Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 © Author(s) 2021. This work is distributed under the Creative Commons Attribution 4.0 License.

Nitrogen isotopic fractionations during nitric oxide production in an agricultural soil

Zhongjie Yu1,2 and Emily M. Elliott1 1Department of Geology and Environmental Science, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, USA 2Department of Natural Resources and Environmental Sciences, University of Illinois Urbana-Champaign, Urbana, Illinois 61801, USA

Correspondence: Zhongjie Yu ([email protected])

Received: 26 September 2020 – Discussion started: 20 October 2020 Revised: 12 December 2020 – Accepted: 22 December 2020 – Published: 4 February 2021

Abstract. Nitric oxide (NO) emissions from agricultural strained to be 15 ‰ to 22 ‰ and −8 ‰ to 2 ‰, respectively. soils play a critical role in atmospheric chemistry and repre- Production of NO in the oxic and hypoxic incubations was + − sent an important pathway for loss of reactive nitrogen (N) contributed by both NH4 oxidation and NO3 consumption, to the environment. With recent methodological advances, with both processes having a significantly higher NO yield there is growing interest in the natural-abundance N isotopic under O2 stress. Under both oxic and hypoxic conditions, NO 15 + 15 composition (δ N) of soil-emitted NO and its utility in pro- production from NH4 oxidation proceeded with a large η viding mechanistic information on soil NO dynamics. How- (i.e., 55 ‰ to 84 ‰) possibly due to expression of multiple ever, interpretation of soil δ15N-NO measurements has been enzyme-level isotopic fractionations during NH+ oxidation − 4 impeded by the lack of constraints on the isotopic fractiona- to NO2 that involves NO as either a metabolic byproduct tions associated with NO production and consumption in rel- or an obligatory intermediate for NO− production. Adding − 2 evant microbial and chemical reactions. In this study, anoxic NO2 to sterilized soil triggered substantial NO production, 15 (0 % O2), oxic (20 % O2), and hypoxic (0.5 % O2) incuba- with a relatively small η (19 ‰). Applying the estimated tions of an agricultural soil were conducted to quantify the 15η values to a previous δ15N measurement of in situ soil 15 net N isotope effects ( η) for NO production in denitrifica- NOx emission (NOx = NO + NO2) provided promising evi- − 15 tion, nitrification, and abiotic reactions of nitrite (NO2 ) us- dence for the potential of δ N-NO measurements in reveal- ing a newly developed δ15N-NO analysis method. A sodium ing NO production pathways. Based on the observational and − nitrate (NO3 ) containing mass-independent oxygen-17 ex- modeling constraints obtained in this study, we suggest that 17 15 15 cess (quantified by a 1 O notation) and three ammonium simultaneous δ N-NO and δ N-N2O measurements can + 15 (NH4 ) fertilizers spanning a δ N gradient were used in soil lead to unprecedented insights into the sources of and pro- incubations to help illuminate the reaction complexity un- cesses controlling NO and N2O emissions from agricultural derlying NO yields and δ15N dynamics in a heterogeneous soils. soil environment. We found strong evidence for the promi- − nent role of NO2 re-oxidation under anoxic conditions in 15 − controlling the apparent η for NO production from NO3 in denitrification (i.e., 49 ‰ to 60 ‰). These results high- 1 Introduction light the importance of an under-recognized mechanism for − Agricultural production of food has required a tremendous the reversible enzyme NO2 oxidoreductase to control the N isotope distribution between the denitrification products. increase in the application of nitrogen (N) fertilizers since the Through a 117O-based modeling of co-occurring denitrifi- 1960s (Davidson, 2009). In order to maximize crop yields, N − 15 − fertilizers are often applied in excess to agricultural soils, re- cation and NO2 re-oxidation, the η for NO2 reduction to NO and NO reduction to nitrous oxide (N2O) were con- sulting in loss of reactive N to the environment (Galloway et al., 2003). Loss of N in the form of gaseous nitric oxide

Published by Copernicus Publications on behalf of the European Geosciences Union. 806 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

(NO) has long been recognized for its adverse impacts on for source partitioning soil-emitted NO under dynamic envi- air quality and human health (Veldkamp and Keller, 1997). ronmental conditions. Once emitted to the atmosphere, NO is rapidly oxidized to Natural-abundance stable N and oxygen (O) isotopes in nitrogen dioxide (NO2), and these compounds (collectively N-containing molecules have long provided insights into the referred to as NOx) drive production and deposition of atmo- sources and relative rates of biogeochemical processes com- − spheric nitrate (NO3 ) (Calvert et al., 1985) and play a critical prising the N cycle (Granger and Wankel, 2016). The unique role in the formation of tropospheric ozone (O3) – a toxic air power of stable isotope ratio measurements stems from the pollutant and potent greenhouse gas (Crutzen, 1979). Despite distinct partitioning of isotopes between chemical species or the observations that emission of NO from agricultural soils phases, known as isotopic fractionation. Thus, in order to ex- can sometimes exceed that of nitrous oxide (N2O) – a climat- tract the greatest information from the distributions of iso- ically important trace gas primarily produced from reduction topic species, a rigorous understanding of the direction and of NO in soils (Liu et al., 2017), NO is frequently overlooked magnitude of isotopic fractionations associated with each rel- in soil N studies due to its high reactivity and transient pres- evant transformation is required. Both kinetic and equilib- ence relative to N2O (Medinets et al., 2015). Consequently, rium isotope effects can lead to isotopic fractionations be- the contribution of soil NO emission to contemporary NOx tween N-bearing compounds in soils (Granger and Wankel, inventories at regional to global scales is highly uncertain 2016; Denk et al., 2017). During kinetic processes, isotopic (e.g., ranging from 3 % to > 30 %) (Hudman et al., 2010; fractionation occurs as a result of differences in the reac- Vinken et al., 2014) and remains the subject of much current tion rates of isotopically substituted molecules (i.e., isotopo- debate (Almaraz et al., 2018; Maaz et al., 2018). logues), leading to either enrichment or, in a few rare cases, As the central hub of the biogeochemical N cycle, NO can depletion of heavy isotopes in the reaction substrate (Fry, be produced and consumed in numerous microbial and chem- 2006; Casciotti, 2009). The degree of kinetic isotope frac- ical reactions in soils (Medinets et al., 2015). Among these tionation can be quantified by a kinetic isotope fractionation processes, nitrification and denitrification are the primary factor (αk), which is often represented by the ratio of reac- sources responsible for NO emission from N-enriched agri- tion rate constants of light isotopologues to that of heavy cultural soils (Firestone and Davidson, 1989). Denitrifica- isotopologues. In this definition, αk is larger than 1 for nor- − − tion is the sequential reduction of NO3 and nitrite (NO2 ) to mal kinetic isotope fractionation. For equilibrium reactions, NO, N2O, and dinitrogen (N2) and can be mediated by a di- equilibrium isotope fractionation arises from differences in versity of soil heterotrophic microorganisms (Zumft, 1997). the zero-point energies of two species undergoing isotopic The enzymatic system of denitrification comprises a series exchange, leading to enrichment of heavy isotopes in the − of dedicated reductases whereby NO2 reductase (NIR) and more strongly bonded form (Fry, 2006; Casciotti, 2009). NO reductase (NOR) are the key enzymes that catalyze pro- In this case, the isotope ratios of two species at equilib- duction and reduction of NO, respectively (Ye et al., 1994). rium are defined by an equilibrium isotope fractionation fac- As such, NO is often viewed as a free intermediate of the tor (αeq), which is also related to the kinetic isotope frac- denitrification process (Russow et al., 2009). In comparison, tionation factors of forward and backward equilibrium reac- nitrification is a two-step aerobic process in which oxida- tions (Fry, 2006). By convention, isotopic fractionation can − tion of ammonia (NH3) to NO2 is mediated by ammonia- be expressed in units of per mill (‰) as an isotope effect oxidizing bacteria (AOB) or archaea (AOA), while the sub- () :  = (α−1)×1000. Nevertheless, in a heterogeneous soil − − sequent oxidation of NO2 to NO3 is performed by nitrite- environment, expression of intrinsic kinetic and equilibrium oxidizing bacteria (NOB) (Lehnert et al., 2018). Although isotope effects for biogeochemical N transformations is of- production of NO during the nitrification process has been ten limited due to transport limitation in soil substrates, the linked to NH3 oxidation (Hooper et al., 2004; Caranto and multi-step nature of transformation processes, and the pres- − Lancaster, 2017) and NO2 reduction by AOB/AOA-encoded ence of diverse soil microbial communities that transform N NIR (Wrage-Mönning et al., 2018), the metabolic role of NO via parallel and/or competing reaction pathways (Maggi and in AOB and AOA remains ambiguous, making it difficult to Riley, 2010). As such, interpretation of N isotope distribution elucidate the enzymatic pathways driving NO release by ni- in soils has largely relied on measuring net isotope effects trification (Beeckman et al., 2018; Stein, 2019). Additionally, (η), which are often characterized by incubating soil samples NO can also be produced from abiotic reactions involving under environmentally relevant conditions, that favor expres- − soil NO2 or its protonated form – nitrous acid (HNO2) (Ven- sion of intrinsic isotope effects for specific N transformations terea et al., 2005; Lim et al., 2018). However, despite empir- (Lewicka-Szczebak et al., 2014). For example, it has been ical evidence for the dependence of soil NO emission on soil shown that the net N isotope effects for N2O production in N availability and moisture content (Davidson and Verchot, soil nitrification, denitrification, and abiotic reactions are dis- 2000), the source contribution of soil NO emission across tinctively different under certain soil conditions (Denk et al., temporal and spatial scales is poorly understood (Hudman 2017), rendering natural-abundance N isotopes of N2O a use- et al., 2012). This is largely due to the lack of a robust means ful index for inferring sources of N2O in agricultural soils (Toyoda et al., 2017).

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 807

While the isotopic dynamics underlying soil N2O emis- tion of NO production with changing O2 availability. By in- sions has been extensively studied, there has been little in- tegrating multi-species measurements of N and O isotopes in vestigation into the N isotopic composition (notated as δ15N an isotopologue-specific modeling framework, we were able 15 in units of ‰; δ = ((Rsample/Rstandard) − 1) × 1000) of soil- for the first time to unambiguously link the yield and δ N emitted NO due to measurement difficulties (Yu and Elliott, variations of soil-emitted NO to nitrification and denitrifica- 2017). Using a tubular denuder that trapped NO released tion carried out by whole soil microbial communities and to + from urea and ammonium (NH4 )-fertilized soils, Li and characterize the net isotope effects for NO production from 15 − + − Wang (2008) revealed a gradual increase in δ N-NO from soil NO3 , NH4 , and NO2 under different redox conditions. −49 ‰ to −19 ‰ and simultaneous 15N enrichment in soil The quantified isotope effects are discussed in the context of + − NH4 and NO3 over a 2-week laboratory incubation. Sim- chemical and enzymatic pathways leading to net NO produc- ilar δ15N variations (i.e., −44 ‰ to −14 ‰) were recently tion in the soil environment and are applied to a previous field reported for in situ soil NOx emission in a manure-fertilized study (Miller et al., 2018) to provide implications for tracing cornfield (Miller et al., 2018). Moreover, the magnitude of the sources of NO emission from agricultural soils. 15 δ N-NOx measured in this study depended on manure appli- cation methods, implying that NOx was mainly sourced from + 2 Materials and methods nitrification of manure-derived NH4 (Miller et al., 2018). Based on a newly developed soil NO collection system that 2.1 Soil characteristics and preparation quantitatively converts soil-emitted NO to NO2 for collec- tion in triethanolamine (TEA) solutions, our previous work Soil samples used in this study were collected in July 2017 15 − demonstrated substantial variations in δ N-NO ( 54 ‰ to from a conventional corn–soybean rotation field in central − 37 ‰) in connection with changes in moisture content in Pennsylvania, USA, managed by the USDA (Agricultural a forest soil (Yu and Elliott, 2017). Furthermore, the mea- Research Service, University Park, PA, USA). The soil is 15 − sured in situ δ N-NO values spanned a wide range ( 60 ‰ a well-drained Hagerstown silt loam (fine, mixed, semiac- to −23 ‰) and were highly sensitive to added N substrates + − − tive, mesic Typic Hapludalfs) with sand, silt, and clay con- (i.e., NH4 , NO3 , and NO2 ), indicating that NO produced tent of 21 %, 58 %, and 21 %, respectively. The sampled sur- 15 from different sources may bear distinguishable δ N im- face layer (0–10 cm) had a bulk density of 1.2 gcm−3 and a prints (Yu and Elliott, 2017). Nevertheless, despite the po- pH (1 : 1 water) of 5.7. Total N content was 0.2 % and δ15N 15 tential of δ N-NO measurements in providing mechanistic of total N was 5.3 ‰. Soil C : N ratio was 11.4 and organic 15 information on soil NO dynamics, interpretation of δ N-NO carbon content was 1.8 %. In the laboratory, soils were ho- has been largely impeded by the knowledge gap as to how mogenized and sieved to 2 mm (but not air-dried) and then 15 δ N-NO is controlled by N isotopic fractionations during stored in resealable plastic bags at 4 ◦C until further analyses NO production and consumption in soils. and incubations. Gravimetric water content of the sieved and To this end, we conducted a series of controlled incuba- −1 + homogenized soils was 0.14 g H2Og . Indigenous NH4 tion experiments to quantify the net N isotope effects for NO − −1 and NO3 concentrations were 0.7 and 19.8 µgNg , re- production in an agricultural soil. Replicate soil incubations spectively. Throughout this paper, soil N concentrations, NO 15 were conducted to measure the yield and δ N of soil-emitted fluxes, and N transformation rates are expressed on the basis NO under anoxic (0 % O2), oxic (20 % O2), and hypoxic ◦ − of soil ovendry (105 C) weight. (0.5 % O2) conditions, respectively. A sodium NO3 fertilizer mined in the Atacama Desert, Chile (Yu and Elliott, 2018), 2.2 Net NO production and collection of NO for δ15N was used to amend the soil in all three incubation experi- analysis − ments. This Chilean NO3 originated from atmospheric de- position and thus contained an anomalous 17O excess (quan- The recently developed soil dynamic flux chamber (DFC) tified by the 117O notation) as a result of mass-independent system was used to measure net NO production rates and isotopic fractionations during its photochemical formation to collect soil-emitted NO for δ15N analysis (Yu and El- in the atmosphere (Michalski et al., 2004). Because iso- liott, 2017). A schematic of the DFC system is shown in − topic fractionations during biogeochemical NO3 production Fig. 1a. Detailed development and validation procedures for and consumption are mass-dependent, 117O-NO− is a con- the NO collection method were presented in Yu and El- −3 servative tracer of gross nitrification and NO3 consump- liott (2017). Briefly, custom-made flow-through incubators tion and provides a quantitative benchmark for disentangling modified from 1 L Pyrex medium bottles (13951 L, Corning, 15 − 18 − isotopic overprinting on δ N-NO3 and δ O-NO3 during USA) were used for all the incubation experiments (Fig. 1b). co-occurring nitrification and denitrification (Yu and Elliott, Each incubator was stoppered with two 42 mm Teflon septa 2018) (see Sect. S1 in the Supplement for more details). secured by an open-topped screw cap and equipped with + As additional tracers, three isotopically different NH4 fer- two vacuum valves for purging and closure of the incu- tilizers were used in parallel treatments of the oxic and hy- bator headspace. To measure net NO production from en- poxic incubations to quantify the nitrifier source contribu- closed soil samples, a flow of NO-free air with desired O2 https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 808 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil content was directed through the incubator into a chemi- was then supplied at 1 SLPM for continuous NO concentra- luminescent NO–NOx–NH3 analyzer (model 146i, Thermo tion measurement and collection. Samples from the replicate Fisher Scientific) (Fig. 1a) (Yu and Elliott, 2017). Outflow batches were measured successively. NO concentration was monitored continuously until steady, Following the completion of measurement and collection and then the net NO production rate was determined from of each sample, the incubator was opened from the top and the flow rate and steady-state NO concentration. To collect the soil was combined with 500 mL deionized water for ex- NO for δ15N analysis, a subsample of the incubator outflow traction of soil NO− and NO− (McKenney et al., 1982). Be- − 3 2 was forced to pass through a NO collection train (Fig. 1a) cause NO2 accumulation was found in pilot experiments, where NO is converted to NO2 by excess O3 (∼ 3 ppm) in deionized water, rather than routinely used KCl solutions, a Teflon reaction tube (9.5 mm i.d., ca. 240 cm length) and was used for the extraction to ensure accurate NO− deter- 2 − subsequently collected in a 500 mL gas washing bottle con- mination (Homyak et al., 2015). To extract soil NO3 and taining a 20 % (v/v, 70 mL) TEA solution (Yu and Elliott, NO−, the soil slurry was agitated vigorously on a stir plate − 2 2017). The collection products were about 90 % NO2 and for 10 min and then centrifuged for 10 min at 3400 g. The re- 10 % NO− (Yu and Elliott, 2017). Results from comprehen- sultant supernatant was filtered through a sterile 0.2 µm filter 3 − sive method testing showed that the NO collection efficiency (Homyak et al., 2015). In light of high NO2 concentrations was 98.5% ± 3.5 % over a wide range of NO concentrations observed in the pilot experiments, the filtrate was divided into (12 to 749 ppb) and environmental conditions (e.g., tempera- two 60 mL Nalgene bottles, with one of the bottles receiving ture from 11 to 31 ◦C and relative humidity of the incubator sulfamic acid to remove NO− (Granger and Sigman, 2009). − 2 − outflow from 27 % to 92 %) (Yu and Elliott, 2017). More- This NO2 -removed sample was used for NO3 isotope anal- over, it was confirmed that high concentrations of ammonia ysis, while the other sample without sulfamic acid treatment − − (NH3) (e.g., 500 ppb) and nitrous acid (HONO) (removed by was used for determining NO2 and NO3 concentrations and an inline HONO scrubber, Fig. 1a) in the incubator outflow combined δ15N analysis of NO− +NO−. Two important con- − − 2 3 do not interfere with NO collection (Yu and Elliott, 2017). trol tests, based on NO2 /NO3 spiking and acetylene (C2H2) addition, were conducted to evaluate the robustness of the 2.3 Anoxic incubation adopted soil incubation and extraction methods. The results confirmed that the water extraction method was robust for To prepare for the anoxic incubation, the soil samples were determining concentrations and isotopic composition of soil − − − + spread out on a covered tray for pre-conditioning under room NO3 and NO2 and that aerobic NO3 production from NH4 temperature (21 ◦C) for 24 h. Next, the soil was amended oxidation was negligible during the soil incubation and ex- − 15 = ± with the Chilean NO3 fertilizer (δ N 0.3‰ 0.1‰, traction procedures (Tables S1 and S2 in the Supplement; see δ18O = 55.8‰±0.1‰, 117O = 18.6‰±0.1‰) to achieve Sect. S2 for more details). − −1 a fertilization rate of 35 µg NO3 -N g and a target soil wa- −1 ter content of 0.21 g H2Og (equivalent to 46 % water-filled 2.4 Oxic and hypoxic incubations pore space, WFPS). The fertilized soil samples were thor- oughly homogenized using a glass rod in the tray. A total of The same pre-conditioning and fertilization protocol de- 100 g (dry-weight equivalent) of soil was then weighed into scribed for the anoxic incubation was used for the oxic + each of eight incubators, resulting in a soil depth of about and hypoxic incubations. Three isotopically different NH4 1.5 cm. The incubators were connected in parallel using a fertilizers were used in parallel treatments of each incuba- 15 + = 15 Teflon purging manifold (Fig. 1c), vacuumed and filled with tion experiment: (1) δ N-NH4 1.9 ‰ (low N enrich- 15 + = 15 ultra-high-purity N2 for three cycles, and incubated in the ment), (2) δ N-NH4 22.5 ‰ (intermediate N enrich- 15 + = 15 dark with a continuous flow of N2 circulating through each ment), and (3) δ N-NH4 45.0 ‰ (high N enrichment). of the eight incubators at 0.015 standard liters per minute An off-the-shelf ammonium sulfate ((NH4)2SO4) reagent 15 + (SLPM). The sample fertilization and preparation procedures was used in the low-δ N-NH4 treatment, while the fertiliz- were repeated three times to establish three batches of repli- ers with intermediate and high enrichment of 15N were pre- cate samples, leading to 24 soil samples in total for the anoxic pared by gravimetrically mixing the (NH4)2SO4 reagent with + 15 + = incubation. NH4 reference materials IAEA-N2 (δ N-NH4 20.3 ‰) 15 + = The first NO measurement and collection event was con- and USGS26 (δ N-NH4 53.7 ‰). In both oxic and hy- 15 + ducted 24 h after the onset of the anoxic incubation, and poxic incubations, each of the three δ N-NH4 treatments daily sampling was conducted thereafter. At each sampling consisted of three replicate sample batches where each batch event, one incubator from each replicate sample batch was consisted of eight samples, resulting in 72 samples for each isolated by closing the vacuum valves, removed from the incubation experiment. purging manifold, and then measured using the DFC system. At the onset of each incubation experiment, soil sam- To prevent O2 contamination by residual air in the DFC sys- ples (100 g dry-weight equivalent) were amended with the + −1 − tem, the DFC system was evacuated and flushed with N2 five desired NH4 fertilizer (90 µgNg ) and the Chilean NO3 −1 times before the vacuum valves were re-opened. A flow of N2 fertilizer (15 µgNg ) to the target soil water content of

