ANME-2a perform methane oxidation in brackish coastal sediments via multiple pathways

Abstract

Methane is a powerful greenhouse gas. Typically, a large fraction of methane formed in coastal sediments is removed via anaerobic oxidation of methane (AOM). Here, we demonstrate the potential for a range of AOM pathways in brackish coastal sediments by ANME-2a archaea. At our study site, geochemical profiles indicate that AOM is primarily restricted to a shallow, metal oxide–rich sulfate-methane transition zone (SMTZ). ANME-2a were the sole methanotrophs detected, and metatranscriptomics showed the highest expression levels of the ANME-2a genes in the SMTZ. AOM activity was observed in sediment incubations with various electron acceptors, including sulfate, metal oxides, and the organic matter analog graphene oxide. Highest potential rates were observed in sediments from below the SMTZ, pointing toward fast stimulation of the deeper methanotrophic community when alleviating the electron acceptor limitation. The variety of AOM pathways and persistence of methanotrophs below the SMTZ likely contribute to the resilience of the microbial methane filter in brackish coastal sediments.

INTRODUCTION

Methane is a key greenhouse gas that, in marine sediments, is mainly produced by methanogenic archaea (1). In most continental margin sediments, methane is efficiently removed through anaerobic and aerobic oxidation (2). In coastal sediments, however, some of methane may escape oxidation, especially when the zone of methanogenesis approaches the sediment-water interface (3, 4). As a consequence, coastal regions may contribute up to 75% of total marine methane emissions, even though these systems account for only a small part of the oceanic surface area (5). This makes it crucial to understand the microbial pathways contributing to methane removal in coastal sediments.

Coastal sediments typically receive high inputs of organic matter, which induce sharp redox gradients near the sediment-water interface (6). Besides oxygen, which generally penetrates down to depths of only a few millimeters, other electron acceptors used to degrade organic matter are, in order of decreasing energy yield, nitrate/nitrite, manganese (Mn) oxide, iron (Fe) oxide, and sulfate [SO₄²⁻; (7)]. When these electron acceptors are depleted, fermentation and methanogenesis are the remaining thermodynamically favorable pathways (8). Most of methane is thought to be removed through anaerobic oxidation of methane (AOM) with sulfate as the electron acceptor (S-AOM) in the so-called sulfate-methane transition zone (SMTZ) (1, 9, 10).

In environments subject to non–steady state deposition, methane may also come into contact with metal oxides, creating an opportunity for a coupling between Mn and Fe oxide reduction and AOM [Mn-AOM and Fe-AOM, respectively; [(11) and references therein]. The first evidence for potential Mn-AOM and Fe-AOM was obtained from incubations of marine methane-seep sediments using ¹³C-labeled methane (12). Follow-up experimental and modeling studies confirmed a potential role for Mn-AOM and/or Fe-AOM in brackish and marine sediments (13–18). Brackish sediments where an increased input of organic matter has led to an upward shift of the SMTZ are particularly conducive to Mn-AOM and Fe-AOM because of the presence of metal oxides in a zone with high methane concentrations (15, 17, 19). Besides direct coupling of metal oxide reduction to AOM in such settings, metal oxides may also drive S-AOM via the production of sulfate in a cryptic sulfur cycle (18, 20–22). Given that there is no conclusive geochemical signature for Mn-AOM, Fe-AOM, or cryptic sulfur cycling, the need to identify the microorganisms involved is important.

Typically, S-AOM is performed by a consortium of anaerobic methanotrophic archaea (ANME) and sulfate-reducing bacteria (SRB) (9, 10). The clade Ca. Methanoperedenaceae (also known as ANME-2d) has been linked to Mn-AOM and Fe-AOM reduction in freshwater sediments (23–25). In brackish and marine sediments, ANME-2d are typically only present at very low abundances and thus are unlikely to contribute to Mn-AOM and Fe-AOM (11, 16, 26, 27). Other subclades of ANME-2 may be involved instead, such as ANME-2ab, now classified as Ca. Methanocomedenaceae (28). ANME-2a were, for example, proposed to be linked to Fe-AOM on the basis of sediment incubations for a marine site in the North Sea (16). Similar depth trends in ANME-2a and total Fe contents (26) for sediments from the brackish Bothnian Sea, for which Fe-AOM activity was reported earlier (15), also point to a potential role for ANME-2a in Fe-AOM. Last, AOM may also be coupled to redox-active humic substances in natural organic matter (NOM-AOM) (29–31). While in freshwater sediments, ANME-2d can be involved in NOM-AOM (32); various other ANME clades have been inferred to carry out the process in brackish and marine environments, including ANME-1b [anoxic marine water; (33)], ANME-2a and ANME-2c [marine methane seep; (29)], and ANME-2ab [brackish canal sediment; (34)].

In this study, we examine the potential and in situ relevance of various AOM pathways and the microbes involved in methane- and metal oxide–rich sediments from a brackish coastal sea (Bothnian Sea). We focus on sediments from within and below the SMTZ and assess the role of the following electron acceptors in AOM: sulfate, Mn and Fe oxides, and natural organic matter. A complementary set of methods was used, including geochemical profiling of sediment and porewater, 16S ribosomal RNA (rRNA) gene amplicon sequencing, metagenomics, metatranscriptomics, and sediment incubations using ¹³C-labeled methane. We find evidence for a range of AOM pathways, performed by ANME2-a, with the highest in situ activity in the SMTZ but the greatest potential activity deeper in the sediment, indicative of a resilient and easily reinvigorated methanotrophic community.

RESULTS

Methane oxidation in a shallow SMTZ

A distinct SMTZ was identified between 12- and 22-cm depths in the sediment at our study site (US2), characterized by strong counter gradients of sulfate and methane in the porewater (Fig. 1). A small fraction of methane passed onward to the zone above the SMTZ. Methane isotopic values of δ¹³C and δD were stable below the SMTZ, at values around −79‰ and −265‰, respectively, and increased in the SMTZ to values of −61‰ and −93‰ (Fig. 1), respectively, indicating methane oxidation. Above the SMTZ, methane isotopic values were stable, apart from the uppermost 2 cm where both values increased.

Fig. 1. Porewater and sediment depth profiles and SRRs for our study site US2.

