Nitrous oxide respiration in acidophilic methanotrophs

Abstract

Aerobic methanotrophic bacteria are considered strict aerobes but are often highly abundant in hypoxic and even anoxic environments. Despite possessing denitrification genes, it remains to be verified whether denitrification contributes to their growth. Here, we show that acidophilic methanotrophs can respire nitrous oxide (N₂O) and grow anaerobically on diverse non-methane substrates, including methanol, C-C substrates, and hydrogen. We study two strains that possess N₂O reductase genes: Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6. We show that N₂O respiration supports growth of Methylacidiphilum caldifontis at an extremely acidic pH of 2.0, exceeding the known physiological pH limits for microbial N₂O consumption. Methylocella tundrae simultaneously consumes N₂O and CH₄ in suboxic conditions, indicating robustness of its N₂O reductase activity in the presence of O₂. Furthermore, in O₂-limiting conditions, the amount of CH₄ oxidized per O₂ reduced increases when N₂O is added, indicating that Methylocella tundrae can direct more O₂ towards methane monooxygenase. Thus, our results demonstrate that some methanotrophs can respire N₂O independently or simultaneously with O₂, which may facilitate their growth and survival in dynamic environments. Such metabolic capability enables these bacteria to simultaneously reduce the release of the key greenhouse gases CO₂, CH_4, and N₂O.

Introduction

Anthropogenic emissions of greenhouse gases (GHGs)—primarily carbon dioxide (CO₂), methane (CH₄), and nitrous oxide (N₂O)—are responsible for a historically rapid increase in Earth’s average annual temperature of more than 0.2 °C per decade^1,2. In addition to achieving net-zero CO₂ emissions by 2050, significant reductions in the emissions of other GHGs including CH₄ and N₂O are now critically needed. Compared to CO₂, the warming effect of CH₄ is around 28 to 34 times greater^3,4. However, its much shorter mean lifetime of approximately 12–13 years⁵ provides an additional opportunity to mitigate future climate change. Like CO₂, N₂O—the third most important GHG—has a long half-life (roughly 120 years) in the atmosphere⁶, and its warming potential is about 300 times greater than CO₂ over a 100-year time scale¹. In addition, N₂O is a major cause of ozone depletion in the stratosphere^7,8.

Although human activities are by far the most important reason for the unprecedented rise in atmospheric GHGs⁹, microbial activities also play a direct role in this rise^10,11. GHG net accumulation is regulated by the biogeochemical source-sink dynamics of GHGs exchanged between terrestrial, marine, and atmospheric reservoirs⁹. GHG production and consumption in both natural and anthropogenic ecosystems are driven primarily by microbes^10,12. Methane fluxes in natural environments are controlled by activities of methane-producing (methanogenic) and methane-consuming (methanotrophic) microorganisms. It is estimated that 69% of the atmospheric CH₄ budget originates from microbial activities (methanogenesis) while about 50−90% of the produced CH₄ is oxidized by methanotrophs before reaching the atmosphere^13,14.

Microbes can oxidize methane under aerobic and anaerobic conditions. Aerobic methanotrophs oxidize methane to methanol by employing either particulate methane monooxygenases (pMMO) or soluble methane monooxygenases (sMMO)¹⁵. There are two ways in which aerobic methanotrophs use molecular oxygen (O₂): as the terminal electron acceptor of aerobic respiration and for methane activation via the methane monooxygenase¹⁵. Under strictly anoxic conditions, anaerobic methanotrophic microorganisms mitigate CH₄ emissions by oxidizing methane with alternative terminal electron acceptors including NO₃⁻, Fe³⁺, Mn⁴⁺, SO₄^2-, and humic acid using reverse methanogenesis pathways^16,17,18,19. Furthermore, intra-aerobic metabolism in the nitrite-dependent anaerobic methane-oxidizing bacterium ‘Candidatus Methylomirabilis oxyfera’ using pMMO was reported²⁰.

Interestingly, the genomes of some aerobic methanotrophs encode denitrification enzymes including nitrate (NO₃⁻), nitrite (NO₂⁻), nitric oxide (NO), and N₂O reductases^21,22,23,24. Surprisingly, however, none of the methanotroph genomes or MAGs known to date encode a complete set of denitrification genes (Supplementary Dataset 1). Kits and colleagues^21,22 demonstrated that some aerobic methanotrophs can couple NO₃⁻ and NO₂⁻ reduction to the oxidation of methane and other electron donors, including methanol, formaldehyde, formate, ethane, ethanol, and ammonia in suboxic conditions. However, whether these aerobic methanotrophs are capable of anaerobic growth with NO₃⁻ and NO₂⁻ as terminal electron acceptors remain to be seen.

More than two-thirds of N₂O emissions arise from bacterial and fungal denitrification and nitrification processes in soils^25,26. N₂O emissions are a major concern in acidic environments due to the high production of N₂O via abiotic reactions and the inhibition of biological N₂O reduction^27,28. Although multiple sources of N₂O exist²⁵, there is only one known sink for N₂O in the biosphere—the microbial reduction of N₂O to N₂, catalyzed by a copper-dependent enzyme, N₂O reductase (N₂OR) encoded by nosZ²⁹. The NosZ enzymes found in prokaryotes are phylogenetically classified into two clades: the canonical NosZ (clade I NosZ), found mostly in denitrifiers³⁰, and the recently described cNosZ (clade II NosZ)³¹, which has an additional c-type heme domain at the C terminus, found commonly in non-denitrifiers^31,32. Thus, bacteria and archaea harboring the nosZ-type genes, in particular those classified as incomplete- or non-denitrifiers because they do not encode the full denitrification pathway, are receiving increasing attention in the search for technologies to combat N₂O emissions³². Previous studies have reported the presence of the nosZ gene in the aerobic methanotrophs, Methylocystis sp. SC2 (ref. ²³) and Methylocella tundrae²⁴. Further genomic analysis from this study suggests that this enzyme is present in some other aerobic methanotrophs, too (Supplementary Dataset 1). Pure culture studies have unequivocally shown that denitrifiers can grow by respiring N₂O (refs. ^33,34). Moreover, an electron sink/spill role for N₂OR has been proposed for Gemmatimonas aurantiaca T-27 (ref. ³⁵) without biomass production (i.e., growth). Despite the presence of N₂OR in Methylocystis sp. SC2, its ability to grow in anoxia under N₂O-reducing conditions is unverified³⁶. Thus, the ability to grow by converting N₂O to N₂ has not yet been reported for any of the known aerobic or anaerobic methanotrophs, even with non-methane substrates such as methanol.

Methanotrophs using MMO enzymes are considered to be obligate aerobes. Paradoxically, however, they are often detected at high relative abundance in extremely hypoxic and even anoxic zones of peat bogs, wetlands, rice paddies, forest soils, and geothermal habitats^37,38. It is therefore critical to investigate the ability of aerobic methanotrophs to use N₂O as the sole terminal electron acceptor for energy conservation and biomass production, a metabolic trait that could allow them to thrive in these anoxic ecosystems. Here, we used a multi-faceted approach to investigate the role of N₂O respiration in defining the physiology and ecology of selected aerobic methanotrophs. Growth experiments demonstrated that the presence of N₂OR in an acidophilic proteobacterial methanotroph, Methylocella tundrae T4, and an extremely acidophilic verrucomicrobial methanotroph, Methylacidiphilum caldifontis IT6, enables these organisms to respire N₂O and to produce biomass while oxidizing a wide variety of electron donors, including methanol, acetol, pyruvate, and hydrogen. In contrast to N₂O, respiration of NO₃⁻ and NO₂⁻ did not support anaerobic growth of these methanotrophs on C1 substrates. We also demonstrate that Methylocella tundrae T4 can reduce both O₂ and N₂O simultaneously, allowing it to oxidize more CH₄ and generate more biomass under O₂-limiting conditions. Our findings significantly expand the potential ecological niche of aerobic methanotrophs and reveal that some methanotrophic microbial strains could be used to mitigate multiple GHG emissions.

Results and discussion

N₂OR-encoding genes in aerobic methanotrophs

To identify methanotrophs capable of using N₂O as an alternative electron acceptor, publicly available genomes and metagenome-assembled genomes (MAGs) of methanotrophs were screened for nosZ genes. We found genes encoding N₂OR in genomes and MAGs of methanotrophs from three bacterial phyla: Pseudomonadota, Verrucomicrobiota, and Gemmatimonadota (Supplementary Dataset 1). They were confined to the alphaproteobacterial methanotrophs and absent in gammaproteobacterial methanotrophs in the case of the phylum Pseudomonadota and represented by only two genera, Methylocella and Methylocystis, which also accounted overall for the majority of the methanotroph genomes encoding nosZ. Similarly, nosZ genes were exclusively found in one representative genome in each of the phyla Verrucomicrobiota (represented by the genus Methylacidiphilum) and Gemmatimonadota (represented by the candidate genus ‘Methylotropicum’), respectively. Phylogenetic analysis of predicted NosZ protein sequences revealed that those found in Methylocella and Methylocystis are from the clade I NosZ lineage, while those found in Methylacidiphilum and ‘Ca. Methylotropicum’ are from the clade II NosZ lineage (Fig. 1, Supplementary Fig. 1).

Fig. 1: Maximum-likelihood phylogenetic tree of derived NosZ proteins, with nos operon arrangements in methanotrophic and non-methanotrophic bacterial strains.

The phylogenetic tree was constructed with IQ-TREE (IQ-TREE options: -B 1000 -m LG + F + R5) using aligned NosZ (details in Materials and Methods) and rooted at the mid-point. Bootstrap values ≥ 70% based on 1000 replications are indicated. The scale bar represents a 0.5 change per amino acid position. Organization of the nos operon in methanotrophic strains (labeled in blue text) and closely related non-methanotrophic bacteria are shown. The genes, represented by arrows, are drawn to scale. Homologs are depicted in identical colors. The NosZ amino acid sequences and gene arrangement information were retrieved using the following genome accessions: GCF_017310505.1, Methylacidiphilum caldifontis IT6; GCF_000010785.1, Hydrogenobacter thermophilus TK-6; GCF_011006175.1, Hydrogenobacter sp. T-8; GCF_900215655.1, Hydrogenobacter hydrogenophilus DSM 2913; GCF_000619805.1, Sulfurihydrogenibium subterraneum DSM 15120; GCF_000021565.1, Persephonella marina EX-H1; GCF_000022145.1, Anaeromyxobacter dehalogenans 2CP1; GCF_000013385.1, Anaeromyxobacter dehalogenans 2CP-C; GCF_003054705.1, Opitutus sp. ER46; GCF_000019965.1, Opitutus terrae PB90-1; GCF_901905185.1, Methylocella tundrae PC4; GCA_901905175.1, Methylocella tundrae PC1; CP139089.1, Methylocella tundrae T4; FO000002.1, Methylocystis sp. SC2; GCF_000025965.1, Aromatoleum aromaticum EbN1; GCF_022760775.1, ‘Candidatus Rhodoblastus alkanivorans’ PC3; GCF_000143145.1, Hyphomicrobium denitrificans ATCC 51888; GCF_000344805.1, Bradyrhizobium oligotrophicum S58; GCF_027923385.1, Methylocystis echinoides LMG27198; GCA_003963405.1, Methylocystis sp. AWTPI-1. * indicates that the nosZ genes are truncated due to genome fragmentation. Source Data contains genome annotation information for Methylocella tundrae T4, Methylacidiphilum caldifontis IT6, Methylocystis spp. (strains IM2, IM3, and IM4), and ‘Ca. Methylotropicum kingii’.

