Unique bacterial communities associated with components of an artificial aquarium ecosystem and their possible contributions to nutrient cycling in this microecosystem.

In order to better understand the bacterial distribution characteristics in a whole microecosystem, the bacterial communities in different components of an artificial aquarium (i.e., plants, fishes, sand and water) were characterized using high throughput sequencing of bacterial 16S rRNA genes. Across all samples, 2873 operational taxonomic units were identified and assigned to 771 genera in 36 phyla. In a principle coordinate analysis, samples clustered according to their origin, indicating that bacterial communities from the same component were most similar. Further taxonomic analysis revealed that most dominant genera, even those with the similar functions, were biased to one component: Nitrospira and Rhodobacter were mainly abundant in plant samples; Rhodococcus, Serratia, Ralstonia, Sphingobacterium and Pseudomonas were most common in sand samples; Cetobacterium and Aeromonas dominated fish samples; and Flavobacterium, Alpinimonas and Limnobacter were especially common in water samples. Functional predictions performed by PICRUSt and the dominant genera exhibited that bacteria detected in each component could participate in all nutrient cycles in the aquarium. However, those involved in carbon and nitrogen cycling were most common in plant and fish samples, while phosphate metabolism-related pathways were more abundant in sand and water samples. Moreover, the aquarium plants, in association with their bacterial communities might be the most important component in the aquarium, as indicated by their highest bacterial richness and diversity. This study adds to our understanding on the differences in the microbiome of different components and their possible contributions to nutrient cycling in a self-sustaining aquarium.