Isotopomeric characterization of nitrous oxide produced by reaction of enzymes extracted from nitrifying and denitrifying bacteria

Nitrous oxide (N2O) is a potent greenhouse gas and produced in denitrification and nitrification by various microorganisms. Site preference (SP) of 15N in N2O, which is defined as the difference in the natural abundance of isotopomers14N15NO and15N14NO relative to14N14NO, has been reported to be a useful tool to quantitatively distinguish N2O production pathways. To determine representative SP values for each microbial process, we firstly measured SP of N2O produced in the enzyme reaction of hydroxylamine oxidoreductase (HAO) purified from two species of ammonia oxidizing bacteria (AOB),Nitrosomonas europaeand Nitrosococcus oceani , and that of nitric oxide reductase (NOR) from Paracoccus denitrificans . The SP value for NOR reaction (−5.9± 2.1 ‰) showed nearly the same value as that reported for N2O produced byP. denitrificansin pure culture. In contrast, SP value for HAO reaction (36.3 ± 2.3 ‰) was a little higher than the values reported for N 2O produced by AOB in aerobic pure culture. Using the SP values obtained by HAO and NOR reactions, we calculated relative contribution of the nitrite (NO−2 ) reduction (which is followed by NO reduction) to N2O production byN. oceanincubated under different O2 availability. Our calculations revealed that previous in vivo studies might have underestimated the SP value for the NH2OH oxidation pathway possibly due to a small contribution of NO−2 reduction pathway. Further evaluation of isotopomer signatures of N 2O using common enzymes of other processes related to N 2O would improve the isotopomer analysis of N 2O in various environments.


Introduction
Nitrous oxide (N 2 O) is a potent greenhouse gas and contributes indirectly to the destruction of the ozone layer in the stratosphere (Ravishankara et al., 2009).Production of N 2 O on earth is mainly controlled by microbial processes that include nitrification and denitrification (Stein and Yung, 2003).In nitrification, autotrophic microorganisms like ammonia oxidizing archaea (AOA) and ammonia oxidizing bacteria (AOB) oxidize ammonia to nitrite (NO − 2 ) and produce N 2 O as a byproduct under mainly aerobic conditions (Casciotti et al., 2011).However, various heterotrophic denitrifying microorganisms such as archaea, bacteria, and fungi reduce nitrate (NO − 3 ) or NO − 2 to dinitrogen (N 2 ) as electron acceptors for anaerobic respiration and produce N 2 O as an intermediate under mainly anaerobic conditions (Hayastu et al., 2008).Thus, production and consumption of N 2 O consist of many microbiological functions, including autotrophic/heterotrophic and oxic/anoxic processes.Investigation of N 2 O production pathways is useful to understand the nitrogen cycle and relevant microorganisms in various environments.
Natural abundance ratios of isotopomers, a set of molecules containing various stable isotopes, have been used to analyze the N 2 O production pathways, and several studies have reported isotopomer ratios of each N 2 O production pathway using several strains belonging to bacteria, archaea, and fungi (Casciotti et al., 2011;Frame and Casciotti, 2010;Santoro et al., 2011;Sutka et al., 2008Sutka et al., , 2006Sutka et al., , 2004;;Toyoda et al., 2005;Yoshida, 1988).These studies showed that the site preference (SP), which was defined as the difference in 15 N enrichment between the center (α) and the terminal (β) N atoms in N 2 O molecules, is a powerful tool to quantitatively distinguish the production pathways such as bacterial NO − 2 reduction and bacterial hydroxylamine (NH 2 OH) oxidation, as the SP value for each production pathway was found to be independent of concentrations or isotope ratios of substrates (Toyoda et al., 2005;Sutka et al., 2003Sutka et al., , 2004Sutka et al., , 2006)).To date, isotope analysis has been applied to N 2 O produced in various ecosystems, including natural and agricultural soils, oceans, rivers, wastewater treatment plants (Goldberg et al., 2010;Koba et al., 2009;Park et al., 2011;Sasaki et al., 2011;Toyoda et al., 2011;Well et al., 2008;Yoshida and Toyoda, 2000).However, as argued by Baggs et al. (2008), the isotopomer signatures were reported for only a few bacterial strains and the variation of the signatures among various species was not fully evaluated.In addition, previous reports on the SP value of N 2 O produced by NH 2 OH oxidation might have been biased by simultaneous N 2 O production by NO − 2 reduction, since some AOB have denitrifying enzymes and produce N 2 O by NO − 2 reduction in addition to NH 2 OH oxidation (Arp et al., 2003).In consideration of this problem, Frame and Casciotti (2010) estimated the SP value of N 2 O produced by NH 2 OH oxidation using the relationship between SP and oxygen isotope ratios in N 2 O obtained in incubation experiments with Nitrosomonas marina C-113a, a marine ammonia-oxidizing bacterium, under various oxygen concentrations.Clearly, more information on SP produced by other strains should be obtained for the better use of SP and precise estimation of the contribution from each N 2 O production pathway.
In this study, we report the isotopomer ratios of N 2 O produced in vitro using enzymes extracted and purified from two strains of AOB (Nitrosococcus oceani and Nitrosomonas europaea) and one species of denitrifying bacteria (Paracoccus denitrificans).Our experiments have the advantage that isotope effects related to the N 2 O production by enzymes (hydroxylamine oxidoreductase (HAO) in AOB and nitric oxide reductase (NOR) in denitrifier) can be directly determined and that the effects related to other processes such as diffusion of substrate/product through cell membrane and reactions mediated by other enzymes can be excluded.We also measure isotopomer ratios of N 2 O produced in vivo by N. oceani under different oxygen concentrations and estimate the relative contribution of NH 2 OH oxidation and NO − 2 reduction using the isotopomer signatures obtained in vitro.

