Pseudomonas chlororaphis subsp. aureofaciens (formerly P. aureofaciens, ATCC 13985) and Pseudomonas chlororaphis subsp. chlororaphis (formerly P. chlororaphis, ATCC 9446) were obtained from the University of Basel Aquatic and Isotope Biogeochemistry culture collection. Growth protocols were adapted from the denitrifier method (Sigman et al., 2001; Casciotti et al., 2002; Weigand et al., 2016).

Two complementary incubation approaches were used to determine how species identity and physiological state influence N₂O isotopic signatures. All experiments were performed in either natural-abundance water (nat) or ¹⁸O-enriched water (en), and all media are described in Sect. 2.1.3.

Experiment codes follow the structure X_species_water, where X can be GF for gas flushed incubations or CB denotes closed-batch incubation, CB-R denotes a resuspension assay in which actively grown cultures were harvested by centrifugation and resuspended in a defined nitrate-containing medium, following an approach adapted from the denitrifier method used to analyse the N and O isotopic composition of nitrate (Sigman et al., 2001; Casciotti et al., 2002; Weigand et al., 2016), CB-RS indicates a resuspension assay using nitrate reference standards. Species can be aur or chlor, which denotes P. aureofaciens and P. chlororaphis, respectively, and water can be nat and en indicating natural-abundance or ¹⁸O-enriched water.

Gas-flushed (GF) incubations (GF_aur_nat, GF_aur_en, GF_chlor_nat) were conducted in a temperature-controlled bioreactor under continuous N₂ stripping, allowing real-time quantification of N₂O concentration and isotopic composition during active growth (Fig. 1).

Closed-batch (CB) incubations (CB_aur_nat, CB_chlor_nat, CB_aur_en, CB_chlor_en, CB-R_aur_nat, CB-R_chlor_nat, CB-RS_aur, CB-RS_chlor) were conducted in sealed vials and served three purposes. First, active-growth closed-batch incubations (CB_aur_nat, CB_chlor_nat, CB_aur_en, CB_chlor_en) were conducted in active-growth medium to generate cumulative N₂O for quantifying water–nitrite oxygen exchange during growth. Second, for stationary-phase resuspension incubations (CB-R_aur_nat, CB-R_chlor_nat), cultures were harvested by centrifugation (4700 rpm, 10 min) and resuspended in a defined nitrate medium prepared in natural-isotope-abundance water. Third, nitrate-standard resuspension assays (CB-RS_aur, CB-RS_chlor) replaced KNO₃ with USGS32, USGS34, IAEA-N3, or ¹⁸O-enriched IAEA-N3 (see Sect. 2.1.4).

Two distinct media formulations were used depending on experimental objective: an active-growth medium for gas-flushed and closed-batch incubations, and a defined resuspension medium for stationary-phase assays and nitrate-standard experiments. Active growth was carried out in a denitrifier-method medium formulation (Weigand et al., 2016) containing final concentrations of 10 mM KNO₃, 15 mM NH4+ supplied as (NH₄)₂SO₄, 30 g L⁻¹ tryptic soy broth (TSB, Merck Germany), and 5 g L⁻¹ K₂HPO₄. Cultures grown in this medium reduced nitrate quantitatively and produced N₂O for isotopic analysis.

For stationary-phase resuspension incubations, actively grown cultures were harvested by centrifugation and resuspended in a simplified medium (Weigand et al., 2016) containing 7.5 mM NH4+ supplied as NH₄Cl, 30 g L⁻¹ TSB, and 5 g L⁻¹ K₂HPO₄ in the final solution. Cell pellets were resuspended in 3 mL of this medium in 20 mL crimp-sealed vials. Vials were purged with N₂ for ∼ 1 h to establish an anoxic headspace and to remove pre-existing gaseous products (including N₂O) prior to initiating the incubation. The resuspension protocol was adapted from the denitrifier method; after purging, 20 nmol KNO₃ was added, and cultures were incubated overnight to allow quantitative conversion to N₂O. Thereafter, N₂O was analyzed by GC-IRMS from the vial headspace.