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 809

Figure 1. (a) Schematic of the DFC system (not to scale) consisting of the following: (1) zero air tank, (2) N2 tank, (3) mass flow controller, (4) Nafion moisture exchanger, (5) flow-through incubator, (6–9) needle valves for controlling vacuum and flushing of the DFC system, (10) HONO scrubber, (11) diaphragm pump, (12) Teflon reaction tube, (13) gas washing bottle containing TEA solution, (14) NO–NOx–NH3 analyzer, (15) O3 generator, and (16) in-line PTFE particulate filter assembly. (b) Photo of the flow-through incubator. (c) Photo of the Teflon purging manifold for connection of the incubators in parallel.

−1 0.21 g H2Og (46 % WFPS). Following the amendment, collection (Yu and Elliott, 2017). The flow rate of purg- two soil samples from each replicate batch were immediately ing air (20 % O2 for the oxic incubation and 0.5 % O2 for extracted – one with 500 mL of deionized water for soil NO− the hypoxic incubation) during the DFC measurement was − 2 and NO3 using the extraction method described above and 0.25 SLPM to each incubator. Following the NO measure- the other one with 500 mL of a 2 M KCl solution for determi- ment and collection, two soil samples from each replicate nation of soil NH+. The remaining samples were incubated batch were extracted for determination of soil NO−/NO− 4 + 3 2 under desired O2 conditions until further measurements. In (500 mL deionized water) and NH4 (500 mL 2M KCl), re- the oxic incubation, the incubators were connected in paral- spectively. Because NO concentrations were too low for reli- lel using the purging manifold and continuously flushed by able NO collection at 72 h after the onset of the incubations, a flow of zero air (20 % O2 + 80 % N2). In the hypoxic in- only net NO production rates were measured using the re- cubation, a flow of synthetic air with 0.5 % O2 content (bal- maining two soil samples in each replicate batch. anced by 99.5 % N2) was used to incubate the soil samples. The synthetic air was generated by mixing the zero air with 2.5 Abiotic NO production ultra-high-purity N2 using two mass flow controllers (model SmartTrak 50, Sierra Instruments). The potential for NO production from abiotic reactions was Replicate NO measurement and collection events were assessed using sterilized soil samples. Soil samples (100 g conducted at 24, 48, and 72 h following the onset of the oxic dry-weight equivalent) were weighed into the incubators and and hypoxic incubations. Because net NO production rates then autoclaved at 121 ◦C and 1.3 atm for 30 min. The au- were low under oxic and hypoxic conditions, all remain- toclaved samples were pre-incubated under oxic and anoxic ing soil samples in each replicate batch were connected in conditions, respectively, for 24 h and then fertilized with parallel for NO measurement and collection using the DFC − − −1 the Chilean NO3 (35 µg NO3 -N g ) or the lab (NH4)2SO4 system. This parallel connection ensured high outflow NO + −1 (90 µg NH4 -N g ). The fertilizer solutions were added to concentrations (i.e., > 30 ppb) required for quantitative NO the soil surface through the Teflon septa using a sterile sy- https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 810 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil ringe equipped with a 25-gauge needle. These samples were 1 : 1 mixture of USGS34 and USGS35. then measured periodically for net NO production. Because   17   18  − 17 δ O δ O NO2 was found to accumulate during the anoxic incubation 1 O = ln + 1 − 0.52ln + 1 × 1000 (1) (see below), four soil samples were sterilized, pre-incubated 1000 1000 under anoxic condition, and then fertilized with a NaNO2 so- 17 − − The precision of the 1 O analysis of USGS35 and the lution (δ15N-NO = 1.4‰ ± 0.2‰) (8 µgNg 1) for imme- 2 − USGS35:USGS34 mixture is ±0.3 ‰ (Yu and Elliott, 2018). diate NO measurement and collection. These NO -amended 2 Following Kaiser et al. (2007), the measured 117O-NO− was samples were thereafter incubated under anoxic conditions 3 used in the reduction of molecular isotope ratios of N O to and measured periodically for net NO production until unde- 2 correct for the isobaric interference (i.e., m/z 45) on the mea- tectable. 15 − sured δ N-NO3 . δ15N of NH+ in the KCl extracts was measured by cou- 2.6 Chemical and isotopic analyses 4 pling the NH3 diffusion method (Zhang et al., 2015) and the − − hypobromite (BrO ) oxidation method (Zhang et al., 2007) Soil NO concentrations were determined using a Dionex 3 with the denitrifier method (Felix et al., 2013). Briefly, an ion chromatograph ICS-2000 with a precision of (1σ ) of + − − aliquot of soil KCl extract with 60 nmol NH was pipetted ±5.0 µgNL 1. Soil NO concentrations were analyzed us- 4 2 into a 20 mL serum vial containing an acidified glass fiber ing the Griess–Ilosvay colorimetric reaction with a precision ± −1 + disk. The solution was made alkaline by adding magnesium of 1.2 µgNL . Soil NH4 concentrations were measured oxide (MgO) to volatilize NH3, which was subsequently cap- using a modified fluorometric o-phthaldialdehyde (OPA) + tured on the acidic disk as NH4 . After incubation under method for soil KCl extracts (Kang et al., 2003) with a ◦ + ± −1 − + − 37 C for 10 d, NH4 was eluted from the disk using deion- precision of 7.0 µgNL . NO2 NO3 concentration in − − ized water, diluted to 10 µM, oxidized by BrO to NO2 , and the TEA collection samples was measured using a modified 15 − − finally measured for δ N as NO at 20 nmol using the deni- spongy cadmium method with a precision of ±1.6 µgNL 1 2 trifier method. International NH+ reference standards IAEA- (Yu and Elliott, 2017). 4 N1, USGS25, and USGS26 underwent the same preparation The denitrifier method (Sigman et al., 2001; Casciotti 15 18 − procedure as the soil KCl extracts and were used along with et al., 2002) was used to measure δ N and δ O of NO in − − − 3 − the NO reference standards to correct for blanks and in- the NO -removed soil extracts and the δ15N of NO +NO 3 2 3 2 strument drift. The precision of the δ15N-NH+ analysis is in the extracts without sulfamic acid treatment. In brief, a 4 ±0.5 ‰ (Yu and Elliott, 2018). denitrifying bacterium (Pseudomonas aureofaciens) lacking δ15N of NO collected in the TEA solution was measured the N2O reductase enzyme was used to convert 20 nmol of − following the method described in Yu and Elliott (2017). NO into gaseous N O. The N O was then purified in a 3 2 2 Briefly, the TEA collection samples were first neutralized series of chemical traps, cryo-focused, and finally analyzed with 12 N HCl to pH ∼ 7, and then 10 to 20 nmol of the on a GV Instruments Isoprime continuous flow isotope ra- collected product NO− + NO− was converted to N O us- tio mass spectrometer (CF-IRMS) at m/z 44, 45, and 46 at 2 3 2 ing the denitrifier method. In light of the low δ15N values the University of Pittsburgh Regional Stable Isotope Labora- of soil-emitted NO and the presence of NO− as the dom- tory for Earth and Environmental Science Research where 2 inant collection product, a low-δ15N-NO− isotopic standard all isotope analyses were conducted for this study. Inter- 2 − (KNO , RSIL20, USGS Reston; δ15N = −79.6 ‰) was used national NO reference standards IAEA-N3, USGS34, and 2 3 together with the international NO− reference standards to USGS35 were used to calibrate the δ15N and δ18O analy- 3 calibrate the δ15N-NO analysis. Following the identical treat- ses. The long-term precision is ±0.3 ‰ and ±0.5 ‰, respec- ment principle, we prepared the isotopic standards in the tively, for the δ15N and δ18O analyses. Because the denitrifier − − same matrix (i.e., 20 % TEA) as the collection samples and method does not differentiate NO and NO for the δ15N − 3 2 matched both the molar N amount and injection volume analysis, δ15N of NO was estimated using an isotopic mass − 2 (±5 %) between the collection samples and the standards to balance when NO accounted for a significant fraction of the − −2 minimize the blank interferences associated with the bacte- total NO + NO pool. 3 −2 rial medium and the TEA solution. The precision and accu- 117O of NO was measured using the coupled bacterial 3 racy of the δ15N-NO analysis, determined by repeated sam- reduction and thermal decomposition method described by pling of an analytical NO tank (δ15N-NO = −71.4‰) un- Kaiser et al. (2007). The denitrifying bacteria were used to − der diverse collection conditions, is ±1.1 ‰ (Yu and Elliott, convert 200 nmol of NO to N O, which was subsequently 3 2 2017). converted to O2 and N2 by reduction over a gold surface at ◦ 800 C. The produced O2 and N2 were separated using a 5 Å molecular sieve gas chromatograph, and the O2 was then an- alyzed for δ17O and δ18O using the CF-IRMS. 117O was calculated from the measured δ17O and δ18O using Eq. (1) (see Sect. S1) and calibrated by USGS34, USGS35, and a

