Archaea catalyze iron-dependent anaerobic oxidation of methane

Significance

Carbon and nitrogen cycles have been altered dramatically by human activities. Methane-producing (methanogenic) and methane-consuming (methanotrophic) microorganisms control the emission of methane, one of the most potent greenhouse gases, to the atmosphere. Earlier studies identified methanotrophic microorganisms that use methane as an electron donor and oxygen, sulfate, nitrite, and nitrate as electron acceptors. Previous research showed that methanotrophy coupled to the reduction of oxidized metals could be important in the environment. In the current paper, we identified archaea of the order Methanosarcinales, related to “Candidatus Methanoperedens nitroreducens,” which couple the reduction of environmentally relevant particulate forms of iron and manganese to methane oxidation, filling one of the remaining lacunas in anaerobic methane oxidation.

Abstract

Anaerobic oxidation of methane (AOM) is crucial for controlling the emission of this potent greenhouse gas to the atmosphere. Nitrite-, nitrate-, and sulfate-dependent methane oxidation is well-documented, but AOM coupled to the reduction of oxidized metals has so far been demonstrated only in environmental samples. Here, using a freshwater enrichment culture, we show that archaea of the order Methanosarcinales, related to “Candidatus Methanoperedens nitroreducens,” couple the reduction of environmentally relevant forms of Fe³⁺ and Mn⁴⁺ to the oxidation of methane. We obtained an enrichment culture of these archaea under anaerobic, nitrate-reducing conditions with a continuous supply of methane. Via batch incubations using [¹³C]methane, we demonstrated that soluble ferric iron (Fe³⁺, as Fe-citrate) and nanoparticulate forms of Fe³⁺ and Mn⁴⁺ supported methane-oxidizing activity. CO₂ and ferrous iron (Fe²⁺) were produced in stoichiometric amounts. Our study connects the previous finding of iron-dependent AOM to microorganisms detected in numerous habitats worldwide. Consequently, it enables a better understanding of the interaction between the biogeochemical cycles of iron and methane.

Methane is of interest to scientists and society for many reasons: It is one of the most important fossil fuels, a significant greenhouse gas, as well as an abundant natural hydrocarbon and substrate for microorganisms. However, due to its high activation energy, it can only be used by a few specialized groups of archaea and bacteria. Although aerobic methane-oxidizing bacteria have been known for 110 years (1), it was only at the turn of the last century that organisms performing anaerobic oxidation of methane (AOM) were first identified (2). These organisms couple AOM to sulfate reduction and belong to distinct clades within the methanogens (phylum Euryarchaeota), microorganisms that normally produce methane under anoxic conditions. These anaerobic methanotrophic (ANME) archaea have vast importance in regulating Earth’s climate, as they prevent the escape of the large quantities of methane contained in the oceans’ sediments and gas hydrates (2). Because CO₂ has a 12 times weaker radiative forcing per molecule than methane (3), converting methane-carbon to CO₂, assimilating it into biomass, or precipitating it as carbonates altogether reduce global warming effects considerably (2).

Until a decade ago, sulfate, a major constituent of seawater, was the only known electron acceptor for anaerobic oxidation of methane (2, 4). However, a variety of other environmentally important oxidized compounds—most notably Fe³⁺, Mn⁴⁺, nitrate (NO₃⁻) (5), and nitrite (NO₂⁻) (6)—are thermodynamically more favorable electron acceptors than sulfate. The lack of known methanotrophs capable of coupling AOM to metal oxide reduction in particular represents a fundamental gap in our understanding of the dynamics of AOM in the environment.

For nitrogenous electron acceptors, two microbial players and processes have by now been identified: Methylomirabilis oxyfera, a bacterium of the NC10 phylum, is capable of producing its own oxygen from nitrite, and uses this oxygen to activate methane, analogous to aerobic methanotrophs (6). In addition, archaea of a new group within Methanosarcinales [termed AOM-associated archaea (AAA) (2), sometimes also referred to as ANME-2d], originally coenriched together with M. oxyfera (7), were recently found to couple the oxidation of methane to the reduction of nitrate to nitrite (5). In reaching these insights, laboratory enrichment cultures of the responsible microorganisms were of crucial importance.

