Carbon monoxide oxidation expands the known metabolic capacity in anaerobic methanotrophic consortia

Abstract

Consortia of anaerobic methane-oxidizing archaea (ANME-2) and sulphate-reducing bacteria (SRB) represent globally relevant syntrophic associations capable of growing with minimal amounts of free energy and can persist when methane becomes limiting. Carbon monoxide (CO) has been reported in seep environments and represents a thermodynamically favourable alternative electron donor due to its low reduction potential. Here, we show that environmental ANME-SRB consortia can oxidize CO in the absence of methane, in anoxic microcosm experiments using a combination of stable isotope geochemical tracers, metatranscriptomics, and single cell activity measurements (FISH–nanoSIMS). The oxidation of CO was coupled with sulphate-reduction by syntrophic consortia, and, in the absence of sulphate, through CO₂ reduction to methane by ANME-2. Under these conditions, the production of methane was one ninth the rate of methanotrophy coupled to sulphate-reduction. Paired single cell FISH-nanoSIMS analysis of anabolic activity indicates that CO respiration appears to support cell maintenance rather than active growth, consistent with the observed down-regulation of energy generating complexes in ANME (e.g., mtr, rnf, etc.). The versatile capability of CO oxidation by anaerobic methanotrophic consortia broadens our understanding of carbon cycling in methane seeps and highlights potential mechanisms of resilience by methanotrophic archaea under changing geochemical regimes.

Consortia of anaerobic methane-oxidizing archaea and sulphate-reducing bacteria can grow with minimal amounts of free energy. Here, Guo et al. show that these consortia can oxidize carbon monoxide, which may help their survival in the absence of methane.

Introduction

The anaerobic oxidation of methane (AOM) coupled to sulfate reduction is an important microbial process for controlling methane flux from ocean sediments worldwide^1,2. This process is largely mediated by anaerobic methanotrophic archaea (ANME) living syntrophically in consortia with sulfate-reducing bacteria (SRB)^3–5. Methanotrophic ANME-2 archaea (Ca. Methanogastraceae, Methanocomedens, and Methanomarinus)⁶ in ocean sediments have been shown to be energetically versatile, capable of using terminal electron acceptors besides sulfate, including metal oxides (manganese (birnessite) and iron (ferrihydrite))⁷, and humic acid-like substances⁸. However, the potential for these archaeal methanotrophs to oxidize electron donors other than methane has not been specifically investigated, beyond experiments assessing possible syntrophic intermediates for SRB in the absence of methane⁹.

Members of the ANME-2 within the order Methanosarcinales are evolutionarily related to methylotrophic methanogens such as Methanosarcina⁶. Previous studies showed that Methanosarcina species were able to grow on carbon monoxide (CO) as an energy source^10,11, suggesting the possibility that related members of ANME-2 may also utilize CO. While infrequently measured in geochemical investigations of seep environments, some reports indicate that CO can be produced in sediments, resulting from the thermal decomposition of organic compounds such as humic acids and phenolic compounds or released as a by-product from microbial pathways^12–16. From a thermodynamic perspective, CO represents a favorable electron donor for ANME with a low reduction potential of −558 mV (pH 7.0, CO₂ /CO)¹⁷, and there are several examples of CO-oxidizing microorganisms within diverse environments. This includes anaerobic bacteria such as Moorella thermoacetica^18,19, the purple sulfur bacteria Rhodopseudomonas rubrum^20,21, Carboxydothermus hydrogenoformans²², selected SRB²³, as well as archaea (Candidatus Hydrothermarchaeota) living within the subseafloor crust ²⁴.

Building from this framework and rich metagenomic data revealing the widespread distribution of anaerobic CO dehydrogenases (CODH and cooS) across diverse ANME lineages, we investigated whether environmental methanotrophic ANME-2 archaea and their syntrophic SRB partners can metabolize CO. We tested this hypothesis using anoxic microcosm experiments integrated with geochemical and isotopic analysis, metatranscriptomics, and single-cell resolved fluorescence in situ hybridization and nanoscale secondary ion mass spectrometry (FISH–nanoSIMS).

Results

Both ANME and SRB encode carbon monoxide dehydrogenases

Anaerobic CO dehydrogenases (CODHs) catalyzing the conversion of CO and CO₂ are comprised of two main groups, the CO dehydrogenase/acetyl-CoA synthase complex (cdhA), present in the Wood–Ljungdahl pathway, and the monofunctional anaerobic CO dehydrogenase (cooS)²⁵. To examine the distribution and diversity of CO dehydrogenases in ANME-SRB consortia, we conducted a phylogenetic analysis of CO dehydrogenases within representative high-quality genomes for all described marine ANME groups, including Candidatus Methanophaga (ANME-1), Ca. Methanocomedens (ANME-2a), Ca. Methanomarinus (ANME-2b), Ca. Methanogaster (ANME-2c), and Ca. Methanovorans (ANME-3)⁶; and syntrophic SRB lineages (Ca. Syntrophophila gen. nov., (Seep-SRB1a), Ca. Desulfomellonium gen. nov., (Seep-SRB1g), Ca. Desulfomithrium gen. nov. (Seep-SRB2), and Ca. Desulfofervidus sp. (HotSeep-1 cluster))²⁶, representing a total of 49 ANME and 65 SRB MAGs. Both cdhA and cooS were detected in ANME and SRB lineages, with the exception of ANME-1 representatives that were lacking cooS homologs (Fig. 1, Supplementary Fig. 1). In addition, several MAGs, including ANME-2a CONS7142G09b1 and Seep-SRB1g str. CR10073A, were found to host two copies of cooS. Both of these were verified to occur on contigs with other characterized ANME-2a or Seep-SRB1g genes and are unlikely to be a product of mis-binning (Fig. 1). These observations illustrate the broad distribution of CO dehydrogenases within diverse ANME and SRB lineages. This genomic evidence motivated our deeper physiological study of the potential for CO metabolism by anaerobic methanotrophic consortia.

Fig. 1. The phylogeny of CO dehydrogenase/acetyl-CoA synthase (CODH/ACS) complex alpha subunit (cdhA) and monofunctional anaerobic CO dehydrogenase (cooS) in ANME and syntrophic SRB lineages.

The CODH homologs were retrieved from metagenome assembled genomes reported in previous studies^6,29. All homolog alignments were examined manually after identification with BLAST. We operationally distinguished cdhA from cooS homologs by the following principle: a hit was labeled cdhA if it was part of a CODH/ACS complex cluster (operon), and cooS if not in the CODH/ACS operon. The names marked in bold correspond to groups most closely related to lineages occurring in the seep sediment used in our study. Green shaded sections represent SRB lineages, and those in warmer colors represent the different ANME clades.

ANME-SRB consortia can respire CO and sulfate

To test whether ANME-SRB consortia can respire CO, we performed a series of anaerobic microcosm experiments using deep-sea methane seep sediments collected from the Costa Rica margin (see “Methods”). The sediments were dominated by members of Methanomarinus (ANME-2b; relative abundance 53.63%) along with Methanocomedans ANME-2a-2b (9.12%) and Methanogaster (ANME-2c; 2.75%) groups, and their sulfate-reducing bacterial partners Seep-SRB1 (6.21%) based on 16S rRNA amplicon sequencing (Supplementary Fig. 2, Supplementary Data 4).

We tested the ability to oxidize CO coupled with sulfate using ¹³CO at three different concentrations, 0.1-, 0.4-, and 1.0-bar partial pressure in biological triplicates. A positive control (methane + sulfate; n = 3) and a killed control (n = 1; 0.1-bar ¹³CO + sulfate) were also run in parallel (Supplementary Table 1). Labeled ¹³CO was used to quantify the rate of CO oxidation by quantifying the increase of ¹³C-DIC over time concomitant with the reduction of sulfate to sulfide (HS⁻) over the course of 24 days. These experiments confirmed active CO oxidation alongside the reduction of sulfate; however, CO oxidation rates showed a decrease with increasing CO partial pressures (Fig. 2). At 0.1 bar, CO was oxidized at a rate of 34.97 ± 0.68 µM d⁻¹ cm⁻³_sed, equating to approximately half the rate of methane oxidation in parallel control incubations with ¹³CH₄ (64.63 ± 0.68 µM d⁻¹ cm⁻³_sed; Supplementary Table 2). This rate decreased further from 23.21 ± 4.67 µM d⁻¹ cm⁻³_sed to 12.00 ± 0.07 µM d⁻¹ cm⁻³_sed with headspace CO concentrations of 0.4 and 1.0 bar, respectively (Supplementary Table 2), possibly attributed to CO toxicity at high concentrations as reported in other organisms ^10,27.

Fig. 2. DIC and sulfide production in the incubation experiments amended with methane or CO as the electron donor and sulfate as the electron acceptor.

a The oxidation of ¹³CH₄ or ¹³CO to ¹³C-dissolved inorganic carbon (DIC); b the reduction of sulfate reflected by the production of hydrogen sulfide (HS⁻). Individual data points are shown to illustrate variability among the biological replicates (n = 3).

