J. G. Kuenen is full professor in general and applied Microbiology and head of the Department of Biotechnology, Delft University of Technology, Delft, and M. S. M. Jetten is full professor in microbial ecology in the Department of Microbiology, University of Nijmegen, Nijmegen, the Netherlands.
Article Links:
Links to Other ASM Pages:
Extraordinary Anaerobic Ammonium-Oxidizing Bacteria
Unusual microbial metabolic process, although slow, can remove ammonium from wastewater in a single, oxygen-limited step
J. G. Kuenen and M. S. M. Jetten
What would you do if someone from industry approached you with a claim that ammonium was disappearing in his anoxic wastewater treatment plant at the expense of nitrite? Some would dismiss this claim as "not in the textbook." Others might remember a prediction by Engelbert Broda some 20 years ago that this should be possible on thermodynamic grounds and challenge the colleague to prove that Broda was right. Well, what happened was that we were asked to take up the challenge. We accepted, not guessing that it would take 10 years to crack the problem of anaerobic ammonium oxidation, later called the anammox process.
In facing the challenge, the first thing we did was to reevaluate the nitrogen balance of the wastewater treatment plant by observing the fate of 15N-labeled NH4+ that we added to the system. On the basis of such analysis, our Ph.D. student Astrid van de Graaf established unequivocally that ammonium combines quantitatively with available 14N-nitrite to give mixed 14-15N2 dinitrogen gas. Soon after, we proved that the reaction is sensitive to heat sterilization, gamma irradiation, and several uncouplers, has a pH optimum at about 8, and a temperature optimum at 37ºC.
Proof that the Reaction Is Biological Launched a Difficult Microbiological Analysis
Although those first studies proved ammonium oxidation is a biological process, they also marked the start of our struggle to explore the microbiological nature of this process. Even today we cannot grow the bacterium responsible for this process in pure culture. Nobody can. But we know its identity: "Candidatus Brocadia anammoxidans." "Brocadia" refers to the place of its discovery, a Gist-Brocades pilot plant where Arnold Mulder, the industrial colleague who first detected the reaction, was working. This bacterium belongs to the exotic order of the Planctomycetales (Fig. 1), organisms that have no peptidoglycan in their cell walls, possess unusual features on the cell surface known as crateriform structures, reproduce by budding, and, most exciting of all, appear to be compartmentalized in membranous structures reminiscent of those found among eukaryotes.
The only way in which we could enrich Candidatus B. anammoxidans was through use of flow-through systems. Other more commonly used techniques, such as batch cultures, proved fruitless for studying these microorganisms. However, once we established the flow-through approach for growing them, we had a practical means for fully characterizing the organism.
Specifically, Marc Strous, at that time working on his Ph.D., developed a successful cultivation system based on a 15-liter anaerobic sequencing batch reactor (SBR) (Fig. 2), which was fed with a mineral medium containing only ammonium, nitrite ( NH4+ + NO2-, 30 mmol/liter each), and bicarbonate. Every day the SBR goes through 2 cycles: each cycle starts with 11.5 hours of gentle stirring, while continuously feeding the medium. After 11.5 hours of feeding, we stop the gentle stirring, let the bacterial flocs settle for 10 minutes in the reactor, and then remove the supernatant in 20 minutes. In this way we efficiently retain biomass; moreover, cells grow exponentially to form a very dense culture of 5 x 1010 cells/ml. (Fig. 2). Although not pure, this culture contains more than 70% coccoid cells of Candidatus B. anammoxidans and is brightly red from the cytochromes that these bacteria contain. The culture grows autotrophically with CO2 as the only carbon source.
