Sulfide Toxicity as Key Control on Anaerobic Oxidation of Methane in Eutrophic Coastal Sediments

Coastal zones account for 75% of marine methane emissions, despite covering only 15% of the ocean surface area. In these ecosystems, the tight balance between methane production and oxidation in sediments prevents most methane from escaping into seawater. However, anthropogenic activities could disrupt this balance, leading to an increased methane escape from coastal sediments. To quantify and unravel potential mechanisms underlying this disruption, we used a suite of biogeochemical and microbiological analyses to investigate the impact of anthropogenically induced redox shifts on methane cycling in sediments from three sites with contrasting bottom water redox conditions (oxic-hypoxic-euxinic) in the eutrophic Stockholm Archipelago. Our results indicate that the methane production potential increased under hypoxia and euxinia, while anaerobic oxidation of methane was disrupted under euxinia. Experimental, genomic, and biogeochemical data suggest that the virtual disappearance of methane-oxidizing archaea at the euxinic site occurred due to sulfide toxicity. This could explain a near 7-fold increase in the extent of escape of benthic methane at the euxinic site relative to the hypoxic one. In conclusion, these insights reveal how the development of euxinia could disrupt the coastal methane biofilter, potentially leading to increased methane emissions from coastal zones.

presented in Figure 1 and Figure 2, 210 Pb data and sulfide exposure calculations.
Supplemental Table 4 (excel spreadsheet).List of MAG 011 ANME-2 genes and corresponding loci identified in this study.

High-resolution depth profiling
High-resolution depth (50 µm) profiles of oxygen and pH in the overlying water and surface sediment were obtained from intact sediment cores, using microelectrodes operated by a motorized micromanipulator (Unisense A.S., Denmark).Oxygen was measured immediately, and pH was measured within 1h of core retrieval.For the oxygen microelectrode, a two-point calibration was applied (100% oxygen saturated and nitrogen purged bottom water) using a calibration vessel (Unisense A.S., Denmark, CAL300).For the calibration of the pH microelectrode, three NBS standards (pH 4, 7 and 10) were used.Then, a TRIS buffer was applied for the pH to correct for signal drift induced by salinity effects 1,2 .

Collection of porewater and sediment samples for geochemical analysis
To determine porewater methane (CH 4 ) concentrations, samples of the bottom water and sediment were taken with cutoff 10 mL syringes via predrilled holes in the core liner with a depth spacing of 2.5 cm directly after coring.Subsequently, 10 mL of sample was transferred into a 65 mL serum bottle filled with saturated salt solution.Finally, the bottles were stoppered, capped and stored upside down until analysis.For further porewater analysis and collection of anoxic sediments, one core was sliced on board the ship in a N 2 -filled portable glove bag (Sekuroka 940x940x640mm; Glas-Col, USA).Prior to anoxic sectioning of the cores, two samples were taken from the overlying water using a 20 mL syringe equipped with a three-way valve.Subsequently, cores were sliced at a resolution of 0.5 cm (0-2 cm), 1 cm (2-10 cm), 2 cm (10-20 cm), 4 cm (20-40 cm) and 5 cm until the bottom of the core.Sediment was collected in 50 mL centrifuge tubes and centrifuged at 3500 rpm for 20 minutes to extract porewater.Bottom and porewater samples were filtered over 0.45 µm filters in a second N 2 -filled portable glove bag (Sekuroka 690x690x380mm; Glas-Col, USA).Subsamples were taken for (1) sulfide -0.5 mL of porewater was added to 2 mL 2% zinc acetate and stored at 4°C; (2) dissolved iron (Fe; assumed to be Fe 2+ ) and manganese (Mn) -1 mL of porewater was acidified with 10 μL 30% suprapur HCl and stored at 4°C; (3) sulfate -0.5 mL sample was collected and stored at 4°C; (4) ammonium -1 mL sample was collected; and (5) nitrate and nitrite -1 mL sample was collected, the latter two were stored at -20°C.The residual anoxic sediment after porewater subsampling was stored in N 2 -flushed gas-tight aluminum bags at -20°C until further analysis.
Samples from the core that was sliced under ambient atmospheric conditions were dried in an oven (~1 week at 60°C) to determine the sediment water content.The porosity of the sediment was then calculated assuming a solid phase density of 2.65 g cm −3 3,4 and a constant density with depth given the absence of lithological changes.

