Phosphorus is an essential nutrient for life. In the ocean, phosphorus burial regulates marine primary production1, 2. Phosphorus is removed from the ocean by sedimentation of organic matter, and the subsequent conversion of organic phosphorus to phosphate minerals such as apatite, and ultimately phosphorite deposits3, 4. Bacteria are thought to mediate these processes5, but the mechanism of sequestration has remained unclear. Here, we present results from laboratory incubations in which we labelled organic-rich sediments from the Benguela upwelling system, Namibia, with a 33P-radiotracer, and tracked the fate of the phosphorus. We show that under both anoxic and oxic conditions, large sulphide-oxidizing bacteria accumulate 33P in their cells, and catalyse the nearly instantaneous conversion of phosphate to apatite. Apatite formation was greatest under anoxic conditions. Nutrient analyses of Namibian upwelling waters and sediments suggest that the rate of phosphate-to-apatite conversion beneath anoxic bottom waters exceeds the rate of phosphorus release during organic matter mineralization in the upper sediment layers. We suggest that bacterial apatite formation is a significant phosphorus sink under anoxic bottom-water conditions. Expanding oxygen minimum zones are projected in simulations of future climate change6, potentially increasing sequestration of marine phosphate, and restricting marine productivity.

MethodsSurficial sediment was sampled during research cruise M76-2 with RV Meteor with a multi-core sampler. Sediment slurries were prepared with bottom water from 1 cm slices of the topmost 6 cm, spiked with carrier-free 33Pi and incubated under anoxic and oxic conditions for 48 h at 4°C, together with killed controls. After incubation, samples were fixed with zinc chloride, immediately frozen and transported to the laboratory for sequential extraction of sedimentary Pi pools. For each of the sequential fractions, we recorded Pi concentrations by photometry and inductively coupled plasma optical emission spectrometry and calculated the size of individual P pools. The redistribution of 33Pi radiolabel among the different P fractions was quantified by liquid scintillation counting of the extractant solutions, after recalculation of the initial activity assuming exponential radioactive decay.

Supplement to: Goldhammer, Tobias; Brüchert, Volker; Ferdelman, Timothy G; Zabel, Matthias (2010): Microbial sequestration of phosphorus in anoxic upwelling sediments. Nature Geoscience, 3(8), 557-561