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 811

3 Results constant over the entire incubations (i.e., variations < 4 %), − and soil NO2 concentration was below the detection limit 15 + Sixty-three NO collection samples were obtained from the in both incubations. In the oxic incubation, δ N-NH4 val- incubation experiments. The NO collection efficiency calcu- ues uniformly increased by 8.6 ‰ to 13.1 ‰ under all three − + − 15 + 15 − lated based on the measured NO2 NO3 concentration in δ N-NH4 treatments (Fig. 3e), while δ N-NO3 values var- 15 + the TEA solution and the theoretical concentration based on ied distinctly, depending on the initial δ N-NH4 values 15 − the measured net NO production rate (Yu and Elliott, 2017) (Fig. 3d). Specifically, δ N-NO3 values increased by 7.8 ‰ ± 15 + was on average 99.1% 3.7%. Out of the 63 collection sam- and decreased by 10.9 ‰ under the high and low δ N-NH4 ples, four samples had a NO collection efficiency lower than treatments, respectively, and remained relatively constant un- 15 + 95 %. These samples were excluded from further data anal- der the intermediate δ N-NH4 treatment (Fig. 3d). Limited 15 + ysis and interpretation. The measured N concentrations, net increases in δ N-NH4 values (< 2 ‰) were observed un- 15 + NO production rates, and isotope data from all the incubation der all three δ N-NH4 treatments in the hypoxic incubation 15 − experiments are available in Tables S5 to S11. (Fig. 3e). Correspondingly, variations in δ N-NO3 values were much smaller in the hypoxic incubation compared to 3.1 Anoxic incubation those revealed in the oxic incubation (Fig. 3d). In both oxic and hypoxic incubations, δ18O-NO− (Fig. 3g) and 117O- − − 3 During the anoxic incubation, soil NO3 concentration NO3 (Fig. 3h) values decreased progressively under all three decreased linearly from 49.3 ± 0.1 to 23.1 ± 0.2 µgNg−1 δ15N-NH+ treatments, although the rates of decrease were − 4 (Fig. 2a), while NO2 concentration increased linearly from significantly higher in the oxic incubation (Fig. 3g and h). 0.4±0.1 to 6.9±0.1 µgNg−1 (Fig. 2b). The net NO produc- The net NO production was significantly higher in the −1 −1 tion rate (fNO-anoxic) increased progressively from the first hypoxic incubation (fNO-hypoxic; 9.0 to 10.4 ngNg h ) −1 −1 −1 −1 sampling day (72 ± 8 ngNg h ) to sampling day 5 and than in the oxic incubation (fNO-oxic; 7.1 to 8.5 ngNg h ) then stabilized at about 82 ngNg−1 h−1 (Fig. 2c). (Fig. 3c). The measured δ15N-NO values ranged from 15 − 15 ± − ± − ± δ N-NO3 and δ N-NO values increased from 4.7‰ 16.8‰ 0.3‰ to 54.9‰ 0.8‰ in the oxic incuba- 0.3‰ to 38.7‰±1.5‰ and −44.7‰±0.3‰ to −22.8‰± tion and from −21.3‰ ± 0.0‰ to −51.4‰ ± 0.4‰ in the 2.2‰, respectively, over the anoxic incubation (Fig. 2d and hypoxic incubation (Fig. 3f). Pooling all the δ15N-NO mea- 15 − 15 15 + f). The difference between δ N-NO3 and δ N-NO values surements, we found that δ N values between NH4 and NO 15 + increased significantly from 49.4 ‰ to 59.5 ‰ toward the differed from 58.9 ‰ to 70.7 ‰ across the three δ N-NH4 end of the incubation (Fig. 2d and f). Based on the closed- treatments in the oxic incubation and from 50.4 ‰ to 69.6 ‰ system Rayleigh model, the apparent N isotopic fractiona- in the hypoxic incubation (Fig. 4). In both incubations, the − ± 15 + tion during NO3 consumption was estimated to be 43.3‰ largest difference was observed under the high-δ N-NH4 15 − 0.9‰ (Fig. S3 in the Supplement). δ N-NO2 was estimated treatment, while the smallest difference was observed un- − 15 + for samples collected in the last 3 sampling days where NO2 der the low δ N-NH4 treatment. Under both oxic and hy- −+ − accounted for > 15 % of the NO3 NO2 pool. The estimated poxic conditions, there was a significant linear relationship 15 − − ± − ± 15 15 + δ N-NO2 values were 6.9‰ 3.7‰, 6.0‰ 2.5‰, between the measured δ N-NO and δ N-NH4 values from − ± 15 + and 0.9‰ 1.3‰, respectively (Fig. 2e). Although lim- all three δ N-NH4 treatments (Fig. 4). The slope of the lin- 15 − ± ± ± ited to the last 3 sampling days, δ N-NO2 was lower than ear relationship is 0.78 0.03 ( 1 SE) and 0.61 0.05 for 15 − δ N-NO3 by 33.6 ‰ to 37.9 ‰ (Fig. 2d and e) but was the oxic and hypoxic incubations, respectively (Fig. 4). higher than the concurrently measured δ15N-NO values by a relatively constant offset of 21.5‰ ± 0.7‰ (Fig. 2e and 3.3 Abiotic NO production 18 − f). Surprisingly, both δ O-NO values (33.4‰ ± 0.2‰ to − + 3− ± 17 ± Addition of NO3 or NH4 to the sterilized soil did not re- 23.1‰ 0.3‰) and 1 O-NO3 values (10.0‰ 0.2‰ to 0.7‰±0.2‰) decreased progressively over the course of the sult in detectable NO production under either oxic or anoxic 15 condition. Immediate NO release was, however, triggered by anoxic incubation and were entirely decoupled from δ N- − − NO2 addition under anoxic conditions (Fig. 5a). The abi- NO3 (Fig. 2g and h). otic NO production rate (fNO-abiotic) reached a steady level ± −1 −1 − 3.2 Oxic and hypoxic incubations of 83 5 ngNg h several minutes after the NO2 addi- tion and then decreased exponentially to < 3 ngNg−1 h−1 + over the following 8 d (Fig. 5a). The natural logarithm of Over the oxic incubation, soil NH4 concentration decreased − fNO-abiotic showed a linear relationship with time (Fig. 5b). linearly with increasing NO concentration under all three − + 3 15 δ15N-NH treatments (Fig. 3a and b). In the hypoxic in- The NO produced following the NO2 addition had a δ N 4 + − − ± 15 cubation, changes in NH and NO concentrations were value of 17.8‰ 0.4‰, giving rise to a δ N offset be- 4 3 − ± more limited, although the linear trends were still evident tween NO2 and NO of 19.2‰ 0.5‰. (Fig. 3a and b). Under both oxic and hypoxic conditions, the + − total concentration of soil NH4 and NO3 remained nearly https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 812 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

− − 15 − − Figure 2. Measured and modeled concentrations of NO3 (a) and NO2 (b); net NO production rate (c); δ N values of NO3 (d), NO2 (e), 18 17 − and NO (f); and δ O (g) and 1 O (h) of NO3 during the anoxic incubation.

4 Discussion

= × × [ −] Because interpretations of the results from the incubation ex- fNO-abiotic sabiotic kabiotic NO2 t (2) periments build upon each other, here we discuss the results − − − × [NO ] = [NO ] e kabiotic t (3) from incubation of the sterilized soils (hereafter, abiotic in- 2 t 2 0 cubation), anoxic incubation, and oxic/hypoxic incubations In Eqs. (2) and (3), t is time; kabiotic is the pseudo-first-order − successively. rate constant for NO loss; sabiotic is the apparent stoichio- 2 − − metric coefficient for NO production from NO2 ; and [NO2 ]t 4.1 Reaction characteristics and N isotopic − − = and [NO2 ]0 are NO2 concentration at time t and t 0 in the fractionation during abiotic NO production sterilized soil, respectively. Combining Eqs. (2) and (3) and − then log-transforming both sides yield The immediate release of NO upon the addition of NO2 highlights the chemically unstable nature of NO− and the 2 ln(f ) = − k × t critical role of chemical NO− reactions in driving soil NO NO-abiotic abiotic 2 + × × [ −] emissions (Venterea et al., 2005; Lim et al., 2018). The strong ln(sabiotic kabiotic NO2 0). (4) linearity between ln(f ) and time (Fig. 5b) suggests NO-abiotic According to Eq. (4), k and s are estimated apparent first-order kinetics for the abiotic NO production abiotic abiotic − using the slope and intercept of the linear regression from NO2 (Eqs. 2 and 3) (McKenney et al., 1990). − = of ln(fNO-abiotic) vs. time (Fig. 5b). Given [NO2 ]0 −1 8 µgNg , sabiotic and kabiotic are estimated to be 0.52±0.05

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 813

− + 15 − + Figure 3. Measured and modeled concentrations of NO3 (a) and NH4 (b); net NO production rate (c); δ N values of NO3 (d), NH4 (e), 18 17 − 15 + and NO (f); and δ O (g) and 1 O (h) of NO3 under the three δ N-NH4 treatments (differed by color) of the oxic (open symbols) and hypoxic (solid symbols) incubations.

± ± −1 − ( SE) and 0.019 0.002 h , respectively, suggesting that 2005). However, given the complete NO2 consumption in ± − NO accounted for 52% 5% of the reacted NO2 during the the abiotic incubation, HNO2 decomposition confined to abiotic incubation. The estimated kabiotic is within the range acidic microsites could not account for all observed NO pro- −1 − (i.e., 0.00055 to 0.73 h ) derived by a recent study based on duction. Under anoxic conditions, NO2 /HNO2 can also be soil samples spanning a wide range of pH values (3.4 to 7.2) stoichiometrically reduced to NO by transition metals (e.g., (Lim et al., 2018). Based on the estimated kabiotic, 97 % of Fe(II)) and diverse organic molecules (e.g., humic and ful- − the added NO2 was lost by the end of the abiotic incubation. vic acids, lignins, and phenols) in a process termed chemo- Several reaction pathways with distinct stoichiometry have denitrification (Zhu-Barker et al., 2015). The produced NO − been proposed for abiotic NO production from NO2 in soils. from chemo-denitrification can undergo further reduction to Under acidic soil conditions, self-decomposition of HNO2 form N2O and N2 (Zhu-Barker et al., 2015). In addition, both − produces NO and nitric acid (HNO3) with a stoichiometric NO2 and NO in soil solution can be consumed as nitroso HNO2-to-NO ratio ranging from 0.5 to 0.66 (i.e., 1 mole donors in abiotic nitrosation reactions, resulting in N incor- of HNO2 produces 0.5 to 0.66 moles of NO) (Van Cleem- poration into soil organic matter (Heil et al., 2016; Lim et al., put and Samater, 1995). Although at pH 5.7 HNO2 consti- 2018). Therefore, our observation that about half of the re- − − tuted < 1 % of the NO + HNO2 pool in this soil, HNO2 acted NO was recovered as NO may result from multiple 2 2 − decomposition can occur on acidic clay mineral surfaces, competing NO2 sinks, parallel NO-producing pathways, and even though bulk soil pH is circumneutral (Venterea et al., possibly abiotic NO consumption in the sterilized soil. The https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 814 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

high diversity of biological NO production pathways in soils (e.g., non-specific reactions catalyzed by extracellular en- zymes) (Medinets et al., 2015). Further research is warranted to compare different sterilization methods for their effects on 15 abiotic NO production and ηNO2/NO(abiotic).

− − 4.2 Reaction reversibility between NO3 and NO2 and − − N isotope distribution between NO3 , NO2 , and NO during the anoxic incubation

−1 −1 The measured fNO-anoxic (72 to 82 ngNg h ) (Fig. 2c) is well within the range reported for anoxic soil incubations (e.g., 5 to 500 ngNg−1 h−1) (Medinets et al., 2015) and is − + − about two-thirds of the net consumption rate of NO3 NO2 15 15 + during the anoxic incubation. That the majority of consumed Figure 4. δ N-NO as a function of δ N-NH4 in the oxic and − + − hypoxic incubations. NO3 NO2 was recovered as NO supports the emerging notion that NO can be the end product of denitrification once limitations on gas diffusion are lifted in soils (Russow − other half of the reacted NO that could not be accounted et al., 2009; Loick et al., 2016). Applying the derived kabiotic 2 − for by the measured NO was likely present in the forms of and sabiotic in the abiotic incubation to the measured NO2 N2O, N2, and/or nitrosated organic compounds in the soil. concentrations under anoxic condition produced a range of 15 − −1 −1 The observed δ N difference between NO2 and NO (i.e., fNO-abiotic from < 4 to 68 ngNg h (Fig. S4). While this 15 = ± ηNO2/NO(abiotic) 19.2‰ 0.5‰) likely reflects a com- modeled fNO-abiotic appears to contribute up to 80 % of the bined N isotope effect for all of the competing NO produc- measured fNO-anoxic (Fig. S4), fNO-anoxic was high and re- tion pathways during the abiotic incubation. While very lit- mained stable even without any significant accumulation of − − tle isotope data exist for abiotic NO2 reactions in the litera- NO2 in the soil (Fig. 2b and c), suggesting that kabiotic was 15 ture, the measured ηNO2/NO(abiotic) in this study is consis- likely overestimated in the abiotic incubation (see above). tent with reported N isotope effects (i.e., 15 ‰ to 25 ‰) for Assuming that net biological NO production was maintained − − abiotic NO2 reduction by Fe(II) at similar NO2 consump- at the level of fNO-anoxic measured during the first sam- −1 tion rates to this study (0.02 to 0.05 h ) (Buchwald et al., pling event and that sabiotic was constant and equal to 0.52, 15 2016). On the other hand, the measured ηNO2/NO(abiotic) is a back-of-the-envelope calculation based on the difference 15 − lower than the reported δ N offsets between NO2 and N2O in fNO-anoxic between the first and last sampling events and 15 − (i.e., ηNO2/N2O(abiotic)) for chemo-denitrification (24 ‰ to the NO2 concentration measured at the end of the anoxic 29 ‰) (Jones et al., 2015; Wei et al., 2019). This seems to incubation indicates that kabiotic was likely on the order of −1 suggest that the observed abiotic NO production was mainly 0.0027 h , or about 7 times lower than the kabiotic derived driven by chemo-denitrification and that accumulation of NO in the abiotic incubation. Although qualitative, this calcula- as an chemo-denitrification intermediate may explain why tion suggests a minor contribution of abiotic NO production 15 the observed ηNO2/N2O(abiotic) was larger than the N isotope to the measured fNO-anoxic (< 12 %; Fig. S4). − 15 − 15 effect for Fe(II)-catalyzed NO2 reduction in previous batch The large increases in δ N-NO3 and δ N-NO values experiments (Jones et al., 2015; Buchwald et al., 2016). Fu- over the anoxic incubation (Fig. 2d and f) are congruent with 15 15 ture studies adopting simultaneous δ N-NO and δ N-N2O strong N isotopic fractionations during microbial denitrifica- measurements will be required to elucidate the role of NO as tion (Mariotti et al., 1981; Granger et al., 2008). However, − the N2O precursor during chemo-denitrification. the observed net isotope effect for NO production from NO3 15 It is important to note that the autoclaving is a harsh ster- (i.e., ηNO3/NO; 49.4 ‰ to 59.5 ‰) is larger than the appar- − ± ilization method and can substantially alter soil physical and ent N isotope effect for NO3 consumption (43.3‰ 0.9 ‰) chemical properties. For example, Buessecker et al. (2019) (Fig. S3). The large magnitude and increasing pattern of 15 − recently showed that autoclaved peat soil had 10-fold-higher ηNO3/NO, together with the accumulation of NO2 in the total fluorescence compared to non-sterilized controls, indi- soil, point to complexity beyond single-step isotopic frac- cating dramatic increases in solubility and lability of organic tionations and highlight the need to carefully examine frac- molecules by autoclaving. Furthermore, autoclaving has also tionation mechanisms for all intermediate steps leading to − − − been shown to substantially increase abiotic N2O produc- net NO production (i.e., NO3 to NO2 , NO2 to NO, and − 18 − tion from NO2 -amended soils (Wei et al., 2019). Conversely, NO to N2O). Moreover, it is surprising that both δ O-NO3 17 − milder sterilization methods (e.g., gamma-irradiation) that and 1 O-NO3 values decreased over the anoxic incuba- presumably cause less alteration of soil properties may not tion (Fig. 2g and h). Interestingly, similar decreasing trends 18 − completely inactivate biological NO production due to the in δ O-NO3 values (e.g., up to 4 ‰ over 25 h) have been