For CB_RS_aur and CB_RS_chlor (closed-batch resuspension assays converting nitrate reference standards with P. aureofaciens and P. chlororaphis, respectively; experiment codes defined in Sect. 2.1.3 and Table 1), the same resuspension procedure was used, except that the nitrate substrate (20 nmol KNO₃) was replaced with 20 nmol of an international nitrate isotope reference standard (USGS32, USGS34, IAEA-N3, or ¹⁸O-enriched IAEA-N3-spike).

Oxygen exchange between reaction intermediates and water in both GF and CB experiments was assessed using replicate incubations in natural-abundance and ¹⁸O-enriched water. The natural water used for all baseline incubations had a δ18O–H₂O = -9.51 ± 0.22 ‰. Two batches of ¹⁸O-enriched water were prepared and used: one with δ18O–H₂O = +68.31 ± 0.17 ‰, and a second with δ18O–H₂O = +73.53 ± 0.29 ‰ (Table 1). For nitrate-standard experiments (CB-RS-aur and CB-RS-chlor, Table 1), an ¹⁸O-enriched IAEA-N3 spike was prepared by mixing 1 mL of ¹⁸O-enriched water (δ18O–H₂O ≈ +780 ‰) with 3 mL of natural water (δ18O–H₂O = -9.5 ‰), yielding a solution with δ18O–H₂O ≈ +187.8 ‰.

The δ18O isotopic composition of both natural and ¹⁸O-enriched waters was measured using a cavity ring-down spectroscopy (CRDS) analyzer (L2130-i, Picarro Inc., USA). Measurements were calibrated on the VSMOW–SLAP scale using three international reference waters: VSMOW2 (δ18O = 0.00 ± 0.02 ‰), SLAP2 (-55.50 ± 0.02 ‰), and IAEA-607 (+99.02 ± 0.13 ‰), which were measured together with the samples during each analytical run to establish a three-point isotope calibration.

Accumulated N₂O produced during incubations was analyzed using gas chromatography coupled to isotope ratio mass spectrometry (GC-IRMS; Delta V Plus coupled to a GasBench system for N₂O purification, Thermo Fisher Scientific, Germany). Isotopic composition was referenced to AIR–N₂ (δ15N) and VSMOW (δ18O) using three laboratory N₂O reference gases (STD1–STD3; ∼ 90 ppm N₂O in N₂) measured alongside samples. Assigned delta values (in ‰) were: STD1: δ15N^α = -22.21 ± 0.39, δ15N^β = -49.28 ± 0.40, SP = 27.08 ± 0.56, δ15N^bulk = -35.74 ± 0.07, δ18O = 26.94 ± 0.23; STD2: δ15N^α = 1.71 ± 0.47, δ15N^β = 94.44 ± 0.70, SP = -92.73 ± 0.84, δ15N^bulk = 48.09 ± 0.23, δ18O = 36.01 ± 0.27; STD3: δ15N^α = 17.11 ± 0.12, δ15N^β = -3.43 ± 0.17, SP = 20.54 ± 0.21, δ15N^bulk = 6.85 ± 0.06, δ18O = 35.39 ± 0.17. Data processing, including scale normalization and correction for ion-source scrambling effects, was performed using the PyIsotopomer package (Kelly et al., 2023).

