Oxygen Isotopic Fractionation of O₂ Consumption by Methane and Ammonia Monooxygenases

Abstract

Understanding stable isotopic fractionation of dissolved O₂ in aquatic environments is crucial to constrain and accurately model the processes responsible for biological O₂ consumption, which are closely linked to the overall health of an ecosystem. This study aimed to investigate whether O₂ consumption by microbial methane and ammonia oxidation may contribute to the observed discrepancy in O₂ isotopic fractionation (¹⁸ϵ) between heterotrophic O₂ respiration in laboratory incubations (−18 to −24 ‰) and in situ measurements of O₂ consumption in lakes and oceans (−10 to −18 ‰). To estimate the in vivo ¹⁸ϵ values of soluble methane monooxygenase (sMMO), particulate methane monooxygenase (pMMO), and ammonia monooxygenase (AMO), which are the first enzymes required for the oxidation of methane and ammonia, experiments were performed with three methanotrophic bacteria and one comammox (complete-ammonia-oxidizing) bacterium. The resulting ¹⁸ϵ values for pMMO and AMO ranged from −18 ± 12 to −24 ± 5 ‰, not significantly different from ¹⁸ϵ values typical for heterotrophic respiration. The ¹⁸ϵ value determined for sMMO (−22 ± 2 ‰) was in the same range, yet more negative than the previously reported ¹⁸ϵ value for the isolated enzyme. Our results provide insights into the potential reaction mechanisms of pMMO and AMO and indicate that O₂ consumption by sMMO, pMMO, or AMO cannot explain the observed discrepancy between in situ and laboratory ¹⁸ϵ values for “community” O₂ consumption in aquatic environments. Instead, the apparent difference may be attributed to aspects involving substrate diffusion limitation.

1. Introduction

Dissolved oxygen (O₂) is the most important oxidizing agent in aquatic ecosystems, and its concentration controls the rate of many biotic and abiotic processes, from cellular metabolism to the biogeochemical cycling of elements. O₂ concentrations and the redox state of an aquatic ecosystem represent important parameters that are directly linked to water quality and trophic status. As such, understanding O₂ consumption and production dynamics in aquatic environments is imperative to comprehensively assess ecosystem functioning, health, and resilience. By analyzing the stable isotopic composition of dissolved O₂ in aquatic environments, quantitative information can be gained on critical ecosystem processes producing or consuming O₂, such as primary production and respiration. , Changes in the isotopic composition of dissolved O₂ (reported in the common δ¹⁸O notation, see eq ) primarily result from the preferential reaction of ¹⁶O¹⁶O in biological and abiotic processes, leaving the remaining O₂ enriched in ¹⁶O¹⁸O. Changes in δ¹⁸O values of O₂, referred to as isotopic fractionation, can be quantified with ¹⁸ϵ values or apparent ¹⁸O kinetic isotope effects (¹⁸O-KIEs) according to eq ,

$${}^{18}ϵ=\frac{\ln \left(\frac{\delta {}^{18}\mathrm{O}+1}{\delta {}^{18}\mathrm{O}_0+1}\right)}{\ln ([{\mathrm{O}}_2]/{[{\mathrm{O}}_2]}_0)}=\frac{1}{{}^{18}\mathrm{O}‐\mathrm{KIE}}-1=\frac{1}{{}^{16}k/{}^{18}k}-1$$ 1

where δ¹⁸O and δ¹⁸O₀ indicate the isotopic composition of O₂ in a sample at a certain time point and in a reference sample representing initial conditions (t ₀), respectively. [O₂]/[O₂]₀ is the fraction of the remaining O₂ concentration upon partial consumption since t ₀, while ¹⁶ k and ¹⁸ k refer to the reaction rates of the light ¹⁶O¹⁶O and heavy ¹⁸O¹⁶O isotopologues, respectively.

Because ¹⁸ϵ values are typically characteristic for different processes causing changes in dissolved O₂ concentrations, i.e., respiration, photosynthesis, and gas–water exchange, they can be used to identify, track, and quantify O₂-consuming or producing processes in the environment. As such, several studies have applied stable isotope analysis of O₂ to study O₂ consumption dynamics in oceans and lakes. − This approach usually involves determining ¹⁸ϵ values for respiration to estimate respiration rates, which then yield information about gross primary production. Typically, ¹⁸ϵ values are obtained from laboratory incubation experiments with water from the study site or by measuring the δ¹⁸O in situ as a function of depth and/or along other O₂ concentration gradients. The “community” ¹⁸ϵ values obtained from incubation experiments have been shown to vary from −18 to −24 ‰, ,− similar to the ¹⁸ϵ values determined for microbial respiration by the terminal cytochrome c oxidase (−18 to −21 ‰). − In contrast, ¹⁸ϵ values determined from water column O₂ isotope measurements tend to be considerably lower, ranging from −10 to −18 ‰. − , This may suggest that the enzyme-level O isotope effect associated with respiration is suppressed in situ, e.g., due to diffusion limitation. Alternatively, the observed lower water column ¹⁸ϵ values may also reflect the combined isotope effects of multiple distinct processes occurring in parallel within the same water mass. It is important to understand which processes modulate the observed water column ¹⁸ϵ value and thereby determine what causes the discrepancy between the ¹⁸ϵ values observed in incubation experiments versus the ¹⁸ϵ values obtained from δ¹⁸O and O₂ concentration profiles in natural marine or lacustrine water columns. A better insight into this difference is essential for assessing O₂-consuming processes and for accurately modeling and estimating O₂ consumption dynamics in aquatic environments. So far, this discrepancy has been suggested to occur due to diffusion-limited O₂ consumption in sediments, , diffusion-limited respiration by microorganisms in sinking organic matter particles, , mixing processes, or reduced biological O isotopic fractionation during O₂ respiration at low temperatures as well as O₂ consumption by biologically produced reactive oxygen species (ROS). Recently, we determined that the biological isotopic fractionation of O₂ can range from −9 to −53 ‰, depending on the type of O₂-consuming enzyme. Although aerobic microbial respiration typically dominates observed biological O₂ consumption, other biological O₂-consuming processes may also contribute significantly to the observed ¹⁸ϵ values, potentially explaining the observed ¹⁸ϵ value discrepancies if they impart different O isotope effects.