Three Methylocella tundrae strains: T4 (re-sequenced genome), PC1 (ref. ³⁹), and PC4 (ref. ³⁹), have nos gene clusters (NGC) (Fig. 1). These are incorporated into nosRZDFYLX operons in strains PC4 and T4 and a nosZDFYLX operon in strain PC1 (Fig. 1). Strain PC1 has truncated nosZ and missing nosR genes. This is most likely due to its genome being highly fragmented into several small contigs containing missing and truncated genes. The NGC composition and operon arrangement, nosRZDFYLX, were largely similar in the genomes of the six N₂OR-containing Methylocystis species (Fig. 1), including Methylocystis sp. SC2 (ref. ²³), Methylocystis echinoides LMG27198, three in-house Methylocystis echinoides-like isolates (strains IM2, IM3, and IM4), and a metagenome-assembled genome (MAG) of a Methylocystis sp. AWTPI-1 recovered from a water treatment facility⁴⁰. A notable feature in their NGC organization was the absence of the gene encoding the membrane-anchored copper chaperon, NosL, which is primarily involved in Cu(I) delivery to apo-NosZ⁴¹. Methanotrophs with pMMO usually possess multiple copper chaperones⁴² that may complement NosL, making it non-essential for NosZ maturation. Altogether, the NGC in these alphaproteobacterial methanotrophs has a similar organization to those of clade I N₂O-reducers (Fig. 1). BLAST results further revealed that the individual nos genes in the Methylocella and Methylocystis strains shared a high degree of similarity to each other and other non-methanotrophic Alphaproteobacteria (Supplementary Dataset 2). Also, their NosZ proteins share high homology with proteins annotated as twin-arginine translocation (Tat)-dependent N₂OR (35–89%) and also possess the Tat signal peptide with a characteristic SRRx[F | L] motif⁴³ found in clade I NosZ³².

The NGC in the genome of Methylacidiphilum caldifontis IT6 (ref. ⁴⁴), comprises a nosCZBLDFYC operon (Fig. 1) but lacks the typical nosX and nosR found in clade I N₂O-reducers^31,32, involved in NosR maturation⁴⁵ and electron transfer to NosZ⁴⁶, respectively. Notably, the NGC (IT6_00904–11) was found within the cluster of genes (IT6_00903, IT6_00912–7) encoding alternative complex III (refer to Source Data for annotation information). Both the aa₃-type and cbb₃-type cytochrome c oxidase-encoding genes are also located next to these genes. Genes encoding two c-type cytochromes (nosC) within the nos operon (Fig. 1) could serve electron transport functions⁴⁷. Interestingly, BLAST and synteny analyses of the NGC show that the individual genes are most closely related to genes found in genomes of extremely thermophilic Hydrogenobacter species of the phylum Aquificota (amino acid identities of 72.41–91.96%) (Supplementary Dataset 2) with a similar genetic organization (Fig. 1). Strain IT6 NosZ shares high similarities to proteins annotated as Sec-dependent N₂OR (35–89%) with an N-terminal Sec-type signal peptide found in clade II NosZ^31,32; the highest identities (79–89%) were with NosZ proteins from other Hydrogenobacter species. Hydrogenobacter thermophilus TK-6, a hydrogen-oxidizing bacterium, can completely denitrify NO₃^- to N₂ gas⁴⁸, indicating the presence of a functional N₂OR. As a result, Methylacidiphilum caldifontis IT6 may also have a functional N₂OR due to the high similarity of its NGC to those of Hydrogenobacter species. Although genomes of other Methylacidiphilum species, including Methylacidiphilum fumariolicum, lacked the gene encoding the N₂OR catalytic subunit, NosZ, some genes encoding Nos accessory proteins were found (Supplementary Dataset 2). Interestingly, the N₂OR genes for Methylacidiphilum caldifontis IT6 were found in a genomic island (Supplementary Dataset 3) and were most likely acquired through horizontal gene transfer, which is consistent with its NosZ phylogeny (Fig. 1, Supplementary Fig. 1). This is not surprising since many key metabolic genes in verrucomicrobial methanotrophs, including those encoding the MMO, are believed to have been acquired through horizontal gene transfer⁴⁹. As a result, Methylacidiphilum fumariolicum strains might have acquired the NGC before losing the key functional genes but retaining some of the accessory genes. Finally, we found a nosZBDF operon in the MAG of the uncultured methanotrophic bacterium ‘Ca. Methylotropicum kingii’⁵⁰ that resembles clade II NGC, with a truncated nosZ and multiple missing genes like nosY, nosL, and nosC (Fig. 1). These are also likely the result of multiple MAG fragmentations. Multiple sequence alignments of the predicted NosZ proteins of methanotrophs and other microorganisms (clade I and II) were constructed. All the expected metal-binding residues present in N₂OR were mostly conserved in the methanotroph NosZ sequences (Supplementary Fig. 2, Supplementary Note 1).

N₂O-dependent anaerobic growth of methanotrophs

The presence of genes predicted to encode N₂OR in the genomes of Methylocella tundrae strains, Methylacidiphilum caldifontis IT6, and Methylocystis strains (SC2, IM2, IM3, and IM4) (Supplementary Datasets 1, 2) led us to investigate whether this enzyme can support the anaerobic growth of these aerobic methanotrophs when N₂O is supplied as their sole electron acceptor. Physiological studies on N₂O reduction by methanotrophs focused on Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6 since preliminary experiments showed that the N₂OR-containing Methylocystis strains, including Methylocystis sp. SC2 and the in-house Methylocystis strains (IM2, IM3, and IM4) failed to reduce N₂O under various anoxic growth conditions. We set up anoxic batch cultures of Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6 using methanol as a sole electron donor with or without N₂O as the sole electron acceptor. For these incubations, 2 mM ammonium (NH₄⁺) was used as the nitrogen source instead of NO₃⁻ to avoid the involvement of dissimilatory nitrate reduction particularly in the Methylocella strains with nitrate-reducing potential. As a negative control, closely related methanotrophs lacking a predicted N₂OR (Methylocella silvestris BL2 and Methylacidiphilum infernorum IT5, respectively) were included in the study design. The growth experiments were conducted in LSM medium at pH 5.5 for Methylocella species (strains T4 and BL2) and at pH 2.0 for Methylacidiphilum species (strains IT5 and IT6). As expected, in control incubations provided with O₂ as the terminal electron acceptor, all four strains grew on CH₃OH (Fig. 2A, D, G, J). In these controls, the maximum specific growth rates (µ_max) of the Methylocella strains (strain T4: µ_max = 2.83 ± 0.03 d⁻¹; strain BL2: µ_max = 1.79 ± 0.05 d⁻¹) were higher than those of the Methylacidiphilum strains (strain IT6: µ_max = 1.57 ± 0.04 d⁻¹; strain IT5: µ_max = 1.49 ± 0.01 d⁻¹).

Fig. 2: Aerobic and anaerobic growth of N₂OR-containing and N₂OR-lacking Methylocella and Methylacidiphilum strains on methanol.

Methylocella tundrae T4, Methylocella silvestris BL2, Methylacidiphilum caldifontis IT6, and Methylacidiphilum infernorum IT5 cells were grown in LSM medium supplemented with 30 mM methanol as the electron donor and NH₄⁺ as the N-source. Aerobic growth of the 4 strains with O₂ (A, D, G, J), anaerobic growth with N₂O (B, E, H, K), and anaerobic growth without N₂O (C, F, I, L) as the sole terminal electron acceptor were determined by optical density measurements at 600 nm, followed by measurements of O₂ and N₂O consumption in the headspaces of the culture bottles. Note that the trace O₂ present at the start of the incubation in the anaerobic cultures without N₂O did not contribute to obvious growth (C, F, I, L). All experiments were performed in triplicates. Data are presented as mean ± 1 standard deviation (SD), and the error bars are hidden when they are smaller than the width of the symbols. Source data are provided as Source Data file.

Under N₂O-containing anoxic conditions, Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6 reduced N₂O and grew on methanol (Fig. 2B, H). When N₂O was depleted, the growth of strains T4 and IT6 ceased. To verify that OD₆₀₀ measurements indicated anaerobic cell growth rather than an artifact such as exopolysaccharide production, we demonstrated that cell counts and counts of 16S rRNA genes increased in parallel with OD₆₀₀ during anaerobic growth (Supplementary Fig. 3). No growth was observed in N₂O-free anoxic conditions used as negative controls (Fig. 2C, I). These results demonstrate that the anaerobic growth of these methanotrophs was dependent on N₂O as the sole electron acceptor. The observed N₂O reduction was catalyzed by a functional respiratory N₂OR, as the N₂OR-lacking relatives (Methylacidiphilum infernorum IT5 and Methylocella silvestris BL2) used as negative controls did not grow or reduce N₂O under anoxic conditions (Fig. 2E, F, K, L). In addition, other known electron donors of Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6, which support their aerobic growth^44,51,52, also supported their growth under anoxic N₂O-reducing conditions (Supplementary Dataset 4). Methylocella tundrae T4 grew on pyruvate and acetol, while Methylacidiphilum caldifontis IT6 grew on acetol under anoxic N₂O-reducing conditions. Further, molecular hydrogen supported the chemolithoautotrophic growth of Methylacidiphilum caldifontis IT6 as the sole electron donor under anoxic N₂O-reducing conditions (Supplementary Fig. 4). The transcriptomic analysis (see below) suggests that the group 1d [NiFe] hydrogenase encoded in the genome of Methylacidiphilum caldifontis IT6 could be involved in chemolithoautotrophic growth under anoxic N₂O respiring conditions.