CO oxidation includes an additional electron sink besides sulfate

An examination of the electron transfer and mass balance for the CO and sulfate redox reaction indicates that there were significantly higher rates of CO oxidation in the incubation experiments than could be explained by sulfate reduction alone. Indeed, the measured rate of CO₂ production (34.97 µM d⁻¹ cm⁻³_sed) was 7 times greater than the predicted rate (5.32 µM d⁻¹ cm⁻³_sed) calculated from sulfide production based on Eq. 1 (Supplementary Table 2). This discrepancy suggests that, under conditions with CO and sulfate, sulfate accounts for only a portion of the electrons from CO oxidation, implying that there must be additional electron acceptor(s) in the sediment microcosm beyond sulfate.

$$4\mathrm{CO}+{\mathrm{SO}}_4^{2-}+{\mathrm{H}}_2\mathrm{O}+3{\mathrm{OH}}^{-}=4{\mathrm{HCO}}_3^{-}+{\mathrm{HS}}^{-}\mathrm{\Delta }G=-83.15\mathrm{kJ}/\mathrm{molCO}$$ 1

Based on this, we identified two likely alternative electron acceptors, CO and CO₂ (present as bicarbonate (HCO₃⁻) in the medium), that were in high enough concentrations to account for this discrepancy. First, CO itself could be reduced by nitrogenases present in ANME and/or SRB^28,29. Electron carrier ferredoxin, being reduced at CO dehydrogenases, could feed the reduction of CO by nitrogenases with an investment of ATP and subsequently result in the production of hydrocarbons (see illustration in Supplementary Fig. 3). Previous work demonstrated that both vanadium- and molybdenum-nitrogenases are able to reduce CO to diverse hydrocarbons including ethylene (C₂H₄), ethane (C₂H₆) and propane (C₃H₈)^30,31. However, we did not detect any of these alkanes when we analyzed our microcosm experiments using GC-MS (data not shown), and we therefore presume that any potential CO reduction by ANME or SRB nitrogenase is minimal to non-existent.

Another possibility is coupling CO with CO₂ (HCO₃⁻) as the electron acceptor, a reaction known to be catalyzed by M. acetivorans. In this study, a number of products were reported, including formate, acetate, and methane¹⁰. Neither formate nor acetate was detected by either NMR or IC in our microcosm experiments (detection limits: formate on NMR, ca. 3 µM; acetate on IC, ca. 5 µM; data not shown). We then focused on methane production. This reaction is predicted to be thermodynamically favorable, with the reduction potential of CO₂/CH₄ (−245 mV at pH 7.0)⁸ higher than that of CO₂/CO (−558 mV, pH 7.0). Microcosm incubations amended with 0.1 bar CO and 5 mM sulfate confirmed that CO continued to be oxidized over time alongside sulfate reduction (Fig. 3a, b). These incubations showed evidence of methane production; however, the concentration did not accumulate above ca. 0.2 µmol in the headspace (Fig. 3c). We hypothesized that this trace methane concentration, lower than predicted, resulted from subsequent oxidation via sulfate-coupled AOM (i.e., cryptic methane cycling)³². To test this idea, we conducted follow-up microcosm experiments amended with ¹³CH₄ and SO₄²⁻, supporting conventional sulfate-coupled AOM, and then spiked in unlabeled CO (¹²CO) at day 40 to track the potential contribution of CO-derived CH₄ above that of ¹³CH₄ to AOM. Newly formed ¹³C- and ¹²C-DIC quantification was used to differentiate the relative proportion of CH₄ oxidation and CO oxidation activity, assuming ¹²C-DIC generation after day 40 was attributed to CO. These results supported the active oxidation of methane by ANME-SRB consortia in the presence of CO, where sulfate remained as the terminal electron acceptor for methane cycling when sulfate and CO were amended (Fig. 4).

Fig. 3. CO oxidation and corresponding sulfide and methane production monitored over time from microcosm incubations amended with sulfate.

a Changes in CO concentrations over time; b corresponding sulfide generation over time; c corresponding methane generation over time. Biological triplicates were used for each experimental condition and shown separately along with a single killed control incubation (× symbol).

Fig. 4. DIC production in microcosm experiments containing sulfate, methane, and CO.

Microcosms were initially amended with ¹³CH₄ and sulfate. ¹²C- and ¹³C-DIC comprised of the total newly formed DIC, where ¹³C-DIC came from ¹³CH₄ oxidation. Unlabeled CO (^naCO, 0.4 bar) was then added at day 40 of the experiment to examine the proportion of DIC sourced from methane vs. CO. A mix of ¹³CH₄ and ^naCO was both present in the headspace after day 40 and served as potential electron donors. Subtracting the ¹³C-DIC from the total concentration of newly formed DIC after day 60 differentiated the proportion of CO oxidation and methane oxidation occurring. This experiment was performed in biological triplicate (mean ± s.d., n = 3). Where not visible, the error bar is smaller than the symbol.

To further investigate the potential for methane production from CO oxidation in the absence of sulfate-coupled AOM, we conducted a second set of microcosm experiments under sulfate-free conditions and amended with CO (0.4 bar partial pressure; Supplementary Table 1). These incubation conditions limit the activity of the syntrophic SRB, minimizing any subsequent oxidation of produced CH₄ through sulfate-coupled AOM. By supplying HCO₃⁻ as the sole added terminal electron acceptor, a continuously increasing amount of methane was produced coupled to CO consumption at a rate of 0.36 µmol CO d⁻¹ cm⁻³_sed (mean, n = 2; Fig. 5). Parallel microcosm experiments without the addition of HCO₃⁻ to the medium were also run. These results showed CO oxidation was also coupled with methane production under these conditions and at comparable rates of 0.35 µmol CO d⁻¹ cm⁻³_sed (mean, n = 3). This suggests that self-generated CO₂ may also serve as the terminal electron acceptor. Most importantly, the ratio of CO oxidation rate to methane production rate calculated from the five replicates was 3.85 ± 1.93, close to the predicted stoichiometry of 4:1 in Eq 2

$$4\mathrm{CO}+12{\mathrm{OH}}^{-}=4{\mathrm{HCO}}_3^{-}+4{\mathrm{H}}_2\mathrm{O}+8{\mathrm{e}}^{-}$$

$${\mathrm{HCO}}_3^{-}+8{\mathrm{e}}^{-}+6{\mathrm{H}}_2\mathrm{O}={\mathrm{CH}}_4+9{\mathrm{OH}}^{-}$$

$$4\mathrm{CO}+3{\mathrm{OH}}^{-}+2{\mathrm{H}}_2\mathrm{O}=3{\mathrm{HCO}}_3^{-}+{\mathrm{CH}}_4\mathrm{\Delta }G=-79.02\mathrm{kJ}/\mathrm{molCO}$$ 2

Fig. 5. CO consumption coupled to methane production over time in microcosms without sulfate and with bicarbonate as the sole added terminal electron acceptor.

Biological duplicates were performed in microcosms with bicarbonate as the sole electron acceptor (solid triangles). Biological triplicates were run without the addition of an external electron acceptor (open triangles). A killed control in duplicate was also run (× symbol). a Changes in CO concentrations over time; b corresponding methane generation over time; c the total DIC concentrations over time. Units were normalized to the total molar amount as µmol. The numbers to the right of each line correspond to individual bottles in this experiment.

To confirm that methane was sourced from CO₂ in these sulfate-free incubations, experiments amended with ¹³CO (≥99 atom% ¹³C) and ¹²C-HCO₃⁻ (100%) were used (incubation set #4; Supplementary Table 1). In these experiments, the detection of ¹³C in the bicarbonate pool is predicted to occur through CO oxidation (Eq. 2 Fig. 5c), and any subsequent methane production from CO₂ reduction will show a mixed ¹²C/¹³C ratio if both processes are occurring (Eq. 3). Results from these isotope labeling experiments were consistent with this hypothesis, producing partially ¹³C-enriched CH₄ with a ¹³C value of 40.64 ± 0.68% (n = 3; Supplementary Fig. 4).

$$4^{13}\mathrm{CO}+{{\mathrm{H}}^{12}\mathrm{CO}}_3^{-}+3{\mathrm{OH}}^{-}+2{\mathrm{H}}_2\mathrm{O}=4{{\mathrm{H}}^{13}\mathrm{CO}}_3^{-}+\left[{{}^{12}\mathrm{CH}}_4,{{}^{13}\mathrm{CH}}_4\right]$$ 3

ANME and SRB consortia show anabolic activity by FISH–nanoSIMS when supplied with CO as the electron donor

Having demonstrated that CO oxidation occurs in seep sediments coupled with the reduction of either sulfate or CO₂, we then asked whether this bulk anaerobic CO respiration in turn supported anabolic activity by ANME-2 and their SRB partners in the absence of methane addition. Single-cell anabolic activity was measured using stable isotope probing with ¹⁵N-ammonium as a marker for biosynthetic activity combined with FISH–nanoSIMS^33–35. In these 4-month CO incubations, sulfate and bicarbonate served as electron acceptors, and ¹⁵N-ammonium was added as a nitrogen source (Supplementary Table 1).

