Production of oceanic nitrous oxide by ammonia-oxidizing archaea

Introduction Conclusions References


Introduction
Atmospheric nitrous oxide (N 2 O) is a strong greenhouse gas (Forster et al., 2007) and a major precursor of stratospheric ozone depleting radicals (Ravishankara et al., 2009).The ocean is a major source of N 2 O contributing approximately 30 % of the N 2 O in the atmosphere (Denman et al., 2007).Oceanic N 2 O is exclusively produced during microbial processes such as nitrification (under oxic to suboxic conditions) and denitrification (under suboxic conditions; Bange et al., 2010;Codispoti, 2010).The formation of N 2 O as a by-product of nitrification (oxidation of ammonia, NH 3 , via hydroxylamine, NH 2 OH to nitrite, NO − 2 ) was reported for ammonia-oxidizing bacteria (AOB) (Frame and Casciotti, 2010;Goreau et al., 1980).In the case of nitrifierdenitrification NO − 2 can further be reduced to nitric oxide (NO) and N 2 O (Poth and Focht, 1985;Shaw et al., 2006).The accumulation of oceanic N 2 O is favored in waters with low oxygen (O 2 ) concentrations, which is attributed to an enhanced N 2 O yield during nitrification (Goreau et al., 1980;Stein and Yung, 2003).The frequently observed linear correlation between N 2 O (i.e.N 2 O excess) and the apparent oxygen utilization (AOU) is usually taken as indirect evidence for N 2 O production via nitrification (Yoshida et al., 1989).
Published by Copernicus Publications on behalf of the European Geosciences Union.

C. R. Löscher et al.: Production of oceanic nitrous oxide by ammonia-oxidizing archaea
The traditional view that oceanic NH 3 oxidation is exclusively performed by AOB has been challenged by (1) the presence of archaeal amoA genes in metagenomes of various environments (Lam et al., 2009;Schleper et al., 2005;Treusch et al., 2005;Venter et al., 2004), (2) the successful isolation of the ammonia-oxidizing archaeon N. maritimus (Könneke et al., 2005) and (3) the fact that archaea capable of ammonia oxidation have been detected in various oceanic regions throughout the water column and in sediments (Church et al., 2009;Francis et al., 2005;Lam et al., 2009;Santoro et al., 2010;Wuchter et al., 2006).Moreover, N. maritimus appears to be adapted to perform ammonia oxidation even under the oligotrophic conditions (Martens-Habbena et al., 2009) that dominate in large parts of the open ocean.These observations point towards an important role of ammonia-oxidizing archaea (AOA, now constituting the novel archaeal lineage of Thaumarchaeota; Brochier-Armanet et al., 2008;Spang et al., 2010) for the oceanic nitrogen (N) cycle, which has been overlooked until recently (Francis et al., 2007;Schleper, 2010).Archaeal N 2 O production has been proposed to contribute significantly to the upper ocean N 2 O production in the central California Current and has recently been demonstrated to occur in two AOA enrichment cultures (Santoro et al., 2011).However, the ability of AOA to independently produce N 2 O as a by-product of nitrification has not been directly demonstrated in pure cultures or in the ocean.
The eastern tropical North Atlantic (ETNA) and the eastern tropical South Pacific (ETSP) Oceans represent two contrasting oceanic O 2 regimes: while O 2 concentrations in the ETNA are commonly above 40 µmol l −1 , the ETSP regime is characterized by a pronounced depletion of O 2 in intermediate waters between ∼ 75 and 600 m, resulting in an oxygen minimum zone (OMZ) with O 2 concentrations close to or even below the detection limit (∼ 2 µmol l −1 ) of conventional analytical methods.
The amoA gene coding for the alpha subunit of the ammonia monooxygenase is present in archaea as well as in βand γ -proteobacterial ammonia-oxidizers and is commonly used as a functional biomarker for this group (Hallam et al., 2006b;Schleper et al., 2005;Treusch et al., 2005;Venter et al., 2004).Thus, in order to identify whether archaeal or bacterial amoA was associated with the maximum in N 2 O concentration in the ocean, we determined the archaeal and bacterial amoA gene abundances and expression in relation to N 2 O concentrations along vertical profiles during three cruises (in February 2007, February 2008, and June 2010) to the ETNA and one cruise (in January 2009) to the ETSP.Further, we demonstrated N 2 O production in a pure culture of N. maritimus SCM1, which was found to be strongly O 2 sensitive and is thus suggested to be of highest impact at times of ocean deoxygenation (Stramma et al., 2010).N 2 O production from pure cultures of the two marine nitrifying bacteria Nitrosococcus oceani NC10 and Nitrosomonas marina NM22 was low compared to the rates achieved by the archaeal isolate in our experiments.