Methylocella tundrae T4 exhibited a higher growth rate (µ_max = 0.47 ± 0.02 d⁻¹) than Methylacidiphilum caldifontis IT6 (µ_max = 0.18 ± 0.01 d⁻¹) on methanol and N₂O. However, these values are approximately 6 and 9 times, respectively, lower than the growth rates measured for both strains under O₂−respiring conditions. Biomass yields Y_x/m (g DW⋅mol⁻¹ N₂O or O₂ reduced) for the methanol-oxidizing cultures of strains T4 and IT6 reducing N₂O as the sole electron acceptor were also lower than for cells reducing O₂ as the sole electron acceptor. The biomass yield of Methylocella tundrae T4 cells grown anaerobically on N₂O (4.64 ± 0.04 g DW⋅mol⁻¹ N₂O reduced) was approximately 45% of that of aerobically grown cells (10.41 ± 0.04 g DW⋅mol⁻¹ O₂ reduced). Similarly, Methylacidiphilum caldifontis IT6 had a biomass yield when grown anoxically on N₂O (2.36 ± 0.04 g DW⋅mol⁻¹ N₂O reduced), which was only about 38% of that achieved by aerobically grown cells (6.27 ± 0.14 g DW⋅mol⁻¹ O₂ reduced). This improved molar yield on O₂ is expected despite the higher reduction potential of N₂O (see Eqs. [1] and [2]), since O₂ respiration accepts twice as many electrons as N₂O respiration (Eq. 1 and 2)⁵³. In addition, the aerobic terminal oxidases of both strains are proton pumps and conserve energy (Supplementary Datasets 5, 6)^54,55, whereas N₂OR does neither⁵⁶. To our knowledge, our results constitute the first report of N₂O reduction coupled with anaerobic growth in any methanotroph.

$${{{{{{\rm{N}}}}}}}_{2}{{{{{\rm{O}}}}}}+{2{{{{{\rm{H}}}}}}}^{+}+{2{{{{{\rm{e}}}}}}}^{-}\to {{{{{{\rm{N}}}}}}}_{2}+{{{{{{\rm{H}}}}}}}_{2}{{{{{\rm{O}}}}}}\,{E}_{0{\prime} }({{{{{\rm{pH}}}}}}7.0)=+ 1.36{{{{{\rm{V}}}}}}$$

(1)

$${{{{{{\rm{O}}}}}}}_{2}+{4{{{{{\rm{H}}}}}}}^{+}+{4{{{{{\rm{e}}}}}}}^{-}\to 2{{{{{{\rm{H}}}}}}}_{2}{{{{{\rm{O}}}}}}+{{{{{{\rm{H}}}}}}}_{2}{{{{{\rm{O}}}}}}\,{E}_{{0}^{{\prime} }}({{{{{\rm{pH}}}}}}7.0)=+ 0.82{{{{{\rm{V}}}}}}$$

(2)

It is well known that N₂O reduction is generally inhibited at acidic pH (<6.0)⁵⁷, resulting in N₂O accumulation in acidic environments^28,58. However, the current study revealed that two acidophilic methanotrophs, Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6 can reduce N₂O in moderately acidic (pH 5.5) and extremely acidic (pH 2.0) conditions, respectively. The existence of acid-tolerant N₂O reducers (pH 4.0 to 6.0) has been proposed in soil microcosm and enrichment experiments^59,60. So far, the only isolate implicated in N₂O reduction at an acidic pH (5.7) is Rhodanobacter sp. C01 isolated from acidic soil in Norway⁶¹. Our study reveals that N₂O reduction can occur even at an extremely acidic pH of 2.0. Furthermore, the conditions required for N₂O reduction in the N₂OR-containing Methylocystis strains remain unresolved. Perhaps some unknown growth or environmental factors are required to stimulate N₂O respiration in these methanotrophs, which will require further investigation.

Nitrate and nitrite reduction in Methylocella species

No anoxic growth of Methylocella species with CH₃OH and NO₃ ⁻

We next tested if the presence of denitrification enzymes in Methylocella tundrae T4 (nitrate reductase [NAR], nitric oxide reductase [NOR] and N₂OR) and Methylocella silvestris BL2 (NAR, nitrite reductase [NIR], and NOR) (Supplementary Dataset 1) can equate to growth when NO₃⁻ or NO₂⁻ is used as the sole terminal electron acceptor. Indeed, the presence of NAR (and NIR) in these methanotrophs resulted in NO₃⁻ (and NO₂⁻) reduction when methanol was provided as the sole electron donor. However, growth was barely detected under these conditions (Fig. 3A, B). Strain T4, which lacks a canonical NIR, reduced all the provided NO₃⁻ stoichiometrically to NO₂⁻ when provided with methanol as the sole electron donor (Fig. 3A). Under the same condition, strain BL2, a NAR and NIR-containing methanotroph, initially reduced the provided NO₃⁻ to NO₂⁻ and eventually, all the accumulated NO₂⁻ was stoichiometrically reduced to N₂O towards the end of the incubation (Fig. 3B). These results demonstrate that these methanotrophs have a functional NAR and or NIR and can utilize NO₃⁻ and/or NO₂⁻ instead of O₂ as a terminal electron acceptor. Nevertheless, these methanotrophs do not appear to rely on these activities for growth. Likewise, other aerobic methanotrophs have demonstrated denitrification activities under suboxic conditions. For example, the gammaproteobacterial methanotrophs Methylomonas denitrificans FJG1 and Methylomicrobium album BG8 were discovered to couple the oxidation of diverse electron donors to NO₃⁻ and NO₂⁻ reduction, respectively^21,22. However, none of these strains was demonstrated to couple this activity to growth, prompting us to investigate the possible reasons behind the lack of growth (see below). It should be noted that the genomes of all known Methylacidiphilum strains lack genes encoding a respiratory NAR (Supplementary Dataset 1).

Fig. 3: Anaerobic growth of Methylocella strains on methanol or pyruvate as the sole electron donor and NO₃⁻ as the terminal electron acceptor.

Methylocella tundrae T4 and Methylocella silvestris BL2 cells were grown in LSM medium supplemented with 30 mM methanol and 2–4 mM NO₃⁻. NH₄⁺ (2 mM) was supplied as the N-source. Anaerobic growth of Methylocella tundrae T4 (A) and Methylocella silvestris BL2 (B) cells on methanol as the sole electron donor with NO₃⁻ as the sole electron acceptor. Anaerobic growth of Methylocella tundrae T4 (C) and Methylocella silvestris BL2 (D) cells on pyruvate as the sole electron donor with NO₃⁻ as the sole electron acceptor. N₂O produced from NO₃⁻ reduction by cells of Methylocella silvestris BL2 grown on methanol or pyruvate is shown as an inset plot within each figure. N₂O production was not observed in strain T4, hence inset plots for N₂O production were not displayed. Lower NO₃⁻ (ca. 2.0 mM) was used in the case of methanol (A) to avoid NO₂⁻ toxicity. Growth was determined by optical density measurements at 600 nm, followed by measurements of NO₃^- and NO₂⁻ concentrations. Data are presented as mean ± 1 SD of triplicate experiments, and the error bars are hidden when they are smaller than the width of the symbols. Source data are provided as Source Data file.

Toxicity of reactive nitrogen species for Methylocella species

Considering that methanol oxidation was coupled to N₂O reduction and led to obvious growth in the N₂OR-containing methanotrophs (Figs 2B, H), the lack of growth during NO₃⁻ reduction by these microorganisms is suspected to be caused by the accumulation of growth-arresting reactive nitrogen species (RNS) like NO₂⁻ and NO (refs. ^62,63). Consistent with this hypothesis, the accumulation of NO₂⁻ in suboxic cultures of Vibrio cholerae and other bacterial species was found to limit population expansion but nitrate reduction still promoted cell viability⁶⁴. NO₂⁻ typically accumulates due to a lack of functional NIR as observed for strain T4 (Fig. 3A) and, to some degree, even transiently accumulates in the presence of a functional NIR, as observed for strain BL2 (Fig. 3B). The impact of NO₂⁻ accumulated from NO₃⁻ reduction might be more severe in acidic environments since protonation of NO₂⁻ leads to the formation of free nitrous acid (FNA), a known inhibitor of microbial anabolic and catabolic processes⁶⁵. In addition, chemodenitrification of NO₂⁻ (ref. ⁶⁶) could result in an accumulation of NO in the cell environment, which is highly toxic to microbial life⁶⁷. To further support the hypothesis of RNS toxicity, strain T4 was cultivated under N₂O-reducing conditions with methanol as the sole electron donor and supplied with NO₃⁻ instead of NH₄⁺ as the N source in the medium (Supplementary Fig. 5). Consistent with the idea that NO₂⁻ accumulation results in growth arrest, the culture growth plateaued at approximately the same time NO₂⁻ accumulated (≥ 0.3 mM NO₂⁻) (Supplementary Fig. 5A), whereas in control cultures containing NH₄⁺ instead of NO₃⁻ as the N-source, NO₂⁻ accumulation was not observed, and the cells were able to reach higher cell densities (Supplementary Fig. 5B). Furthermore, the effect of NO₂⁻ stress induced in strain T4 was verified by adding varying NO₂⁻ concentrations (0, 0.01, 0.03, 0.1, 0.3, and 1 mM) to aerobic (Supplementary Fig. 6A) and anaerobic N₂O-respiring cultures (Supplementary Fig. 6B). Nitrite, particularly at concentrations higher than 0.3 mM at pH 5.5, induced stress in Methylocella tundrae T4, resulting in growth inhibition (Supplementary Fig. 6). These results are comparable to that of Methylophaga nitratireducenticrescens JAM1, a facultative methylotroph, which, when grown aerobically on methanol at pH 7.4, had a four-fold decrease in biomass in the presence of 0.36 mM NO₂⁻ and did not grow in the presence of 0.71 mM NO₂⁻ (ref. ⁶⁸). Taken together, our data suggest that the failure of NO₃⁻/NO₂⁻-reducing methanotrophs to grow on methanol may result from RNS toxicity. On the other hand, when N₂O is reduced to N₂ by N₂O-reducing methanotrophs, the creation of these RNS is avoided, which may explain the disparity in growth with N₂O as the terminal electron acceptor compared to NO₃⁻ and NO₂⁻.

