Metal-Driven Anaerobic Oxidation of Methane as an Important Methane Sink in Methanic Cold Seep Sediments

Anaerobic oxidation of methane (AOM) coupled with reduction of metal oxides is supposed to be a globally important bioprocess in marine sediments. However, the responsible microorganisms and their contributions to methane budget are not clear in deep sea cold seep sediments. ABSTRACT Anaerobic oxidation of methane (AOM) coupled with reduction of metal oxides is supposed to be a globally important bioprocess in marine sediments. However, the responsible microorganisms and their contributions to methane budget are not clear in deep sea cold seep sediments. Here, we combined geochemistry, muti-omics, and numerical modeling to study metal-dependent AOM in methanic cold seep sediments in the northern continental slope of the South China Sea. Geochemical data based on methane concentrations, carbon stable isotope, solid-phase sediment analysis, and pore water measurements indicate the occurrence of anaerobic methane oxidation coupled to metal oxides reduction in the methanic zone. The 16S rRNA gene and transcript amplicons, along with metagenomic and metatranscriptomic data suggest that diverse anaerobic methanotrophic archaea (ANME) groups actively mediated methane oxidation in the methanic zone either independently or in syntrophy with, e.g., ETH-SRB1, as potential metal reducers. Modeling results suggest that the estimated rates of methane consumption via Fe-AOM and Mn-AOM were both 0.3 μmol cm−2 year−1, which account for ~3% of total CH4 removal in sediments. Overall, our results highlight metal-driven anaerobic oxidation of methane as an important methane sink in methanic cold seep sediments. IMPORTANCE Anaerobic oxidation of methane (AOM) coupled with reduction of metal oxides is supposed to be a globally important bioprocess in marine sediments. However, the responsible microorganisms and their contributions to methane budget are not clear in deep sea cold seep sediments. Our findings provide a comprehensive view of metal-dependent AOM in the methanic cold seep sediments and uncovered the potential mechanisms for involved microorganisms. High amounts of buried reactive Fe(III)/Mn(IV) minerals could be an important available electron acceptors for AOM. It is estimated that metal-AOM at least contributes 3% of total methane consumption from methanic sediments to the seep. Therefore, this research paper advances our understanding of the role of metal reduction to the global carbon cycle, especially the methane sink.

In the methanic zone, methane concentrations fluctuated between 53 mM and 920 mM, with a notable decrease from 781 mM at 210 cmbsf to 53 mM at 330 cmbsf ( Fig. 1a; Table S1). The stable carbon isotope ratios (d 13 C) of CH 4 became heavier from 268.77% to 264.33% when CH 4 decreased (Table S1), which is consistent with the preferential use of lighter isotopic values by microbes leading to residual inorganic carbon enriched in d 13 C-CH 4 (25). In accordance with this, the dissolved inorganic carbon (DIC) values increased from 250 cmbsf (10.38 mM) to 330 cmbsf (17.27 mM) (Fig. 1b), implying the formation of HCO 3 2 in the methanic zone (37), which can be also evidenced by the increase in total alkalinity (38) from 16.01 mM to 28.94 mM (Table S1). The measured d 13 C DIC values were maintained at lower than 242.30% (Fig. 1b) in the methanic zone and were much more 13 C-depleted than that of typical marine organic matters (approximately 220%) in this sea area (39). As microbes preferentially use the lighter carbon isotopes (d 12 C), AOM usually results in 13 C-depleted DIC and slightly heavier 13 C values of the residual CH 4 (25). In this cold seep site, the observed high concentrations of DIC, extreme 13 C-depletion d 13 C DIC , and heavier d 13 C values of residual CH 4 indicate that the DIC increase was probably caused by microbial methane oxidation T and MnO 2 T in sediments. (e to h) Sequential extraction of iron minerals in sediments. Fe carb , carbonate-associated Fe; Fe ox1 , amorphous iron (oxyhydr)oxides; Fe mag , magnetite Fe; Fe py , pyrite Fe. Zones A, B, and C are suggested as the sulfate reduction zone, the sulfatemethane transition zone, and the methanic zone.
rather than microbial degradation of other organic matter. Additionally, low concentration profiles of phosphate (PO 4 32 ) and ammonium (NH 4 1 ), lower than 41.65 mM and 56.72 mM, respectively (Table S1), also support that organic matter degradation was not the main reason for increased DIC concentrations in these sediment samples (40).