Even though rock weathering supplies ample amounts of iron, manganese, and other metal oxides suitable as electron acceptors to freshwater bodies (8), and eventually oceans, evidence for the occurrence of iron- and manganese-dependent methane oxidation in nature is scarce. By geochemical profiling or isotope tracer studies, the coupling of methane oxidation to metal reduction has been demonstrated or suggested in lake sediments (9–11), lake water (12), a contaminated aquifer (13), a mud volcano (14), and marine sediments (15–18). The possible role of a cryptic sulfur cycle supplying sulfate via Fe³⁺-mediated reoxidation of sulfides (19, 20) could often not be excluded in these studies, and none of them identified the microorganisms responsible for the process nor succeeded in obtaining an enrichment culture. Only two studies investigated methane oxidation coupled to Fe³⁺ or Mn⁴⁺ reduction in naturally enriched samples of ANME-2 archaea. Using marine samples, Nauhaus et al. reported that methane oxidation was not coupled to the reduction of ferrihydrite, iron citrate, humic acids, and anthraquinone-2,6-disulfonate (AQDS) (21), whereas recently Scheller et al. showed that marine samples containing ANME-2a, ANME-2c, and other archaea oxidized methane in the presence of chelated iron and could couple methane oxidation activity to AQDS reduction (22). Nevertheless, in nature, metals predominantly occur in particulate forms, which are more difficult to access (15, 16, 19, 23). In short, the microorganisms that can couple the oxidation of methane to the reduction of environmentally relevant species of oxidized metals remain unknown.

Here we demonstrate Fe³⁺- and Mn⁴⁺-dependent methane oxidation in a freshwater enrichment culture of archaeal methanotrophs enriched from the sediment of a Dutch canal (Twentekanaal) with a continuous supply of methane, nitrite, and nitrate (7). A subculture with nitrate as the only electron acceptor in the mineral medium resulted in a microbial community dominated by archaea [ca. 40–50%, belonging to the AAA cluster, 95% similar to Methanoperedens nitroreducens (3), and previously referred to as M. nitroreducens MPEBLZ (24)] and a microorganism closely related to M. oxyfera (ca. 40%) as judged by FISH microscopy (Fig. 1). The enrichment culture coupled methane oxidation [assessed as production of ¹³CO₂, 180 nmol mg_(protein)⁻¹ h⁻¹] to the reduction of nitrate [360 nmol mg_(protein)⁻¹ h⁻¹], with dinitrogen gas (N₂) as the main end product [153 nmol mg_(protein)⁻¹ h⁻¹] (Fig. 2A), similar to the theoretical stoichiometry:

$$5{\text{CH}}_4+8{\text{NO}}_3^{-}+8{\text{H}}^{+}\to 5{\text{CO}}_2+4{\text{N}}_2+14{\text{H}}_2\text{O}.$$ [1]

The discrepancy (5:10 instead of 5:8) is most likely due to carbon assimilation into biomass and the conversion of about 10% of nitrate to ammonium [2] (Fig. 2B):

$${\text{CH}}_4+{\text{NO}}_3^{-}+2{\text{H}}^{+}\to {\text{CO}}_2+{\text{NH}}_4^{+}+2{\text{H}}_2\text{O}.$$ [2]

This observation adds a new aspect to the current understanding of AAA physiology (5): In the absence of ammonium-oxidizing bacteria, ammonium production in our enrichment culture could be measured. Despite M. oxyfera-like bacteria reducing nitrite to N₂ a transient accumulation of nitrite was also often observed, although it was consumed when nitrate became limiting (Fig. 2B). Nitrate reduction in the enrichment culture was largely (>95%) dependent on methane. All genes necessary for nitrate reduction to nitrite and nitrite reduction to ammonium are present in the genome of AAA (5, 24). Together, these results suggest that AAA most likely couples methane oxidation to dissimilatory nitrate reduction to ammonium via nitrite, a trait rarely observed in archaea (25).

Fig. 1.