Toxicity of C1 metabolites in nitrate-reducing Methylocella species

Aside from the inhibitory effects of RNS, toxic intermediates from methanol metabolism might synergistically contribute to the inability of methanotrophs to grow when respiring NO₃⁻/NO₂⁻. Although formaldehyde is a key intermediate in the C1 metabolic pathway in many methylotrophs, it is highly toxic⁶⁹. Therefore, in situations where biomass production is limited due to RNS toxicity, it is likely that formaldehyde further retards the growth of denitrifying methanotrophs. To investigate this mechanism, we grew Methylocella strains under NO₃⁻-reducing conditions using a C-C electron donor, pyruvate, which does not generate formaldehyde as a major metabolite (Figs. 3C, 3D). Eventually, nearly all the supplied NO₃⁻ was stoichiometrically converted to NO₂⁻ and N₂O in strains T4 and BL2, respectively. In contrast to the lack of growth on methanol, pyruvate supported the growth of both Methylocella strains under NO₃⁻-reducing conditions (Fig. 3C, D). Growth was more pronounced in strain BL2 than in strain T4 (Fig. 3C, D), possibly due to the presence of NIR and NOR in addition to NAR in strain BL2, which limited NO₂⁻ accumulation (Fig. 3D). Nonetheless, no further growth on pyruvate was observed in strain BL2 after day 5, despite reduction of the accumulated NO₂⁻ (~2.5 mM) to N₂O (Fig. 3D). It is worth noting that the accumulated NO₂⁻ concentration (Fig. 3D) is higher than the 0.3 mM concentration that inhibited Methylocella tundrae T4 (Supplementary Fig. 6) and may also be responsible for the lack of growth in strain BL2.

Overall, these results demonstrate that in the tested Methylocella strains: (i) RNS have a major inhibitory effect on growth under denitrifying conditions; (ii) there are no growth benefits from methanol oxidation coupled to NO₃⁻ reduction, probably due to toxic C1 metabolic intermediates as well as RNS; and (iii) anaerobic growth is observed when NO₃⁻ reduction is coupled to the oxidation of pyruvate, a C-C electron donor; although the amount of growth is dependent on the completeness of the denitrification pathway and the accumulation of RNS. These propositions are supported by increased expression of genes involved in RNS and C1 metabolite detoxification under denitrifying conditions (see transcriptomic analysis below). Taken together, these results may explain why methanotrophs that couple methanol oxidation to NO₃⁻ or NO₂⁻ reduction show no clear signs of growth due to this process. Most methanotrophs can only utilize methane and its C1 derivatives as energy sources⁷⁰ and thus should not be able to grow under denitrifying conditions^21,22. On the other hand, versatile facultative methanotrophs of the genus Methylocella are potentially able to grow in strictly anoxic habitats when alternative multi-carbon substrates are available. In terrestrial environments, various nitrogen oxides, originating from nitrification and denitrification processes, coexist and are spatiotemporally dynamic⁷¹. Thus, depending on the versatility of NO₂⁻ and NO reduction potential of methanotrophs as well as their coexistence with other NO₂⁻ and NO-reducing microorganisms, N₂O respiration can be supported or compromised (see Supplementary Figs 5, 6).

N₂O reduction coupled with CH₃OH or CH₄ oxidation

N₂O reduction kinetics

We investigated N₂O respiration kinetics using resting cells of anaerobic N₂O-respiring cultures (CH₃OH + N₂O) in a microrespirometry (MR) chamber. Harvested cells of strains Methylacidiphilum caldifontis IT6 and Methylocella tundrae T4 were dispensed into a closed 10-mL MR chamber outfitted with O₂ and N₂O-detecting microsensors, supplied with CH₃OH (2 mM) and N₂O as a sole electron donor and acceptor, respectively, and incubated anoxically. The N₂O respiration kinetics followed Michaelis-Menten kinetics (Supplementary Fig. 7, Supplementary Note 2). The cells of strains T4 and IT6 grown at anoxic CH₃OH + N₂O conditions reduced N₂O at a maximum rate of 1.122 ± 0.005 mmol N₂O·h⁻¹·g DW⁻¹ (Supplementary Fig. 7A) and 0.414 ± 0.003 mmol N₂O·h⁻¹·g DW⁻¹ (Supplementary Fig. 7B), respectively. The molar ratios of CH₃OH to O₂ and CH₃OH to N₂O consumed were approximately 1:1.0 ( ± 0.05; n = 3) and 1:2.04 ( ± 0.17; n = 3), respectively, which coincide with the theoretical values obtained from Eqs. 3 and 4.

$${{{{{{\rm{CH}}}}}}}_{3}{{{{{\rm{OH}}}}}}+{{{{{{\rm{O}}}}}}}_{2}\to 0.5{{{{{{\rm{CO}}}}}}}_{2}+1.5{{{{{{\rm{H}}}}}}}_{2}{{{{{\rm{O}}}}}}+0.5{{{{{{\rm{CH}}}}}}}_{2}{{{{{\rm{O}}}}}}({{{{{\rm{biomass}}}}}})$$

(3)

$${{{{{{\rm{CH}}}}}}}_{3}{{{{{\rm{OH}}}}}}+2{{{{{{\rm{N}}}}}}}_{2}{{{{{\rm{O}}}}}}\to 0.5{{{{{\rm{C}}}}}}{{{{{{\rm{O}}}}}}}_{2}+1.5{{{{{{\rm{H}}}}}}}_{2}{{{{{\rm{O}}}}}}+2{{{{{{\rm{N}}}}}}}_{2}+0.5{{{{{\rm{C}}}}}}{{{{{{\rm{H}}}}}}}_{2}{{{{{\rm{O}}}}}}({{{{{\rm{biomass}}}}}})$$

(4)

Sensitivity of N₂OR to O₂

While O₂ is well known to impair N₂OR activity⁷², some bacterial strains have been reported to reduce N₂O in the presence of O₂ (refs. ^73,74). We therefore tested the capacity of strains IT6 and T4 to reduce N₂O in the presence of O₂ by using resting cells of anoxic CH₃OH + N₂O cultures. After spiking O₂ to strain IT6 cells respiring N₂O in the anoxic MR chamber, N₂O-respiration ceased: dropping from the maximum (0.4–0.5 mmol N₂O·h⁻¹·g DW⁻¹) to zero (Fig. 4A, Table 1). N₂O reduction activity only started when the dissolved O₂ concentration was below ca. 3 µM, suggesting the N₂O reduction activity of this strain is highly sensitive to O₂. In contrast, when O₂ ( ~ 14 and 30 µM) was added to N₂O-respiring cells of strain T4, simultaneous reduction of N₂O and O₂ was observed (Fig. 4B). However, the N₂O respiration rates dropped to 0.24 and 0.13 mmol N₂O·h⁻¹·g DW⁻¹ after spiking ~14 and 30 µM O₂, respectively, which were approximately 34 and 20% of the maximum rate before O₂ introduction (0.64–0.68 mmol N₂O·h⁻¹·g DW⁻¹). These results suggest that in contrast to strain IT6, N₂O reduction in strain T4 is not highly impaired by O₂. N₂OR activity fully recovered in both strains after O₂ was depleted. Because the N₂OR of strain IT6 was found to be highly sensitive to O₂, further characterization of methanotroph N₂OR activity in response to O₂ exposure was limited to strain T4.

Fig. 4: Microrespirometry-based N₂O and O₂ reduction during methanol oxidation by N₂OR-containing methanotrophs.

N₂O and O₂ reduction by cells of Methylacidiphilum caldifontis IT6 (A) and Methylocella tundrae T4 (B) during methanol oxidation. Filled blue dots represent dissolved N₂O, filled orange dots represent dissolved O₂, and filled black dots represent N₂O reduction rates. Experiments were performed in a microrespirometry (MR) chamber fitted with O₂ and N₂O microsensors. The red arrows mark the addition of 14–33 µM O₂ into the MR chamber. The red- and green-marked numbers close to the red and green lines represent the N₂O reduction rates before and during O₂ reduction (gray-shaded area) in the MR chamber, respectively. Source data are provided as Source Data file.

Table 1 Microrespirometry-based substrate-specific N₂O- or O₂-reduction rate by Methylocella tundrae T4 cells grown under anoxic and suboxic growth conditions

Considering these results, we set out to see if cells of strain T4 could continue N₂O respiration while using O₂ for CH₄ oxidation in the MR chamber. The cells used for this experiment were cultured in suboxic conditions with starting gas mixing ratios (v/v) of 1% O₂, 5% N₂O, and 20% CH₄ (i.e., CH₄ + O₂ + N₂O condition). Similar to the anoxic CH₃OH + N₂O-adapted cells described above, the suboxic CH₄ + O₂ + N₂O-adapted cells co-respired O₂ and N₂O after injecting CH₄ (~406 µM) into a 5-mL MR chamber containing O₂ (~30 µM) and N₂O (~480 µM) (Fig. 5A). Interestingly, the maximum N₂O respiration rates during each O₂ spike were 1.4 to 2 times higher (1.58–2.47 mmol N₂O·h⁻¹·g DW⁻¹) in the suboxic CH₄ + O₂ + N₂O-adapted cells (Fig. 5B, Table 1) than in the anoxic CH₃OH + N₂O-adapted cells (1.12 ± 0.01 mmol N₂O·h⁻¹·g DW⁻¹) (Table 1, Supplementary Fig. 7B), suggesting that the cells can modulate the rates of N₂O reduction in response to O₂ availability.

Fig. 5: Simultaneous N₂O and O₂ reduction by Methylocella tundrae T4 cells during CH₄ oxidation in microrespirometry (MR) and growth experiments.

A MR experiment showing the simultaneous reduction of N₂O and O₂ by Methylocella tundrae T4 cells during CH₄ oxidation. B N₂O and O₂ reduction rates by cells of strain T4 during CH₄ oxidation calculated from (A). The filled orange and blue dots in the upper (A) represent the concentrations of dissolved O₂ and N₂O, respectively. The filled orange and blue dots in the bottom (B) represent the rates of O₂ and N₂O reduction, respectively. Experiments were performed in a MR chamber fitted with O₂ and N₂O microsensors. The red arrow marks the addition of CH₄ (~406 µM) into the MR chamber. The black arrow marks the addition of ~26 µM or ~60 µM O₂ into the MR chamber. The gray-shaded area represents points where N₂O and O₂ are reduced simultaneously. C Growth experiment showing Methylocella tundrae T4 cells reducing N₂O and O₂ simultaneously during CH₄ oxidation. The culture was grown in 2-liter sealed bottles (triplicates) containing 60 mL of LSM medium with 2 mM NH₄⁺ as the N-source. The headspace of the bottles was composed of CH₄ (5%, v/v), O₂ (0.5%, v/v), N₂O (1.4%, v/v), and CO₂ (5%, v/v) and supplemented with additional O₂ (~0.5%, v/v) before its depletion. The incubation period shown in (C) is after the initial 20-day incubation period. After the depletion of O₂, additional O₂ was spiked to observe the simultaneous reduction of O₂ and N₂O during CH₄ oxidation. Data are presented as the mean ± 1 SD of a triplicate experiment, and the error bars are hidden when they are smaller than the width of the symbols. Source data are provided as Source Data file.