A total of 4,979 FISH-identified cells within 42 multi-celled consortia of ANME-2b (ANME-2b-729 oligonucleotide probe) and their SRB partner (likely Seep-SRB1g²⁹; hybridized with a Desulfosarcina/Desulfococcus group probe (DSS658)) were analyzed from the Set #1 incubation experiments (Supplementary Table 1). Cellular ¹⁵N enrichment was measured in both ANME-2b and syntrophic SRB cells (Fig. 6a, b, Table 1), demonstrating that both were anabolically active during CO oxidation coupled with sulfate reduction in the absence of external methane addition. Notably, high concentrations of CO in the headspace resulted in lower levels of cellular ¹⁵N enrichment, consistent with some degree of toxicity (Fig. 6b, Table 1). Anabolic activity decreased to a greater extent with increasing concentrations of CO for both ANME (1.0 bar CO) and SRB (0.4 bar and 1.0 bar), with values below the baseline ¹⁵N enrichment (here determined to be 0.53 atom%; Table 1). This observation is consistent with reports from CO-metabolizing methanogens, where higher levels of CO inhibited methanogenic activity^10,36,37. When CO concentrations were provided at 0.4 bar, however, approximately 22% of the 831 ANME-2b cells analyzed remained anabolically active at a considerable level, while the fraction of active SRB cells dropped dramatically (6 out of 488 cells). This suggests a higher CO tolerance by ANME-2b relative to their SRB partner.

Fig. 6. Anabolic activity summarized from paired FISH–nanoSIMS experiments for ANME-2b archaea and its syntrophic SRB (Seep-SRB1g).

In (a, c), the top row shows FISH images for ANME-2b (red) and their syntrophic SRB partner (green); the bottom shows a paired representative nanoSIMS image for each microcosm condition. The ¹⁵N fractional abundance in all cell ROIs analyzed in this study is summarized in (b, d), where the dashed line represents the minimum threshold considered for cell ¹⁵N enrichment/activity (0.59 atom% ¹⁵N). Statistical information for these analyses is summarized in Table 1. a, b Samples from sulfate/HCO₃-microcosm incubations were analyzed after 122 days, and c, d sample from HCO₃⁻ microcosms (sulfate-free) were analyzed after 90 days. The numbers above (b, d) represent the number of ANME and SRB cells (ROIs) analyzed in each treatment. Scale bar is 5 µm in all panels. Box plot in (b) is depicted through their quartiles (25th, 50th, and 75th percentiles are shown).

Table 1.

¹⁵N enrichment of ANME-SRB consortia at single cell level from FISH–nanoSIMS analysis

	e− acceptor	Sulfate	Sulfate HCO3−	Sulfate HCO3−	Sulfate HCO3−	Sulfate HCO3−	HCO3−
	e− donor	CH4	CO_0.1 bar	CO_0.4 bar	CO_1.0 bar	CO_killed	CO_0.4 bar
ANME-2b	15N fractional abundancea % of active cellsb no. of cells analyzedc	13.65 ± 8.33 (100%) (368)	1.16 ± 0.40 (94.27%) (576	0.63 ± 0.17 (22.38%) (831)	0.49 ± 0.09 (3.85%) (805)	0.53 ± 0.03 (776)	0.67 ± 0.26 (27.91%) (1,684)
SRB	15N fractional abundance % of active cells no. of cells analyzed	16.72 ± 6.29 (100%) (194)	1.05 ± 0.71 (64.82%) (452)	0.46 ± 0.08 (1.23%) (488)	0.44 ± 0.05 (1.23%) (489)		0.44 ± 0.05 (0.77%) (780)
	# of aggregates	5	10	10	10	7	20

^a15N fractional abundance: ¹⁵N/(¹⁵N + ¹⁴N)×100%.

^bThe percentage of active cell ROIs in each condition (the ones with ¹⁵N fractional abundance higher than any ROI in the killed condition, thresholding value 0.59%).

^cThe number of single cells analyzed in each treatment.

To further examine whether ANME-2b were anabolically active and responsible for the measured CO oxidation and methane production rather than catalyzed by other unidentified CO-oxidizing microorganisms in the sediment, additional FISH–nanoSIMS experiments were performed using sulfate-free conditions over a 3-month period, here designed to limit the activity of the SRB partner and potential for cryptic methane cycling through syntrophic AOM (Fig. 6c, d, Table 1). Cellular ¹⁵N enrichment for 2,464 ROIs (cells) belonging to ANME-2b and their SRB partners were resolved. ANME-2b cells showed an average ¹⁵N enrichment of 0.67 ± 0.26 atom% (470 out of 1,684), while co-occurring SRB cells were anabolically inactive as expected, with the exception of a few cells (6 out of 780; Fig. 6d). The observation of ANME-2b anabolic activity in the absence of activity by their SRB partner supported their direct involvement in methane production during respiration of CO as an energy source.

We assessed the relative differences between CO oxidation and anaerobic methanotrophy supporting ANME growth using the cell-specific growth rates from FISH–nanoSIMS and bulk sediment respiration rates (Table 2). With CO addition (0.4 bar), methane production by ANME-2b occurred with an average rate that was approximately 1/9 of the sulfate-coupled AOM rate determined in parallel incubations supplied with methane and sulfate (Table 2). However, the apparent cellular energy gain from growth rate was minimal, with ANME-2b growth on CO (0.4 bar) and CO₂ nearly 20 times lower than the canonical AOM coupled sulfate reduction (Table 2). This implies that the observed ANME-2b CO oxidation is likely used to support cellular maintenance processes, rather than growth.

Table 2.

The ratio of single-cell ¹⁵N assimilation to bulk potential energy generation by redox reaction

Condition	CO/bicarbonateb	CH4/sulfate
CH4 production or oxidation rate (nmol CH4 d−1 cm−3sed)	0.061 ± 0.054	0.52 ± 0.02
ANME-2b single cell 15N (atom%) assimilation per bulk catabolic energy yield (15F J−1 cm−3sed)a	0.010 ± 0.0038c	0.21 ± 0.13d
number of cells analyzed	1,684	368

^a15F = ¹⁵N/(¹⁵N + ¹⁴N).

^bData from Fig. 5 with 0.4 bar CO.

^cThe bulk energy yield was based on the theoretical redox reaction from Eq 2.

^dThe bulk energy yield was based on the reaction: CH₄ + SO₄²⁻ = HCO₃⁻ + HS⁻ + H₂O, ΔG = −17 kJ/mol CH₄.

Metatranscriptomic analysis of ANME-2b/SRB response to CO

Metatranscriptomics was performed to investigate the underlying metabolic response of CO metabolism compared with methane (1.5 bar methane, 0.1 and 1.0 bar CO) using either sulfate or bicarbonate as the electron acceptors (n = 3 biological triplicates for each treatment, 9 sediment incubations total). Nearly 90% of the reads from these microcosm experiments were affiliated with the dominant AOM consortia members, ANME-2b and syntrophic Seep-SRB1g, consistent with the 16S rRNA amplicon sequencing results (Supplementary Fig. 2). In the remaining pool of transcripts, we were unable to resolve the taxonomic identity of any other microbes at a diagnostic level, however targeted screening of expressed methyl-coenzyme M reductase (mcr) from methanogens only recovered ANME-affiliated mcr transcripts. This finding further decreased the likelihood of hidden CO-metabolizing methanogens in the sediment incubations. The dominance of transcripts affiliated with ANME-SRB consortia members supports our single-cell isotopic data and independently supports ANME-2b and their syntrophic SRB partner as the organisms responsible for the observed CO oxidation. Subsequent transcriptomic analyses were then focused on ANME-2b and Seep-SRB1g. To better illustrate the transcriptomic activity, both differential gene expression (hereafter abbreviated DE) (DESeq2) and percentile ranking (Sleuth TPMs (transcripts per million reads)) were analyzed.

For ANME-2b, the expression of most genes did not change significantly between CO and methane treatments according to the overall transcripts per million reads (TPM) distribution (Supplementary Fig. 6). Approximately 2.79% of ANME genes were significantly up-regulated and 13.21% down-regulated in the microcosms with 0.1 bar CO compared with the control AOM conditions supplied with CH₄ and sulfate (Supplementary Fig. 7). Of the genes found to be significantly up-regulated in the CO condition compared to the CH₄ condition, most were associated with genes generally affiliated with stress response, including defense mechanisms, toxin/antitoxin systems, and stress-related ABC exporters (ABC-type antimicrobial peptide transport system; Supplementary Data 1). This indicates that CO is a less favorable substrate for ANME-2b respiration compared with CH₄ and is consistent with some degree of toxicity.

The ANME-2b carbon monoxide dehydrogenase, cooS, was actively expressed, consistent with involvement in CO metabolism. Unexpectedly, however, the expression levels were found to be equivalent between the methane/sulfate vs. CO/sulfate or CO/bicarbonate conditions, with mean values of 52.3 ± 5.1% (CH₄), 48.2 ± 3.4% (0.1 bar CO), and 55.1 ± 3.4% (1.0 bar CO), respectively (Supplementary Data 1). In contrast to cooS, significantly lower expression levels were found for the full cdh operon, which is an integral component of the Wood–Ljungdahl pathway (66.3 ± 13.6% (CH₄) vs. 49.9 ± 11.6% (0.1 bar CO), p < 0.001 (two-tailed t-test); Supplementary Data 1). This suggests a different response between the two types of ANME-2b-encoded CO dehydrogenases when supplied with an external source of CO.