Sulfate (SO₄²⁻) and methane (CH₄) (A), δ¹³C-CH₄ and δD-CH₄ (B), SRRs (C), total S (D), Mn oxides (E), dissolved Mn (F), Fe oxides (G), dissolved Fe (H), C_org (I), NH₄⁺ (J), DIC and alkalinity (K), and δ¹³C-DIC (L). The zone indicated with gray shading indicates the SMTZ. The extraction step used to determine Mn oxides may include Mn phosphates. Additional sediment Mn and Fe forms are shown in fig. S4. The unit for the solid phase refers to the dry weight of the sediment. V-SMOW, Vienna standard mean ocean water; VPDB, Vienna Pee Dee belemnite; d, day.

Sulfate reduction rates (SRRs) were the highest just above the SMTZ at a depth of 9 cm, reaching values of up to 352 nmol cm⁻³ day⁻¹ (Fig. 1). A second maximum in SRRs was observed at the bottom of the SMTZ at a depth of 22 cm, reaching values of up to 157 nmol cm⁻³ day⁻¹. Free sulfide was always below the detection limit (1 μM) in the porewater. Below the SMTZ, solid-phase total S reached a maximum concentration of 342 μmol g⁻¹ dry sediment at the 27-cm depth. A substantial part of the peak in total S, up to 155 μmol g⁻¹ dry sediment, consisted of iron sulfide (FeS) (Fig. 1).

Surface sediments were enriched in Mn and Fe oxides, reaching concentrations of up to 106 and 379 μmol g⁻¹ dry sediment, respectively (Fig. 1). At the upper boundary of the SMTZ, a peak in Mn oxides was present, which may partly consist of Mn phosphates dissolved in the same extraction step as the Mn oxides (fig. S3 and table S1). Within the SMTZ, Mn oxide contents decreased to around 20 μmol g⁻¹ dry sediment, and Fe oxide contents decreased from 46 μmol g⁻¹ dry sediment to 0 μmol g⁻¹ dry sediment. Mn oxides remained around 20 μmol g⁻¹ dry sediment, but Fe oxides emerged again below the 29-cm depth, reaching concentrations of up to 29 μmol g⁻¹ dry sediment. Dissolved Mn and Fe increased rapidly in the upper few centimeters of the sediment to around 400 and 200 μM, respectively, and subsequently decreased but at different depths (Fig. 1). Deeper in the sediment, the concentrations of both Mn and Fe increased again, with dissolved Mn and Fe reaching maxima of 566 and 670 μM, respectively.

The C_org content was ∼2.5 wt % in the top 20 cm of the sediment and decreased to around 2 wt % below a depth of 20 cm (Fig. 1). Concentrations of ammonium, dissolved inorganic carbon (DIC), and alkalinity increased strongly in the top 20 cm of the sediment, indicating the highest rates of organic matter degradation in this zone (Fig. 1). Below 20 cm, ammonium continued to increase with depth, whereas DIC and alkalinity concentrations stabilized around 12 and 13 mM, respectively. Values of δ¹³C of DIC were close to −5.5‰ near the sediment water interface and decreased to a minimum of −15.4‰ within the SMTZ (Fig. 1). Below the SMTZ, the δ¹³C of DIC increased to 0.7‰. Porosity decreased from 0.95 at the sediment surface to ∼0.90 at the 5-cm depth and remained near this value at a greater depth (fig. S3).

Diversity of the microbial community

The 16S rRNA gene amplicon sequences revealed distinct communities of both archaea and bacteria, which in some cases were related to the boundaries of the SMTZ (Figs. 1 and 2). Ammonia-oxidizing archaea, unclassified Nitrosopumilaceae and “Candidatus Nitrosopumilus,” covered >65% of all archaeal reads in the top 12 cm. In the SMTZ (12 to 22 cm), ANME-2ab were the only methanotrophs detected at a maximum abundance of 31%. No ANME-2b [Candidatus (Ca.) Methanomarinus] were found in the metagenomic analysis (table S4); therefore, we will refer to ANME-2a in the following text. Unexpectedly, ANME-2a even accounted for up to 61% of all archaeal reads in the interval from 24 to 34 cm below the SMTZ. We detected four abundant ANME-2a amplicon sequence variants (ASVs) with slightly different niches, with ASV_1 dominating down to a depth of 34 cm, whereas the other ASVs were either more variable or restricted to the sediments below the SMTZ (fig. S5). Below 28 cm, methanogenic genera Methanosarcina, Methanoregula, and Methanosaeta, together with Bathyarchaeaia, appeared and became abundant from a depth of 34 cm onward.

Fig. 2. Results of 16S rRNA gene amplicon sequencing for sediments from our study site US2.

Relative abundances of archaeal DNA (A) and bacterial DNA (B). All taxa with <2% abundance are grouped in “Others.”

The bacterial community was highly diverse, with more than 50% of taxa having a relative abundance of less than 2%. Despite this diversity, there was a clear difference in the most abundant taxa with sediment depth. Above and in the SMTZ (0 to 22 cm), Nitrosomonadaceae, Nitrospiraceae, Flavobacteriaceae, Woeseiaceae, unclassified Gammaproteobacteria, and the unclassified phylum NB1-j made up about 25% of all bacterial reads. Between 18- and 34-cm depths, the SRB Desulfosarcinaceae was the most abundant bacterial group after the MBNT15 phylum, reaching up to 7% in relative abundance. Below 32 cm, bacterial diversity increased, with Anaerolineaceae and Solirubrobacterales being the most abundant taxa.

Metabolic potential for sulfur, C1, and metal cycling

We sequenced the metagenomes from two depths, 16 to 20 cm (SMTZ) and 28 to 32 cm (below the SMTZ), and in total, 479 medium-quality (>50% completeness and <10% contamination) metagenomically assembled genomes (MAGs) were obtained. These MAGs covered 38% (SMTZ) and 34% (below the SMTZ) of the total metagenomic reads; thus, a substantial part of the community was not assembled and thus not analyzed. On the basis of the mapping of metagenomic reads (Fig. 3), the MAG with the highest relative abundance at both depths was Defferimicrobium (MAG US2_17), which accounted for 4.4% of reads in the SMTZ and 2.0% below the SMTZ.

Fig. 3. Most abundant MAGs (>0.2% abundance at one depth) and their metabolic potential (cycling of S and C1; metal dynamics).

S cycling genes: dsrA, aprA, soxA, sqr, and asrC. C1 cycling genes: cdhA, cooF, mcrA, mch, mtaA, mtbA, and moxA. Genes involved in metal reduction/oxidation or EET: cyc1, cyc2, mtoA, mtrA, mtrC, omcZ, omcS, and omcF. The MAGs are named on the basis of the lowest taxonomic rank with a classified name. The relative abundance of each MAG at the two depths is indicated in the two columns before the genome name.