Cultivation of the bacterial strains
Nitrosococcus oceani strain NS58 was kindly supplied by Dr. H. Urakawa (Florida Gulf Coast University).Phylogenetic and morphological analyses indicated a close systematic relationship of the bacterium with N. oceani ATCC19707 (Hozuki et al., 2010).The bacterium was cultivated in the (NH 4 ) 2 SO 4 -supplemented artificial seawater (37.8 mM NH + 4 ), of which the pH was buffered to 7.8 by 50 mM MOPS (3-morpholinopropanesulfonic acid) as described in detail in a previous report (Hozuki et al., 2010).The inoculated medium in the conical flask was shaken reciprocally at 120 rpm at 25 • C in a dark condition.After cultivation for 7-10 days and the concentration of NO − 2 in the medium had reached 25-30 mM, the culture of the bacterium in the late-log or stationary phase was used for the in vivo experiments as described below.Large-scale cultivation of the bacterium to prepare the starting material for purification of HAO was carried out with the same procedure, but using a glass bottle 10 L in volume with vigorous air ventilation through a sterilized air filter.
Cultivated cells of Nitrosomonas europaea ATCC19718, which was obtained by large-scale cultivation in accordance to a previous study (Yamanaka and Shinra, 1974), were kindly supplied by Dr. Y. Fukumori (Graduate School of Natural Science and Technology, Kanazawa University) and were used as a starting material for the purification of HAO.Paracoccus denitrificans ATCC35512 was cultured under denitrifying conditions and used for preparing a purified NOR according to the previous report by Fujiwara and Fukumori (1996).

Purification of hydroxylamine oxidoreductase (HAO)
HAO was purified from the cultivated NS58 cells through three preparative steps, including (NH 4 ) 2 SO 4 fractionation, gel-filtration and hydrogen-bonding chromatography according to a previous report (Hozuki et al., 2010).Catalytic activity of HAO was analyzed by spectrophotometrical measurement of the NH 2 OH-dependent reduction of potassium ferricyanide as reported previously (Hozuki et al., 2010).Briefly, the purified enzyme was mixed with 1 mL of the reaction solution containing 0.1 M sodium phosphate buffer (pH 7.8), 20 µM NH 2 OH and 100 µM potassium ferricyanide.Then, the decreasing rate of absorbance at 420 nm of the solution was monitored using a cuvette with an optical path length of 1cm and a spectrophotometer (MPS-2000, Shimadzu, Kyoto, Japan).The difference in the millimolar extinction coefficient ( ε mM ) at 420 nm between ferricyanide and ferrocyanide was 1.02 mM −1 cm −1 .We obtained 1.8 mg of the purified HAO with enzymatic activity of 37 unit mg protein −1 (1 unit is equivalent to the activity where 1 µmol of NH 2 OH is oxidized in a minute) and used it for the assay of N 2 Ogenerating activity.HAO was also purified from N. europaea according to the method of Yamanaka et al. (1979) with some modifications.We found lower catalytic activity for the enzyme purified from N. europaea (6.7 unit mg protein −1 ); this is probably because we used an old stock of the cultivated bacterial cells which had been kept for about 10 years at −30 • C.

Quantitative and isotopomeric analysis of N 2 O produced during oxidation of NH 2 OH with HAO
In a 69 mL glass vial (Maruemu Corp., Osaka, Japan), 10 mL of substrate solution was prepared so that it contains 0.1-3 mM hydroxylamine, 1 mM potassium ferricyanide as an electron acceptor, and 10 mM sodium phosphate as buffer (pH = 7.8).After the vial was sealed with a butyl rubber stopper and an aluminum cap, air in the headspace was replaced with pure N 2 (Shizuoka Sanso Co., Shizuoka, Japan), and then the reaction was started by injecting 0.1 unit of the HAO extracted from N. oceani or N. europaea.Experiments with 3 mM NH 2 OH were conducted on four different dates (A-D).In experiments C and D, the lot number of NH 2 OH reagent (hydroxylamine hydrochloride, Wako Pure Chemical Industries, Ltd., Osaka, Japan) was different from that used in experiments A and B. After incubating the vial for 2 h at 25 • C or for 12 h on ice, gas samples were extracted and analyzed as described below.Concentration of NO − 2 was determined spectrophotometrically by a diazo-coupling method (Nicholas and Mason 1957).