Methane oxidation is coupled to metal oxide reduction in the methanic zone. In the methanic zone, dissolved ferrous iron (Fe 21 ) and manganese (Mn 21 ) concentrations in pore water were found to reach up to 148 mM at 370 cmbsf and as high as 2,289 mM at 340 cmbsf, respectively ( Fig. 1c; Table S1). The Spearman correlation (Fig. 3a) results further show that Fe 21 Table S7 in the  Table S8 in the supplemental material). Therefore, these data implied sufficient supplies of reactive Fe-oxides and the occurrence of Fe authigenic minerals (carbonate Fe/Mn and magnetite) as the products of iron reduction (46,47). Similar to that of iron, total manganese (MnO 2 T ) is also elevated from 0.04% in the SMTZ to 0.06% in the methanic zone ( Fig. 1d; Table S7). Given the elevated MnO 2 T and extremely high dissolved Mn 21 (up to 2,289 mM), the contribution of manganese reduction to AOM cannot be ignored in this seep. Overall, porewater and solid-phase profiles support metal-driven methane oxidation in methanic sediments. Potential microorganisms involved in dissimilatory metal reduction. For indepth understanding of Fe(III)/Mn(IV)-dependent AOM in Haima cold seep, it is critical to identify the indigenous microorganisms responsible for this process. Members in different ANME clades are suggested to mediate metal-driven AOM by extracellular electron transfer (EET) to Mn(IV)/Fe(III) (oxyhydr)oxides or metal-reducing partners. In iron-reducer Geobacter sulfurreducens, the process of EET is carried out via MHCs during metal reduction (27,48). For ANME-2d from freshwater Fe-AOM enrichment, a set of MHCs for extracellular dissimilatory Fe(III) reduction were highly expressed (12,49,50). Here, all analyzed ANME genomes were found to contain the genes encoding several c-type and periplasmic cytochromes (see Table S9 in the supplemental material). Among all ANME genomes, three MAGs, S11_2_24, S11_5_37 and S11_6_25, belonging to ANME-2c, also encode S-layer-associated multiheme c-type cytochromes, implying a role of ANME-2c archaea with an S-layer protein in conducting electron derived from reverse methanogenesis shuttling from the archaeal membrane to the outside of the cell (50). S11_12_8 and S11_6_25 affiliated with the family ANME-2c also encode outer membrane cytochrome Z (omcZ) gene (see Table S10 in the supplemental material), which plays an important role in Fe(III) reduction (51). Furthermore, S11_12_8 not only actively expressed at zone C with the maximum of MAG's abundance and TPM values with mcrB gene but also had a significantly positive relation with CH 4 (r = 0.790; P , 0.01), Fe 21 (r = 0.734; P , 0.01), and Mn 21 (r = 0.601; P , 0.05) (Fig. 3b).
To identify potential dissimilatory metal reducers in the methanic zone, we performed sequencing of bacterial 16S rRNA gene and transcript amplicons, and the 16S  rRNA gene from metagenome ( Fig. S2 and Table S2). The typical partner SRB of ANME-1, i.e., family Desulfobacteraceae clustering into the SEEP-SRB1 (seep-endemic sulfatereducing bacteria) clade (52), were much more abundant in the methanic zone where sulfate was depleted (7 to 26% in DNA libraries, 9 to 42% in RNA libraries, and 7 to 35% in metagenomic libraries) than other zones ( Fig. 2a; Table S3). Additionally, 2 to 4% in DNA libraries, 2 to 12% in RNA libraries, and 1 to 4% in metagenomic libraries of bacterial sequences in the sulfate zone and SMTZ were identified as Desulfatiglans (family Desulfarculaceae), which is another common SRB associated with ANMEs in methane seep environments (38). Besides, the SEEP-SRB2 clade was detected with 2 to 4% in DNA libraries, ,0.7% in metagenomic libraries, but 22 to 33% in RNA sequence libraries in the SMTZ (Fig. 2a), implying the metabolic activity of SEEP-SRB2 involved in Sulfate-AOM (52). Therefore, according to 16S rRNA gene and transcript amplicons, and 16S rRNA metagenomes in sediment samples, members of the SEEP-SRB1 clade were the dominant and active bacteria in the methanic zone, potentially involved in dissimilatory metal reduction in situ.