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 815

− Figure 5. (a) Net NO production rate (fNO-abiotic) of the NO2 -amended sterilized soil as a function of time. (b) Plot of the natural logarithm of fNO-abiotic vs. time showing first-order decay of fNO-abiotic. reported by Lewicka-Szczebak et al. (2014) for two anoxi- with 117O = 0 (Brunner et al., 2013). However, due to the 18 + cally incubated agricultural soils amended with a high-δ O low indigenous NH4 concentration, anammox is considered − 18 − = Chilean NO3 fertilizer similar to ours (i.e., δ O-NO3 not pertinent during the anoxic incubation. Given the com- 17 − − 56 ‰), although 1 O-NO3 was not reported in this previ- plete recovery of NO3 concentrations and isotopes in the 18 − ous study. The decreasing δ O-NO3 values, observed here control experiments (Tables S1 and S2), as well as the sig- 15 − and by Lewicka-Szczebak et al. (2014), appear to contradict nificantly increased δ N-NO3 values during the anoxic in- 15 − − + the well-established paradigm that variations in δ N-NO3 cubation, we excluded NO3 production from aerobic NH4 18 − and δ O-NO3 values follow a linear trajectory with a slope oxidation as a possible explanation for the observed declines − 18 − 17 − of 0.5 to 1 during dissimilatory NO3 reduction (Granger in δ O-NO3 and 1 O-NO3 values. 17 − et al., 2008). Furthermore, as 1 O-NO3 is in theory not Therefore, having ruled out the above possibilities led us altered by microbial denitrification – a mass-dependent frac- to postulate that the decreasing δ18O-NO− and 117O-NO− − 3 3 tionation process (Michalski et al., 2004; Yu and Elliott, values may result from anaerobic NO2 oxidation mediated 2018), the decreasing 117O-NO− values observed in this by NOB in the soil. The enzyme catalyzing NO− oxidation 3 − − 2 study indicate that processes capable of diluting or erasing to NO3 in NOB – NO2 oxidoreductase (NXR) – is metabol- 17 − the 1 O signal may occur concurrently with denitrification ically versatile and has been shown to catalyze NO3 reduc- during the anoxic incubation. Importantly, if this dilution or tion under anoxic conditions by operating in reverse (Fried- removal of the 117O signal was accompanied by N isotopic man et al., 1986; Freitag et al., 1987; Bock et al., 1988; Koch fractionations, there may be cascading effects on the distri- et al., 2015). Moreover, during NXR-catalyzed NO− oxida- − − 2 bution of N isotopes between NO3 , NO2 , and NO. tion, the required O atom originates from H2O molecules 18 − 17 − − − The decreasing δ O-NO3 and 1 O-NO3 values could (Reaction R1), so that NO2 can in theory be oxidized to NO3 be potentially explained by an O isotope equilibration be- without the presence of O2 by donating electrons to redox- − tween NO and soil H2O, catalyzed either chemically or bi- active intracellular components (Wunderlich et al., 2013) or 3 − − ologically via a reversible reaction between NO3 and NO2 alternative electron acceptors in niche environments (Babbin (Granger and Wankel, 2016). However, it has been shown et al., 2017). − in controlled laboratory experiments that dissimilatory NO − + − − 3 NO + 2H + 2e ⇔ H O + NO (R1) reduction catalyzed by bacterial nitrate reductase (NAR) 3 2 2 is irreversible at the enzyme level (Treibergs and Granger, In a denitrifying environment, anaerobic oxidation of − denitrification-produced NO− back to NO− (i.e., NO− re- 2017) and that abiotic O isotope exchange between NO3 and 2 3 2 9 ◦ 18 − 17 − H2O is extremely slow (half-life > 10 years at 25 C and oxidation) can dilute δ O-NO3 and 1 O-NO3 values by pH 7) and therefore irrelevant under natural soil conditions incorporating a “new” O atom from H2O into the reacting − (Kaneko and Poulson, 2013). Although fungi use a distinct NO3 pool (Reaction R1) (Granger and Wankel, 2016). Under enzyme system for denitrification (Shoun et al., 2012), there acidic and circumneutral pH conditions, this dilution effect is no evidence for enzymatic reversibility of fungal NAR in can be further enhanced by chemically and perhaps biolog- + − ically catalyzed O isotope equilibration between NO− and the literature. Furthermore, by converting NH4 and NO2 2 − + H O (Casciotti et al., 2007; Buchwald and Casciotti, 2010), simultaneously to N2 and NO3 , anaerobic NH4 oxidation 2 (anammox) could dilute the 117O signal by producing NO− which effectively erases the isotopic imprints of denitrifica- 3 − tion on NO2 prior to its re-oxidation. The reversibility of https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 816 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

NXR and its direct control on O isotopes in NO− have been the alternative electron acceptors that coupled with anaero- 3 − convincingly demonstrated by Wunderlich et al. (2013) using bic NO2 oxidation in the studied soil, we followed the reac- a pure culture of Nitrobacter vulgaris. By incubating N. vul- tion scheme proposed by Wunderlich et al. (2013) and Ke- − garis in a NO3 solution under anoxic conditions, Wunderlich meny et al. (2016) (Reaction R1) to parameterize the NXR- et al. (2013) showed that NO− was produced in the solution catalyzed NO− re-oxidation as the backward reaction of a 2 2 − − by N. vulgaris and that N. vulgaris promoted incorporation of dynamic equilibrium between NO3 and NO2 (Fig. 6) – that 18 − − amended O-H2O labels into NO through a re-oxidation of is, the NXR-catalyzed NO re-oxidation (backward reac- − 3 2 − the accumulated NO2 (Wunderlich et al., 2013). tion) is balanced by an NXR-catalyzed NO3 reduction (for- Importantly, there is mounting evidence from the marine ward reaction), leading to no net NO− oxidation or NO− re- − 2 3 N cycle community that NO2 re-oxidation plays a critical duction in the soil. Importantly, this representation is consis- role in the N isotope partitioning between NO− and NO−. tent with the observation that both NO− consumption and − 3 2 − 3 At the process scale, NO2 re-oxidation co-occurring with NO2 accumulation followed a pseudo-zero-order kinetics − 15 dissimilatory NO3 reduction can lead to a large δ N dif- over the anoxic incubation (Fig. 2a and b), which implies no ference between NO− and NO− beyond what would be ex- net contribution from the NO−/NO− interconversion. Given 3 − 2 3 2 pected to result from NO3 reduction alone (Gaye et al., 2013; previous findings that the NXR-catalyzed O exchange be- Dale et al., 2014; Dähnke and Thamdrup, 2016; Peters et al., tween NO− and NO− depends on NO− availability (Wun- 3 2 −2 2016; Martin and Casciotti, 2017; Buchwald et al., 2018). derlich et al., 2013), the backward NO2 re-oxidation was as- 15 − This large δ N difference is thought to arise from a rare, sumed to be first order (with respect to NO2 ), defined by a but intrinsic, inverse kinetic isotope effect associated with first-order rate constant, kNXR(b). With respect to the O iso- − − NO re-oxidation (e.g., −13 ‰) (Casciotti, 2009). As such, tope equilibration between H2O and the reacting NO pool, 2 − 2 in a net denitrifying environment, NO2 re-oxidation func- we considered two extreme-case scenarios: (1) no exchange tions as an apparent branching pathway along the sequential and (2) complete exchange. In the no-exchange scenario, reduction of NO−, preferentially re-oxidizing 15NO− back the imprints of denitrification on δ18O-NO− and 117O-NO− − 3 2 2 2 to NO3 . At the enzyme scale, the bidirectional NXR enzyme values are preserved, such that only one H2O-derived O atom has been proposed to catalyze intracellular coupled NO− re- is incorporated into NO− with each NO− molecule being re- − 3 3 2 duction and NO2 oxidation (i.e., bidirectional interconver- oxidized (Reaction R1). In the complete-exchange scenario, − − 18 17 − sion of NO3 and NO2 ), facilitating expression of an equilib- δ O and 1 O values of NO2 always reflect those of soil − − 18 17 rium N isotope effect between NO and NO (Reaction R2) H2O(δ O-H2O ≈ −10 ‰, 1 O-H2O = 0 ‰) (Fig. 6), and 3 2 − − (Wunderlich et al., 2013; Kemeny et al., 2016). therefore all three O atoms in NO3 produced from NO2 re- oxidation originate from H2O. Furthermore, we considered 14 − + 15 − ⇔ 15 − + 14 − NO2 NO3 NO2 NO3 (R2) both abiotic NO production and denitrification as the source of NO during the anoxic incubation (Fig. 6). To account for Evidence from pure culture studies of anammox bacteria the potential overestimation in kabiotic(see above), we used a −1 carrying the NXR enzyme (Brunner et al., 2013) and the- reduced kabiotic (0.0027 h ) to model net abiotic NO pro- − 15 oretical quantum calculations (Casciotti, 2009) suggest that duction from NO2 , while sabiotic and ηNO2/NO(abiotic) were this N isotope equilibration favors partitioning of 14N into fixed at 0.52 ‰ and 19.2 ‰, respectively. With respect to − − 15 NO2 with an equilibrium isotope effect ranging from 50 ‰ δ N of denitrification-produced NO, we assumed that NIR- − − to 60 ‰ (negative sign is used to denote that this N iso- catalyzed NO2 reduction to NO and NOR-catalyzed NO re- 14 tope equilibration partitions N to the left side of Reac- duction to N2O were each associated with a kinetic N iso- − − 15 15 tion R2). This NXR-catalyzed NO3 /NO2 interconversion tope effect ( ηNIR and ηNOR). The closed-system Rayleigh 15 − was invoked to explain the extremely low δ N-NO2 values equation was then used to simulate the coupled NO pro- relative to δ15N-NO− (up to 90 ‰) in the surface Antarc- duction and reduction in denitrification at each model time 3 − tic Ocean, where aerobic NO2 oxidation is inhibited by low interval (Lewicka-Szczebak et al., 2014). Detailed model nutrient availability (Kemeny et al., 2016). Hypothetically, derivation and formulation are provided in the Supplement if expressed at either the process or the enzyme level, the (Sect. S3.1). − − N isotope effect for NO2 re-oxidation could propagate into With this model of co-occurring denitrification and NO2 denitrification-produced NO, giving rise to an increased δ15N re-oxidation, we first solved for the rates of denitrifier- − 15 − − difference between NO and NO ( ηNO /NO). catalyzed NO (RNAR), NO (RNIR), and NO (RNOR) reduc- 3− 3 3 2 − To test whether NO2 re-oxidation can explain the ob- tions and kNXR(b) (four unknowns) using the measured NO3 18 − 17 − − 17 − served declines in δ O-NO3 and 1 O-NO3 values and and NO2 concentrations, fNO-anoxic, and 1 O-NO3 values 15 − − δ N distribution between NO3 , NO2 , and NO, we modi- (four measured variables). This first modeling step was ro- 14 15 16 17 17 − fied an isotopologue-specific (i.e., N, N, O, O, and bustly constrained by the measured 1 O-NO3 , which es- 18O) numerical model previously described by Yu and El- sentially functions as a 15NO− tracer (Yu and Elliott, 2018) − 3 − liott (2018) to simulate co-occurring denitrification and NO2 and is therefore particularly sensitive to NO2 re-oxidation. 15 − 15 re-oxidation in two steps. Without a clear identification of In the second modeling step, the measured δ N-NO3 , δ N-

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 817

−1 −1 −1 −1 (0.158 µgNg h ), RNIR (0.112 µgNg h ), and RNOR (0.039 µgNg−1 h−1) can be well described by zero-order ki- netics and are not sensitive to model scenarios for O ex- − change between NO and H2O (Table 1). Moreover, the − 2 observed NO2 accumulation and fNO-anoxic dynamics can be well reproduced using the modeled denitrification rates and the downward adjustment of kabiotic (Fig. 2b and c). −1 kNXR(b) was estimated to be 0.64 and 0.25 h under the no-exchange and complete-exchange scenarios, respectively 17 − (Table 1). Under both scenarios, the simulated 1 O-NO3 values exhibit a characteristic decreasing trend and are 17 − in excellent agreement with measured 1 O-NO3 values − Figure 6. Model structure of co-occurring denitrification and NO (Fig. 2h). The larger kNXR(b) under the no-exchange sce- 2 nario is expected and can be explained by the faster back re-oxidation and associated N isotope effects. Nitrogen transforma- − tions driven by denitrifiers and nitrifiers are shown by solid black reaction (i.e., NO2 re-oxidation) required to reproduce the − − and red arrows, respectively, and abiotic O exchange between NO observed dilution of 117O-NO , because only one new O 2 −3 − and H2O by the solid blue arrow. The dashed blue arrow denotes atom is incorporated into NO with each NO molecule − 3 2 net NO yield from abiotic NO reactions. 18 − 2 being re-oxidized. Although the measured δ O-NO3 val- ues did not provide quantitative constraints for the model optimization, the isotopologue-specific model with the op- − 15 NO2 , and δ N-NO values (three measured variables) were timized denitrification rates and kNXR(b) was run forward to 18 − used to optimize the kinetic N isotope effects for NAR- test whether the decreasing δ O-NO3 values can also be − 15 15 15 − catalyzed NO reduction ( ηNAR), ηNIR, ηNOR, and the possibly explained by co-occurring denitrification and NO 3 − − 2 equilibrium N isotope effect for NXR-catalyzed NO3 /NO2 re-oxidation (details are provided in Sect. S4). The results 15 − 18 − interconversion ( ηNXR(eq)) (Reaction R2; Fig. 6) (four un- showed that NO3 reduction (acting to increase δ O-NO3 − 18 − knowns). This modeling system is under-determined (num- values) and NO2 re-oxidation (acting to decrease δ O-NO3 ber of measured variables is less than the number of un- values) have counteracting effects on the forward-modeled knowns) and thus cannot be solved uniquely. Thus, instead δ18O-NO− (Fig. S2) and that the decreasing trend in δ18O- − 3 of definitively solving for the four unknown isotope effects, NO3 values can be well reproduced under both no-exchange we explored their best combination to fit the measured δ15N and complete-exchange scenarios with a reasonable assump- − − − values of NO3 , NO2 , and NO. Specifically, to reduce the tion on the net O isotope effects for denitrification and NO2 15 number of unknowns for model optimization, ηNAR and re-oxidation (Fig. S2; see Sect. S4) (Granger and Wankel, 15 15 ηNXR(eq) were treated as known values, and ηNIR and 2016). Therefore, although kNXR(b) cannot be definitively 15 ηNOR were solved by mapping through the entire space quantified in this study due to the unknown degree of O 15 15 − of ηNAR and ηNXR(eq) (at a resolution of 1 ‰), defined exchange between NO2 and H2O, these simulation results by their respective widest range of possible values. We used provide confidence in our hypothesis that the observed de- 15 18 − 17 − a range of 5 ‰ to 55 ‰ for ηNAR, consistent with a re- creases in δ O-NO3 and 1 O-NO3 values were driven cent compilation based on soil incubations and denitrifier by the reversible action of the NXR enzyme. It is impor- pure cultures (Denk et al., 2017). Given the existing obser- tant to note that the estimated kNXR(b) is fairly large even − vational and theoretical constraints (Casciotti, 2009; Brun- under the complete-exchange scenario. Based on the NO2 ner et al., 2013), a range of −60 ‰ to 0 ‰ was assigned to concentration measured at the end of the anoxic incubation 15 −1 −1 − ηNXR(eq), which is equivalent to the argument that the im- (6.9 µgNg ), a kNXR(b) of 0.25 h would require a NO2 pact of NO−/NO− interconversion on the N isotope distribu- re-oxidation rate (1.7 µgNg−1 h−1) that is 1 order of mag- 3 2− − tion between NO3 and NO2 can vary from null to a strong nitude higher than the estimated RNAR and RNIR. However, 14 − − partitioning of N to NO2 . We further defined the lower the inferred maximum NO2 re-oxidation rate under either percentile 2.5 of the error-weighted residual sum of squares model scenario (1.7 to 4.4 µgNg−1 h−1) is still within the re- 15 − − (RSS) between simulated and measured δ N values of NO3 , ported range for aerobic NO2 oxidation in agricultural soils − −1 −1 NO2 , and NO as the threshold for selection of the best-fit (e.g., up to 6–7 µgNg h ) (Taylor et al., 2019), which is models. Detailed information regarding model optimization indicative of high NOB activity even under anoxic condi- can be found in the Supplement (Sect. S3.2). tions (Koch et al., 2015). It is also noteworthy that 117O Results from the first modeling step are summarized in analysis of NO− can in theory provide quantitative con- 2 − Table 1, and the best-fit models were plotted in Fig. 2 straint on the degree of O isotope exchange between NO2 to compare with the measured data. Because the NXR- and H2O during the anoxic incubation, as has been previ- − − 17 catalyzed NO /NO interconversion was assumed to re- ously demonstrated by 1 O analysis of N2O to determine 3 2 − − sult in no change in NO3 and NO2 concentrations, RNAR O exchange between N2O and H2O during denitrification https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 818 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