Fluorescence in situ hybridization of biomass from the enrichment culture of AAA and M. oxyfera-like bacteria. AAA were hybridized with Cy3-labeled AAA-specific probe S-*-AAA-FW-641 (red) (33); M. oxyfera-like bacteria were hybridized with Fluos-labeled NC10 bacteria-specific probe S-*-DBACT-1027-a-A-18 (yellow) and a Cy5-labeled mixture of EUB I to III and V (blue; overlay appears green). Other bacteria only hybridizing with EUB I to III and V appear blue. (Scale bar, 5 µm.)

Fig. 2.

(A) Stoichiometric coupling of [¹⁵N]nitrate reduction (open squares) to ³⁰N₂ (open triangles) and ¹³CO₂ (filled circles) production in the presence of ¹³CH₄ in batch incubations. Data of CO₂ and N₂ are the average of four replicate incubations measured twice each; nitrate values represent two replicates. Each incubation contained 77.0 ± 3.6 μg protein. Error bars indicate SD. (B) Transient formation of nitrite (open diamonds) and accumulation of ammonium (open circles) after repeated pulse feeding (arrows) of nitrate (open squares) in the continuous culture reactor with constant methane supply. Note the different scale for nitrite.

We incubated this enrichment culture with AAA as the only archaea with a range of oxidized metal compounds. In the absence of an electron acceptor in batch incubations, methane oxidation did not occur, but activity began immediately after addition of ferric citrate (Fig. 3A). Recently, using a sample that contained ANME-2a and ANME-2c, Scheller et al. reported ¹³CO₂ production from ¹³CH₄ oxidation in the presence of ferric citrate (22). Due to the notorious difficulties associated with measuring dissolved ferrous (Fe²⁺) iron in complex media or natural samples, the stoichiometric relationship between the reduction of Fe³⁺ and oxidation of CH₄ was not shown (22). Here, in a setup with destructive sampling, where total CO₂ and Fe²⁺ could both be measured, we observed a near-perfect fit to the theoretically expected stoichiometry of 1:8 (Fig. 3B):

$${\text{CH}}_4+8{\text{Fe}}^{3+}+2{\text{H}}_2\text{O}\to {\text{CO}}_2+8{\text{Fe}}^{2+}+8{\text{H}}^{+}\hspace{1em}\left(\mathrm{\Delta }{\text{G}}^{\text{0}\mathrm{\prime }}=-454\hspace{0.17em}\text{kJ}\hspace{0.17em}{\text{mol}}^{-1}\hspace{0.17em}{\text{CH}}_4\right).$$ [3]

In nature, iron reduction is often limited by the bioavailability of iron oxides, and soluble or complexed Fe³⁺ is scarce (19, 23). Therefore, we also incubated the enrichment culture with more environmentally relevant forms of oxidized metals: nanoparticulate ferrihydrite (Fe³⁺) and birnessite (Mn⁴⁺). These were suitable as electron acceptors for the oxidation of methane to CO₂ by the enrichment culture, albeit at lower rates (Fig. 3C): 6 and 21 nmol CO₂ mg_(protein)⁻¹ h⁻¹, compared with 80 and 180 nmol CO₂ mg_(protein)⁻¹ h⁻¹, with soluble Fe³⁺ (ferric citrate) and nitrate as electron acceptors, respectively.

Fig. 3.

Metal-dependent oxidation of ¹³C-labeled CH₄ to ¹³CO₂ by an enrichment culture of AAA. (A) Fe³⁺-citrate addition at the start of the experiment and after 150 h (arrow) stimulates methane oxidation. Control incubations of the enrichment culture without any addition, sulfate or citrate only, as well as an archaea-free enrichment culture of M. oxyfera (34) incubated with Fe³⁺-citrate were not active. Data are the average of two incubations containing 1 ± 0.1 mg protein each. (B) Stoichiometric coupling of CH₄ oxidation to Fe²⁺ production. Methane oxidation was deduced from ¹³CO₂ production; data points represent the average of four incubations of 77 ± 3.6 µg protein each. (C) Stimulation of ¹³CO₂ production from ¹³CH₄ by the addition of nanoparticulate suspensions of ferrihydrite (Fe³⁺) and birnessite (Mn⁴⁺). Protein content in these experiments was 77.0 ± 3.6 µg per vial. Error bars indicate SD.