For ANME-2b, we observed a high level of expression for most genes associated with methanotrophy/methanogenesis and carbon fixation pathways within the CO/sulfate and bicarbonate conditions, which was at a comparable level with the methane/sulfate treatment supporting AOM (Fig. 7, Supplementary Data 1). These genes included the methyl-coenzyme M reductase (mcr) operon (>99.8% among all conditions), 5,10-methylenetetrahydromethanopterin reductase (mer; mean: 99.1% (CH₄), 97.4% (0.1 bar CO), 96.8% (1.0 bar CO)), and formyl-methanofuran dehydrogenase (fwd) operon (mean: 96.6% (CH₄), 94.0% (0.1 bar CO), 93.8% (1.0 bar CO); Fig. 7, Supplementary Data 1). In contrast, N⁵-methyl-H₄SPT:HS-CoM methyltransferase (mtr) operon, a key energy-converting step in the methanogenesis pathway, had a significantly lower expression level in CO/sulfate/bicarbonate conditions compared to the parallel methane/sulfate treatment (0.1 bar CO vs. CH₄; percentile: 84.7 ± 3.3% vs. 98.2 ± 0.6%, mean ± s.d.; differential analysis: ca. 25-fold decrease on average; Fig. 7, Supplementary Fig. 8, Supplementary Data 1).

Fig. 7. Transcriptomic data for selected genes from ANME-2b and syntrophic Seep-SRB1g involved in energy metabolism.

Transcriptomic analysis for seep sediment microcosms was conducted after 6 days of incubation with either methane or CO as the electron donor and sulfate as the electron acceptor. a Expression of selected genes associated with ANME-2b: CO dehydrogenase, methanogenesis pathway, energy conservation, syntrophic electron transfer, and carbon fixation. b Expression of selected genes in Seep-SRB1g: CO dehydrogenases, core respiratory pathways, syntrophic electron transfer, and carbon fixation. The detailed profiling of these genes is summarized in Supplementary Data 1 and 2. Symbol colors here correspond to the microcosms shown in Fig. 2. This experiment was performed in biological triplicate (dot represents mean value, n = 3). Where not visible, the error bar is smaller than the symbol.

Other key genes involved in the pathways of energy conservation and syntrophic electron transfer⁶ also displayed a lower expression level in treatments with CO versus methane, both in terms of ranking and DE analysis (Fig. 7, Supplementary Fig. 8). For example, the NADH-ubiquinone/plastoquinone oxidoreductase (fpo) operon (F₄₂₀H₂:methanophenazine oxidoreductase), involved in F₄₂₀ conversion, was down-regulated under CO/sulfate/bicarbonate treatments (ranking: 84.9% (0.1 bar CO) and 83.9% (1.0 bar CO) vs. 96.7% (CH₄)), with a 9-fold decrease in differential expression (DE) (0.1 bar CO vs. CH₄; Fig. 7, Supplementary Fig. 8, Supplementary Data 1). In addition, the expression of genes associated with the Na⁺-translocating ferredoxin:NAD⁺ oxidoreductase (rnf) was also significantly lower in CO/sulfate/bicarbonate treatments (ranking: mean, 71.4% (CH₄), 43.4% (0.1 bar CO), 49.4% (1.0 bar CO); DE: ca. 4-fold decrease (0.1 bar CO vs. CH₄); Fig. 7, Supplementary Fig. 8, Supplementary Data 1). The decrease in transcription of mtr, fpo, and rnf indicates that while these energy-conserving pathways are expressed during CO oxidation in ANME-2b, this respiratory process clearly deviates from their primary methanotrophic metabolism.

In comparison with ANME-2b, the syntrophic Seep-SRB1g partner revealed an opposite transcriptional response to CO, where 21.00% of expressed genes were significantly up-regulated in treatments with 0.1 bar CO with either sulfate or bicarbonate as the electron acceptor compared with the methane/sulfate treatment (Supplementary Fig. 6), and only 1.52% of genes were significantly down-regulated (Supplementary Fig. 7). This expression pattern changed substantially in the high CO condition (1.0 bar CO), where the proportion of down-regulated genes increased to 14.42%, and only 2.32% of genes were highly expressed (Supplementary Fig. 7). Like ANME-2b, Seep-SRB1g actively expressed CO dehydrogenases among CH₄ and CO treatments, where both the cooS and the cdh operons had comparable rankings in expression between methane/sulfate and CO/sulfate/bicarbonate (0.1 bar CO) condition (Fig. 7, Supplementary Data 2). At higher CO concentrations, however, the cdh operon gene expression was significantly down-regulated (1.0 bar CO).

In the CO incubations with sulfate, the Seep-SRB1g partner showed active expression of respiratory pathways, potentially enabling Direct Interspecies Electron Transfer (DIET) between the syntrophic partners³⁵. This included significant up-regulation of genes encoding extracellular multi-heme cytochrome c proteins predicted to be involved in extracellular electron transfer²⁶ including omcX (fold change: 2.72 (1.0 bar CO vs. CH₄), P_adj = 5.78E-08) and tmlA (fold change: 3.15 (1.0 bar CO vs. CH₄), P_adj = 0.027; Supplementary Fig. 8, Supplementary Data 2). Most genes involved in the dissimilatory sulfate-reduction pathway were also highly expressed in all three conditions (percentile ranking >80%), including the quinone-interacting membrane-bound oxidoreductase complex (qmo), dissimilatory-type sulfite reductase (dsr), sulfate adenylyltransferase (sat), and adenylyl-sulfate reductase (apr) (Fig. 7, Supplementary Data 2). In addition, the expression level of genes belonging to the multi-subunit Na⁺/H⁺ antiporter (mnh) operon changed dramatically between the methane/sulfate condition (mean ranking: 9.7%, CH₄) and CO/sulfate/bicarbonate treatments (mean ranking 86.5%, 0.1 bar CO), representing an average 188-fold increase in expression (0.1 bar CO vs. CH₄; Fig. 7, Supplementary Fig. 8, Supplementary Data 2). Other genes, such as the Flx-Hdr complex, a key component in the cycling of DsrC in the CO metabolism in SRB cells²⁶, showed comparable expression levels with CO vs. methane (Fig. 7, Supplementary Fig. 8, Supplementary Data 2). SRB genes previously identified to be involved in syntrophic electron transfer and carbon fixation (via Wood–Ljungdahl pathway) also showed equivalent expression levels in treatments with 0.1 bar CO vs. CH₄/sulfate (Fig. 7, Supplementary Fig. 8, Supplementary Data 2).

Discussion

Syntrophic ANME-SRB consortia within deep-sea sediments are well-characterized for their role in sulfate-coupled AOM. Although environmental, genomic, and thermodynamic evidence has suggested that marine methanotrophic ANME may also use alternative electron donors such as CO, this has not been experimentally tested. Here, using complementary techniques including long-term microcosm experiments, multi-omics, and single-cell analyses, we demonstrate that ANME and their syntrophic SRB partners are capable of actively oxidizing CO in seep sediments.

CO oxidation by ANME-2b/syntrophic SRB

We initially confirmed that CO oxidation was carried out by anaerobic methanotrophic ANME-2b/ SRB consortia rather than by co-occurring methanogens or other carboxydotrophic microbes, within the complex and diverse microbial community present within our sediment microcosms. We briefly summarize the lines of evidence from our geochemical, transcriptomics, and single-cell isotopic analyses, which support this conclusion.

First, 16S rRNA amplicon sequencing did not recover known methanogens within our microcosm experiments after 4 months of incubation (Supplementary Fig. 2). Transcripts from CO dehydrogenases or mcr genes phylogenetically associated with methanogenic archaea were also absent from our metatranscriptomic analysis, which was overwhelmingly dominated by ANME-2b and Seep-SRB1g. Finally, while related CO-oxidizing methanogens like Methanosarcina sp. are known to produce acetate and formate as major products, along with methane, during growth on CO¹⁰, neither metabolite was detected in our microcosm experiments. Consistent with this, no expression of genes associated with acetate metabolism was detected. Collectively, these findings reduce the likelihood that non-ANME-SRB microbes contributed meaningfully to CO oxidation in our experiments. Instead, they support the conclusion that anaerobic ANME and SRB consortia, specifically ANME-2b/Seep-SRB1g, are catalyzing CO oxidation in our seep sediment microcosm experiments.

The second line of evidence came from direct analysis of the ANME-2b and Seep-SRB1g consortia themselves. In experiments where CO was provided as the sole electron donor, CO oxidation occurred alongside the reduction of both sulfate and CO₂ (bicarbonate), resulting in the production of sulfide and methane. Despite the lack of methane headspace in these treatments, FISH–nanoSIMS confirmed that ANME-SRB consortia were anabolically active, while GC-MS analysis revealed methane production linked to CO₂ reduction. Still, the detection of activity by ANME-SRB on its own does not unequivocally identify them as the primary CO oxidizers, as it remained possible that an as-yet-unidentified seep sediment microorganism could have oxidized CO and generated methane, subsequently fueling sulfate-coupled AOM by ANME-SRB (e.g., Hadarchaeota MAG, Guaymas_P_008, as previously described in the Guaymas Basin)¹⁸. To directly test whether ANME-2b was capable of CO oxidation independent of sulfate-coupled AOM, we incubated the same seep sediments in a sulfate-free medium and evaluated the activity of ANME. Under these conditions, with CO₂ as the sole electron acceptor, we again observed CO oxidation and methane production and, importantly, the ANME-2b cells were anabolically active, supporting their role as the primary CO oxidizer. While previous studies have reported that ANME archaea are unable to oxidize methane with CO₂ as the electron acceptor³⁸, our results demonstrate that ANME-2b can use CO₂ to oxidize CO, producing methane, and that ANME-SRB consortia remain active even in the absence of an external methane source.

