Can we use functional genetics to constrain the fate of nitrate in estuaries?

Estuaries, the interface between land and ocean, are among the most heavily populated and perturbed parts of the world. A major concern is the input of nitrogen (N) which is predicted to increase in the future. Molecular indicators for N pathways are increasingly measured and analysed, but how robust and universal are they to quantify estuarine ‘health’? This talk focuses on how we can use the diversity and community composition of sediment bacteria to constrain N removal (via denitrification; nirS/nirK amplicons) or N retention (via DNRA; nrfA amplicon) in 11 temperate estuaries covering 4 types of land use in Australia. We noted significant differences in the community composition for all assessed N-cycling clades between the 11 estuaries (at a zero-radius OTU level, but also at 88% similarity level to emphasize ecologically relevant sequences).

N removal and retention rates were measured, and regression models showed that NO₃^-, Fe²⁺ and NH₄⁺ were the best predictor variables. In addition to the high-throughput sequencing we also quantified 16S DNA and cDNA copy numbers via qPCR. The integration of these results into regression models improved the predictability of N removal rates. The same parameters also structured the N-cycling communities. Interestingly, no clear bacterial indicators were found between the three estuaries with high N removal rates. Network analyses were used to identify functional N-cycling OTU hubs (indicating network key stone species) in different land uses. Urban sites revealed to have a higher number of hubs compared to the pristine sites. Yet, the network modularity (controversially linked to resilience) in the pristine estuaries were higher compared to the urban sites, suggesting a more effective ‘information transfer’ between hubs in the pristine sites.

We conclude that, from our data set, the predictive capability of 16S qPCR results, rather than high-throughput sequencing of functional N-genes, revealed a more cost and time effective way to better predict the spatial distribution of relative N removal and retention rates.