Negative controls with ambient sulfate concentrations (∼15 µM) or added sulfate (1 mM) (Fig. 3A) excluded the possibility of a cryptic sulfur cycle (19, 20) as an indirect route to (re)generate electron acceptors (sulfate). Such a mechanism would also not fit the observed stoichiometry [3]. In addition, iron reduction was dependent on the presence of methane: In controls without methane (<15 ppm), the enrichment culture did not produce any Fe²⁺ (above the background of ∼2 µM remaining from 3.75 µM FeSO₄ supplied as a micronutrient), the end product of Fe³⁺ reduction in incubations with methane. Finally, an enrichment culture containing M. oxyfera as the only methanotroph (26) did not oxidize methane coupled to the reduction of Fe³⁺. Taken together, these findings strongly indicated that the AAA in the enrichment culture directly coupled methane oxidation to Fe³⁺ and Mn⁴⁺ reduction.

AAA and an M. oxyfera-like microorganism dominated our enrichment culture (Fig. 1). The analyses of the 16S rRNA gene reads of the metagenome (DDBJ/EMBL/GenBank accession no. LKCM01000000) revealed that other archaea or known iron-reducing microorganisms [Geobacter or Shewanella (27)] were absent from the enrichment culture; this was expected, as the culture had never been exposed to iron exceeding the above-mentioned 3.75 µM FeSO₄ in the medium. Furthermore, the immediate onset of Fe³⁺ reduction in batch incubations and its linear course excluded the growth of an iron-reducing subpopulation interacting with a methanotroph being responsible for AOM in our culture.

Rather, the analysis of the partially reconstructed AAA genome (24) supported that the AAA alone were able to couple methane oxidation to the reduction of different electron acceptors. All genes of the reverse methanogenesis pathway (28, 29) were present in the enriched strain of AAA, in complete agreement with its closest relative, M. nitroreducens (5, 24). Nitrate reduction could be accomplished by two candidate enzyme complexes, one encoded by the narGH genes (MPEBLZ_02035 and 2036), closely related to the ones of M. nitroreducens (81 and 83% identity, respectively), and the other by an nxrAB-related operon (MPEBLZ_01127 and 1129) absent from M. nitroreducens and similar to that in nitrite-oxidizing bacteria (68/78% similarity to Nitrolancea hollandica and 68/69% to Nitrobacter winogradskyi). The further reduction of nitrite to N₂ gas could be accomplished by the M. oxyfera-like microorganism as previously demonstrated (6) or, alternatively, ammonium could be formed by a NrfAH-like protein complex (MPEBLZ_01114 and 1115) transcribed by the AAA (24). As the likely genetic basis for Fe³⁺ reduction, we identified a wealth of genes encoding heme c-containing proteins (c-type cytochromes), which are generally recognized to function as electron shuttles from the cell to soluble or solid electron acceptors such as iron minerals (27, 30). The AAA in our enrichment culture encode numerous (41 species) multiheme c-type cytochromes (Table S1). This number is much higher than for methanogens, of which only Methanosarcinales encodes few multiheme cytochromes, but similar to M. nitroreducens (38 species) (5), the sulfate-dependent ANME-1 (11 species) (28), ANME-2a (16 species) (29, 31), as well as the iron-reducing archaea Ferroglobus placidus (22 species) and Geoglobus (15 species) (32).

Table S1.

Multi(>2)-heme c-type cytochromes predicted from the contigs assigned to AAA in the metagenome of the enrichment culture