Besides aerobic microbial respiration, methane and ammonia oxidation are prevalent biological processes in aquatic environments, which may contribute significantly to the total “community” O₂ consumption. − For example, estimates of global methane oxidation in oceans and freshwaters translate to an annual O₂ consumption of approximately 8·10¹³ g (assuming a 1O₂:1CH₄ stoichiometry). This value is 7 times larger than estimates of annual O₂ consumption by photochemical reactions in marine waters, which is considered to be a major sink of O₂ at least in surface waters. Aerobic methane- and ammonia-oxidizing microorganisms are not only widespread in the environment but also play dominant roles in specific environmental niches, where their metabolic activities significantly impact the biogeochemical cycling of carbon and nitrogen, respectively. − Aerobic methane oxidation is the process by which methanotrophic bacteria oxidize methane (CH₄) to carbon dioxide using O₂ in a 2O₂:1CH₄ stoichiometry (see Figure ). − The first step of CH₄ oxidation, i.e., to methanol (CH₃OH), involves stoichiometric O₂ consumption and is catalyzed by a methane monooxygenase (MMO) enzyme, either soluble methane monooxygenase (sMMO) or particulate methane monooxygenase (pMMO). sMMO is a cytosolic iron-dependent O₂-consuming enzyme, whereas pMMO is membrane-bound, copper-dependent, and the most prevalent MMO among methanotrophic microorganisms. In aerobic ammonia oxidation, nitrifying microorganisms utilize O₂ to oxidize ammonia (NH₃) first to nitrite (NO₂ ^–) and subsequently to nitrate (NO₃ ^–). The overall stoichiometries of the consumption of O₂ consumption vs NH₃ oxidation are 1.5:1 and 2:1 for the oxidation to nitrite and for complete oxidation to nitrate, respectively (see Figure ). The latter can involve two distinct organisms (ammonia-oxidizing bacteria or archaea in conjunction with nitrite-oxidizing bacteria) or a single comammox (complete-ammonia-oxidizing) bacterium. − The first step of nitrification, ammonia oxidation to hydroxylamine (NH₂OH), concomitant with O₂ consumption, is performed by the membrane-bound, copper-dependent enzyme ammonia monooxygenase (AMO). ,

1.

Key reaction steps in the oxidation of methane (CH₄) by methanotrophic bacteria and ammonia (NH₃) by ammonia-oxidizing microorganisms and comammox bacteria. Dashed-lined boxes highlight the reactions and overall stoichiometry of incomplete CH₄ oxidation and cell biomass ({CH₃O}) formation via methanol (CH₃OH) and formaldehyde (CH₃O), and NH₃ oxidation to nitrite (NO₂ ^–) via hydroxylamine (NH₂OH), respectively. Solid-lined boxes highlight the reactions and overall stoichiometries of complete CH₄ oxidation to CO₂ via formate (HCOO^–) and NH₃ oxidation to nitrate (NO₃ ^–), respectively. Enzymes involved in these reactions are methane monooxygenase (MMO), methanol dehydrogenase (MDH), formaldehyde dehydrogenase (FADH), formate dehydrogenase (FDH), ammonia monooxygenase (AMO), hydroxylamine oxidoreductase (HAO), nitrite oxidoreductase (NXR), and aerobic terminal oxidases. *We note that HAO only catalyzes the reaction of NH₂OH to nitric oxide (NO), which is subsequently converted to NO₂ ^–.

Notably, pMMO and AMO share many characteristics, including membrane association, copper dependency, substrate specificity, putative subunit compositions, and a similar DNA sequence. ,, Because they are membrane-bound, both enzymes have been proven difficult to purify while maintaining their activity, ,, which has limited our understanding of their structure, mechanism, and the O isotopic fractionation they impart. In contrast, the cytosolic iron-dependent sMMO has been successfully purified with activity, and ¹⁸ϵ values between −15 and −17 ‰ have been determined for the isolated sMMO of Methylococcus capsulatus. Despite their prevalence and environmental relevance, ¹⁸ϵ values for pMMO or AMO have, to the best of our knowledge, not yet been determined. Other copper-dependent O₂-consuming enzymes have been shown to express ¹⁸ϵ values between −9 and −22 ‰. An ¹⁸ϵ value closer to −9 ‰ for pMMO or AMO could explain the apparent under-expression of the respiration O isotope effect of approximately −20 ‰ in environments where aerobic methane and ammonia oxidation represent important processes. In contrast, an ¹⁸ϵ value closer to −22 ‰ for pMMO or AMO would make the O isotope effects for methane and ammonia oxidation indistinguishable from that of respiration, minimizing their potential to suppress the O isotope effect of water column respiration.

In this study, we determined the in vivo ¹⁸ϵ values for pMMO, sMMO, and AMO by performing experiments with the pMMO-containing Methylomonas lenta and Methylotetracoccus oryzae, sMMO-containing Methylocella silvestris, and with the AMO-containing Nitrospira inopinata, respectively. While M. lenta possesses the genes for both pMMO and sMMO, when grown in the presence of excess copper (typical culturing conditions), it expresses only pMMO. M. oryzae expresses only pMMO and M. silvestris only sMMO. , By measuring the O₂-to-CH₄ and O₂-to-NH₃ consumption stoichiometries, we first determined the contribution of O₂ consumption by pMMO, sMMO, and AMO, respectively, to the overall O₂ consumption (i.e., relative to terminal respiratory oxidase activity). Second, based on the observed stoichiometries and measured ¹⁸ϵ values for total O₂ consumption determined for each microorganism, we were able to calculate in vivo ¹⁸ϵ values for pMMO, sMMO, and AMO. Additionally, by supplying the methanotrophs with methanol as the substrate, we determined an isolated respiration value by the terminal cytochrome c oxidase. Our work provides novel insights into the previously unexplored O isotopic fractionation associated with the activity of two environmentally relevant O₂-consuming enzymes, MMO and AMO. These new insights will be discussed in the context of whether O₂ consumption by methane and/or ammonia oxidation in natural environments can explain the previously observed discrepancies in the “community” ¹⁸ϵ values for biological O₂ consumption in the water column of lakes or oceans.