15 (Lewicka-Szczebak et al., 2016). However, in this study, ro- nitude of the best-fit ηNXR(eq) may be due to the partial 17 − − − − bust 1 O-NO2 analysis was confounded by the low NO2 rate limitation by transport of NO2 /NO3 to the active site of − 15 concentrations as well as the fact that NO2 can undergo O NXR. As such, the best-fit ηNXR(eq) should provide a con- exchange with H2O during sample processing and storage servative estimate of the intrinsic equilibrium isotope effect. (Casciotti et al., 2007). Future development in soil 117O- Thus, the results from the anoxic incubation underscore the − 17 NO2 analysis and calibration will benefit the use of 1 O to important, yet previously unrecognized, role of the reversible disentangle NO− reaction complexity in soil environments. NO−/NO− interconversion in controlling the δ15N dynam- 2 3 2 − Based on the modeled denitrification rates and kNXR(b), ics of soil NO3 and its denitrification products. Substantial 15 − the best-fit ηNXR(b) was confined to a narrow range from re-oxidation of NO2 under anoxic conditions seems para- −40 ‰ to −35 ‰ (Fig. 7a and b) and was not sensitive to doxical but is underpinned by the increasingly recognized − model scenarios for O equilibration between NO2 and H2O high degree of metabolic versatility of NOB, including si- 15 15 − (Fig. 8b). While the best-fit ηNAR and ηNXR(b) were pos- multaneous oxidation of an organic substrate and NO , as − 2 itively correlated, especially under the complete-exchange well as parallel use of NO3 and O2 as electron acceptors 15 scenario (Fig. 7a and b), the best-fit ηNAR spanned a wide (Koch et al., 2015). In the absence of O2, few electron accep- range (5 ‰ to 45 ‰) and was significantly lower under the tors exist at common environmental pH that have a higher − − no-exchange scenario (RSS-weighted mean: 19 ‰) relative redox potential than the NO3 /NO2 pair (Wunderlich et al., to the complete-exchange scenario (RSS-weighted mean: 2013; Babbin et al., 2017). It is therefore likely that NOB 15 30 ‰) (Fig. 8a). On the other hand, the best-fit ηNIR (15 ‰ would gain energy by performing the intracellular coupled 15 − − to 22 ‰) and ηNOR (−8 ‰ to 2 ‰) did not vary sub- oxidation of NO and reduction of NO to survive periods 2 3 − stantially and were similar between the two model scenar- of O2 deprivation. Although anaerobic NO2 oxidation until ios (Figs. 7c–d and 8c–d). Under both model scenarios, the now has been conclusively shown only in anoxic ocean water 15 − 15 − 15 measured δ N-NO3 , δ N-NO2 , and δ N-NO values can columns (Sun et al., 2017; Babbin et al., 2017) and aquatic be well simulated using the RSS-weighted mean 15η values sediments (Wunderlich et al., 2013), soils host a huge diver- from the best-fit models (Fig. 2d to f). Specifically, the mod- sity of coexisting NOB (Le Roux et al., 2016) and the phys- 15 − 15 − − eled difference between δ N-NO3 and δ N-NO2 values iological flexibility of NOB beyond aerobic NO2 oxidation increased from about 29 ‰ at the beginning of the incuba- may contribute to the unexpected higher abundances and ac- tion to about 38 ‰ at the end of the incubation (Fig. 2d and tivities of NOB relative to AOB and AOA in agricultural soils e), whereas a constant δ15N offset of about 20 ‰ was re- (Høberg et al., 1996; Ke et al., 2013). Using the modified 15 − 15 vealed between the modeled δ N-NO2 and δ N-NO val- isotopologue-specific model, we demonstrate the possibility 15 15 ues (Fig. 2e and f). Therefore, the modeled η values and that large ηNAR can be an artifact of an isotopic equilib- 15 − − − δ N-NO2 dynamics reveal important new information for rium between NO3 and NO2 , occurring in connection with 15 understanding the increasing ηNO3/NO over the anoxic in- the bifunctional NXR enzyme. Therefore, effective expres- 15 15 cubation. During the early phase of the incubation, the N iso- sions of ηNXR(eq) in concurrence with ηNAR may explain − − 15 tope partitioning between NO3 , NO2 , and NO was mainly why ηNAR values estimated by some anoxic soil incuba- controlled by denitrification and its associated isotope ef- tions (e.g., 25 ‰ to 65 ‰) are far larger than those reported 15 15 15 − fects (i.e., ηNAR, ηNIR, and ηNOR). With the increas- by studies of denitrifying and NO -reducing bacterial cul- − 3 ing accumulation of NO2 in the soil, the dominant control tures (e.g., 5 ‰ to 30 ‰) (Denk et al., 2017) and why the on the δ15N distribution shifted to the N isotope exchange slope of δ18O-NO− vs. δ15N-NO− values during denitrifica- − − 3 3 between NO3 and NO2 , so that the difference between the tion in many field studies was not constant and rarely close 15 − 15 − δ N-NO3 and δ N-NO2 values was primarily determined to unity as observed in pure denitrifying cultures (Granger 15 by ηNXR(eq) (−40 ‰ to −35 ‰). The revealed positive cor- and Wankely, 2016). Indeed, evidence for a reversible enzy- 15 15 − − relation between the best-fit ηNAR and ηNXR(b) (Fig. 7a matic pathway linking NO3 and NO2 under anoxic condi- 15 and b) and the significantly lower ηNAR under the no- tions has already been documented in previous soil studies exchange scenario (Fig. 8a) essentially reflect a trade-off be- (e.g., Kool et al., 2011; Lewicka-Szcebak et al., 2014), im- 15 15 15 tween ηNAR and ηNXR(b) in controlling the δ N differ- plying its wide occurrence in soils. More studies using soils ence between NO− and NO− – that is, when the intercon- from a broad range of environments are needed to pinpoint 3 − 2 − − version between NO3 and NO2 is fast and the magnitude of the exact mechanisms by which NO2 can be anaerobically 15 15 17 − ηNXR(eq) is large (i.e., very negative), only a small ηNAR oxidized in soils. To that end, 1 O-NO3 can be used as a is required to sustain the large δ15N difference between NO− powerful benchmark for disentangling co-occurring NO− re- − 3 − 3 and NO2 over the course of the anoxic incubation. duction and NO2 re-oxidation. 15 15 The estimated ηNXR(eq) from the best-fit models is higher The best-fit ηNIR (15 ‰ to 22 ‰) falls within the range (i.e., closer to zero) than that derived from theoretical cal- derived in anoxic soil incubations (11 ‰ to 33 ‰) (Mariotti culations and pure culture studies (−50 ‰ to −60 ‰) (Cas- et al., 1982) and is consistent with results based on denitrify- ciotti, 2009; Brunner et al., 2013). Given the heterogeneous ing bacteria carrying copper-containing NIR (22 ‰) (Martin distribution of substrates in soils, the lower absolute mag- and Casciotti, 2016). Under both model scenarios, the best-fit

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 819

− Table 1. Means and 95 % confidence intervals of modeled denitrification rates and NO2 re-oxidation rate constants under the no-exchange and complete-exchange scenarios.

Parameter Description No exchange Complete exchange Mean 95 % CI Mean 95 % CI − −1 −1 RNAR Zero-order rate for NO3 reduction (µgNg h ) 0.158 0.157 to 0.160 0.158 0.157 to 0.160 − −1 −1 RNIR Zero-order rate for NO2 reduction (µgNg h ) 0.112 0.111 to 0.113 0.112 0.111 to 0.113 −1 −1 RNOR Zero-order rate for NO reduction (µgNg h ) 0.039 0.038 to 0.040 0.039 0.038 to 0.040 − −1 kNXR(b) First-order rate constant of NO2 re-oxidation (h ) 0.64 0.61 to 0.66 0.25 0.24 to 0.26

Figure 7. Contour maps showing variations in error-weighted residual sum of squares (RSS) between simulated and measured δ15N values, 15 15 15 15 modeled ηNIR, and modeled ηNOR as a function of prescribed ηNAR and ηNXR under the no-exchange (a, c, e) and complete- exchange (b, d, f) model scenarios. Bold contour lines encompass the best-fit models defined by the lower percentile 2.5 of the error-weighted RSS.

https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 820 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

15 incubation. Alternatively, the model-inferred ηNOR might reflect a balance between enzymatic and diffusion isotope effects, as has been previously demonstrated for N2O reduc- tion in soil denitrification (Lewicka-Szczebak et al., 2014). Because diffusion would be expected to have a small and − normal kinetic isotope effect, if NO2 reduction was limited by NO diffusion out of soil denitrifying sites, the estimated 15 ηNOR would be shifted toward the isotope effect for NO diffusion. Diffusion might be particularly important in this study due to the flow-through condition during the anoxic incubation and the low solubility of NO, both of which fa- vor gas diffusion while preventing re-entry of escaped NO 15 to denitrifying cells. Thus, the small ηNOR inferred from the best-fit models is likely a combination of diverse NO re- duction pathways in this agricultural soil, as well as limited expression of enzymatic isotope effects imposed by NO dif- fusion. Regardless, the empirical finding of this study sug- 15 15 gests that due to the small ηNOR, the bulk δ N values of denitrification-produced N2O should not be significantly al- tered by accumulation and diffusion of NO during denitrifi- cation. 15 Figure 8. Frequency distributions of the best-fit ηNAR (a), 4.3 NO source contribution and N isotope effects for 15η (b), 15η (c), and 15η (d) under the no-exchange NXR(eq) NIR NOR NO production from NH+ oxidation under oxic (red) and complete-exchange (blue) model scenarios. Dashed verti- 4 cal lines denote the RSS-weighted mean 15η values from the best-fit and hypoxic conditions models under the two model scenarios. The coupled decrease in NH+ concentrations and increase − 4 in NO3 concentrations (Fig. 3a and b) indicate active nitri- 15 fication in both oxic and hypoxic incubations. Moreover, the ηNOR (−8 ‰ to 2 ‰) is relatively small and more normal 15 two oxidation steps of nitrification were tightly coupled, re- (i.e., η value closer to zero) than the bulk N isotope effect − − for NO reduction to N O catalyzed by purified fungal NOR sulting in no accumulation of NO2 in the soil. Because NO3 2 17 (P450nor) (−14 ‰) (Yang et al., 2014). During P450nor- produced from nitrification has a zero 1 O value, the ac- catalyzed NO reduction, two NO molecules are sequentially tive nitrification was also reflected in the progressive dilution 17 − bonded to the Fe active site of P450nor, and the observed in- of 1 O-NO3 under both oxic and hypoxic conditions (Yu and Elliott, 2018). Based on the measured concentrations and verse isotope effect was proposed to arise from a reversible + − bonding of the first NO molecule (Yang et al., 2014). To date, isotopic composition of NH4 and NO3 , the isotopologue- the N isotope effect for NO reduction catalyzed by bacterial specific model previously developed by Yu and Elliott (2018) NORs has not yet been quantified. Unlike P450nor, which was used to estimate the rates and net N isotope effects of net R 15η + contains only a single heme Fe at the active site, the ac- mineralization ( OrgN/NH4 and OrgN/NH4 ), gross NH4 ox- − R 15η − tive site of bacterial NORs has two Fe atoms (i.e., binuclear idation to NO3 ( NH4/NO3 and NH4/NO3 ), and gross NO3 R 15η center). Therefore, three classes of mechanisms have been consumption ( NO3comp and NO3comp) during the oxic and proposed for the two-electron reduction of NO by bacterial hypoxic incubations. As has been discussed above, this nu- 17 NORs, including sequential bonding of two NO molecules merical model relies on the conservative nature of 1 O- NO− and its powerful application in tracing co-occurring to either Fe catalytic center and simultaneous bonding of two 3 − − NO molecules to both Fe centers (Kuypers et al., 2018; Lehn- nitrification and NO3 consumption (consisting of NO3 im- ert et al., 2018). Although the precise catalytic mechanism mobilization and denitrification in this case) (Yu and Elliott, remains uncertain, site-specific measurements of N isotopes 2018). Detailed model derivation, formulation, and optimiza- tion have been documented in Yu and Elliott (2018) and are in N2O (i.e., N2O isotopomers) produced from denitrifying bacteria indicate a similar magnitude for isotopic fractiona- also briefly summarized in Sect. S5. The modeling results 15 + tions during the reduction of two NO molecules, in support based on the low-δ N-NH4 treatment in the oxic incubation of the simultaneous binding theory (Sutka et al., 2006; Ya- were reported by Yu and Elliott (2018). Here, we used data 15 + mazaki et al., 2014). Thus, if the bulk N isotope effect for from all three δ N-NH4 treatments to more robustly con- bacterial NO reduction is higher than that for fungal NO re- strain the N transformation rates and net N isotope effects 15 for each incubation experiment (i.e., oxic and hypoxic). duction, the best-fit ηNOR may reflect a mixed contribution of bacteria and fungi to NO consumption during the anoxic

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 821

Table 2. Means and 95 % confidence intervals of modeled gross N transformation rates, NO yields, and net N isotope effects in the oxic and hypoxic incubations.

Parameter Description Oxic Hypoxic Mean 95 % CI Mean 95 % CI −1 −1 − ROrgN/NH4 Zero-order rate for net mineralization (µgNg h ) 0.014 0.013 to 0.016 0.012 0.011 to 0.038 −1 −1 RNH4/NO3 Zero-order rate for gross nitrification (µgNg h ) 0.458 0.455 to 0.460 0.111 0.110 to 0.113 − −1 −1 RNO3comp Zero-order rate for gross NO3 consumption (µgNg h ) 0.071 0.070 to 0.072 0.070 0.049 to 0.091 15 − − ηOrgN/NH4 Net N isotope effect for net mineralization 2 ‰ 27 ‰ to 31 ‰ 0 ‰ 18 ‰ to 17 ‰ 15 ηNH4/NO3 Net N isotope effect for gross nitrification 28 ‰ 27 ‰ to 30 ‰ 23 ‰ 12 ‰ to 33 ‰ 15 − − − ηNO3comp Net N isotope effect for gross NO3 consumption 5 ‰ 16 ‰ to 20 ‰ 7 ‰ 9 ‰ to 23 ‰ fNH4 Fraction of net NO production from nitrification 0.72 0.65 to 0.78 0.58 0.55 to 0.61 YNH /NO NO yield in nitrification 1.3 % 1.2 % to 1.4 % 5.2 % 4.8 % to 5.5 % 4 − YNO3/NO NO yield in NO3 consumption 3.2 % 2.5 % to 4.0 % 6.1 % 4.3 % to 9.3 % 15 + ηcomb Combined net isotope effect for NO production from NH4 56 ‰ 54 ‰ to 58 ‰ 51 ‰ 50 ‰ to 52 ‰ − and NO3 Mean 95 % CI 15 + ηNH4/NO Net isotope effect for NO production from NH4 oxidation 66 ‰ 59 ‰ to 85 ‰ 15 − ηNO3/NO Net isotope effect for NO production from NO3 consumption 30 ‰ 1 ‰ to 42 ‰