For continuous N₂O isotope monitoring under cultivation conditions, strains were grown in the active-growth medium described in Sect. 2.1.2 (Fig. 1). Incubations were performed at 30 ± 1 °C for P. aureofaciens and 20 ± 1 °C for P. chlororaphis, reflecting optimal conditions for growth and N₂O production. Cultures were stirred using a magnetic stirrer to ensure homogeneous mixing, and maintained anaerobic by continuous N₂ flushing (100 mL min⁻¹). Dissolved oxygen, pH (Mettler-Toledo, Switzerland), and optical density (OD₅₅₀) were monitored throughout the experiments. Substrate consumption was monitored by daily nitrite measurements. Incubations were typically completed within 5–8 d. Isotopic measurements were stopped when N₂O concentrations dropped below 50 ppm, whereas concentration measurements by FTIR continued until N₂O levels dropped below detection limits (0.0075 ppm N₂O), confirming complete turnover of nitrogen substrates (Fig. 2). Gas exiting the bioreactor was conditioned prior to analysis; carbon dioxide and water vapor were removed using two consecutive washing bottles containing 1 M NaOH, followed by permeation drying (Perma Pure, USA) (Fig. 1). The CO₂-free, dehumidified sample gas was subsequently filtered through a sintered metal filter (2 µm pore size) and analyzed for trace gas concentrations (N₂O, NO) using an online FTIR analyzer (CX-4000 or CX-4015, Gasmet Technologies, Finland) (Wunderlin et al., 2012). Three-minute average N₂O concentration values were used to calculate dilution ratios and regulate the MFC settings to maintain an approximately constant N₂O concentration (∼ 50 ppm) in the diluted sample gas, which minimized concentration effects on isotope ratio measurements by QCLAS (Wunderlin et al., 2013). Choosing 50 ppm as target concentration for N₂O isotope analysis was a compromise between data coverage and response time of the experimental setup but does not represent a lower threshold of the analytics.

Automated sample gas dilution and calibration of the QCLAS system were controlled via a LabVIEW interface (National Instruments Corp., USA) connected to mass flow controllers (Vögtlin Instruments Inc., Switzerland) and electromagnetic valves (Series 9, Parker Hannifin, USA).

Temporally-resolved (1 Hz) N₂O isotopic analysis in the diluted sample gas was performed using a compact quantum cascade laser absorption spectrometer (mini-TILDAS, pathlength 76 m, Aerodyne Research Inc., USA), operated in flow-through mode at a sample pressure of 26.6 hPa (20 Torr) (Ibraim et al., 2018). The instrument simultaneously quantified the concentrations of the four most abundant N₂O isotopologues in the ν3 absorption band (2203 cm⁻¹): ¹⁴N¹⁴N¹⁶O, ¹⁴N¹⁵N¹⁶O (¹⁵N^α), ¹⁵N¹⁴N¹⁶O (¹⁵N^β), and ¹⁴N¹⁴N¹⁸O. Raw concentrations were logged in ASCII format. Time series for each isotopologue were averaged to 60 s intervals as the first step in the data-processing workflow, prior to any correction or calibration. Isotopologue ratios (¹⁴N¹⁵N¹⁶O/14N¹⁴N¹⁶O, ¹⁵N¹⁴N¹⁶O/14N¹⁴N¹⁶O, and ¹⁴N¹⁴N¹⁸O/14N¹⁴N¹⁶O) were then calculated from these averaged concentrations using a customized R script. Allan deviations at a 60 s averaging time were used to propagate uncertainties in isotope delta values (see Appendix A).

Two N₂O isotope reference gases diluted in N₂ were used (Mohn et al., 2022); CG1 (RM1A, anchor gas, in ‰): δ15N^α = -0.22 ± 0.46, δ15N^β = 0.84 ± 0.46, δ18O = 39.22 ± 0.15; and CG2 (RM3A, span gas, in ‰): δ15N^α = 50.96 ± 0.47, δ15N^β = 53.06 ± 0.47, δ18O = 103.04 ± 0.16. CG1 was measured every 40 min for 10 min, to correct for instrumental drift, while the calibration span was determined twice daily analyzing CG1 and CG2. From the measured delta values, the following parameters were calculated: δ15N-bulk = (δ15N^α + δ15N^β)/2 and SP = δ15Nα-δ15N^β, following the nomenclature of Yoshida and Toyoda (2000). Data processing steps are detailed in Appendix A.

For validation an independent N₂O isotope reference gas (RM3B) with assigned values of δ15N-bulk = 16.08 ± 0.0 5 ‰, δ18O = 55.17 ± 0.15 ‰, and SP = -0.68 ± 0.91 ‰, was diluted using a two-step diluter (METAS, Switzerland) to a target concentration of 50 ppm N₂O. Measurements were conducted during a short validation campaign in April–May 2024. RM3B was analyzed within the existing CG1/CG2 calibration framework, and the measured values agreed with assigned values within <0.5 ‰ across all isotope parameters.

