control, adding E. coli O157:H7 alone at ensiling did not affect bacterial diversity or composition but adding EC+LP or EC+LB reduced the Shannon index, a measure of diversity (3.21 vs. 2.63 or 2.80, respectively). The relative abundance of Firmicutes (69.2 and 68.8%) was reduced, whereas that of Proteobacteria (24.0 and 24.9%) was increased by EC+LP and EC+PA treatments, relative to those of the control (79.5 and 16.5%) and EC+LB (77.4 and 18.5%) silages, respectively. Compared with the control, treatment with EC+LP increased the relative abundance of Lactobacillus, Sphingomonas, Pantoea, Pseudomonas , and Erwinia by 426, 157, 200, 194, and 163%, respectively, but reduced those of Pediococcus, Weissella , and Methylobacterium by 5,436, 763, and 250%, respectively. Relative abundance of Weissella (9.19%) and Methylobacterium (0.94%) were also reduced in the EC+LB silage compared with the control (29.7 and 1.50%, respectively). Application of propionic acid did not affect the relative abundance of Lactobacillus, Weissella , or Pediococcus. Lactate concentration correlated positively (r = 0.56) with relative abundance of Lactobacillus and negatively (r = −0.41) with relative abundance of Pediococcus . Negative correlations were detected between ammonia-N concentration and relative abundance of Sphingomonas (r = −0.51), Pantoea (r = −0.46), Pseudomonas (r = −0.45), and Stenotrophomonas (r = −0.38). Silage pH was negatively correlated with relative abundance of Lactobacillus (r = −0.59), Sphingomonas (r = −0.66), Pantoea (r = −0.69), Pseudomonas (r = −0.69), and Stenotrophomonas (r = −0.50). Future studies should aim to speciate, culture, and determine the functions of the unknown bacteria detected in this study to elucidate their roles in silage fermentation.