New lipid-based proxies for past cyanobacterial abundance

Throughout the past century, eutrophication and climate change have strongly impacted temperate lakes, resulting in greater algal productivity and shifts in phytoplankton community composition. In many lakes cyanobacterial blooms have become more common, reducing water quality, negatively impacting aquatic food webs, and affecting the cycling of carbon and other nutrients. Appropriate management of aquatic systems to mitigate and avoid these problems is informed by monitoring data, but direct observations are often limited to recent decades. Paleolimnological approaches can extend the observational window and contextualize long-term changes of algal communities in response to climatic and environmental forcings. However, it remains challenging to reconstruct changes in algal productivity and community assembly, particularly the relative abundance of cyanobacteria to eukaryotic algae, throughout the geologic past.

Here, we present two recently developed lipid-based proxies that can be used to reconstruct broad shifts in algal community composition: (1) the Phytol:Sterol Index (PSI), which represents the relative abundance of the chlorophyll side-chain, phytol, to phytosterols from eukaryotic algae and (2) hydrogen isotope offsets between phytol and the common membrane lipid C16 fatty acid (δ²H_{C16:0 Acid/Phytol}), which is higher for lipids produced by cyanobacteria and green algae than for other eukaryotic algae. We demonstrate the utility of these proxies in a collection of short sediment cores from lakes in the Swiss Plateau (Murtensee, Greifensee, and two hydrologically distinct basins of Zugersee), all of which experienced extreme eutrophication in the mid- to late 20^th century, followed by partial recovery to lower nutrient levels. We found significant changes in lipid distributions coincident with the main period of increasing total phosphorus inputs. During this time, PSI increased in all four lake records, indicating that more of the algal biomass accumulating in the sediments was derived from cyanobacteria. In Murtensee, PSI and δ²H_{C16:0 Acid/Phytol} co-varied, while in Greifensee the initial increase in cyanobacteria was followed by a period of low PSI and high δ²H_{C16:0 Acid/Phytol} values, consistent with observations of abundant green algae during this later period.

We cross-compared our lipid biomarker data with cyanobacterial and plastid 23S rRNA amplicon sequencing variants (ASVs) of DNA extracted from the cores. In general, there was good agreement between PSI and the abundance of cyanobacterial ASVs. However, during periods when the cyanobacterial DNA was primarily from small-celled taxa such as Synechococcus, such as the early 20^th century in Murtensee, PSI was low relative to the abundance of cyanobacterial ASVs. This suggests that small but numerous cyanobacteria might be overrepresented in sedimentary DNA relative to their biomass, likely related to the polyploidy of their chromosomes.

Overall, sedimentary PSI appears to be a robust and analytically straight-forward indicator of cyanobacterial abundance. Due to the greater mass of phytol needed for δ²H measurements, chromatographical challenges can limit the application of δ²H_{C16:0 Acid/Phytol} in some sediments, such as those from Zugersee. The combination of these new lipid-based proxies with other tools, including sedimentary DNA, pigments, and microfossil analyses can provide the most comprehensive picture of past algal community composition.