Hydrographic parameters and nutrients
Samples for salinity, O 2 concentrations and nutrients were taken from a 24-Niskin-bottle rosette equipped with a CTDsensor.Oxygen concentrations were determined following the Winkler method using 50 or 100 ml sampling volumes, and salinity and nutrient concentrations were determined as described in Grasshoff et al. (1999).

Determination of dissolved N 2 O concentrations
Triplicate samples for N 2 O analysis were taken from CTD casts during the cruises P348 (February 2007), ATA03 (February 2008), P399 (June 2010) to the ETNA and M77/3 (January 2009) to the ETSP.N 2 O concentrations were measured with a GC headspace equilibration method as described in Walter et al. (2006); N 2 O and AOU were calculated as described therein.

Sampling
Seawater samples were taken from a minimum of 12 depths from the CTD casts.For the extraction of DNA and RNA a volume of about 2 l seawater was rapidly filtered (∼ 30 min filtration time for samples from the ETNA, for samples from the ETSP exact filtration volumes and times were determined and recorded continuously) through 0.2 µm polyethersulfone membrane filters (Millipore, Billerica, MA, USA).The filters were immediately frozen and stored at −80 • C until further analysis.

Nucleic acid purification
DNA and RNA was extracted using the Qiagen DNA/RNA AllPrep Kit (Qiagen, Hilden, Germany) according to the manufacturers protocol with a small modification.A lysozyme treatment (50 µg ml −1 for 10 min at room temperature) followed by a proteinase K treatment was performed prior to starting the extraction.Extracts of DNA and RNA were quantified fluorometrically using a NanoDrop 2000 (Thermo Scientific Fischer).A treatment with Dnase I (Invitrogen, Carlsbad, CA) was performed with the extracted RNA to remove any residual DNA; purity of RNA was checked by 16S rDNA PCR amplification before reverse transcription.

PCR and quantitative PCR
The extracted RNA was reverse transcribed to cDNA by using the QuantiTect ® Reverse Transcription Kit (Qiagen, Hilden, Germany) following the manufacturers' protocol.Bacterial and archaeal amoA as marker genes for nitrification were PCR-amplified from DNA and cDNA.PCR and quantitative PCR conditions and primers are listed in Table 1.nirS, nirK and nosZ as marker genes for denitrification and the key functional marker gene for anammox, hzo, were PCR amplified according to established protocols (Lam et al., 2007;Schmid et al., 2008).The presence of key genes for anammox and denitrification was tested by PCR in the ETNA but quantified exclusively in samples of the ETSP.Assuming that the PCR detection system has a detection limit comparable to the respective qPCR (using the same Primers and PCR conditions), it should be in the range of 1 copy l −1 for nirS up to 4 copies l −1 in case of the other genes (deduced from the standard calibration curve in the qPCR assays).All PCRs were performed using 0.1 µl FlexiTaq (Promega Corporation, USA).
Absolute quantification of bacterial and archaeal amoA was performed with standard dilution series; quantification was performed in duplicates.Standards for quantitative PCRs were obtained from Nitrosococcus oceani NC10, Nitrosomonas marina NM22 and NM51 for bacterial amoA and from an environmental clone for archaeal amoA (GenBank accession number JF796147).The specifity of the newly developed qPCR primers detecting archaeal amoA was checked according to the Miqe guidelines (Bustin et al., 2009) by cross amplification tests and re-cloning and sequencing of the products.Reactions were performed in a final volume of 25 µl using 0.5 µl of each primer, 6.5 µl nuclease free water and 12.5 µl SYBR qPCR Supermix W/ROX (Invitrogen, Carlsbad, CA).Reactions were performed using an ABI 7300 Real Time PCR system (Applied Biosystems, Carlsbad, CA) according to Lam et al. (2007).

