Biofilm communities in biofilters are structuring the rearing water microbiota in recirculating aquaculture systems (RAS) through dispersal of bacteria

Nitrifying biofilters are mandatory in recirculating aquaculture systems (RAS). In addition to nitrification, they can influence the rearing microbiota through consumption of organic matter and the release of bacteria, thereby influencing the microbial ecology of the system. However, these aspects of biofilters are poorly understood. The aim of this study was to investigate the potential direct influence of the biofilter microbiota on the rearing water microbiota through the release of bacteria into the rearing water. We carried out three similar independent experiments in lab-scale RAS with Atlantic salmon fry, each comprising two identical RAS that differed only in the source of the biofilter biofilm carriers. They originated from geographically distant commercial RAS facilities and were known to harbor distinct biofilm communities. The experimental design included the exchange of biofilters between the two RAS units (after one month in Experiment 1, and 3 weeks in Experiment 2 and 3). Biofilm carriers and rearing water were sampled at the start of the experiment, directly before the biofilter exchange, and at the end of the experiments. Bacterial communities were characterized by 16S rRNA gene amplicon sequencing. In all three experiments, the biofilter communities were relatively stable and differed significantly between the two RAS units throughout the experiments, independent of in which RAS they were located. The biofilter communities were also significantly different from rearing water microbiota within the respective RAS at all sampling timepoints in all experiments. Prior to biofilter exchanges, the rearing water microbiota differed significantly between the two RAS, potentially reflecting an effect of the distinct biofilter microbiotas. However, this observation could also be a consequence of stochastic processes. To examine the potential influence of biofilter on the rearing water microbiota, we exchanged the biofilters between the two RAS. After this exchange, the rearing water microbiota became significantly more similar to the newly introduced biofilters in all experiments, with mean Bray–Curtis similarities increasing from 0.05 to 0.17 to 0.21–0.30 across all experiments. Concurrently, the Bray–Curtis similarity between the microbiota of the rearing water and the biofilter present prior to the exchange, significantly decreased in all RAS, with mean values declining from 0.24 to 0.41 to 0.09–0.22. Moreover, after the biofilter exchanges, we observed the appearance of specific ASVs in the rearing water that originated from the newly introduced biofilter. These observations demonstrate that biofilter biofilms exert a direct influence on rearing water microbiota in RAS through dispersal of bacteria.