Paracoccus denitrificans
NorBC-type NO reductase was purified from the cultivated P. denitrificans cells according to a previous report (Fujiwara and Fukumori 1996) with some modification.The membrane fraction was prepared from the bacterial cells that had been cultivated anaerobically in the presence of NO − 3 .NO reductase was extracted from the membrane fraction by treating with the 1 % (w/v) 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS; Dojindo Lab., Kumamoto, Japan).It was then fractionated by anion-exchange chromatography using DEAE-Toyopearl 650 M gel (Tosoh, Tokyo, Japan).
NO-reducing activity of the purified enzyme was measured spectrometrically by monitoring NO-dependent oxidation of horse ferrocytochrome c (Nacalai Tesque, Kyoto, Japan).Experimental procedure was detailed in a pre-vious report (Fujiwara and Fukumori 1996).Before starting the assay of NO-reducing activity, dissolved oxygen in the solution was enzymatically removed using D-glucose oxidase/catalase system in order to prevent its reaction with NO.A stock solution of NO-saturated ethanol (11.9 mM) was prepared by purging ethanol with pure NO gas (Sumitomo Seika Chemicals Co., Ltd.Osaka, Japan) (Seidell and Linke, 1965).Oxidation rate of ferrocytochrome c was measured by monitoring the decrease in the absorbance at 550 nm with a spectrophotometer (MPS-2000, Shimadzu, Kyoto, Japan), whereby the difference in millimolar extinction coefficient ( ε mM , 21.0 mM −1 cm −1 ) at 550 nm between the reduced and oxidized form of horse ferro-and ferricytochrome c was used.NO-reducing activity of the purified enzyme was estimated to be 35 unit mg protein −1 (1 unit is equivalent to the activity where 1 µmol of NO is reduced in a minute).

Isotopomeric analysis of N 2 O produced by enzymatic reduction of NO
In the same glass vial that had been used for NH 2 OH oxidation experiments, a 10 mL solution was prepared so that it contained 10 mM sodium phosphate buffer (pH 6.0), 2.9 µM horse cytochrome c, 1.0 mM sodium ascorbate and 0.1 mM N, N, N , N -tetramethyl-p-phenylenediamine (TMPD).The reaction was started by injecting the NO-saturated ethanol to yield final concentration of 50 µM, and the purified NO reductase (0.005 unit) by using a gastight syringe (VICI Precision Sampling Inc., Baton Rouge, LA).Experiments were conducted on three different dates (A, C, and D).After incubating the vial for 2 h at 25 • C or for 12 h on ice, the gas phase in the headspace of the reaction vial was extracted and the isotopomeric analysis of the produced N 2 O was conducted as mentioned below.

Quantitative and isotopomeric analysis of N 2 O produced from cultivated N. oceani cells
The 70 mL culture of the N. oceani NS58 was centrifuged at 9800 g for 60 min at 4 • C (refrigerated centrifugator model 3700, Kubota Corp., Tokyo, Japan).Obtained pelleted cells were suspended in the 70 mL of the pH-buffered and (NH 4 ) 2 SO 4 -supplemented artificial seawater (see Sect. 2.1), then were incubated with gentle stirring at 25 • C in dark for 30 min to remove NO − 2 that accumulate in the medium during the cultivation.The resulting cell suspension was centrifuged again under the same condition, and the cell pellet was resuspended in the same volume of the freshly prepared culture medium and used as the washed cell suspension.
Reaction solution was prepared in a glass vial by mixing 1.0 mL of the washed cell suspension and 9.0 mL of the cultivation medium.Cell density of the medium in the reaction vial was estimated at 5.8 × 10 6 cells mL −1 by direct counting of the fixed and 4 ,6-diamidino-2-phenylindole (DAPI)stained cells using an epifluorescent microscope according to  Takenaka et al. (2007).After sealing, the headspace (59 mL in volume) of the vial was replaced with a pure N 2 or O 2 /N 2 mixture that contained 2 % (v/v) O 2 (Shizuoka Sanso Co., Shizuoka, Japan) by gently bubbling for 15 min.In addition to the anaerobic (0 % O 2 ) and the microaerobic (2 % O 2 ) vials, an aerobic vial without gas replacement was prepared.Incubation of the bacterial cells in the vials was carried out by gentle shaking at 25 • C in dark.After 24 or 48 h from starting the incubation, 50 µM of 10 M NaOH solution was added into the vials to stop the microbial reaction.Quantification and measurement of isotopomer ratios of N 2 O gas released into the headspace of the incubation vial were performed as described below.Concentration of accumulated NO − 2 was also measured as described above.