Locus tag	Protein length, no. of amino acids	CxxCH and other motifs	No. of transmembrane helices	Best hit identifier	Best hit organism	Best hit identity
MPEBLZ_00008	358	4	0	gi|851163073|ref|WP_048092705.1|	Candidatus Methanoperedens nitroreducens	44.5
MPEBLZ_00010	190	6	1	gi|851163076|ref|WP_048092707.1|	Ca. M. nitroreducens	60.5
MPEBLZ_00012	319	4	1	gi|851163096|ref|WP_048092715.1|	Ca. M. nitroreducens	61.2
MPEBLZ_00015	253	7	1	gi|851163104|ref|WP_048092719.1|	Ca. M. nitroreducens	60.8
MPEBLZ_00016	343	4	0	gi|851163073|ref|WP_048092705.1|	Ca. M. nitroreducens	45.3
MPEBLZ_00238	317	11	2	gi|630830727|gb|KCZ72437.1|	Ca. M. nitroreducens	60.6
MPEBLZ_00452	577	9	0	gi|851163118|ref|WP_048092727.1|	Ca. M. nitroreducens	50.4
MPEBLZ_00526	864	16	1	gi|851162463|ref|WP_048092373.1|	Ca. M. nitroreducens	37.9
MPEBLZ_00803	1,846	7*	1	gi|851163104|ref|WP_048092719.1|	Ca. M. nitroreducens	54.5
MPEBLZ_00816	316	6†	0	gi|851164534|ref|WP_048093459.1|	Ca. M. nitroreducens	71.9
MPEBLZ_00817	309	6†	2	gi|851164537|ref|WP_048093461.1|	Ca. M. nitroreducens	63.2
MPEBLZ_01070	678	9‡	0	gi|851160641|ref|WP_048091456.1|	Ca. M. nitroreducens	60.6
MPEBLZ_01114	437	5*	1	gi|851165424|ref|WP_048093932.1|	Ca. M. nitroreducens	74.5
MPEBLZ_01115	167	4	0	gi|630828680|gb|KCZ70396.1|	Ca. M. nitroreducens	61.1
MPEBLZ_01126	288	2	2	gi|490648195|ref|WP_004513190.1|	Geobacter metallireducens	43.4
MPEBLZ_01329	202	5	1	gi|851154723|ref|WP_048088734.1|	Ca. M. nitroreducens	57.4
MPEBLZ_01361	1,137	16	1	gi|851162463|ref|WP_048092373.1|	Ca. M. nitroreducens	40.3
MPEBLZ_01741	696	21	0	gi|851163401|ref|WP_048092876.1|	Geoglobus acetivorans	28.6
MPEBLZ_01742	561	11	0	gi|851155354|ref|WP_048089050.1|	Ca. M. nitroreducens	31.9
MPEBLZ_01743	914	18	1	gi|851149563|ref|WP_048086421.1|	Ferroglobus placidus	30.0
MPEBLZ_01744	606	6†,‡	1	gi|851163118|ref|WP_048092727.1|	Ca. M. nitroreducens	59.3
MPEBLZ_01877	828	16‡	1	gi|851163401|ref|WP_048092876.1|	G. acetivorans	38.2
MPEBLZ_01950	432	8	0	gi|851165403|ref|WP_048093919.1|	Ca. M. nitroreducens	69.1
MPEBLZ_02010	573	6	0	gi|851163118|ref|WP_048092727.1|	Ca. M. nitroreducens	51.1
MPEBLZ_02042	403	5	2	gi|851156284|ref|WP_048089356.1|	Ca. M. nitroreducens	67.5
MPEBLZ_02500	373	7	1	gi|851153793|ref|WP_048088239.1|	Ca. M. nitroreducens	74.2
MPEBLZ_02501	179	5	0	gi|851153793|ref|WP_048088239.1|	Ca. M. nitroreducens	67.1
MPEBLZ_02503	469	8	0	gi|851153795|ref|WP_048088240.1|	Ca. M. nitroreducens	69.1
MPEBLZ_02608	209	5	0	gi|851160641|ref|WP_048091456.1|	Ca. M. nitroreducens	51.5
MPEBLZ_02874	305	7‡	2	gi|630831567|gb|KCZ73276.1|	Ca. M. nitroreducens	41.3
MPEBLZ_02879	267	7	2	gi|851162380|ref|WP_048092326.1|	Ca. M. nitroreducens	54.6
MPEBLZ_03194	318	6†	1	gi|630831820|gb|KCZ73529.1|	Ca. M. nitroreducens	71.2
MPEBLZ_03195	149	4‡	1	gi|630831821|gb|KCZ73530.1|	Ca. M. nitroreducens	70.5
MPEBLZ_03279	655	2	1	gi|635652775|ref|WP_024299646.1|	Methylosarcina lacus	68.3
MPEBLZ_03539	237	5‡	1	gi|630831567|gb|KCZ73276.1|	Ca. M. nitroreducens	78.1
MPEBLZ_04272	521	12*	1	gi|851149560|ref|WP_048086419.1|	F. placidus	27.5
MPEBLZ_04274	227	4	1	gi|851166986|ref|WP_048094540.1|	Geoglobus ahangari	35.6
MPEBLZ_04276	224	5	1	gi|851163417|ref|WP_048092885.1|	G. acetivorans	27.7
MPEBLZ_04299	324	6†	2	gi|851155273|ref|WP_048089031.1|	Ca. M. nitroreducens	61.9
MPEBLZ_04300	364	13†	1	gi|851155276|ref|WP_048089032.1|	Ca. M. nitroreducens	73.0
MPEBLZ_04301	246	2	1	gi|258591543|emb|CBE67844.1|	Ca. Methylomirabilis oxyfera	26.2