2. Materials and Methods

2.1. Bacterial Cultures

Pure cultures of M. lenta (LMG26260) and M. oryzae were grown in 60 to 150 mL serum vials, containing 15 to 60 mL of dilute nitrate mineral salt solution (dNMS: 0.2 g L^–1 MgSO₄·7H₂O, 0.006 g L^–1 CaCl₂·2H₂O, 0.2 g L^–1 NaNO₃), supplemented with 5 mM phosphate buffer (pH 7.0), 0.1% (v/v) trace element solution I (10 g L^–1 trisodium nitrilotriacetate and 5 g L^–1 FeSO₄·7H₂O), and II (0.43 g L^–1 ZnSO₄·7H₂O, 0.24 g L^–1 CoCl₂·6H₂O, 0.99 g L^–1 MnCl₂·4H₂O, 0.25 g L^–1 CuSO₄·5H₂O, 0.22 g L^–1 NaMoO₄·2H₂O, 0.19 g L^–1 NiCl₂·6H₂O, 0.21 g L^–1 NaSeO₄·10H₂O, 0.14 g L^–1 H₃BO₄, and 0.24 g L^–1 CeCl·6H₂O), and a headspace of 20% CH₄ in air. , Cells were grown at 22 °C, with shaking at 150 rpm for 1 to 2 weeks. M. silvestris (DSMZ 15510) was cultured under similar conditions but at pH 5.8 and 26 °C. A pure culture of N. inopinata was grown aerobically in 1000 mL Schott bottles containing 200 mL of mineral medium (0.05 g L^–1 KH₂PO₄, 0.075 g L^–1 KCl, 0.05 g L^–1 MgSO₄·7H₂O, 0.584 g L^–1 NaCl) supplemented with 4 mM HEPES (pH 7.8) and 0.1% (v/v) of trace metal solutions (see composition above), at 37 °C, in the dark, without shaking. The pH was maintained by the addition of Na₂CO₃. All experiments were performed with actively growing batch cultures.

2.2. O₂ Consumption Experiments with Methanotrophs

2.2.1. Experiments with Methane as the Substrate

To determine the ¹⁸ϵ values of O₂ consumption by different methanotrophs during CH₄ oxiation, experiments were conducted in duplicate for all three methanotrophic species in a 60 mL incubation chamber that housed a membrane-inlet mass spectrometer (MIMS) probe following previously described procedures. , The probe (3 mm in diameter with 4–16 perforated holes of 1 mm diameter each) was covered by silicon tubing and connected to the mass spectrometer, which was operated at a 40 μA emission current, via a 1/8- or 1/16-in. stainless-steel tube.

To monitor dissolved O₂ concentrations the incubation chamber was equipped with an optical O₂ sensor spot read by a FireStingO₂ pro meter (PyroScience GmbH) with automated pressure, humidity, and temperature correction. A calibration point for 100% dissolved O₂ was achieved by purging the water-filled chamber with air until stable readings (i.e., equilibrium concentrations) were obtained at the set experimental temperatures (268 and 260 μM at 25 and 28 °C, respectively; salinity 0 ppm). 0% dissolved O₂ was assigned to stable low-level readings achieved after biological O₂ consumption with excess substrate.

The chamber was placed on a magnetic stirring plate, and temperature was regulated by an external water bath connected to the chamber’s water jacket. Experiments with M. lenta and M. oryzae were conducted at 28 °C, experiments with M. silvestris at 25 °C. Liquid and gas could be introduced into the chamber via a central hole in the piston (0.8 mm diameter), using metal capillaries or gastight syringes. Any gas headspace was eliminated before the start of the measurement. The central piston hole was sealed with a 1/16-in. PEEK tube (0.76 mm inner diameter) Luer-locked to a stainless-steel needle (0.41 mm o.d.), through which samples were obtained by lowering the piston. When taking a sample, the first mL of solution was always discarded, then the PEEK tube-attached needle was injected into a 12 mL Exetainer (Labco Limited), and 3 to 7 mL of the sample was transferred by lowering the piston. Prior to starting an experiment, Exetainers were closed with chlorobutyl septa and screw caps, purged with He gas for 1 h, and amended with 200 μL of 3.2 M HCl to inactivate the bacteria upon sample transfer. To ensure equal headspace pressure in the Exetainers with different sample volumes, Exetainer septa were pierced with a stainless-steel needle (0.45 mm o.d.) connected with a T-piece to a slow He flow, and an open outlet submerged under 10 cm of water during sample injection, as described in de Carvalho et al. Procedural blanks were prepared by filling the incubation chamber with N₂-purged (<0.1% O₂) water and transferring 3 to 7 mL of this O₂-free water to He-purged Exetainers containing HCl, as described above. After injection, Exetainers containing samples or procedural blanks were mixed gently and stored upside down in the dark, with the lid submerged under water, until isotope analysis was performed within 16 days after sample/blank collection (see Section ). We experimentally verified that within this time frame isotopic fractionation of O₂ consumption can be determined reliably.

CH₄ concentrations were determined from the MIMS mass-to-charge (m/z) signal at m/z 15 (i.e., CH₃ ⁺, a fragment formed in the ion source). m/z 15 was selected over m/z 16 due to lower background signals. A calibration was performed by sequentially injecting two 1 mL aliquots of CH₄-saturated water (1.5 mM CH₄) into the water-filled, initially CH₄-free chamber while continuously recording a signal. This procedure resulted in three calibration points at 0, 30, and 60 μM dissolved CH₄. The CH₄-saturated water was prepared at room temperature in a sealed bottle under a known pressure (1.10–1.14 bar) of pure CH₄. Accuracy at higher CH₄ concentrations was verified by equilibrating the liquid filled chamber with 30% CH₄ in air, resulting in a dissolved CH₄ concentration of 450 μM.

For each experiment, the chamber was filled with 52–57 mL of dNMS medium. The medium was purged with air to obtain air-saturated O₂ concentrations. CH₄ gas was introduced into a 1 to 2 mL headspace to achieve an initial dissolved CH₄ concentration between 250 and 450 μM. The generation of a CH₄ headspace led to a decrease of the dissolved O₂ concentrations to 184 to 214 μM. Once the desired experimental starting conditions were achieved, the CH₄ headspace was removed, and 0.1 to 1 mL of a bacterial suspension (obtained by centrifugation of 30 to 120 mL of active culture at 4000 rpm and subsequent resuspension in fresh medium) was added to start the reaction. Experiments were stirred at 500 rpm, except when adding bacterial suspensions. The first sample (3 mL) was taken immediately after adding bacterial suspension at a dissolved O₂ concentration of 193 ± 8 μM. Four additional samples (3–7 mL) were taken at decreasing O₂ concentrations down to 53 ± 2 μM.