The modeling results are summarized in Table 2. Excellent nificantly lower than unity under both oxic and hypoxic con- agreement was obtained between the observed and simulated ditions (Fig. 4). This suggests that sources other than NH+ + − 4 concentrations and isotopic composition of NH4 and NO3 oxidation contributed to the observed net NO production. Al- for both oxic and hypoxic incubations (Fig. 3). RNH4/NO3 can though NO can be produced by numerous microbial and abi- be well described by zero-order kinetics and was estimated otic processes (Medinets et al., 2015), we argue that the other to be 0.46 and 0.11 µgNg−1 h−1 for the oxic and hypoxic major NO source is mostly likely related to NO− consump- 3 − incubations, respectively (Table 2). The lower RNH4/NO3 in tion. This is based on the observation of high NO3 concen- the hypoxic incubation indicates that nitrification was lim- trations in both oxic and hypoxic incubations, as well as the ited by low O2 availability. Under both oxic and hypoxic estimated low ROrgN/NH (Table 2), which indicates a low + − 4 conditions, oxidation of NH4 to NO3 was associated with availability of labile organic N – another potential substrate 15 a large ηNH /NO (23 ‰ to 28 ‰; Table 2), consistent with for NO production (Stange et al., 2013) – in this agricultural 4 3 + the N isotope effects for NH3 oxidation in pure cultures of soil. Therefore, based on the assumption that NH oxidation − 4 AOB and AOA (e.g., 13 ‰ to 41 ‰) (Mariotti et al., 1981; and NO3 consumption were the two primary NO sources Casciotti et al., 2003; Santoro and Casciotti, 2011). On the during the oxic and hypoxic incubations, a two-source iso- 15 other hand, the estimated ROrgN/NH4 and RNO3comp were low tope mixing model was used to relate the measured δ N-NO and not significantly different between the two incubation values to the concurrently measured δ15N-NH+ and δ15N- − 4 experiments (Table 2). Nevertheless, while RNO3comp was NO3 values: only 16 % of RNH4/NO3 in the oxic incubation, RNO3comp ac- δ15 = f × (δ15 +−15η ) counted for a much larger fraction (63 %) of RNH4/NO3 in N-NO NH4 N-NH4 NH4/NO 15 − 15 the hypoxic incubation, mainly due to the reduced RNH4/NO3 + − × − (1 fNH4 ) (δ N-NO3 ηNO3/NO), (5) under the low-O2 condition. Due to the low magnitude of 15 15 15 ROrgN/NH4 and RNO3comp, the estimated ηOrgN/NH4 and where ηNH4/NO and ηNO3/NO are the net isotope effects 15 + − ηNO3comp are associated with large errors and not signifi- for NO production from NH4 oxidation and NO3 consump- cantly different from zero (Table 2). tion, respectively. Rearranging Eq. (5) yields Eq. (6): By using three isotopically different NH+ fertilizers in 4 15 = × 15 + + − × 15 − parallel treatments, we are able to quantify the fractional δ N-NO fNH4 δ N-NH4 (1 fNH4 ) δ N-NO3 + contribution of NH oxidation to the measured net NO pro- − [ ×15 + − 4 fNH4 ηNH4/NO (1 fNH4 ) duction (fNH ). Specifically, if NO was exclusively produced 4+ ×15 ] 15 ηNO3/NO , (6) from soil NH4 , we would expect to see a constant δ N dif- + 15 + 15 = ×15 ference between NH4 and NO across the three δ N-NH4 ηcomb fNH4 ηNH4/NO 15 treatments. In fact, the observed δ N differences were not + − ×15 15 + 15 (1 fNH4 ) ηNO3/NO, (7) constant, and the slope of δ N-NH4 vs. δ N-NO was sig- https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 822 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

tion, whereas NO produced from denitrification is further re- duced to N2O before it escapes to the soil surface (Kester et al., 1997; Skiba et al., 1997). The minor role of denitrifi- cation is largely deduced from the supposition that denitri- fication is activated only under wet soil conditions (David- son and Verchot, 2000). However, based on our δ15N-based NO source partitioning, about 30 % of the net NO produc- − tion was contributed by NO3 consumption under oxic con- dition, highlighting the potential importance of denitrifica- tion in driving soil NO emissions under conditions not typ- ically conducive to its occurrence. There is growing evi- dence that extensive anoxic microsites can develop in oth- erwise well-aerated soils due to micro-scale variability of O2 demand and soil texture-dependent gas diffusion limita- tions (Keiluweit et al., 2018). Although we would not pre- dict high rates of heterotrophic respiration in this agricul- tural soil with low organic carbon, it is possible that rapid O2 Figure 9. Hole-in-the-pipe illustration of NO production from gross consumption by nitrification may outpace O2 supply through − nitrification and NO3 consumption under oxic and hypoxic condi- diffusion in soil microsites, fostering development of anoxic tions. “OrgN” denotes organic nitrogen. niches in close association with nitrification hot spots (Kre- men et al., 2005). Based on 15N labeling and direct 15NO measurements using a gas chromatograph–quadrupole mass spectrometer, Russow et al. (2009) demonstrated that nitri- fication contributed about 70 % of net NO production in a 15 15 + 15 − δ N-NO = fNH × δ N-NH + (1 − fNH ) × δ N-NO + 4 4 4 3 well-aerated, NH4 -fertilized silt loam, in strong agreement 15 15 − ηcomb. (8) with our results based on natural-abundance δ N measure- ments. An even lower contribution to NO production, e.g., Equation (6) essentially dictates that the δ15N-NO values can 26 % to 44 %, has been reported for nitrification in organic, 15 + 15 − be modeled from the δ N-NH4 and δ N-NO3 values us- N-rich forest soils incubated under oxic conditions (Stange ing a hypothetical isotope effect for NO production from the et al., 2013). The persistence of denitrifying microsites in the + − 15 combined soil NH4 and NO3 pool ( ηcomb; the last term studied soil is further corroborated by the nearly doubled net 15 15 − in Eq. 6) that is a mixing of ηNH4/NO and ηNO3/NO con- NO production from NO3 consumption in the hypoxic incu- 15 trolled by fNH4 (Eq. 7). Thus, assuming fNH4 and ηcomb bation (Fig. 9). Importantly, the actual NO yield in denitrifi- were constant in each incubation experiment, fNH4 and cation might be much higher than those estimated for gross 15 15 15 − ηcomb can be solved using the measured δ N-NO, δ N- NO3 consumption during the oxic and hypoxic incubations + 15 − 15 + NH4 , and δ N-NO3 values from all three δ N-NH4 treat- (i.e., 3.2 % and 6.1 %), as denitrification occurring in anoxic ments (Eq. 8). fNH4 was estimated to be 0.72 under the oxic niches might only comprise a small fraction of the estimated incubation (Table 2), indicating that 72 % of the measured net RNO comp. + 3 NO production was sourced from NH4 oxidation, with the Interestingly, while RNH4/NO3 was significantly lower in remainder being ascribed to NO− consumption. Under the the hypoxic incubation, the net NO production from NH+ 3 + 4 hypoxic condition, the share of NH4 oxidation decreased to oxidation was similar between the two incubation experi- 15 58 % (Table 2). ηcomb was estimated to be 56 ‰ under the ments, indicating a higher NO yield in nitrification when O2 oxic condition and 51 ‰ under the hypoxic condition (Ta- availability became limited (Fig. 9). However, mechanisms ble 2). Combining the δ15N-based NO source partitioning underlying the differential NO yield in nitrification are diffi- with the estimated RNH /NO and RNO comp, we further es- cult to elucidate owing to the high complexity of biochemical 4+ 3 3 − timated NO yield in NH4 oxidation and NO3 consumption, pathways of NO production by AOB and AOA. In AOB, the respectively, and where the results are illustrated according to prevailing view of NH3 oxidation is that it occurs via a two- the classic “hole-in-the-pipe” (HIP) concept (Fig. 9) (David- step enzymatic process, involving hydroxylamine (NH2OH) son and Verchot, 2000). NO yield was 1.3 % in NH+ oxida- as an obligatory intermediate (Fig. 10). The first step is cat- − 4 tion and 3.2 % in NO3 consumption in the oxic incubation alyzed by NH3 monooxygenase (AMO), which uses copper (Fig. 9; Table 2). Under the hypoxic condition, NO yield was and O2 to hydroxylate NH3 to NH2OH. Next, a multiheme + − increased to 5.2 % in NH oxidation and 6.1 % in NO con- enzyme, NH2OH oxidoreductase (HAO), catalyzes the four- 4 3 − sumption (Fig. 9; Table 2). electron oxidation of NH2OH to NO2 via enzyme-bound Most previous laboratory and field studies suggest that nitroxyl ([HNO-Fe]) and nitrosyl ([NO-Fe]) intermediates soil NO emissions are predominately driven by nitrifica- (Lehnert et al., 2018) (Fig. 10). Under this NH2OH obli-

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 823

Figure 10. The three enzymatic pathways for NO production during − NH3 oxidation to NO2 by AOB: the NH2OH obligatory interme- diate pathway is indicated by blue circle 1, the NH2OH/NO oblig- atory intermediate pathway is indicated by blue circle 2, and the nitrifier-denitrification pathway is indicated by blue circle 3. Square brackets enclose proposed enzyme-bound intermediates [HNO-Fe] and [NO-Fe] of the NH OH obligatory intermediate pathway. The 2 Figure 11. Relative magnitude of net N isotope effects for NO pro- role of AOB-encoded nitrite reductase (NIR) in catalyzing NO oxi- + 15 − − duction from NH4 oxidation ( ηNH4/NO) and NO3 consumption dation to NO in the NH2OH/NO obligatory intermediate pathway 2 (15η ) in the oxic and hypoxic incubations. is hypothetical. NO3/NO

ately consumed with tighter control in AOA than in AOB gate intermediate model, NO emission was proposed to re- (Kozlowski et al., 2016). sult from dissociation of NO from the enzyme-bound nitrosyl To shed further light on the inner workings of net + complex under high-NH3 and/or low-O2 conditions (Fig. 10) NO production from NH4 , we turn to constraining 15 (Hooper et al., 2004; Beeckman et al., 2018). However, there ηNH4/NO. Specifically, the inherent linkage between 15 15 15 is recent strong evidence that HAO generally catalyzes the ηcomb, ηNH4/NO, and ηNO3/NO (Eq. 7) allows one to 15 15 three-electron oxidation of NH2OH to NO under both aero- probe the relative magnitude of ηNH4/NO and ηNO3/NO 15 − bic and anaerobic conditions; the HAO-produced NO is fur- using the determined ηcomb and fNH . Given that NO was − 4 2 ther oxidized to NO2 by an unknown enzyme (Caranto and absent in the soil and that NO reduction in denitrification was 15 Lancaster, 2017). In this way, NO would not be a byproduct likely associated with a small isotope effect (i.e., ηNOR; 15 of incomplete NH2OH oxidation but rather required as an see above), ηNO3/NO in the oxic and hypoxic incubations − 15 obligatory intermediate for NO2 production (Fig. 10). It was should mainly reflect ηNAR. Thus, by assigning the entire 15 further proposed that AOB-encoded copper-containing NIR possible range of the best-fit ηNAR derived in the anoxic − 15 15 may catalyze the final one-electron oxidation of NO to NO2 incubation (5 ‰ to 45 ‰; Fig. 7a) to ηNO3/NO, ηNH4/NO by operating in reverse (Lancaster et al., 2018). Under this was estimated to range from 60 ‰ to 76 ‰ in the oxic in- NH2OH/NO obligate intermediate model, high intracellular cubation and from 55 ‰ to 84 ‰ in the hypoxic incubation NO concentrations arise when the rate of NO production out- (Fig. 11). If we take one step further by assuming that both − 15 15 paces the rate of its oxidation to NO2 , leading to NO leak- ηNO3/NO and ηNH4/NO were identical between the oxic 15 15 age from cells. Consequently, under O2 stress, decreases in and hypoxic incubations, then ηNO /NO and ηNH /NO − 3 4 the rate of NO oxidation to NO2 might be expected, and this could be uniquely determined to be 30 ‰ and 66 ‰, respec- may explain the observed increase in nitrification NO yield tively (Fig. 11; Table 2). Thus, the relative magnitude of 15 15 in the hypoxic incubation. Additionally, some AOB strains ηNO3/NO and ηNH4/NO provides insights into the differ- 15 + 15 can produce NO in a process termed nitrifier denitrification, ential relationship between δ N-NH4 and δ N-NO across − 15 + in which NO is produced through NIR-catalyzed NO2 re- the three δ N-NH4 treatments in the oxic and hypoxic in- duction and can be further reduced to N2O by AOB-encoded cubations (Fig. 4). In the oxic incubation, if we assume that 15 = 15 = 15 NOR (Wrage-Mönning et al., 2018) (Fig. 10). Compared to ηNH4/NO 66 ‰ and ηNO3/NO 30 ‰, the δ N of NO + 15 + AOB, the NH3 oxidation pathway in AOA remains unclear produced from NH4 oxidation under the low δ N-NH4 (Beeckman et al., 2018). The current model is that NH3 is treatment (about −60 ‰) would be much lower than the 15 − − first oxidized by an archaeal AMO to NH2OH and subse- δ N of NO from NO3 consumption (about 38 ‰). How- − 15 + 15 + quently converted to NO2 by an unknown HAO counterpart ever, under the high-δ N-NH4 treatment, the δ N of NH4 - (Kozlowski et al., 2016). NO seems to be mandatory for ar- produced NO would increase to about −14 ‰ and be higher 15 − chaeal NH2OH oxidation and has been proposed to act as than δ N values of NO -produced NO (about −26 ‰). − 3 − a co-substrate for the NO2 production (Kozlowski et al., Consequently, the production of NO from NO3 consumption 2016). Consequently, NO is usually produced and immedi- would dilute the δ15N of total net NO production, pulling it to https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 824 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

: 15 + 15 fall below the 1 1 line between the δ N-NH4 and δ N-NO count for the residual isotope effect for net NO production values in Fig. 4. This dilution effect was more pronounced in from NH+. This inferred isotope effect is of similar magni- 4 − the hypoxic incubation due to the lower fNH4 (i.e., higher tude to that reported for NXR-catalyzed NO2 oxidation (i.e., − − contribution of NO3 -produced NO) (Fig. 4). 13 ‰) (Casciotti, 2009). However, to unambiguously de- 15 Therefore, under either oxic or hypoxic condition, the net termine the mechanisms giving rise to the large ηNH /NO, + 4 NO production from NH4 oxidation proceeded with a large further biochemical analyses will be needed to clarify the en- 15 ηNH4/NO. As NH3 oxidation to NH2OH was likely the rate- zymatic pathways responsible for NO production by AOB limiting step for the entire nitrification process, a fraction of and AOA under relevant soil conditions. Nonetheless, the re- 15 the inferred large ηNH4/NO can be accounted for by the iso- sults presented here provide evidence that production of NO 15 tope effect for NH3 oxidation to NH2OH, which should be with low δ N values may be a characteristic feature of nitri- 15 + similar to the estimated ηNH4/NO3 (e.g., 23 ‰ to 28 ‰). fication in NH4 -fertilized agricultural soils under both oxic The residual isotope effect, on the order of 40 ‰, must there- and hypoxic conditions. fore stem from additional bond forming/breaking during net NO production in NH3 oxidation. This additional N isotope − effect could be explained by NO2 reduction catalyzed by 5 Implications for NO emission from agricultural soils AOB-encoded NIR if NO was dominantly produced through the nitrifier-denitrification pathway (Fig. 10). However, pro- In this study, the net production rates and δ15N values of NO vided that the two oxidation steps of nitrification were tightly were measured under a range of controlled laboratory con- coupled under both oxic and hypoxic conditions, it is un- ditions. The results provide insights into how stable N and − likely that NO2 would accumulate to high enough intracellu- O isotopes can be effectively used to understand the reac- lar concentrations to trigger nitrifier denitrification (Wrage- tion mechanisms by which NO is produced and consumed Mönning et al., 2018). Similarly, we would not expect any in soils. While nitrification is the commonly cited source for substantial isotope fractionations to result from accumula- NO emissions from agricultural soils, the measured net NO tion of intracellular NH2OH or enzyme-bound intermedi- production rates in this study highlight the great potential of − ate species (e.g., [HNO-Fe] and [NO-Fe]). Thus, we are left abiotic NO2 reduction and denitrification in driving NO pro- with either a large and normal isotope effect for NO disso- duction and release from agricultural soils and thus should ciation from its enzyme-bound precursor if NO production not be overlooked when attributing field soil NO emissions. was mainly routed through the NH2OH obligate intermediate Indeed, because NO is a direct product or free intermediate − pathway or an inverse isotope effect associated with NO ox- in these processes, abiotic NO2 reduction and denitrifica- idation if NO itself was an obligatory intermediate required tion may inherently have a larger NO yield – that is, a big- − for NO2 production (Fig. 10). With respect to the first pos- ger “hole” for NO leaking in the HIP model (Davidson and sibility, if NO dissociation from the Fe active site of HAO Verchot, 2000). We conclude that the isotope-based measure- is mainly controlled by an equilibrium reaction between NO ment and modeling framework established in this work is a and enzyme-bound nitrosyl species, the forward and back- powerful tool to bridge NO production with gross N trans- ward reactions may occur with distinctively different isotope formation processes in agricultural soils, thereby providing a effects, giving rise to an equilibrium isotope effect that fa- quantitative way to parameterize the HIP model for modeling vors partitioning of 14N to the dissociated NO. However, ex- soil NO emissions under dynamic environmental conditions pression of this equilibrium isotope effect would be largely (e.g., varying temperature and soil moisture content). suppressed by limited isotope exchange between the two N The differences in the net isotope effects for NO produc- − pools due to the presumably transient presence of nitrosyl in- tion from abiotic NO2 reduction, denitrification, and nitri- termediate. Therefore, a partial expression of a large equilib- fication revealed in this study (Fig. 12a) suggest that δ15N- rium isotope effect (e.g., > 40 ‰) would be required to ex- NO is a useful tracer for informing NO production pathways plain the residual N isotopic fractionation during NO produc- in agricultural soils. Specifically, the relatively small magni- 15 15 tion in NH3 oxidation. Alternatively, in regards to the second tude of ηNO2/NO(abiotic) indicates that δ N-NO is particu- possibility, if we assume that the enzyme-catalyzed oxida- larly useful in probing the relative importance of NO produc- − tion of NO to NO2 proceeds via an enzyme-bound transition tion from abiotic vs. microbial reactions, lending support to state and that the transition state contains the newly formed our previous finding based on rewetting of a dry forest soil N-O bond, an inverse isotope effect may result from more that high δ15N values of rewetting-triggered NO pulses were − strongly bonded N atom in the transition state, for which mainly contributed by chemical NO2 reduction (Yu and El- − 15 there is precedent in the literature (i.e., NO2 oxidation to liott, 2017). Moreover, the large ηNH4/NO revealed in the NO−; see above) (Casciotti, 2009). Moreover, the small NO oxic and hypoxic incubations provides an empirical basis for 3 + − yield observed in the oxic and hypoxic incubations would in- discerning the relative role of NH4 oxidation and NO3 re- dicate a large consumption of NO (i.e., 95 % to 99 %). With duction in driving soil NO production and emissions. Inter- this high level of NO consumption, an inverse isotope effect estingly, comparing the measured net isotope effects for NO − − − on the order of 13 ‰ to 9 ‰ would be sufficient to ac- production from abiotic NO2 reduction, denitrification, and

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 825

Figure 12. (a) Comparison of net isotope effects for NO production estimated in this study to net isotope effects for N2O production 15 reported in the literature. (b) Comparison of in situ δ N of NOx emission from a manure-fertilized soil (reported by Miller et al., 2018) to nitrification and denitrification δ15N-NO end-members derived using the estimated net isotope effects for NO production in the oxic and hypoxic incubations.