Prediction of multiheme cytochrome c proteins was done using a perl script (prot_motif_search.pl) available at https://github.com/dspeth followed by BLAST analysis of conservation of the heme-binding motifs in related sequences. Prediction of transmembrane helices was performed using TMHMM (www.cbs.dtu.dk/services/TMHMM/) (49). Each superscript denotes one divergent heme binding motif.

^*CxxCK.

^†CxxxCH.

^‡CxxxxCH.

This genomic blueprint suggests that AAA, and possibly also the cytochrome-rich ANME-1 (28) and ANME-2a (29, 31) archaea, known for sulfate-dependent methane oxidation, could be versatile methanotrophs that are able to switch between different electron acceptors, depending on environmental conditions (17). It is also conceivable that ANME-2a and ANME-2c archaea, which were shown to couple the reduction of solubilized iron species (iron citrate) to methane oxidation recently (22), could also be capable of converting particulate Fe³⁺ or Mn⁴⁺. Compared with M. oxyfera, the ecological niche of AAA probably extends into more reduced environments, as these microorganisms can also use oxidized metals. However, AAA can form stable cocultures with microorganisms related to the nitrite-dependent methanotroph M. oxyfera, as previously reported (7). While AAA supply nitrite, which is often a limiting electron acceptor in nature, to M. oxyfera-like bacteria, or as in the culture described by ref. 5, to anaerobic ammonium-oxidizing (anammox) bacteria, the bacterial partner in these cocultures keeps the concentration of the potentially toxic nitrite low. A bioreactor containing all three microorganisms could be applied in wastewater treatment to simultaneously convert ammonium, methane, and oxidized nitrogen species to harmless N₂ and CO₂.

Molecular surveys of 16S rRNA or mcrA genes in the environment have often retrieved AAA-affiliated sequences in ecosystems poor in nitrate but rich in oxidized metals (14, 33) (Fig. S1), so far without an ecophysiological explanation. Our results show that the AAA directly coupled the reduction of these environmentally relevant, particulate oxidized metal species to methane oxidation and could play an as-yet largely overlooked role in decreasing methane emissions from anoxic environments to the atmosphere. Given the 8:1 stoichiometry of Fe³⁺ reduced to methane oxidized [3], their contribution to iron cycling, thereby impacting also the geochemistry of phosphorus and sulfur, may be even more important than their impact on the carbon cycle (11, 16, 18, 20). The enrichment of at least one of the microbial players not only now enables a targeted investigation of iron-dependent methane oxidation in physiological and molecular detail, but may also shed light on the long-standing discussion about Fe²⁺-producing processes on early Earth, when AAA-related organisms may have thrived under the methane-rich atmosphere in the ferruginous Archean oceans (12, 18, 30).

Fig. S1.