As seen from the 16S rRNA gene amplicon profile of the sediment, the most abundant taxa differed significantly between the two depths. In the SMTZ, Deferrimicrobiaceae (US2_15), Burkholderiales (US2_35), Solirubrobacterales, and Gammaproteobacteria (US2_34) were the only bins with >1% abundance. The sediment below the SMTZ was more diverse, and only Deferrimicrobium and Solirubrobacterales reached >1% abundance. Curiously, MAGs representing likely SRBs (phylum Desulfobacterota) were higher in abundance below than in the SMTZ (Fig. 3). We recovered one ANME-2a MAG (US2_17) from the newly categorized genus Ca. Methanoborealis (35) that, in line with 16S rRNA amplicon reads, was more abundant below the SMTZ (0.8%) when compared to the SMTZ (0.2%).

We screened the MAGs with >0.2% abundance at either depth for marker genes in S cycling, one-carbon (C1) cycling, and potential for extracellular electron transport (EET) or metal reduction/oxidation (Fig. 3). S cycling potential was widespread across the MAGs, with dissimilatory sulfate reduction genes dsrA and/or aprA found in multiple MAGs that were more abundant below the SMTZ, such as the two Desulfobacterales MAGs. However, the sulfur-oxidizing Sulfomarinibacterales MAG (US2_11) also encoded these genes and was more abundant in the SMTZ. In contrast, the sulfide:quinone oxidoreductase sqr was found in MAGs that were more abundant in the SMTZ and potentially used for sulfide detoxification. C1 cycle genes were scarce, indicating that the sediments harbor mainly heterotrophic clades. Only two MAGs with the marker gene for (reverse) methanogenesis, mcrA, were present: ANME-2a (Ca. Methanoborealis) and methanogenic Methanoregula (US2-13). Putative genes involved in EET and metal reduction/oxidation were found in multiple MAGs, which were often more abundant in the SMTZ (Fig. 3). Woeseiaceae MAG (US2_43) encoded multiple cytochrome c or membrane proteins involved in electron transfer across the membranes, i.e., RnfC from a complex that translocates proteins across the membrane and MtrA, MtrB, and MtrC from decaheme-cytochrome c that transfers electrons to the outer membrane. OmcS, OmcF, and OmcZ, outer membrane cytochromes linked to metal reduction via EET, were present in a few genomes such as Desulfobacterota MAGs, Desulfobacterales_UBA11574 that encoded OmcS, and the most abundant Deferrimicrobium that encoded OmcF. The ANME-2a genome also showed potential for EET and metal reduction as RnfC and OmcZ homologs were detected in their genome.

Metatranscriptomic analysis shows active ANME-2a in the SMTZ

To identify the metabolically active members of the community that contribute to biogeochemical cycling in the SMTZ (16- to 20-cm depth) and the ANME-rich sediment below the SMTZ (28 to 32 cm), metatranscriptomes were analyzed from three replicates for each of the two depths. The analysis of mapped coding DNA sequence (CDS) reads revealed that the most active taxa were notably different from the most abundant taxa identified in the metagenomic DNA data (Figs. 3 and 4), as has been observed in other studies [e.g., (36)]. The proportion of unmapped reads was even greater in the metatranscriptomic dataset than in the metagenomic dataset, as ∼71 to 75% of all transcripts did not map to the medium-quality MAGs analyzed. In the SMTZ, most transcripts mapped to Aminicenantia_bin2 (MAG US2_47), which covered 12 to 22% of all CDS reads, and Xanthomonadales_bin3 (US2_56) with 7 to 8% relative abundance. Three putative SRBs, Desulfobacterales_JAFDCJ01 (US2_46), Desulfobacterales_UBA11574 (US2_27), and Desulfurivibrionaceae (US2_55), covered ∼6% of the reads in the SMTZ, although they were either not in the top taxa at all or more abundant below the SMTZ in the metagenomic reads (fig. S6). The Ca. Methanoborealis (ANME-2a) MAG had higher expression levels in the SMTZ than below the SMTZ with >6 and <1% relative abundance, respectively (fig. S7).

Fig. 4. Relative abundance of expressed CDS reads of top MAGs at both depths [SMTZ: 16 to 20 cm; below the SMTZ in the methanogenesis zone (MZ): 28 to 32 cm].

Three replicates from each depth were sequenced for total RNA. The relative abundance here represents the percentage of total counts of coding genes per MAG detected in the transcriptome of each depth.

The active community below the SMTZ differed greatly from that in the SMTZ (Fig. 4). In general, archaeal MAGs were much more abundant in the deeper layer, such as Bathyarchaeia and Thermoplasmata bins. However, almost a third of the transcripts (∼20%) were covered by two Zixibacteria MAGs, which were barely present in the SMTZ (0.001 to 0.2%).

Increased methane oxidation upon addition of electron acceptors

The potential of methane oxidation coupled to oxygen, sulfate, Mn and Fe oxides, and graphene oxide, as an analog for natural organic matter, was tested in sediment both from within the SMTZ (16 to 20 cm) and below the SMTZ (28 to 32 cm; Fig. 5). All incubations supplied with an extra electron acceptor showed an increase in ¹³C-CO₂/¹²C-CO₂ ratio relative to the control where no electron acceptor was added, indicating active methane oxidation. In general, the signal for methane oxidation was the largest in the sediment from below the SMTZ (Fig. 5 and fig. S8). Oxygen was the most effective electron acceptor added, inducing a methane oxidation rate of around 216 μmol cm⁻³ year⁻¹ at both depths (Fig. 5 and fig. S8). After oxygen, the highest methane oxidation rates were obtained with sulfate as the electron acceptor with 28.6 ± 0.4 and 8.0 ± 0.8 μmol cm⁻³ year⁻¹ for sediments within and below the SMTZ, respectively. The methane oxidation rates in the Mn-AOM, Fe-AOM, and NOM-AOM incubations were within a similar range, varying from 1.2 to 3.3 μmol cm⁻³ year⁻¹ for sediments in the SMTZ and from 9.4 to 15.5 μmol cm⁻³ year⁻¹ for those below the SMTZ.

Fig. 5. ¹³C-CO₂/¹²C-CO₂ ratios in the headspace of the incubation experiments for sediments from within the SMTZ (16 to 20 cm; circles) and from below the SMTZ (28 to 32 cm; squares).

