Water Microbiota in Greenhouses With Soilless Cultures of Tomato by Metabarcoding and Culture-Dependent Approaches

Water supply, in hydroponic greenhouses, can originate from groundwater, surface water or rainwater stored in open tanks. To limit contamination of water supply, several methods have been used including active and passive methods such as slow filtration techniques which consist in passing the nutrient solutions slowly through filters. The purpose of this study was to describe the microbiota associated with water sampled before entering greenhouses and in recirculating nutrient solutions, either before or after running through a biofiltration system. Metabarcoding analysis revealed that water ecosystems were unique niches for diverse bacterial and fungal communities. Microbial composition varied greatly across storage conditions (groundwater vs. rainwater) and among greenhouses, suggesting that water microbiota is site- and storage-condition-specific. Nonetheless, we found that microbiota structure in open-stored water (either coming from ground or rain) shared a higher degree of similarity than with water directly pumped out of the ground. Open-stored waters were characterized by predominant taxa, notably those involved in aerobic chemoheterotrophy, such as the Sphingomonadaceae and Hyphomicrobiaceae families. Water directly collected from the ground showed the lowest levels of fungal and bacterial richness while also characterized by a significantly higher level of bacterial equitability and an enrichment in taxa involved in N-cycling. Slow filtration allowed reducing cultivable bacterial loads as well Pythium spp. and Fusarium oxysporum propagules, based on culture-dependent results, without compromising microbiota richness and diversity. Although compositional structure was similar following biofiltration, significant differences in bacterial (but not fungal) taxa abundance were reported, with primarily an enrichment of Chelativorans, Mycobacterium, and Gemmata as well as a depletion of Rhodobacter, Aminobacter, and Ellin329. The exact mechanisms by which such taxa would be favored at the expense of other remained unknown. Besides the accurate description of microbiota found in water at both taxonomical and predicted functional levels, our study allowed comparing the water microbiota between various storage system and following biofiltration. Although preliminary, our results provide a first insight into the potential microbial diversity, which can increase ecosystem functionality.