Accordingly, the maximum rates of N₂O reduction (1.58–2.47 mmol N₂O·h⁻¹·g DW⁻¹) and O₂ reduction (0.98–2.37 mmol O₂·h⁻¹·g DW⁻¹) by the suboxic CH₄ + O₂ + N₂O-adapted cells were comparable (Fig. 5B, Table 1). As the O₂ concentration and reduction rate decreased, the N₂O reduction rate also decreased (Fig. 5A, B), revealing that activation of CH₄ by O₂ is required for stimulating N₂O respiration by CH₄ + O₂ + N₂O-adapted cells. Based on these results, we conclude that, under suboxic conditions, both aerobic CH₄ oxidation and N₂O reduction were operating in concert: O₂ was needed for the monooxygenase, but the N₂OR remained active and was able to accept electrons released downstream in the C1 oxidation pathway. This adds to the evidence that aerobic N₂O respiration occurs in strain T4 and is linked to aerobic CH₄ oxidation.

Finally, we estimated the O₂ concentration range at which the suboxic CH₄ + O₂ + N₂O-adapted cells of strain T4 show N₂O-reducing activity. At a O₂ concentration of 170 µM, O₂ and N₂O were reduced simultaneously (Supplementary Fig. 8A, B). The maximum N₂O reduction rate (Table 1) was nearly constant (1.32 ± 0.25 mmol N₂O·h⁻¹·g DW⁻¹) across the O₂ concentration range of 5–170 µM (Supplementary Fig. 8B, C) and was about 1.4 times higher than the maximum O₂ reduction rates (0.95 ± 0.09 mmol O₂·h⁻¹·g DW⁻¹). This means that even when exposed to high levels of O₂, the N₂OR in the suboxic CH₄ + O₂ + N₂O-adapted cells remained functional and could reduce N₂O at high rates. Other bacterial strains’ N₂OR activities have been reported at O₂ concentrations between 100 and 260 µM (refs. ^73,74), indicating that their N₂OR activity is similarly O₂-tolerant⁷³ as that of strain T4. According to the findings of Wang and colleagues⁷³, N₂O reducers with an O₂ tolerant N₂OR maintain low internal O₂ concentrations in their cells by rapidly consuming O₂, allowing the N₂OR to remain active. However, it remains unclear if Methylocella tundrae T4 employs a similar strategy to maintain an O₂-tolerant N₂OR.

Improved methanotrophic growth of Methylocella tundrae in the presence of N₂O

Based on the MR experiments showing the simultaneous reduction of O₂ and N₂O by CH₄-fed cells of strain T4, alongside the clear N₂O-dependent anaerobic growth, we hypothesized that strain T4 growth can be enhanced when it oxidizes CH₄ by simultaneously reducing O₂ and N₂O under suboxic conditions. Using fed-batch growth experiments, we verified that strain T4 grows by CH₄ oxidation coupled with co-respiration of N₂O and O₂ (Fig. 5C, Table 2), strongly supporting the MR results above. Cells grown under the suboxic CH₄ + O₂ + N₂O condition consumed roughly the same amount of O₂ and N₂O (Fig. 5C, Table 2), and these values were comparable to what CH₄ + O₂ + N₂O-grown cells consumed in the MR experiments (see Fig. 5A). Consequently, our results demonstrate that in an O₂-limited environment, the cells can benefit energetically by directing more O₂ to the monooxygenase step of CH₄ oxidation, and simultaneously running a hybrid (O₂ + N₂O) electron transport system as shown in Table 2 and Fig. 5C.

Table 2 The effect of N₂O addition on CH₄-oxidizing cultures of Methylocella tundrae T4 growing in suboxic conditions

The data showed unequivocally that the total electron equivalents released during CH₄ oxidation to CO₂ could account for the total electron acceptor (O₂ + N₂O) reduced. Based on a CH₄ to O₂ ratio of 1:1.57 (ref. ⁷⁵), the total amount of O₂ reduced (13.96 mmol·L^-1) by the suboxic CH₄ + O₂ + N₂O cultures could theoretically only account for 8.89 mmol·L^-1 oxidized CH₄. However, a larger total of 12.19 mmol·L^-1 CH₄ was oxidized by this culture (Table 2), and the excess 3.29 mmol·L^-1 must have required an additional electron acceptor (i.e., N₂O). Consistently, about 10.15 mmol·L^-1 N₂O was reduced by the suboxic CH₄ + O₂ + N₂O cells, equivalent to 5.08 mmol·L^-1 O₂, since half as many electrons are consumed per mol during N₂O reduction to N₂ compared to O₂ reduction to H₂O. By running the N₂O respiration system, the cells lower their O₂-demand for respiration by the aerobic terminal oxidase and maximize O₂ use by the methane monooxygenase⁷⁶. Due to having more CH₄ oxidized per O₂ reduced (~37%) when N₂O is present, higher cell densities (OD₆₀₀) per O₂ reduced (~34%) were reached in the suboxic CH₄ + N₂O + O₂ cultures than in the O₂-replete CH₄ + O₂ cultures (Table 2), further demonstrating the beneficial contribution of N₂O reduction to growth on CH₄ at suboxic conditions.

Transcriptomics

The overall regulation of key genes involved in denitrification and methane oxidation is depicted in Fig. 6 as well as in the supplementary material (Supplementary Figs. 9, 10, 11, Supplementary Datasets 5, 6, 7). Differences in expression were considered significant if the Log₂FC was higher than [0.85] or lower than [-1.0] with an adjusted p ≤ 0.05.

Fig. 6: Metabolic reconstruction and transcriptional response of Methylocella tundrae T4 cells to O₂-replete (CH₃OH + O₂) and anoxic (CH₃OH + N₂O) methanol-oxidizing growth conditions.

The genes used to reconstruct the metabolic pathway are listed in Table S5. The gene products are shaded according to the relative fold change (Log₂FC) in gene expression between cells grown under anoxic (CH₃OH + N₂O) and O₂-replete (CH₃OH + O₂) conditions. Genes up-regulated in CH₃OH + N₂O-grown cells are shown in teal green, while those up-regulated in CH₃OH + O₂-grown cells are shown in purple. Note that proteins are not drawn to scale. Methanol oxidation: Methanol is oxidized to formaldehyde in the periplasmic space by the PQQ-dependent methanol dehydrogenase (Xox- and Mxa-type), T4_03519–21, T4_00353–5, and T4_01862–76. The NAD(P)⁺-dependent alcohol dehydrogenase (T4_03199) may also be involved in methanol oxidation to formaldehyde in the cytoplasmic space during anaerobic growth on methanol. Formaldehyde oxidation to formate then proceeds via the tetrahydromethanopterin (H₄MPT) pathway, and C1 incorporation into the serine cycle is mediated by the tetrahydrofolate (H₄F) carbon assimilation pathway. The Calvin-Benson-Bassham pathway is also a possible route for CO₂ fixation. Nitrous oxide reduction: N₂O is reduced to N₂ through the activity of nitrous oxide reductase in the periplasmic space. Electron transfer to NosZ occurs via cytochrome c from the cytochrome bc1 (Qcr) complex^136,137. Electron transfer to the NosZ may also involve direct interaction with methanol dehydrogenase C-type cytochrome (XoxG, MxaG). The NosR protein may be involved in the transfer of electrons to NosZ (refs. ^136,137).

N₂OR (O₂ replete vs. anoxic conditions)

The transcript levels of the N₂OR-encoding genes (T4_03941–7), nosRZDFYLX, were 2- to 4.7-fold higher in strain T4 cells respiring N₂O in the anoxic CH₃OH + N₂O conditions compared to strain T4 cells respiring O₂ in the O₂-replete CH₃OH + O₂ conditions (Fig. 6 Supplementary Datasets 5, 6). Cells of strain IT6 respiring N₂O in the anoxic CH₃OH + N₂O conditions showed transcriptional upregulation (1.9–6.7-fold) of four nos genes (nosC1BZC2) under anoxic conditions (Supplementary Fig. 10, Supplementary Dataset 7). Other nos operon genes (nosYFDL; IT6_00904–11) were expressed constitutively under both the anoxic CH₃OH + N₂O and O₂-replete CH₃OH + O₂ conditions. The NosC1 and NosC2 proteins of Wolinella succinogenes were predicted to facilitate electron transfer from menaquinol to the periplasmic NosZ during the reduction of N₂O to N₂ (ref. ⁴⁷) and are likely to play a similar role in strain IT6. Although the exact function of NosB has yet to be elucidated, Hein and colleagues⁷⁷ used a non-polar nosB deletion mutant of Wolinella succinogenes to show that it is necessary for N₂O respiration. Overall, increased expression of N₂OR-encoding genes in Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6 cells during anaerobic growth indicates that the N₂OR is functional in these methanotrophs and supports their ability to respire and grow using N₂O as a terminal electron acceptor.

N₂OR (O₂ replete vs. suboxic conditions)

Transcript levels of N₂OR-encoding genes were 2–10.7-fold higher in strain T4 cells grown under suboxic CH₄ + O₂ + N₂O conditions than in cells grown under O₂-replete CH₄ + O₂ conditions (Supplementary Fig. 9, Supplementary Datasets 5, 6). This finding is consistent with the N₂O respiration activity and growth of strain T4 under suboxic CH₄ + O₂ + N₂O conditions (Fig. 5), in which the cells can efficiently oxidize more CH₄ (see Table 2), most likely because the use of N₂O for cellular respiration allows them to devote more O₂ to CH₄ oxygenation.

Methanol dehydrogenase (O₂ replete vs. anoxic conditions)

In methanotrophs, methanol oxidation occurs in the periplasmic space by PQQ (pyrroloquinoline quinone)-dependent methanol dehydrogenase (MDH). Seven PQQ-dependent alcohol dehydrogenases (ADHs)⁷⁸ are encoded in the genome of strain T4 (Supplementary Dataset 5). Five are type I ADHs (quinoproteins), which include one calcium-dependent MDH (MxaF-type MDH), and four lanthanide-dependent MDHs (XoxF-type MDH), divided into clades 1 (XoxF1), 3 (XoxF3), and 5 (XoxF5; 2 copies) (Supplementary Fig. 12). The other two are type II ADHs (quinohemoproteins). In addition to the PQQ-dependent ADH, Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6 genomes contain genes encoding cytosolic Zn²⁺-dependent ADH, which are part of a large family of enzymes that oxidize alcohols to aldehydes or ketones and reduce NAD(P)⁺ or a similar cofactor⁷⁹ (Supplementary Datasets 5, 6).