Measurement of N 2 O concentration and isotopomer ratios
Concentrations and isotopomer ratios of N 2 O were measured using an online analytical system that originally was developed for N 2 O dissolved in water samples (Toyoda et al., 2002).The system consists of a gas extraction chamber with a septum for syringe injection, traps made of stainless steel tubing or glass, a gas chromatograph (Agilent 6890, Agilent Technologies Japan, Ltd., Tokyo) and an isotope-ratio monitoring mass spectrometer (MAT 252, ThermoFisher Scientific KK, Yokohama) equipped with GC interface.Using a gastight syringe, 0.1 to 1.5 mL of the gas was extracted from the headspace of the sample vial and was injected into the gas extraction chamber.The sample gas was then transferred by He carrier gas to chemical traps (Mg(ClO 4 ) 2 and NaOH on support) to remove H 2 O and CO 2 , and N 2 O was concentrated on glass beads packed in a U-shaped trap at liquid N 2 temperature.After further purification on GC, N 2 O was introduced into the mass spectrometer for isotope ratio monitoring.Site-specific nitrogen isotope analysis was conducted with ion detectors modified for mass analysis of fragment ions of N 2 O (NO + ) that contain an N atom in the center position of the N 2 O molecules (Toyoda and Yoshida, 1999).Concentration and bulk nitrogen and oxygen isotope ratio of N 2 O was determined by analysis of N 2 O molecule ions (N 2 O + ).Pure N 2 O was used as a reference gas for isotopomer ratios.Notation of isotopomer ratios of N 2 O is shown below.
In Eqs. ( 1) and (2), 15 R and 18 R represent 15 N / 14 N ratio and 18 O / 16 O ratio, respectively.Subscript "sample" and "standard" indicate isotope ratios for sample and the standard (atmospheric N 2 for nitrogen and Vienna Standard Mean Ocean Water (VSMOW) for oxygen), respectively.Superscript i is α, β, or bulk, which respectively designates central, peripheral, or average isotope ratios in nitrogen atom(s) in N 2 O molecules.We also define the 15 N site preference (SP) as an illustrative parameter of intramolecular distribution of 15 N (Yoshida and Toyoda, 1999).The precision of measurement is better than 0.5 ‰ for δ 15 N bulk and δ 18 O, and better than 1.0 ‰ for δ 15 N α and δ 15 N β .
15 N − Site preference The δ 15 N of NH 2 OH was measured using an elemental analyzer coupled with isotope ratio mass spectrometer.Statistical analysis was performed using Excel 2011 (Microsoft, USA).The statistical difference was determined by a twosided Student's t test.A difference of p < 0.05 was considered significant.

Concentrations and isotopomer ratios of N 2 O produced during oxidation of NH 2 OH with HAO
The purity of HAO obtained from N. oceani NS58 was confirmed by the SDS-PAGE technique.As shown in Supplement Fig. S1, SDS-PAGE results of the purified sample revealed a single protein band with a molecular weight of approximately 182 000, following pretreatment with SDS (Lane 2).Upon treatment of the enzyme with both SDS and β-mercaptoethanol, the major band accompanied by minor protein bands with molecular weights of 140 000 and 60 000 appeared on the gel (Lane 1).The chemical cleavage of the thioester bond between heme c and a Cys residue in HAO resulted in the appearance of a protein band of molecular weight 60 000 (Lane 3), suggesting that the enzyme is composed of a homotrimer of subunit proteins covalently linked together and mediated by prosthetic heme c molecules, as previously reported for N. europaea HAO (Terry and Hooper, 1981;Igarashi et al., 1997).The two minor protein bands observed in Lane 1 possibly correspond to the monomeric and dimeric configurations of the HAO subunit.In general, AOB possesses two types of hydroxylamine-oxidizing enzymes: HAO, with a trimeric configuration of the octaheme c subunit and cytochrome P460, a monoheme protein with a molecular weight of only approximately 18 000 (Numata et al. 1990).
Gel filtration was used for purification of N. oceani NS58 HAO; thus, contamination of the cytochrome P460 in the enzyme preparation should be improbable.Results of electrophoretic analysis are consistent with the absence of components that may interfere with the correct interpretation of our experimental findings (Lanes 1 and 2, Fig. S1).
Figure 1 shows the amount of NO − 2 and N 2 O produced during the reaction catalyzed by HAO from N. oceani with different initial concentration of NH 2 OH (0.1-3 mM).About 5.1-29.0% of NH 2 OH was converted to NO − 2 or N 2 O after the reaction, and the ratio of produced NO − 2 and N 2 O decreased with initial NH 2 OH concentration.Production of N 2 O in the blank runs with 1 and 3 mM NH 2 OH were 5.9 and 7.1 % of that in the presence of HAO (HAO+), respectively (Table 1).With lower NH 2 OH concentrations of 0.1 and 0.3 mM, however, production of N 2 O in the blank runs was 32.2 and 179.1 % of that observed in the HAO+ runs, respectively (Table 1).For this reason, further experiments were conducted at an initial NH 2 OH concentration of 3 mM to examine repeatability of the reaction and effect of bacterial strain and N isotope ratio of NH 2 OH.The δ 15 N bulk and δ 18 O of N 2 O showed lower values at high (3 mM) NH 2 OH concentration compared to low (1 mM) concentration (Table 1).SP value was independent of substrate concentration or the degree of the reaction, since the difference in SP values between experiments with 1 and 3 mM NH 2 OH was insignificant (p > 0.05).
Results obtained with an initial NH 2 OH concentration of 3 mM showed that δ 15 N bulk of N 2 O produced by N. oceani (experiments A and B) and N. europaea (experiment A) was almost constant and about 10 ‰ lower than δ 15 N of NH 2 OH (−7.0 ‰) (Table 2).Similar δ 15 N difference (9 ‰) between N 2 O and NH 2 OH was observed in experiment D, in which NH 2 OH with lower δ 15 N value (−43.9 ‰) was used.In experiment C of both N. europaea and N. oceani, however, δ 15 N of N 2 O was higher than that of NH 2 OH by 1.8-9.9‰ (Table 2).The δ 18 O also showed a variation among experiments A-D of N. oceani, although it is not clear whether the δ 18 O of NH 2 OH was different between experiments A/B and C/D.When we compare the results from experiment A, the δ 15 N bulk and δ 18 O of N 2 O were not significantly different between the two strains.
In contrast, the SP value of N 2 O produced by NH 2 OH oxidation with HAO was independent of initial concentration and δ 15 N value of NH 2 OH or degree of reaction process (Tables 1 and 2).Moreover, the difference in SP between experiments with HAO extracted from N. oceani in experiments A, B , C and D (average: 36.2 ± 1.7 ‰, n = 7) and N. europaea in experiments A and C (average: 36.6 ± 3.3 ‰, n = 4) was insignificant (p > 0.05).