2.2.2. Experiments with Methanol as the Substrate

To determine ¹⁸ϵ values representative for O₂ consumption by methanotrophic respiration alone (i.e., excluding O₂ consumption by pMMO or sMMO), methanol was used as the substrate. Additionally, to assess the potential impact of background O₂ consumption by the MIMS probe (up to 5%) on the measured δ¹⁸O values, replicate experiments were conducted in two distinct setups: (1) in the 60 mL incubation chamber with an active MIMS probe, and (2) in a 50 mL gastight glass syringe, where background O₂ consumption by the MIMS probe was absent.

Experiments were conducted in both the incubation chamber with the MIMS probe, as described above, but with 2.3 mM methanol instead of CH₄, and in the gastight glass syringe, containing a stir bar and optical sensor spots for measuring O₂ concentrations and temperature (PyroScience GmbH), at room temperature as described in de Carvalho et al. In brief, the syringe was filled with 50 mL of air-saturated dNMS medium containing 0.1 to 1 mL of concentrated bacterial biomass. A 100 μL gastight glass syringe was used to add an aqueous solution of methanol (for a final concentration of 2.3 mM) through the Luer-tip opening of the syringe to initiate the reaction. Immediately after methanol addition, a stainless-steel needle (0.8 mm o.d.) was attached to the Luer-Lock tip of the syringe and pushed into a 12 mm thick chlorobutyl stopper to limit O₂ exchange with the atmosphere. Six samples of 3–7 mL were taken from this syringe reactor and transferred to He-purged Exetainers as described in Section . Residual O₂ concentration in the samples typically ranged from 200 to 50 μM. A control sample was prepared by transferring 3 mL of remaining assay solution without methanol with a 10 mL gastight glass syringe into a He-purged Exetainer containing 200 μL of 3.2 M HCl. These control samples were assumed to be representative of the initial O₂ concentration (270 ± 5 μM) and isotopic composition. Blanks were prepared by transferring 3 to 7 mL of N₂-purged water with a 50 mL gastight glass syringe from closed serum bottles into He-purged Exetainers. Exetainers containing samples and blanks were stored as described above.

2.3. O₂ Consumption Experiments with N. inopinata

To determine the ¹⁸ϵ values of O₂ consumption by the comammox bacterium N. inopinata during NH₃ oxidation, experiments were conducted in a 50 mL gastight glass syringe, as described in Section . The assay solution was prepared by adding 15 mL of ammonium (NH₄ ⁺)-free mineral medium supplemented with 20 mM HEPES (pH 7.8) to 45 mL of an active N. inopinata culture. The syringe was filled headspace-free and placed onto a stirring plate (500 rpm) inside an incubator (37 °C). The assay solution contained approximately 100 μM NH₄ ⁺, but an additional 500 μM NH₄ ⁺ was introduced through the Luer-lock tip of the syringe at the start of the reaction. Six 2 mL samples were taken for NH₃, NO₃ ^–, and NO₂ ^– concentration measurements, and six additional samples (3–7 mL) were taken for O₂ isotope analysis. Samples for NH₃, NO₃ ^–, and NO₂ ^– concentration measurements, collected in 2 mL Eppendorf tubes, were always taken directly before the samples for isotope analysis. The latter were collected into 12 mL He-purged Exetainers as described in Section , containing 50 μL saturated HgCl₂ solution. Adding HCl did not completely stop the reaction; instead, it led to additional O₂ consumption likely due to the reaction with nitrous acid. All handling of HgCl₂ was performed with appropriate caution, and all solutions were disposed of as hazardous waste. The first samples were taken immediately (i.e., approximately 30 s) after the NH₄ ⁺ addition, representing the initial O₂, (NH₄ ⁺ + NH₃), NO₃ ^–, and NO₂ ^– concentrations and δ¹⁸O values of O₂. The remaining five samples were collected at O₂ concentrations of approximately 150, 120, 90, 70, and 50 μM. Samples collected for the dissolved inorganic nitrogen species were immediately passed through a sterile membrane filter (Sarsted, Filtropur, 0.2 μM pore size). Concentrations of (NH₄ ⁺ + NH₃) were determined by Hach Lange ammonium tests (0.2 to 2 mg L^–1 NH₄ ⁺ + NH₃, HACH). NO₂ ^– was determined colorimetrically using sulfanilamide and N-(1-Naphthyl)­ethylenediamine according to Hansen and Koroleff. NO _x (i.e., NO₂ ^– + NO₃ ^–) was determined using an NO _x TELEDYNE T200 Analyzer (TELEDYNE API, CA, USA) involving the reduction of NO _x with hot acidic Vanadium­(III) solution to NO gas and subsequent chemiluminescence detection.

2.4. Stable Isotope Analysis

The δ¹⁸O values of O₂ were measured in the headspace of the 12 mL Exetainers using a GasBench II coupled via a Conflo IV to a Delta V Plus isotope ratio mass spectrometer (Thermo Fisher Scientific) as described by de Carvalho et al. and reported in permil (‰ ± one standard deviation) relative to the international measurement standard Vienna Standard Mean Ocean Water (VSMOW), according to eq .

$$\delta {}^{18}\mathrm{O}=\frac{{({}^{18}\mathrm{O}/{}^{16}\mathrm{O})}_{\mathrm{sample}}}{{({}^{18}\mathrm{O}/{}^{16}\mathrm{O})}_{\mathrm{VSMOW}}}-1$$ 2

where (¹⁸O/¹⁶O)_sample and (¹⁸O/¹⁶O)_VSMOW are the ratios of heavy to light isotopes in O₂ in a sample and in VSMOW, respectively. In brief, seven 100 μL injections were made from each Exetainer headspace onto a 60 m Rt-Molsieve 5 Å PLOT column (Restek from BGB Analytik, 0.32 mm ID, 30 μm film thickness) maintained at 25 °C. Each GC/IRMS measurement sequence comprised 7–16 experimental samples, 10–15 procedural blanks, five water standards, and three air standards. Air standards were evenly distributed across the sequence and consisted of 150 μL of ambient air in 12 mL of He. Air standards were used to verify the absence of any instrument drift and to perform a one-point calibration of the δ¹⁸O values on the VSMOW scale. In recent work, we could demonstrate that the error introduced when using a one-point calibration with air is negligible for δ¹⁸O values of O₂. The δ¹⁸O value of O₂ in air was assumed to be 23.8 ‰. − Procedural blanks were used to correct the measured δ¹⁸O values for blank contributions. Water standards (Exetainers containing different amounts of air-saturated water) were used to correct δ¹⁸O values for instrument linearity (change in δ values with signal size).