Among the four XoxF-type MDHs encoded in the genome of strain T4, genes in a xoxFGJ operon (T4_03519–21), which include a gene encoding a XoxF5 enzyme, were found to be constitutively transcribed at high levels in cells grown under both O₂-replete CH₃OH + O₂ and anoxic CH₃OH + N₂O conditions (Fig. 6, Supplementary Datasets 5, 6). Thus, the xoxF5 gene likely encodes the predominant MDH used by strain T4 in both O₂-respiring and N₂O-respiring cells. The other singleton xoxF5 gene (T4_03691) and a xoxF3 gene found in a separate xoxFGJ cluster (T4_00353–5) were also significantly upregulated in cells grown under anoxic CH₃OH + N₂O conditions in comparison to cells grown under O₂-replete CH₃OH + O₂ conditions (Fig. 6, Supplementary Datasets 5, 6). Furthermore, we observed a significant upregulation (2- to 22-fold) of the genes encoding MxaFI-type MDH (T4_01872–6) in the anoxic CH₃OH + N₂O-grown cells (Fig. 6, Supplementary Datasets 5, 6). Thus, our results indicate the use of various MDHs by strain T4 during anaerobic growth. In strain IT6 a xoxF gene encoding a XoxF2-type MDH is present as part of the xoxGJF operon (IT6_00336–8) (Supplementary Dataset 7) and the expression of the xoxF2 gene was 2-fold upregulated in the N₂O-respiring cells (Supplementary Fig. 10, Supplementary Dataset 7).

A cytosolic Zn²⁺-dependent ADH bound to NAD(P)⁺ is known to perform the oxidation of methanol in Gram-positive methylotrophs⁸⁰. A Zn²⁺-dependent ADH (T4_03199) of strain T4 was significantly upregulated (13.8-fold) in the anoxic CH₃OH + N₂O-grown cells compared to the O₂-replete CH₃OH + O₂-grown cells (Fig. 6, Supplementary Datasets 5, 6). Strain IT6 genome also contained three copies of genes encoding enzymes annotated as Zn²⁺-dependent ADH (Supplementary Dataset 7). The expression of two of these genes (IT6_01501 and IT6_01931) were 3.9-fold and 2.5-fold upregulated in N₂O-respiring cells compared to cells respiring O₂ (Supplementary Fig. 10, Supplementary Dataset 7). Even though PQQ-dependent MDHs have a high affinity for and activity with methanol as a substrate, their use in strictly anoxic conditions will be limited because PQQ biosynthesis requires molecular oxygen⁸¹. Thus, PQQ-dependent MDHs are suggested to be functional at completely anoxic conditions only when PQQ is carried over from an aerobic growth stage or provided externally⁸². On the other hand, Zn²⁺-dependent MDHs have the advantage of utilizing a ubiquitous cofactor, NAD(P)⁺, and can be functional during anaerobic growth⁸³. This finding raises the possibility that strains T4 and IT6 can employ alternative ADHs such as the Zn²⁺-dependent ADH to facilitate methanol oxidation in strict anoxia. Some genes required for the subsequent steps of C1 metabolism, i.e., formaldehyde and formate dehydrogenases, were also upregulated in strain T4 (but not IT6) growing anaerobically. These are depicted in Fig. 6 and supplementary materials (Supplementary Fig. 9, Supplementary Dataset 5).

Methanol dehydrogenase (O₂ replete vs. suboxic conditions)

Furthermore, we also examined expression levels of genes encoding MDHs in strain T4 cells grown under suboxic CH₄ + O₂ + N₂O conditions (Supplementary Fig. 9, Supplementary Datasets 5, 6). Genes in the cluster T4_01862–76, which encodes the calcium-dependent MDH (MxaF-type MDH), had the highest transcript expression among all MDH-encoding genes in CH₄-oxidizing cells grown under suboxic CH₄ + O₂ + N₂O conditions. When compared to O₂-replete CH₄ + O₂ conditions, the expression of genes within this cluster was 1.8- to 371.5-fold upregulated (Supplementary Fig. 9, Supplementary Datasets 5, 6). This is unexpected since genes encoding the Mxa-type MDH are typically downregulated in the presence of lanthanides⁸⁴; which we also included (2 µM each of cerium and lanthanum) in the growth medium. Their apparent upregulation (even when lanthanides are present) suggests that this enzyme might play an important role in CH₄ metabolism in the presence of N₂O and suboxic conditions. As observed above, genes in the xoxFGJ operon (T4_03519–21) were also highly expressed at the suboxic conditions (Supplementary Fig. 9, Supplementary Datasets 5, 6), suggesting that this key MDH is used by strain T4 in all three conditions. Genes in the cluster T4_01892–4 including the gene encoding the XoxF1 MDH were also significantly upregulated (18- to -35-fold) in the suboxic CH₄ + O₂ + N₂O-grown cells compared to the O₂-replete CH₄ + O₂-grown cells. The operon T4_02097–8, which encodes a cytochrome c550 (T4_02097) and a type II ADH (T4_02098), exhibited 6-fold and 30.6-fold upregulation, respectively, in cells grown under suboxic CH₄ + O₂ + N₂O conditions as opposed to cells grown under O₂-replete CH₄ + O₂ conditions. In addition, two Zn²⁺-dependent ADHs (T4_03097 and T4_03199) were significantly upregulated (3.5-fold and 46-fold, respectively) in strain T4 cells grown under suboxic CH₄ + O₂ + N₂O conditions compared to cells grown under O₂-replete CH₄ + O₂ conditions. Overall, it appears that cells oxidizing methanol under anoxia (CH₃OH + N₂O-grown cells) or those oxidizing methane under suboxia (CH₄ + O₂ + N₂O-grown cells) use a distinct set of MDHs from those they use during O₂ respiration.

Methane monooxygenase

The genomes of Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6 contain genes that encode sMMO and pMMO, respectively. In the suboxic CH₄ + O₂ + N₂O conditions, all the genes (mmoXYBZDCRG) in the gene cluster T4_01946–54 displayed a high degree of transcriptional upregulation (18.7–96-fold) compared to O₂-replete CH₄ conditions (Supplementary Fig. 9, Supplementary Datasets 5, 6). In a previous study⁸⁵, Methylosinus trichosporium OB3b sMMO activity and protein expression were found to be significantly elevated under hypoxic conditions (24 µM) compared to higher O₂ conditions (188 µM). Furthermore, Methylosinus trichosporium OB3b sMMO’s catalytic activity in the degradation of dichloroethane was enhanced at low O₂ levels and impaired at elevated O₂ levels⁸⁶. Thus, in methanotrophs, upregulation of methane monooxygenase genes under O₂ limiting conditions might be a strategy to produce more methane monooxygenase. This will lead to increased methane oxidation and thus provide stronger competition for the limited O₂ with the terminal oxidase. Aside from the methane monooxygenase genes, group II and III truncated hemoglobin encoding genes were upregulated in Methylocella tundrae T4 (T4_02445, T4_02637, and T4_00400; 4- to 12-fold) and Methylacidiphilum caldifontis IT6 (IT6_00149; 3-fold) cells in response to suboxia or anoxia (Supplementary Datasets 5, 6). These truncated hemoglobins are thought to transport O₂ to the methane monooxygenase²². Compared to methane, methanol resulted in lower transcript levels of sMMO genes in Methylocella tundrae T4 (Supplementary Fig. 11, Supplementary Dataset 5), with much lower levels in the O₂ replete CH₃OH + O₂ conditions compared to the anoxic CH₃OH + N₂O conditions (Fig. 6, Supplementary Datasets 5, 6). Transcriptional repression of sMMO genes by growth substrates other than methane has been observed in Methylocella silvestris BL2 (refs. ^87,88). The expression of genes encoding denitrification enzymes, their transcriptional regulators, and terminal oxidase is described in Supplementary Note 3.

Ecological relevance

Our findings revealed that certain methanotrophic strains, particularly those from the genera Methylocella and Methylocystis, which are commonly found in acidic and neutral terrestrial environments based on ecological meta-data from the BacDive database^89,90, have the ability to reduce N₂O. Wetlands, such as acidic peatlands and paddy fields, are significant contributors to the release of CH₄ and N₂O (refs. ^27,91,92). Although active N₂O consumption has been observed in acidic wetlands⁹³, little is known about the microbial mechanisms that drive these processes. In a recent study²⁷ wherein active N₂O consumption was observed in peatlands (pH 6.4–3.7) located in Central and South America, Methylocystis species accounted for over 20% of the N₂O-reducing microbial community based on nosZ gene amplicon sequence variants. This implies that N₂OR-containing methanotrophs might make significant contributions to N₂O reduction in these environments. The current prevailing perception of N₂OR containing methanotrophs as a phylogenetically narrow group with limited ecological impact might be heavily biased by the scarcity of cultured methanotrophs with such metabolic capabilities. Thus, additional in situ and ecogenomic-based investigations are needed to more precisely quantify the contribution of known methanotrophs to N₂O reduction as well as to uncover other novel N₂O-reducing methanotrophs, such as those belonging to the Gemmatimonadota phylum⁵⁰.

Short-term or seasonal water table fluctuations caused by either natural or anthropogenic desiccation influence the transition zone from oxic to anoxic conditions in wetlands^94,95,96. In the deeper, water-filled anoxic layer of wetlands⁹⁷, and even in oxygenated wetland soils⁹⁸, methanogens produce CH₄. N₂O can be produced from denitrification processes, especially by incomplete denitrifiers which are frequently abundant in environments^30,31,32. Nitrifiers also produce a significant amount of N₂O as a byproduct of ammonia oxidation in the suboxic layers⁹⁹. Furthermore, NO₂⁻ produced from nitrogen cycling processes can be abiotically reduced to N₂O through chemodenitrification due to the stability of Fe²⁺ in acidic peat soils. At the oxic-anoxic interface, where CH₄ and O₂ gradients overlap, N₂O-respiring methanotrophs will have simultaneous access to both CH₄ and N₂O. Although the CH₄-O₂ counter gradient is dynamic and O₂-respiring organisms can rapidly deplete the limited O₂, these N₂O-respiring methanotrophs can use a growth strategy that involves respiring both N₂O and O₂ and coupling it to CH₄ oxidation. This unique lifestyle, combined with the potential ability to respire N₂O solely with non-methane substrates such as C1, C-C compounds^51,100 as well as H₂ (refs. ^52,100), can confer a selective growth advantage, facilitate their niche expansion to suboxic and anoxic zones, and make them resilient in such environments.