Concentration and isotopomer ratios of N 2 O produced during reduction of NO with NOR
Table 3 shows concentration and isotopomer ratios of

Isotopomer ratio of N 2 O produced by N. oceani under different initial O 2 concentration
The amount of N 2 O produced by N. oceani was 2.9 ± 0.9 (n = 6), 8.9 ± 0.5 (n = 2), and 18.1 ± 3.6 (n = 5) nmol for O 2 concentrations of 0, 2, and 21 %, respectively.Production of both N 2 O and NO − 2 was much higher under aerobic conditions than under anaerobic conditions (p < 0.05, Fig. 2).Isotopomer ratios of N 2 O showed a strong dependence on initial O 2 concentration (Fig. 3).They were more enriched under aerobic conditions compared to anaerobic conditions, and positive correlation between the isotopomer ratios was observed (Fig. 3, R 2 = 0.89, p < 0.05).

Discussion
The characteristic SP value of N 2 O produced during the in vitro oxidation of NH 2 OH with HAO from the two strains of AOB (36.3 ± 2.3 ‰) indicates that this parameter is determined by the enzymatic reaction step and is not affected by other factors such as concentration, the extent of reaction, and nitrogen isotope ratios of substrate (Table 2).This study demonstrates the first direct evidence of isotopomeric fractionation of N 2 O produced by enzymatic reactions, confirming the robustness of SP of N 2 O as predicted by previous studies (e.g., Frame and Casciotti. 2010).
Observed dependence of product NO − 2 / N 2 O ratio on initial NH 2 OH concentration could be caused by the availability of electron acceptors.The oxidation of NH 2 OH with HAO mainly provides NO − 2 as a product when enough electron acceptors are supplied, while the reaction is likely to produce N 2 O when the amount of electron acceptors is not enough to complete the reaction as proposed by Yamanaka and Sakano (1980)     to initial NH 2 OH), and it must have been produced by noncatalytic reactions.
The yield of N 2 O produced by the reduction of NO with NOR exceeded 100 %, which could be caused by underestimation of initial NO concentration.Since the NO-saturated ethanol solution was prepared by purging the ethanol with pure NO gas in a sealed vial at room temperature (25 • C), we calculated the NO concentration based on NO solubility at 25 • C and 1 atm.However, there is a possibility that the vial inside was pressurized and the excess NO dissolved in ethanol during the preservation at 4 • C.
The SP value of N 2 O produced during the in vitro reduction of NO with NOR from P. denitirifcans (−5.9 ± 2.1 ‰) agrees with the value reported for N 2 O produced in vivo from NO − 3 reduction by the same species (−5.1 ± 1.8 ‰, Toyoda et al. 2005).Therefore, our results prove that the factorcontrolling SP value of N 2 O produced in NO − 3 reduction is the reaction with NOR rather than other reaction steps, including diffusion of substrate and product through cell membranes.However, the SP value is slightly lower than SP values reported for N 2 O produced from NO − 3 or NO − 2 reduction by other denitrifying bacteria or some species of AOB (−0.8 to +0.1 ‰) (Sutka et al., 2004(Sutka et al., , 2006)), and slightly higher than the value estimated for N 2 O produced from NO − 2 reduction by N. marina 113a, oceanic AOB (−10.7 ± 2.9 ‰; Frame and Casciotti 2010).This implies that SP might depend on a small structural difference in the active center of NOR between the studied species, or that previously reported SP for N 2 O from NO − 3 / NO − 2 reduction could have been affected by other processes like NH 2 OH oxidation.
It is known that AOB produce N 2 O as a byproduct during the oxidation of NH 2 OH to NO − 2 with HAO and that some species of AOB can also produce N 2 O from nitrite by the sequential reduction of nitrite to NO (using NIR) followed by reduction of NO (using NOR) (Klotz et al., 2006;Arp et al., 2003).The latter pathway is often referred to as nitrifierdenitrification (Wrage et al., 2001) and is believed to occur under anaerobic conditions.Although elucidating the reason why those AOB function as nitrifier-denitrification requires study of a wider spectrum of strains, detoxification of accumulated nitrite, co-respiration of oxygen and nitrite, and other mechanisms have been proposed (Beaumont et al., 2004;Stein, 2011).