(A) Phylogenetic position of AOM-associated archaea (AAA) (red) among other archaeal methanotrophs (blue) and methanogens (gray) based on sequences of 16S rRNA gene sequences obtained from GenBank. (B) Detailed phylogenetic tree of 16S rRNA genes of the AAA group within Methanosarcinales. Sequences from enrichment cultures are highlighted in bold letters; sequences from sites or cultures with demonstrated AOM activity are shown in blue; those from particularly metal-rich or iron-reducing environments are shown in orange. The evolutionary analysis was conducted with the maximum-likelihood algorithm implemented in MEGA6 (39). There were a total of 949 positions in the final dataset. Bootstrap support >90 is shown at the branches. (Scale bar, number of substitutions per site.)

Methods

Enrichment of the Coculture of AAA and the M. oxyfera-like Bacteria.

The enrichment of the biomass used here was described in brief previously (24). Biomass from the settler of an M. oxyfera-dominated reactor (26) containing about 10% (of the total microbial community) AAA was used to inoculate a continuous culture reactor (2-L working volume) to further enrich AAA. The reactor was run at room temperature with a pH of 7.5 (±0.15) and stirred gently with a magnetic bar (100 rpm). Initially, the reactor was coupled to the effluent of an M. oxyfera reactor, receiving once per day the methane-saturated effluent containing low nitrite (0 to 50 µM) and high nitrate (2 to 3 mM) concentrations. Along with the growth of biomass in the reactor, nitrite and nitrate consumption rates increased. Therefore, increasing amounts of nitrate were manually injected into the reactor (from 100 to 1,000 µM final concentration) when nitrite and nitrate were below 100 µM. After about 1 year of incubation, the enrichment culture was uncoupled from the M. oxyfera reactor. Methane was provided by constant sparging with CH₄:CO₂ (95:5, vol:vol; ca. 10 mL⋅min⁻¹). Nitrate (0.5 to 1.5 mM final concentration) as the sole electron acceptor was injected daily into the reactor for over 2 years. Supernatant (ca. 0.8 L) was removed and replenished with fresh anoxic mineral medium (as previously described by ref. 34, omitting nitrate and nitrite) after settling of biomass every 2 weeks.

Nitrate Reduction and Nitrite and Ammonium Production by the Enrichment Culture.

Conversion of nitrogenous compounds was measured in the continuous culture while sparging the reactor with CH₄:CO₂ (95:5, vol:vol; ca. 10 mL⋅min⁻¹). After nitrate depletion (to <10 µM), fresh nitrate was added to a final concentration of ∼1 mM on 3 consecutive days. Liquid samples (1.5 mL) were taken every 3 to 4 h from the reactor and centrifuged at 14,000 × g, and supernatants were stored at 4 °C before nitrate, nitrite, and ammonium analysis.

Activity Tests with Different Electron Acceptors.

To test the effect of various electron acceptors, the enrichment culture was left under continuous CH₄:CO₂ (95:5, vol:vol) sparging but without nitrate supply until residual nitrate was depleted. Twenty milliliters of culture liquid (protein content ca. 1.0 mg) was taken and anaerobically transferred to a sealed, evacuated serum bottle (50 mL). After removal of unlabeled methane by flushing with Ar:CO₂ (90:10, vol:vol), 10 mL ¹³CH₄ (>99.99% ¹³C; Sigma-Aldrich) was injected into the 1-bar Ar:CO₂ headspace. Ferric citrate (pH 6.8) and sodium sulfate were added from anoxic stock solutions to final concentrations of 2 and 1 mM, respectively. Particulate iron and manganese were added as ferrihydrite or birnessite nanoparticles from anoxic suspensions. The incubation of nitrite-depleted M. oxyfera biomass (34) with ferric citrate was prepared in the same manner as described above. To control for the effect of electron acceptor addition, additional incubations were prepared accordingly, but either without any additions or with sodium citrate only (2 mM final concentration). All serum bottles were incubated at room temperature and shaken gently at 50 rpm. Intervals between headspace sampling for concentration and isotopic composition of CH₄ and CO₂ were chosen according to the observed activity.

Stoichiometry Determination in Activity Tests with Destructive Sampling.