In conclusion, we revealed that sMMO- and pMMO-containing acidophilic methanotrophs of the genera Methylocella and Methylacidiphilum can grow anoxically by respiring N₂O using clade I and II NosZ, respectively. N₂O reduction was detected at an extremely acidic pH of 2.0, which is by far the lowest pH reported for this process^27,92. Further, N₂O reduction can improve the growth yields of these bacteria under O₂-limiting conditions and provide a competitive advantage. This study significantly expands our perception of the potential ecological niches of aerobic methanotrophs. In addition to mitigating CH₄ and CO₂ emissions, aerobic methanotrophs potentially play a role in reducing the emission of the climate-active and ozone-depleting gas N₂O, particularly in low pH environments.

Methods

Bacterial strains and growth conditions

The methanotrophic bacterial strains used for the experiments include Methylacidiphilum caldifontis IT6, Methylacidiphilum infernorum IT5, Methylocella tundrae T4 ( = KCTC 52858 ^T), Methylocella silvestris BL2 ( = KCTC 52857 ^T), Methylocystis sp. SC2, and three in-house Methylocystis echinoides-like isolates (strains IM2, IM3, and IM4). The Methylacidiphilum strains are also in-house strains isolated previously from a mud-water mixture taken from Pisciarelli hot spring in Pozzuoli, Italy⁴⁴. The Methylocella strains were obtained from the Korean Collection for Type Cultures (KCTC). Growth of the bacterial strains was performed using a low salt mineral (LSM) medium. The medium contained 0.4 mM MgSO₄·7H₂O, 0.2 mM K₂SO₄, and 0.1 mM CaCl₂·2H₂O and was supplemented with filter-sterilized solutions of 2 mM (NH₄)₂SO₄, 0.1 mM KH₂PO₄, 1 µM CeCl₃, 1 µM LaCl₃, 1 mL (1×) vitamin, and 1 mL (1×) trace element solutions¹⁰¹ per liter. The pH of the medium was adjusted to pH 2.0 with concentrated sulfuric acid (filter-sterilized) for the Methylacidiphilum strains and to pH 5.5 with 20 mM 2-morpholinoethanesulfonic acid (filter-sterilized) for the Methylocella strains. The cultures were incubated at 52 °C for Methylacidiphilum strains (IT5 and IT6) and 28 °C for Methylocella strains (T4 and BL2) with shaking at 160 rpm. Unless stated otherwise, ammonium sulfate, (NH₄)₂SO₄, was used as the nitrogen source.

Enrichment and isolation of Methylocystis strains

The N₂OR-containing Methylocystis strains were isolated from an acidic forest soil in Chungcheongbuk-do, South Korea (36°55'31” N 127°54'86” E). The soil sample preparation and initial enrichment of the methanotrophs¹⁰², as well as the isolation of methanotrophic strains through repeated serial dilution of the enrichment cultures¹⁰³, have all been described previously. Briefly, the most diluted culture exhibiting methane oxidation was serially diluted and filtered through 0.2-μm Track-Etch membrane polycarbonate filters (Whatman). The filters were placed on LSM medium (pH 5.5) in Petri dishes and incubated at 30 °C in airtight containers containing CH₄ (10%, v/v) and CO₂ (5%, v/v). Colonies that appeared on the filters after 3 weeks of incubation were transferred to fresh LSM medium in 160-mL serum vials with the same gas composition. Three individual methanotrophic isolates were identified by sequencing the 16 S rRNA gene with the 27 F/1492 R primer set¹⁰⁴. The purity of the isolates was confirmed by seeding aliquots of the CH₄-grown cultures into the LSM medium with 0.05% (w/v) yeast extract, tryptic soy broth, and Luria-Bertani broth without CH₄ and incubating at 30 °C. Three methanotrophic isolates, IM2, IM3, and IM4, shared 99.46% 16 S ribosomal RNA (rRNA) gene-sequence identity with the alphaproteobacterial methanotroph Methylocystis echinoides LMG27198. The three strains share average nucleotide identity values ranging from 81.85–81.93 with Methylocystis echinoides LMG27198, implying that they represent a new species in the genus Methylocystis.

DNA isolation, genomic and phylogenetic analyses

High-molecular-weight genomic DNA was extracted using a modified CTAB method¹⁰⁵, from 200 mL amounts of Methylocella tundrae T4 grown in methanol, and the Methylocystis isolates (strain s IM2, IM3, and IM4) grown in CH₄. The genomes of Methylocella tundrae T4 and Methylocystis sp. IM3 were sequenced at LabGenomics (Seongnam, Republic of Korea) and Macrogen (Seoul, Republic of Korea) using the PacBio RS II (long-read sequencing) and Illumina HiSeq (2 × 150 bp) platforms, respectively. The genomes of Methylocystis sp. IM2 and Methylocystis sp. IM4 were sequenced using a MinION R10.4.1 flow cell (FLO-MIN114, Oxford Nanopore Technologies). The PacBio reads were assembled with the Trycycler pipeline (v0.5.4)¹⁰⁶. Filtered reads were subsampled and assembled using Miniasm/Minpolish (v0.3-r179)¹⁰⁷, Flye (v2.9.2)¹⁰⁸, and Raven (v1.8.3)¹⁰⁹ assemblers. The consensus contigs were polished with Illumina short reads using Polypolish (v0.5.0)¹¹⁰ and POLCA (v4.0.5)¹¹¹. The circularity was confirmed during the Trycycler pipeline assembly and again by mapping the Illumina reads backward. De novo genome assembly of the MinION long reads was accomplished using Flye (v2.9.2)¹⁰⁸. Annotation of methanotrophs’ genomes was performed with the Prokka annotation pipeline (v1.14.6)¹¹² and NCBI Prokaryotic Genome Annotation Pipeline (PGAP; v4.2)¹¹³. Functional assignment of the predicted genes was improved using a set of public databases (InterPro¹¹⁴, GO^115,116, PFAM¹¹⁷, CDD¹¹⁸, TIGRFAM¹¹⁹, and EggNOG¹²⁰). Prediction of signal peptides and transmembrane helices was performed using the web-based services SignalP (v6.0)¹²¹ and TMHMM (v2.0)¹²² with default settings.

The distribution of denitrification genes in methanotroph isolates or metagenome-assembled genomes (MAG) (meeting the following CheckM (v1.2.2) criteria: completeness > 60% and contamination <10%) was examined using genomic data from the NCBI assembly database. Reference protein sequences of denitrification enzymes (NapA, NapB, NarG, NarH, NarI, NirS, NirK, NorB, NorC, and NosZ) were obtained from the NCyc¹²³ and BV-BRC¹²⁴ databases. The annotated protein sequences of methanotrophs were re-annotated against the obtained reference sequences from the NCyc¹²³ and BV-BRC¹²⁴ databases. The identities of the obtained denitrification protein sequences in methanotrophs were verified using manual alignment and tree-building procedures with reference sequences. Sequences incorrectly annotated as denitrification genes were removed, and only candidate genes that clustered with reference sequences were counted as true hits.

For phylogenetic analyses of the NosZ proteins and methanol dehydrogenases of strains T4 and IT6, representative amino acid sequences of the genes of related taxa were obtained from NCBI. The derived amino acid sequences of the NosZ and methanol dehydrogenases (XoxF and MxaF) were aligned using MAFFT (v7.511)¹²⁵. Maximum-likelihood trees were inferred with IQ-TREE (v1.6.12). The constructed trees and operon arrangements were visualized using iTOL (v.6.7.2)¹²⁶ and used for annotation. Genomic islands were predicted using the IslandViewer web server¹²⁷.

Anoxic growth coupled with N₂O reduction

To demonstrate the ability of N₂OR-containing methanotrophs to grow using N₂O as the electron acceptor, we established anoxic batch cultures of Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6 in 160-mL bottles containing 20 mL of LSM media and inoculated with 1–5% (v/v) actively growing-cells from the log phase (starting OD₆₀₀ values ≤ 0.05). To remove oxygen, nitrogen gas (N₂, purity >99.999%) was introduced into the bottles via a long needle (18 G). Following that, the bottles were flushed with N₂ gas for 20 min before being sealed with gas-tight butyl rubber stoppers and aluminum crimp seals to prevent O₂ leakage. We used contactless trace-range oxygen sensor spots (TROXSP5) to monitor O₂ contamination (<0.10%, v/v) in the culture bottles incubated after N₂-flushing (see Analytical methods, for details). These spots have a detection limit of 20 nM O₂. Chemical-reducing agents, Na₂S (0.5, 1, and 2 mM), cysteine (0.5 mM), DTT (0.5 mM), and titanium citrate (0.5 and 1 mM) in the media resulted in severe cell toxicity, hindering their use in this study as previously reported for N₂OR reducer Anaeromyxobacter dehalogenans¹²⁸. When the cultures were incubated without the chemical-reducing agents, the cells completely depleted the trace O₂ concentration present in the culture bottles in less than 24 h as measured by the oxygen sensor spots.

The N₂OR-lacking methanotrophs Methylocella silvestris BL2 and Methylacidiphilum infernorum IT5, which are closely related to Methylocella tundrae T4 and Methylacidiphilum caldifontis IT6, respectively, were used as negative controls. Methanol (30 mM), N₂O (5%, v/v), and CO₂ (5%, v/v) were used as the energy source, electron acceptor, and carbon source, respectively. In addition, pyruvate (10 mM) and hydroxyacetone (acetol) (10 mM) were tested as the sole C-C electron donors in strains T4 and IT6, respectively. Furthermore, strain IT6 cells were investigated to grow chemolithoautotrophically in sealed 1-liter bottles (duplicate) containing 20 mL of LSM medium at pH 2.0 on H₂ (10% v/v) with and without N₂O (5% v/v). As part of the control experiments, we incubated cells from the four strains in LSM media under anoxic conditions (without N₂O) to assess the contribution of the initial trace O₂ present in the culture bottles to biomass increase. The increase in biomass as OD₆₀₀ by the trace O₂ in the control cultures was negligible when compared to cultures growing with N₂O as the sole electron acceptor (see Fig. 2C, F, I, L). Positive control experiments with methanol (30 mM) and O₂ (5%, v/v) as the electron donor and electron acceptor, respectively, were conducted for each strain. The concentrations of H₂, O₂, N₂O, NO₃⁻, and NO₂⁻ were monitored at intervals during incubations (described in Analytical methods). Cell growth was also evaluated using optical density measurements (λ = 600 nm), direct microscopic cell counts, and real-time quantification of 16 S rRNA gene abundance (described in Analytical methods). All growth experiments were performed in triplicates unless otherwise stated.