2.5. Data Analysis

¹⁸ϵ values were determined from the slope obtained through the linear regression fit according to eq and are expressed in permil (‰). For experiments conducted in the 60 mL incubation chamber (Section ), the consumption rates of O₂ and CH₄ were corrected for background consumption by the MIMS probe. The O₂-to-CH₄ and O₂-to-NH₃ consumption stoichiometries were determined based on the measured concentrations of O₂ versus CH₄ or NH₃, respectively. All linear regressions were performed with R version 4.3.1 and the lm and confint functions. Unless noted otherwise, errors are reported as 95% confidence intervals.

3. Results and Discussion

3.1. Isotopic Fractionation Associated with O₂ Consumption by Methanotrophs

During O₂ consumption experiments with methanotrophs when using CH₄ as the substrate, the δ¹⁸O values of the unreacted dissolved O₂ increased with progressing O₂ consumption, consistent with the Rayleigh-type dynamics predicted by eq . As an example, Figure shows data from one experiment with M. lenta, including the measured O₂ and CH₄ concentrations (Figure A) and the changes in δ¹⁸O of the remaining O₂ over time (Figure C). An O₂:CH₄ consumption stoichiometry of 1.49 ± 0.01 was determined based on the slope of the linear regression between O₂ and CH₄ concentrations (see Figure B). The fact that this stoichiometry is between 1 and 2 indicates that both complete CH₄ oxidation to CO₂ (for energy generation) and incomplete CH₄ oxidation in combination with cell biomass formation (see Figure ) occurred during this experiment. In addition, Figure D shows the linearized isotope and concentration data, from which an ¹⁸ϵ_bulk value of −23 ± 3 ‰ was calculated. ¹⁸ϵ_bulk values, which represent bulk O isotopic fractionation of total O₂ consumption, were between −20 ± 1 ‰ and −23 ± 3 ‰ for all experiments with methanotrophs using CH₄ as the substrate (see upper-row panels in Figure and Table ). ¹⁸ϵ_resp values determined in experiments with methanotrophs using methanol as the substrate ranged between −19.0 ± 0.7 ‰ and −22 ± 3 ‰ (see bottom-row panels in Figure and Table ), in agreement with previous studies on terminal respiration in microbial cultures and pure enzymes. ,,,, The ¹⁸ϵ_resp value determined for M. lenta in the presence of the MIMS probe (−22 ± 3 ‰, n = 1; Figure , red circles) was identical within error to the ¹⁸ϵ_resp value determined in the syringe reactor (−20 ± 1 ‰, n = 2; Figure , orange and green circles), confirming that O₂ consumption by the MIMS probe (up 5% of the observed O₂ consumption rate) did not significantly affect the measured isotopic fractionation of O₂. For further considerations, average ¹⁸ϵ_resp values per species were used (see Table ).

2.

O₂ and CH₄ consumption and isotopic fractionation of O₂ during a single experiment with M. lenta. (A) Change in concentrations of dissolved O₂ (dashed line) and CH₄ (solid line) over time. The dip in CH₄ concentration between 5 and 10 min was due to a measurement artifact when stirring was paused to add bacterial biomass to start the reaction. (B) O₂ versus CH₄ concentrations from panel A. The black line indicates a linear regression fit with a slope of 1.49 ± 0.01 representing the O₂:CH₄ consumption stoichiometry. (C) δ¹⁸O values of unreacted O₂ measured in discrete samples over time. (D) Linearized and normalized data (δ¹⁸O vs [O₂]) from panels A and C, where [O₂]₀ and δ¹⁸O₀ represent the concentration and δ¹⁸O value, respectively, of O₂ at the beginning of the experiment. The solid line shows a linear regression fit according to eq , from which an ¹⁸ϵ value of −23 ± 3 ‰ was obtained.

3.

Log-normalized change in δ¹⁸O values vs log-normalized concentrations of O₂ ([O₂]) from experiments with different methanotrophs, where [O₂]₀ and δ¹⁸O₀ are the concentration and δ¹⁸O value of O₂ at the beginning of an experiment. The upper-row panels show results from experiments performed with CH₄ as the substrate. The bottom-row panels show results from experiments performed with methanol as the substrate. Replicate experiments are colored green, orange, and red. Experiments with CH₄ and one replicate experiment with methanol and M. lenta (red circles) were performed in the incubation chamber with the MIMs probe. All other experiments with methanol were performed in a syringe reactor (see Section for details). The slopes of the linear regressions are reported for each experiment in ‰ as ¹⁸ϵ_bulk (upper row) and ¹⁸ϵ_resp (lower row). All linear regressions were significant (p < 0.001) and had adjusted R ² values ≥ 0.99.

1. O₂:CH₄ Consumption Stoichiometries (Δ­[O₂]/Δ­[CH₄]), the Fraction of O₂ Consumed by MMO (f _MMO), ¹⁸ ϵ Values Determined with CH₄ (¹⁸ ϵ _bulk) and Methanol (¹⁸ ϵ _resp) as the Substrate, and the Calculated O Isotope Effects Imparted by Methane Monooxygenase (as ¹⁸ ϵ _MMO and Average Apparent ¹⁸O-KIE, see eq ) for Three Methanotrophic Bacteria.

species	Δ[O2]/Δ[CH4] (−)	f MMO (−)	18ϵbulk (‰)	18ϵresp (‰)	18ϵMMO (‰)	18O-KIE (−)
M. lenta	1.358 ± 0.009	0.736 ± 0.005	–21 ± 2	–20.1 ± 0.9	–22 ± 3	1.023 ± 0.004
	1.49 ± 0.01	0.670 ± 0.005	–23 ± 3		–24 ± 5
M. oryzae	1.266 ± 0.007	0.790 ± 0.004	–21.9 ± 0.3	–19.0 ± 0.7	–23 ± 1	1.023 ± 0.002
	1.323 ± 0.009	0.756 ± 0.005	–21 ± 1		–22 ± 2
M. silvestris	1.49 ± 0.02	0.673 ± 0.007	–21.2 ± 0.5	–19.9 ± 0.5	–22 ± 3	1.022 ± 0.002
	1.341 ± 0.008	0.746 ± 0.004	–20 ± 1		–21 ± 1