Isotope fractionation during denitrification was assessed using a Rayleigh-type cumulative product model, following the definitions of enrichment factors (ε) provided by Mariotti et al. (1981) and Sutka et al. (2006), which describe the isotopic discrimination between the substrate and the accumulated product over the course of the reaction.

Changes in the isotopic composition of N₂O were modeled as a function of nitrate reduction using the equation: 1y=a+b×[x] where y is the isotopic composition (δ value) of the N₂O product, a and b are fitted coefficients, and [x] represents the reaction progress. The variable [x] was calculated using the Rayleigh equation (Sutka et al., 2006): 2x=-f×ln⁡f(1-f) Here, f is the fraction of remaining nitrate, calculated as: 3f=1-2×nN2OnNO3- where n denotes the molar quantities of accumulated N₂O and initial NO3-, respectively. The slope b corresponds to the apparent enrichment factor ε (in ‰), while the intercept a represents the initial isotopic composition of the product pool. This interpretation assumes Rayleigh-type fractionation behavior, i.e. a closed substrate pool with unidirectional consumption and negligible back-reaction. Microbial incubations featuring multistep processes may deviate from these conditions, particularly when O isotope exchange with water or partial recycling of intermediates occurs. Nevertheless, the Rayleigh model provides a useful first-order approximation for comparing fractionation patterns across experiments.

More complex Rayleigh formulations have been proposed to explicitly account for intramolecular isotope effects during N₂O formation, including site-specific treatments of δ15N^α and δ15N^β (e.g., Rivett et al., 2025). In the present study, Rayleigh modeling was applied only to δ15N-bulk, for which the standard cumulative product formulation is sufficient.

Fractionation calculations were performed using a custom R pipeline that included uncertainty propagation for N₂O gas concentration, sample gas flow, and the initial substrate amount (NO3-). Full computational details are provided in the Appendix A.

During gas-flushed incubations, NO3- and NO2- concentrations were monitored daily to track substrate depletion and the progression of nitrate reduction. Nitrite was measured spectrophotometrically immediately after sampling using the Griess assay (Hansen and Koroleff, 1999). For nitrate concentration analysis, 1 mL samples were collected daily, filtered through 0.22 µm syringe filters, and stored in Eppendorf tubes at -18 °C prior to analysis. Nitrate concentrations were determined by quantitative reduction to nitric oxide (NO) using acidic vanadium(III), followed by chemiluminescence detection (Braman and Hendrix, 1989).

The isotopic composition of the nitrate substrates used to prepare incubation media was determined prior to the experiments. In addition, selected samples from GF incubations were analyzed to assess changes in the isotopic composition of the residual nitrate pool during denitrification. Nitrate isotope analyses were not performed for closed-batch incubations.

For nitrate isotopic analysis, aliquots containing 20 nmol N were converted to N₂O using the denitrifier method (Sigman et al., 2001; Casciotti et al., 2002). The nitrate stock solution used for all active-growth media had a δ15N–NO3- = +5.7 ± 0.3 ‰ and a δ18O–NO3- = +12.4 ± 0.3 ‰ (n=3).

Calibration was performed against international nitrate reference materials USGS32 (KNO₃, δ15N = +180 ‰, δ18O = +25.4 ± 0.2 ‰), USGS34 (KNO₃, δ15N = -1.8 ± 0.1 ‰, δ18O = -27.78 ± 0.37 ‰), and IAEA-N3 (KNO₃, δ15N = +4.7 ± 0.2 ‰, δ18O = +25.6 ± 0.4 ‰). Additional reference materials included UBN-1 (δ15N = +14.15 ‰, δ18O = +25.7 ‰), Deep Pacific nitrate (δ15N ≈ +5.0 ‰, δ18O ≈ +2.0 ‰), and an IAEA-N3-spike standard (IAEA-N3 dissolved in water with δ18O–H₂O = +780 ‰, see Sect. 2.1.4).

