Rates and regulation of anaerobic ammonium oxidation and denitrification in the Black Sea

We performed incubation experiments with 15N-labeled nitrogen compounds to investigate the vertical distribution of pathways of N2 production through the suboxic zone of the central Black Sea and the impact of oxygen and sulfide on the anammox process. Anammox rates increased with depth through the upper suboxic zone and reached a maximum of ,11 nmol N2 L21 d21 at the sharp interface between nitrate and ammonium, below which rates decreased toward the depth of sulfide accumulation. Heterotrophic denitrification was not detected, and therefore anammox was the prevailing sink for fixed nitrogen in the central Black Sea. In incubations with low oxygen concentrations, anammox activity was only partially inhibited, with a decrease in anammox rates to ,70% and 50% of the anoxic level at ,3.5 and ,8 mmol L21 O2, respectively, and complete inhibition at ,13.5 mmol L21 O2. Thus, the anammox process is not constrained to anoxic marine waters. This increases the volume of the major open-ocean oxygen-deficient zones, where anammox is potentially active, which has important implications for the contribution of anammox to the marine nitrogen cycle. We observed an inhibitory effect of micromolar sulfide concentrations on anammox activity, indicating that the vertical and likely horizontal distribution of active anammox bacteria is constrained to nonsulfidic water layers, which may explain the absence of the process in sulfidic basins with no suboxic zone. The discovery of anaerobic ammonium oxidation (anammox) in wastewater treatment systems (Mulder et al. 1995) and natural aquatic environments (Thamdrup and Dalsgaard 2002) resolved the mystery of ammonium deficiency in anoxic waters and challenged the preeminence of microbial denitrification as the only significant pathway for the removal of fixed nitrogen in the oceans (Devol 2003; Dalsgaard et al. 2005; Kuypers et al. 2006). Anaerobic ammonium oxidation with nitrite as an electron acceptor is mediated by a monophyletic group of bacteria that branches deeply in the phylum Planctomycetes (Strous et al. 1999; Schmid et al. 2003). The one-to-one coupling of nitrogen from ammonium and nitrite into gaseous N2, NH z4 + NO { 2 R N2 + 2H2O (Van De Graaf et al. 1995), distinguishes the anammox process from denitrification, where two molecules of nitrate are combined to N2 in a stepwise pathway (2NO 3 R 2NO { 2 R 2NO R N2O R N2) (Zumft 1997). Although a number of studies demonstrate the importance of anammox bacteria in the biological nitrogen cycle (Dalsgaard et al. 2005; Kuypers et al. 2006; Thamdrup et al. 2006), little is known about the main factors that control the distribution and magnitude of the process. Expectedly, oxygen is such an important regulator. Experimental work with enrichments of anammox bacteria from laboratory wastewater bioreactors has shown that anammox activity is reversibly inhibited by oxygen levels as low as 1 mmol L21 (Strous et al. 1997), indicating that the process is active only under strictly anoxic conditions. Still, in the Benguela upwelling system off Namibia, the observed dominance of anammox was suggested to result from anammox being less sensitive than denitrification toward oxygen (Kuypers et al. 2005). There, anammox bacteria were immediately active when waters containing up to 9 mmol L21 of oxygen in situ were incubated under anoxic conditions, while denitrification remained undetectable. These observations raise the question whether anammox could indeed be active at low oxygen levels in the oceans. The sensitivity of anammox bacteria toward oxygen is an important issue for un1 Present address: Max Planck Institute for Marine Microbiology, Nutrient Group, Celsiusstr. 1, 28359 Bremen, Germany. 2 Corresponding author ([email protected]). Acknowledgments We sincerely thank Carsten Schubert for organizing the research expedition and the captain and the crew on board the R/V Professor Vodyanitskiy for their assistance and enthusiastic support. We are grateful to Daniel McGinnis for generously providing CTD and oxygen data and G. Klockgether for valuable analytical assistance. We thank two anonymous reviewers for their helpful evaluations of the manuscript. M.M.J. and B.T. were supported by the Danish National Research Council and the Danish National Research Foundation, M.M. M.K. and G.L. by the Max Planck Society. Limnol. Oceanogr., 53(1), 2008, 23–36 E 2008, by the American Society of Limnology and Oceanography, Inc.