Physiologically, both catabolic and anabolic activity of ANME-SRB decreased with increasing CO concentration, although cell-specific activity remained high at lower CO concentrations (0.1 bar). This trend is consistent with previous reports of microbial inhibition in response to CO exposure^10,39 and is also supported by the higher expression levels of stress-related genes under CO conditions. A CO partial pressure of 0.1 bar in the headspace corresponds to 112.83 µM dissolved in seawater at 4 °C^40,41, which is two to three orders of magnitude higher than the standing CO concentrations reported from methane-rich deep-sea hydrothermal sediments from Guaymas Basin (hundreds of nM)⁴² and from organic rich marine coastal sediments of the East China Sea (98.3–333.7 nM)⁴³ and from a Mediterranean seagrass meadow (17–51 nM)⁴⁴. Thus, it is likely that CO metabolism by anaerobic methanotrophs may be more pronounced in the natural environment, where CO concentrations and toxic effects are lower.

Growth or maintenance?

Methane production by ANME-2b was observed at a rate approximately one-ninth that of AOM in parallel incubations with 0.4 bar CO. This observation raised the question of whether ANME-2b archaea are obligate methanotrophs and whether they can sustain anabolism through CO oxidation coupled to methane generation. Quantitative estimates indicate that methane production from CO (0.4 bar) and CO₂ yields minimal biomass per unit of potential energy gained, approximately 20 times lower than that of canonical sulfate-coupled AOM. These results suggest that CO metabolism primarily supports maintenance of ANME-2b activity rather than enabling significant growth. This interpretation is consistent with a recent report showing that, in the phylogenetically distinct soil bacterium Mycobacterium smegmatis, CO metabolism does not support growth but instead enhances survival and long-term persistence under nutrient limitation⁴⁵. Several factors may explain the low efficiency of energy utilization for biosynthesis for ANME-2b.

The low biomass yield associated with CO-dependent methane production indicates that CO oxidation provides limited energy for growth in ANME-2b. This observation motivates an examination of whether canonical acetyl-CoA-based pathways, common in CO metabolism in related methanogens, are also operative in ANME. Studies of CO metabolism by Methanosarcina acetivorans have shown that acetate and formate production via the acetyl-CoA pathway, coupled to substrate-level phosphorylation, is the predominant CO-dependent catabolic pathway, rather than methanogenesis^10,27. In these studies, more than 75% of the CO was converted to acetate, while less than 10% was released as methane during carboxydotrophic growth.

In contrast, in our microcosm experiments with ANME-2b consortia, neither acetate nor formate was detected during CO oxidation. Consistent with this, phosphotransacetylase (pta) and acetate kinase (ackA), two critical genes involved in substrate-level phosphorylation in the conversion of acetyl phosphate from acetyl-CoA to acetate, are absent from ANME genomes⁶, nor was their expression found in the metatranscriptomic sequencing data, suggesting that the Pta-AckA pathway is not active. An alternative route from acetyl-CoA to acetate involves AMP-forming acetyl-CoA synthetase (Acs) and ADP-forming acetate-CoA ligase (Acd). A previous study proposed that marine ANME-2a might perform acetogenesis via Acd⁴⁶; however, this reaction yields little energy (4.6 kJ/mol acetate) to sustain microbial growth. This interpretation is also further supported by enzymatic characterization of ANME-2a Acs, which suggested that acetate conversion, if existing in ANME, is more likely associated with anabolic processes than energy conservation⁴⁷. These observations indicate that energy conservation via the acetyl-CoA pathway in ANME-2b is unlikely to occur through the same mechanism as described for related methanogens, which is distinct from what has been recently reported for freshwater Methanoperedens (ANME-2d)⁴⁸.

Another possibility is the presence of an uncharacterized cytoplasmic enzyme associated with the methanogenesis pathway that catalyses the methyl-H₄SPT reaction without contributing to energy conservation. During CO oxidation, we observed a substantial down-regulation of genes from the methanogenesis pathway (Supplementary Fig. 8). In particular, the N⁵-methyl-H₄SPT:HS-CoM methyltransferase (mtr) operon, which catalyses the energy-conserving (sodium-pumping) methyl transfer reaction, exhibited significantly lower expression under CO conditions (Supplementary Fig. 8, Supplementary Data 1). This suggests a diminished role for the Mtr complex compared with other steps in the methanogenic pathway, potentially explaining the limited energy conservation available for biosynthesis.

The observed ANME-2b expression pattern is consistent with reports of a substantial decrease in mtr transcripts for M. acetivorans when grown on CO compared to methanol as the energy substrate⁴⁹. Concurrent with the decrease in mtr, the expression levels of the mts system in M. acetivorans (mtsD, mtsF, and mtsH) were significantly up-regulated under carboxydotrophic conditions⁴⁹. Based on these findings, Ferry et al. proposed a potential bypass for Mtr via direct cytoplasmic methyl transfer at the expense of minimal energy conservation⁵⁰. Mts genes encode a predicted cytoplasmic methyl-tetrahydromethanopterin:CoM (CH₃-THMPT:HS-CoM) methyltransferase, which was hypothesized to serve as an alternative pathway when mtr expression is down-regulated under CO conditions. While biochemical enzymatic assays using exogenous protein expression demonstrated CH₃-THMPT:HS-CoM methyltransferase activity⁵¹, subsequent in vivo mutation and physiological experiments did not support this model, concluding that Mtr remains central to methanogenic metabolism and is not easily bypassed ⁵².

Supporting this interpretation, homologs to mts were not detected in the ANME-2b genome, indicating that the proposed bypass mechanism in refs. ^49,50 is unlikely to compensate for the down-regulation of mtr in ANME-2b. Consequently, the question of why ANME-2b mtr expression is low with CO, and whether this step can be bypassed, remains unresolved. Overall, the decreased expression of key energy conservation machinery in ANME-2b under CO conditions likely accounts for the low efficiency of energy conservation for anabolism, which in slow-growing methane-oxidizing ANME archaea is already comparatively minor (~1%)⁵³. Instead, moderate methane production from CO₂ reduction coupled with CO oxidation may function primarily as a mechanism for CO detoxification and maintenance of metabolic activity under conditions unfavorable for AOM. Similar low assimilation efficiencies have been reported in other microorganisms, such as SRB during respiration of extracellular metal oxides^54–56, and methanogenic hydrocarbon-oxidizing archaea metabolizing long-chain alkanes to methane^57–59. CO oxidation by ANME now adds to the list of low-carbon assimilation and maintenance metabolisms documented among slow-growing anaerobes.

The question of whether methanotrophic ANME lineages possess net methanogenic capabilities has been a topic of debate for the past decade^60–63. To date, there is no direct evidence supporting energy-conserving net methanogenesis in ANME lineages under in situ conditions, although some ANME-1 lineages contain hydrogenases that could, in principle, support hydrogenotrophic methane production⁶. However, the majority of ANME-2 lineages (e.g., ANME-2b) lack hydrogenases and do not appear to be capable of using H₂, acetate, or methyl compounds as an energy source⁶. Our experiments instead support CO-dependent methane production by ANME-2b, coupled with CO₂ reduction, as a mechanism to sustain basic metabolic activity (e.g., maintenance energy), rather than supporting sufficient anabolic activity for cell division at high CO levels (e.g., partial pressures of 0.1 bar or 0.4 bar). In contrast, the methylotrophic methanogen M. acetivorans grown with 1 bar of CO is capable of methanogenic growth, with a generation time of 24 h^10,27. Given the lack of growth observed for ANME-2b in our experiments, CO-dependent methane production by these archaea should not be regarded as conventional methanogenesis, but rather as a mechanism for maintenance and cellular redox balancing. While high CO levels appeared to trigger a stress response in the ANME consortia based on metatranscriptomic data, ANME growth rates and metabolism may be more favorable at the lower CO concentrations likely present in natural environments, consistent with a potential hormetic response to CO exposure⁶⁴. Accordingly, future long-term cultivation experiments with a low sustained CO supply would be beneficial for assessing ANME-SRB growth under more environmentally relevant conditions.

CO oxidation and energy metabolism in ANME

ANME-2b expressed CO dehydrogenases (i.e., cooS) at a comparable level in the CO vs. methane treatments (Fig. 7, Supplementary Fig. 8). While these transcriptomics results were unexpected, similarly unintuitive cooS expression patterns were reported for the methanogen M. acetivorans when grown on acetate or CO, where a copy of cooS was down-regulated on CO relative to acetate⁴⁹. More recent work using direct enzymatic assays rather than mRNA expression, however, has shown that CO-metabolizing M. acetivorans exhibits approximately three-fold higher CO dehydrogenase activity over methanol-grown cultures¹⁰. This suggests that, as in M. acetivorans, CO-respiring ANME-2b archaea may regulate cooS activity at the post-transcriptional or enzymatic level rather than through changes in gene expression.

Carbon monoxide oxidation by CO dehydrogenases in ANME-2b is the initial step in electron flow and subsequently proceeds via various intracellular electron carriers like ferredoxin and F₄₂₀. Concomitant with the cycling of reducing power and maintenance of cellular redox balance, we considered two plausible sinks for CO-derived electrons in ANME-2b archaea: (i) through cryptic AOM via DIET with syntrophic SRB partners, consistent with its main mode of energy conservation³⁵, or (ii) via the methanogenesis pathway for CO₂ reduction (Fig. 8).