Construction of clone libraries and phylogenetic analysis
Cloning of PCR amplicons was performed using the Topo TA Cloning ® Kit (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions.
Sequencing was carried out by the Institute of Clinical Molecular Biology, Kiel.Sequences for archaeal amoA were analyzed using the ClustalW multiple alignment tool on a 495 bp fragment (sequences were submitted to Gen-Bank under accession numbers JF796145-JF796179); sequence differences were set on a minimum of 5 %, and phylogenetic trees were made using distance-based neighbourjoining analysis (Saitou and Nei, 1987).

Seawater incubations
Seawater incubations were performed at three different stations in the ETNA (cruise P399).The 25 ml serum bottles were filled with seawater from 200-250 m depth from the CTD casts, closed with an air-tight butyl rubber stopper and aluminum crimp-capped.Triplicate samples were taken to determine the initial N 2 O concentration; six bottles were incubated, one triplicate as a control and one triplicate was treated with 1 mM of the hypusination inhibitor N1-guanyl-1,7-diaminoheptane (GC 7 ) (Jansson et al., 2000).Prior to the experiment, the sensitivity of AOA and AOB was checked using Nitrosopumilus maritimus SCM1, Nitrosococcus oceani NC10 and Nitrosomonas marina NM22 pure cultures.Different concentrations up to 1.5 mM GC7 were applied to the cultures, which did not affect ammonia oxidation www.biogeosciences.net/9/2419/2012/Biogeosciences, 9, 2419-2429, 2012 or growth behavior in AOB.In contrast, N. maritimus showed a decrease in ammonia oxidation and growth when applying GC 7 concentrations higher than 0.2 mM and shut down nitrification when applying GC 7 in a concentration of ∼ 0.8 mM.
Incubations were kept for the duration of the experiment (24 h) in the dark at 8 • C. The experiment was stopped by HgCl 2 addition, followed by the determination of the final N 2 O concentrations.

Culture experiments
Pure cultures of N. maritimus, N. oceani and N. marina were grown in triplicates in 125 ml serum bottles (containing 75 ml culture and 50 ml headspace) at 28 • C according to Könneke et al. (2005) and Goreau et al. (1980).Serum bottles were closed with an air-tight butyl rubber stopper and aluminum crimp-capped.Cultures were kept for the duration of the experiment in the dark.Cell abundances from the triplicate samples were monitored by flow cytometry (FACScalibur, Becton, Dickinson) after staining with Sybr Green I (Invitrogen, Carlsbad).The accuracy of the flow cytometry was previously checked microscopically after staining the cells with the fluorescent DNA-binding dye 40, 6 0-diamidino-2-phenylindole (DAPI).Cultures were checked for contaminants microscopically and by 16S rDNA analysis.While tests for the bacterial 16S rDNA gene were negative, the analysis of 84 clones of archaeal 16S rDNA showed that all analyzed sequences matched the identity of the N.maritimus culture.The 16S rDNA gene was PCR amplified with universal primers, followed by Topo TA cloning and sequencing.Sequence analysis did not show any contaminants.
N 2 O concentrations were measured by gas chromatography using the headspace method as described above, oxygen concentrations were determined using Winkler titration in 50mL Winkler bottles.NH + 4 and NO − 2 concentrations were determined at several time points over the complete incubation time frame (Grasshoff et al., 1999).In order to exclude chemical N 2 O production from the medium, cultures toxified with mercury chloride were measured in parallel; no N 2 O production could be detected.
Isotopomeric studies were performed with cultures of 0.  3 (Nevison et al., 2003) was found for the ETNA (Fig. 2), indicating that nitrification was the likely pathway for N 2 O production.A comparable pattern of the distribution of archaeal amoA genes and N 2 O was observed in the water column of the ETNA (Fig. 3), strongly suggesting a correlation between AOA abundance and [N 2 O] accumulation (Pearson correlation coefficient r = 0.63, statistical significance is indicated) in the layers with low O 2 (Fig. 4).The key genes for denitrification and anammox (nirS and nirK coding for nitrite reductases and hzo coding for the hydrazine oxidoreductase; Lam et al., 2007;Schmid et al., 2008) were not present in detectable amounts.A cooccurrence of N 2 O and archaeal amoA genes was detected at certain depths, e.g. at profile V at 100 m water depth (Fig. 1) in the ETSP, but was not a general feature, possibly resulting from additional N 2 O production via other processes such as denitrification, nitrifier-denitrification or anammox (Kartal et al., 2007) at present suboxic conditions.The presence of key genes of anammox and denitrification assayed and predominantly detected at coastal stations of the ETSP but also present in large parts of the area off Peru further points to an active contribution of mixed processes to N 2 O production in the ETSP (the complete dataset of the ETSP can be seen in Löscher (2011).N 2 O production by mixed processes may explain the lack of correlation between N 2 O and AOU as well as NO − 3 in the ETSP (Ryabenko et al., 2012).Gene abundances of archaeal amoA in the ETNA and ETSP were detectable throughout the water column and reached values of up to 1.9 × 10 5 and 6 × 10 4 copies ml −1 , respectively (Fig. 1).Gene abundances of β-and γ -proteobacterial amoA were much lower (up to 950 and 1178 copies ml −1 in the ETNA and ETSP, respectively; Fig. S1 in the Supplement).This is in line with previous studies reporting 1-4 orders of magnitude higher abundances of AOA than AOB in various oceanic regions (Wuchter et al., 2006;Santoro et al., 2010;Lam et al., 2009;Francis et al., 2005;Church et al., 2009).Thus, we hypothesize that a significant part of the N 2 O production occurs via archaeal nitrification in the ETNA and might also be present in parts of the water column of the ETSP.A difference of one order of magnitude between archaeal amoA copies in RNA and in DNA has been observed in vertical profiles of the ETSP, with copy numbers up to 7 × 10 4 copies ml −1 in the DNA and up to 1.5 × 10 3 copies ml −1 in the RNA.A similar tendency is detectable in the ETNA; however, the difference is less pronounced compared to the ETSP.This discrepancy, already reported by Lam et al. (2009), is hypothesized to be due to a diurnal cycle of ammonia-oxidation and therefore changing amoA expression.Moreover, a sampling bias due to comparably long filtration times (up to 30 min) might have led to RNA degradation, as previous studies reported transcript half-lives of down to 0.5 min in Prochlorococcus (Steglich et al., 2010).