Control (A), oxygen (B), sulfate (C), manganese (D), iron (E), and natural organic matter (F). The methane oxidation rates are based on the increase in ⁴⁵C-CO₂ in the headspace and liquid of the incubations (fig. S8). Note the different y axes at the plot showing the O₂-MO incubations. yr, year.

Accumulation of sulfate occurred in the O₂-MO (max. 2.7 mM within the SMTZ and 1.1 mM below the SMTZ), Mn-AOM (max. 1 mM within the SMTZ and 0.3 mM below the SMTZ), and NOM-AOM (max. 0.3 mM within the SMTZ and 0.5 mM below the SMTZ) incubations (fig. S9). In the controls and Fe-AOM incubations, no sulfate accumulated. Accumulation of sulfide was only observed in the S-AOM incubation from below the SMTZ (fig. S9).

In the incubations where Mn oxide was added, dissolved Mn concentrations increased up to around 2 mM within the first 30 days and subsequently stabilized for sediments from both depths (fig. S9). Determination of the redox state of the dissolved Mn in incubations where Mn oxides were added showed that the accumulated dissolved Mn was almost exclusively Mn(II), indicating that reduction of Mn(IV) to Mn(II) occurs in one step and likely does not form dissolved Mn(III) as a reactive intermediate (fig. S9). In the incubations with oxygen as the electron acceptor, accumulation of dissolved Mn up to a concentration of 0.8 mM was also observed (fig. S9). The other incubations had relatively stable dissolved Mn concentrations varying between 0.2 and 0.4 mM.

In the incubations with Fe oxides, dissolved Fe increased from 0.5 to 0.9 mM in the SMTZ and from 0.2 to 0.5 mM in the incubation with sediment from below the SMTZ (fig. S9). When oxygen was added as an electron acceptor, dissolved Fe accumulated to values of up to 1.1 and 0.6 mM in the incubations with sediments from within and below the SMTZ, respectively. In the other incubations, dissolved Fe was relatively stable or absent, apart from the incubations with graphene oxide, where the dissolved Fe concentrations fluctuated around 0.2 mM.

ANME archaea perform methane oxidation in the AOM incubations

ANME-2ab were, on the basis of 16S rRNA analysis, the most abundant archaeal taxa in the incubations with the deeper sediment layer, covering ∼55% of the total archaeal reads (fig. S10). Given that all ANME in the metagenomic analysis of the source sediment belong to ANME-2a (table S4), we assume that all ANME in the incubation are ANME-2a. In the anoxic incubations, ANME reads increased substantially in the electron acceptor–amended samples but not in the control samples, underlining their role in AOM (Fig. 6A). Upon addition of Mn and Fe oxides, the relative archaeal abundance of ANME-2a increased from 55 to 75%. Methanosarcina was the second most abundant archaea in all samples and increased the most in the control and graphene oxide treatments (three- and twofold, respectively).

Fig. 6. Changes in the relative abundance of microbial taxa in the sediment from below the SMTZ (28 to 32 cm) after the incubation with different electron acceptors compared to that at the start.

Archaea (A) and bacteria (B). GO is used as an abbreviation for graphene oxide.

The 16S rRNA gene amplicon analysis revealed distinct differences between the microbial community of incubations with oxygen versus other electron acceptors (Fig. 6). With oxygen, the largest changes in the bacterial community were seen in Methylomonas and Methylotenera, which increased in relative abundance by 5 to 10%, indicating that dormant buried cells were activated upon oxygen amendment (Fig. 6B). In addition, potential sulfur-oxidizing taxa Sulfurimonas, Sulfurovum, and Magnetovibrio increased in these samples. The changes in the bacterial community composition were highly substrate-dependent, although Syntrophotalea reads, of phylum Desulfobacterota, increased in all samples from <0.01% by 7 to 18%, the biggest increase observed in the Fe-AOM and NOM-AOM treatments (Fig. 6B and fig. S11). Also, other Desulfobacterota taxa increased, with Trichloromas and MSBL7 increasing in all but the control samples, whereas changes in other genera were more substrate-specific. Curiously, SEEP-SRB1, a putative SRB partner for ANME, increased only with sulfate, whereas other putative AOM syntrophs, uncharacterized Desulfobulbales, increased in Mn-AOM and Fe-AOM samples. The changes in heterotrophic bacteria involved in organic matter breakdown via fermentation, such as Bacteroidetes, were also substrate-specific (Fig. 6B).

DISCUSSION

Two distinct zones of AOM

Our integrated analysis of sediments from the Bothnian Sea site reveals two vertically distinct zones of AOM, which are both characterized by the presence of a metabolically versatile community of ANME-2a archaea but differ in their porewater and sediment chemistry. The first is the canonical SMTZ (16 to 20 cm), where AOM is primarily coupled to sulfate reduction. Here, multiple indicators point toward AOM in situ. The second is a deeper, previously unrecognized zone below the SMTZ (18 to 32 cm), where ANME-2a persist and have the potential to couple AOM to a wide range of electron acceptors, although their in situ activity appears to be limited. This finding is supported by a combination of geochemical profiling, DNA and RNA sequencing, and sediment incubation experiments, as discussed in detail below.

Active AOM in the SMTZ, which is located between depths of 12 and 22 cm (Fig. 1), is evident from the strong isotopic enrichment of ¹³C and D in the residual methane and depletion of ¹³C in DIC (37). This is corroborated by a peak in sulfate reduction and change in the gradient in porewater sulfate at the lower boundary of this zone, which is a typical feature in sediments where S-AOM occurs (38). Molecular analysis confirms these observations, with a high relative abundance of ANME-2a archaea (a maximum of 31% of archaeal sequences) co-occurring with putative partner SRB from the Desulfosarcinaceae family (Fig. 2) (39, 40). Furthermore, metatranscriptomic data reveal that the Ca. Methanoborealis gene transcripts are highly expressed within the SMTZ, accounting for more than 6% of mapped transcripts. Besides sulfate, the SMTZ at our site also contains Mn and Fe oxides and abundant organic matter (Fig. 1), which may all act as potential electron acceptors in AOM (Fig. 5). The addition of sulfate, Mn oxides, Fe oxides, and graphene oxide to sediment incubations all stimulated AOM at rates comparable to previous incubation studies with sulfate [range of 2 to 52 μmol CH₄ cm⁻³_sed year⁻¹; (12, 13, 16)], Fe oxides [range of 0.035 to 6 μmol CH₄ cm⁻³_sed year⁻¹; (12, 13,15, 16, 41)], and Mn oxides [range of 0.20 to 14 μmol CH₄ cm⁻³_sed year⁻¹; (12, 13, 41)]. However, the accumulation of sulfate in the incubations amended with Mn and graphene oxide makes it difficult to distinguish direct metal-AOM from AOM coupled to a cryptic sulfur cycle in these treatments [e.g., (20, 22)]. Together, these results provide strong indications that ANME-2a are the primary microorganisms actively mediating AOM in the SMTZ, with S-AOM as the dominant pathway.