As stated above, ¹⁸ϵ_bulk values represent the bulk O isotopic fractionation of total O₂ consumption, which integrates the O isotopic fractionation associated with both MMO activity (¹⁸ϵ_MMO) and terminal respiration (¹⁸ϵ_resp). In M. lenta and M. oryzae, MMO activity is ascribed to pMMO, while in M. silvestris, it is ascribed to sMMO. Using the observed O₂:CH₄ consumption stoichiometry, we estimated the relative contribution of the two O₂ consuming processes (i.e., the fractional contribution of MMO, f _MMO, versus respiration, f _resp) to total O₂ consumption, which was then used to calculate ¹⁸ϵ_MMO from ¹⁸ϵ_bulk and ¹⁸ϵ_resp as outlined in eq .

$${}^{18}ϵ_{\mathrm{bulk}}=f_{\mathrm{MMO}}·{}^{18}ϵ_{\mathrm{MMO}}+f_{\mathrm{resp}}·{}^{18}ϵ_{\mathrm{resp}}=f_{\mathrm{MMO}}·{}^{18}ϵ_{\mathrm{MMO}}+(1-f_{\mathrm{MMO}})·{}^{18}ϵ_{\mathrm{resp}}$$ 3

We considered the plausible stoichiometries for CH₄ oxidation to range from 1CH₄:2O₂, during complete oxidation of CH₄ to CO₂, to 1CH₄:1O₂, for incomplete oxidation of CH₄ and cell biomass formation (see Figure ). The O₂:CH₄ consumption stoichiometries, determined in this study, ranged from 1.266 ± 0.007 to 1.49 ± 0.02 (see Table ), indicating a mix of complete and incomplete CH₄ oxidation, consistent with previous studies. , We can use these stoichiometries as a direct proxy for the fraction of O₂ consumption by MMO (see eq ) because based on the reactions shown in Figure , the amount of CH₄ consumed (Δ­[CH₄]) is equal to the amount of O₂ consumed by MMO (Δ­[O₂]_MMO) and the amount of total O₂ consumed (Δ­[O₂]) is equal to the sum of O₂ consumed by respiration (Δ­[O₂]_resp) and MMO.

$$\frac{\mathrm{\Delta }[{\mathrm{O}}_2]}{\mathrm{\Delta }[{\mathrm{CH}}_4]}=\frac{\mathrm{\Delta }{[{\mathrm{O}}_2]}_{\mathrm{MMO}}+\mathrm{\Delta }{[{\mathrm{O}}_2]}_{\mathrm{resp}}}{\mathrm{\Delta }{[{\mathrm{O}}_2]}_{\mathrm{MMO}}}=\frac{1}{f_{\mathrm{MMO}}}$$ 4

We can thus calculate O isotopic fractionation associated with O₂ consumption by MMO for the three methanotrophs with eqs and and the measured parameters O₂:CH₄ stoichiometry, ¹⁸ϵ_bulk and ¹⁸ϵ_resp. The resulting average ¹⁸ϵ_MMO values were −22 ± 2 ‰ for M. silvestris, −23 ± 2 ‰ for M. oryzae, and −23 ± 4 ‰ for M. lenta (see Table ). While not significantly different, there is a slight difference in isotopic fractionation between pMMO (¹⁸ϵ_MMO = −23 ± 3 ‰) and sMMO (¹⁸ϵ_MMO = −22 ± 2 ‰).

3.2. Isotopic Fractionation Associated with O₂ Consumption by N. inopinata

¹⁸ϵ_bulk values associated with O₂ consumption by the complete-ammonia oxidizer N. inopinata were determined in experiments analogously to those with methanotrophs, except that NH₃ consumption was determined from NH₄ ⁺ concentrations in discrete samples at different time points. The O₂:NH₃ consumption stoichiometries were 1.7 ± 0.5 and 1.5 ± 0.7, respectively, and the corresponding ¹⁸ϵ_bulk values were −19 ± 2 ‰ and −19.1 ± 0.7 ‰ (see Figure and Table ), similar to the ¹⁸ϵ_bulk for methanotrophic O₂ consumption. The fact that average NO₂ ^– (62 ± 54 μM) and NO _x production (56 ± 103 μM) were both comparable to average NH₃ consumption (96 ± 43 μM) together with an average O₂:NH₃ stoichiometry of 1.6 suggests that the majority of NH₃ was oxidized only to NO₂ ^– (partial nitrification) during the course of the experiments (see Figure ). Note that the poor precision of the N-species concentration measurements are due to the fact that initial concentrations of NH₄ ⁺, NO₂ ^–, and NO₃ ^– (approximately 0.6, 1.3, and 3.9 mM, respectively) were high relative to the changes measured in these concentrations.

4.

(A) O₂ vs (NH₄ ⁺ + NH₃) concentrations in two replicate experiments with N. inopinata. The NH₄ ⁺ concentrations were determined from duplicate measurements at different time points of the reaction (error bars represent standard deviations). The solid lines show a linear regression fit of the data in each experiment, with the slopes representing the respective O₂:NH₃ consumption stoichiometries. (B) Log-normalized change in δ¹⁸O values vs log-normalized concentration of O₂ for experiments with N. inopinata. The solid lines show a linear regression fit according to eq , with slopes indicating ¹⁸ϵ_bulk values.

2. O₂:NH₃ Consumption Stoichiometries (Δ­[O₂]/Δ­[NH₃]), the Fraction of O₂ Consumed by AMO (f _AMO), ¹⁸ ϵ Values Determined In Vivo (¹⁸ ϵ _bulk), and the Calculated O Isotope Effects Imparted by Ammonia Monooxygenase (as ¹⁸ ϵ _AMO and Average Apparent ¹⁸O-KIE, see eq ) for N. inopinata .

species	Δ[O2]/Δ[NH3] (−)	f AMO (−)	18ϵbulk (‰)	18ϵAMO (‰)	18O-KIE (−)
N. inopinata	1.5 ± 0.7	0.7 ± 0.3	–19.1 ± 0.7	–19 ± 14	1.019 ± 0.014
	1.7 ± 0.5	0.6 ± 0.2	–19 ± 2	–18 ± 12