Next, we checked the anoxic growth of Methylocella strains on NO₃⁻ (2 to 4 mM KNO₃) as the terminal electron acceptor instead of N₂O. Methanol (30 mM) was used as the sole electron donor and 2 mM NH₄⁺ was used as the N-source. To compare the effect of electron donors on NO₃⁻ and NO₂⁻ reduction, Methylocella strains were also anoxically grown in LSM medium containing a C-C substrate, pyruvate (10 mM). Cells of strain T4 were grown under O₂-replete (O₂; 21%, v/v) or anoxic conditions (O₂; 0%, v/v, N₂O; 5%, v/v) for the NO₂⁻ toxicity test (triplicates) with varying NO₂⁻ (KNO₂) concentrations (0, 0.01, 0.03, 0.1, 0.3, and 1 mM).

Analytical methods

A YL 6100 gas chromatograph (YL Instrument Co., Anyang, South Korea) with a flame ionization detector (FID) and a thermal conductivity detector (TCD) was used to analyze the mixing ratios of CH₄, N₂O, and H₂ in the headspace of the sealed bottles used to cultivate the Methylocella and Methylacidiphilum strains. Using a Hamilton glass syringe, 100 µL of the sealed bottle headspaces were injected into a gas chromatograph equipped with MolSieve 5 A column (3Ft, 1/8, 2 mm, 60/80 SST, Agilent Technologies, Inc., CA, USA; for separating H₂, O₂, and N₂O) and Haysep N column (7Ft, 1/8, 2 mm, 60/80 SST, Agilent Technologies, Inc., CA, USA; for separating CO₂ and CH₄) to determine the gases present. Helium was used as the carrier gas, with a flow rate of 15 mL·min⁻¹. By utilizing pure gases of known concentrations, a calibration curve of the gases used as substrates was generated. The bottles were fitted with contactless trace range oxygen sensor spots (TROXSP5, PyroScience, Germany) calibrated at 0% and ambient air (21% oxygen), and a FireSting-Pro multi-analyte meter (FSPRO-4, PyroScience, Germany) was used to measure the O₂ concentration in the sealed bottles. Acidic Griess reagent and VCl₂/Griess reagent were used for photometric quantification of NO₂⁻ and NO₃⁻ concentrations¹²⁹, respectively, using a SpectraMax M2 microplate reader (Molecular Devices, USA). Cell growth was assessed by measuring changes in OD₆₀₀ using a spectrophotometer (Optizen 2120UV, Mecasys Co., Daejeon, Korea). Real-time quantification of the 16 S rRNA gene was performed using the 518 F/786 R primer set¹³⁰. The total cell number was determined by counting cells stained with DAPI (4,6-diamidino-2-phenylindole) using an epifluorescence microscope (AxioScope.A1; Carl Zeiss, Oberkochen, Germany).

Kinetic analysis using microrespirometry (MR)

For kinetic analysis using microrespirometry (MR), Methylocella tundrae T4 cells were grown under three different O₂ conditions: O₂-replete (CH₃OH + O₂), suboxic (CH₄ + O₂ + N₂O), and anoxic (CH₃OH + N₂O). Methylacidiphilum caldifontis IT6 cells were grown under O₂-replete (CH₃OH + O₂) and anoxic (CH₃OH + N₂O) conditions. The O₂-replete growth conditions included ambient air (21% O₂, v/v) and CH₃OH (30 mM) as the sole electron donor. The suboxic cell cultures were grown under a condition that included CH₄ (5%, v/v) as the sole electron donor and O₂ (0.5%, v/v) with N₂O (1%, v/v) as terminal electron acceptors. O₂ (0.5%, v/v) was resupplied intermittently before its depletion. Anaerobically grown cells were cultured in bottles containing 30 mM CH₃OH as the sole electron donor and 5% (v/v) N₂O as the terminal electron acceptor. The cultures were monitored daily and harvested as soon as active consumption of electron donors and acceptors was detected. After being collected by centrifugation (5000 × g, 30 min, 25 °C), the cells were washed twice with substrate- and N-source-free MES-buffered LSM (20 mM MES; pH 5.5) or H₂SO₄-buffered LSM (4 mM H₂SO₄; pH 2.0) and then resuspended in 20 mL of the same media without electron donors and acceptors. In the cultures grown under anoxic and suboxic conditions, the cell suspensions were transferred to sealed 20-mL bottles and flushed with nitrogen gas (N₂, purity >99.999%) before use. The cell suspensions were dispensed into a double-port MR chamber (no headspace) with a capacity of 5 or 10 mL outfitted with O₂ and N₂O-detecting microsensors, two MR injection lids, and two glass-coated stir bars. Kinetics and stoichiometry of N₂O and O₂ reduction coupled to CH₃OH oxidation were estimated using anoxic CH₃OH + N₂O- and oxic CH₃OH + O₂-grown cells, respectively. Anoxic CH₃OH + N₂O-grown cells were used to test CH₃OH-dependent O₂ and N₂O uptake by strains IT6 (starting OD₆₀₀ = 0.96) and T4 (starting OD₆₀₀ = 0.79). The effect of O₂ to N₂OR activities of strains T4 and IT6 was determined by spiking varying O₂ to the N₂O respiring cells. In a 5-mL MR chamber, suboxic CH₄ + O₂ + N₂O-grown cells of strain T4 (starting OD₆₀₀ = 1.0) were used to test the CH₄-dependent simultaneous respiration of O₂ and N₂O.

All MR experiments were performed in a recirculating water bath at 27 °C and 50 °C for strains T4 and IT6, respectively. A 10-µL or 50-µL syringe (Hamilton, Reno, USA) fitted with a 26 G needle was used to inject the substrate (CH₄, CH₃OH, N₂O, or O₂) into the chamber via an injection port. Concentrations of O₂ and N₂O were measured using an OX-MR oxygen microsensor (OX-MR-202142, Unisense, Aarhus, Denmark) and a N₂O-MR sensor (N2O-MR-303088, Unisense), respectively. The detection limits of the OX-MR and N₂O-MR microsensors are 0.3 µM O₂ and 0.1 µM N₂O, respectively. The OX-MR and N₂O-MR microsensors were directly plugged into a microsensor multimeter before being polarized for more than a day and calibrated according to the manufacturer’s instructions. All data from the microsensor multimeter was logged onto a laptop using SensorTrace Suite software (v.3.3.0, Unisense). Anoxically prepared aliquots of N₂O, CH₄, and CH₃OH were injected into the MR chamber via the injection port with a 10-µL syringe (Hamilton, Reno, USA). Anoxic substrate-free LSM media (at pH 2.0 and 5.5) were prepared by sparging the solutions with N₂ gas for 1 h before use. Anoxic saturated-aqueous CH₄ and N₂O solutions were made in capped 160-mL bottles containing 100 mL of LSM medium and pressurized with CH₄ or N₂O (1, 2, or 3 atm; 100%, v/v). Saturated-aqueous O₂ solutions were prepared in capped 160-mL bottles containing 100 mL of LSM medium and pressurized with O₂ (1, 2, and 3 atm; 100%, v/v).

Growth based on CH₄ oxidation coupled with co-respiration of O₂ and N₂O

Suboxic cultivations were carried out to investigate the growth of Methylocella tundrae T4 by oxidizing methane with simultaneous respiration of O₂ and N₂O. The experiments were conducted in N₂-flushed 2-liter sealed bottles containing 60 mL of LSM medium with 2 mM NH₄⁺ as the N-source. The headspace of the bottles was composed of CH₄ (5%, v/v), O₂ (0.5%, v/v), N₂O (1%, v/v), and CO₂ (5%, v/v) and supplemented with additional O₂ (~0.5%, v/v) before its depletion. The headspace gas (CH₄, N₂O, and O₂) mixing ratios were monitored at intervals during incubations as described above in Analytical methods. To investigate the growth benefits of cells of strain T4 respiring N₂O in tandem with O₂ during CH₄ oxidation, an O₂-replete culture was included for comparison (triplicates). The apparent increase in cell densities of both growth conditions was compared using OD₆₀₀ measurements.

Transcriptome analysis

Cells of strains T4 and IT6 were cultured in 60 mL of LSM medium at pH 5.5 and pH 2.0 in sealed 2-liter bottles (4 or 5 replicates) for transcriptome analyses. Strain T4 cells were cultured under three different O₂ levels, with the first setting being O₂-replete (CH₄ + O₂ and CH₃OH + O₂), the second being suboxic (CH₄ + O₂ + N₂O), and the third being anoxic (CH₃OH + N₂O). Strain IT6 was cultivated in O₂-replete CH₃OH + O₂ and anoxic CH₃OH + N₂O conditions. Cells were grown anaerobically in bottles containing 30 mM CH₃OH as the sole electron donor and 5% N₂O as the terminal electron acceptor. The O₂-replete growth conditions were made up of ambient air (21% O₂, v/v) with CH₄ (5%, v/v) or CH₃OH (30 mM) serving as the sole electron donor. The suboxic growth conditions were made up of a mixture of CH₄ (5% v/v) as the sole electron donor and O₂ (0.5% v/v) and N₂O (1% v/v) as terminal electron acceptors. Before the depletion of O₂, additional O₂ was resupplied intermittently at a mixing ratio of 0.5% (v/v). Contactless trace-range oxygen sensor spots (TROXSP5) were installed into the culture bottles to monitor O₂ concentration.

The cells were harvested during the mid-exponential phase by centrifugation at 5000 × g for 10 min at 25 °C. Total RNA was extracted from the cells in four replicates using the AllPrep DNA/RNA Mini Kit (Qiagen) according to the manufacturer’s protocol. RNA quality was checked with the Agilent 2100 Expert Bioanalyzer (Agilent), and cDNA libraries were prepared from the RNA samples using the Nugen Universal Prokaryotic RNA-Seq Library Preparation Kit. The cDNA libraries were sequenced using NovaSeq6000 (Illumina) at LabGenomics (Seongnam, Korea). Read quality was evaluated with FastQC (v0.11.8)¹³¹. Trimmomatic (v0.36)¹³² was used to trim reads with the options: SLIDINGWINDOW:4:15 LEADING:3 TRAILING:3 MINLEN:38 HEADCROP:13. Reads mapped to strains T4 and IT6 rRNA sequences were removed with SortMeRNA (v4.3.6)¹³³. The remaining reads were aligned to the genomes of strains T4 and IT6 using Bowtie2 (v2.4.4)¹³⁴, and the reads mapped to each gene were counted using HTSeq (v0.12.3)¹³⁵. Expression values are presented as transcripts per kilobase million (TPM). The statistical analysis of differentially expressed genes was performed using the DESeq2 package in R (v4.3.2). A two-sided Wald test was used to calculate the p values, and multiple-comparison adjustments were made using the Benjamini-Hochberg method by default in DESeq2 (v1.40.2).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.