Fig. 8. Overview of proposed carbon monoxide oxidation and energy metabolism in ANME-SRB consortia.

a Simple schematic of CO metabolism in ANME-SRB consortia and b detail of hypothesized energy metabolism in ANME-2b and SRB consortia. In (b), proposed catalyzing steps connected with black arrows are the conventional energy metabolic pathways of AOM; green arrows indicate the electron flow across ANME-SRB consortia with CO and sulfate as the electron donor and acceptor. CODH, carbon monoxide dehydrogenase (more likely cooS); CoM-S-S-CoB, heterodisulfide; F₄₂₀/F₄₂₀H₂, oxidized/reduced cofactor F₄₂₀; Fd/Fd²⁻, oxidized/reduced ferredoxin; Mcr, methyl-S-CoM reductase; Mtr, N⁵-methyl-H₄SPT:CoM methyltransferase; Mer, N⁵,N¹⁰-methylene-H₄SPT reductase; Mtd, N⁵,N¹⁰-methylene-H₄SPT dehydrogenase; Mch, N⁵,N¹⁰-methenyl-H₄SPT cyclohydrolase; Ftr, formyl-MF:H₄SPT formyltransferase; Fmd, formyl-methanofuran dehydrogenase; Hdr, heterodisulfide reductase; HS-CoB, coenzyme B; HS-CoM, coenzyme M; Mp/MpH₂, oxidized/reduced methanophenazine; Fpo, F₄₂₀H₂:methanophenazine oxidoreductase; Rnf, rhodobacter nitrogen fixation, energy converting Fd²⁻: acceptor oxidoreductase; Mnh, multi-subunit Na⁺/H⁺ antiporters; Tmc, transmembrane channel-like protein; Qrc, quinone-reductase complex; Dsr, dissimilatory sulfite reductase; Qmo, quinol oxidizing complex; Mrp, multiple resistance and pH adaptation antiporters; Flx, flavin oxidoreductase; Apr, adenosine-5′-phosphosulphate reductase; APS, adenylyl-sulfate/adenosine-5′-phosphosulphate; Sat, sulfate adenylyltransferase; Q/QH₂, oxidized/reduced quinol.

DIET coupled with CO metabolism in ANME-2b is predicted to share the same pathway as methane oxidation (AOM) for shuttling electrons across the cytoplasmic membrane and into the extracellular space (see Fig. 8b, green arrows in the upper panel). In this scenario, membrane-bound enzymes Rnf and/or Fpo reduce ferredoxin (Fd) and F₄₂₀, which subsequently deliver electrons to the membrane-associated electron carrier methanophenazine (Mp). Mp is then predicted to transfer electrons to membrane-bound multi-heme cytochromes c, which facilitate electron transfer to the syntrophic partner via DIET³⁵. Notably, both the energy-conserving Rnf and Fpo operons were significantly down-regulated under the sulfate-CO condition relative to the sulfate-CH₄ treatment (Supplementary Fig. 8, Supplementary Data 1), suggesting a metabolic preference for methane. Despite the comparatively low expression levels, however, these pathways appeared to remain active. Therefore, it is conceivable that electrons derived from CO were transferred extracellularly to the syntrophic SRB partner, reducing sulfate as the electron sink for the consortia.

Under the ANME methanogenesis scenario, expression patterns of genes involved in methane production with CO₂ as the terminal electron sink were examined. Here, electrons carried by electron donors Fd²⁻ and F₄₂₀H₂ are likely transferred from formyl-methanofuran dehydrogenase (Fwd) to F₄₂₀-dependent N⁵,N¹⁰-methylene-H₄MPT dehydrogenase (Mtd) and methylene-H₄MPT reductase (Mer), following the conventional methanogenesis pathway in M. acetivorans^10,27. Importantly, the ANME-2b mcr operon exhibited one of the highest percentile expression rankings in the CO treatments (ca. 99.8%, 0.1 bar CO) (Fig. 7, Supplementary Data 1), remaining essentially unchanged relative to the methane condition. The high transcriptional activity of mcr is consistent with its involvement in methane generation by ANME-2b and, possibly, in the cryptic oxidation of newly synthesized methane under conditions where both CO and sulfate were provided (Fig. 3). While the metabolic contributions of individual ANME-2b cells to each process remains unknown, a spatial or temporal division of labor among co-occurring ANME-2b cells likely supports cryptic cycling rather than both processes occurring simultaneously within the same cell.

CO oxidation and hypothesized energy metabolism in the syntrophic Seep-SRB1g

Under incubation conditions with sulfate and CO, the syntrophic SRB partner may receive electrons extracellularly via DIET from CO-oxidizing ANME, and/or intracellularly through direct CO oxidation catalyzed by its own CO dehydrogenase. We assess the likely biochemical pathways involved for both possibilities below.

In the first DIET scenario, respiratory electron transfer would follow pathways analogous to those described for conventional syntrophic sulfate-coupled methane oxidation. In this model, electrons are transferred from ANME to their syntrophic SRB partner via multi-heme cytochrome c proteins, which then couple these electrons to sulfate reduction²⁶. Transcriptomic data from the Seep-SRB1g bacterial partner support this scenario, with genes encoding extracellular multi-heme cytochrome c proteins significantly up-regulated under CO conditions. These include omcX (fold change: 2.72 (1.0 bar CO vs. CH₄), P_adj = 5.78E-08) and tmlA (fold change: 3.15 (1.0 bar CO vs. CH₄), P_adj = 0.027; Supplementary Fig. 8, Supplementary Data 2), both predicted to be involved in extracellular electron transfer in Seep-SRB1g²⁶. Electrons delivered via extracellular cytochrome c proteins are likely routed through the periplasm to reduce quinone pools in the cytoplasmic membrane and ultimately used for energy generation through dissimilatory sulfate reduction.

As the Seep-SRB1g partner also encodes its own complement of CO dehydrogenases that were actively expressed, it is possible that the syntrophic SRB also directly couples CO oxidation to sulfate reduction (Fig. 7, Supplementary Data 2). CO oxidation in this scenario is predicted to occur through the SRB’s cytoplasmic CO dehydrogenases, serving as the primary source of electrons for downstream sulfate reduction. The direct use of CO would require a physiological shift in the syntrophic Seep-SRB1g partner, where respiratory electron transfer occurs intracellularly rather than, or in addition to, extracellularly via DIET. This change in the direction of electron flow during CO oxidation likely requires metabolic adjustment to optimize the electron transport pathways in the SRB. Here, electrons from CO would be used to reduce ferredoxin, and the reduced ferredoxins, and subsequently NAD⁺, are then used to generate sodium motive force through membrane-bound Rnf and Mnh complexes²⁶. Both Rnf and Mnh were significantly up-regulated by SRB, suggesting these energy-conserving complexes were active in treatments with CO and sulfate (Supplementary Fig. 8, Supplementary Data 2). Additionally, Flx-Hdr was also actively expressed at a comparable expression level with CO amendment (Fig. 7, Supplementary Data 2). This complex could serve as a key component in the cycling of DsrC during CO metabolism in SRB, by oxidizing two molecules of NADH to reduce one molecule each of ferredoxin and DsrC²⁶. The predicted coupling by Flx-Hdr effectively recycles ferredoxin from Rnf, while generating reduced DsrC for dissimilatory sulfate reduction. We hypothesize that the SRB’s use of ferredoxin, reflected in the high expression of Mnh genes along with the endogenous cycling of DsrC, represents a potential mechanism for the syntrophic Seep-SRB1g to oxidize CO coupled with sulfate reduction independent of electron transfer by ANME (see supplement for additional details). Targeted physiological and biochemical studies of the SRB partner will be necessary for deconvolving their direct and indirect interaction(s) with CO.

The combined evidence from our nanoSIMS experiments and transcriptomics results indicates that syntrophic Seep-SRB1g cells were engaged in both extracellular electron transfer, involving the passage of CO-derived electrons from ANME to SRB, and in the direct oxidation of CO, potentially lessening their energetic reliance on ANME (Table 1, Supplementary Fig. 5, Supplementary Note 2). Developing experimental strategies that enable the partial or complete decoupling of metabolic activity by the SRB from its ANME partner, akin to the previously demonstrated use of AQDS to decouple methane oxidation by ANME from SRB^8,38, would enable new opportunities in which to investigate the underlying physiology of this enigmatic syntrophic partnership. Future studies using optimized CO concentrations and controlled delivery mechanisms that minimize toxicity may reveal more about the metabolic flexibility of the SRB partners, including the extent to which these syntrophs can respire CO relative to their dependency on ANME.

In this study, we used microcosm incubations, combined with geochemical analysis, metatranscriptomics, and single-cell activity measurements by FISH–nanoSIMS to demonstrate that anaerobic ANME-SRB consortia can oxidize carbon monoxide (CO) in the absence of methane. Transcriptomic results suggest that CO oxidation in the ANME-2b archaea is catalyzed by CO dehydrogenases (likely cooS) and used to reduce CO₂, resulting in methane production through the seven-step methanogenesis/methanotrophy pathway. This redox process appears to be used primarily for maintenance by ANME-2, supporting cellular redox balance, rather than growth. In the presence of sulfate, however, the methane generated from CO oxidation in ANME fuels cryptic methane cycling through DIET with their syntrophic sulfate-reducing bacterial partners. Direct CO respiration by the sulfate-reducing bacterial partner also appears likely, allowing for some degree of energetic decoupling between ANME and SRB, while also representing an additional mechanism for energy maintenance within the consortia when methane concentrations may be limiting in the environment. The discovery that environmental AOM consortia can oxidize CO in the absence of methane may enable these consortia to remain viable under changing environmental conditions. This finding deepens our understanding of the complex carbon cycling interactions by these syntrophic archaeal-bacterial associations in deep-sea seep ecosystems.