Similar to O₂ consumption by methanotrophs, the average ¹⁸ϵ_bulk value for O₂ consumption by N. inopinata reflects the combined O isotopic fractionation imparted by both AMO and respiration. As with methanotrophs, where experiments with methanol were used to isolate the respiratory ¹⁸ϵ_resp value by excluding MMO-driven O₂ consumption, a similar approach could theoretically be applied to N. inopinata by using hydroxylamine as an intermediate substrate to bypass AMO activity. However, due to the reactivity of hydroxylamine, an isolated respiration ¹⁸ϵ_resp value for N. inopinata could not be determined. Nevertheless, ¹⁸ϵ_resp for N. inopinata can be reasonably approximated based on known properties of its respiratory machinery. Notably, N. inopinata lacks the canonical terminal copper-heme-dependent cytochrome c oxidase, and instead expresses a heme-dependent cytochrome bd-like oxidase. , Stolper et al. demonstrated that the ¹⁸ϵ_resp values obtained for Escherichia coli mutant strains containing either only cytochrome c oxidase, only cytochrome bd-I oxidase, or only cytochrome bd-II oxidase, were indistinguishable within error, ranging from −14.9 to −15.5 ‰. Although these ¹⁸ϵ_resp values are lower than those typically reported for isolated cytochrome c oxidase, or microbial respiration more broadly, ,,, they nevertheless suggest that the cytochrome bd-like oxidase in N. inopinata may impart ¹⁸ϵ_resp values similar to those of more common respiratory terminal oxidases found in most bacteria (i.e., approximately −20 ‰).

Similar to the approach used for the methanotrophs in this study, the O₂:NH₃ consumption stoichiometry was employed to calculate the relative contributions of AMO activity versus respiration to total O₂ consumption and, in turn, to calculate ¹⁸ϵ_AMO from the observed ¹⁸ϵ_bulk values and an assumed ¹⁸ϵ_resp of −20 ‰ with eq .

$${}^{18}ϵ_{\mathrm{bulk}}=f_{\mathrm{AMO}}·{}^{18}ϵ_{\mathrm{AMO}}+f_{\mathrm{resp}}·{}^{18}ϵ_{\mathrm{resp}}=\frac{\mathrm{\Delta }[{\mathrm{NH}}_3]}{\mathrm{\Delta }[{\mathrm{O}}_2]}·{}^{18}ϵ_{\mathrm{AMO}}+\left(1-\frac{\mathrm{\Delta }[{\mathrm{NH}}_3]}{\mathrm{\Delta }[{\mathrm{O}}_2]}\right)·{}^{18}ϵ_{\mathrm{resp}}$$ 5

We note that the calculated average ¹⁸ϵ_AMO value of −19 ± 13 ‰ (Table ) is contingent on the assumed ¹⁸ϵ_resp, which was not independently determined (unlike in our methanotroph experiments) but instead derived from the existing literature. Nevertheless, varying the assumed ¹⁸ϵ_resp value by ±5 ‰ still yields ¹⁸ϵ_AMO estimates (−15 to −22 ‰) that fall within the typical range reported for terminal aerobic respiration (¹⁸ϵ_resp). The high uncertainty of the ¹⁸ϵ_AMO estimate is due to the elevated 95% confidence intervals of the O₂:NH₃ stoichiometries (Table ), and thus it is a consequence of the errors associated with NH₃ concentration measurements.

3.3. Kinetic Isotope Effects and O₂ Activation Mechanisms of Enzymes Involved in Methane and Ammonia Oxidation

The in vivo ¹⁸ϵ_MMO, ¹⁸ϵ_AMO, and ¹⁸ϵ_resp values determined in this study can be converted into average apparent ¹⁸O-KIEs of 1.023 ± 0.003, 1.022 ± 0.002, 1.019 ± 0.014, and 1.020 ± 0.001 (Tables and ) for the reduction of O₂ by pMMO, sMMO, AMO, and terminal oxidase, respectively. ¹⁸O-KIEs can be used as mechanistic probes to assess the rate-limiting steps in O₂-consuming enzymatic reactions and thus shed light on the catalytic cycle of such enzymes. This is of particular interest for pMMO and AMO, whose catalytic cycles have not been well studied. Because pMMO, sMMO, AMO, and terminal oxidase represent three different classes of O₂-consuming enzymes with distinct active-site structures, it is surprising to see their respective ¹⁸O-KIEs fall within error of each other. To better link these values with the potentially rate-limiting reaction steps, a closer look at these enzyme’s catalytic cycles is required.

Cytochrome c oxidase, the most prevalent terminal oxidase, is a well-studied heme-copper-dependent enzyme, in which the heme a3 subunit initially binds O₂ reversibly, forming an iron-superoxo intermediate (see Figure a). In the subsequent step, which is suggested to be rate-limiting, a hydrogen atom abstraction from an adjacent tyrosine residue by this iron-superoxo species, concomitant with O–O bond cleavage, leads to the formation of an iron-oxo and a copper-hydroxo species. The measured ¹⁸O-KIE of 1.020 ± 0.001 for terminal oxidases is in agreement with such a rate-limiting hydrogen atom abstraction step by a metal-superoxo intermediate as well as with a rate-limiting iron-oxo species formation. ,

5.

Key steps in the O₂ activation mechanisms of (a) cytochrome c oxidase (adapted from ref , copyright 2015 American Chemical Society), (b) soluble methane monooxygenase (sMMO, adapted from ref , copyright 2024 American Chemical Society, and with permission under a Creative Commons CC-BY 4.0 license from ref , copyright 2001 ASBMB), and (c) particulate methane monooxygenase (pMMO, adapted from ref , copyright 2024 American Chemical Society and with permission under a Creative Commons CC-BY 4.0 license from ref , copyright 2024 P.E.M. Siegbahn). Potentially rate-limiting steps according to measured ¹⁸O kinetic isotope effects are marked with an asterisk (*).