Pore water analysis
Pore water sulfide was determined spectrophotometrically using phenylenediamine and ferric chloride 5 .Dissolved Fe and Mn were measured by Inductively Coupled Plasma-Optimal Emission Spectroscopy (ICP-OES) with a Perkin Elmer Avio 500 (Perkin Elmer, USA).
Samples were measured with a radial plasma.Argon was used as the plasma gas (10 L min -1 ), nebulizer gas (0.75 L min -1 ) and auxiliary gas (0.2 L min -1 ).Nitrogen was used as the shear gas (25 L min -1 ).The rate of the peristaltic pump was 1.5 mL min -1 and the radio frequency generator was set at 1500 watt.Sulfate concentrations were determined by ion chromatography (IC) with a 930 Compact IC Flex (Metrohm, Switzerland), equipped with a Metrosep A Supp 5-150/4.0Guard column.Blanks and three different quality controls (QCs) were run five times each per series to monitor the detection limit and reproducibility.Average analytical uncertainty based on duplicates was <1%.Ammonium concentrations were determined colorimetrically using indophenol-blue 6 .Nitrate and nitrite were determined with a Gallery™ Automated Chemistry Analyzer type 861 (Thermo Fisher Scientific, USA) 7 in accordance with NEN-ISO 15923-1 guidelines (https://www.nen.nl/en/nen-iso-15923-1-2013-en-190502).

Approximated sedimentary sulfide exposure
An approximation of the exposure of the sediments, hence the microorganisms in it, to sulfide was calculated based on sedimentation rates and pore water sulfide depth profiles (Supplemental Table 3).We calculated the cumulative concentration of sulfide that the sediments are exposed to at 30 cm depth and the average sulfide exposure per year.
Sedimentation rates were calculated based on 210 Pb data (Supplemental Table 3) as previously described 8 .

Sediment analysis
Total organic carbon (TOC) was determined on a subsample of the oxic fraction as previously described 9 .Briefly, about 200 to 300 mg of sample was decalcified with 1M HCl, after which the residues were dried, powdered and weighed for analysis with a Fisons Instruments NA 1500 NCS analyzer (Carlo Erba, France).TOC was calculated after correction for the weight loss upon decalcification and the salt content of the freeze-dried sediment.
The total sedimentary concentrations of Fe and Mn were determined by digestion of ca.125 mg of the freeze-dried oxic subsample in a mixture of strong acids as previously described 8,10 and analyzed for their elemental composition by ICP-OES.The accuracy (recovery) for Fe and Mn was 91 and 93%, respectively.The average analytical uncertainty based on duplicates (n=14) was <2 % for both Fe and Mn.Two subsamples of the freeze-dried anoxic sample of about 50 mg each were subjected to two different sequential extraction procedures to determine the different solid phase forms of Fe and Mn, respectively.
For Fe, a previously described sequential extraction procedure 11  concentrations in all extraction steps was determined with the colorimetric phenanthroline method 12 .Average analytical uncertainty for all fractions, based on duplicates (n=12), was <9% for Fe.

DNA extractions, amplicon sequencing and 16S rRNA gene analyses
Sediments for DNA sequencing were immediately frozen at -20ºC after on board core slicing under nitrogen atmosphere and were stored at -20ºC for four months until thawing at room temperature for DNA extractions.DNA was extracted from 73 sediment samples retrieved from three cores in total, one core per site, with a depth resolution of 0.5 cm for the top 2 cm, 1 cm until 10 cm depth, 2 cm until 20 cm depth, and 4 cm below that.DNA extractions were performed with the DNeasy Power Soil Kit (Qiagen, Germany) according to the manufacturer's instructions with two modifications.For the bead-beating step, a TissueLyser LT (Qiagen, Germany) was used for 10 minutes at 50 Hz, and DNA was eluted in 30 µL of autoclaved ultrapure water (Milli-Q Reference Water Purification System, Merck & Co., USA).