The overall anammox process consists of the following two reactions:
Catabolic reaction:
NH4+ + NO2- ® N2 + 2H2O (1)Anabolic reaction:
CO2 +2NO2- + H2O ® CH2O (biomass) + 2NO3- (2)Thus, NO2- is not only the e-acceptor in reaction (1) but also the e-donor for CO2 fixation (2). In fact, the latter reaction is analogous to the carbon assimilation of aerobic, chemolithoautotrophic, nitrite-oxidizing bacteria. Indeed, these reactions—one carried out by aerobic ammonium-oxidizing bacteria such as Nitrosomonas europaea and the other by Candidatus B. anammoxidans—have many features in common (see table).
For instance, the thermodynamics and biomass yields are very similar. However, the big difference is in their aerobic and anaerobic rates of ammonium oxidation and in their maximum specific growth rates (µmax). Candidatus B. anammoxidans is a very slowly growing obligately anaerobic bacterium, with doubling times of 10-14 days. The aerobic nitrifiers, such as Nitrosomonas, are facultatively anaerobic, growing relatively fast under oxic conditions (td = 0.7 day). These organisms can also metabolize ammonium and nitrite anaerobically, but their rate is at least 30 times lower than that of Candidatus B. anammoxidans.
Mechanism of Anaerobic Ammonium Oxidation: a Working Model
Experiments with 15N-labeled nitrogen compounds indicated that in "Candidatus B. anammoxidans," ammonium and hydroxylamine (NH2OH) combine during a condensation reaction. When we feed NH2OH to the culture, very surprisingly hydrazine (N2H4) transiently accumulates, indicating that this compound is an intermediate. To our knowledge, this rocket fuel is not known as a free intermediate in any other biological system.
Our finding this hydrazine intermediate led to our working model (Fig. 3) for energy transduction and the buildup of an electrochemical gradient in this microorganism. In this scheme, the hydrazine-forming enzyme, hydrazine hydrolase (HZF), remains hypothetical. However, an enzyme capable of oxidizing hydrazine (HZO) accounts for 10% of the total protein of Candidatus B. anammoxidans enrichment cultures, according to our Ph.D. student Jos Schalk, who purified the enzyme to homogeneity.
At first we thought this enzyme resembles the hydroxyamine oxidoreductase (HAO) of N. europaea. However, the HZO/HAO of Candidatus B. anammoxidans has a lower molecular mass, and peptides obtained after trypsin digestion show no homology with the proteins represented in the database of N. europaea or any other organisms. The enzyme of Candidatus B. anammoxidans is active only under anoxic conditions and produces NO and N2O from hydroxylamine, whereas the N. europaea enzyme oxidizes hydroxylamine aerobically to nitrite. Both enzyme systems can oxidize hydrazine to N2.
In terms of our model, HZO/HAO would function as the hydrazine-oxidizing enzyme (hydrazine-cytochrome c oxidoreductase). HZO/HAO contains 3 a subunits and carries approximately 24 c-type cytochromes. Similar to the N. europaea enzyme, there is an additional very characteristic cytochrome peak at 468 nm that can be (partly) reduced by hydrazine and hydroxylamine. The high HZO/HAO content of the Candidatus B. anammoxidans culture also explains the bright red color of the biomass (see photograph). According to our model (Fig. 3), four electrons from hydrazine would be needed to reduce the nitrite to hydroxylamine. Thus a closed cyclic system would lead to the buildup of a proton-motive force across the membrane, which, in turn, would allow ATP synthesis. We are continuing to study this pathway.
Frustrations with Efforts To Obtain Pure Cultures
Frustrated in all our attempts to purify Candidatus B. anammoxidans, we tried using DNA-based methods to learn more about it. The standard way for obtaining pure DNA from an organism growing in mixed culture is to isolate RNA or DNA from the mix and then amplify and clone the 16S rDNA via reverse transcriptase and polymerase chain reaction, or RT-PCR. However, the presumed universal 16S rDNA primer set was not so universal—instead, as we found out later, having three mismatches with that of Candidatus B. anammoxidans. Hence, we did not find the expected (70%) dominant 16S-rDNA on gels or in our clone-libraries.