Phylogenetic diversity of archaeal amoA
The diversity of AOA present in the ETNA was determined based on the analysis of ∼ 300 amoA sequences from 15 stations of 3 cruises (P348, ATA03, and P399).Sequences were derived from up to 3 depths between the ocean surface and 1000 m, which showed archaeal amoA presence by PCR.The sequences split into two main clusters, with sequences from the O 2 minimum clustering mainly in cluster B (Fig. 5, Table S1 in the Supplement).Only 11.5 % of sequences derived from samples from the O 2 minimum fall into cluster A. Sequences derived from depths between the surface and the upper oxycline were present in both clusters to equal amounts (Fig. 5, Table S1 in the Supplement).In the ETSP, sequences from within as well as from depths above the OMZ separated into both clusters, with the majority of absolute sequence numbers from the OMZ affiliating with cluster B (Fig. 5, Table S1 in the Supplement), as already observed for the sequences from the O 2 minimum in the Atlantic Ocean.

Potential importance of cluster B affiliated Thaumarchaeota for N 2 O production
The distribution of archaeal amoA genotypes along vertical profiles in the ETNA with the majority of cluster B sequences present in clone libraries from the OMZ suggest a production of N 2 O by Thaumarchaeota affiliated with cluster B, previously reported to be a deep marine cluster (Hallam et al., 2006a) associated mainly with O 2 and NH + 4 poor waters (Molina et al., 2010).A niche separation based on O 2 concentrations of cluster B affiliated AOA in the ETNA seems to be very likely, which is consistent with our data from the ETSP.Regarding the on-going decrease in dissolved O 2 concentrations in tropical ocean areas (Stramma et al., 2010), we hypothesize that cluster B affiliated AOA might dominate the production of N 2 O and the balance between reduced and oxidized nitrogen species in the ocean, as those organisms are likely more adapted to low O 2 concentrations.
Both observed clusters were present in a similar distribution along vertical profiles during all cruises to the ETNA (Figs. 4 and 5, Table S1 in the Supplement).The community of AOA in this area appears therefore to be stable over the time investigated.

N 2 O production in the ETNA
At three different stations in the ETNA, 24h seawater incubations using seawater from the N 2 O maximum (at the depth of the OMZ) were performed.In two out of three experiments, N 2 O production was significantly lower in samples treated with N 1 -guanyl-1,7-diaminoheptane (GC 7 ) (Fig. 6), a hypusination inhibitor shown to selectively inhibit the cell cycle of archaea (Jansson et al., 2000), but which appears not to affect AOB (for detailed experimental data see Sect. 2).In the third experiment performed at a coastal station, a similar trend was observed; however it was not statistically significant.These findings further support our hypothesis that N 2 O production in large parts of the ETNA occurs within the OMZ and is mainly carried out by archaea.