Unexpectedly, the relative abundance of ANME-2a below the SMTZ (28- to 32-cm depth) is high when compared to the sediment within the SMTZ (16- to 20-cm depth; Fig. 2 and fig. S6). The sediment below the SMTZ is depleted in sulfate but contains solid-phase Mn and, in a few depth intervals, Fe oxides that have been buried below the current SMTZ. Metatranscriptomics show the activity of ANME-2a in this sediment zone (Fig. 4), but the methane isotope profile does not show a clear signal for oxidation in this deeper layer (Fig. 1). Although, thermodynamically, AOM and methanogenesis cannot coexist, small-scale spatial variations within sediment layers may facilitate both processes in close vicinity (42). Previous work has suggested that methanogenesis could mask AOM signals in methane isotope profiles under such conditions (43, 44). Methanogens appear in the 16S rRNA profiles at a depth of 28 cm, and metatranscriptomics indicate that they are active in the zone from the 28- to 32-cm depth (Figs. 2 and 4), which makes it likely that methanogenesis is occurring at these depths. The activity of ANME-2a in combination with active methanogens as reflected in metatranscriptomics points toward the occurrence of both methanogenesis and methanotrophy in the same sediment layer below the SMTZ, even though this process is not visible in the isotopic CH₄ profiles.

Insight in the metabolic potential of this deeper community is obtained from our sediment incubation experiments. The addition of sulfate, Mn oxides, Fe oxides, and graphene oxides all stimulated AOM, with consistently higher rates compared to the sediment within the SMTZ (Fig. 5), possibly linked to a higher initial ANME biomass (Fig. 2 and fig. S6). In every anaerobic incubation for the sediment interval from 28 to 32 cm, the relative abundance of ANME-2a increased significantly, particularly in the presence of metal oxides, where they lastly constituted up to 75% of the archaeal community (Fig. 6). The sediment incubations indicate that ANME-2a are capable of oxidizing methane using sulfate and the alternative electron acceptors tested. The lack of sulfate accumulation in the Fe-amended incubations could indicate a direct coupling between AOM and Fe reduction. However, a cryptic sulfur cycle cannot be fully excluded in any of the incubations. Metagenomic analyses on the genome of ANME-2a (Ca. Methanoborealis) show a high potential for metal-AOM (35). Direct metal-AOM by ANME-2a is also supported by our genomic data, which show that the ANME-2a MAG recovered from this site encodes a homolog for OmcZ (Fig. 3), a multiheme cytochrome known to be involved in extracellular electron transfer to metals in bacteria (45), although its function in ANME is not confirmed. Notably, the ANME-2a in this deeper community were detectable through metatranscriptomics (fig. S7), although their activity was lower compared to the community within the SMTZ. This strongly suggests that this deep ANME-2a population is not dormant.

Environmental relevance

The existence of a deep, active ANME-2a community may be explained by the recent environmental history of the Bothnian Sea region (15, 19, 46). Geochemical evidence suggests that at this site, the lower boundary of the SMTZ has become shallower over the past decades. This is based on a previously observed depth of ∼28 cm for the lower boundary of the SMTZ at this site (47) and on the mismatch between the current position of the SMTZ and the FeS and total S peak in the sediment (Fig. 1). In a setting where sulfide is largely restricted to the SMTZ or, as is the case here, is absent from the porewater, such solid-phase sulfur peaks are expected within the SMTZ and not below (14, 15). The upward shift of the SMTZ at our study site has left a large population of ANME-2a stranded in a now sulfate-depleted environment. Our data suggest that this community has remained active using one or more solid-phase electron acceptors, possibly partly through a cryptic sulfur cycle, thereby establishing a microbial filter for methane that extends below the SMTZ. Such a low rate of AOM fueled by other electron acceptors than sulfate could also explain the presence of ANME-2a below the SMTZ at sites in the North Sea (16, 48) and Stockholm Archipelago (49). In conclusion, microbial methane removal below the SMTZ could be a more general feature in coastal sediments than previously recognized.

Our findings have significant implications for our understanding of coastal methane cycling. Eutrophication is expected to continue to increase globally (50) and typically leads to a shallowing of the SMTZ (4, 15). Variations in river runoff resulting from climate change may lead to a fluctuating supply of organic matter and metal oxides to coastal systems, especially at high latitudes (46, 51). Combined eutrophication and changes in river runoff may lead to a large temporal variability in the depth of the SMTZ and the availability of metal oxides, and potentially, humic substances, to methanotrophs. The versatility of ANME-2a in using these electron acceptors is expected to contribute to a more resilient methane filter in coastal brackish sediments in two ways. First, removal of methane by ANME below the SMTZ via AOM pathways other than S-AOM could be quantitatively relevant, given the relatively large overlap of methane with electron acceptors like Fe and Mn oxides at many coastal locations (15, 16, 17, 52). Second, a downward shift of the SMTZ driven by a decrease in rates of organic matter degradation, as predicted by reactive transport models upon a decrease in organic matter supply (19), could directly enable S-AOM by the ANME-2a community deeper in the sediment. Methanotrophic archaea have notoriously low growth rates (53), and S-AOM is thus often limited by ANME biomass (17, 54). Hence, a rapid start of S-AOM by ANME-2a upon a downward shift of the SMTZ could increase the efficiency of the methane filter.

ANME-2a are widespread and, apart from Bothnian Sea sediments (this study), have been observed in sediments from the continental margin of Gabon (Atlantic) (55) and the Mediterranean Sea (56), besides the North Sea (16, 27, 35, 48) and wider Baltic Sea (48, 49). Further research is needed on the occurrence of ANME-2a and other ANME types in relation to environmental conditions. The persistence and versatility of the ANME community in using a range of electron acceptors in AOM should be incorporated in computer models designed to predict methane dynamics in coastal sediments (17, 57). Future research should also focus on quantifying the in situ rates of Mn-, Fe-, and NOM-driven AOM in brackish coastal sediments and on determining the global distribution of such AOM systems to refine estimates and models of the marine methane budget in a changing world (11, 58).