Metagenomic sequencing and data analyses
DNA was extracted and quantified as abovementioned from four homogenized sediment samples from each of the three sites with a depth resolution of 4 cm: 0-4 cm, 9-12 cm, 21-24 cm, and 33-36 cm.These 12 samples were sequenced by Macrogen Europe BV (Amsterdam, Netherlands) using the TruSeq Nano DNA library with an insert size of 350bp on an Illumina NovaSeq6000 platform, producing 2x151bp paired-end reads (5 Gbp per sample).
MAGs were annotated with DRAM v1.0 28 with default options, except -min_contig_size 2500 bp for MAGs and 5000 bp for unbinned contigs, and genes of interest were searched in annotation files.Additionally, methyl-coenzyme M reductase alpha subunitencoding mcrA genes were also searched with the HMM PF02745.15MCR_alpha_N and iron metabolism genes were searched with FeGenie 29 v1.0.Only high and medium quality MAGs (>50% complete and less than 10% contaminated) were included in genome-centric analyses, and the entire dataset (binned and unbinned contigs) was considered in gene-centric analyses.
For this, genomes were gene called with Prodigal v2.6.3 34 , and amino acid fasta files were used as input.The abundance of MAGs was inferred from normalized genome coverage (ngCOV), that is, MAG coverage normalized to total metagenome size.Briefly, this was calculated as the total base pairs of summed metagenome reads that mapped to MAGs multiplied by 1 Gbp divided by genome length and total summed metagenome base pairs, as previously described 35 .Normalized mapped read (NMR) values for specific genes were calculated as follows: NMR = the number of mapped reads to the gene /(the length of the gene in bp*10 3 )*(the total number of reads (mapped + unmapped) in the metagenome/10 6 ).This is similar to values of reads per kilobase of gene per million mapped reads (RPKM), except that, instead of using only the number of metagenome mapped reads as denominator, NMR uses the total number of reads in the metagenome (the sum of mapped and unmapped reads), as in ngCOV, improving normalization for metagenome size and therefore cross-sample comparisons.
concentrations were proportionally adjusted to gas-phase methane concentrations.Gas-phase and liquid-phase methane amounts were summed to obtain the total amount of methane in each bottle for each time point.Potential rates of methane production were calculated using linear regression of methane measurements obtained in the first 8-9 days of sediment incubation.All calculations and measurements are provided as Supplemental Information.

Sulfide toxicity experiment and AOM rate measurements
Sediments for incubations to measure AOM rates under different sulfide concentrations were retrieved from previously stored, anoxically sealed aluminum bags kept in the dark at 4ºC for approximately two years after sample collection until bottles were assembled.For these incubations, Site 5 was selected due to highest coverages of the methanotroph genome (MAG 011 ANME-2), and samples were combined and homogenized, resulting in a mixture spanning the depth of 8 to 28 cm.Five grams of wet sediments were placed into 60 mL-serum bottles and mixed with 5 mL of a solution containing MgSO 4 8mM and 0.5% NaCl, to achieve a final concentration of 4 mM sulfate and a pH buffered to 7.05.Bottles were degassed with argon gas and a headspace at 0.5 bar of overpressure was created, containing also 0.5% of N 2 , 0.5% CO 2 ), and 20% 13 CH 4 .To one set of triplicate bottles, sulfide was added to a final concentration of 2 mM, while the other bottles did not received sulfide.After 13 weeks of incubation in the dark at 4ºC, AOM was confirmed by detection of 13 CO 2 production (data not shown).Then, sulfide was added to the remaining bottles, in triplicate incubations, targeting final concentrations of 0, 0.5, 1, 2 and 4 mM.To monitor 13 CO 2 production, bottle pressure was monitored using a GHM 3111 Digital Pressure Meter with a GMSD 2 BR -K31 sensor, and 50 µl of headspace was injected into an Agilent 6890 series gas chromatograph coupled to a mass spectrometer equipped with a Porapak Q column heated at 80°C with helium as the carrier gas as previously described 39 .Liquid-dissolved 13 CO 2 was estimated with the equation ∑ 13 CO 2 = 13 CO 2(g) [1 +   kRT V liquid /V gas (1 + K Z /[H + ])] 40 and summed to headspace 13 CO 2 derived from calibration curves for AOM rate calculations (Supplementary Table 4).Bottles were sampled for sulfide determination as described in the section collection of porewater and sediment samples for geochemical analysis.For that, approximately 1 mL of slurry was anoxically withdrawn and filtered through a 0.2µm Nylon syringe filter.Approximately 300 µL of filtrate was mixed with 1.2 mL of an anoxic 2% zinc acetate solution and stored at 4 ºC until analysis.Removed liquid volumes were taken into account for AOM rate calculations.Finally, sulfide concentrations were measured as described in the section "pore water analyses".

23 Supplementary Figure 1 .
Bottom water redox conditions, salinity and temperature from 1998 24 until in the three sites investigated in this study presented both as a summary of data as 25 well as for individual sites.Long-term seasonal monitoring data was obtained from the Swedish 26 Meteorological and Hydrological Institute -SMHI (https://sharkweb.smhi.se/hamta-data/).