Methods

Sediment collection and processing

Methane seep sediments from a white sulfur-oxidizing microbial mat habitat were collected by push core (PC6) with DSV Alvin during dive #AD4912 on 27 May 2017 on research cruise AT37-13 on the R/V Atlantis. Samples were collected from the Jaco Scar site, off Costa Rica, at 1811m water depth (lat 9.1163, lon 284.8372). The sediment core was maintained at 4 °C and processed shipboard upon collection, by upward extrusion and sectioning the sediments into 3-cm-thick horizons, which were stored individually in Whirl-Pak bags and then sealed in gas-tight mylar bags flushed with argon. These samples were maintained under anoxic conditions at 4 °C until processing back at the shore-based laboratory. Parallel cores were taken for geochemical and microbiological context at this site ⁶⁵.

Medium composition

Artificial seawater was used in the study, modified from ref. ⁸. The composition of the medium was: NaCl 457 mM, MgCl₂ 47 mM, Na⁺-HEPES (pH 7.5) 25 mM, KCl 7.0 mM, NaHCO₃ 5.0 mM, CaCl₂ 1.0 mM, K₂HPO₄ 1.0 mM, SeO₃²⁻ 0.01 μM, WO₄²⁻ 0.007 μM, 0.1% trace element solution, containing per liter: nitrilotriacetic acid 150 mg, MnCl₂·4H₂O 610 mg, CoCl₂·6H₂O 420 mg, ZnCl₂ 90 mg, CuCl₂·2H₂O 7 mg, AlCl₃ 6 mg, H₃BO₃ 10 mg, Na₂MoO₄·2H₂O 20 mg, SrCl₂·6H₂O 10 mg, NaBr 10 mg, KI 70 mg, FeCl₃·6H₂O 500 mg, NiCl₂·6H₂O 25 mg.

Microcosm setup and sampling

All manipulations of the sediment incubations were done anaerobically. Sediments from PC6 (serial number #10079) from Jaco Scar were first divided into two 100 ml serial vials and maintained with sulfate/methane prior to establishment of the formal microcosm experiments. Sediments were evenly allocated into 30 ml serum vials with 2 ml sediment (~3 g wet weight) distributed per vial and then overlaid with 8 ml artificial seawater to a final volume of 10 ml (1:5 dilution, see Supplementary Materials and Methods). As a killed control, slurried sediment was first fixed with 4% paraformaldehyde overnight, and then heat-treated at 90 °C for 4 h. Electron donors and acceptors were added separately to each incubation bottle. The electron acceptor, Na₂SO₄, was added at 5.0 mM final concentration. Methane (1.0 bar, with 5% of ¹³C-methane, Sigma-Aldrich) or ¹³CO (100%, 0.1 bar, 0.4 bar, and 1.0 bar, Sigma-Aldrich) was added as the electron donor with N₂ as the balance gas to a final total overpressure of 1.0 bar in the headspace. Isotopically labeled ¹⁵NH₄⁺ (final concentration 1 mM, 99% ¹⁵N, Cambridge Isotope Laboratories, Tewksbury, MA, USA) was also added as a tracer for single-cell stable isotope probing experiments using FISH–nanoSIMS. All the microcosm incubations were maintained in the dark at 4 °C, and 0.5 ml aliquot of the supernatant was taken every 6 days from the bottles using syringes and needles without opening the bottles. All the sampling operations were done in an anaerobic chamber (Coy) to avoid any potential oxidation of compounds of interest in the liquid. 20 µl of the filtered supernatant was immediately added to 380 μl Zn(OAc)₂ (500 mM) to preserve sulfide as a Zn precipitate for analysis by Cline. Approximately 100 µl of filtered supernatant was dispensed into sealed DIC vials, previously flushed with He and containing 200 µl of H₃PO₄ (42.5%). ¹³C-DIC analysis was used to quantify methane consumption in the microcosm experiments⁸. The remainder of the filtered supernatant was flash frozen in liquid nitrogen and stored at −80 °C for IC analysis (~200 µl). Microcosm incubations were refreshed with new ASW medium every month prior to sulfate depletion.

Microcosm incubations ran for 4 months. Slurry samples (1 ml from each vial) were taken and fixed in 2% PFA overnight at 4 °C (final concentration 1% PFA). The fixed sediment was subsequently washed 3 times with 3X PBS, followed by a single wash in EtOH/PBS (1:1) and then re-suspended in EtOH/PBS (1:1) to a final volume of 0.5 ml that was used for microscopy and nanoSIMS. The rest of the slurry was flash frozen in liquid nitrogen and kept in −80 °C for DNA extraction.

Microcosm manipulation experiments for monitoring methane production and CO consumption were performed in a separate set of incubations. Here, sediments from the same core (PC6) were mixed with ASW as described in the Supplementary Materials and Methods. Incubations were set up in 10-ml serum vials, under a N₂ headspace and amended with ¹³CO and Ar. The addition of Ar gas was used as an internal standard for the GC. DIC samples were taken before GC measurement at the various sampling time points.

Geochemistry determination

Analysis of sulfide was carried out using the colorimetric methylene blue method in technical triplicate⁶⁶. Sulfide concentrations were measured using a plate reader (TECAN Sunrise^TM) by monitoring the absorbance at 670 nm and quantified by comparing to known standards prepared in the same anoxic seawater medium. Quantification of newly formed DIC from ¹³CH₄ was measured using a GC-IR-MS GasBench II (Thermo Scientific) following protocols described in ref. ⁸. Briefly, the newly formed ¹³C-DIC concentration (∆[DIC](t_n)) was calculated from the measured ¹³C fractional abundance (¹³F) following the equation:

$$\mathrm{\Delta }\left[\mathrm{DIC}\right]\left(t_n\right)=\left[\mathrm{DIC}\right]\left(t_0\right)\times \frac{{}^{13}\mathrm{F}\left(t_n\right)-{}^{13}\mathrm{F}\left(t_0\right)}{{}^{13}\mathrm{F}\left(\mathrm{CO}\mathrm{or}{\mathrm{CH}}_4\right)-{}^{13}\mathrm{F}\left(t_n\right)}$$

where, [DIC](t₀) is 5 mM, ¹³F(t₀) is 0.1153, ¹³F(CO) is 1, and ¹³F(CH₄) is 0.05. For the total DIC concentration analysis, a standard curve over a series of known concentrations of DIC was determined from 0.1 to 10 µmol. The total DIC concentrations were analyzed from the measured peak areas corresponding to the total DIC. For analysis in Fig. 4, the newly formed ¹³C-DIC represented the amount of DIC generated from the ¹³CH₄ oxidation, and the difference between the total DIC and ¹³C-DIC showed the activity of CO oxidation in the incubations. Gas samples were quantified by a GC-TCD equipped with a HP-MOLSIV column (SN US3351826H, Agilent Technologies).

To test whether nitrogenases reduced CO in our microcosm experiments, GC-MS analysis was performed using an Agilent/HP 5890 GC–5972 MSD system equipped with a CP-Molsieve 5A column (Agilent Technologies). Multiple hydrocarbons were monitored, including methane (CH₄), ethylene (C₂H₄), ethane (C₂H₆), and propane (C₃H₈) (hydrocarbon detection limit: low ppb to low ppm range)^30,31. To test the production of ¹²C- and ¹³C-methane from sulfate-free bottles, a capillary HP-Plot/Q column was used on the same GC-MS system.

NMR spectra were acquired on a Bruker NEO 400 NMR spectrometer. D₂O was added to aqueous samples to a final concentration of 10% D₂O. To check for the presence of formate, a 1H spectrum was acquired using the standard Bruker parameter set WATER, employing a 1D NOESY pulse sequence with a 10 ms mixing time. To confirm the identity of formate in our incubations and to estimate its initial concentration, a sample containing formate was spiked with a known formate standard to raise the sample concentration by 100 µM, and the WATER experiment was repeated.

Extracellular metabolites determination

Extracellular metabolites were quantified via parallel Dionex Integrion HPIC (ThermoFisher, Waltham, MA, USA) ion chromatography systems with anion and cation columns running in parallel, housed at the Resnick Water and Environment Lab (WEL) at Caltech. The ion chromatography method was run as previously described^67,68 with the following modifications.