MATERIALS AND METHODS

Study site

The Bothnian Sea is a brackish oligotrophic coastal basin in the northern part of the Baltic Sea. Rivers are the main source of organic matter in the region. Inputs of this terrestrial organic matter vary greatly between years, linked to variations in rainfall and river discharge, and appear to be increasing, linked to climate change (46, 59). Land uplift in the Bothnian Sea region contributes to the transport of this material from shallow areas to the deeper basins (60). The input of marine organic matter has also gained importance over the past decades as a result of anthropogenic eutrophication (61). The increase in deposition of both terrestrial organic matter and marine organic matter is thought to be responsible for the observed shallowing of the SMTZ at many locations in the Bothnian Sea (15, 19, 47). Notably, variations in organic matter input related to river discharge may also lead to a deepening of the SMTZ (46). Bothnian Sea sediments are metal oxide–rich, with frequent burial of metal oxides into and below the SMTZ (15, 46, 47).

The sampling location, US2 (62°50.99′N; 18°53.53′E; fig. S1), has a permanently oxygenated water column and a sediment accumulation rate of 1.6 cm year⁻¹ and is characterized by a high input of metal oxides and a shallow SMTZ (47). These characteristics make it representative for many locations in the Bothnian Sea (15, 19, 46, 47). Macrofaunal activity, mainly by burrowing crustaceans and polychaetes, affects the sediment via bioturbation and bioirrigation (62).

Sediment and porewater collection

In May 2022, six sediment cores were collected on board research vessel KBV 181 using a Gemini corer system with transparent core liners of 8 cm in inner diameter: one core for porewater and solid phase analysis, one for methane, one for SRRs, one for porosity and DNA/RNA samples, and two to collect material for incubation experiments.

Two bottom water samples were taken from the core used for porewater and solid-phase analysis using a two-way valve syringe. The core was then sectioned under a N₂ atmosphere in a glove bag at a 1-cm depth resolution until the 10-cm depth and a 2-cm depth resolution below the 10-cm depth. The sediment was transferred into 50-ml centrifuge tubes and centrifuged at 4500 rpm for 20 min to separate the sediment from the porewater. The sediment was stored in N₂-purged aluminum bags at −20°C until further processing for solid-phase analysis. The supernatant was filtered (0.45 μm) in a N₂-filled glove bag and subsampled for analysis of alkalinity, sulfate, ammonium (NH₄⁺), dissolved Mn, Fe, phosphorus (P), sulfide (H₂S), and DIC, including the DIC C-isotopic signature. Aliquots for alkalinity and sulfate were stored in polyethylene vials at 4°C until analysis. Samples for ammonium analysis were stored at −20°C. Samples for total dissolved Mn, Fe, and P were acidified with 10 μl of 30% suprapur HCl per milliliter of sample and stored at 4°C. Samples for sulfide analysis were diluted five times in 2% Zn acetate in a glass vial and stored at 4°C. Samples for DIC and DIC C-isotopic signal were stored in airtight glass vials without a headspace and with 10 μl of HgCl₂ per 2 ml of sample at 4°C.

Samples for the determination of methane, including C- and D-isotopes, and SRRs were taken directly upon recovery from a core liner with predrilled holes at a 2.5-cm depth resolution covered with tape before coring. For methane, 10 ml of sediment was transferred directly into a 65-ml glass bottle filled with a saturated NaCl solution using precut plastic syringes. Bottles were immediately stoppered, capped, and stored upside down until analysis. For SRRs, 5 ml of sediment was sampled using precut syringes, which were sealed off with parafilm and stored in a N₂-flushed aluminum bag at 4°C.

For porosity analysis and DNA/RNA sample collection, a core was sectioned under atmospheric conditions in sections of 1 cm in the top 10 cm of the core and in sections of 2 cm below 10 cm. Part of the sediment was placed in preweighed centrifuge tubes and stored at 4°C until porosity analysis. Another part was conveyed to autoclaved Eppendorf tubes, frozen with liquid N₂, and stored at −20°C for DNA analysis and at −80°C for RNA analysis.

The remaining cores were stored at 4°C and sectioned within 2 weeks after retrieval under a N₂ atmosphere at a 4-cm resolution. The sediment was placed in plastic bags and stored in N₂-flushed aluminum bags at 4°C until use in sediment incubations.

Chemical analysis of solid phase

Sample residues for the solid-phase analysis were freeze dried and ground with an agate mortar and pestle, all under a N₂ atmosphere. Aliquots of the sediment (∼300 mg) were used to determine C_org contents. The sediment was decalcified with 1 M HCl [via a two-step wash; (63)], dried, weighed, and powdered in an agate mortar and pestle. The powdered sediment was analyzed using an elemental analyzer (Fison Instruments model NA 1500 NCS). The C content was corrected for weight loss during decalcification, and an internationally certified soil standard IVA2 was used to determine the accuracy and precision of the analysis. The certified C content in IVA2 was 0.732 wt %, and the measured mean value for C in IVA2 (n = 7) was 0.716 wt % with a standard deviation of 0.004 wt %.

Another aliquot (∼50 to 100 mg) was used to determine the mineral phases of Mn and Fe via a sequential extraction. For Mn, we used the extraction scheme as reported in (64), and for Fe, we used the procedure described in (65). Details are provided in tables S1 and S2. In the first step of the Fe extraction (1 M HCl), Fe(II) and Fe(III) were measured separately via the spectrophotometric method using 1,10-phenantroline (66) to distinguish Fe oxides from reduced Fe minerals such as FeS and Fe carbonate. In the other steps, extracted Mn and Fe were measured via inductively coupled plasma optical mission spectroscopy (ICP-OES; PerkinElmer Avio; detection limit: 0.1 μM for Fe and 0.02 μM for Mn). The standard deviation values over all extraction steps based on duplicates (n = 3) were 2.5% for Fe and 2.4% for Mn.

FeS concentrations were determined via the passive diffusion method as described in (67) on sediment aliquots of ∼300 mg. The standard deviation based on duplicates (n = 3) was 5.0%.

Total sediment Mn, Fe, P, and S were determined via sediment digestion of ∼100 mg of sediment in 2.5 ml of mixed acid (HNO₃:HClO⁻ = 2:3) and 2.5 ml of 40% HF at 90°C. After evaporation of the acid mixture, the residue was redissolved in 1 M HNO₃. The solution was analyzed using ICP-OES with the following recoveries: 106% for Mn, 107% for Fe, 104% for P, and 101% for S. The standard deviation (n = 2) values were 1% for Mn, 1.1% for Fe, 1.1% for P, and 2.1% for S. Porosity was determined on the basis of the weight loss of the sample after drying in an oven at 60°C.

