Freshwater lakes represent large methane sources that, in contrast to the Ocean, significantly contribute to non-anthropogenic methane emissions to the atmosphere. Particularly mixed lakes are major methane emitters, while permanently and seasonally stratified lakes with anoxic bottom waters are often characterized by strongly reduced methane emissions. The causes for this reduced methane flux from anoxic lake waters are not fully understood. Here we identified the microorganisms and processes responsible for the near complete consumption of methane in the anoxic waters of a permanently stratified lake, Lago di Cadagno. Interestingly, known anaerobic methanotrophs could not be detected in these waters. Instead, we found abundant gamma-proteobacterial aerobic methane-oxidizing bacteria active in the anoxic waters. In vitro incubations revealed that, among all the tested potential electron acceptors, only the addition of oxygen enhanced the rates of methane oxidation. An equally pronounced stimulation was also observed when the anoxic water samples were incubated in the light. Our combined results from molecular, biogeochemical and single-cell analyses indicate that methane removal at the anoxic chemocline of Lago di Cadagno is due to true aerobic oxidation of methane fuelled by in situ oxygen production by photosynthetic algae. A similar mechanism could be active in seasonally stratified lakes and marine basins such as the Black Sea, where light penetrates to the anoxic chemocline. Given the widespread occurrence of seasonally stratified anoxic lakes, aerobic methane oxidation coupled to oxygenic photosynthesis might have an important but so far neglected role in methane emissions from lakes.

Distributions of isoprenoid (isoGDGT) and branched glycerol dialkyl glycerol tetraethers (brGDGTs) were measured in the water column and sediments of the eutrophic Lake Lugano and the oligotrophic Lake Brienz, Switzerland. Absolute concentrations of isoprenoid, i.e. archaeal GDGTs, were highest in the euphotic zone of both lakes, as well as in sediments deposited at times when lake eutrophication occurred. This indicates that GDGT concentrations may be used as indicators for primary productivity. Both lakes, including the anoxic bottom water of Lake Lugano, are characterised by GDGT distributions typical for group I Crenarchaeota with GDGT-0/crenarchaeol ratios of around 1. Comparison of the distribution of brGDGTs with isoGDGTs and other terrestrial biomarkers throughout the Lake Lugano water column, together with CBT/MBT-derived temperatures that resemble that of the lake, suggest significant in situ production. BIT index values for Lake Brienz sediments (ca. 0.4) were significantly higher than water column values (ca. 0.1), most probably because terrestrial run off events were not captured during water sampling. TEX₈₆ – derived temperatures reflect surface water conditions to within a few degrees, while lower values obtained from deeper water layers suggest a contribution of in situ produced isoGDGTs. For both lake sediments, TEX₈₆-derived temperatures could be matched reasonably with mean annual lake surface water temperature variation, albeit with a larger offset for Lake Lugano. This suggests that absolute temperatures can only be reconstructed from lake sediments for which a local calibration is known.

Here we provide the first direct evidence for the anammox process (anaerobic ammonium oxidation) in a lacustrine system, Lake Tanganyika, the second largest lake in the world. Incubations with ¹⁵N labelled nitrate showed that anammox occurred in the suboxic water layer at 100-110 m water depth. Anammox rates up to 10 nM N₂ h^-1 are comparable to those reported for the marine water column. Up to ∼13% of produced N₂ could be attributed to the anammox process whereas the remainder was related to denitrification. Typical lipid biomarkers characteristic of f anammox bacteria were found in filtered water from the depths where anammox occurred, thus supporting the presence of anammox bacteria. Further evidence is provided by fluorescence in situ hybridization (FISH), revealing up to 13 000 anammox bacteria cells per ml or 1.4% of all DAPI (4′-6-Diamidino-2-phenylindole)-stained cells. Phylogenetic analyses of partial 16S rRNA genes indicated the presence of sequences most closely related to the known anammox bacterium Candidatus 'Scalindua brodae' (95.7% similarity). Using the incubation results, a total loss of 0.2 Tg N₂ per year linked to anammox was estimated for the Northern basin of Lake Tanganyika.