− nitrification with those previously quantified for N2O pro- tures. Because NO2 accumulation was not reported by Miller duction in soil incubations and pure cultures (Denk et al., et al. (2018), we consider nitrification and denitrification to 2017, and references therein; Jones et al., 2015; Wei et al., be the primary sources for the observed NO (and, to a much 15 2019), a similar pattern is evident across these three common lesser extent, NO2) emissions. Therefore, the ηNH4/NO and 15 production pathways for NO and N2O (Fig. 12a). This simi- ηNO3/NO values derived in the oxic and hypoxic incuba- larity reflects the intimate connection between NO and N2O tions (i.e., 66 ‰ and 30 ‰, respectively) were used in com- 15 + − turnover within each reaction pathway and provides strong bination with the δ N values of soil NH4 and NO3 reported 15 15 15 evidence that simultaneous δ N-NO and δ N-N2O mea- in Miller et al. (2018) to calculate the δ N end-members for + − surements can potentially yield unprecedented insights into NO produced from NH4 oxidation and NO3 reduction. As 15 the sources and processes controlling NO and N2O emis- shown in Fig. 12b, comparing the in situ δ N-NOx mea- sions from agricultural soils. However, on the other hand, the surements with the estimated isotopic end-members provides − − demonstrated reaction reversibility between NO2 and NO3 a compelling picture of soil NO dynamics following ma- 15 under anoxic conditions is a new complication that needs to nure application. Notably, the initial low δ N-NOx values 15 be considered when using δ N to examine soil NO and N2O reported by Miller et al. (2018) might indicate a mixed con- − + − emissions. As NO2 is often accumulated in agricultural soils tribution of NH4 oxidation and NO3 reduction to soil NOx 15 following fertilizer application (Venterea et al., 2020), ex- emissions (Fig. 12b). Nevertheless, the increase in δ N-NOx − pression of the equilibrium isotope effect between NO2 and values measured 4 to 11 d after manure application may re- − 15 NO3 in redox-dynamic surface soils may render δ N-NO flect a shift in the dominant NO production pathway to den- 15 − and δ N-N2O less useful in tracing NO and N2O sources. itrification, in line with the increasing accumulation of NO3 Given that high soil NO− concentrations can trigger emis- supplied by nitrification in the soil (Miller et al., 2018). Al- 2 − sion pulses of NO and N2O (Venterea et al., 2020), NO2 though data-limited, this example provides promising initial accumulation should be taken as a critical sign for careful evidence for the ability of multi-species δ15N measurements evaluation of the reaction complexity underlying δ15N distri- to provide mechanistic information on soil NO dynamics and butions among the denitrification products. its environmental controls. Further experimental constraints To further assess the potential utility of δ15N measure- on soil δ15N-NO variations can build on the measurement ments in source partitioning NO emissions from agricultural and modeling framework developed in this study to advance soils, we applied the estimated N isotope effects to the in situ our understanding of soil NO source contributions over a 15 δ N-NOx measurements reported by Miller et al. (2018). wide range of environmental conditions and soil types. Importantly, the soil used in this study was collected from the same farm where Miller et al. (2018) conducted their field measurements (e.g., the USDA-managed corn–soybean field Code and data availability. The datasets generated for this in central Pennsylvania, USA). Hence, the derived isotope study and documentation about the equations and param- effects may be particularly relevant to their reported δ15N- eters of the isotopologue-specific models are available in the Supplement. The MATLAB codes for the isotopologue- NOx values due to similar soil microbial community struc- specific models are available at https://github.com/zjyuuiuc/ https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 826 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

Isotopologue-specific-models (last access: 1 February 2021) and at by Fe (II) and its production of N2O, Geochim. Cosmochim. Ac., https://doi.org/10.5281/zenodo.4495715 (Yu, 2020). 186, 1–12, 2016. Buchwald, C., Homola, K., Spivack, A. J., Estes, E. R., Mur- ray, R. W., and Wankel, S. D.: Isotopic constraints on nitrogen Supplement. The supplement related to this article is available on- transformation rates in the deep sedimentary marine biosphere, line at: https://doi.org/10.5194/bg-18-805-2021-supplement. Global Biogeochem. Cy., 32, 1688–1702, 2018. Buessecker, S., Tylor, K., Nye, J., Holbert, K. E., Urquiza Muñoz, J. D., Glass, J. B., Hartnett, H. E., and Cadillo- Author contributions. ZY and EME designed the study; ZY con- Quiroz, H.: Effects of sterilization techniques on chemodenitri- ducted the experiments and analyzed the data; ZY and EME wrote fication and N2O production in tropical peat soil microcosms, the paper. Biogeosciences, 16, 4601–4612, https://doi.org/10.5194/bg-16- 4601-2019, 2019. Calvert, J. G., Lazrus, A., Kok, G. L., Heikes, B. G., Walega, J. G., Lind, J., and Cantrell, C. A.: Chemical mechanisms of acid gen- Competing interests. The authors declare that they have no conflict eration in the troposphere, Nature, 317, 27–35, 1985. of interest. Caranto, J. D. and Lancaster, K. M.: Nitric oxide is an obligate bac- terial nitrification intermediate produced by hydroxylamine oxi- doreductase, P. Natl. Acad. Sci. USA, 114, 8217–8222, 2017. Acknowledgements. The authors thank Curtis Dell (USDA-ARS) Casciotti, K. L.: Inverse kinetic isotope fractionation during bacte- for helping with the field soil sampling and Katherine Redling, Vi- rial nitrite oxidation, Geochim. Cosmochim. Ac., 73, 2061–2076, vian Feng, Madeline Ellgass, and Madeline Gray (University of 2009. Pittsburgh) for assistance with the isotopic analyses. Casciotti, K. L., Sigman, D. M., Hastings, M. G., Böhlke, J. K., and Hilkert, A.: Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method, Financial support. This work was supported by a National Science Anal. Chem., 74, 4905–4912, 2002. Foundation CAREER award (grant no. 1253000) to Emily M. El- Casciotti, K. L., Sigman, D. M., and Ward, B. B.: Linking diversity liott. and stable isotope fractionation in ammonia-oxidizing bacteria, Geomicrobiol. J., 20, 335–353, 2003. Casciotti, K. L., Böhlke, J. K., McIlvin, M. R., Mroczkowski, S. J., Review statement. This paper was edited by Andreas Richter and and Hannon, J. E.: Oxygen isotopes in nitrite: analysis, calibra- reviewed by Wolfgang Wanek and one anonymous referee. tion, and equilibration, Anal. Chem., 79, 2427–2436, 2007. Crutzen, P. J.: The role of NO and NO2 in the chemistry of the troposphere and stratosphere, Annu. Rev. Earth Pl. Sc., 7, 443– 472, 1979. References Davidson, E. A.: The contribution of manure and fertilizer nitrogen to atmospheric nitrous oxide since 1860, Nat. Geosci., 2, 659– Almaraz, M., Bai, E., Wang, C., Trousdell, J., Conley, S., 662, 2009. Faloona, I., and Houlton, B. Z.: Agriculture is a major Dähnke, K. and Thamdrup, B.: Isotope fractionation and isotope source of NOx pollution in California, Sci. Adv., 4, 3477, decoupling during anammox and denitrification in marine sedi- https://doi.org/10.1126/sciadv.aao3477, 2018. ments, Limnol. Oceanogr., 61, 610–624, 2016. Babbin, A. R., Peters, B. D., Mordy, C. W., Widner, B., Cas- Dale, A. W., Sommer, S., Ryabenko, E., Noffke, A., Bohlen, L., ciotti, K. L., and Ward, B. B.: Multiple metabolisms constrain Wallmann, K., Stolpovsky, K., Greinert, J., and Pfannkuche, O.: the anaerobic nitrite budget in the Eastern Tropical South Pa- Benthic nitrogen fluxes and fractionation of nitrate in the Mauri- cific, Global Biogeochem. Cy., 31, 258–271, 2017. tanian oxygen minimum zone (Eastern Tropical North Atlantic), Beeckman, F., Motte, H., and Beeckman, T.: Nitrification in agricul- Geochim. Cosmochim. Ac., 134, 234–256, 2014. tural soils: impact, actors and mitigation, Curr. Opin. Biotech., Davidson, E. A. and Verchot, L. V.: Testing the hole-in-the-pipe 50, 166–173, 2018. model of nitric and nitrous oxide emissions from soils using the Bock, E., Wilderer, P. A., and Freitag, A.: Growth of Nitrobacter in TRAGNET database, Global Biogeochem. Cy., 14, 1035–1043, the absence of dissolved oxygen, Water Res., 22, 245–250, 1988. 2000. Brunner, B., Contreras, S., Lehmann, M. F., Matantseva, O., Rol- Denk, T. R., Mohn, J., Decock, C., Lewicka-Szczebak, D., Har- log, M., Kalvelage, T., Klockgether, G., Lavik, G., Jetten, M. S., ris, E., Butterbach-Bahl, K., Kiese, R., and Wolf, B.: The ni- Kartal, B., and Kuypers, M. M.: Nitrogen isotope effects induced trogen cycle: A review of isotope effects and isotope modeling by anammox bacteria, P. Natl. Acad. Sci. USA, 110, 18994– approaches, Soil Biol. Biochem., 105, 121–137, 2017. 18999, 2013. Felix, D. J., Elliott, E. M., Gish, T. J., McConnell, L. L., and Buchwald, C. and Casciotti, K. L.: Oxygen isotopic fractiona- Shaw, S. L.: Characterizing the isotopic composition of atmo- tion and exchange during bacterial nitrite oxidation, Limnol. spheric ammonia emission sources using passive samplers and a Oceanogr., 55, 1064–1074, 2010. combined oxidation-bacterial denitrifier approach, Rapid Com- Buchwald, C., Grabb, K., Hansel, C. M., and Wankel, S. D.: Con- mun. Mass Sp., 27, 2239–2246, 2013. straining the role of iron in environmental nitrogen transforma- − tions: dual stable isotope systematics of abiotic NO2 reduction