Chemical analysis of porewater

Porewater Fe, Mn, and P were measured via ICP-OES (PerkinElmer Avio; detection limit: 0.1 μM for Fe, 0.02 μM for Mn, and 3.1 μM for P). A 10-ml N₂ headspace was injected into the methane bottles, and after an equilibration time of 7 days, the headspace was analyzed on a Thermo Finnigan Trace gas chromatograph (flame ionization detector; limit of detection: 0.02 μM). The isotopic composition of methane (δ¹³C-CH₄ and δD-CH₄) was analyzed by continuous flow isotope ratio mass spectrometry, as described in (68, 69). Porewater sulfate was determined via ion chromatography (IC; Metrohm 930 Compact IC Flex; detection limit: 10 μM). Porewater sulfide was measured spectrophotometrically using the phenylendiamine and ferric chloride method [(70); detection limit: 1 μM]. Alkalinity was measured within 24 hours after sample collection through titration with 0.01 M HCl. Porewater DIC including its isotopic composition was determined after acidification of the sample with 85% H₃PO₄ in an argon-purged exetainer and measurement of the headspace on a GasBench-Isotope Ratio Mass Spectrometer (GasBench II Delta V Advantage IRMA, Thermo Fisher Scientific). Concentrations of ammonium were determined via the indophenol blue spectrophotometric method (71).

Determination of SRRs

To determine SRRs, the samples were injected with 100 kBq ³⁵S-SO₄²⁻ 6 days after core retrieval, sealed again, and incubated for 24 hours in N₂-purged aluminum bags at 4°C. The sediment was then transferred into a 50-ml centrifuge tube containing 20 ml of oxygen-free 20% zinc acetate to precipitate any sulfide formed and terminate microbial activity (72, 73). The samples were stored at −20°C in aluminum nitrogen-purged bags. Upon analysis, the samples were washed twice with oxygen-free bottom water and centrifuged to remove porewater and unreacted ³⁵S-SO₄²⁻ (74). The reduced S was then extracted with an acidic chrome chloride solution for 48 hours (73) and captured in 20% zinc acetate via the passive diffusion method, as described in (67). The formed radioactive sulfide was determined by mixing 20% zinc acetate (1:2 vol:vol) with Ecoscint XR (NAT1396, Scientific Laboratory Supplies, UK) and analysis on an automatic triple-to-double coincidence ratio liquid scintillation counter (Hidex 600 SL, LabLogic Systems Limited, UK). SRRs were quantified by comparing the activity (decays per minute) of the radiolabeled total reduced inorganic sulfur (α_TRIS) to the total sulfate radiotracer (α_TOT), as described in (73)

$$\text{SRR}=\left[\mathrm{S}{\mathrm{O}}_4^{2-}\right]\ast \mathrm{\phi }\ast \frac{{\mathrm{\alpha }}_{\text{TRIS}}}{{\mathrm{\alpha }}_{\text{TOT}}}\ast \frac{1}{t}\ast 1.06$$

where φ is the measured porosity, t is the incubation time in days, and 1.06 is the correction factor for the expected isotopic fractionation (73, 75).

DNA extraction and 16S rRNA amplicon sequencing and analysis

DNA was isolated with the DNeasy PowerSoil Pro DNA isolation kit (Qiagen, Venlo, Netherlands) according to the manufacturer’s instructions after bead-beating the samples on a TissueLyser LT (Qiagen) for 10 min at 50 Hz. Samples from the incubations were pre-incubated at 65°C for 10 min in the kit’s C1 solution before bead-beating to increase archaeal DNA yields. 16S rRNA gene amplicon sequencing was performed on the Illumina MiSeq Next Generation Sequencing platform by Macrogen (Seoul, South Korea) using the Herculase II Fusion DNA Polymerase Nextera XT Index Kit V2, yielding 2 × 300–base pair (bp) paired-end reads. Archaeal primers used were Arch349F (5′-GYGCASCAGKCGMGAAW-3′) and Arch806R [5′-GGACTACVSGGGTATCTAAT-3′; (76)], and for bacteria, Bac341F [5′-CCTACGGGNGGCWGCAG-3′; (77)] and Bac806R [5′-GGACTACHVGGGTWTCTAAT-3′; (78)] were used. The optimal trimming parameters were determined with Figaro (79), and trimming and further processing were done using the DADA2 pipeline [version 1.8; (80)] in Rstudio (version 4.1.1; R Core Team, 2019), as explained previously (27). Raw reads are accessible on the National Center for Biotechnology Information (NCBI) website under the accession number PRJNA1306556.

Metagenomic sequencing and analysis

The DNA from sediment depths of 16 to 20 cm and 28 to 32 cm was used for short-read metagenomic sequencing using a TruSeq DNA PCR-free library with an insert size of 550 bp on the NovaSeq6000 (Illumina) platform by Macrogen (Seoul, South Korea), producing 2 × 150-bp paired-end reads [100 giga–base pairs (Gbp) per sample]. For long read sequencing with Nanopore, library preparation was done starting with 150 to 400 ng of DNA. The quality of the DNA was checked by agarose gel electrophoresis. DNA library construction was performed using the Native Barcoding Kit 24 V14 (SQK-NBD114-24), according to the manufacturer’s protocol (Oxford Nanopore Technologies, Oxford, UK). The library was loaded on a Flow Cell (R10.4.1) and sequenced using the MinION Mk1C device (Oxford Nanopore Technologies, Oxford, UK), according to the manufacturer’s instructions. The quality of the sequence reads was analyzed using fastqc (81). We used SingleM (82) to characterize the samples taxonomically. Quality control, trimming, assembly, and binning were done using the Aviary pipeline [version 0.11.0; (83)]. The reads were trimmed with fastp [version 0.24.0; (84)], assembled with metaSPAdes (version 4.0.0; (85)], mapped with CoverM [version 0.7.0; (86)], and binned with MetaBAT and MetaBAT2 [version 2.15; (87, 88)], Semibin (89), Rosella (version 0.5.5; https://rhysnewell.github.io/rosella/), and DASTool [version 1.1.2.; (90)]. The MAGs were taxonomically classified with GTDB-Tk release 220 [version 2.4.0; (91)], and the MAG completeness and contamination were estimated with CheckM2 [version 1.0.2; (86)]. We also used differential coverage binning (92) as described previously (27, 35) to recover one ANME MAG. Genomes were annotated with DRAM version 1.0 (93) and metascan (94). Genes involved in metal reduction were retrieved with FeGenie (95). Proteins were identified with prodigal [version 2.6.3; (96)] and used to search for proteins with heme-binding motifs (>3).