Closed-batch incubations were used to generate cumulative N₂O under controlled growth conditions. All experiments were conducted in 500 mL Wheaton bottles containing 400 mL of active-growth medium (Sect. 2.1.2) and incubated at 23 °C for 7 d. Experiments performed in natural-abundance water (CB_aur_nat, CB_chlor_nat) and ¹⁸O-enriched-water (CB_aur_en, CB_chlor_en) differed only in the isotopic composition of the water used to prepare the medium. The enriched-water incubations produced δ18O–N₂O endpoint values that were regressed against δ18O–H₂O to quantify oxygen atom exchange between nitrite and water. After complete nitrate reduction, accumulated N₂O was analyzed by GC-IRMS (Sect. 2.1.5).

To obtain stationary-phase δ18O–N₂O endpoint values, actively grown cultures were pelleted and resuspended in the defined nitrate medium described in Sect. 2.1.2. These resuspension assays (CB-R_aur_nat, CB-R_chlor_nat) were performed with stationary-phase cultures resuspended in fresh nitrate medium. The experiments were conducted only in natural-abundance water. After overnight nitrate reduction, the accumulated N₂O was analyzed by GC-IRMS (Sect. 2.1.5).

Additional stationary-phase incubations were performed to quantify oxygen-atom exchange between nitrite and water using nitrate standards with contrasting δ18O–NO3- values (CB-RS_aur, CB-RS_chlor): USGS32, USGS34, IAEA-N3. These were accompanied by an ¹⁸O-enriched IAEA-N3-spike to match the typical approach for quantifying oxygen-atom exchange in the denitrifier method (Sect. 2.1.2). Following quantitative reduction of the added nitrate, δ18O–N₂O was measured by GC-IRMS (Sect. 2.1.5). The resulting δ18O–N₂O values were regressed against the known δ18O–NO3- values of the standards. The δ18O–N₂O versus δ18O–NO3- relationship was applied to derive stationary-phase oxygen-exchange fractions.

This nitrate-standard approach provides an independent estimate of oxygen exchange complementary to the enriched-water experiments (Sect. 2.3.1). Whereas regressions of δ18O–N₂O versus δ18O H₂O quantify exchange from the sensitivity of N₂O to water isotopic composition, regressions of δ18O–N₂O versus δ18O–NO3- quantify the O-atom exchange from the degree of scale compression relative to nitrate isotope standards. Under conditions of no oxygen exchange, δ18O–N₂O scales directly with δ18O–NO3- (slope ≈ 1), whereas increasing exchange reduces this dependence (slope < 1). The fraction of oxygen atoms in N₂O derived from water can therefore be estimated from the regression slope (Casciotti et al., 2002; Snider et al., 2009; Lewicka-Szczebak et al., 2016).

A wide range of δ18O–N₂O endpoint values was observed in natural-abundance water incubations under the two active-growth modes (gas-flushed and closed-batch), as summarized in Fig. 6A. In closed-batch incubations, P. aureofaciens produced relatively low δ18O–N₂O values (∼ 9 ‰), whereas P. chlororaphis yielded substantially higher values (∼ 47 ‰). Under gas-flushed conditions, δ18O–N₂O endpoint values shifted in a strain-specific manner: values for P. aureofaciens increased to an intermediate range (26 ‰–42 ‰), while P. chlororaphis produced consistently elevated values (57 ‰–67 ‰). These contrasts indicate a strong incubation-mode sensitivity in P. aureofaciens, whereas P. chlororaphis maintained consistently elevated δ18O–N₂O values with a less pronounced response to incubation configuration. Time-resolved δ18O–N₂O trajectories during gas-flushed incubations further highlights this strain-dependent behavior (Fig. 3C). In P. aureofaciens, δ18O–N₂O starts at high values (∼ 50 ‰–60 ‰) early in nitrate reduction and declines progressively toward ∼ 30 ‰ as the reaction proceeds. In contrast, P. chlororaphis maintains elevated δ18O–N₂O values throughout the reaction, varying within a narrow band around 60 ‰. Thus, under identical conditions, P. aureofaciens shows strong variability in δ18O–N₂O, whereas P. chlororaphis maintains consistently high values with comparatively little temporal change.

