Bacterial community analysis of purulent material from liver abscesses of crossbred cattle and Holstein steers fed finishing diets with or without tylosin

Abstract

Liver abscesses in feedlot cattle are polymicrobial infections. Culture-based studies have identified Fusobacterium necrophorum as the primary causative agent, but a number of other bacterial species are frequently isolated. The incidence of liver abscesses is highly variable and is affected by a number of factors, including cattle type. Holstein steers raised for beef production have a higher incidence than crossbred feedlot cattle. Tylosin is the commonly used antimicrobial feed additive to reduce the incidence of liver abscesses. The objective of this study was to utilize 16S ribosomal RNA amplicon sequence analyses to analyze the bacterial community composition of purulent material of liver abscesses of crossbred cattle (n = 24) and Holstein steers (n = 24), each fed finishing diet with or without tylosin. DNA was extracted and the V3 and V4 regions of the 16S rRNA gene were amplified, sequenced, and analyzed. The minimum, mean, and maximum sequence reads per sample were 996, 177,070, and 877,770, respectively, across all the liver abscess samples. Sequence analyses identified 5 phyla, 14 families, 98 genera, and 102 amplicon sequence variants (ASV) in the 4 treatment groups. The dominant phyla identified were Fusobacteria (52% of total reads) and Proteobacteria (33%). Of the top 25 genera identified, 17 genera were Gram negative and 8 were Gram positive. The top 3 genera, which accounted for 75% of the total reads, in the order of abundance, were Fusobacterium, Pseudomonas, and Bacteroides. The relative abundance, expressed as percent of total reads, of phyla, family, and genera did not differ (P > 0.05) between the 4 treatment groups. Generic richness and evenness, determined by Shannon-Weiner and Simpson's diversity indices, respectively, did not differ between the groups. The UniFrac distance matrices data revealed no clustering of the ASV indicating variance between the samples within each treatment group. Co-occurrence network analysis at the genus level indicated a strong association of Fusobacterium with 15 other genera, and not all of them have been previously isolated from liver abscesses. In conclusion, the culture-independent method identified the bacterial composition of liver abscesses as predominantly Gram negative and Fusobacterium as the dominant genus, followed by Pseudomonas. The bacterial community composition did not differ between crossbred and Holstein steers fed finishing diets with or without tylosin.

Keywords: 16S ribosomal RNA; Holsteins; bacteriome; crossbred cattle; liver abscesses; tylosin.

Figure 1.

Relative abundance of the bacterial phyla in liver abscesses from crossbred cattle and Holstein steers fed high-grain finishing diets with or without tylosin.

Figure 2.

Heatmap of relative abundance, based on read counts, of the top 25 genera in liver abscesses from crossbred (CB) and Holstein (HF) steers fed high-grain finishing diets with or without tylosin.

Figure 3.

Boxplots showing Shannon–Weiner’s and Simpson’s α diversity indices of bacterial communities in liver abscesses from crossbred (CB) cattle and Holstein (HF) steers fed high-grain finishing diets with or without tylosin. Treatment means were compared using Kruskal–Wallis test to identify differences in species richness and evenness between treatment groups. No significant differences between treatment groups were detected (P > 0.05).

Figure 4.

Beta diversity metrics of bacterial communities in liver abscesses from crossbred (CB) cattle and Holstein (HF) steers fed high-grain finishing diets with or without tylosin. Principle coordinate plot using weighted UniFrac distances are shown. Each point represents a liver abscess bacteriome of a sample. Samples are colored by treatment subgroups of tylosin-fed and no tylosin-fed cattle. Axes represent the amount of variation explained in the first 2 principle coordinates. No significant differences between treatment groups were detected (P > 0.05).

Figure 5.

Heatmap of differentially abundant ASV between tylosin-fed and no tylosin-fed cattle (A) and between crossbred (CB) cattle and Holstein (HF) steers (B). Deferential ASV were identified using DESeq2 (see Materials and Methods). Tiles in the heatmap are colored by variance stabilization transformation of read counts, with the deeper color indicating a higher abundance of the organism within sample. Differential ASV were clustered using hierarchical clustering of variance stabilized read count data.

Figure 6.

Co-occurrence network of liver abscess bacteriome at genus and family level using SparCC network analysis. Each SparCC analysis was performed using 20 iterations, and correlations below a threshold of 0.3 were considered as zero for the inner SparCC loop. Clusters of closely associated taxa are represented by halos around the nodes, calculated using walktrap community analysis. Nodes with read counts collapsed to the genus level taxonomy (A) and to the family level (B). Black lines represent within cluster associations, while red lines represent between cluster associations. The genus Fusobacterium (16; purple circle) was associated with other 15 genera indicated in upper case letters (A). The family Fusobacteriaceae (12; purple circle) was associated with 15 other families indicated in upper case letters (B).