A Dionex AS-DV autosampler loaded with samples diluted 1:50 in 18 MΩ water to serial anion and cation columns and detectors, which are maintained at 30 °C. Anions were resolved by a 2 × 250 mm Dionex IonPac AS19-4 µm analytical column protected by a 2 × 50 mm Dionex IonPac AG19-4µm guard column (ThermoFisher, Waltham, MA, USA). A potassium hydroxide eluent generator cartridge generated a hydroxide gradient that was pumped at 0.25 mL min⁻¹. The gradient was constant at 10 mM for 5 min, increased linearly to 48.5 mM at 27 min, then increased linearly to 50 mM at 40 min. A Dionex AERS 500 suppressor provided suppressed conductivity detection running in recycle mode with an applied current of 31 mA. Cations were resolved by a 2 × 250 mm Dionex IonPac CS16-4 µm analytical column protected by a 2 × 50 mm Dionex IonPac CG16-4 µm guard column. A methanesulfonic acid eluent generator cartridge generated a methanesulfonic acid gradient that was pumped at 0.16 ml min⁻¹. The gradient was constant at 10 mM for 5 min, nonlinearly increased to 20 mM at 20 min (Chromeleon curve 7, concave up), and nonlinearly increased to 40 mM at 40 min (Chromeleon curve 1, concave down). A Dionex CERS 2 mm suppressor provided suppressed conductivity detection with an applied voltage of 4.2 V. Chromatographic peaks were integrated by Chromeleon 7.2.9 using the Cobra algorithm and were correlated with concentration by running known standards and generating 4-point calibration curves.

Sample preparation for aggregate embedding, sectioning, FISH, and nanoSIMS

Sample preparation was carried out following a protocol described previously³⁵. Briefly, paraformaldehyde-fixed consortia in the sediment slurry were detached from the sediment particles via gentle sonication on ice (10 s sonication/10 s break, 3 times, 4 W) with a microtip probe (Branson). AOM aggregates were separated using Percoll density centrifugation and concentrated onto a 5 μm filter, covered in molten noble agar (3% in 1×PBS), and embedded in glycol methacrylate (Heraeus Kulzer - Technovit® 8100). Sections of ca. 1 μm thickness were cut and stretched on a DI water droplet on a polylysine-coated slide with teflon wells (Tekdon Inc) and analyzed by fluorescence in situ hybridization (FISH) on a Zeiss LSM 980 confocal microscope with 63X objective using a general archaeal probe, an ANME-2b-specific probe, and a sulfate-reducing bacterial probe. Images of the FISH-stained consortia were collected, and the location of these consortia was mapped for subsequent nanoSIMS analysis as described below.

FISH conditions and oligonucleotide probes used in this study

The FISH hybridization followed the published protocol described in ref. ⁸. 2% PFA fixed sample aliquots were hybridized in a hybridization buffer containing 35% formamide and incubated at 46 °C for 2 h, followed by a wash step at 48 °C for 15 min.

The FISH probes used in this study (5 ng/μl for each) include:

ARCH915 (Alexa488), dual labeled: 5′ to 3′ = GTGCTCCCCCGCCAATTCCT⁶⁹;

ANME-2b-729 (Cy3), dual labeled: 5′ to 3′ = CGTTCTCGTAGGGCGCCT³⁵;

DSS658 (Alexa647), dual labeled: 5′ to 3′ = TCCACTTCCCTCTCCCAT⁷⁰.

NanoSIMS data acquisition

Isotope enrichment data were collected on a CAMECA nanoSIMS 50L in the Centre for Microanalysis at the California Institute of Technology. Gold-coated samples were pre-sputtered with a 70-pA primary Cs⁺ ion beam with aperture diaphragm D1 = 1 until the ¹⁴N¹²C⁻ ion counts stabilized. Data were collected using a 2.5-pA beam with D1 = 2 and entrance slit (ES) = 2. Four masses were collected corresponding to the ¹²C⁻, ¹³C⁻, ¹⁴N¹²C⁻, and ¹⁵N¹²C⁻ ions, for the determination of ¹³C/¹²C and ¹⁵N/¹⁴N ratios, respectively, using a tuning protocol described previously^8,35,71. Acquisitions were collected from different-sized square raster areas (ranging from 10 to 35 µm²), with 512 × 512 pixels, and 1 to 2 planes collected per area. Dwell time settings resulted in acquisition times of 30 min to 1 h per plane, depending on the size of the raster.

NanoSIMS data processing

All data processing and analysis were done in MATLAB. NanoSIMS.im data files were initially processed using the Look@NanoSIMS Matlab GUI⁷² to align planes and export raw data. To align isotopic enrichment data of nanoSIMS with the taxonomic information in the space resolution, the processed nanoSIMS raw ¹⁴N¹²C^– and ¹⁵N¹²C^– images were used for the region of interest (ROI) drawing by hand, combining with FISH images as reference. This was performed in the software Fresco (Adobe) on an iPad tablet. Regions of acquisitions that contained ANME-SRB cells were outlined based on the manual cell ROIs, and the elemental information was assigned to each ROI, which generated the single-cell level isotopic enrichment of ¹⁵F = ¹⁵N¹²C^–/(¹⁵N¹²C^– + ¹⁴N¹²C^–). All representative nanoSIMS images were illustrated by Limage PV-WAVE (v9.00).

DNA extraction and 16S rRNA iTag community analysis

DNA was extracted from the sediments stored after 4-month incubation using DNeasy PowerSoil Kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions. Illumina iTag 16S rRNA gene sequencing protocol was followed³⁸. Briefly, PCR amplification, barcoding, and sequencing of the 16S rRNA hypervariable region (V4-V5) were performed using PCR primers 515f and 906r and amplification conditions already described⁷³. Sequencing data were processed using QIIME v1.8.0⁷⁴, clustered at 99% sequencing identity using UCLUST v7.0.1001⁷⁵, and the taxonomic identity of the most abundant sequence in each cluster was assigned using a custom SILVA database modified from SILVA Ref NR 99 Database Release 115^76,77. iTag community analysis was performed using dada2 and R.

RNA extraction and metatranscriptomic analysis

To obtain fresh material for RNA analysis, incubations using sediment aliquots from PC6 were developed in parallel following the same protocol as mentioned above (section Microcosm setup). For RNA analysis, electron donors included 1.5 bar methane, 0.1 bar CO, and 1.0 bar CO. Each condition was run in biological triplicate. Sediment samples were taken on day 6 after the start of the microcosm experiment (~3 g wet weight), immediately flash frozen in liquid nitrogen, and extracted for total RNA the same day.

RNA was extracted using RNA PowerSoil Total RNA Isolation Kit (Qiagen, Valencia, CA, USA). TURBO DNA-free Kit (ThermoFisher) was used to remove genomic DNA, and purified using RNeasy Plus MicroKit (Qiagen). rRNA was removed using Pan-Pro (RNA-Seq) riboPOOL 24 reaction Kit (siTOOLs Biotech). Then, RNA was prepared for sequencing using Superscript IV VILO Master Mix (ThermoFisher/Invitrogen). The library was prepared with the Next® Ultra™ II Directional RNA Library Prep Kit (NEB), and sequencing was conducted on a NextSeq2000 (Illumina) platform in the Millard and Muriel Jacobs Genetics and Genomics Laboratory at the California Institute of Technology, generating 2 × 150 bp paired-end reads with an average insert length of 300 bp. A depth of 40 M reads per sample was sequenced.

Transcript abundance was quantified using kallisto⁷⁸ against a single reference database consisting of one ANME-2b MAG (Ga0402030_bin1) and one SRB1g MAG (Ga0402030_bin2) generated previously by a single-aggregate sequencing study^6,26. These MAGs were the best representative as they originate from an adjacent horizon of the same sediment core and are the best-quality pair of ANME-SRB MAGs (99.4% complete, 1.3% redundant, and 98.7% complete, 2.2% redundant, respectively). Counts were read into DESeq2⁷⁹ in R for normalization and analysis for differential abundance results presented. Percentile rankings of each gene were also converted from the kallisto TPM data on a per-sample basis. Focal genes were annotated with a combination of KEGG^80–82 and Pfam databases⁸³ and manual BLAST⁸⁴ to known sequences.

CO dehydrogenase homology searches and phylogeny analysis

A local database of genome sequences from ANME and syntrophic SRB was assembled from two previous studies^6,26 from our lab, representing all the genomes of ANME-SRB consortia available currently. CO dehydrogenase/acetyl-CoA synthase complex subunit alpha (cdhA, MA_RS05310) and anaerobic carbon monoxide dehydrogenase catalytic subunit (cooS, MA_RS06785) from Methanosarcina acetivorans C2A were used as templates to search for homologs in ANME genomes; anaerobic carbon monoxide dehydrogenase catalytic subunit (cooS, DVU_RS09895) from Desulfovibrio vulgaris Hildenborough was used to blast against all syntrophic SRB genomes for homologs. Protein sequences were aligned using MUSCLE⁸⁵, and then a phylogenetic tree was built using IQ-Tree2 using -m MFP with 1000 bootstraps⁸⁶. The tree was visualized using iTOL ⁸⁷.

Reporting summary

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

Author contributions

V.J.O. conceived and supervised the project; Y.G. and V.J.O designed the methodology; Y.G. performed the experiments; D.R.U. and Y.G. performed the metatranscriptomic data treatment and analysis; R.M., V.J.O., and Y.G. performed the metabolic analysis of ANME-2b and syntrophic SRB; Y.G. generated the original draft of the manuscript in consultation with V.J.O, and all authors provided advice and contributed to the final drafting of the manuscript.

Peer review

Peer review information

Nature Communications thanks Niculina Musat, Xiuran Yin, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The raw 16S rRNA amplicon sequencing and metatranscriptomic sequencing reads generated in this study have been deposited in the National Centre for Biotechnology Information database under BioProject ID PRJNA1227537. Database IDs for each MAG in this study are provided in the Supplementary Data 3.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-71433-9.