In our bacterial culture experiments, production of both NO − 2 and N 2 O are enhanced under aerobic conditions (Fig. 2) because O 2 is required for ammonium oxidation, the first step of the successive reaction to NO − 2 .Although we did not measure O 2 concentration during the aerobic incubation (21 % O 2 ), oxygen concentration must have been kept high enough to maintain the aerobic condition for the following reasons: (1) it is reported that the measured ratio of consumed O 2 and produced NO − 2 during aerobic incubation of N. oceani (1.35-1.43) is consistent with the ratio calculated from the stoichiometry of ammonia oxidation to nitrite (1.50) (Gundersen, 1966); (2) oxygen consumption by respiration using endogenous organics is low (Gundersen, 1966); and (3) a quantitative estimate based on the stoichiometry and the amount of the product (NO − 2 and N 2 O) in this study indicates that the oxygen consumption was relatively small (e.g., oxygen concentration should have decreased from 21 % to ca. 18 %).
Furthermore, observed co-variation of SP values of N 2 O and oxygen concentrations (Fig. 3) implies that relative contributions from the two production pathways are sensitive to oxygen availability.It has been shown that SP of N 2 produced by nitrifier-denitrification and by NH 2 OH oxidation are approximately 0 and 30 ‰, respectively (Sutka et al., 2006).Consequently, the lower SP values produced during nitrification in this study are most likely the result of the addition of N 2 O from nitrifier-denitrification. It turns out that NH 2 OH oxidation is the dominant N 2 O production pathway under aerobic conditions, while NO − 2 reduction is dominant under anaerobic conditions (Fig. 3).We note that even in aerobic conditions (21 % O 2 ), the SP value of N 2 O produced by N. oceani is lower than that of N 2 O produced by in vitro NH 2 OH oxidation with HAO extracted from the same bacteria (Fig. 3).This suggests the NO − 2 reduction pathway is not negligible under aerobic conditions and thus previous studies based on pure culture incubation of AOB underestimated the SP value for N 2 O from NH 2 OH oxidation.
We further show quantitative estimation of the contribution of NO − 2 reduction to N 2 O production.Here we assume that only NH 2 OH oxidation and nitrifier-denitrification were the N 2 O production pathways, although there was a possi-bility that N 2 O was produced by other unknown processes.This assumption allows a simple interpretation of the relative contribution of hydroxylamine oxidation and nitrifierdenitrification to N 2 O production on the basis of SP as follows.
In Eq. ( 4), SP measured , SP NOR , and SP HAO represent SP values of observed N 2 O and N 2 O produced in enzymatic reaction with NOR and HAO, respectively, and X NOR indicates the relative contribution of the NOR-mediated pathway.As shown in Table 4, average X NOR is calculated as 82 ± 18 %, 42 ± 10 %, and 22 ± 4 % under 0, 2, and 21 % initial O 2 concentrations, respectively.These values were similar to those measured by Frame and Casciotti (2010).Using the total amount of N 2 O produced in each experiment, the N 2 O produced from NO − 2 reduction is estimated at 2.5 ± 1.0 nmol (n = 6), 3.8 ± 0.7 nmol (n = 2), 3.9 ± 0.7 nmol (n = 5) under 0, 2, and 21 % initial O 2 concentrations, respectively.This means that the rate of N 2 O production via the NO − 2 reduction pathway does not depend on the amount of NO − 2 produced, which showed an increase of more than 20 times under aerobic conditions compared to anaerobic conditions.This might indicate that activities of NIR or NOR were not enhanced in this study, although we did not directly measure the regulation and activity of the enzymes.The process information provided by SP values enables us to estimate bulk 15 N-enrichment