To interpret the observed variability in δ18O–N₂O, three reference scenarios can be defined. Following the framework of Casciotti et al. (2002, 2007), nitrate reduction to nitrite is associated with an oxygen isotope fractionation of approximately 25 ‰–30 ‰, while NO2- reduction to N₂O is associated with a branching isotope effect of approximately 10 ‰–15 ‰. In the absence of oxygen atom exchange between nitrite and water, δ18O–N₂O is therefore expected to be approximately 50 ‰, reflecting the nitrate-source signature (δ18O–NO3- = +12.4 ± 0.3 ‰) shifted by the combined isotope effects described above. If nitrite fully equilibrates with water prior to reduction, its isotopic composition approaches δ18O–NO2-eq = δ18O–H₂O + ε18eq (here +4.5 ‰), resulting in δ18O–N₂O values of +14.5 ‰ to +19.5 ‰ after addition of the branching isotope effect (Casciotti et al., 2002; 2007). The δ18O of water itself (-9.5 ‰) represents a conceptual lower limit approached only by direct incorporation of O atoms from water with no kinetic isotope effects or equilibrium exchange. Because oxygen exchange requires sufficient residence time of nitrite, the transient NO2- accumulation observed during active growth (Fig. 2E) provides a mechanistic basis for such exchange.

Against this framework, the two strains exhibit contrasting behavior. P. aureofaciens, in closed-batch incubations, yields δ18O–N₂O values below the full exchange plus branching range, indicating extensive oxygen exchange, while gas-flushed conditions produce intermediate values (26 ‰–42 ‰), consistent with reduced but still substantial exchange. In contrast, P. chlororaphis yields consistently high δ18O–N₂O values under both incubation modes (≈47 ‰–67 ‰), clustering near the nitrate-dominated regime and indicating limited equilibration with water.

These patterns are consistent with previous observations under comparable active-growth conditions (Rohe et al., 2017; Haslun et al., 2018), but here can be interpreted explicitly in terms of oxygen exchange dynamics. Taken together, these results show that δ18O–N₂O reflects the combined influence of nitrate-derived oxygen, oxygen-atom exchange with water, and branching fractionation, which vary with strain and incubation mode. To quantify these contributions, we derived oxygen exchange rates using ¹⁸O-enriched water experiments (Sect. 3.4.2).

To quantify oxygen exchange between nitrite and water during active growth under gas-flushed conditions, P. aureofaciens (NirK) was incubated in ¹⁸O-enriched water (GF_aur_en; Table 1). δ18O–N₂O values began near ∼ 40 ‰, well below the δ18O of the enriched water (+68 ‰; Fig. 7). As nitrate reduction progressed, δ18O–N₂O increased steadily, crossed the δ18O-H₂O level, and approached 80 ‰ toward the end of the incubation. Throughout the experiment, δ18O–N₂O remained below values, anticipated for δ18O-NO2- under the nitrite–water equilibrium (δ18O-NO₂eq ≈ +82 ‰) and well below the expected range for δ18O-N₂O derived from fully equilibrated nitrite water combined with branching fractionation (δ18O-NO2-eq + εB18 ≈ +91 ‰ to +97 ‰; Fig. 7).

This trajectory – beginning below δ18O-H₂O, rising above it, and stabilizing below δ18O-NO₂eq – indicates that the N₂O product contains a mixture of nitrate-derived and water-exchanged O atoms, with incomplete approach to full nitrite–water equilibration. The initial δ18O–N₂O values reflect expression of the branching isotope effect under minimal water exchange, whereas the later ¹⁸O enrichment reflects the combined influence of branching fractionation and increasing nitrite–water exchange. This behavior is consistent with the elevated δ18O–N₂O endpoints observed for gas-flushed incubations in natural-abundance water (Figs. 3C; 6A).