To overcome this difficulty, Katinka van de Pas-Schoonen and Marc Strous painstakingly developed a Percoll density gradient centrifugation procedure to separate cells of Candidatus B. anammoxidans from other members of the community. Eventually, they obtained 99.6% pure cultures. From the DNA of these purified cultures, we identified the dominant bacterium as a member of Planctomycetales, branching very deep within this lineage (Fig. 1). The purified individual cells responded to eight specific fluorescent in situ hybridization (FISH) probes that were designed on the basis of the 16S rDNA-gene sequence. Very importantly, the purified cells also showed active anaerobic ammonium oxidation with nitrite as well as the expected simultaneous CO2 fixation. In this way we unequivocally established that Candidatus B. anammoxidans is responsible for the anammox reaction.
Anammox activity of the purified suspensions of Candidatus B. anammoxidans could be restored only by reconcentrating cells back to the original density of 1010-1011 cells/ml and adding micromolar quantities of either hydroxylamine or hydrazine. In trying to explain the need to concentrate the cells, we consider it very unlikely that the 1 in 500 contaminating cells contribute significantly to the anammox activity of the purified preparations. For one thing, they would need to have enzymes with extremely high activities. Furthermore, although we can detect a few contaminating cells, they are very diverse in terms of morphology.
However, although we believe that the purified Candidatus B. anammoxidans is responsible for the anammox reaction and for fixing CO2, we also believe that other bacteria in the community might play a vital role during prolonged growth of this organism, perhaps providing it vitamins or removing toxic intermediates. It is conceivable that cell-cell communication through molecular signal compounds is required for activity. Another possibility is that the hydrazine intermediate can diffuse out of the cells relatively easily and needs to be maintained at a critical concentration—hence, the need for high cell densities.
Diversity—Is Everything Everywhere?
Although other investigators initially could not duplicate our enrichment culture-based findings, once they took the slow growth rate into account, enrichment cultures with similar but not identical bacteria were established in several different laboratories. The first confirmations of the anammox reaction came from a team in Germany who used samples taken from pilot-wastewater plants in which there was high ammonium loading.
To characterize the microorganisms responsible for those reactions, we and our collaborators Markus Schmid and Michael Wagner in the Department of Microbiology of the Technical University of Munich, Germany, used the same molecular probes that helped us earlier to establish the identity of Candidatus B. anammoxidans. This series of 16S-rDNA-probe-binding experiments established that a range of similar, but not identical, bacteria can carry out the anammox reaction.
Indeed, the anammox bacteria from the wastewater treatment plants in Germany are not closely related to Candidatus B. anammoxidans. Instead, they appear to belong to a separate monophyletic cluster (90% similarity to that of Candidatus B. anammoxidans), which was given the name "Candidatus Kuenenia stuttgartiensis." FISH probes indicate that these bacteria are dominant in communities responsible for high nitrogen turnover in wastewater treatment plants. Follow-up studies with functional probes could help us to learn more about the environmental variables that give rise to such extensive diversity of anammox bacteria among these disparate communities.
Morphology and Electron Microscopy
With John Fuerst and his colleagues at the University of Queensland in Brisbane, Australia, we showed that the Candidatus B. anammoxidans cells contain at least three membrane-bound compartments (Fig. 4): the outer compartment containing the paryphoplasm; a second inner compartment, containing the ribosomes and the condensed, or nucleoid, DNA; and a third inner-membrane-bound compartment making up 30-60% of the Candidatus B. anammoxidans volume.
Immunogold-labeling experiments indictate that this third compartment, or organelle, contains the HZO/HAO enzyme, leading us to name it the anammoxosome. Although the function of all three compartments remains uncertain, we speculate that the anammoxosome generates an electrochemical gradient across its membrane (Fig. 3). The big anammoxosome gives a special feature to FISH-stained cells, setting off the ribosomes in the compartment surrounding it in a fluorescent circle. Because FISH probes reveal a similar fluorescent ring in Candidatus K. stuttgartiensis, we suspect it has a similar ultrastructure.