Metatranscriptomics

RNA was extracted from freeze-dried sediment from depths of 16 to 20 cm and 28 to 32 cm with the RNeasy PowerSoil Total RNA Kit (Qiagen, Hilden, Germany) using the protocol with the following modifications to increase the RNA yield: The input amount was increased to ∼0.5 g of sediment; in step 2, we added an extra 0.5 ml of diethyl pyrocarbonate–treated water (Invitrogen, Carlsbad, US) to increase the volume of the aquatic phase; in step 3, we added 4 ml of phenol/chloroform/isoamyl alcohol to counteract for the increased humic acids; in step 9, 1 ml of diethyl pyrocarbonate–treated water was added before incubating for 30 min at −20°C; and the final extraction was done with 25 μl of SR7 to increase the RNA concentration. DNA contamination was removed with the Invitrogen DNA removal kit (Thermo Fisher Scientific, US). The quality of the RNA was checked using the 2100 Bioanalyzer instrument (Agilent Technologies, US) and the Agilent RNA 6000 Nano Kit according to the manufacturer’s instructions. Metatranscriptomic sequencing was performed using a TruSeq stranded with NEB rRNA depletion kit (bacteria) (Illumina, US) on a NovaSeqX (Illumina) platform by Macrogen (Seoul, South Korea), generating 150-bp paired-end reads with ∼15-Gbp throughput per sample. For both depths, three biological replicates per depth from different extractions were sequenced. Raw sequences were quality trimmed, filtered for rRNA encoding reads, and mapped against the DRAM-generated scaffolds, and transcripts per million values were generated using transcriptm version 0.5 (https://github.com/sternp/transcriptm). Relative abundance was calculated using CoverM version 0.6.1. Further analysis, including data visualization, was done in RStudio version 4.4.0. Map reads corresponding to CDSs were converted into transcripts per million and pooled per genome to compare the relative abundance of MAGs expressed per sample.

Incubation experiment

The methane oxidation potential of various electron acceptors, namely oxygen, sulfate, Mn oxide, Fe oxide, and graphene oxide, as an analog for humic acid, was tested via batch incubations using sediments from within the SMTZ (16 to 20 cm) and below the SMTZ (28 to 32 cm). For both intervals, methane and metal oxides overlapped in situ. Although oxygen is not expected at these depths, it was included to test whether potential microbes involved in aerobic methane oxidation were still present. The incubations were started within 6 months of sediment collection.

The sediment was diluted at a 1:4 ratio with Hepes-buffered artificial sulfate-free seawater (ASW; for the composition, see table S3) under an anoxic atmosphere. The sediment slurry was distributed over sterile 120-ml serum bottles to a final amount of 60 g of slurry per bottle. In the incubations with sulfate, 10 mM Na₂SO₄ solution was added. Manganese was added in the form of birnessite (7.5 mM; MnO₂·nH₂O) synthesized according to the protocol in (97), and iron was added in the form of ferrihydrite [7.5 mM; FeO(OH)] synthesized as described in (98). Both metal oxides were analyzed via x-ray diffraction to characterize the minerals (fig. S2). Graphene oxide (C_xH_yO_z) was added at a concentration of 200 mg liter⁻¹ after diluting graphene oxide paste (Merck KGaA, Darmstadt, Germany) in demineralized water and deoxygenating the solution. The bottles were closed with capped rubber stoppers, and in the incubation with oxygen, air was injected to achieve 4% oxygen in the headspace. The sediment was first pre-incubated for 1 week, with shaking at 90 rpm at room temperature with a 100% N₂ headspace, to oxidize sediment FeS and minimize the release of sulfate in a cryptic sulfur cycle in the real incubations. After 1 week, the ASW was replaced with freshly prepared ASW under an anoxic atmosphere, and new electron acceptors were added to each bottle at the same concentrations as in the pre-incubations. The headspace was replaced by a headspace of 76% N₂, 4% CO₂, and 20% ¹³C-labeled methane (Cambridge Isotope Laboratories Inc., Andover, US). The headspace in O₂-MO incubation was amended with 50% air, 4% CO₂, and 20% ¹³C-CH₄. The bottles were then placed at room temperature, with shaking at 90 rpm for 94 days. In the O₂-MO incubation, new oxygen was added in the form of pure oxygen gas to a concentration of 4% of the headspace after 28 and 49 days. In the NOM-AOM incubations with sediment from below the SMTZ, a new substrate was added to a concentration of 200 mg liter⁻¹ after 56 days when the activity stagnated likely because of substrate depletion.

The headspace was measured weekly for ¹²C-CO₂ and ¹³C-CO₂ concentrations on a gas chromatograph coupled to a mass spectrometer (GC-MS; Agilent 5975C inert MSD; the detection limit for accurate CO₂ isotope concentrations is 40 nmol CO₂) to monitor ¹³C-CH₄ oxidation. Upon addition of Mn oxides, the headspace concentration of CO₂ decreased, possibly because CO₂ was drawn into the dissolved phase upon an increase in the pH during Mn oxide reduction (99). To prevent CO₂ concentrations below the detection limit, new CO₂ (4% of the headspace) was added to Mn-AOM incubation with sediment from below the SMTZ after 49 days.

Every other week, a sample was taken from the medium with a N₂-purged syringe and needle through the stopper, which was used for the analysis of dissolved Mn and Fe (via ICP-OES; iCAP 6300 with detection limits of 0.02 and 0.18 μM for Mn and Fe, respectively), sulfate (via IC), sulfide, and dissolved Mn(II) and dissolved Mn(III) [spectrophotometrically, as described in (100, 101)]. Every 4 weeks, a sample was taken to determine dissolved CO₂, including the CO₂ isotopes. A liquid sample of 0.5 ml was injected into an argon-flushed 3-ml exetainer containing 1 ml of 85% phosphoric acid, and the headspace of the exetainer was injected into the GC-MS.

Samples for 16S rRNA amplicon sequencing were obtained at the start of the pre-incubations and from each bottle separately at the end of the incubation experiment of sediment from below the SMTZ (28 to 32 cm). These samples were processed as described in the “DNA extraction and 16S rRNA amplicon sequencing and analysis” section.

Supplementary Materials

The PDF file includes:

Figs. S1 to S11

Tables S1 to S4

Legend for data file S1

References

Other Supplementary Material for this manuscript includes the following:

Data file S1