Based on the metabolic pathways with metal reduction (Table S6), gene encoding metal (iron/manganese) reduction enzymes, such as decaheme c-type cytochrome (mtrC), were present in S11_6_22 and Co_S11_566 affiliated with ETH-SRB1 (ethane-dependent sulfate-reducing bacteria) from the order Desulfobacterales, which were identified as the marine SEEP-SRB1 group of Desulfosarcina-affiliated sulfate-reducing Deltaproteobacteria (53). The two MAGs (S11_6_22 and Co_S11_566) have the higher abundance in the methanic zone (mean, 0.20% and 0.12%) than the SMTZ (mean, 0.18% and 0.07%) (Table S4). Besides, Spearman's correlation results (Fig. 3b) show that Co_S11_566 closely related with concentrations of Fe 21 (r = 0. 699) and Mn 21 (r = 0. 650). Consistent with the results of microbial communities based on 16S rRNA gene and 16S rRNA transcripts, metagenomic and metatranscriptomic evidence indicated ETH-SRB1 (identified as the marine SEEP-SRB1 group) probably act as the role of metal reducing bacteria in the methanic zone. We also found the presence of hypothetical proteins attributed to porins, cytochrome c binding motif sites (CxxCH), and Geobacter-related gene markers (omc) for iron reduction in Co_S11_933 (Table S10), belonging to Zixibacteria, which was reported with pathways of either oxidation or reduction of ferric/ferrous iron and arsenate/arenite and nitrate/nitrite (54). Co_S11_933 also displayed a higher abundance in the methanic zone (0.03 to 0.06%) than in other zones (Table S4).
Contribution of metal-AOM to methane consumption. Geochemical observations and microbiological analyses support that Fe and Mn oxides reduction is coupled to methane oxidation in the methanic zone. We then used reactive transport numerical modeling to predict their contributions to methane consumption. Sensitivity tests of the model results suggest that the modeled profiles are insensitive to the changes of sedimentation rates (see Fig. S3 in the supplemental material). This is because the major porewater profiles in the methane seeps are controlled by methane supply and AOM rather than particulate organic carbon degradation. Constrained by the measured porewater data and Fe leaching experiments ( Fig. 1; Tables S1, S7, and S8), the results of the reaction-transport modeling predict the model-derived rates for Fe-AOM of up ;0.02 mmol CH 4 cm 23 year 21 in the methanic zone ( Fig. 4 and Table 1). Our estimated Fe-AOM rate is lower than those derived from stimulated microbial communities in laboratory incubations with the sufficient supply of substrates (CH 4 and Fe oxides) (Table 1) (5,24,25). Despite that, it is more than 20 times as big as the estimated potential Fe-AOM rates by kinetic modeling from in situ marine methanic sediments with a much higher Fe 21 concentration (approximately 180 to 800 mM) (20-23) (Table 1) because of much higher total AOM rates related to intense methane bubbling in the Haima seep.
As Mn speciation data were not available, we used the diffusive Mn 21 flux calculated based on the quasilinear concentration gradient at the depth interval of ;250 cm and 350 cm to represent the rate of Mn-AOM. Based on the porewater profiles, our estimated diffusive flux for Mn 21 [19]. This is a minimum estimate as the potential Mn 21 consumption by authigenic minerals is not taken into account. Depth-integrated rates of Fe-AOM and Mn-AOM are both 0.3 mmol cm 22 year 21 in the methanic zone, which are considerably lower than the S-AOM rate (;20 mmol cm 22 year 21 ) and account for ;1.5% of total CH 4 removal by microbial metabolism, respectively ( Table 1). The high S-AOM rate is caused by methane bubble dissolution, while its upward-ascending and enhanced sulfate supply from seawater is due to bubble irrigation. The estimated depth-integrated rate of Fe-AOM in the Haima seep falls within the range of those reported in different environments globally (20)(21)(22)(23). These data from the Haima cold seep provide the first in situ evidence for quantitatively significant manganese-dependent AOM in marine sediments. Given an apparent elevated sedimentary manganese content in the methanic zone (from 0.04% to 0.06%) and high concentration of dissolved Mn 21 (up to 2,289 mM), the contribution for Mn-AOM consumed by authigenic minerals could have been underestimated.