factors (ε, which is approximately equal to δ 15 N product − δ 15 N substrate under the excess supply of substrate) for N 2 O production from NH + 4 by NORand HAO-mediated pathways.If N 2 O is produced only by a NH 2 OH oxidation pathway (X NOR = 0), SP measured = SP HAO and δ 15 N bulk of N 2 O is estimated at −32.9 ‰ from the linear relationship between SP and δ 15 N bulk (Fig. 3).However, if N 2 O is produced only by a NO − 2 reduction pathway (X NOR = 1), SP measured = SP NOR and δ 15 N bulk of N 2 O is estimated at −67.5 ‰.Combining these values with δ 15 N of (NH 4 ) 2 SO 4 used in this study (= −0.34 ‰), we obtain ε HAO = −32.6 ‰ and ε NOR = −67.2‰.The ε N OR value is about 10 ‰ lower than the value estimated from pure culture incubation of Nitrosomonas marina C-113a (Frame and Casciotti, 2010) under several O 2 concentrations ( 15 ε ND = −56.9‰).The cause of the difference could be different experimental approaches (with/without enzymatic reactions) or different species studied.However, it is evident that ε is significantly different between the NH 2 OH oxidation pathway and the NO − 2 reduction pathway.Finally, we discuss the mechanisms that control SP values of N 2 O produced during the enzymatic reactions.Both HAO and NOR enzymes are known to have Fe atoms as active centers, but their structures are different according to functional types.Because the catalytic site of HAO has a single Fe atom (one nuclear center) (Igarashi et al., 1997), it is likely that a single NH 2 OH molecule binds to the center and is oxidized 0 % 2.9 ± 0.9 1.5 ± 0.5 0.82 ± 0.18 2.5 ± 1.0 2 % 8.9 ± 0.5 4.7 ± 0.3 0.42 ± 0.10 3.8 ± 0.7 21 % 18.1 ± 3.6 9.5 ± 1.9 0.22 ± 0.04 3.9 ± 0.7 The catalytic center of the N. europaea HAO for hydroxylamine oxidation is formed by the P460 molecule and the side-chains of three amino acids (Asp291, His292, Tyr358; N. europaea HAO numbering) that surround the distal side of the heme (Igarashi et al. 1996).Although two of the three residues (Asp291 and His292) are completely conserved among all the HAOs, the other (Tyr358) is replaced by Asn in the enzymes of the gamma-AOBs, including the NS58.Recently, the crystal structure of the N. europaea HAO in the substrate-binding state was reported by Maalcke et al. (2014).They proposed that the side-chain of Tyr358 would participate in the second step of the two-step reaction shown in Fig. 4, namely the addition of water molecules to the intermediate species bound on the P460.The replacement of the Tyr to Asn in N. oceani NS58 HAO (Asn368) may obstruct the addition of H 2 O to the intermediate in the HAO reaction.For that reason, it is probable that the N 2 O-generating pathway shown in the Fig. 4 would be more likely to occur in the case of N. oceani NS58 HAO compared to the case of N. europaea HAO.In addition, this explanation is consistent with the present result that N. oceani NS58 generates more N 2 O under the aerobic condition than the anaerobic condition (Fig 3), showing a contrast to the case of N. europaea (Poth and Focht., 1985).
In the case of NOR, three types are known for bacteria: cNOR that accepts electrons from cytochrome c, and qNOR and qCuNOR that accept electrons from quinols (Zumft et al. 2005).Bacterial denitrification is considered to be catalyzed by cNOR, although qNOR and qCuNOR may be responsible for detoxification of NO produced in environments (Hendriks et al. 2000).The active site of NOR enzymes has two Fe atoms (binuclear center) and have similarity among NOR types.In the case of cNOR, the binuclear center consists of non-heme iron (Fe b ) and heme b 3 (Hino et al. 2010), and it is proposed that two NO molecules respectively bind to the two centers simultaneously to form N 2 O (Watmough et al. 2009).This parallel binding mechanism could bring about nearly the same isotope effect for the two N atoms in intermediate ONNO, and if the elimination of the O atom from N-O bonding does not fractionate 15 N within the intermediate molecule, then the SP value of N 2 O would be nearly 0 ‰.