N 2 O production in N. maritimus
Direct evidence for the production of N 2 O by archaea was obtained from experiments with pure cultures of N. maritimus.The purity of the cultures was confirmed by extensive 16S rDNA analysis (see Sect. 2).N. maritimus cultures grew at comparable rates under the varying O 2 conditions and showed similar nitrification rates.Production of N 2 O in N. maritimus cultures was inversely correlated to O 2 concentrations (Fig. 7) which were chosen according to the O 2 concentrations present along the oxycline in the ETNA (112, 223 and 287 µmol l −1 , Fig. 1).N 2 O production rates from two AOB cultures (Nitrosomonas marina NM22 and Nitrosococcus oceani NC10) were considerably lower compared to the N 2 O produced by N. maritimus (Fig. 7, Table 2).The N 2 O yields (N 2 O/NH + 4 ), which appear to result from ammonia oxidation, ranged from 0.002 %-0.03 %  Goreau et al. (1980), even though a different experimental setup was used.Culture experiments, such as those presented here, were performed with AOB cell densities (∼ 10 5 cells ml −1 ), which are much higher than usually found in the ocean (10 2 -10 3 cells ml −1 ) (Wuchter et al., 2006;Lam et al., 2009).Thus, the N 2 O production rates from the AOB cultures are probably overestimated and not representative as N 2 O production per cell by AOB also depends on the present cell densities (Frame and Casciotti, 2010), with high cell densities leading to enhanced N 2 O production.In contrast, the AOA cell densities in our culture experiment (∼ 10 5 -10 7 cells ml −1 ) were comparable to those present in the oceanic environment (∼ 10 5 cells ml −1 ) and thus seem to be reasonably representative of the rates expected in natural populations of AOA.
Using the observed archaeal N 2 O production rate for low O 2 conditions derived from our experimental results (140 nmol l −1 d −1 ; normalized to 10 6 cells ml −1 yielding ∼ 24 nmol l −1 d −1 , see Table 2), an upper estimate for the potential archaeal N 2 O production would be around 14 nmol m −2 s −1 (however, NH + 4 concentrations in our culture experiments were significantly higher than in the environment), assuming a thickness of about 50 m for the low O 2 layer as typically found in the ETNA.Compared to estimates of N 2 O emissions from the ETNA to the atmosphere of up to 2 nmol m −2 s −1 (Wittke et al., 2010), potential oceanic archaeal N 2 O production might be indeed significant.

Potential pathways for archaeal N 2 O production
AOB can produce N 2 O from NH 2 OH during nitrification or from NO − 2 during nitrifier-denitrification (Kool et al., 2010;Shaw et al., 2006).In AOA however, the pathway of ammonia oxidation is yet not understood.So far, no equivalent to the hydroxylamine-oxidoreductase, which catalyses the oxidation of NH 2 OH to NO − 2 during nitrification in AOB, has been identified (Könneke et al., 2005;Martens-Habbena et al., 2009;Walker et al., 2010) (Santoro et al., 2011), which is in agreement with the SP N 2 O of ∼ 33 ‰ typically found in AOB cultures performing ammonia oxidation (Sutka et al., 2006) (for comparison: nitrifier-denitrification of AOB results in a SP N 2 O of about 0 ‰).Thus, our dataset points towards a production of N 2 O via the oxidation of NH + 4 to NO − 2 , potentially via an unknown intermediate as we were not able to detect NH 2 OH in N. maritimus cultures using the method described in Schweiger et al. (2007).However, taking δ 18 O data into account, Santoro et al. (2011) suggested a reduction of NO − 2 to N 2 O.As we have not performed O 2 isotopomeric studies, we cannot exclude N 2 O production via nitrifier-denitrification, particularly when O 2 becomes limiting as previously described for the Arabian Sea (Nicholls et al., 2007) where O 2 concentrations drop far more than in our experiments (lowest O 2 concentration ∼ 112 µM).