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 827

Firestone, M. K. and Davidson, E. A.: Microbiological basis of NO as markers of chemodenitrification, Environ. Sci. Technol., 49, and N2O production and consumption in soil, in: Exchange of 3444–3452, 2015. trace gases between terrestrial ecosystems and the atmosphere, Kaiser, J., Hastings, M. G., Houlton, B. Z., Röckmann, T., and edited by: Andreae, M. O., Schimel, D. S., and Robertson, G. P., Sigman, D. M.: Triple oxygen isotope analysis of nitrate using John Wiley and Sons Ltd, Berlin, 7–21, 1989. the denitrifier method and thermal decomposition of N2O, Anal. Freitag, A., Rudert, M., and Bock, E.: Growth of Nitrobacter by Chem., 79, 599–607, 2007. dissimilatoric nitrate reduction, FEMS Microbiol. Lett., 48, 105– Kaneko, M. and Poulson, S. R.: The rate of oxygen isotope ex- 109, 1987. change between nitrate and water, Geochim. Cosmochim. Ac., Friedman, S. H., Massefski, W., and Hollocher, T. C.: Catalysis of 118, 148–156, 2013. intermolecular oxygen atom transfer by nitrite dehydrogenase of Kang, H., Stanley, E. H., and Park, S. S.: A sensitive method for the Nitrobacter agilis, J. Biol. Chem., 261, 10538–10543, 1986. measurement of ammonium in soil extract and water, Commun. Fry, B.: Stable Isotope Ecology, Vol. 521, Springer, New York, Soil Sci. Plan., 34, 2193–2201, 2003. 2006. Ke, X., Angel, R., Lu, Y., and Conrad, R.: Niche differentiation of Galloway, J. N., Aber, J. D., Erisman, J. W., Seitzinger, S. P., ammonia oxidizers and nitrite oxidizers in rice paddy soil, Envi- Howarth, R. W., Cowling, E. B., and Cosby, B. J.: The nitrogen ron. Microbiol., 15, 2275–2292, 2013. cascade, Bioscience, 53, 341–356, 2003. Keiluweit, M., Gee, K., Denney, A., and Fendorf, S.: Anoxic mi- Gaye, B., Nagel, B., Dähnke, K., Rixen, T., and Emeis, K. C.: Evi- crosites in upland soils dominantly controlled by clay content, dence of parallel denitrification and nitrite oxidation in the ODZ Soil Biol. Biochem., 118, 42–50, 2018. of the Arabian Sea from paired stable isotopes of nitrate and ni- Kemeny, P. C., Weigand, M. A., Zhang, R., Carter, B. R., trite, Global Biogeochem. Cy., 27, 1059–1071, 2013. Karsh, K. L., Fawcett, S. E., and Sigman, D. M.: Enzyme-level Granger, J. and Sigman, D. M.: Removal of nitrite with sulfamic interconversion of nitrate and nitrite in the fall mixed layer of acid for nitrate N and O isotope analysis with the denitrifier the Antarctic Ocean, Global Biogeochem. Cy., 30, 1069–1085, method, Rapid Commun. Mass Sp., 23, 3753–3762, 2009. 2016. Granger, J. and Wankel, S. D.: Isotopic overprinting of nitrification Kester, R. A., Meijer, M. E., Libochant, J. A., Boer, W. D., and on denitrification as a ubiquitous and unifying feature of environ- Laanbroek, H. J.: Contribution of nitrification and denitrification mental nitrogen cycling, P. Natl. Acad. Sci. USA, 113, E6391– to the NO and N2O emissions of an acid forest soil, a river sed- E6400, 2016. iment and a fertilized grassland soil, Soil Biol. Biochem., 29, Granger, J., Sigman, D. M., Lehmann, M. F., and Tortell, P. D.: Ni- 1655–1664, 1997. trogen and oxygen isotope fractionation during dissimilatory ni- Koch, H., Lücker, S., Albertsen, M., Kitzinger, K., Herbold, C., trate reduction by denitrifying bacteria, Limnol. Oceanogr., 53, Spieck, E., Nielsen, P. H., Wagner, M., and Daims, H.: Ex- 2533–2545, 2008. panded metabolic versatility of ubiquitous nitrite-oxidizing bac- Heil, J., Vereecken, H., and Brüggemann, N.: A review of chemical teria from the genus Nitrospira, P. Natl. Acad. Sci. USA, 112, reactions of nitrification intermediates and their role in nitrogen 11371–11376, 2015. cycling and nitrogen trace gas formation in soil, Eur. J. Soil Sci., Kool, D. M., Wrage, N., Oenema, O., Van Kessel, C., and Van 67, 23–39, 2016. Groenigen, J. W.: Oxygen exchange with water alters the oxy- Højberg, O., Binnerup, S. J., and Sørensen, J.: Potential rates of am- gen isotopic signature of nitrate in soil ecosystems, Soil Biol. monium oxidation, nitrite oxidation, nitrate reduction and deni- Biochem., 43, 1180–1185, 2011. trification in the young barley rhizosphere, Soil Biol. Biochem., Kozlowski, J. A., Stieglmeier, M., Schleper, C., Klotz, M. G., and 28, 47–54, 1996. Stein, L. Y.: Pathways and key intermediates required for obli- Homyak, P. M., Vasquez, K. T., Sickman, J. O., Parker, D. R., and gate aerobic ammonia-dependent chemolithotrophy in bacteria Schimel, J. P.: Improving nitrite analysis in soils: Drawbacks of and Thaumarchaeota, ISME J., 10, 1836–1845, 2016. the conventional 2 M KCl extraction, Soil Sci. Soc. Am. J., 79, Kremen, A., Bear, J., Shavit, U., and Shaviv, A.: Model demonstrat- 1237–1242, 2015. ing the potential for coupled nitrification denitrification in soil Hooper, A. B., Arciero, D., Bergmann, D., and Hendrich, M. P.: The aggregates, Environ. Sci. Technol., 39, 4180–4188, 2005. Oxidation of Ammonia as an Engergy Source in Bacteria, in: Kuypers, M. M., Marchant, H. K., and Kartal, B.: The microbial Respiration in archaea and bacteria, Springer, Dordrecht, 121– nitrogen-cycling network, Nat. Rev. Microbiol., 16, 263–276, 147, 2004. https://doi.org/10.1038/nrmicro.2018.9, 2018. Hudman, R. C., Russell, A. R., Valin, L. C., and Cohen, R. C.: Inter- Lancaster, K. M., Caranto, J. D., Majer, S. H., and Smith, M. A.: annual variability in soil nitric oxide emissions over the United Alternative bioenergy: updates to and challenges in nitrification States as viewed from space, Atmos. Chem. Phys., 10, 9943– metalloenzymology, Joule, 2, 421–441, 2018. 9952, https://doi.org/10.5194/acp-10-9943-2010, 2010. Lehnert, N., Dong, H. T., Harland, J. B., Hunt, A. P., and Hudman, R. C., Moore, N. E., Mebust, A. K., Martin, R. V., Rus- White, C. J.: Reversing nitrogen fixation, Nat. Rev. Chem., 2, sell, A. R., Valin, L. C., and Cohen, R. C.: Steps towards a mech- 278–289, 2018. anistic model of global soil nitric oxide emissions: implementa- Le Roux, X., Bouskill, N. J., Niboyet, A., Barthes, L., Di- tion and space based-constraints, Atmos. Chem. Phys., 12, 7779– jkstra, P., Field, C. B., Hungate, B. A., Lerondelle, C., 7795, https://doi.org/10.5194/acp-12-7779-2012, 2012. Pommier, T., Tang, J., and Terada, A.: Predicting the re- Jones, L. C., Peters, B., Lezama Pacheco, J. S., Casciotti, K. L., sponses of soil nitrite-oxidizers to multi-factorial global and Fendorf, S.: Stable isotopes and iron oxide mineral products change: a trait-based approach, Front. Microbiol., 7, 628, https://doi.org/10.3389/fmicb.2016.00628, 2016.

https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021 828 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil

Lewicka-Szczebak, D., Well, R., Köster, J. R., Fuß R., Sen- tion in a complex semiarid ecosystem using 117O, Environ. Sci. bayram, M., Dittert, K., and Flessa, H.: Experimental determina- Technol., 38, 2175–2181, 2004. tions of isotopic fractionation factors associated with N2O pro- Miller, D. J., Chai, J., Guo, F., Dell, C. J., Karsten, H., and Hast- duction and reduction during denitrification in soils, Geochim. ings, M. G.: Isotopic Composition of In Situ Soil NOx Emissions Cosmochim. Ac., 134, 55–73, 2014. in Manure-Fertilized Cropland, Geophys. Res. Lett., 45, 12–058, Lewicka-Szczebak, D., Dyckmans, J., Kaiser, J., Marca, A., Au- 2018. gustin, J., and Well, R.: Oxygen isotope fractionation during N2O Peters, B. D., Babbin, A. R., Lettmann, K. A., Mordy, C. W., Ul- production by soil denitrification, Biogeosciences, 13, 1129– loa, O., Ward, B. B., and Casciotti, K. L.: Vertical modeling of 1144, https://doi.org/10.5194/bg-13-1129-2016, 2016. the nitrogen cycle in the eastern tropical South Pacific oxygen Li, D. and Wang, X.: Nitrogen isotopic signature of soil-released deficient zone using high-resolution concentration and isotope nitric oxide (NO) after fertilizer application, Atmos. Environ., measurements, Global Biogeochem. Cy., 30, 1661–1681, 2016. 42, 4747–4754, 2008. Russow, R., Stange, C. F., and Neue, H. U.: Role of nitrite and nitric Lim, N. Y., Frostegård, Å., and Bakken, L. R.: Nitrite kinetics dur- oxide in the processes of nitrification and denitrification in soil: ing anoxia: The role of abiotic reactions versus microbial reduc- Results from 15N tracer experiments, Soil Biol. Biochem., 41, tion, Soil Biol. Biochem., 119, 203–209, 2018. 785–795, 2009. Liu, S., Lin, F., Wu, S., Ji, C., Sun, Y., Jin, Y., Li, S., Li, Z., and Santoro, A. E. and Casciotti, K. L.: Enrichment and characterization Zou, J.: A meta-analysis of fertilizer-induced soil NO and com- of ammonia-oxidizing archaea from the open ocean: phylogeny, bined NO+N2O emissions, Glob. Change Biol., 23, 2520–2532, physiology and stable isotope fractionation, ISME J., 5, 1796– 2017. 1808, https://doi.org/10.1038/ismej.2011.58, 2011. Loick, N., Dixon, E. R., Abalos, D., Vallejo, A., Matthews, G. P., Shoun, H., Fushinobu, S., Jiang, L., Kim, S. W., and Wakagi, T.: McGeough, K. L., Well, R., Watson, C. J., Laughlin, R. J., and Fungal denitrification and nitric oxide reductase cytochrome Cardenas, L. M.: Denitrification as a source of nitric oxide emis- P450nor, Philos. T. R. Soc. B, 367, 1186–1194, 2012. sions from incubated soil cores from a UK grassland soil, Soil Sigman, D. M., Casciotti, K. L., Andreani, M., Barford, C., Biol. Biochem., 95, 1–7, 2016. Galanter, M. B. J. K., and Böhlke, J. K.: A bacterial method for Maaz, T. M., Waldo, S., Bruulsema, T., and Mikkelsen, R.: In- the nitrogen isotopic analysis of nitrate in seawater and freshwa- consistencies undermine the conclusion that agriculture is a ter, Anal. Chem., 73, 4145–4153, 2001. dominant source of NOx in California, Sci. Adv., 4, 4706, Skiba, U., Fowler, D., and Smith, K. A.: Nitric oxide emissions from https://doi.org/10.1126/sciadv.aat4706, 2018. agricultural soils in temperate and tropical climates: sources, Maggi, F. and Riley, W. J.: Mathematical treatment of isotopologue controls and mitigation options, Nutr. Cycl. Agroecosys., 48, and isotopomer speciation and fractionation in biochemical ki- 139–153, 1997. netics, Geochim. Cosmochim. Ac., 74, 1823–1835, 2010. Stange, C. F., Spott, O., Arriaga, H., Menéndez, S., Estavillo, J. M., Mariotti, A., Germon, J. C., Hubert, P., Kaiser, P., Letolle, R., and Merino, P.: Use of the inverse abundance approach to iden- Tardieux, A., and Tardieux, P.: Experimental determination of ni- tify the sources of NO and N2O release from Spanish forest soils trogen kinetic isotope fractionation: some principles; illustration under oxic and hypoxic conditions, Soil Biol. Biochem., 57, 451– for the denitrification and nitrification processes, Plant Soil, 62, 458, 2013. 413–430, 1981. Stein, L. Y.: Insights into the physiology of ammonia-oxidizing mi- Mariotti, A., Leclerc, A., and Germon, J. C.: Nitrogen isotope frac- croorganisms, Curr. Opin. Chem. Biol., 49, 9–15, 2019. − → tionation associated with the NO2 N2O step of denitrification Sun, X., Ji, Q., Jayakumar, A., and Ward, B. B.: Dependence of in soils, Can. J. Soil Sci., 62, 227–241, 1982. nitrite oxidation on nitrite and oxygen in low-oxygen seawater, Martin, T. S. and Casciotti, K. L.: Nitrogen and oxygen iso- Geophys. Res. Lett., 44, 7883–7891, 2017. topic fractionation during microbial nitrite reduction, Limnol. Sutka, R. L., Ostrom, N. E., Ostrom, P. H., Breznak, J. A., Oceanogr., 61, 1134–1143, 2016. Gandhi, H., Pitt, A. J., and Li, F.: Distinguishing nitrous oxide Martin, T. S. and Casciotti, K. L.: Paired N and O isotopic analysis production from nitrification and denitrification on the basis of of nitrate and nitrite in the Arabian Sea oxygen deficient zone, isotopomer abundances, Appl. Environ. Microb., 72, 638–644, Deep-Sea Res. Pt. I, 121, 121–131, 2017. 2006. McKenney, D. J., Shuttleworth, K. F., Vriesacker, J. R., and Find- Taylor, A. E., Myrold, D. D., and Bottomley, P. J.: Temperature af- lay, W. I.: Production and loss of nitric oxide from denitrification fects the kinetics of nitrite oxidation and nitrification coupling in anaerobic Brookston clay, Appl. Environ. Microb., 43, 534– in four agricultural soils, Soil Biol. Biochem., 136, 107523, 541, 1982. https://doi.org/10.1016/j.soilbio.2019.107523, 2019. McKenney, D. J., Lazar, C., and Findlay, W. J.: Kinetics of the nitrite Toyoda, S., Yoshida, N., and Koba, K.: Isotopocule analysis of bio- to nitric oxide reaction in peat, Soil Sci. Soc. Am. J., 54, 106– logically produced nitrous oxide in various environments, Mass 112, 1990. Spectrom. Rev., 36, 135–160, 2017. Medinets, S., Skiba, U., Rennenberg, H., and Butterbach-Bahl, K.: Treibergs, L. A. and Granger, J.: Enzyme level N and O isotope ef- A review of soil NO transformation: Associated processes and fects of assimilatory and dissimilatory nitrate reduction, Limnol. possible physiological significance on organisms, Soil Biol. Oceanogr., 62, 272–288, 2017. Biochem., 80, 92–117, 2015. Van Cleemput, O. and Samater, A. H.: Nitrite in soils: accumulation Michalski, G., Meixner, T., Fenn, M., Hernandez, L., Sirulnik, A., and role in the formation of gaseous N compounds, Fert. Res., 45, Allen, E., and Thiemens, M.: Tracing atmospheric nitrate deposi- 81–89, 1995.

Biogeosciences, 18, 805–829, 2021 https://doi.org/10.5194/bg-18-805-2021 Z. Yu and E. M. Elliott: Nitrogen isotopic fractionations during NO production in an agricultural soil 829

Veldkamp, E. and Keller, M.: Fertilizer-induced nitric oxide emis- Yang, H., Gandhi, H., Ostrom, N. E., and Hegg, E. L.: Isotopic frac- sions from agricultural soils, Nutr. Cycl. Agroecosys., 48, 69–77, tionation by a fungal P450 nitric oxide reductase during the pro- 1997. duction of N2O, Environ. Sci. Technol., 48, 10707–10715, 2014. Venterea, R. T., Rolston, D. E. E., and Cardon, Z. G.: Effects of soil Ye, R. W., Averill, B. A., and Tiedje, J. M.: Denitrification: produc- moisture, physical, and chemical characteristics on abiotic nitric tion and consumption of nitric oxide, Appl. Environ. Microb., 60, oxide production, Nutr. Cycl. Agroecosys., 72, 27–40, 2005. 1053–1058, 1994. Venterea, R. T., Coulter, J. A., and Clough, T. J.: Ni- Yu, Z.: Isotopologue-specific models, Zenodo, trite accumulation and nitrogen gas production increase https://doi.org/10.5281/zenodo.4495715, 2021. with decreasing temperature in urea-amended soils: Exper- Yu, Z. and Elliott, E. M.: Novel method for nitrogen isotopic analy- iments and modeling, Soil Biol. Biochem., 142, 107727, sis of soil-emitted nitric oxide, Environ. Sci. Technol., 51, 6268– https://doi.org/10.1016/j.soilbio.2020.107727, 2020. 6278, 2017. Vinken, G. C. M., Boersma, K. F., Maasakkers, J. D., Adon, M., and Yu, Z. and Elliott, E. M.: Probing soil nitrification and nitrate con- 17 Martin, R. V.: Worldwide biogenic soil NOx emissions inferred sumption using 1 O of soil nitrate, Soil Biol. Biochem., 127, from OMI NO2 observations, Atmos. Chem. Phys., 14, 10363– 187–199, 2018. 10381, https://doi.org/10.5194/acp-14-10363-2014, 2014. Zhang, L., Altabet, M. A., Wu, T., and Hadas, O.: Sensitive mea- +15 14 15 + Wei, J., Ibraim, E., Brüggemann, N., Vereecken, H., and Mohn, J.: surement of NH4 N/ N(δ NH4 ) at natural abundance lev- First real-time isotopic characterisation of N2O from chemoden- els in fresh and saltwaters, Anal. Chem., 79, 5297–5303, 2007. itrification, Geochim. Cosmochim. Ac., 267, 17–32, 2019. Zhang, S., Fang, Y., and Xi, D.: Adaptation of micro-diffusion Wrage-Mönnig, N., Horn, M. A., Well, R., Müller, C., Velthof, G., method for the analysis of 15N natural abundance of ammonium and Oenema, O.: The role of nitrifier denitrification in the pro- in samples with small volume, Rapid Commun. Mass Sp., 29, duction of nitrous oxide revisited, Soil Biol. Biochem., 123, A3– 1297–1306, 2015. A16, 2018. Zhu-Barker, X., Cavazos, A. R., Ostrom, N. E., Horwath, W. R., Wunderlich, A., Meckenstock, R. U., and Einsiedl, F.: A mixture and Glass, J. B.: The importance of abiotic reactions for nitrous of nitrite-oxidizing and denitrifying microorganisms affects the oxide production, Biogeochemistry, 126, 251–267, 2015. δ18O of dissolved nitrate during anaerobic microbial denitrifica- Zumft, W. G.: Cell biology and molecular basis of denitrification, tion depending on the δ18O of ambient water, Geochim. Cos- Microbiol. Mol. Biol. R., 61, 533–616, 1997. mochim. Ac., 119, 31–45, 2013. Yamazaki, T., Hozuki, T., Arai, K., Toyoda, S., Koba, K., Fuji- wara, T., and Yoshida, N.: Isotopomeric characterization of ni- trous oxide produced by reaction of enzymes extracted from nitrifying and denitrifying bacteria, Biogeosciences, 11, 2679– 2689, https://doi.org/10.5194/bg-11-2679-2014, 2014.

https://doi.org/10.5194/bg-18-805-2021 Biogeosciences, 18, 805–829, 2021