Sedimentary DNA records long-term changes in a lake bacterial community in response to varying nutrient availability

Microbial communities play important roles in lake ecosystems and are sensitive to environmental change. However, our understanding of their responses to long-term change such as eutrophication is limited as long-term lake monitoring is rare, and traditional palaeolimnological (e.g. microfossil) techniques are restricted to organisms with well-preserved structures. Sedimentary DNA (sedDNA) is a promising technique to reconstruct past microbial communities in sediments, but taphonomic processes and the ability of sedDNA to record bacterial pelagic history accurately is largely unknown. Here we sequenced the 16S rRNA gene in triplicate sediment cores from Esthwaite Water (English Lake District) which has concurrent long-term monitoring and observational data. The sediment record spanned 113 years and included an episode of increased nutrient availability from the 1970s, followed by a more recent decline. Trends in bacterial community composition were homogenous among the three sediment cores, and there was no evidence of a DNA degradation bias. Cyanobacterial richness in the sediment cores correlated significantly with that of cyanobacteria in a 65-year microscopy-based monitoring record, and known pelagic bacterial taxa were detected in the sediment. sedDNA revealed distinct shifts in community composition in response to changing conditions. The relative abundance of cyanobacteria closely reflected nutrient enrichment, and Proteobacteria, Bacteroidetes and Verrucomicrobia were more abundant in recent sediments, while Chloroflexi, Firmicutes, Acidobacteria, Nitrospirae, Spirochaetes and Planctomycetes declined in more recent sediments. Following lake restoration efforts to reduce nutrient enrichment, the relative abundance of cyanobacteria returned to pre-1970 levels, but the bacterial community did not fully recover from the period of intense eutrophication within the time-scale of our study. These results suggest that sedDNA is an effective approach to reconstruct lake microbial communities and their responses to past conditions.