Diversity and functional analysis of rumen and fecal microbial communities associated with dietary changes in crossbreed dairy cattle

Abstract

The objective of this study was to investigate the effect of varying roughage and concentrate proportions, in diet of crossbreed dairy cattle, on the composition and associated functional genes of rumen and fecal microbiota. We also explored fecal samples as a proxy for rumen liquor samples. Six crossbred dairy cattle were reared on three diets with an increasing concentrate and reducing roughage amount in three consecutive 10-day periods. After each period, individual rumen liquor and fecal samples were collected and analyzed through shotgun metagenomic sequencing. Average relative abundance of identified Operational Taxonomic Units (OTU) and microbial functional roles from all animals were compared between diets and sample types (fecal and rumen liquor). Results indicated that dietary modifications significantly affected several rumen and fecal microbial OTUs. In the rumen, an increase in dietary concentrate resulted in an upsurge in the abundance of Proteobacteria, while reducing the proportions of Bacteroidetes and Firmicutes. Conversely, changes in microbial composition in fecal samples were not consistent with dietary modification patterns. Microbial functional pathway classification identified that carbohydrate metabolism and protein metabolism pathways dominated microbial roles. Assessment of dietary effects on the predicted functional roles of these microbiota revealed that a high amount of dietary concentrate resulted in an increase in central carbohydrate metabolism and a corresponding reduction in protein synthesis. Moreover, we identified several microbial stress-related responses linked to dietary changes. Bacteroides and Clostridium genera were the principal hosts of these microbial functions. Therefore, the roughage to concentrate proportion has more influence on the microbial composition and microbial functional genes in rumen samples than fecal samples. As such, we did not establish a significant relationship between the rumen and fecal metagenome profiles, and the rumen and fecal microbiota from one animal did not correlate more than those from different animals.

Fig 1. Phylogenetic classification of the rumen microbiome as revealed by metagenomic analysis.

a. Pie chart for microbial classification at the domain level. Averages are from all fecal and rumen liquor samples. b. Stacked bar plot of relative abundance of the top 10 bacterial phyla in the three diets for fecal and rumen liquor samples. c. Stacked bar plot of relative abundance of the dominant phyla in Archaeal domain. d. Stacked bar plot of relative abundance of one dominant phylum in viral domain, and e. Stacked bar plot of relative abundance of the dominant phyla in the Eukaryota domain.

Fig 2. Alpha and beta diversity analysis.

a. Box-and-whiskers plot for estimated alpha diversity indices in the fecal and rumen liquor microbiota. NS = Not Significant, * = Significant at p = 0.05, ** = Significant at p = 0.01, *** = Significant at p = 0.001. b. principal coordinate analysis (PCOA) of fecal and rumen liquor microbial communities. The axis of principal component 1 (PC1) described 51.8% and principal component 2 (PC2) described 25.64% of total data variability. PCOA was performed using PAST v3.13 [35]. S = sample type, D = diet and SXD = sample type X diet interaction.

Fig 3. Comparison of relative abundance of most abundant taxa within diets in each sample type and between fecal and rumen liquor sample types.

a. heatmap hierarchical clustering of most abundant phyla. In the heatmap, F = fecal, RL = rumen liquor, S = sample type, D = diet and SXD = sample type X diet interaction. Phyla color correspond to domain, blue = Bacteria, red = Archaea, green = Eucharyota and brown = Virus. NS = Not Significant, * = Significant at p = 0.05, ** = Significant at p = 0.01, *** = Significant at p = 0.001. b. scatter plot of 20 most abundant genera, and c. bar chart of 15 most abundant species. In fig b and c, a log transformed scale of the relative abundances is used for beater visualization of low abundance taxa.

Fig 4. Cladograms and histograms illustrating significantly different microbial taxa in LEfSe analysis.

a. Diets within fecal sample type, b. Diets within rumen liquor sample type, c. Fecal and rumen liquor sample types. The cladogram illustrates the presence of microbial phyla and genera that are significantly different based on linear discriminant analysis (LDA) scores in the histogram. The LDA score at log 10 > 4 is set as threshold and the length of each bin, i.e., LDA score represents the extent to which the microbial taxa differ among the groups.

Fig 5. Correlation heatmap representing correlations.

Correlation heatmap between fecal and rumen liquor microbial abundance in all the animals. F = fecal, RL = rumen liquor, D = Diet and numbers 1 to 6 represent the individual animals. Fecal -rumen liquor pairs from the same animal are highlighted by a black square. Overall comparison between fecal and rumen liquor sample types was not significant p = 0.91.

Fig 6. Stacked bar charts of relative abundance (%) of predominant microbiota associated with various functional pathways.

a. protein biosynthesis, b. central carbohydrate metabolism, c. sulfur metabolism, d. oxidative stress, e. cell wall and capsule and f. dormancy and sporulation pathways.