Ecology, Combined Anoxic and Oxic Ammonium Oxidation
The wide distribution of bacteria capable of oxidizing ammonium under anoxic conditions raises questions about their natural habitat and about conditions that favor their growth in wastewater treatment systems. We know from experiments that the activity of the anammox enrichment cultures is very sensitive to oxygen. Even 0.5 % of air saturation (1 µM oxygen) completely stops the reaction.
However, low oxygen levels block the reaction without killing the organisms. After anoxic conditions are restored, cultures immediately resume oxidizing ammonium at the expense of nitrite. This behavior is in sharp contrast to that of cultures of aerobic nitrifiers such as N. europaea, which are capable of very slow anaerobic metabolism. Even if these organisms are provided with 25 ppm of nitrogen dioxide to stimulate anaerobic metabolism, their rate is 30 times lower (see table) than that of the Candidatus B. anammoxidans culture. Notably, in our anammox enrichment cultures we consistently detect low levels of aerobic nitrifiers, confirming earlier observations that these organisms can survive anoxic condition for prolonged periods.
Based on such observations, the oxic-anoxic interface would appear to be the preferred habitat for Candidatus B. anammoxidans. At such sites, oxygen-limited cells of ammonium oxidizers such as N. europaea would be available to provide nitrite that, in the absence of oxygen, could be used by the anaerobic ammonium oxidizers to consume remaining ammonium.
With this in mind, Olav Sliekers and Katie Third in our laboratory created a coculture of the highly enriched anammox culture with aerobic nitrifiers in a sequencing batch reactor. To do so, they provided increasing quantities of limiting oxygen to the Candidatus B. anammoxidans enrichment while ensuring that all oxygen is immediately consumed by the culture. These conditions lead to a gradual increase of aerobic nitrifiers in the culture, eventually reaching a stage in which 30 mmol/liter of ammonium is converted into mainly (>80%) dinitrogen gas and some nitrate according to equation (3)
2.5 NH4+ + 2.1 O2 ® 0.2 NO3- + 1.15 N2 + 3.6 H2O + 2.8 H+ (3)
Part (1.15/2.5 * 30 mmol/liter) of the ammonium presumably is oxidized to nitrite by the aerobic nitrifiers and immediately converted to dinitrogen gas by the anammox bacteria.An additional portion (0.2/2.5 * 30 mmol/liter) is oxidized to nitrite by the same aerobic nitrifiers and then on to nitrate by the anammox bacteria to provide the reducing power for CO2-fixation by the latter organisms [see reaction (2)]. When biomass from such cultures is made aerobic (up to 15% air saturation), the anammox reaction stops, and ammonium is converted to nitrite but not to nitrate. Apparently, when oxygen is limited, the aerobic nitrite oxidizers are not enriched.
Tests with FISH probes that are specific for the aerobic and anaerobic nitrifiers confirm that Candidatus B. anammoxidans at first dominates the culture for more than 70%, but, with time, Nitrosomonas type organisms take on an increasingly substantial role. We did not detect aerobic nitrite-oxidizing bacteria, such as Nitrobacter or Nitrospira species, suggesting that the combined Nitrosomonas/Brocadia cultures outcompete the nitrite oxidizers for both oxygen and nitrite. With Renee Wijffels and his colleagues at the Department of Process Engineering, Wageningen University and Research Center, the Netherlands, we immobilized both anammox and nitrifying cells together in alginate beads, providing a protoype of an air loop reactor capable of converting ammonium in nitrogen gas.
A similar set of conditions likely exists in nature—for instance, at the oxic-anoxic interface between water bodies and sediments. Presumably, ammonium that forms under anoxic conditions would diffuse to the interface where it would be oxidized, in part, to nitrite. The nitrite could diffuse back into the anoxic zone and be reduced with the remaining ammonium.