Conclusions. Methane oxidation occuring in the methanic zone driven by sedimentary microbial communities is an important mechanism that controls natural emissions of methane from the gas hydrate-bearing area. It happens mainly due to the presence of alternative electron acceptors other than sulfate to react with methane. Abundant Fe/Mn-(oxyhydr)oxides preserved in the shelf sediments might be migrated into the study region due to the rapid increase of anomalous subsidence toward the deep water areas in the Qiongdongnan basins (Fig. 5). Therefore, high amounts of buried reactive Fe(III)/Mn(IV) minerals seem to be important available electron acceptors for AOM in the methanic zone of the Haima methane seep, accompanied by the generation of highly alkaline, extremely d 13 C DIC -depleted and Fe(II)/Mn(II)-enriched pore waters, abundant Fe-Mn carbonates, along with authigenic magnetite by microbial iron/manganese reduction. In methanic sediments, abundant active ANME groups (ANME-1 and ANME-2c) and potential dissimilatory iron reducers (e.g., ETH-SRB1) are potentially involved in metal-AOM in situ. Mechanistically, the apparent ability of ANME-2c to oxidize methane via the release of single electrons in this study should also be able to respire solid electron acceptors directly via extracellular metal reduction, which would confirm the presence of previously reported methane oxidation coupled to insoluble Fe(III)/Mn(IV) reduction. It is estimated that metal-AOM at least contributes 3% CH 4 removal from methanic sediments to the seep. Overall, metal-AOM could significantly impact the biogeochemical cycles in consuming CH 4 in modern marine seep sediments.

MATERIALS AND METHODS
Sampling and geochemical analyses. Sediment core HM-S11 with a length of 430 cm was obtained by a gravity piston sampler at the ROV1 site of Haima cold seeps (29) (see Fig. S1 in the supplemental material) during the Haiyangdizhi10 cruise in June 2019 by the Guangzhou Marine Geological Survey, China Geological Survey.
The pore water was extracted from each sediment segment with an interval of 40 cm except the top 60 cm (20-cm interval) on board at room temperature by a vacuum apparatus. The concentrations of Fe 21 and Mn 21 in pore water were immediately on board determined by UV-visible (UV-Vis) spectrophotometer (Beijing Purkinje, China) using the 1,10-phenanthroline colorimetric method and potassium periodate oxidation spectrophotometry, respectively. SO 4 22 concentrations were measured by an ICS-1100 ion chromatography (Thermo Fisher, USA) with an analytical error of 61%. Concentrations of Ca 21 and Mg 21 in pore water were determined by the ICS-1100 ion chromatography with an analytical precision of ,10%. The concentrations of DIC and d 13 C DIC values in pore water were analyzed by a multiflow-isotope ratio mass spectrometer (Delta V Advantage; Thermo Fisher, USA), with an analytical error of 60.2%. Concentrations of PO 4 32 and NH 4 1 were photometrically measured on board using a UV-Vis spectrophotometer (Hitachi U5100; Hitachi Limited, Tokyo, Japan) with an analytical error of 63.0%. Porosity was determined from the weight loss before and after freeze-drying of the wet sediments (39) using a cutting ring with the volume of 5 mL on board, assuming a density of the porewater of 1.0 g cm 23 .
The C 1 ;C 3 concentrations of the headspace gas samples were determined using an Agilent 6850 gas chromatograph (Thermo, USA) with a Porapak Q column. The detection limit for all gases is 10 ppm, and the quantification limit is 30 ppm. The d 13 C values of the methane were measured using gas chromatography-isotope ratio-mass spectrometry (GC-IR-MS) (Thermo, USA) and are reported relative to the Vienna Peedee Belemnite standard (V-PDB), with an analytical error of 60.5%.
The major element composition of sediments was determined by an iCAP 7200 inductively coupled plasma (ICP) optical emission spectrometer (OES) (Thermo, USA). The contents of different iron phase mineral components were determined by Infinite M200Pro multifunction enzyme marker (TECAN, USA) with a sequential extraction method (55) and measured at the light absorption wavelength of 510 nm. The accuracy and repeatability of the absorption wavelength of the instrument are less than 60.5 nm (l . 315 nm).

FIG 5
Simplified scenario for how buried reactive Fe(III)/Mn(IV) minerals offer electron acceptors for AOM in methanic sediments under the high seepage flux of methane. At high seepage activities, methane gas and fluids move along migration pathways from deep sediments to the seabed in cold seeps with gas hydrate reservoirs. When high amounts of buried reactive Fe/Mn(oxyhydr)oxides from the slope sediments are exposed to methanic environments, diverse ANME groups actively mediated methane oxidation coupled to insoluble Fe(III)/Mn(IV) reduction either independently or in syntrophy with metal reducers. It results in authigenic Fe/Mn mineral (carbonate Fe/Mn and magnetite) precipitation and extremely d 13 C DIC -depleted and Fe(II)/Mn(II)-enriched pore waters in the methanic zone. EET, extracellular electron transfer.