Summary
Taken together, the high abundance of archaeal amoA relative to AOB, the frequently obtained comparable patterns of N 2 O accumulation and archaeal amoA, the inhibition of N 2 O production in seawater experiments in the presence of the archaeal inhibitor GC 7 as well as the N 2 O production by N. maritimus add to the mounting evidence that, in large parts of the ocean, N 2 O is produced by archaeal rather than by Biogeosciences, 9, 2419-2429, 2012 www.biogeosciences.net/9/2419/2012/bacterial nitrification.Further, the archaeal N 2 O production appears to be highly sensitive to the dissolved O 2 concentration, with highest N 2 O production rates at low O 2 concentrations such as those present in the OMZ of the ETNA.The predicted expansion of OMZs in the future in many parts of the ocean (Stramma et al., 2008) may lead to an enhanced N 2 O production in the ocean (Naqvi et al., 2010) and therefore may have severe consequences for the budget of N 2 O in the atmosphere as well.

Fig. 1 .
Fig. 1.Maps of sea surface temperatures (A) from the eastern tropical North Atlantic Ocean and (B) from the eastern tropical South Pacific Ocean.The locations of sampling stations are indicated with asterisks in the maps.Selected vertical profiles (I-VI) of O 2 , NO − 2 , N 2 O (measured in triplicates) and archaeal amoA (measured in duplicates by qPCR) are shown; (I) and (IV) are located offshore, (II) and (V) are located on the continental slope, and (III) and (VI) are onshore/coastal stations.

Fig. 2 .
Fig. 2. N 2 O versus the apparent oxygen utilization (AOU) and nitrate in the ETNA (data from cruises ATA03, P348 and P399), the potential density is colour-coded.

Fig. 4 .
Fig. 4. Archaeal amoA and N 2 O versus O 2 in the ETNA (data from the cruises ATA03, P348 and P399/2).A similar trend has been detected during the three cruises.

Fig. 5 .
Fig. 5. Distance-based neighbour-joining phylogenetic tree of archaeal amoA sequences from the ETNA (cruises ATA03 and P348) and ETSP (cruise M77/3).Sequences derived from the oxygen minimum zone (OMZ) of the ETNA are in italics, bold and marked with solid stars; sequences from above the OMZ are in italics.Sequences from the OMZ of the ETSP are in italics and bold; sequences from above the OMZ are in italics.

Fig. 6 .
Fig. 6.N 2 O production determined from seawater incubations at three different stations (1-3) from the ETNA (P399).N 2 O was calculated as the difference of N 2 O concentrations over the incubation time (i.e.24 h).Open columns represent samples with no inhibitor, filled columns represent samples with the archaeal inhibitor GC 7 .Error bars indicate the standard deviation of three technical replicates.

Fig. 7 .
Fig. 7. N 2 O (A), O 2 (B), NH + 4 and NO − 2 (C) as well as cell abundances (D) determined from incubation experiments with pure cultures of N. maritimus.Experiments are colour-coded according to their initial O 2 concentrations: red (112 µmol l −1 ); violet (223 µmol l −1 ); and blue (287 µmol l −1 ).N 2 O and cell abundances were measured in triplicates and the associated error ranges are indicated (please note that in the most cases the error bars are too small to be visible in the figure).

Table 1 .
Primers and PCR conditions: for real-time qPCR the initial denaturing step was 10 min at 95 • C, annealing temperatures were the same as in the end point PCRs, no final extension step took place, 40 cycles were performed followed by melting curve analysis.A fragment of 175 bp was amplified in qPCRs of archaeal amoA.

Table 2 .
N 2 O production in culture experiments: mean O 2 and N 2 O concentrations (in triplicate samples) of pure cultures of N. maritimus, N. marina and N. oceani after 264 h incubation, the initial NH + 4 concentration (∼ 1 mmol l −1 ) was completely converted to NO − 2 in the end of the experiment in AOB cultures, NO − 2 was below the detection limit at the initial time point of the incubation.
, indicating that AOA likely use a different pathway than AOB do when producing N 2 O.The detection of the nitrite reductase gene nirK in the sequenced genomes of cultured Thaumarchaeota (Walker et al., 2010) led to the theory that AOA might produce N 2 O by nitrifier-denitrification, which might particularly impact at low O 2 concentrations.To identify the origin of N 2 O formation, isotopomeric studies were performed with N. maritimus pure cultures.Using the lowest O 2 concentration of the three chosen (112 µM), a 15 N site preference (SP N 2 O ) in N 2 O of 34 ± 12 ‰ was detected, consistent with results from AOA enrichments