These observations also could prove useful for redesigning wastewater treatment plants. Thus, for example, it may be possible to remove ammonium from wastewater in a single, oxygen-limited step by means of a recently patented process, "completely autotrophic N-removal over nitrite," or CANON. This name also alludes to the subtle interplay of the oxic- and anoxic- ammonium oxidizing reactions, which can run simultaneously and in tune.
ACKNOWLEDGMENTS
The research on anaerobic ammonium oxidation was financially supported by the Foundation for Applied Sciences (STW), the Foundation of Applied Water Research (STOWA), the Netherlands Foundation for Life Sciences (NWO-ALW), the Royal Netherlands Academy of Arts and Sciences (KNAW), Gist-brocades, DSM, Paques BV and Grontmij consultants. We gratefully acknowledge the contributions of various coworkers and students over the years.
SUGGESTED READING
Fuerst, J. A. 1995. The planctomycetes: emerging models for microbial ecology, evolution and cell biology. Microbiology 141:1493-1506.
Jetten, M. S. M., M. Strous, K.T. van de Pas-Schoonen, J. Schalk, L. van Dongen, A. A. van de Graaf, S. Logemann, G. Muyzer, M. C. M. van Loosdrecht, and J. G. Kuenen. 1999. The anaerobic oxidation of ammonium. FEMS Microbiol. Reviews 22:421-437.
Jetten, M. S. M., M. Wagner, J.A. Fuerst, M. C. M. van Loosdrecht, J. G. Kuenen, and M. Strous. 2001. Microbiology and application of the anaerobic ammonium oxidation (anammox) process. Curr. Opinion Biotechnol. 12:283-288.
Lindsay, M. R., R. I. Webb, M. Strous, M. S. M. Jetten, M. K. Butler, and J. A. Fuerst. 2001. Cell compartmentalization in planctomycetes: novel types of structural organization for the bacterial cell. Arch. Microbiol. 175:413-429.
Schalk, J., S. Devries, J. G. Kuenen, and M. S. M. Jetten. 2000. A novel hydroxylamine oxidoreductase involved in the Anammox process. Biochemistry 39:5405-5412.
Schmid, M., U. Twachtmann, M. Klein, M. Strous, S. Juretschko, M.S.M. Jetten, J. Metzger, K. H. Schleifer, and M. Wagner. 2000. Molecular evidence for genus level diversity of bacteria capable of catalyzing anaerobic ammonium oxidation. Syst. Appl. Microbiol. 23:93-106.
Schmidt, I., and E. Bock. 1997. Anaerobic ammonia oxidation with nitrogen dioxide by Nitrosomonas eutropha. Arch. Microbiol. 167:106-111.
Strous, M., J. J. Heijnen, J. G. Kuenen, and M. S. M. Jetten. 1998. The sequencing batch reactor as a powerful tool to study very slowly growing microorganisms. Appl. Microbiol. Biotechnol. 50:589-596.
Strous, M., J. G. Kuenen, and M. S. M. Jetten. 1999. Key physiology of anaerobic ammonium oxidation. Appl. Environ. Microbiol. 65:3248-3250.
Strous, M., J. A. Fuerst, E. H. M. Kramer, S. Logemann, G. Muyzer, K. T. van de Pas-Schoonen, R. Webb, J. G. Kuenen, and M. S. M. Jetten. 1999. Missing lithotroph identified as new planctomycete. Nature 400:446-449.
Van Dongen, L., M. S. M. Jetten, and M. C. M. Van Loosdrecht. 2001. The Sharon-Annamox process for treatment of ammonium rich waste water. Water Sci. Technol. 44:153-160.
PATENTS
Mulder, A. 1992. Anoxic Ammonium Oxidation US patent 427849(5078884).
Dijkman, H. and M. Strous. 1999. Process for ammonia removal from wastewater. Patent PCT/NL99/00446.
van Loosdrecht, M. C. M., and M. S. M. Jetten. 1997. Method for treating ammonia-comprising wastewater. Patent PCT/NL97/00482.
Email: webmaster@asmusa.org