To quantify these processes, δ18O–N₂O values from natural-water and ¹⁸O-enriched incubations were combined in δ18O–N₂O vs. δ18O–H₂O regressions (Fig. 8), following the framework of Casciotti et al. (2002). In this approach, the regression slope represents the fraction of oxygen atoms in N₂O derived from water during nitrite reduction, whereas the y-intercept at δ18O–H₂O = 0 ‰ isolates the component of the N₂O oxygen isotope signature independent of water isotopic composition. Under conditions of complete nitrite–water equilibration prior to reduction, the intercept approximates the sum of the nitrite–water equilibrium isotope effect and the branching isotope effect (i.e., b ≈ ε18eq + ε18B; Casciotti et al., 2002, 2007). Deviations from this expectation indicate incomplete equilibration and/or retention of nitrate-derived oxygen in the N₂O product.

Closed-batch incubations with P. chlororaphis (NirS) exhibited a slope of 0.67 ± 0.05, consistent with partial water exchange (Fig. 8). Its intercept (53.7 ± 4.1 ‰) is strongly elevated relative to ε18eq + ε18B (+24 ‰ to +30 ‰), reflecting combined contributions from nitrate-derived oxygen, incomplete exchange, and branching effects.

For P. aureofaciens, closed-batch incubations yielded a slope of 1.03 ± 0.06, indicating complete oxygen-atom exchange of nitrite with water during active growth. The corresponding intercept (18.7 ± 4.5 ‰) is at the lower end of values expected for ε18eq + ε18B. In contrast, gas-flushed incubation of P. aureofaciens yielded a slope of 0.38 ± 0.01, indicating only partial oxygen-atom exchange. The intercept (42.6 ± 1.4 ‰) is substantially higher than both the ε18eq + ε18B expectation and the fully exchanged closed-batch intercept, indicating strong retention of nitrate-derived oxygen and enhanced expression of branching fractionation under gas-flushed conditions.

These results demonstrate that oxygen isotope exchange in P. aureofaciens during active growth is highly sensitive to incubation conditions, spanning from complete to strongly limited exchange. In contrast, P. chlororaphis maintained consistently high oxygen-atom exchange across the experimental conditions investigated. These contrasts suggest that the canonical low-exchange behavior attributed to NirK is not an intrinsic property of the enzyme, but instead a condition-specific outcome of the denitrifier method.

To assess whether exchange patterns observed during active growth reflect intrinsic enzyme behavior or experimental conditions, we performed resuspension experiments in which actively grown cultures were pelleted and transferred into a defined nitrate medium. This setup places both strains into a stationary-phase reduction regime closely resembling the denitrifier method (Casciotti et al., 2002; Weigand et al., 2016).

Resuspension in natural-abundance water produced high and strain-specific δ18O–N₂O endpoints (Fig. 6B): 59.6 ± 2.5 ‰ for P. aureofaciens and 50.3 ± 0.04 ‰ for P. chlororaphis. These values lie well above the range expected under full nitrite–water equilibration (∼ 14.5 ‰–19.5 ‰), indicating limited oxygen exchange and strong retention of nitrate-derived oxygen, consistent with denitrifier-method observations.

Because endpoint values do not directly constrain exchange, nitrate standards spanning a range of δ18O–NO3- values can be applied to quantify exchange from δ18O–N₂O vs. δ18O–NO3- regressions. Across both nitrate-standard pairs (USGS32–USGS34 and IAEA-N3–¹⁸O-enriched IAEA-N3), the two strains reproduced the canonical low- versus high-exchange behavior reported previously (Casciotti et al., 2002; Appendix B Fig. B2). P. aureofaciens exhibited low oxygen exchange (1.5 %–4.3 %), whereas P. chlororaphis showed substantially higher exchange (64.7 %–76.6 %), in close agreement with published ranges. These results confirm that resuspension experiments recover the expected strain-specific exchange behavior under denitrifier-method conditions.