38
Assaying the metabolic network of the bovine rumen microbiome
Diverse populations of microbes inhabit the ruminant’s digestive system helping to degrade substrates that are inaccessible to the enzymes encoded in the host’s genome. Variation in the metagenome is associated with factors including feed efficiency, methane production, and overall animal health. Gene transfer among microbes in rumen makes the taxonomic structure of the ecosystem an imperfect representation of its functional structure. The objective of this study is to elucidate the metabolic interface between the ruminal metagenome and the host metabolic network. Rumen fluid samples were collected from two crossbred steers fed a concentrate feedlot diet. DNA was extracted using the QIAamp DNA Stool Mini Kit, libraries were constructed, and shotgun sequencing was completed on the Illumina GAII platform. Paired-end reads were quality filtered. Filtered reads were compared to a reference database of 16S rDNA genes. Reads with ≥97% sequence identity defined operational taxonomic units (OTU), or groups of like species. To infer the metabolic network of the metagenome, all six frame-translations were computed from each paired-end read. Any translation where both pairs had open reading frames (ORF) longer than 45 amino acids in opposite orientations were retained and queried against an enzyme database allowing comparison of the animals’ similarity in OTU distributions and ruminal enzymes. To assess the metabolic interface between the microbiome and the host animal, a host metabolic network was defined. Using comparative genomics, bovine orthologs of the human enzymes in MetaCyc were linked to the microbial enzymes with a list of potentially shared metabolites. OTU distributions were not identical for the two animals, Spearman’s ρ=0.45 (p<10-6); however, the metabolism of the metagenomes was more similar, Spearman’s ρ=0.93 (p<10-10). A distance metric on the metabolic network was calculated by using the list of metabolites potentially taken from the microbiome by the host. While the two animals were quite similar in copy number for enzymes with potential interactions with the host metabolic network, they were more dissimilar in more distant parts of the metabolic network. Thus, the taxonomic studies of microbial ecosystems may overstate the amount of functional variation in those ecosystems; however, despite this, the ecosystems do differ in their enzymatic structure. The nature of those differences suggest that even when there is strong external pressure, it is possible to construct ecosystem-scale metabolic networks that differ in their composition but where those differences are partly masked by a common metabolic interface.
Keywords:
metagenomics, metabolic network