To be able to precisely determine the stoichiometry of methane oxidation coupled to nitrate or Fe³⁺ reduction and minimize the biomass needed per experiment, the experimental setup was miniaturized and incubations were performed in 3-mL Exetainer vials (Labco). Before distribution of biomass over the vials, residual methane was stripped from nitrate-depleted enrichment culture by stirring and concomitant flushing with He:CO₂ (90:10, vol:vol) in a previously evacuated and sealed serum bottle. The biomass was then transferred to an anaerobic chamber with an Ar:H₂ (95:5, vol:vol) atmosphere (O₂ in the anaerobic chamber was removed by passing of the gas over a Pd catalyst resulting in a residual O₂ concentration of ≤0.2 ppm) and amended with an electron acceptor (nitrate or ferric citrate). The biomass was then equally distributed to aliquots of 1.5 mL in Exetainers, resulting in a protein content of 77.0 ± 3.6 µg per vial. Exetainers were sealed and 1 mL of ¹³CH₄:CO₂ (80:20, vol:vol) was injected into the headspace to provide an electron donor, to buffer pH, and create overpressure. The vials were incubated at room temperature in an overhead shaker inside the anaerobic cabinet. For incubations with ferric citrate, three vials were harvested every 3 h and acidified with 0.5 mL anaerobic HCl solution (final concentration 0.5 mM) and kept in the anaerobic tent for at least 0.5 h before GC/MS measurement to allow CO₂ to equilibrate with the headspace. After gas measurements, biomass from each vial was transferred to a 1.5-mL tube and centrifuged, and the supernatant was stored at 4 °C until Fe²⁺ determination. For incubations with nitrate, four vials were harvested every 2 h, two of which were acidified with 0.5 mL anaerobic HCl solution. CO₂ production in all four vials’ headspace was measured. The biomass from the nonacidified vials was transferred to a 1.5-mL cup and centrifuged, and supernatant was stored at 4 °C until nitrate determination by HPLC.

Analytical Procedures and Synthesis of Reagents.

Quantity and isotopic composition of CH₄, CO₂, and N₂ in the headspace were measured by GC/MS (GC 6890 coupled to MSD 5975C; Agilent) as previously described (34). Total amounts indicated in the incubations take into account the dissolved fraction of the respective gases. Nitrate and nitrite were measured after a reduction step to NO with a Sievers Nitric Oxide Analyzer (NOA 280i; Analytix). Nitrate and nitrite were both reduced with a saturated solution of vanadium (III) chloride in hydrochloric acid (1 M) at 90 °C. Nitrite was colorimetrically measured by the Griess reaction as previously described (35). Nitrate concentrations were obtained by subtracting the nitrite concentration from the sum of nitrate and nitrite. Ammonium was measured colorimetrically at 623 nm after a 30-min reaction of 100-µL samples with alkaline phenol catalyzed by sodium nitroprusside at 37 °C (36). For the miniaturized setup (Exetainer incubations), nitrate was quantified with HPLC using a conductivity detector (26). Fe²⁺ was measured using the ferrozine method (37).

Ferrihydrite and birnessite were synthesized according to protocols described previously (23, 38). Briefly, for ferrihydrite, 21 g of ferric citrate (Sigma) was dissolved in 300 mL MilliQ water by vigorous stirring while being kept at 80 °C. After cooling the ferric citrate solution to room temperature, the pH was quickly adjusted to 8.0 with 10 M NaOH while stirring rigorously for 30 min. The supernatant of the resulting suspension was discarded after centrifuging at 8,000 × g at 4 °C for 1 h, and the pellet was resuspended in 500 mL MilliQ water. After four washing cycles, the pellet was finally resuspended in a small volume of MilliQ water. For birnessite, concentrated HCl (46%) was slowly added to a boiling and rigorously stirring 0.4 M potassium permanganate solution to a final concentration of 0.8 M in a fume hood. After further boiling and stirring for 10 min, the suspension was cooled to room temperature and centrifuged. Washing and resuspension were performed as described for ferrihydrite. Approximate iron and manganese concentrations in the suspensions were determined by measuring the dry weight of an aliquot.

Fluorescence in Situ Hybridization.

Biomass from the enrichment culture (2 mL) was harvested by centrifugation, fixed, hybridized, and visualized as previously described (26) with additional use of the probe S-*-AAA-FW-641 (33), specifically targeting AAA.