Conclusions
We presented direct evidence that SP values of N 2 O produced by bacterial nitrification and denitrification are controlled by the enzymatic reactions of HAO and NOR during NH 2 OH oxidation and NO reduction, respectively.The SP value does not depend on factors like concentration and isotope ratios of substrate and degree of reaction progress.Using the distinct SP values for HAO-and NOR-related processes, we evaluated the relative contributions from the two pathways of N 2 O production by AOB in pure culture, and showed that they are sensitive to oxygen concentration and that NO − 2 reduction could occur under aerobic conditions.Although further studies are required, this study demonstrates that isotopomer analysis constitutes a powerful tool to investigate N 2 O production pathways in various environments.It will be important to expand these observations beyond nitrifying and denitrifying bacteria to examine the generality of these results.For example, isotopomeric characteristics on fungal denitrification and archaeal nitrification have been reported only recently (Santro et al., 2011;Sutka et al., 2008), and those on fungal co-denitrification are unknown.Combined analysis of several isotopomer ratios should be developed to distinguish NH 2 OH oxidation and fungal denitrification, because SP values of N 2 O from the two pathways have been found to be nearly the same (Sutka et

Figure 1 .
Figure 1.Amounts of N 2 O and NO − 2 produced by HAO extracted from N. oceani under different substrate (NH 2 OH) concentration (expressed as amount in the reaction vessel).The error bar indicates standard deviation for three replicates.
N 2 O produced by the reduction of NO with NOR.About 65.0-149.2% of NO was converted to N 2 O.The concentration of N 2 O was 17 times and 7.3 times higher in the presence of NOR (NOR+) compared to the blank runs without NOR.SP values of N 2 O from the enzymatic reaction showed little variation (−5.9 ± 2.1 ‰) and were lower than SP values observed in the control runs (15.1-16.8‰) (p < 0.05).
Average 4.6 ± 1.4 117.9 ± 36.5 15.6 ± 3.0 12.6 ± 4.1 18.5 ± 2.1 22.4 ± 2.5 −5.9 ± 2.1 Experiments A, C and D were conducted on different dates.Concentration is that in the gas phase (head space) in the vial.N 2 O yield represents the fraction of N atoms converted to N 2 O from substrate NO ([N 2 O-N / NO-N]).Isotopomer ratios for NOR+ were corrected for the blank and were averaged for each experiment.SD represents standard deviation for each experiment.For experiment C, results of experiment D were adopted as the blank, because the blank δ 18 O value obtained in experiment C was significantly different from that obtained in experiments A and D.

Figure 2 .
Figure 2. Amount of NO − 2 and N 2 O produced from 378 µmol of NH + 4 in concentrated cell suspensions of N. oceani under different initial O 2 concentrations (n = 2).

Figure 3 .
Figure 3. Site preference and δ 15 N of N 2 O in concentrated cell suspensions of N. oceani under different initial O 2 concentrations.Error bar indicates the measurement error.

N 2 O
yield represents the fraction of N atoms converted to N 2 O from substrate NH + 4 ([N 2 O-N] / [NH + 4 -N]).* Contribution of NO − 2 reduction to N 2 O production.* * Amount of N 2 O produced by NO − 2 reduction.toNO − 2 .To explain the high SP values of the N 2 O generated by HAO reactions, we propose a two-step reaction mechanism as indicated in Fig.4.If substrate NH 2 OH is supplied in excess or the electron acceptor for the reaction is lacking, however, another NH 2 OH molecule will bind to the same center to form an N-N bond.In this case, the primary-binding NH 2 OH molecule could be more depleted in 15 N than the secondary-binding NH 2 OH molecule by the kinetic isotope effect.The observed positive SP might indicate that the peripheral (β) and central (α) N atoms in the product N 2 O derive from the primary and secondary NH 2 OH, respectively.

Figure 4 .
Figure 4. Two-step model of NH 2 OH oxidation catalyzed by HAO.

Table 1 .
Concentration and isotopomer ratios of N 2 O produced during NH 2 OH oxidation by HAO enzyme extracted from N. oceani with different concentrations of NH 2 OH.Concentration is that in the gas phase (head space) in the vial.N 2 O yield represents the fraction of N atoms converted to N 2 O from substrate NH 2 OH ([N 2 O-N] / [NH 2 OH-N]).Isotopomer ratios for "HAO+" were corrected for the blank.Average and SD (standard deviation) were calculated for samples, including HAO in each condition (n = 3).

Table 2 .
Concentration and isotopomer ratios of N 2 O produced during NH 2 OH oxidation by HAO enzyme extracted from Nitrosococcus oceani and Nitrosomonas europaea with 3 mM of NH 2 OH. 2 O became closer to that produced in the blank experiments.The amount of blank N 2 O was almost constant (70-90 nmol; note that the "yield" listed in Table 1 shows the amount of N 2 O-N relative ferricyanide) might have been depleted under conditions of 1 and 3 mM NH 2 OH because its initial concentration was kept constant.With lower NH 2 OH concentrations (0.1 and 0.3 mM), the production rate of N 2 O was smaller than that of NO − 2 , and the amount of produced N

Table 3 .
Concentration and isotopomer ratios of N 2 O during NO reduction by NOR extracted from Paracoccus denitrificans.

Table 4 .
Estimated contribution of NO − 2 reduction catalyzed by NOR to N 2 O production by N. oceani under different initial O 2 concentrations.

www.biogeosciences.net/11/2679/2014/ Biogeosciences, 11, 2679-2689, 2014 2008
al.,   ).As for the characterization of various microbial N 2 O production processes, studies focused on enzymatic reactions would be effective, as shown in this work.Although the enzymatic pathway for N 2 O production by AOA is still uncertain, combined analysis of genome sequences, physiology, and isotopomer ratios would promise further understanding of microbial N 2 O production mechanisms.