The ¹⁸O-KIE determined for the diiron enzyme sMMO in this study with M. silvestris (1.022 ± 0.002) is larger than previously reported values for the isolated enzyme from M. capsulatus (1.015–1.017). While natural variability in isotope effects for a given enzyme is not uncommon, it is also possible that the isolated enzyme exhibits a different ¹⁸O-KIE outside its cellular environment, compared to in its in vivo state. In sMMO, it is proposed that O₂ activation is initiated with a reversible O₂ binding to one of the active-site iron atoms, followed by the formation of a diiron-bridged peroxo species (3 in Figure b). One of the O atoms is subsequently released from the active site as H₂O, while the other O atom may form a high-valent diiron-bridged oxo species (4 in Figure b). The final steps in the catalytic cycle of sMMO involve hydrogen atom abstraction from CH₄ by 4, rebound of the methyl radical, release of methanol, and reduction of the active-site iron atoms (see Figure b). While Stahl et al. suggested the formation of a diiron-bridged peroxo species (3) to be the rate-limiting step based on their ¹⁸O-KIEs of 1.015–1.017, the value measured in this study (1.022 ± 0.002) is in better agreement with a rate-limiting formation of an iron-bound oxo species (4). A common rate-limiting iron-oxo species formation would also be compatible with the identical ¹⁸O-KIEs of sMMO and the terminal oxidase measured in this study. Further studies should focus to elucidate whether the difference in isotope effects of sMMO is due to experimental variables (in vitro vs in vivo) or the different bacterial species (M. silvestris vs M. capsulatus).

The ¹⁸O-KIEs determined for pMMO (1.023 ± 0.003) and AMO (1.019 ± 0.014) in this study align with the upper range of values previously reported for copper-dependent O₂-consuming enzymes. Generally, O₂ activation by copper-dependent enzymes with apparent ¹⁸O-KIEs between 1.017 and 1.022 is associated with a rate-limiting hydrogen atom abstraction, leading to the formation of a copper-hydroperoxo species. ,, Given the similarities in substrate specificity, subunit compositions, and DNA sequences of pMMO and AMO, a common O₂ activation mechanism is plausible. A complete catalytic cycle has not been resolved for either of the two enzymes, but experimental and computational evidence from studies with pMMO can be used, together with the measured ¹⁸O-KIEs, to suggest two possible O₂ activation mechanism (see Figure C). While there is still debate regarding the exact location of the active site in pMMO, current evidence suggests a mononuclear copper center to be the active site (7 in Figure c). In analogy to O₂ activation mechanisms of other mononuclear copper enzymes with similar ¹⁸O-KIEs, the first step of the catalytic cycle likely involves reversible O₂ binding and formation of a copper-superoxo intermediate (8 in Figure c). This could be followed by a rate-limiting hydrogen atom abstraction by 8 directly at the substrate (CH₄) and thus result in a Cu-hydroperoxo intermediate and a methyl radical (9 in Figure c). The catalytic cycle can then be completed through a radical rebound step and the release of the products methanol and H₂O. Computational studies, however, suggest that direct hydrogen atom abstraction from CH₄, achieved enzymatically only by MMO and AMO, requires a more reactive intermediate, such as Cu­(II)-(hydr)­oxo or Cu­(III)-oxo species (12a–12c in Figure c). These species could form through a rate-limiting O–O bond cleavage step from a Cu-hydroperoxo intermediate (11 in Figure c). While a Cu­(III)-oxo species has not yet been detected in enzymes, the analogous intermediate in iron-dependent enzymes (an Fe­(IV)-oxo species formed in a rate-limiting O–O bond cleavage step) is associated with a ¹⁸O-KIE of 1.022, similar to pMMO and AMO here. Similar to the mechanism mentioned above, this second possible catalytic cycle of pMMO (and AMO) is concluded with hydrogen atom abstraction from CH₄, radical rebound, and release of the products methanol and H₂O (see Figure c). The key difference between the two proposed mechanisms (8 → 9 → 10 → 7 vs 8 → 11 → 12 → 13 → 14 → 7 in Figure ) is the timing of hydrogen atom abstraction from CH₄ relative to the first irreversible step of O₂ activation. In the former pathway, hydrogen atom abstraction occurs during the first irreversible step, whereas in the latter, it occurs after the first irreversible step of the O₂ activation. The nature of the rate-limiting step of pMMO could potentially be determined by comparing ¹⁸O-KIEs using protiated and deuterated substrates (CH₄ or CD₄). Deuteration of the substrate would alter the ¹⁸O-KIE onyl if hydrogen atom abstraction occurs during the first irreversible step of O₂ activation. However, this analysis was beyond the scope of the current study. Nonetheless, the ¹⁸O-KIEs determined here for the first time for O₂ activation by pMMO and AMO provide important evidence to better refine the catalytic cycle of these environmentally relevant enzymes.

4. Conclusions

The in vivo ¹⁸ϵ values for pMMO, sMMO, and AMO determined in this study ranged from −18 ± 12 ‰ to −24 ± 5 ‰. These values are not significantly different from the in vivo ¹⁸ϵ_resp values we report here for methanotrophs (−19.0 ± 0.7 ‰ to −22 ± 3 ‰) or from typical aquatic respiration ¹⁸ϵ values reported in previous studies (−18 to −24 ‰). ,,, This apparent similarity implies that O₂ consumption by the widespread enzymes pMMO or AMO cannot account for the discrepancy observed between in situ and laboratory-derived ¹⁸ϵ values for O₂ consumption in aquatic environments. Consequently, the ¹⁸ϵ values estimated for aerobic CH₄ or NH₃ oxidation indicate that these processes cannot be uniquely identified or quantified solely on the basis of O₂ isotope analysis in the environment. Nevertheless, it would be interesting to conduct further experiments, particularly with other ammonia-oxidizers or mixed cultures, to test the robustness of the ¹⁸ϵ values determined in this study. We consider it unlikely that other biological (i.e., beyond respiration and CH₄ or NH₃ oxidation) or abiotic O₂-consuming processes, such as H₂S or Fe­(II) oxidation, ,, which are characterized by low ¹⁸ϵ values, are pervasive enough in natural aquatic environments to significantly mask the isotopic imprint of heterotrophic O₂ respiration at the ecosystem scale. And while photochemical reactions can have a significant impact on the isotopic fractionation of O₂ consumption in surface waters, , these processes unlikely mask the isotopic imprint of heterotrophic O₂ respiration in lower parts of the water column, where in situ isotopic fractionation of O₂ consumption is typically determined. Hence, the observed discrepancy between “community” in situ and laboratory ¹⁸ϵ values for O₂ consumption is most plausibly explained by diffusion limitation at different spatial scales in natural environments or O₂ consumption via biological ROS formation, as previously proposed.

The data underlying this study are openly available at 10.5281/zenodo.17415999.

The authors declare no competing financial interest.

Published as part of ACS Environmental Au special issue “2025 Rising Stars in Environmental Research”.