DNA and RNA extraction. Total genomic DNA and RNA of each sample was extracted using a soil DNA kit and soil RNA minikit (Omega Bio-Tek Inc., Norcross, GA) according to the manufacturer's instructions, respectively. DNA concentration and purity were measured by TBS-380 (Turner Biosystems, CA, USA) and NanoDrop ND-2000 (Thermo Fisher Scientific, Waltham, USA), respectively. DNA extract quality, RNA degradation, and contamination were monitored on 1% agarose gels. RNA quantity was measured using Qubit 2.0 (Thermo Fisher Scientific, MA, USA) and NanoDrop One (Thermo Fisher Scientific, MA, USA) at the same time. RNA integrity was accurately detected using the Agilent 2100 system (Agilent Technologies, Waldbronn, Germany).
Amplicon analysis of 16S rRNA genes and transcripts. The DNA and RNA for each sample were amplified in triplicate using primers 338F/806R for bacteria and Arch344F/Arch915R for archaea. Their PCR products were pooled and purified. The same amount of the PCR product from each sample was mixed to construct a sequencing library. High-throughput sequencing was carried out on the Illumina MiSeq sequencing platform using the PE300 chemical at Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China).
After demultiplexing, the resulting sequences were merged with FLASH (v1.2.11) (56) and quality filtered with fastp (v0.19.6) (57). Then the high-quality sequences were denoised using the DADA2 (58) plugin in the Qiime2 (v2020.2) (59) pipeline with recommended parameters, which obtains single nucleotide resolution based on error profiles within samples. DADA2 denoised sequences are usually called amplicon sequence variants (ASVs). To minimize the effects of sequencing depth on alpha and beta diversity measure, the number of sequences from each sample was rarefied to 4000, which still yielded an average Good's coverage of 97.90%. Taxonomic assignment of ASVs was performed using the BLAST consensus taxonomy classifier implemented in Qiime2 and the SILVA 16S rRNA database (v138).
Metatranscriptomic sequencing. Whole mRNAseq libraries were generated by Guangdong Magigene Biotechnology Co., Ltd. (Guangzhou, China) using NEB Next Ultra Nondirectional RNA Library Prep Kit for Illumina (New England Biolabs, MA, USA) following manufacturer's recommendations. Briefly, the bacterial and archaeal 16S rRNA transcripts in total RNA samples were reduced by Ribo-zero rRNA removal kit. Fragmentation was carried out using NEB Next First Strand synthesis reaction buffer. The first strand cDNA was synthesized using random hexamer primer and M-MuLV reverse transcriptase (RNase H), and the second strand cDNA synthesis was subsequently performed using DNA polymerase I and RNase H. Remaining overhangs were converted into blunt ends via exonuclease/polymerase activities. After adenylation of 39 ends of DNA fragments, NEB Next Adaptor with hairpin loop structure were ligated to prepare for hybridization. In order to select cDNA fragments of preferentially 150 to 200 bp in length, the fragments were selected with AMPure XP beads (Beckman Coulter, Beverly, USA). Then, PCR was performed with Phusion high-fidelity DNA polymerase, Universal PCR primers, and Index (X) primer. At last, PCR products were purified with AMPure XP beads, and library insert size was assessed on the Agilent 2100 system (Agilent Technologies, Waldbronn, Germany). The clustering of the index-coded samples was performed on a cBot Cluster Generation System. After cluster generation, the library was sequenced on an Illumina HiSeq X Ten platform, and 150-bp paired-end reads were generated.
Numerical modeling. A one-dimensional, reaction-transport model (22,73) was applied to simulate two solid phases (reactive Fe oxides and Fe carbonate) and six dissolved species (SO 4 22 , CH 4 , DIC, Ca 21 , Mg 21 , and Fe 21 ) (see Tables S1 and S7 in the supplemental material). The reactions and their kinetic rate expressions considered in the model are listed in Table S11 in the supplemental material. Net reaction terms for all species and model parameters are listed in Tables S12 and S13 in the supplemental material, respectively. Detailed model construction can be found in the supplemental material.
Flux calculations. Diffusive fluxes J (mmol cm 22 year 21 ) of dissolved Mn 21 in the methanic zone were calculated using its pore water concentration gradients according to Fick's first law of diffusion. The algorithms are described in detail in the supplemental material.
Statistical analysis. The Spearman's linear correlation among the geochemical parameters and microbial abundances of the subsamples was analyzed using the R package Vegan (74)